River Restoration Managing the Uncertainty in Restoring ... - Inecol
River Restoration Managing the Uncertainty in Restoring ... - Inecol
River Restoration Managing the Uncertainty in Restoring ... - Inecol
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
<strong>River</strong> <strong>Restoration</strong><br />
<strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong><br />
Restor<strong>in</strong>g Physical Habitat<br />
Editors<br />
Stephen Darby<br />
School of Geography, University of Southampton, UK<br />
and<br />
David Sear<br />
School of Geography, University of Southampton, UK
<strong>River</strong> <strong>Restoration</strong>
<strong>River</strong> <strong>Restoration</strong><br />
<strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong><br />
Restor<strong>in</strong>g Physical Habitat<br />
Editors<br />
Stephen Darby<br />
School of Geography, University of Southampton, UK<br />
and<br />
David Sear<br />
School of Geography, University of Southampton, UK
Copyright © 2008 John Wiley & Sons Ltd, The Atrium, Sou<strong>the</strong>rn Gate, Chichester,<br />
West Sussex PO19 8SQ, England<br />
Telephone (+44) 1243 779777<br />
Email (for orders and customer service enquiries): cs-books@wiley.co.uk<br />
Visit our Home Page on www.wileyeurope.com or www.wiley.com<br />
All Rights Reserved. No part of this publication may be reproduced, stored <strong>in</strong> a retrieval system or transmitted <strong>in</strong> any form or by any<br />
means, electronic, mechanical, photocopy<strong>in</strong>g, record<strong>in</strong>g, scann<strong>in</strong>g or o<strong>the</strong>rwise, except under <strong>the</strong> terms of <strong>the</strong> Copyright, Designs and<br />
Patents Act 1988 or under <strong>the</strong> terms of a licence issued by <strong>the</strong> Copyright Licens<strong>in</strong>g Agency Ltd, 90 Tottenham Court Road, London W1T<br />
4LP, UK, without <strong>the</strong> permission <strong>in</strong> writ<strong>in</strong>g of <strong>the</strong> Publisher. Requests to <strong>the</strong> Publisher should be addressed to <strong>the</strong> Permissions Department,<br />
John Wiley & Sons Ltd, The Atrium, Sou<strong>the</strong>rn Gate, Chichester, West Sussex PO19 8SQ, England, or emailed to permreq@wiley.co.uk, or<br />
faxed to (+44) 1243 770620.<br />
Designations used by companies to dist<strong>in</strong>guish <strong>the</strong>ir products are often claimed as trademarks. All brand names and product names used <strong>in</strong><br />
this book are trade names, service marks, trademarks or registered trademarks of <strong>the</strong>ir respective owners. The Publisher is not associated<br />
with any product or vendor mentioned <strong>in</strong> this book.<br />
This publication is designed to provide accurate and authoritative <strong>in</strong>formation <strong>in</strong> regard to <strong>the</strong> subject matter covered. It is sold on <strong>the</strong><br />
understand<strong>in</strong>g that <strong>the</strong> Publisher is not engaged <strong>in</strong> render<strong>in</strong>g professional services. If professional advice or o<strong>the</strong>r expert assistance is<br />
required, <strong>the</strong> services of a competent professional should be sought.<br />
The Publisher and <strong>the</strong> Author make no representations or warranties with respect to <strong>the</strong> accuracy or completeness of <strong>the</strong> contents of this<br />
work and specifi cally disclaim all warranties, <strong>in</strong>clud<strong>in</strong>g without limitation any implied warranties of fi tness for a particular purpose. The<br />
advice and strategies conta<strong>in</strong>ed here<strong>in</strong> may not be suitable for every situation. In view of ongo<strong>in</strong>g research, equipment modifi cations,<br />
changes <strong>in</strong> governmental regulations, and <strong>the</strong> constant fl ow of <strong>in</strong>formation relat<strong>in</strong>g to <strong>the</strong> use of experimental reagents, equipment, and<br />
devices, <strong>the</strong> reader is urged to review and evaluate <strong>the</strong> <strong>in</strong>formation provided <strong>in</strong> <strong>the</strong> package <strong>in</strong>sert or <strong>in</strong>structions for each chemical, piece<br />
of equipment, reagent, or device for, among o<strong>the</strong>r th<strong>in</strong>gs, any changes <strong>in</strong> <strong>the</strong> <strong>in</strong>structions or <strong>in</strong>dication of usage and for added warn<strong>in</strong>gs and<br />
precautions. The fact that an organization or Website is referred to <strong>in</strong> this work as a citation and/or a potential source of fur<strong>the</strong>r <strong>in</strong>formation<br />
does not mean that <strong>the</strong> author or <strong>the</strong> publisher endorses <strong>the</strong> <strong>in</strong>formation <strong>the</strong> organization or Website may provide or recommendations it<br />
may make. Fur<strong>the</strong>r, readers should be aware that Internet Websites listed <strong>in</strong> this work may have changed or disappeared between when this<br />
work was written and when it is read. No warranty may be created or extended by any promotional statements for this work. Nei<strong>the</strong>r <strong>the</strong><br />
Publisher nor <strong>the</strong> Author shall be liable for any damages aris<strong>in</strong>g herefrom.<br />
O<strong>the</strong>r Wiley Editorial Offi ces<br />
John Wiley & Sons Inc., 111 <strong>River</strong> Street, Hoboken, NJ 07030, USA<br />
Jossey-Bass, 989 Market Street, San Francisco, CA 94103-1741, USA<br />
Wiley-VCH Verlag GmbH, Boschstr. 12, D-69469 We<strong>in</strong>heim, Germany<br />
John Wiley & Sons Australia Ltd, 42 McDougall Street, Milton, Queensland 4064, Australia<br />
John Wiley & Sons (Asia) Pte Ltd, 2 Clementi Loop #02-01, J<strong>in</strong> X<strong>in</strong>g Distripark, S<strong>in</strong>gapore 129809<br />
John Wiley & Sons Ltd, 6045 Freemont Blvd, Mississauga, Ontario L5R 4J3, Canada<br />
Wiley also publishes its books <strong>in</strong> a variety of electronic formats. Some content that appears <strong>in</strong> pr<strong>in</strong>t may not be available <strong>in</strong> electronic books.<br />
Library of Congress Catalog<strong>in</strong>g-<strong>in</strong>-Publication Data<br />
Darby, Stephen.<br />
<strong>River</strong> restoration : manag<strong>in</strong>g <strong>the</strong> uncerta<strong>in</strong>ty <strong>in</strong> restor<strong>in</strong>g physical habitat / Stephen Darby and David Sear.<br />
p. cm.<br />
Includes bibliographical references and <strong>in</strong>dex.<br />
ISBN 978-0-470-86706-8 (cloth)<br />
1. Stream restoration. 2. Riparian restoration. I. Sear, David (David A.) II. Title.<br />
TC409.D28 2008<br />
627′.12–dc22<br />
2007030210<br />
British Library Catalogu<strong>in</strong>g <strong>in</strong> Publication Data<br />
A catalogue record for this book is available from <strong>the</strong> British Library<br />
ISBN-13 978-0-470-86706-8 (HB)<br />
Typeset <strong>in</strong> 9/11 pt Times by SNP Best-set Typesetter Ltd., Hong Kong<br />
Pr<strong>in</strong>ted and bound <strong>in</strong> Great Brita<strong>in</strong> by Antony Rowe Ltd, Chippenham, Wiltshire
Contents<br />
Preface vii<br />
List of Contributors xi<br />
Section I Introduction: The Nature and Signifi cance of <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong><br />
<strong>River</strong> <strong>Restoration</strong><br />
1 <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> 3<br />
J. Lemons and R. Victor<br />
2 Sources of <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> Research 15<br />
W. L. Graf<br />
3 The Scope of Uncerta<strong>in</strong>ties <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> 21<br />
J.M. Wheaton, S.E. Darby and D.A. Sear<br />
Section II Plann<strong>in</strong>g and Design<strong>in</strong>g <strong>Restoration</strong> Projects<br />
4 Plann<strong>in</strong>g <strong>River</strong> <strong>Restoration</strong> Projects: Social and Cultural Dimensions 43<br />
G.M. Kondolf and C-N. Yang<br />
5 Conceptual and Ma<strong>the</strong>matical Modell<strong>in</strong>g <strong>in</strong> <strong>River</strong> <strong>Restoration</strong>: Do We Have<br />
Unreasonable Confi dence? 61<br />
M. Stewardson and I. Ru<strong>the</strong>rfurd<br />
6 <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Riparian and Floodpla<strong>in</strong> <strong>Restoration</strong> 79<br />
F.M.R. Hughes, T. Moss and K.S. Richards<br />
7 Hydrological and Hydraulic Aspects of <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 105<br />
N.J. Clifford, M.C. Acreman and D.J. Booker<br />
8 <strong>Uncerta<strong>in</strong>ty</strong> Surround<strong>in</strong>g <strong>the</strong> Ecological Targets and Response of <strong>River</strong> and Stream <strong>Restoration</strong> 139<br />
M.R. Perrow, E.R. Skeate, D. Leem<strong>in</strong>g, J. England and M.L. Toml<strong>in</strong>son<br />
Section III The Construction and Post-Construction Phases<br />
9 Construct<strong>in</strong>g <strong>Restoration</strong> Schemes: <strong>Uncerta<strong>in</strong>ty</strong>, Challenges and Opportunities 167<br />
J. Mant, R. Richardson and M. Janes
vi Contents<br />
10 Measures of Success: <strong>Uncerta<strong>in</strong>ty</strong> and Defi n<strong>in</strong>g <strong>the</strong> Outcomes of <strong>River</strong> <strong>Restoration</strong> Schemes 187<br />
K. Sk<strong>in</strong>ner, F.D. Shields, Jr and S. Harrison<br />
11 Methods for Evaluat<strong>in</strong>g <strong>the</strong> Geomorphological Performance of Naturalized <strong>River</strong>s:<br />
Examples from <strong>the</strong> Chicago Metropolitan Area 209<br />
B.L. Rhoads, M.H. Garcia, J. Rodriguez, F. Bombardelli, J. Abad and M. Daniels<br />
12 <strong>Uncerta<strong>in</strong>ty</strong> and <strong>the</strong> Management of <strong>Restoration</strong> Projects: The Construction and Early<br />
Post-Construction Phases 229<br />
A. Brookes and H. Dangerfi eld<br />
Section IV <strong>Uncerta<strong>in</strong>ty</strong> and Susta<strong>in</strong>ability: <strong>Restoration</strong> <strong>in</strong> <strong>the</strong> Long Term<br />
13 The Susta<strong>in</strong>ability of Restored <strong>River</strong>s: Catchment-Scale Perspectives on Long Term Response 253<br />
K.J. Gregory and P.W. Downs<br />
14 <strong>Uncerta<strong>in</strong>ty</strong> and The Susta<strong>in</strong>able Management of Restored <strong>River</strong>s 287<br />
M.D. Newson and M.J. Clark<br />
Index 303
For many years scientists and river practitioners have<br />
recognised <strong>the</strong> severity and extent to which aquatic ecosystems<br />
have been degraded by a variety of human disturbances<br />
and activities (Gregory and Park, 1974; Sear and<br />
Arnell, 2006). In turn, realisation of <strong>the</strong> widespread nature<br />
of <strong>the</strong> problem has more recently elicited a surge of <strong>in</strong>terest<br />
<strong>in</strong> <strong>the</strong> possibility of undertak<strong>in</strong>g corrective <strong>in</strong>terventions,<br />
such as fl ow restoration and channel modifi cations,<br />
to restore or rehabilitate lost and/or damaged ecosystem<br />
functions (Brookes and Shields, 1996; Wissmar and<br />
Bisson, 2003). Indeed, <strong>the</strong>re is now a substantial volume<br />
of literature on <strong>the</strong> broad topic of river restoration, much<br />
of which suggests that, to be susta<strong>in</strong>able, river restoration<br />
projects should be designed to recreate functional characteristics<br />
with<strong>in</strong> a context of physical (i.e. geomorphic)<br />
stability. It is true that <strong>the</strong> emphasis on stable channel<br />
design may refl ect <strong>the</strong> traditional discipl<strong>in</strong>es of many of<br />
<strong>the</strong> river eng<strong>in</strong>eers who have now turned <strong>the</strong>ir attentions<br />
to restoration. Whatever <strong>the</strong> provenance and merits of this<br />
approach, a focus on stable channel design requires <strong>the</strong><br />
application of geomorphic and eng<strong>in</strong>eer<strong>in</strong>g design tools<br />
(models) that are for <strong>the</strong> most part ei<strong>the</strong>r entirely empirical<br />
or empirically calibrated. As a result, different results are<br />
obta<strong>in</strong>ed when different models are applied to <strong>the</strong> same<br />
problem. Fur<strong>the</strong>rmore, <strong>the</strong> data required to apply morphological<br />
models to restoration design are often absent,<br />
<strong>in</strong>complete, or subject to measurement error. F<strong>in</strong>ally, even<br />
when a restoration design is completed, it is usually not<br />
possible to predict <strong>the</strong> precise sequence of fl ood events.<br />
In <strong>the</strong> long term fur<strong>the</strong>r variability is <strong>in</strong>troduced by climatic<br />
or catchment changes (e.g. <strong>in</strong> land use), or unanticipated<br />
social or cultural changes, all of which might shift<br />
<strong>the</strong> basic premise(s) of <strong>the</strong> design. It is evident that <strong>the</strong><br />
designers and managers of stream restoration projects are<br />
<strong>in</strong>evitably confronted with uncerta<strong>in</strong>ty.<br />
Despite, or perhaps because of, this challeng<strong>in</strong>g situation<br />
<strong>the</strong> restoration literature, albeit with some notable<br />
Preface<br />
exceptions (Wissmar and Bisson, 2003), has not yet<br />
devoted consideration to identify<strong>in</strong>g associated uncerta<strong>in</strong>ties,<br />
let alone seek<strong>in</strong>g to quantify, manage, or – where<br />
appropriate (see below) – constra<strong>in</strong> <strong>the</strong>m. Ra<strong>the</strong>r, <strong>the</strong> discipl<strong>in</strong>e<br />
has <strong>in</strong>stead tended to focus on management<br />
responses (e.g. post-project appraisal, adaptive management<br />
strategies) that only implicitly confront assumed<br />
sources of variability and uncerta<strong>in</strong>ty. Our concern is that<br />
a collective discipl<strong>in</strong>ary failure to recognise, communicate<br />
and deal appropriately with uncerta<strong>in</strong>ties might, at some<br />
time <strong>in</strong> <strong>the</strong> future, underm<strong>in</strong>e <strong>in</strong>stitutional and public confi<br />
dence <strong>in</strong> river restoration. In a fi rst attempt to address<br />
<strong>the</strong>se issues, we (toge<strong>the</strong>r with Dr Andrew Collison and<br />
Dr Sean Bennett) convened a special session on <strong>Uncerta<strong>in</strong>ty</strong><br />
<strong>in</strong> <strong>River</strong> <strong>Restoration</strong> at <strong>the</strong> 2002 Fall Meet<strong>in</strong>g of <strong>the</strong><br />
American Geophysical Union (AGU) <strong>in</strong> San Francisco,<br />
California. While recognis<strong>in</strong>g that no s<strong>in</strong>gle volume can<br />
ever cover all aspects of such a multi-faceted discipl<strong>in</strong>e as<br />
river restoration, <strong>the</strong> positive response to <strong>the</strong> topic at that<br />
AGU symposium prompted us to seek to explore it fur<strong>the</strong>r<br />
<strong>in</strong> this volume. All <strong>the</strong> chapters for this book were, <strong>the</strong>refore,<br />
specially commissioned <strong>in</strong> an attempt to provide a<br />
coherent narrative structure that offers a rational <strong>the</strong>oretical<br />
analysis of <strong>the</strong> uncerta<strong>in</strong> basis of restoration, while<br />
simultaneously provid<strong>in</strong>g practical guidance on manag<strong>in</strong>g<br />
<strong>the</strong> implications of that uncerta<strong>in</strong>ty.<br />
The result<strong>in</strong>g book is structured <strong>in</strong>to four ma<strong>in</strong> sections.<br />
Each offers a range of case studies <strong>in</strong> an attempt to ensure<br />
a wide geographic coverage. Likewise, <strong>the</strong> authorship is<br />
drawn from a range of countries and discipl<strong>in</strong>es, <strong>in</strong> an<br />
attempt to br<strong>in</strong>g a range of perspectives to <strong>the</strong> table.<br />
Section I comprises three chapters that review <strong>the</strong> nature<br />
and signifi cance of uncerta<strong>in</strong>ty <strong>in</strong> river restoration, provid<strong>in</strong>g<br />
a context for <strong>the</strong> rema<strong>in</strong>der of <strong>the</strong> book. In Chapter 1<br />
Lemons and Victor focus on <strong>the</strong> specifi c nature of scientifi<br />
c uncerta<strong>in</strong>ty <strong>in</strong> restoration, while Graf (Chapter 2)<br />
expands on this <strong>the</strong>me, identify<strong>in</strong>g a series of sources of
viii Preface<br />
uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong>ory, research and communication. In<br />
Chapter 3 Wheaton et al. syn<strong>the</strong>sise and extend <strong>the</strong>se<br />
analyses, present<strong>in</strong>g a classifi cation that suggests uncerta<strong>in</strong>ty<br />
fundamentally arises ei<strong>the</strong>r through limited knowledge<br />
or through natural system variability. This is an<br />
important dist<strong>in</strong>ction, not least because it helps to discrim<strong>in</strong>ate<br />
between those sources of uncerta<strong>in</strong>ty (limited<br />
knowledge) which should, where possible, be constra<strong>in</strong>ed<br />
(e.g. by scientifi c progress) from those sources (e.g. natural<br />
variability) that should be embraced to promote healthy<br />
system function<strong>in</strong>g. The management implications associated<br />
with each form of uncerta<strong>in</strong>ty are <strong>the</strong>refore dist<strong>in</strong>ct,<br />
but recognition that embrac<strong>in</strong>g certa<strong>in</strong> types of uncerta<strong>in</strong>ty<br />
may be both necessary and desirable to assure susta<strong>in</strong>ability<br />
is a <strong>the</strong>me that runs throughout many of <strong>the</strong> contributions<br />
here<strong>in</strong>.<br />
The book is subsequently structured to address <strong>the</strong> discrete<br />
stages <strong>in</strong> <strong>the</strong> life span of a typical restoration project,<br />
cover<strong>in</strong>g <strong>the</strong> plann<strong>in</strong>g and design activities associated with<br />
<strong>the</strong> pre-construction phase (Section II), <strong>the</strong> construction<br />
phase itself (Section III) and <strong>the</strong> long term post-construction<br />
phase (Section IV). Section II (Chapters 4 to 8) presents<br />
contributions cover<strong>in</strong>g various aspects of plann<strong>in</strong>g<br />
and design associated with restoration projects. In Chapter<br />
4 Kondolf and Yang’s review rem<strong>in</strong>ds us that restoration<br />
is fundamentally a social and cultural process, with variability<br />
<strong>in</strong> cultural values act<strong>in</strong>g as a signifi cant contributor<br />
of uncerta<strong>in</strong>ty. Present<strong>in</strong>g an Australian case study, where<br />
<strong>the</strong> aim was to restore fl ows capable of fl ush<strong>in</strong>g fi ne sediment<br />
from river gravels, Stewardson et al. (Chapter 5)<br />
identify <strong>the</strong> limits <strong>in</strong> our understand<strong>in</strong>g of hydrological,<br />
hydraulic and geomorphic processes and how <strong>the</strong>se constra<strong>in</strong><br />
our ability to model river system dynamics. In terms<br />
of <strong>the</strong> uncerta<strong>in</strong>ty classifi cation discussed <strong>in</strong> Chapter 3,<br />
<strong>the</strong>ir focus is essentially on quantify<strong>in</strong>g <strong>the</strong> magnitude of<br />
uncerta<strong>in</strong>ty due to limited knowledge. Their results (that<br />
<strong>the</strong> magnitude of designed fl ush<strong>in</strong>g fl ows is subject to<br />
uncerta<strong>in</strong>ty estimates approximately twice that of <strong>the</strong> fl ow<br />
itself) re<strong>in</strong>force <strong>the</strong> earlier suggestion that restoration is<br />
<strong>in</strong>deed an uncerta<strong>in</strong> discipl<strong>in</strong>e. Whe<strong>the</strong>r this really means<br />
that we should have ‘unreasonable confi dence’ <strong>in</strong> restoration,<br />
as suggested by <strong>the</strong>ir provocative sub-title, is a <strong>the</strong>me<br />
that is cont<strong>in</strong>ued throughout <strong>the</strong> book.<br />
In contrast to <strong>the</strong> focus on uncerta<strong>in</strong>ty due to limited<br />
knowledge expounded <strong>in</strong> Chapter 5, Chapter 6 (Hughes<br />
et al.) reviews some of <strong>the</strong> diffi culties associated with<br />
extend<strong>in</strong>g restoration <strong>in</strong>to complex riparian and fl oodpla<strong>in</strong><br />
habitats, emphasis<strong>in</strong>g that <strong>in</strong> <strong>the</strong>se systems uncerta<strong>in</strong>ty (<strong>in</strong><br />
this case <strong>in</strong> <strong>the</strong> form of physical diversity and variability)<br />
is necessary to underp<strong>in</strong> <strong>the</strong> successful restoration of<br />
forest fl oodpla<strong>in</strong> ecosystems. In Chapter 7, Clifford et al.’s<br />
comprehensive review of how <strong>the</strong> restoration of fl ow<br />
hydrology and hydraulics can be used to enhance aquatic<br />
habitats also recognises <strong>the</strong> importance of restor<strong>in</strong>g natural<br />
variability, and provides recommendations on how such<br />
variability can be <strong>in</strong>terpreted <strong>in</strong> modell<strong>in</strong>g <strong>in</strong>vestigations.<br />
The <strong>the</strong>me of uncerta<strong>in</strong>ty associated with ecological<br />
targets is explored fur<strong>the</strong>r <strong>in</strong> Chapter 8 (Perrow et al.),<br />
where <strong>the</strong> paradox that uncerta<strong>in</strong>ty is often viewed as a<br />
pejorative term is aga<strong>in</strong> highlighted, even if it is uncerta<strong>in</strong>ty<br />
(<strong>in</strong> <strong>the</strong> form of natural variability) that is <strong>the</strong> key<br />
mechanism for susta<strong>in</strong><strong>in</strong>g healthy ecosystems. How uncerta<strong>in</strong>ty<br />
due to <strong>the</strong> lack of understand<strong>in</strong>g of a discipl<strong>in</strong>e<br />
(ecology) by river managers has led to a lack of us<strong>in</strong>g<br />
available science with<strong>in</strong> restoration process is also<br />
highlighted.<br />
Section III (Chapters 9 to 12) addresses <strong>the</strong> construction<br />
phase of a restoration project, which is defi ned <strong>in</strong> this book<br />
as extend<strong>in</strong>g up to one or two years after completion of<br />
<strong>the</strong> project. The contributions <strong>in</strong> this section are written<br />
primarily by river practitioners, who employ <strong>the</strong>ir collective<br />
experience to offer a range of perspectives on uncerta<strong>in</strong>ties<br />
encountered dur<strong>in</strong>g this key stage of restoration.<br />
Mant et al. (Chapter 9) review <strong>the</strong> diffi culties encountered<br />
dur<strong>in</strong>g construction and note that strong teamwork skills<br />
are required to ensure that <strong>the</strong> design concepts provided<br />
by geomorphologists and ecologists are correctly translated<br />
<strong>in</strong>to practice by those responsible for construction.<br />
A diffi culty here is that restoration is seen as a relatively<br />
new facet of civil eng<strong>in</strong>eer<strong>in</strong>g, such that contractors may<br />
not always have <strong>the</strong> experience necessary to recognise that<br />
variability, ra<strong>the</strong>r than uniformity (<strong>the</strong>ir experience to<br />
date), is often necessary. To this end it is essential that<br />
designers <strong>in</strong>form <strong>the</strong> workforce of <strong>the</strong> specifi c requirements<br />
of <strong>the</strong> river restoration project, while project managers<br />
must also take responsibility for monitor<strong>in</strong>g<br />
construction as it progresses. This po<strong>in</strong>ts to <strong>the</strong> importance<br />
of ensur<strong>in</strong>g that <strong>the</strong> constructed project does <strong>in</strong>deed<br />
conform to <strong>the</strong> design specifi cations, rais<strong>in</strong>g <strong>the</strong> issue of<br />
evaluat<strong>in</strong>g project outcomes.<br />
This subject is <strong>the</strong> <strong>the</strong>me of <strong>the</strong> next three chapters.<br />
Sk<strong>in</strong>ner et al. (Chapter 10) review post-project appraisals<br />
with reference to both physical and ecological measures<br />
of success, whilst Rhoads et al. (Chapter 11) focus on<br />
methods for evaluat<strong>in</strong>g <strong>the</strong> geomorphological performance<br />
of restored rivers, provid<strong>in</strong>g examples from heavily urbanised<br />
catchments <strong>in</strong> Ill<strong>in</strong>ois <strong>in</strong> <strong>the</strong> USA. Both contributions<br />
emphasise <strong>the</strong> key need for both pre and post-project monitor<strong>in</strong>g,<br />
even if only to a m<strong>in</strong>imum standard. This is viewed<br />
as necessary to verify that projects are constructed accord<strong>in</strong>g<br />
to <strong>the</strong>ir design, as well as an <strong>in</strong>tegral tool of adaptive<br />
management that can make project adjustments <strong>in</strong> <strong>the</strong> face<br />
of uncerta<strong>in</strong>ties <strong>in</strong>troduced by variable post-project conditions.<br />
This <strong>the</strong>me is fur<strong>the</strong>r explored by Brookes and
Dangerfi eld (Chapter 12), who propose that managers<br />
should adopt cont<strong>in</strong>uous improvement as an overall operational<br />
philosophy for restor<strong>in</strong>g rivers. The term cont<strong>in</strong>uous<br />
improvement is widely documented <strong>in</strong> human resource,<br />
organisational management and environmental management<br />
literature, and is taken to be a philosophy of learn<strong>in</strong>g<br />
dur<strong>in</strong>g <strong>the</strong> construction and post-construction phases (and<br />
mak<strong>in</strong>g adaptations to a particular project as necessary)<br />
for <strong>the</strong> benefi t of <strong>the</strong> cont<strong>in</strong>u<strong>in</strong>g work and future practice.<br />
This appears to be a robust approach that has <strong>the</strong> potential,<br />
over time, to reduce uncerta<strong>in</strong>ties associated with limited<br />
knowledge while simultaneously provid<strong>in</strong>g a framework<br />
for adaptive response to uncerta<strong>in</strong>ties associated with<br />
natural variability.<br />
Section IV (Chapters 13 and 14) addresses <strong>the</strong> challenge<br />
of <strong>the</strong> need to assure <strong>the</strong> long term susta<strong>in</strong>ability of restoration<br />
projects <strong>in</strong> <strong>the</strong> face of uncerta<strong>in</strong> futures. There is a<br />
clear recognition that as <strong>the</strong> time scales over which project<br />
outcomes should be considered <strong>in</strong>crease, <strong>the</strong>re is a concomitant<br />
need to address <strong>in</strong>creased spatial scales. Specifi -<br />
cally, <strong>the</strong>re is a need to consider how catchment-scale<br />
processes (which <strong>in</strong>fl uence <strong>the</strong> fl uxes of water and sediment<br />
supplied to restoration reaches) are to be susta<strong>in</strong>ed<br />
<strong>in</strong> <strong>the</strong> long term. Clearly, as spatial and temporal scales<br />
<strong>in</strong>crease, <strong>the</strong>n so do <strong>the</strong> uncerta<strong>in</strong>ties particularly, but not<br />
exclusively so, those associated with <strong>in</strong>creases <strong>in</strong> spatial<br />
and temporal variability. In Chapter 13 Downs and Gregory<br />
br<strong>in</strong>g a hydromorphological perspective to <strong>the</strong>se issues,<br />
suggest<strong>in</strong>g that <strong>the</strong> bounds of <strong>the</strong>se uncerta<strong>in</strong>ties can be<br />
evaluated with reference to long term (palaeo)hydrological<br />
and geomorphological evidence of past catchment response<br />
– <strong>in</strong> effect advocat<strong>in</strong>g a more precise defi nition of <strong>the</strong><br />
uncerta<strong>in</strong>ty due to natural variability.<br />
The fi nal chapter (Chapter 14; Newson & Clark) provides<br />
an apt conclusion. Recognis<strong>in</strong>g that uncerta<strong>in</strong>ty (<strong>in</strong><br />
<strong>the</strong> form of natural variability) is both endemic and necessary,<br />
it is noted that <strong>the</strong>re is a confl ict between <strong>the</strong> precautionary<br />
pr<strong>in</strong>ciple – a cornerstone of susta<strong>in</strong>able th<strong>in</strong>k<strong>in</strong>g<br />
– and uncerta<strong>in</strong>ty. Newson and Clark recognise that all<br />
restorations have outcomes that are to some extent unpredictable,<br />
and <strong>the</strong> precautionary pr<strong>in</strong>ciple thus becomes a<br />
recipe for <strong>in</strong>action. <strong>Uncerta<strong>in</strong>ty</strong> is <strong>the</strong>refore simultaneously<br />
necessary for, but also a barrier to, susta<strong>in</strong>ability.<br />
They attempt to resolve this particular problem by identify<strong>in</strong>g<br />
management and restoration opportunities that are<br />
susta<strong>in</strong>able despite be<strong>in</strong>g uncerta<strong>in</strong>, not<strong>in</strong>g that <strong>in</strong> practical<br />
terms it is to adaptive management that we most often turn<br />
for a way forward, re<strong>in</strong>forc<strong>in</strong>g a series of conclusions from<br />
earlier chapters.<br />
Do we have unreasonable confi dence <strong>in</strong> restoration,<br />
based on <strong>the</strong> state of <strong>the</strong> art, or are we happy to boldly go<br />
Preface ix<br />
with <strong>the</strong> uncerta<strong>in</strong> ebb and fl ow of natural variability?<br />
Perhaps <strong>the</strong> way forward is through a clearer and more<br />
transparent approach to communicat<strong>in</strong>g uncerta<strong>in</strong>ty, such<br />
that all participants – and especially stakeholders – understand<br />
that while every effort can and should be made to<br />
identify, and where possible reduce, scientifi c and methodological<br />
uncerta<strong>in</strong>ties, uncerta<strong>in</strong>ty due to natural variability<br />
is both welcome and necessary to susta<strong>in</strong> healthy<br />
aquatic ecosystems. This implies a concerted approach on<br />
two fronts: scientists can cont<strong>in</strong>ue to refi ne <strong>the</strong> knowledge<br />
base (and managers need to recognise <strong>the</strong> value of this<br />
research for <strong>the</strong>ir (adaptive) management practices), while<br />
managers must work harder to build and accommodate<br />
variability <strong>in</strong>to projects. <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> river restoration is<br />
endemic but clearly offers opportunities, not just as a<br />
rationale for fur<strong>the</strong>r research, but fundamentally for more<br />
susta<strong>in</strong>ably managed restoration projects. Just as life goes<br />
on with little confi dence or ability to predict <strong>the</strong> future, so<br />
restoration must cont<strong>in</strong>ue to evolve and adopt an approach<br />
that is consistent with <strong>the</strong> uncerta<strong>in</strong> function<strong>in</strong>g of river<strong>in</strong>e<br />
ecosystems.<br />
In clos<strong>in</strong>g this preface, we would like to acknowledge<br />
those who have made signifi cant contributions dur<strong>in</strong>g <strong>the</strong><br />
production of this book. Firstly, we would like to thank<br />
those numerous professionals who provided detailed peer<br />
reviews of each chapter, often work<strong>in</strong>g to tight deadl<strong>in</strong>es.<br />
The anonymous nature of peer review means that we are<br />
unable to identify <strong>the</strong>m here, but you know who you are!<br />
F<strong>in</strong>ally, Tim Aspden and <strong>the</strong> staff of <strong>the</strong> Cartographic Unit<br />
at <strong>the</strong> School of Geography, University of Southampton,<br />
provided guidance on, and help with, <strong>the</strong> production of<br />
much of <strong>the</strong> artwork. A fi nal acknowledgment must go to<br />
our families who have put up with longer hours than<br />
normal (or natural!) <strong>in</strong> <strong>the</strong> drive for completion.<br />
Stephen Darby and David Sear<br />
December 2007<br />
REFERENCES<br />
Brookes A, Shields FD. 1996. <strong>River</strong> Channel <strong>Restoration</strong>:<br />
Guid<strong>in</strong>g Pr<strong>in</strong>ciples for Susta<strong>in</strong>able Projects. John Wiley &<br />
Sons Ltd: Chichester, UK.<br />
Gregory KJ, Park CC. 1974. Adjustment of river channel capacity<br />
downstream from a reservoir. Water Resources Research 10:<br />
840–873.<br />
Sear DA, Arnell NW. 2006. The application of palaeohydrology<br />
to river management. Catena 66: 169–183.<br />
Wissmar RC, Bisson PA (Eds). 2003c. Strategies for Restor<strong>in</strong>g<br />
<strong>River</strong> Ecosystems: Sources of Variability and <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong><br />
Natural and Managed Systems. American Fisheries Society:<br />
Be<strong>the</strong>sda, Maryland, USA.
Jorge Abad, Department of Civil and Environmental Eng<strong>in</strong>eer<strong>in</strong>g,<br />
University of Ill<strong>in</strong>ois at Urbana-Champaign,<br />
Urbana, Ill<strong>in</strong>ois 61801, USA.<br />
Dr Mike Acreman, Water Resources & Environment Division,<br />
Centre for Ecology & Hydrology, Maclean Build<strong>in</strong>g,<br />
Crowmarsh Gifford, Wall<strong>in</strong>gford, Oxfordshire OX10 8BB,<br />
UK (man@ceh.ac.uk).<br />
Professor Fabian Bombadelli, Department of Civil and<br />
Environmental Eng<strong>in</strong>eer<strong>in</strong>g, University of California at<br />
Davis, Davis, California 95616, USA (fabombardelli@<br />
ucdavis.edu).<br />
Dr Douglas Booker, National Institute of Water and<br />
Atmospheric Research, 10 Kyle St., Riccarton, Christchurch,<br />
8011, New Zealand (d.booker@niwa.co.nz).<br />
Dr Andrew Brookes, Jacobs (UK), School Green, Sh<strong>in</strong>fi<br />
eld, Read<strong>in</strong>g RG2 9HL, UK (andrew.brookes@jacobs.<br />
com).<br />
Professor Mike Clark, School of Geography, University of<br />
Southampton, Highfi eld, Southampton SO17 1BJ, UK<br />
(M.J.Clark@soton.ac.uk).<br />
Professor Nick Clifford, School of Geography, The University<br />
of Nott<strong>in</strong>gham, University Park, Nott<strong>in</strong>gham NG7<br />
2RD, UK (Nick.Clifford@nott<strong>in</strong>gham.ac.uk).<br />
Dr Helen Dangerfi eld, Royal Haskon<strong>in</strong>g, 4 Dean’s<br />
Yard, London, SW19 3NL, UK (Helen.dangerfi eld@<br />
royalhaskon<strong>in</strong>g.com).<br />
List of Contributors<br />
Dr Mel<strong>in</strong>da Daniels, Department of Geography, Kansas<br />
State University, USA (mel<strong>in</strong>da.daniels@uconn.edu).<br />
Dr Stephen Darby, School of Geography, University of<br />
Southampton, Highfi eld, Southampton SO17 1BJ, UK<br />
(S.E.Darby@soton.ac.uk).<br />
Dr Peter Downs, Stillwater Sciences, 2855 Telegraph<br />
Avenue, #400, Berkeley, California 94705 USA (downs@<br />
stillwatersci.com).<br />
Dr Judy England, Environment Agency, Apollo Court, 2<br />
Bishop’s Square, St. Albans Road West, Hatfi eld AL10<br />
9EX, UK.<br />
Professor Marcelo Garcia, Department of Civil and Environmental<br />
Eng<strong>in</strong>eer<strong>in</strong>g, University of Ill<strong>in</strong>ois at Urbana-<br />
Champaign, Urbana, Ill<strong>in</strong>ois 61801, USA (mhgarcia@<br />
uiuc.edu).<br />
Professor William Graf, Department of Geography, University<br />
of South Carol<strong>in</strong>a, Columbia, South Carol<strong>in</strong>a<br />
29208, USA (graf@sc.edu).<br />
Professor Ken Gregory, School of Geography, University<br />
of Southampton, Highfi eld, Southampton SO17 1BJ, UK<br />
(Ken.Gregory@bt<strong>in</strong>ternet.com).<br />
Dr Simon Harrison, Department of Zoology, Ecology and<br />
Plant Sciences, University College Cork, Lee Malt<strong>in</strong>gs,<br />
Prospect Row, Cork, Ireland (s.harrison@ucc.ie).
xii List of Contributors<br />
Dr Franc<strong>in</strong>e Hughes, Department of Life Sciences, Anglia<br />
Rusk<strong>in</strong> University, East Road, Cambridge CB1 1PT, UK<br />
(f.hughes@anglia.ac.uk).<br />
Mart<strong>in</strong> Janes, <strong>River</strong> <strong>Restoration</strong> Centre Manager, The<br />
<strong>River</strong> <strong>Restoration</strong> Centre, Build<strong>in</strong>g 53, Cranfi eld University,<br />
Cranfi eld, Bedfordshire MK43 0AL, UK (rrc@<strong>the</strong>rrc.<br />
co.uk).<br />
Professor Mat Kondolf, Department of Landscape Architecture<br />
and Environmental Plann<strong>in</strong>g, University of California,<br />
Berkeley, California 94720-2000, USA (kondolf@<br />
ucl<strong>in</strong>k.berkeley.edu).<br />
David Leem<strong>in</strong>g, Consultant Ecologist, Sp<strong>in</strong>dlewood, 45<br />
West End, Ashwell, Hertfordshire SG7 5QY, UK.<br />
Professor John Lemons, Department of Environmental<br />
Studies, University of New England,11 Hills Beach Road,<br />
Biddeford, Ma<strong>in</strong>e 04005, USA (jlemons@une.edu).<br />
Dr Jenny Mant, The <strong>River</strong> <strong>Restoration</strong> Centre, Build<strong>in</strong>g<br />
53, Cranfi eld University, Cranfi eld, Bedfordshire MK43<br />
0AL, UK (rrc@<strong>the</strong>rrc.co.uk).<br />
Dr Tim Moss, Institute for Regional Development and<br />
Structural Plann<strong>in</strong>g (IRS), Flakenstrasse 28–31, 15537<br />
Erkner, Germany. (MossT@irs-net.de).<br />
Professor Malcolm Newson, Department of Geography,<br />
University of Newcastle-Upon-Tyne, Newcastle-Upon-<br />
Tyne NE1 7RU, UK (M.D.Newson@ncl.ac.uk).<br />
Dr Mart<strong>in</strong> Perrow, ECON, Ecological Consultancy,<br />
Norwich Research Park, Colney Lane, Norwich, Norfolk<br />
NR4 7UH, UK (m.perrow@econ-ecology.com).<br />
Professor Bruce Rhoads, Department of Geography, University<br />
of Ill<strong>in</strong>ois at Urbana-Champaign, Room 220 Davenport<br />
Hall, 607 South Ma<strong>the</strong>ws Avenue, Urbana, Ill<strong>in</strong>ois<br />
61801-3671, USA (brhoads@uiuc.edu).<br />
Professor Keith Richards, Department of Geography, University<br />
of Cambridge, Down<strong>in</strong>g Place, Cambridge CB2<br />
3EN, UK (keith.richards@geog.cam.ac.uk).<br />
Dr Roy Richardson, Scottish Environment Protection<br />
Agency, Burnbrae, Mossilee Road, Galashiels TD1 1NF,<br />
UK.<br />
Dr Jose Rodriguez, Faculty of Eng<strong>in</strong>eer<strong>in</strong>g and Built Environment,<br />
University of Newcastle, Newcastle, New South<br />
Wales 2308, Australia (jose.rodriguez@newcastle.edu.<br />
au).<br />
Dr Ian Ru<strong>the</strong>rfurd, Geography Program, School of<br />
Resource Management, University of Melbourne, Melbourne,<br />
Victoria 3010 Australia (idruth@unimelb.edu.au).<br />
Professor David Sear, School of Geography, University of<br />
Southampton, Highfi eld, Southampton SO17 1BJ, UK<br />
(D.Sear@soton.ac.uk).<br />
Dr F. Doug Shields Jr, Water Quality and Ecology<br />
Research Unit, National Sedimentation Laboratory, USDA<br />
Agricultural Research Service, National Sedimentation<br />
Laboratory, PO Box 1157, Oxford, Mississippi, USA<br />
(dshields@ars.usda.gov).<br />
Eleanor R. Skeate, ECON, Ecological Consultancy,<br />
Norwich Research Park, Colney Lane, Norwich, Norfolk<br />
NR4 7UH, UK (e.skeate@econ-ecology.com).<br />
Dr Kev<strong>in</strong> Sk<strong>in</strong>ner, Pr<strong>in</strong>cipal Geomorphologist, Jacobs<br />
Babtie, School Green, Sh<strong>in</strong>fi eld, Read<strong>in</strong>g RG2 9HL UK<br />
(Kev<strong>in</strong>.sk<strong>in</strong>ner@jacobs.com).<br />
Dr Michael Stewardson, Department of Civil and Environmental<br />
Eng<strong>in</strong>eer<strong>in</strong>g and eWater CRC, University of<br />
Melbourne, Melbourne, Victoria, 3010 Australia (mjstew@<br />
unimelb.edu.au).<br />
Mark L. Toml<strong>in</strong>son ECON, Ecological Consultancy,<br />
Norwich Research Park, Colney Lane, Norwich, Norfolk<br />
NR4 7UH, UK (m.toml<strong>in</strong>son@econ-ecology.com).<br />
Professor Reg<strong>in</strong>ald Victor, Centre for Environmental<br />
Studies and Research, c/o Department of Biology, Sultan<br />
Qaboos University, PO Box 36, Al-Khod, PC 123, Muscat<br />
Sultanate of Oman (rvictor@squ.edu.om).<br />
Joseph M. Wheaton, Institute of Geography and Earth<br />
Sciences, University of Wales, Aberystwyth SY23 3DB,<br />
UK (Joe.Wheaton@aber.ac.uk).<br />
Dr Chia-N<strong>in</strong>g Yang, Department of Landscape Ar -<br />
chitecture, California State Polytechnic University,<br />
Pomona, California 91768, USA (cnyang@csupomona.<br />
edu).
SECTION I<br />
Introduction: The Nature and Signifi cance<br />
of <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>River</strong> <strong>Restoration</strong>
<strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat Edited by Stephen Darby and David Sear<br />
© 2008 John Wiley & Sons, Ltd<br />
1<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>River</strong> <strong>Restoration</strong><br />
John Lemons 1 and Reg<strong>in</strong>ald Victor 2<br />
1 Department of Environmental Studies, University of New England, USA<br />
2 Center for Environmental Studies and Research, Department of Biology, Sultan Qaboos University, Sultanate of Oman<br />
1.1 INTRODUCTION<br />
As we are well aware, rivers fundamentally shape <strong>the</strong><br />
planet and human life. Both ancient and modern societies<br />
have developed and fl ourished <strong>in</strong> <strong>the</strong> proximity of rivers<br />
and this trend has cont<strong>in</strong>ued till modern times. Nienhuis<br />
and Leuven (2001) summarize how humans have spatially<br />
and temporally altered rivers over a 6000-year period by<br />
various anthropogenic activities. For example, <strong>in</strong>tensive<br />
use of European rivers started over 500 years ago lead<strong>in</strong>g<br />
to <strong>the</strong> loss of <strong>the</strong>ir ecological <strong>in</strong>tegrity (Smits et al., 2000).<br />
Some rivers were altered for navigation, fl ood control,<br />
agriculture and reclamation of land for urban development,<br />
while most were used as chutes for waste disposal<br />
<strong>in</strong>clud<strong>in</strong>g sewage, <strong>the</strong>rmal effl uents and both nontoxic and<br />
toxic chemicals; some rivers were also rout<strong>in</strong>ely dredged<br />
to facilitate <strong>the</strong> transport and storage of timber, while<br />
o<strong>the</strong>rs were heavily fi shed (Ward and Stanford, 1979; De<br />
Wall et al., 1995; Eiseltova and Biggs, 1995).<br />
Large river systems (stream order >8) all over <strong>the</strong> world<br />
have been extensively dammed for hydroelectric power,<br />
recreation, fl ood control and to divert water to support<br />
agriculture. Impacts of large dams <strong>in</strong>clude <strong>the</strong> loss of<br />
fi sheries and <strong>the</strong> ecological collapse of <strong>the</strong> entire river<br />
regime (Balon and Coche, 1974; Rzoska, 1976; Obeng,<br />
1981). Extensive series of levees built along large rivers<br />
have caused major losses of ecosystem structure and function.<br />
Along <strong>the</strong> Mississippi <strong>River</strong>, <strong>the</strong> largest river <strong>in</strong> North<br />
America, levees threaten federal plans to protect endangered<br />
species (EPA, 2004). The effects of impound<strong>in</strong>g<br />
small rivers (stream order 4–8) are even more drastic. In<br />
some West African small rivers entire fi sh communities<br />
had changed due to impoundment and <strong>the</strong> ecological perturbations<br />
extended for considerable distances downstream<br />
(Victor and Tetteh, 1988; Victor and Meye, 1994; Victor<br />
and Onomivbori, 1996). Gopal (2003) describes how<br />
rivers <strong>in</strong> arid and semi-arid regions <strong>in</strong> Asia are be<strong>in</strong>g<br />
degraded due to overexploitation of natural resources,<br />
sal<strong>in</strong>ization, pollution and <strong>in</strong>troduction of exotic species.<br />
Just as rivers have undergone alteration, so too have<br />
<strong>the</strong>re been efforts to restore <strong>the</strong>m <strong>in</strong> order to provide benefi<br />
ts to <strong>the</strong> environment and/or human health, as this book<br />
attests (see also MacMahon and Holl, 2001). Obviously,<br />
scientifi c research contributes to river restoration by: provid<strong>in</strong>g<br />
reliable and needed explanatory or heuristic knowledge<br />
and understand<strong>in</strong>g of restoration problems; help<strong>in</strong>g<br />
to identify and defi ne new research needs and directions<br />
through <strong>the</strong> acquisition of factual <strong>in</strong>formation; and <strong>in</strong>form<strong>in</strong>g<br />
policy and decision mak<strong>in</strong>g (Caldwell, 1996).<br />
A major premise of this book is that to be susta<strong>in</strong>able,<br />
river restoration projects need to effectively recreate a<br />
rivers’ functional characteristics tak<strong>in</strong>g <strong>in</strong>to account <strong>the</strong><br />
dynamic geomorphic characteristics. While many restoration<br />
projects have benefi ted environmental and/or human<br />
health, understudied sources of uncerta<strong>in</strong>ty limit confi -<br />
dence <strong>in</strong> predict<strong>in</strong>g <strong>the</strong> outcomes of restoration activities<br />
and programs. Specifi c examples of uncerta<strong>in</strong>ty <strong>in</strong> river<br />
restoration discussed <strong>in</strong> this book <strong>in</strong>clude those <strong>in</strong>herent<br />
<strong>in</strong>: river management processes; <strong>the</strong> plann<strong>in</strong>g and design<br />
phases of restoration projects; hydraulic and hydrological<br />
aspects of restoration; water quantity issues; identify<strong>in</strong>g<br />
appropriate ecological characteristics and predict<strong>in</strong>g <strong>the</strong>ir
4 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
responses <strong>in</strong> restoration designs; and <strong>the</strong> construction and<br />
post-construction phases of restoration projects. The<br />
sources of uncerta<strong>in</strong>ties <strong>in</strong>clude: lack of scientifi c and<br />
o<strong>the</strong>r <strong>in</strong>formation; limitations of analytical methods<br />
and tools; complexities of river systems; and needs<br />
to make value-laden judgments at all stages of river restoration<br />
problem identifi cation, analysis and solution<br />
implementation.<br />
Beg<strong>in</strong>n<strong>in</strong>g <strong>in</strong> and s<strong>in</strong>ce <strong>the</strong> early 1990s some philosophers,<br />
scientists and public policy experts concluded that<br />
<strong>the</strong> sources and implications of scientifi c and o<strong>the</strong>r uncerta<strong>in</strong>ty<br />
<strong>in</strong> environmental problem solv<strong>in</strong>g, <strong>in</strong>clud<strong>in</strong>g restoration,<br />
have been understudied and, as a consequence,<br />
not suffi ciently taken <strong>in</strong>to account by researchers, public<br />
policy makers and decision makers (Mayo and Hollander,<br />
1991; Cranor, 1993; Shrader-Frechette and McCoy, 1993;<br />
Funtowicz and Ravetz, 1995; Lemons and Brown, 1995;<br />
Lemons, 1996; EEA, 2001; Kriebel et al., 2001; Tickner,<br />
2002, 2003).<br />
In agree<strong>in</strong>g with this conclusion, <strong>the</strong> objective <strong>in</strong> this<br />
chapter is <strong>the</strong>refore to fi rst discuss various broad views<br />
about scientifi c uncerta<strong>in</strong>ty and <strong>in</strong>dicate how and why<br />
<strong>the</strong>se need to be taken <strong>in</strong>to greater account by scientists,<br />
policy makers and decision makers. (O<strong>the</strong>r chapters<br />
address uncerta<strong>in</strong>ty and analyze <strong>in</strong> more concrete detail<br />
how it <strong>in</strong>teracts with <strong>the</strong> specifi c <strong>the</strong>ories and practices of<br />
river restoration.). Discussion <strong>the</strong>n focuses on what might<br />
constitute ‘good’ science when science is used to <strong>in</strong>form<br />
policy and decision mak<strong>in</strong>g under conditions of scientifi c<br />
uncerta<strong>in</strong>ty. Value-laden sources and implications of<br />
uncerta<strong>in</strong>ty <strong>in</strong> river restoration are <strong>the</strong>n discussed because<br />
<strong>the</strong>y are both important but understudied. Discussion of<br />
<strong>the</strong> value-laden sources and implications of uncerta<strong>in</strong>ty is<br />
followed with: a brief discussion of some of <strong>the</strong> practical<br />
and policy implications of uncerta<strong>in</strong>ty <strong>in</strong> river restoration,<br />
and, fi nally, a brief case study of river restoration <strong>in</strong> order<br />
to communicate our views with a practical example. For<br />
reasons of brevity <strong>the</strong> case study communicates views<br />
about some, but not all, aspects of uncerta<strong>in</strong>ty <strong>in</strong> river<br />
restoration.<br />
Paren<strong>the</strong>tically, here it is necessary to comment on defi -<br />
nitions of ‘restoration’ when used <strong>in</strong> <strong>the</strong> context of river<br />
restoration. The fi eld of restoration ecology suffers from<br />
a lack of conceptual clarity concern<strong>in</strong>g its mean<strong>in</strong>g, goals<br />
and objectives. S<strong>in</strong>ce about <strong>the</strong> mid-1980s, <strong>the</strong> fi eld of<br />
river restoration has <strong>in</strong>creas<strong>in</strong>gly evolved <strong>in</strong> an attempt to<br />
better meet societies’ needs to more effectively repair<br />
damage to rivers (e.g., Cairns and Heckman, 1996; Karr<br />
and Chu, 1999; Cairns, 2001). The Society of Wetlands<br />
Scientists (SWS, 2000) defi ned restoration as ‘actions<br />
taken <strong>in</strong> a converted or degraded natural wetland that<br />
result <strong>in</strong> <strong>the</strong> re-establishment of ecological processes,<br />
function, and biotic/abiotic l<strong>in</strong>kages and lead to a persistent,<br />
resilient system <strong>in</strong>tegrated with<strong>in</strong> its landscape.’ In<br />
2002, <strong>the</strong> Society for Ecological <strong>Restoration</strong> (SER, 2002)<br />
defi ned restoration as <strong>the</strong> ‘. . . process of assist<strong>in</strong>g <strong>the</strong><br />
recovery of an ecosystem that has been degraded, damaged,<br />
or destroyed.’ Regardless of <strong>the</strong>se defi nitions, <strong>the</strong> goals<br />
and objectives of river restoration are not clear.<br />
Rolston (1988) believes that where possible ecosystems<br />
should be returned to <strong>the</strong>ir ‘natural’ or ‘orig<strong>in</strong>al’ condition.<br />
Westra (1995) argues that restoration should focus on<br />
restor<strong>in</strong>g ecosystems’ abilities to cont<strong>in</strong>ue <strong>the</strong>ir ongo<strong>in</strong>g<br />
change and development unconstra<strong>in</strong>ed by human <strong>in</strong>terruptions<br />
past or present. The United States National<br />
Research Council (NRC, 1999) defi ned restoration as<br />
‘<strong>the</strong> return of an ecosystem to a close approximation of<br />
its condition prior to disturbance.’ This defi nition was<br />
expanded by Cairns (2001), who asserted that <strong>the</strong> goal of<br />
restoration should be devoted to ‘return<strong>in</strong>g damaged ecosystems<br />
to a condition that is structurally and functionally<br />
similar to <strong>the</strong> predisturbance state.’<br />
Alternatively, o<strong>the</strong>rs <strong>in</strong>volved <strong>in</strong> <strong>the</strong> fi eld of restoration<br />
ecology provide defi nitions for restoration that more<br />
explicitly focus on historical, social, cultural, political,<br />
aes<strong>the</strong>tic and moral aspects. For example, Sweeney (2000)<br />
argues that restoration should focus on <strong>the</strong> value-laden<br />
social and ethical perspectives regard<strong>in</strong>g what constitutes<br />
a ‘restored’ ecosystem. Some o<strong>the</strong>rs ma<strong>in</strong>ta<strong>in</strong> that conservation<br />
and, by implication, restoration goals should take<br />
<strong>in</strong>to account <strong>the</strong> views and practices of rural and <strong>in</strong>digenous<br />
people who depend on <strong>the</strong> ecosystems for <strong>the</strong>ir<br />
physical and cultural subsistence, and should also <strong>in</strong>clude<br />
scientifi c and nonscientifi c considerations (Gomez-Pompa<br />
and Kaus, 1992; Westra, 1995; Light and Higgs, 1996;<br />
Higgs, 1997; Chauhan, 2003). Regier (1995) proposes an<br />
abstract defi nition for restoration that is dependent on<br />
what people believe as foster<strong>in</strong>g a state of ‘well-be<strong>in</strong>g.’<br />
Obviously, lack of conceptual clarity about restoration<br />
<strong>in</strong>troduces an element of uncerta<strong>in</strong>ty <strong>in</strong>to restoration<br />
problem solv<strong>in</strong>g. In this chapter, while be<strong>in</strong>g m<strong>in</strong>dful of<br />
<strong>the</strong> unresolved problems of conceptual clarity regard<strong>in</strong>g<br />
‘restoration’ o<strong>the</strong>r sources and implications of uncerta<strong>in</strong>ty<br />
and <strong>the</strong>ir relevance to river restoration are focused upon.<br />
1.2 BROAD PHILOSOPHICAL VIEWS ABOUT<br />
SCIENTIFIC UNCERTAINTY<br />
Dur<strong>in</strong>g <strong>the</strong> 19th century <strong>the</strong>re was a high degree of confi -<br />
dence <strong>in</strong> <strong>the</strong> methods and tools of science and technology<br />
to <strong>in</strong>crease understand<strong>in</strong>g of <strong>the</strong> natural world and enable<br />
robust predictions of its future states. This confi dence <strong>in</strong><br />
science contributed to beliefs that ‘nature’ could be controlled<br />
and rendered useful to humank<strong>in</strong>d (Latour, 1988).
Contribut<strong>in</strong>g to <strong>the</strong>se beliefs were philosophers and scientists<br />
(so-called ‘logical positivists’) who proposed that<br />
an important goal of science should focus on formulat<strong>in</strong>g<br />
hypo<strong>the</strong>ses and conduct<strong>in</strong>g observations to test <strong>the</strong>m,<br />
develop<strong>in</strong>g an understand<strong>in</strong>g of processes and l<strong>in</strong>kages<br />
among variables, and develop<strong>in</strong>g conclusions and predictions<br />
about which <strong>the</strong>re is a high degree of confi dence.<br />
More specifi cally, <strong>the</strong> logical positivistic view of science<br />
assumes that: knowledge is founded on experience; concepts<br />
and generalizations only represent <strong>the</strong> particulars<br />
from which <strong>the</strong>y have been abstracted; mean<strong>in</strong>g is<br />
grounded <strong>in</strong> observation; <strong>the</strong> sciences are unifi ed accord<strong>in</strong>g<br />
to <strong>the</strong> methodology of <strong>the</strong> natural sciences and <strong>the</strong><br />
ideal pursued <strong>in</strong> knowledge is <strong>the</strong> form of ma<strong>the</strong>matically<br />
formulated universal science deducible from <strong>the</strong> smallest<br />
number of possible axioms; and values are not facts<br />
grounded <strong>in</strong> observation and <strong>the</strong>refore cannot be <strong>in</strong>cluded<br />
as a part of scientifi c knowledge. One <strong>the</strong> one hand, while<br />
logical positivism has <strong>in</strong>fl uenced <strong>the</strong> th<strong>in</strong>k<strong>in</strong>g of modern<br />
scientists public policy makers, and decision makers, on<br />
<strong>the</strong> o<strong>the</strong>r it does not enjoy wide support from contemporary<br />
scientifi c philosophers (Hull, 1974).<br />
Scientists typically are conservative <strong>in</strong>sofar as <strong>the</strong>y provisionally<br />
reject a null hypo<strong>the</strong>sis only if <strong>the</strong> probability<br />
of mak<strong>in</strong>g a type I error is fi ve percent or less (Cranor,<br />
1993; Lemons et al., 1997). This scientifi c conservatism<br />
is consistent with <strong>the</strong> logical positivist goal of develop<strong>in</strong>g<br />
conclusions about which <strong>the</strong>re is a high degree of confi -<br />
dence. With respect to <strong>the</strong> use of science as a basis for<br />
public policy and decision mak<strong>in</strong>g, <strong>the</strong>re are those who<br />
hold that scientifi c methods and tools are capable of yield<strong>in</strong>g<br />
<strong>in</strong>formation about which <strong>the</strong>re is a high degree of scientifi<br />
c confi dence and, <strong>the</strong>refore, it is this <strong>in</strong>formation and<br />
not more speculative <strong>in</strong>formation that should be used as<br />
<strong>the</strong> basis for policy and decision mak<strong>in</strong>g (Peters, 1991;<br />
Sunste<strong>in</strong>, 2002). This latter view is a component of <strong>the</strong><br />
fi eld of environmental and human health risk assessment,<br />
which has developed to help <strong>in</strong>form public policy and<br />
decision makers about <strong>the</strong> risks from threats from both<br />
natural phenomena and human activities, <strong>in</strong>clud<strong>in</strong>g assess<strong>in</strong>g<br />
whe<strong>the</strong>r to undertake some river restoration projects.<br />
Components of risk analysis <strong>in</strong>clude: identify<strong>in</strong>g <strong>the</strong><br />
sequence of events through which exposure to risk could<br />
occur; determ<strong>in</strong><strong>in</strong>g <strong>the</strong> number and k<strong>in</strong>ds of people or<br />
environmental resources exposed to <strong>the</strong> risk; determ<strong>in</strong><strong>in</strong>g<br />
<strong>the</strong> adverse effects of exposure to <strong>the</strong> risks; and communicat<strong>in</strong>g<br />
risk assessment fi nd<strong>in</strong>gs to decision makers and<br />
<strong>the</strong> public. Although risk assessors acknowledge scientifi c<br />
uncerta<strong>in</strong>ty, <strong>the</strong>y often hold that scientifi c methods and<br />
tools can identify <strong>the</strong> risks and enable <strong>the</strong> calculation of<br />
<strong>the</strong> probabilities of <strong>the</strong>ir occurrence, <strong>in</strong>clud<strong>in</strong>g <strong>the</strong> bound<strong>in</strong>g<br />
of <strong>the</strong> probabilities with confi dence limits. For <strong>in</strong>-<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> 5<br />
depth discussions on <strong>the</strong> role of scientifi c <strong>in</strong>formation <strong>in</strong><br />
policy and decision mak<strong>in</strong>g, see Peters (1991), Shrader-<br />
Frechette, (1994), Caldwell (1996), Lemons (1996), and<br />
Kaiser and Storvik (2003).<br />
Historically, logical positivism and its outgrowths also<br />
have <strong>in</strong>fl uenced <strong>the</strong> th<strong>in</strong>k<strong>in</strong>g of some scientists and policy<br />
makers <strong>in</strong> o<strong>the</strong>r ways by <strong>in</strong>culcat<strong>in</strong>g <strong>the</strong> view that ‘good’<br />
science is objective <strong>in</strong>sofar as it is not biased by <strong>the</strong> values<br />
of <strong>the</strong> scientists. Accord<strong>in</strong>gly, this view holds that <strong>the</strong><br />
proper role of science <strong>in</strong> policy and decision mak<strong>in</strong>g is to<br />
provide factual <strong>in</strong>formation to decision makers, and that<br />
any controversies about <strong>the</strong> factual <strong>in</strong>formation should be<br />
left to members of <strong>the</strong> scientifi c community competent<br />
<strong>in</strong> evaluat<strong>in</strong>g <strong>the</strong> scientifi c bases of <strong>the</strong> controversies<br />
(Shrader-Frechette, 1982). Consequently, <strong>the</strong> conclusions<br />
of scientifi c analyses do not become a part of broader<br />
public policy debates such as those that might perta<strong>in</strong> to<br />
such issues as what level of risk is acceptable. Practically<br />
speak<strong>in</strong>g, proponents of this view believe that <strong>the</strong> scientifi<br />
c and technical problems of manag<strong>in</strong>g large scale and<br />
complex problems are enormous and that <strong>the</strong> public cannot<br />
be expected to grasp <strong>the</strong> many scientifi c and technical<br />
issues <strong>in</strong>herent <strong>in</strong> understand<strong>in</strong>g and resolv<strong>in</strong>g <strong>the</strong> problems.<br />
Fur<strong>the</strong>r, <strong>the</strong> fundamental differences people have<br />
about how problems should be handled generate endless<br />
debate and controversy. This implies that while people and<br />
local governmental representatives with different <strong>in</strong>terests<br />
may review and comment on scientifi c and technical<br />
documents, <strong>the</strong>y would not be brought <strong>in</strong>to <strong>the</strong> actual<br />
decision- mak<strong>in</strong>g process regard<strong>in</strong>g <strong>the</strong> complex scientifi c<br />
dimensions of problems (Lemons et al., 1997).<br />
Despite <strong>the</strong> high degree of confi dence held by some<br />
people <strong>in</strong> scientifi c methods, confi dence <strong>in</strong> <strong>the</strong> power of<br />
science to understand and predict natural phenomena has<br />
been underm<strong>in</strong>ed by general relativity <strong>the</strong>ories, quantum<br />
<strong>the</strong>ories and chaos <strong>the</strong>ories (Brown, 1987). Rorty (1979)<br />
notes that <strong>the</strong>re is no evidence that science develops better<br />
and more accurate ‘mirrors’ with which to view nature. In<br />
his classic work, Kuhn (1962) describes how on <strong>the</strong> one<br />
hand <strong>the</strong> level of confi dence <strong>in</strong> models used by members<br />
of <strong>the</strong> scientifi c community <strong>in</strong>creases with evidence that<br />
supports <strong>the</strong> underly<strong>in</strong>g hypo<strong>the</strong>ses of <strong>the</strong> models, and on<br />
<strong>the</strong> o<strong>the</strong>r <strong>the</strong> scientists’ use of <strong>the</strong> models cannot be<br />
expected to produce consistently better and cumulatively<br />
more truthful descriptions of <strong>the</strong> way <strong>the</strong> world works.<br />
Accord<strong>in</strong>g to Kuhn, <strong>the</strong> reason is because predictive successes<br />
of scientifi c <strong>the</strong>ories do not guarantee <strong>the</strong>ir metaphysical<br />
accuracy because ‘paradigm shifts’ subsequently<br />
change scientists’ views of nature. O<strong>the</strong>r critics have<br />
po<strong>in</strong>ted out that so-called scientifi c truths of historical<br />
periods are social constructs <strong>in</strong>fl uenced by <strong>the</strong> dom<strong>in</strong>ant<br />
cultural and political powers of those periods (Briggs and
6 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Peat, 1982; Funtowizc and Ravetz, 1995). Some postmodern<br />
critics argue that Western science has been permeated<br />
by a variety of biases (e.g., ‘free market’ economics and<br />
<strong>in</strong>dustrialism, racism, religion, patriarchy) that while<br />
serv<strong>in</strong>g powerful <strong>in</strong>terests have not led to <strong>the</strong> generation<br />
and use of more ‘objective’ or value-free scientifi c<br />
knowledge (Sirageld<strong>in</strong>, 2002).<br />
More practically speak<strong>in</strong>g, scientifi c <strong>in</strong>stitutions as well<br />
as <strong>in</strong>dividual scientists <strong>in</strong>creas<strong>in</strong>gly hold <strong>the</strong> view that<br />
scientifi c uncerta<strong>in</strong>ty regard<strong>in</strong>g environment and human<br />
health problems is so pervasive and value laden that many<br />
conclusions about <strong>the</strong> problems cannot be made with<br />
a high degree of scientifi c confi dence (Cranor, 1993;<br />
Shrader-Frechette and McCoy, 1993; Lemons and Brown,<br />
1995; Lemons, 1996; EEA, 2001; Kriebel et al., 2001;<br />
Tickner, 2002, 2003). This view is based on empirical<br />
studies focus<strong>in</strong>g on: exposure to radiation from nuclear<br />
facilities and nuclear waste; manag<strong>in</strong>g large-scale ecosystems<br />
such as <strong>the</strong> Florida Everglades, agricultural lands,<br />
mar<strong>in</strong>e and freshwater oil spills; biodiversity protection<br />
and management of biological reserves; ocean dump<strong>in</strong>g<br />
of sewage sludge; sulfur dioxide and protection of human<br />
lungs to remote lake restoration; antifoul<strong>in</strong>g pa<strong>in</strong>ts on<br />
ships (e.g. tributylt<strong>in</strong>); estuar<strong>in</strong>e eutrophication; protection<br />
and management of mar<strong>in</strong>e fi sheries; extrapolat<strong>in</strong>g<br />
from toxicological responses <strong>in</strong> laboratory systems to both<br />
human health and to <strong>the</strong> responses of natural systems;<br />
management of fresh water resources; benzene <strong>in</strong> occupational<br />
sett<strong>in</strong>gs; <strong>the</strong> use and health impacts of asbestos;<br />
risks from polychlor<strong>in</strong>ated biphenyls (PCBs); halocarbons<br />
and <strong>the</strong> ozone layer; diethylstilbestrol (DES) and longterm<br />
consequences of prenatal exposure; human health<br />
effects of lead <strong>in</strong> <strong>the</strong> environment; methyl tertiary-butyl<br />
e<strong>the</strong>r (MBTE) <strong>in</strong> petrol as a substitute for lead; chemical<br />
contam<strong>in</strong>ation <strong>in</strong> <strong>the</strong> Great Lakes; hormones as growth<br />
promoters <strong>in</strong> animals used for food; and global climate<br />
change.<br />
1.3 WHAT IS ‘GOOD’ SCIENCE UNDER<br />
CONDITIONS OF UNCERTAINTY?<br />
Here, <strong>the</strong> question discussed is: What is ‘good’ science<br />
when science is used <strong>in</strong> try<strong>in</strong>g to solve river restoration<br />
problems under conditions of scientifi c uncerta<strong>in</strong>ty?<br />
A traditional and commonly accepted goal of science is<br />
that <strong>the</strong> probabilities of add<strong>in</strong>g speculative <strong>in</strong>formation to<br />
<strong>the</strong> body of scientifi c knowledge should be m<strong>in</strong>imal (Hull,<br />
1974; Peters, 1991). For this reason, scientists typically<br />
are conservative <strong>in</strong>sofar as <strong>the</strong>y provisionally reject a null<br />
hypo<strong>the</strong>sis if <strong>the</strong>re is a fi ve percent or less chance of<br />
reject<strong>in</strong>g it when it is true; this criterion is known as a<br />
normal standard of scientifi c proof or so-called ‘n<strong>in</strong>ety-<br />
fi ve percent confi dence rule.’ With respect to <strong>the</strong> science<br />
used to <strong>in</strong>form certa<strong>in</strong> types of river restoration policies<br />
and decisions, an example of a null hypo<strong>the</strong>sis is that <strong>the</strong>re<br />
is no effect on rivers or <strong>the</strong>ir resources from exist<strong>in</strong>g or<br />
proposed human activities. A type I error is to accept a<br />
false positive result, that is, to conclude that <strong>the</strong>re is harm<br />
to rivers or <strong>the</strong>ir resources when <strong>in</strong> fact <strong>the</strong>re is none. A<br />
type II error is to accept a false negative result, that is, to<br />
conclude <strong>the</strong>re is no harm when <strong>in</strong> fact <strong>the</strong>re is.<br />
Many environmental laws and regulations place <strong>the</strong><br />
burden of proof for demonstrat<strong>in</strong>g harm to <strong>the</strong> environment<br />
or human health on government regulatory agencies<br />
or o<strong>the</strong>rs attempt<strong>in</strong>g to demonstrate harm from development<br />
activities and, often, <strong>the</strong> standard that is used to meet<br />
<strong>the</strong> burden of proof test is <strong>the</strong> normal standard of scientifi c<br />
proof (Brown, 1995). When this standard is adopted as a<br />
basis for environmental decisions <strong>the</strong> scientifi c uncerta<strong>in</strong>ty<br />
that pervades many environmental problems means<br />
that <strong>the</strong> burden of proof usually will not be met, despite<br />
<strong>the</strong> fact that some <strong>in</strong>formation or even <strong>the</strong> weight of evidence<br />
might <strong>in</strong>dicate <strong>the</strong> existence of harm to <strong>the</strong> environment<br />
or human health. Consequently, <strong>in</strong> public policy and<br />
decision mak<strong>in</strong>g if <strong>the</strong> data show that some factor or perturbation<br />
has had an effect on <strong>the</strong> environment or human<br />
health but, say, only at <strong>the</strong> 70–90% confi dence level <strong>the</strong><br />
null hypo<strong>the</strong>sis that <strong>the</strong>re is no effect from <strong>the</strong> factor or<br />
perturbation is accepted. In such cases <strong>the</strong>re is a tendency<br />
by decision makers and o<strong>the</strong>rs to assume not only that<br />
<strong>the</strong>re was not enough evidence to reject <strong>the</strong> null hypo<strong>the</strong>sis<br />
but that <strong>the</strong>re was no effect when, <strong>in</strong> fact, <strong>the</strong> experimental<br />
design or test could have been too weak or <strong>the</strong> data too<br />
variable or too close for an effect to be demonstrated even<br />
if <strong>the</strong>re had been one (a type II error).<br />
M<strong>in</strong>imiz<strong>in</strong>g a type II error requires <strong>the</strong> statistical power<br />
of a research design or hypo<strong>the</strong>sis test to be calculated. In<br />
contrast to confi dence, which is designed to m<strong>in</strong>imize type<br />
I error, power depends on <strong>the</strong> magnitude of <strong>the</strong> hypo<strong>the</strong>sized<br />
change to be detected, <strong>the</strong> sample variance, <strong>the</strong><br />
number of replicates and <strong>the</strong> signifi cance value. The power<br />
of a test is <strong>the</strong> probability of reject<strong>in</strong>g a null hypo<strong>the</strong>sis<br />
when it is <strong>in</strong> fact false and should be rejected. The larger<br />
<strong>the</strong> detected change, <strong>the</strong> larger is <strong>the</strong> power. In situations<br />
where <strong>the</strong> detected changes are relatively small, statistical<br />
power is <strong>in</strong>creased by <strong>in</strong>creased sampl<strong>in</strong>g size but this<br />
<strong>in</strong>volves additional costs, research facilities and time.<br />
Analysis of variance <strong>in</strong> assess<strong>in</strong>g threats to environmental<br />
and human health problems shows that <strong>the</strong> number of<br />
samples required to yield a power of 0.95 <strong>in</strong>creases rapidly<br />
if changes smaller than 50% of <strong>the</strong> standard deviation are<br />
to be detected (Cranor, 1993). If <strong>the</strong> sample size stays<br />
<strong>the</strong> same <strong>the</strong> probability of a type I error is <strong>in</strong>creased if<br />
<strong>the</strong> probability of a type II error is decreased. A practical
problem <strong>in</strong> river restoration is that a desired emphasis on<br />
avoid<strong>in</strong>g type II error must be balanced aga<strong>in</strong>st o<strong>the</strong>r<br />
opportunities to use limited scientifi c resources to address<br />
o<strong>the</strong>r environmental and human health problems.<br />
Decisions about water management <strong>in</strong> <strong>the</strong> Klamath<br />
Bas<strong>in</strong> along <strong>the</strong> California and Oregon border <strong>in</strong> <strong>the</strong><br />
United States show some of <strong>the</strong> types of consequences that<br />
can happen when <strong>the</strong> law or decision makers require <strong>the</strong><br />
use of scientifi c <strong>in</strong>formation that meets <strong>the</strong> normal standard<br />
of scientifi c proof. In decades long disputes about<br />
water management <strong>in</strong> <strong>the</strong> bas<strong>in</strong>, federal biologists have<br />
been try<strong>in</strong>g to save three species of endangered fi sh by<br />
call<strong>in</strong>g for diversions of water from irrigation <strong>in</strong>to <strong>the</strong><br />
bas<strong>in</strong> to reduce <strong>the</strong> frequency of fi sh kills dur<strong>in</strong>g low water<br />
periods (over 30 000 Ch<strong>in</strong>ook salmon died dur<strong>in</strong>g a fi sh<br />
kill <strong>in</strong> 2002) (Service, 2003). As would be expected, a<br />
recommendation to reduce <strong>the</strong> amount of water available<br />
for irrigation met with strong opposition by ranchers and<br />
farmers <strong>in</strong> <strong>the</strong> bas<strong>in</strong>. However, failure of <strong>the</strong> biologists to<br />
meet normal scientifi c standards of proof demonstrat<strong>in</strong>g<br />
that releas<strong>in</strong>g more water <strong>in</strong>to <strong>the</strong> bas<strong>in</strong> would help <strong>the</strong><br />
fi sh has been cited by <strong>the</strong> United States Department of<br />
Interior (DOI) <strong>in</strong> its recent refusal to restrict <strong>the</strong> amount<br />
of water farmers can remove from waterways <strong>in</strong> <strong>the</strong> bas<strong>in</strong><br />
(NRC, 2004). It is important to understand that <strong>the</strong> DOI<br />
was not criticiz<strong>in</strong>g <strong>the</strong> scientists for do<strong>in</strong>g poor science;<br />
ra<strong>the</strong>r, it concluded that <strong>the</strong> normal standard of proof was<br />
not met. The DOI noted that factors such as nutrient runoff<br />
from natural sources as well as farms and ranches, algae<br />
blooms and dams that restrict access to fi shes’ spawn<strong>in</strong>g<br />
grounds complicate and <strong>in</strong> fact might preclude demonstrat<strong>in</strong>g<br />
<strong>the</strong> relation of water fl ow <strong>in</strong>to <strong>the</strong> bas<strong>in</strong> and <strong>the</strong><br />
health of <strong>the</strong> fi sh populations with a higher degree of<br />
scientifi c confi dence.<br />
The question of how to protect endangered species <strong>in</strong><br />
<strong>the</strong> Klamath Bas<strong>in</strong> and manage water resources raises a<br />
fundamental dilemma that those <strong>in</strong>volved <strong>in</strong> river restoration<br />
have to confront. On <strong>the</strong> one hand, traditional scientifi<br />
c norms call for mak<strong>in</strong>g conclusions on <strong>in</strong>formation<br />
about which <strong>the</strong>re is a high degree of confi dence. In <strong>the</strong><br />
Klamath Bas<strong>in</strong> example, adher<strong>in</strong>g to traditional scientifi c<br />
norms constra<strong>in</strong>s decisions to protect endangered fi sh<br />
under conditions of uncerta<strong>in</strong>ty but, at <strong>the</strong> same time, <strong>in</strong><br />
<strong>the</strong> absence of decisions to protect endangered fi sh <strong>the</strong><br />
threats cont<strong>in</strong>ue. In this type of situation, when science is<br />
used for public policy and decision mak<strong>in</strong>g, scientists<br />
might wish to consider whe<strong>the</strong>r and to what extent <strong>the</strong>y<br />
should be more comfortable with mak<strong>in</strong>g conclusions<br />
based on <strong>the</strong> weight of evidence ra<strong>the</strong>r than based solely<br />
or primarily on high levels of confi dence, especially s<strong>in</strong>ce<br />
public policy decisions are not based simply upon probabilistic<br />
considerations but ra<strong>the</strong>r <strong>in</strong>volve mak<strong>in</strong>g discrete<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> 7<br />
and explicit choices among specifi c alternatives, <strong>in</strong>clud<strong>in</strong>g<br />
those with political, economic and ethical ramifi cations<br />
(Bella et al., 1994; Lemons et al., 1997). Admittedly, this<br />
could create a tension between do<strong>in</strong>g ‘good’ science as<br />
traditionally defi ned because scientists would be mak<strong>in</strong>g<br />
more speculative conclusions; however, <strong>in</strong> <strong>the</strong>ir attempt to<br />
make science rigorous <strong>in</strong> <strong>the</strong> sense of not want<strong>in</strong>g to add<br />
speculation to <strong>the</strong> body of scientifi c knowledge as required<br />
by <strong>the</strong> scientifi c profession <strong>the</strong> regulatory questions for<br />
which <strong>the</strong> studies are done may be frustrated.<br />
1.4 VALUE-LADEN DIMENSIONS OF SCIENCE<br />
AND UNCERTAINTY<br />
In addition to <strong>the</strong> policy and management problems that<br />
arise from <strong>the</strong> use of traditional scientifi c norms for<br />
mak<strong>in</strong>g conclusions <strong>in</strong> river restoration, o<strong>the</strong>r value-laden<br />
dimensions of science and policy both contribute to uncerta<strong>in</strong>ty<br />
and raise complicated questions about how it should<br />
be handled <strong>in</strong> public policy.<br />
Westra and Lemons (1995) and Lemons (1996) conta<strong>in</strong><br />
papers analyz<strong>in</strong>g both philosophical and scientifi c concepts<br />
used to <strong>in</strong>form ecological restoration science and<br />
practice. The concepts are diverse and <strong>in</strong>clude bas<strong>in</strong>g restoration<br />
on: ecosystems’ abilities to function successfully<br />
<strong>in</strong> a way deemed satisfactory by society; ecosystems’<br />
abilities to ma<strong>in</strong>ta<strong>in</strong> a balanced, <strong>in</strong>tegrated, adaptive<br />
community of organisms hav<strong>in</strong>g species composition,<br />
diversity and functional organization comparable to that<br />
of ‘natural’ habits of <strong>the</strong> region; ecosystems’ abilities to<br />
regenerate <strong>the</strong>mselves and withstand anthropogenic stress;<br />
and ecosystems’ abilities to approach optimum capacity<br />
for ecological succession development options. One<br />
problem with all <strong>the</strong>se defi nitions is that <strong>the</strong>y are <strong>in</strong>complete,<br />
general and qualitative <strong>in</strong>sofar as <strong>the</strong>y fail to provide<br />
precise pr<strong>in</strong>ciples that would make <strong>the</strong>m operational.<br />
In his analysis of value-laden issues <strong>in</strong> restoration<br />
for ecological as opposed to primarily or exclusively<br />
economic development goals, Cairns (2003) focuses on<br />
several types of problems. Firstly, some restoration projects<br />
are carried out on habitats different <strong>in</strong> k<strong>in</strong>d from those<br />
altered or destroyed. For example, an upland forest may<br />
be destroyed <strong>in</strong> order to partially restore river systems and<br />
wetlands that once occupied a particular lowland area.<br />
Despite <strong>the</strong> fact that restoration of rivers and/or wetlands<br />
has ecological value, sacrifi c<strong>in</strong>g a relatively undamaged<br />
habitat to restore ano<strong>the</strong>r k<strong>in</strong>d may cause unanticipated<br />
ecological change or harm. Secondly, with few exceptions<br />
most river and o<strong>the</strong>r ecological restoration projects are<br />
done to support <strong>the</strong> anthropocentric commodity or utilitarian<br />
values <strong>the</strong>y offer humans and this poses confl icts with<br />
restoration goals for nonanthropcentric reasons. Thirdly,
8 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
river restoration has uncerta<strong>in</strong> outcomes because of unpredictable<br />
events like fl oods or droughts, and because of <strong>the</strong><br />
limitations of <strong>the</strong> methods and tools of science to predict<br />
long-term outcomes. Fourthly, restoration efforts focus<strong>in</strong>g<br />
on s<strong>in</strong>gle species or ecosystem attributes might elim<strong>in</strong>ate<br />
those species that had <strong>in</strong>itially colonized disturbed areas<br />
and were at <strong>the</strong> same time able to tolerate anthropocentric<br />
stress. However, restoration projects might result <strong>in</strong> <strong>the</strong><br />
displacement of species tolerant to human activities with<br />
those less tolerant, at least <strong>in</strong> <strong>the</strong> short term. Fifthly, ecological<br />
restoration often takes place with species that<br />
tolerate anthropocentric stress and <strong>the</strong> ultimate succession<br />
processes and states will be human dom<strong>in</strong>ated or dependent.<br />
Most likely, a return to <strong>in</strong>digenous species would<br />
require cont<strong>in</strong>ual <strong>in</strong>tervention by researchers and<br />
environmental decision makers on behalf of <strong>the</strong>ir reestablishment.<br />
While science is not determ<strong>in</strong>ative to how<br />
<strong>the</strong> issues are resolved, robust scientifi c <strong>in</strong>formation is<br />
needed to help <strong>in</strong>form satisfactory policy judgments.<br />
Mayo and Hollander (1991), Cranor (1993), Shrader-<br />
Frechette and McCoy (1993) and Lemons and Brown<br />
(1995) analyzed how and why numerous value-laden<br />
judgments, evaluations, assumptions and <strong>in</strong>ferences are<br />
embedded <strong>in</strong> scientifi c methods perta<strong>in</strong><strong>in</strong>g to <strong>the</strong> study<br />
and management of ecosystems, <strong>in</strong>clud<strong>in</strong>g geohydrological<br />
and o<strong>the</strong>r water resources. For example, people have<br />
to decide <strong>the</strong> ecosystem parameters that are more important<br />
to base judgments on, often with little or no empirical<br />
<strong>in</strong>formation available. Assumptions have to be made, often<br />
without direct empirical evidence, whe<strong>the</strong>r ecosystem<br />
parameters should be considered <strong>in</strong>dependently or synergistically,<br />
and whe<strong>the</strong>r threshold values for environmental<br />
or health impacts exist and, if so, what such values should<br />
be. In addition, a lack of empirical data cannot be separated<br />
entirely from practical limitations imposed on environmental<br />
scientists. Decision makers require <strong>in</strong>formation<br />
<strong>in</strong> a relatively short period and at reasonable cost. These<br />
factors constra<strong>in</strong> <strong>the</strong> focus of most restoration studies to<br />
<strong>the</strong> short term, relatively small spatial areas and measurement<br />
of a relatively small number of samples and parameters.<br />
Fur<strong>the</strong>r, <strong>the</strong> above commentators conclude that<br />
many of <strong>the</strong> value-laden dimensions of scientifi c methodology<br />
and <strong>in</strong>formation not only are not fully recognized<br />
by scientists, policy and decision makers, but that <strong>the</strong><br />
failure to suffi ciently recognize <strong>the</strong> value-laden dimensions<br />
of science casts serious doubts about even <strong>the</strong> best<br />
and most thorough scientifi c and technical studies used to<br />
<strong>in</strong>form decisions about problems such as river restoration.<br />
In o<strong>the</strong>r words, unless <strong>the</strong> value-laden dimensions of scientifi<br />
c studies are disclosed <strong>the</strong> positions of decision<br />
makers will appear to be justifi ed on value-neutral scientifi<br />
c reason<strong>in</strong>g and will appear to be more certa<strong>in</strong> than<br />
warranted when, <strong>in</strong> fact, <strong>the</strong> positions will be based, <strong>in</strong><br />
part, on often controversial and confl ict<strong>in</strong>g values of scientists<br />
and decision makers (see also Fleck, 1979).<br />
One of <strong>the</strong> most common ways <strong>in</strong> which value issues<br />
are hidden <strong>in</strong> public policy concern<strong>in</strong>g issues such as river<br />
restoration develops out of <strong>the</strong> expectation that technical<br />
analysts can isolate and apply <strong>the</strong> facts under dispute <strong>in</strong><br />
a manner consistent with policy directives or legislative<br />
mandates. This separation of facts and values is highly<br />
problematic. For example, consider <strong>the</strong> use of safety<br />
factors <strong>in</strong> river water quality regulations as a means of<br />
extra protection for human or environmental health.<br />
Implicit <strong>in</strong> <strong>the</strong> choice of safety factors is an asymmetric<br />
cost function with health costs ris<strong>in</strong>g more steeply than<br />
costs for over-treatment. Implicit <strong>in</strong> <strong>the</strong> magnitude of a<br />
safety factor are signifi cant uncerta<strong>in</strong>ties <strong>in</strong> health impacts<br />
and a steeper cost function for health effects from undertreatment<br />
than for over-treatment. When <strong>the</strong>se issues<br />
rema<strong>in</strong> implicit <strong>in</strong> <strong>the</strong> use of safety factors (as <strong>the</strong>y typically<br />
are) <strong>the</strong> real issues of knowledge and uncerta<strong>in</strong>ty are<br />
obscured for decision makers and <strong>the</strong> public. Often, <strong>the</strong>se<br />
issues rema<strong>in</strong> implicit or hidden because safety factors<br />
and cost factors are described <strong>in</strong> quantitative terms perta<strong>in</strong><strong>in</strong>g<br />
to risks or cost–benefi t calculations. This <strong>in</strong>creases<br />
<strong>the</strong> likelihood of <strong>the</strong> misuse of conclusions by decision<br />
makers who do not understand <strong>the</strong> basis for deriv<strong>in</strong>g safety<br />
factors (Brown, 1987).<br />
1.5 PRACTICAL AND POLICY ASPECTS<br />
OF UNCERTAINTY<br />
Cairns (2001) analyzed how most complex environmental<br />
problems transcend <strong>the</strong> capabilities of any s<strong>in</strong>gle discipl<strong>in</strong>e<br />
but at <strong>the</strong> same time and all too often research teams<br />
are not suffi ciently <strong>in</strong>terdiscipl<strong>in</strong>ary to deal adequately<br />
with <strong>the</strong> problems. In addition, problem solv<strong>in</strong>g often does<br />
not provide a balanced mix of academicians, public policy<br />
and decision makers, representatives from private <strong>in</strong>dustry<br />
or bus<strong>in</strong>ess and nongovernmental organizations. As a<br />
result, <strong>the</strong> fram<strong>in</strong>g of problems and <strong>the</strong>ir solution is too<br />
often fragmented and <strong>in</strong>effectual and biased towards one<br />
or a few discipl<strong>in</strong>ary approaches or stakeholder groups<br />
(Nienhuis and Leuven, 2001; Benyam<strong>in</strong>e, 2002).<br />
Some scientists and policy makers <strong>in</strong>volved <strong>in</strong> environmental<br />
problem solv<strong>in</strong>g have argued for syn<strong>the</strong>siz<strong>in</strong>g<br />
analyses and alternatives to solutions of environmental<br />
resource problems (Lubchenco et al., 1991; Bella et al.,<br />
1994; Lemons and Brown, 1995; Caldwell, 1996). In practice,<br />
at least three levels of syn<strong>the</strong>sis may be identifi ed.<br />
The fi rst is conceptual syn<strong>the</strong>sis and occurs when <strong>the</strong><br />
diverse and often disparate elements of a problem situation<br />
are pulled toge<strong>the</strong>r <strong>in</strong>tuitively, <strong>the</strong>n tested and <strong>in</strong>tegrated
to form a coherent research design. Follow<strong>in</strong>g analysis of<br />
<strong>the</strong> problem and identifi cation of its causes and consequences,<br />
a second level of syn<strong>the</strong>sis <strong>in</strong>volves del<strong>in</strong>eation<br />
of <strong>the</strong> fi nd<strong>in</strong>gs of <strong>the</strong> scientifi c research. A third level of<br />
syn<strong>the</strong>sis can occur when research fi nd<strong>in</strong>gs are evaluated<br />
and consolidated <strong>in</strong> decid<strong>in</strong>g a course of action by<br />
decision makers.<br />
Despite <strong>the</strong> need for greater syn<strong>the</strong>sis of research<br />
methods and <strong>in</strong>formation, syn<strong>the</strong>sis itself <strong>in</strong>troduces additional<br />
value-laden dimensions and uncerta<strong>in</strong>ties <strong>in</strong>to environmental<br />
problem solv<strong>in</strong>g. Caldwell (1996) and Brown<br />
(1995) discuss how decision makers must syn<strong>the</strong>size a<br />
policy (<strong>in</strong> part) from <strong>the</strong> scientifi c <strong>in</strong>formation available<br />
even when <strong>the</strong> <strong>in</strong>formation often is <strong>in</strong>complete. When<br />
science is used to <strong>in</strong>form policy decisions such decisions<br />
also <strong>in</strong>clude economic, legal, adm<strong>in</strong>istrative and cultural<br />
parameters and, <strong>the</strong>refore, are based on human values<br />
and judgments. Benyam<strong>in</strong>e (2002) discusses how disagreements<br />
about scientifi c <strong>the</strong>ories that are used as a<br />
basis for <strong>in</strong>form<strong>in</strong>g public policy and decision mak<strong>in</strong>g<br />
become entangled with economic, legal and ideological<br />
issues. Sometimes, <strong>the</strong> disagreements rema<strong>in</strong> largely confi<br />
ned to <strong>the</strong> scientifi c community, while at o<strong>the</strong>r times <strong>the</strong><br />
public knows about <strong>the</strong>m. When scientists and/or decision<br />
makers know <strong>the</strong> underly<strong>in</strong>g <strong>the</strong>oretical bases for disagreements,<br />
this knowledge can <strong>in</strong>fl uence <strong>the</strong> scientifi c<br />
arguments about <strong>the</strong> disagreements. However, some confl<br />
ict<strong>in</strong>g arguments and <strong>the</strong>ir underly<strong>in</strong>g <strong>the</strong>oretical support<br />
can be under recognized or little understood by <strong>the</strong> nonscientifi<br />
c communities as well as by scientists whose specialized<br />
fi elds are outside <strong>the</strong> discipl<strong>in</strong>e where debates<br />
about <strong>the</strong>ories are tak<strong>in</strong>g place. When this happens, confl<br />
ict<strong>in</strong>g scientifi c arguments will not have much <strong>in</strong>fl uence<br />
on <strong>the</strong> disagreements.<br />
There is debate with<strong>in</strong> <strong>the</strong> scientifi c and public policy<br />
communities regard<strong>in</strong>g approaches to deal with uncerta<strong>in</strong>ties<br />
(Bradshaw and Borchers, 2000). For example, one<br />
approach might be to attempt to <strong>in</strong>crease scientifi c confi -<br />
dence by <strong>in</strong>creas<strong>in</strong>g scientifi c confi rmation of hypo<strong>the</strong>ses.<br />
In this way, scientists can decrease uncerta<strong>in</strong>ty suffi ciently<br />
to allow more precise estimates of risk for policy and<br />
decision makers. A second approach might be to <strong>in</strong>crease<br />
<strong>the</strong> knowledge of sources of uncerta<strong>in</strong>ty by enhanc<strong>in</strong>g<br />
education and communication between scientists, policy<br />
and decision makers and <strong>the</strong> general public. A benefi t of<br />
this approach is that when scientists and decision makers<br />
are <strong>in</strong>volved with <strong>the</strong> public <strong>the</strong>re is greater opportunity<br />
for consensus build<strong>in</strong>g and less risk of legal challenges<br />
from disaffected stakeholders. A third approach might be<br />
to foster <strong>the</strong> view that scientifi c uncerta<strong>in</strong>ty should be<br />
regarded <strong>in</strong> public policy and decision mak<strong>in</strong>g as it is<br />
with<strong>in</strong> <strong>the</strong> scientifi c community, namely, as <strong>in</strong>formation<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> 9<br />
for hypo<strong>the</strong>sis build<strong>in</strong>g and test<strong>in</strong>g. Consequently, calls<br />
for faster and more ‘certa<strong>in</strong>’ scientifi c conclusions to<br />
<strong>in</strong>form public policy and decision mak<strong>in</strong>g would be tempered<br />
with a better understand<strong>in</strong>g of <strong>the</strong> limitations and<br />
capabilities of science to provide <strong>in</strong>formation about which<br />
<strong>the</strong>re is a high degree of confi dence.<br />
Still ano<strong>the</strong>r approach might be for society to require<br />
procedural rules for mak<strong>in</strong>g decisions under conditions<br />
of scientifi c uncerta<strong>in</strong>ty to take <strong>in</strong>to account confl ict<strong>in</strong>g<br />
po<strong>in</strong>ts of view, possible consequences to welfare, as well<br />
as various ethical and legal obligations such as those<br />
<strong>in</strong>volv<strong>in</strong>g free <strong>in</strong>formed consent and due process (Shrader-<br />
Frechette, 1996). This approach could <strong>in</strong>clude greater use<br />
of <strong>the</strong> precautionary pr<strong>in</strong>ciple by help<strong>in</strong>g to ensure that<br />
when <strong>the</strong>re is substantial scientifi c uncerta<strong>in</strong>ty about <strong>the</strong><br />
risks and benefi ts of a proposed activity, policy decisions<br />
should be made <strong>in</strong> a way that errs on <strong>the</strong> side of caution<br />
with respect to <strong>the</strong> environment and <strong>the</strong> health of <strong>the</strong><br />
public (Kriebel et al., 2001; Tickner, 2003).<br />
1.6 CASE STUDY OF SCIENTIFIC<br />
UNCERTAINTY IN RIVER RESTORATION<br />
The example discussed here is based on ecological studies<br />
conducted from 1980–1989 <strong>in</strong> a small (4th order), black<br />
water West African river, <strong>the</strong> <strong>River</strong> Ikpoba fl ow<strong>in</strong>g through<br />
Ben<strong>in</strong> City, Sou<strong>the</strong>rn Nigeria (Victor and Dickson, 1985;<br />
Victor and Ogbeibu, 1985, 1986, 1991; Victor and Tetteh,<br />
1988; Ogbeibu and Victor, 1989; Victor and Brown, 1990;<br />
Victor and Meye, 1994; Victor and Onomivbori, 1996;<br />
Victor, 1998). The stretch of river studied was affected by<br />
a variety of urban perturbations such as damm<strong>in</strong>g, water<br />
extraction, po<strong>in</strong>t and nonpo<strong>in</strong>t source pollution, sand<br />
dredg<strong>in</strong>g and agriculture. As a result of government<br />
policies and directives mandat<strong>in</strong>g river clean-up activities,<br />
<strong>the</strong>re was a rare opportunity to study river restoration by<br />
recovery processes. Scientifi c results of this study were<br />
published <strong>in</strong> <strong>the</strong> series of publications listed above and<br />
provide one of <strong>the</strong> bases of our focus on uncerta<strong>in</strong>ties<br />
associated with <strong>the</strong> restoration process.<br />
The fi rst logical step was to <strong>in</strong>vestigate recovery processes.<br />
Geomorphologic changes of <strong>the</strong> river channel and<br />
<strong>the</strong> entire riparian corridor <strong>in</strong>fl uenced by urban development<br />
could not be reversed (e.g. <strong>the</strong> presence of a dam,<br />
water extraction for human consumption) and <strong>the</strong>refore<br />
complete restoration would not be possible. Removal of<br />
human <strong>in</strong>fl uences where possible would permit recovery,<br />
but <strong>the</strong> rates limit<strong>in</strong>g recovery <strong>in</strong> different sections would<br />
not only depend on <strong>the</strong> type of <strong>in</strong>fl uence (e.g. sand extraction,<br />
car wash<strong>in</strong>g), but would also be complicated by<br />
natural events such as fl oods. Thus <strong>the</strong> optimum threshold<br />
for <strong>the</strong> recovery process <strong>in</strong> this study at various sections
10 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
of <strong>the</strong> river cont<strong>in</strong>uum was unpredictable and uncerta<strong>in</strong>.<br />
O<strong>the</strong>r signifi cant uncerta<strong>in</strong>ties were: <strong>the</strong> role of early<br />
recoloniz<strong>in</strong>g species affect<strong>in</strong>g <strong>the</strong> trajectory of recovery;<br />
<strong>the</strong> successional sequence of species re-establish<strong>in</strong>g; and<br />
<strong>the</strong> establishment of appropriate abiotic conditions and<br />
<strong>the</strong> establishment of previously non-exist<strong>in</strong>g non-native<br />
species like <strong>the</strong> water hyac<strong>in</strong>th.<br />
The next group of uncerta<strong>in</strong>ties was related to <strong>the</strong> analysis<br />
and syn<strong>the</strong>sis of data. Removal of a particular human<br />
<strong>in</strong>fl uence (e.g. discharge of untreated sewage) <strong>in</strong> one<br />
section signifi cantly <strong>in</strong>creased <strong>the</strong> presence of a parameter,<br />
say i (P < 0.05), show<strong>in</strong>g that this parameter was a good<br />
<strong>in</strong>dicator of recovery. But <strong>the</strong> same parameter did not<br />
<strong>in</strong>crease signifi cantly <strong>in</strong> an adjacent section with a similar<br />
problem (P > 0.05) show<strong>in</strong>g its uncerta<strong>in</strong> predictive status.<br />
Graphical exam<strong>in</strong>ation of associations between specifi c<br />
human <strong>in</strong>fl uences (e.g. removal of detergent contam<strong>in</strong>ation)<br />
and biological parameters like taxa richness and<br />
abundance showed positive relationships, but statistically<br />
<strong>the</strong>se relationships as evaluated by Pearson’s r or Spearman’s<br />
r s were not signifi cant (P > 0.05). Thus, correlation<br />
matrices generated for evaluat<strong>in</strong>g relationships between<br />
<strong>the</strong> removal of perturbation <strong>in</strong>fl uences and <strong>the</strong> recovery of<br />
both biotic and abiotic parameters were diffi cult to <strong>in</strong>terpret.<br />
Interpretation us<strong>in</strong>g traditional statistical norms and<br />
acceptable levels of signifi cance were ecologically and<br />
rationally highly problematic.<br />
Fur<strong>the</strong>r uncerta<strong>in</strong>ties arose while consider<strong>in</strong>g <strong>the</strong> temporal<br />
and spatial scale of <strong>the</strong> recovery process. The recovery<br />
process was happen<strong>in</strong>g <strong>in</strong> an urban sett<strong>in</strong>g with a new<br />
land use matrix, far different from prist<strong>in</strong>e or semiprist<strong>in</strong>e<br />
natural conditions that previously existed. Therefore, comparison<br />
of <strong>the</strong> restored river sections to that of ‘undisturbed’<br />
sections upstream was not valid and new basel<strong>in</strong>e<br />
standards had to be established for future monitor<strong>in</strong>g.<br />
Even <strong>the</strong>se were extremely site specifi c with very limited<br />
potential for use <strong>in</strong> o<strong>the</strong>r sections of <strong>the</strong> study stretch.<br />
Because of <strong>the</strong> uncerta<strong>in</strong>ties <strong>in</strong>volved, <strong>the</strong> scale needed for<br />
manag<strong>in</strong>g temporal and spatial variability <strong>in</strong> restoration<br />
was not apparent. ‘Rules of thumb’ based on value judgments<br />
had to be made to evaluate recovery <strong>in</strong> specifi c<br />
sections of <strong>the</strong> river stretch with specifi c types of perturbations.<br />
The magnitude of uncerta<strong>in</strong>ties <strong>in</strong>volved render <strong>the</strong><br />
comb<strong>in</strong>ation of tools used here (e.g. sampl<strong>in</strong>g duration,<br />
sampl<strong>in</strong>g frequencies, choice of methods, size of samples,<br />
analytical models) <strong>in</strong>adequate to evaluate recovery processes<br />
<strong>in</strong> o<strong>the</strong>r rivers of similar stream order, larger rivers<br />
with higher stream order and even <strong>the</strong> same river 100 km<br />
downstream where its stream order is >8.<br />
Implementation and analysis of monitor<strong>in</strong>g were also<br />
wrought with uncerta<strong>in</strong>ties. For example, fi ve different<br />
sections of <strong>the</strong> river stretch were monitored for restoration<br />
by recovery. Each section was characterized by its own set<br />
of physical and biological parameters that were good<br />
<strong>in</strong>dicators of recovery at <strong>the</strong> time of <strong>the</strong> study. Due to<br />
limitations of fund<strong>in</strong>g, personnel and <strong>the</strong> required cost<br />
effectiveness of <strong>the</strong> monitor<strong>in</strong>g program, proposals had<br />
to identify common parameters that would monitor <strong>the</strong><br />
overall health of <strong>the</strong> study stretch <strong>in</strong> <strong>the</strong> long term. As<br />
discussed earlier, uncerta<strong>in</strong>ties associated with <strong>the</strong> analysis<br />
and syn<strong>the</strong>sis of data did not permit <strong>the</strong> ready identifi<br />
cation of common parameters. Even if <strong>the</strong>re was an<br />
agreement on us<strong>in</strong>g different sets of parameters for different<br />
sections of <strong>the</strong> stretch, <strong>the</strong>re was no certa<strong>in</strong>ty that <strong>the</strong>se<br />
parameters (e.g. BOD, nitrate–N, fi sh diversity) will cont<strong>in</strong>ue<br />
to serve as good <strong>in</strong>dicators of recovery <strong>in</strong> <strong>the</strong> long<br />
term. It was also possible that a parameter considered<br />
trivial and not <strong>in</strong>cluded <strong>in</strong> <strong>the</strong> monitor<strong>in</strong>g program (e.g.<br />
dissolved organic matter, haptobenthos) may become<br />
important <strong>in</strong> <strong>the</strong> long term, which <strong>in</strong> itself cannot be<br />
defi ned clearly. ‘Long term’ <strong>in</strong> this case at least did not<br />
refer to an <strong>in</strong>defi nite period and envisaged monitor<strong>in</strong>g<br />
programs were not relatively open-ended, as often is <strong>the</strong><br />
case <strong>in</strong> countries with limited resources. Policy and decision<br />
makers considered what seemed to be a comprehensive<br />
proposal for monitor<strong>in</strong>g <strong>in</strong> <strong>the</strong> view of scientists as<br />
not be<strong>in</strong>g practical.<br />
Policy questions concern<strong>in</strong>g river restoration <strong>in</strong> <strong>the</strong><br />
geopolitical context were plagued with more uncerta<strong>in</strong>ties<br />
than scientifi c questions. The political climate of <strong>the</strong> study<br />
area at that time was unstable and government changed<br />
hands frequently. For example, one government downgraded<br />
<strong>the</strong> priority given to environmental issues, such as<br />
river restoration, by <strong>the</strong> previous government if personal<br />
<strong>in</strong>terests and political expediency demanded it. Assum<strong>in</strong>g<br />
no change <strong>in</strong> policies with change <strong>in</strong> governments, <strong>the</strong>re<br />
were uncerta<strong>in</strong>ties concern<strong>in</strong>g fund<strong>in</strong>g tools that would<br />
ensure <strong>the</strong> long term success of restoration, design of<br />
legislation to accommodate river restoration without compromis<strong>in</strong>g<br />
susta<strong>in</strong>able development and coord<strong>in</strong>ation of<br />
policies and legislation to devise strategies for river restoration<br />
<strong>in</strong> a broader context of <strong>the</strong> adm<strong>in</strong>istrative region<br />
(e.g. district, state, country). The management of restored<br />
or recovered river as a water resource for domestic use,<br />
agriculture, fi sheries and recreation was not considered<br />
<strong>in</strong>tentionally. For scientifi c uncerta<strong>in</strong>ty concern<strong>in</strong>g water<br />
resources management, see Canter (1996).<br />
1.7 CONCLUSION<br />
Scientifi c and o<strong>the</strong>r uncerta<strong>in</strong>ty is pervasive <strong>in</strong> environmental<br />
problem solv<strong>in</strong>g, and river restoration is no exception.<br />
When <strong>the</strong> traditional scientifi c standard of proof is<br />
used as a basis for river restoration decisions, <strong>the</strong> scientifi c
uncerta<strong>in</strong>ty that pervades many restoration problems<br />
means that <strong>the</strong> standard usually will not be met, despite<br />
<strong>the</strong> fact that some <strong>in</strong>formation or even <strong>the</strong> weight of evidence<br />
might <strong>in</strong>dicate <strong>the</strong> existence of harm and <strong>the</strong>refore<br />
<strong>the</strong> need for restoration. A high degree of confi dence <strong>in</strong><br />
river restoration science, as <strong>in</strong> o<strong>the</strong>r sciences, unfortunately<br />
seems to h<strong>in</strong>ge on conventional statistical decision<br />
rules such as when, for example, river monitor<strong>in</strong>g dur<strong>in</strong>g<br />
restoration strives to detect human-<strong>in</strong>fl uenced factors<br />
that caused deviations from basel<strong>in</strong>e conditions. The major<br />
concern here will be ecological change and not how large<br />
or small <strong>the</strong> P-values are (Yoccoz, 1991; Stewart-Oaten,<br />
1996). Most statistical decision rules are too simplistic and<br />
mislead<strong>in</strong>g <strong>in</strong>sofar as <strong>the</strong>ir assumptions that lack of statistical<br />
signifi cance means lack of environmental signifi -<br />
cance (Karr and Chu, 1999). Accord<strong>in</strong>g to Yoccoz (1991),<br />
Kriebel et al. (2001), and Lemons et al. (1997) ecologists<br />
tend to over-use tests of signifi cance and restoration ecologists<br />
are no exception to this rule. Karr and Chu (1999)<br />
suggest that it would be wiser to decide what is ecologically<br />
relevant fi rst and <strong>the</strong>n use hypo<strong>the</strong>sis test<strong>in</strong>g to detect<br />
ecologically relevant effects; <strong>the</strong> use of o<strong>the</strong>r statistical<br />
tools such as power analysis and decision <strong>the</strong>ory also is<br />
recommended (Hilborn, 1997).<br />
Cairns and Heckman (1996) state that restoration<br />
ecology <strong>in</strong> general ‘is a bridge between <strong>the</strong> social and<br />
natural sciences.’ In this chapter it has been shown that it<br />
is impossible to separate scientifi c and policy questions <strong>in</strong><br />
restoration ecology and this, <strong>in</strong> and of itself, <strong>in</strong>troduces<br />
uncerta<strong>in</strong>ty <strong>in</strong>to what o<strong>the</strong>rwise might be viewed as value–<br />
neutral or ‘objective’ scientifi c conclusions.<br />
As discussed more generally <strong>in</strong> this chapter and shown<br />
more specifi cally <strong>in</strong> <strong>the</strong> case study section, scientifi c<br />
research is both value-laden and is used to support politically-driven<br />
river restoration policies and decision mak<strong>in</strong>g<br />
(see also Shrader-Frechette, 1994). For example, historical<br />
or descriptive research is <strong>in</strong>tended to reveal or expla<strong>in</strong> <strong>the</strong><br />
dynamics of a given policy and to explore its orig<strong>in</strong> and<br />
evolution. Prescriptive or advocacy research defends a<br />
conclusion or possibly even a preconceived policy, and<br />
also is characterized by publicized disputes among, e.g.,<br />
scientists. Decision-<strong>in</strong>form<strong>in</strong>g or predictive research typically<br />
is fi nanced by grants or contracts lead<strong>in</strong>g to conclusions<br />
supportive of a predeterm<strong>in</strong>ed policy preference,<br />
sponsor bias, or predilections with<strong>in</strong> a research peer group.<br />
Consequently, <strong>the</strong> focus of this research does not attempt<br />
to analyze all feasible alternative policy choices and <strong>the</strong><br />
probable consequences. Because <strong>the</strong> focus of this research<br />
is on applicability for a particular policy its fi nd<strong>in</strong>gs are<br />
presented <strong>in</strong> <strong>the</strong> form of propositions upon which decisions<br />
can be made. The effi cacy of <strong>the</strong> policy towards<br />
which <strong>the</strong> research is focused depends on <strong>the</strong> validity,<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> 11<br />
reliability and persuasiveness of <strong>the</strong> research and <strong>the</strong><br />
extent of political public receptivity.<br />
It is important to clearly dist<strong>in</strong>guish between <strong>the</strong> use of<br />
methods and tools of science to understand <strong>the</strong> phenomena<br />
of nature and <strong>the</strong> acquisition of scientifi c <strong>in</strong>formation<br />
about a restoration issue and <strong>the</strong> sett<strong>in</strong>g of policy; but <strong>in</strong><br />
practice, <strong>the</strong>re is not always an unambiguous demarcation.<br />
Policy makers set agendas that determ<strong>in</strong>e <strong>the</strong> questions<br />
that are asked of scientists; scientists formulate hypo<strong>the</strong>ses<br />
<strong>in</strong> ways limited by <strong>the</strong>ir tools and <strong>the</strong>ir imag<strong>in</strong>ations<br />
and discipl<strong>in</strong>ary conventions. Consequently, <strong>the</strong> <strong>in</strong>formation<br />
<strong>the</strong>y provide to <strong>the</strong> policy makers is limited and<br />
socially determ<strong>in</strong>ed to a degree and <strong>the</strong>refore <strong>the</strong>re is a<br />
complicated feedback relation between <strong>the</strong> discoveries of<br />
science and <strong>the</strong> sett<strong>in</strong>g of policy. While attempt<strong>in</strong>g to be<br />
objective and focus on understand<strong>in</strong>g river restoration<br />
phenomena, scientists and o<strong>the</strong>r researchers should be<br />
aware of <strong>the</strong> policy uses of <strong>the</strong>ir work and of <strong>the</strong>ir social<br />
responsibility to carryout science that protects <strong>the</strong> environment<br />
and human health (Kriebel et al., 2001). In try<strong>in</strong>g<br />
to fulfi ll this responsibility, scientifi c and o<strong>the</strong>r uncerta<strong>in</strong>ty<br />
needs to be taken <strong>in</strong>to greater account.<br />
The discussion of some of <strong>the</strong> value-laden decisions and<br />
judgments scientists and o<strong>the</strong>r researchers make is not a<br />
criticism. Ra<strong>the</strong>r, <strong>the</strong> issue is discussed because a failure to<br />
recognize <strong>the</strong> existence of <strong>the</strong> value-laden dimensions of<br />
science casts serious doubt about even <strong>the</strong> best and most<br />
thorough of scientifi c and technical studies used to <strong>in</strong>form<br />
decisions about river restoration. In o<strong>the</strong>r words, unless <strong>the</strong><br />
value-laden dimensions of scientifi c and technical studies<br />
used to derive <strong>in</strong>formation are disclosed, <strong>the</strong> positions<br />
of policy makers and decision makers will appear to be<br />
justifi ed on objective or value–neutral scientifi c reason<strong>in</strong>g<br />
when, <strong>in</strong> fact, <strong>the</strong>y will be based <strong>in</strong> part on often controversial<br />
or confl ict<strong>in</strong>g values of scientists <strong>the</strong>mselves.<br />
REFERENCES<br />
Balon EK, Coche AG. 1974. Lake Kariba, A Man-Made Tropical<br />
Ecosystem <strong>in</strong> Central Africa. DW Junk Publishers: The Hague,<br />
The Ne<strong>the</strong>rlands.<br />
Bella DA, Jacobs R, Hiram L. 1994. Ecological <strong>in</strong>dicators of<br />
global climate change: A research framework. Environmental<br />
Management 18: 489–500.<br />
Benyam<strong>in</strong>e M. 2002. Theoretical Disputes and Practical Environmental<br />
Dilemmas. Orebro Studies <strong>in</strong> Environmental Science<br />
3. Orebro University: Orebro.<br />
Bradshaw GA, Borchers JG. 2000. <strong>Uncerta<strong>in</strong>ty</strong> as <strong>in</strong>formation:<br />
Narrow<strong>in</strong>g <strong>the</strong> Science – Policy gap. Conservation Ecology<br />
4: 7. [onl<strong>in</strong>e] URL: http://www.consecol.org/vol4/iss1/art7/)<br />
(accessed 23 March 2004)<br />
Briggs J, Peat F. 1982. Look<strong>in</strong>g Glass Universe. Simon and<br />
Schuster: New York.
12 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Brown D. 1987. Ethics, science, and environmental regulation.<br />
Environmental Ethics 9: 331–350.<br />
Brown D. 1995. The role of ethics <strong>in</strong> susta<strong>in</strong>able development and<br />
environmental protection decisionmak<strong>in</strong>g. In Susta<strong>in</strong>able<br />
Development: Science, Ethics, and Public Policy, Lemons J,<br />
Brown D (Eds). Kluwer Academic Publisher: Dordrecht, The<br />
Ne<strong>the</strong>rlands; 39–51.<br />
Cairns (Jr) J. 2001. Rationale for restoration. In Handbook of<br />
Ecological <strong>Restoration</strong>, Vol. 1: Pr<strong>in</strong>ciples and <strong>Restoration</strong>,<br />
Davy AJ, Perry, J (Eds). Cambridge University Press:<br />
Cambridge; 10–23.<br />
Cairns (Jr) J. 2003. Ethical issues <strong>in</strong> ecological restoration. Ethics<br />
<strong>in</strong> Science and Environmental Politics 3: 50–61.<br />
Cairns J (Jr), Heckman JR. 1996. <strong>Restoration</strong> ecology: The<br />
state of an emerg<strong>in</strong>g fi eld. Annual Review of Energy and<br />
Environment 21: 167–187.<br />
Caldwell LK. 1996. In Scientifi c <strong>Uncerta<strong>in</strong>ty</strong> and Environmental<br />
Problem Solv<strong>in</strong>g, Lemons J (Ed). Blackwell Science,<br />
Cambridge; 394–422.<br />
Canter LW. 1996. Scientifi c uncerta<strong>in</strong>ty and water resources management.<br />
In Scientifi c <strong>Uncerta<strong>in</strong>ty</strong> and Environmental Problem<br />
Solv<strong>in</strong>g, Lemons J (Ed). Blackwell Science Inc.: Cambridge;<br />
264–297.<br />
Chauhan M. 2003. Conserv<strong>in</strong>g biodiversity <strong>in</strong> arid regions: Experiences<br />
with protected areas <strong>in</strong> India. In Conserv<strong>in</strong>g Biodiversity<br />
<strong>in</strong> Arid Regions. Lemons J, Victor R, Schaffer D (Eds).<br />
Kluwer Academic Publishers: Boston.<br />
Cranor C. 1993. Regulat<strong>in</strong>g Toxic Substances: A Philosophy of<br />
Science and Law. Oxford University Press, New York.<br />
De Wall LC, Large ARG, Gippel CJ, Wade PM. 1995. <strong>River</strong> and<br />
fl oodpla<strong>in</strong> rehabilitation <strong>in</strong> Western Europe: Opportunities and<br />
constra<strong>in</strong>ts. Archives Hydrobiologia Supplement 101 – Large<br />
<strong>River</strong>s 9: 679–693.<br />
Eiseltova M, Biggs J. 1995. <strong>Restoration</strong> of stream ecosystems: An<br />
<strong>in</strong>tegrated approach. International Waterfowl and Wetlands<br />
Research Bureau Publication 37: 1–170.<br />
(EEA) European Environment Agency. 2001. Late Lessons From<br />
Early Warn<strong>in</strong>gs: The Precautionary Pr<strong>in</strong>ciple 1896–2000,<br />
Environmental Issue Report No. 22. European Environment<br />
Agency: Luxembourg.<br />
(EPA) U.S. Environmental Protection Agency. 2004. Lower Mississippi<br />
Valley Ecosystem <strong>Restoration</strong> Initiative. [onl<strong>in</strong>e] URL:<br />
http://www.epa.gov/region4/programs/cbep/lowmiss.html<br />
(accessed 25 March 2004).<br />
Fleck L. 1979. Genesis and Development of a Scientifi c Fact.<br />
University of Chicago Press: Chicago.<br />
Funtowicz SO, Ravetz JR. 1995. Science for <strong>the</strong> post normal age.<br />
In Perspectives on Ecological Integrity, Westra L, Lemons<br />
J (eds). Kluwer Academic Publishers: Dordrecht, The<br />
Ne<strong>the</strong>rlands, 146–161.<br />
Gomez-Pompa A, Kaus A. 1992. Tam<strong>in</strong>g <strong>the</strong> wilderness myth.<br />
BioScience 42: 271–279.<br />
Gopal B. 2003. Aquatic biodiversity <strong>in</strong> arid and semi-arid zones<br />
of Asia and water management. In Conserv<strong>in</strong>g Biodiversity <strong>in</strong><br />
Arid Regions, Lemons J, Victor R, Schaffer D (Eds). Kluwer<br />
Academic Publishers: Norwell; 199–216.<br />
Higgs ES. 1997. What is good ecological restoration? Conservation<br />
Biology 11: 338–348.<br />
Hilborn R. 1997. Statistical hypo<strong>the</strong>sis test<strong>in</strong>g and decision<br />
<strong>the</strong>ory <strong>in</strong> fi sheries science. Fisheries 22: 19–20.<br />
Hull D. 1974. Philosophy of Biological Science. Prentice–Hall:<br />
Englewood Cliffs.<br />
Kaiser M, Storvik H (Eds). 2003. The Precautionary Pr<strong>in</strong>ciple:<br />
Between Research and Politics. The National Research<br />
Ethical Committee for Natural Science and Technology:<br />
Oslo, Norway.<br />
Karr JR, Chu EW. 1999. Restor<strong>in</strong>g Life <strong>in</strong> Runn<strong>in</strong>g Waters. Better<br />
Biological Monitor<strong>in</strong>g. Island Press: Wash<strong>in</strong>gton, DC.<br />
Kriebel D et al. 2001. The precautionary pr<strong>in</strong>ciple <strong>in</strong> environmental<br />
science. Environmental Health Perspectives 109: 871–876.<br />
Kuhn T. 1962. The Theory of Scientifi c Revolutions. University of<br />
Chicago Press: Chicago.<br />
Latour B. 1988. The politics of explanation: An alternative. In<br />
Knowledge and Refl exivity: New Frontiers <strong>in</strong> <strong>the</strong> Sociology<br />
of Knowledge, Woolgar S (Ed). Sage Publications: London;<br />
155–176.<br />
Lemons J (Ed). 1996. Scientifi c <strong>Uncerta<strong>in</strong>ty</strong> and Environmental<br />
Problem Solv<strong>in</strong>g. Blackwell Science: Cambridge.<br />
Lemons J, Brown D (Eds). 1995. Susta<strong>in</strong>able Development:<br />
Science, Ethics, and Public Policy. Kluwer Academic<br />
Publishers: Dordrecht, The Ne<strong>the</strong>rlands.<br />
Lemons J, Shrader-Frechette KS, Cranor C. 1997. The precautionary<br />
pr<strong>in</strong>ciple: Scientifi c uncerta<strong>in</strong>ty and type I and type II<br />
errors. Foundations of Science 2: 207–236.<br />
Light A, Higgs ES. 1996. The politics of ecological restoration.<br />
Environmental Ethics 18: 227–247.<br />
Lubchenco J et al. 1991. The susta<strong>in</strong>able biosphere <strong>in</strong>itiative: An<br />
ecological research agenda. Ecology 72: 371–412.<br />
MacMahon JA, Holl KD. 2001. Ecological restoration. A key to<br />
conservation biology’s future. In Conservation Biology.<br />
Research Priorities for <strong>the</strong> Next Decade, Soule ME, Orians GH<br />
(Eds). Island Press: Wash<strong>in</strong>gton, DC; 245–269.<br />
Mayo DG, Hollander RD (Eds). 1991. Acceptable Evidence.<br />
Oxford University Press: Oxford.<br />
(NRC) National Research Council. 1999. New Strategies for<br />
America’s Watersheds. National Academy Press, Wash<strong>in</strong>gton,<br />
DC.<br />
(NRC) National Research Council. 2004. Endangered and<br />
Threatened Fishes <strong>in</strong> <strong>the</strong> Klamath <strong>River</strong> Bas<strong>in</strong>: Causes of<br />
Decl<strong>in</strong>e and Strategies for Recovery. Committee on Endangered<br />
and Threatened Species <strong>in</strong> <strong>the</strong> Klamath <strong>River</strong> Bas<strong>in</strong>,<br />
Wash<strong>in</strong>gton, DC.<br />
Nienhuis PH, Leuven RSEW. 2001. <strong>River</strong> restoration and fl ood<br />
protection: Controversy or synergism? Hydrobiologia 444:<br />
85–99.<br />
Obeng LE. 1981. Man’s Impact on Tropical <strong>River</strong>s. In Perspectives<br />
<strong>in</strong> Runn<strong>in</strong>g Water Ecology, Lock MA, Williams DD (Eds).<br />
Plenum Press: New York; 265–288.<br />
Ogbeibu AE, Victor R. 1989. The effects of road and bridge<br />
construction on <strong>the</strong> bank–root macrobenthic <strong>in</strong>vertebrates<br />
of a sou<strong>the</strong>rn Nigerian stream. Environmental Pollution 56:<br />
85–100.
Peters RH. 1991. A Critique for Ecology. Cambridge University<br />
Press: Cambridge.<br />
Regier HA. 1995. Ecosystem <strong>in</strong>tegrity <strong>in</strong> a context of ecostudies<br />
as related to <strong>the</strong> Great Lakes Region. In Perspectives on Ecological<br />
Integrity, Westra L, Lemons J (Eds). Kluwer Academic<br />
Publishers: Dordrecht, The Ne<strong>the</strong>rlands, 88–101.<br />
Rolston H (III). 1988. Environmental Ethics, Duties to and Values<br />
<strong>in</strong> <strong>the</strong> Natural World. Temple University Press: Philadelphia.<br />
Rorty R. 1979. Philosophy and <strong>the</strong> Mirror of Nature. Pr<strong>in</strong>ceton<br />
University Press: New Jersey.<br />
Rzoska J. 1976. The Nile: Biology of an Ancient <strong>River</strong>. DW Junk<br />
Publishers: The Hague, The Ne<strong>the</strong>rlands.<br />
(SER) Society for Ecological <strong>Restoration</strong> Science and Policy<br />
Work<strong>in</strong>g Group. 2002. The SER Primer on Ecological<br />
<strong>Restoration</strong>. [onl<strong>in</strong>e] URL: www.ser.org (accessed 25 August<br />
2004).<br />
Service RF. 2003. NRC backs ecosystem-wide changes to save<br />
Klamath fi sh. Science 302: 765.<br />
Shrader-Frechette KS. 1982. Environmental impact assessment<br />
and <strong>the</strong> fallacy of unfi nished bus<strong>in</strong>ess. Environmental Ethics<br />
4: 37–48.<br />
Shrader-Frechette KS. 1994. Ethics of Scientifi c Research. Rowan<br />
& Littlefi eld Publisher, Inc.: Lanham.<br />
Shrader-Frechette KS. 1996. Methodological rules for four<br />
classes of scientifi c uncerta<strong>in</strong>ty. In Scientifi c <strong>Uncerta<strong>in</strong>ty</strong> and<br />
Environmental Problem Solv<strong>in</strong>g, Lemons J (Ed). Blackwell<br />
Science: Cambridge; 12–39.<br />
Shrader-Frechette KS, McCoy E. 1993. Method <strong>in</strong> Ecology.<br />
Cambridge University Press: Cambridge.<br />
Sirageld<strong>in</strong> I. 2002. Human prospects <strong>in</strong> an age of uncerta<strong>in</strong>ty and<br />
decl<strong>in</strong>e of rationality. Newsletter of <strong>the</strong> Economic Research<br />
Forum for <strong>the</strong> Arab Countries, Iran and Turkey 2: 9–11.<br />
Smits AJ, Niehuis PH, Leuven RSEW. 2000. New Approaches<br />
to <strong>River</strong> Management. Backhuys Publishers: Leiden, The<br />
Ne<strong>the</strong>rlands.<br />
Stewart-Oaten A. 1996. Goals <strong>in</strong> environmental monitor<strong>in</strong>g. In<br />
Detect<strong>in</strong>g Ecological Impacts: Concepts and Applications <strong>in</strong><br />
Coastal Habitats, Schmitt RJ, Osenbert CW (Eds). Academic<br />
Press: San Diego; 17–28.<br />
Sunste<strong>in</strong> C. 2002. Risk and Reason. Cambridge University Press.<br />
Cambridge.<br />
Sweeney S. 2000. Different means, shared ends: Environmental<br />
restoration and restoration ecology. Biologica 38: 129–139.<br />
(SWS) Society of Wetland Scientists position paper on <strong>the</strong> defi nitions<br />
of wetlands restoration. 2000. Ecological Applications 6:<br />
84–93.<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> 13<br />
Tickner J. (Ed). 2002. Environmental Science and Preventive<br />
Public Policy. Island Press: Wash<strong>in</strong>gton DC.<br />
Tickner JA. 2003. Precaution: Environmental Science and<br />
Preventive Public Policy. Island Press: Wash<strong>in</strong>gton, DC.<br />
Victor R. 1998. Fish community changes associated with a small<br />
African reservoir. International Review of Hydrobiologia 83:<br />
603–610.<br />
Victor R, Brown CA. 1990. The food and feed<strong>in</strong>g habits of two<br />
species of characid fi sh <strong>in</strong> a perturbed West Africa river.<br />
Journal of African Zoology 104: 97–108.<br />
Victor R, Dickson DT. 1985. Macrobenthic <strong>in</strong>vertebrates of a<br />
perturbed stream <strong>in</strong> Sou<strong>the</strong>rn Nigeria. Environmental Pollution<br />
Series A 38: 99–107.<br />
Victor R, Meye J. 1994. Fur<strong>the</strong>r studies on <strong>the</strong> fi sh communities<br />
of a perturbed stream <strong>in</strong> Sou<strong>the</strong>rn Nigeria. Journal of Tropical<br />
Ecology 10: 627–632.<br />
Victor R, Ogbeibu AE. 1985. Macrobenthic <strong>in</strong>vertebrates of<br />
a stream fl ow<strong>in</strong>g through farmlands <strong>in</strong> Sou<strong>the</strong>rn Nigeria.<br />
Environmental Pollution Series A 39: 337–349.<br />
Victor R., Ogbeibu AE. 1986. Reconlonization of macrobenthic<br />
<strong>in</strong>vertebrates <strong>in</strong> a Nigerian stream after pesticide treatment and<br />
associated disruption. Environmental Pollution Series A 41:<br />
125–137.<br />
Victor R, Ogbeibu AE. 1991. Macrobenthic <strong>in</strong>vertebrates <strong>in</strong> <strong>the</strong><br />
erosional biotope of a Nigerian river. Tropical Zoology 4:<br />
1–12.<br />
Victor R. Onomivbori O. 1996. The effects of urban perturbations<br />
on <strong>the</strong> benthic macro<strong>in</strong>vertebrates of a sou<strong>the</strong>rn Nigerian<br />
stream. In Perspectives <strong>in</strong> Tropical Limnology, Schiemer F,<br />
Boland T (Eds). SPB Academic Publish<strong>in</strong>g: Amsterdam, The<br />
Ne<strong>the</strong>rlands; 233–238.<br />
Victor R, Tetteh JO. 1988. Fish communities of a perturbed<br />
stream <strong>in</strong> Sou<strong>the</strong>rn Nigeria. Nigeria Journal Tropical Ecology<br />
4: 49–59.<br />
Ward JV, Stanford JA. 1979. The Ecology of Regulated Streams.<br />
Plenum Press: New York.<br />
Westra L. 1995. Ecosystem <strong>in</strong>tegrity and susta<strong>in</strong>ability: The foundational<br />
value of <strong>the</strong> wild. In Perspectives on Ecological Integrity,<br />
Westra L, Lemons J (Eds). Kluwer Academic Publishers:<br />
Dordrecht, The Ne<strong>the</strong>rlands; 12–33.<br />
Westra L, Lemons J (Eds). 1995. Perspectives on Ecological<br />
Integrity. Kluwer Academic Publishers: Dordrecht, The<br />
Ne<strong>the</strong>rlands.<br />
Yoccoz NG. 1991. Use, overuse and misuse of signifi cance tests<br />
<strong>in</strong> evolutionary biology and ecology. Bullet<strong>in</strong> Ecological<br />
Society America 71: 106–111.
<strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat Edited by Stephen Darby and David Sear<br />
© 2008 John Wiley & Sons, Ltd<br />
2<br />
Sources of <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>River</strong><br />
<strong>Restoration</strong> Research<br />
2.1 INTRODUCTION: GENERAL SOURCES<br />
OF UNCERTAINTY<br />
The practice of science <strong>in</strong> support of river restoration is<br />
subject to four primary sources of uncerta<strong>in</strong>ty (see Chapters<br />
1 and 3 for additional/alternative views) so signifi cant<br />
that <strong>the</strong>y may prevent <strong>the</strong> restoration from achiev<strong>in</strong>g its<br />
goals. Firstly, <strong>the</strong> underly<strong>in</strong>g <strong>the</strong>ory applied by <strong>in</strong>vestigators<br />
to particular problem cases is imperfect and conta<strong>in</strong>s<br />
substantial gaps <strong>in</strong> explanatory and predictive capability.<br />
Secondly, <strong>the</strong> research process itself is subject to a variety<br />
of operational problems that <strong>in</strong>troduce uncerta<strong>in</strong>ty to <strong>the</strong><br />
use of science. Thirdly, <strong>the</strong> communication of scientifi c<br />
results to decision makers is often fraught with ambiguity<br />
derived from <strong>the</strong> scientifi c sender as well as <strong>the</strong> policy<br />
receiver. Fourthly, <strong>the</strong> scientists <strong>the</strong>mselves are subject to<br />
bias that generates doubt <strong>in</strong> <strong>the</strong> outcome of generat<strong>in</strong>g<br />
scientifi c products and apply<strong>in</strong>g <strong>the</strong>m. In <strong>the</strong> follow<strong>in</strong>g<br />
sections <strong>the</strong> issues for each of <strong>the</strong>se sources of uncerta<strong>in</strong>ty<br />
is outl<strong>in</strong>ed.<br />
2.2 UNCERTAINTY IN THEORY<br />
All science <strong>in</strong> support of river restoration beg<strong>in</strong>s with<br />
<strong>the</strong>ory, because it is <strong>the</strong>ory that allows <strong>the</strong> <strong>in</strong>vestigator to<br />
identify what to measure and how to construct a conceptual<br />
model that connects <strong>the</strong> measurements toge<strong>the</strong>r.<br />
Investigators perceive only those aspects of <strong>the</strong> river and<br />
its operations that <strong>the</strong>ory allows <strong>the</strong>m to see. Practitioners<br />
of fl uvial geomorphology tend to revere exist<strong>in</strong>g <strong>the</strong>ory as<br />
a sacrosanct start<strong>in</strong>g po<strong>in</strong>t but, like all sciences, geomorphology<br />
is <strong>in</strong> a state of constant change and revision. The<br />
William L. Graf<br />
Department of Geography, University of South Carol<strong>in</strong>a, USA<br />
change is sometimes gradual, as with <strong>the</strong> development and<br />
application of fundamental hydraulics to expla<strong>in</strong> river<br />
behavior that evolved over a period of several decades<br />
(Chang, 1998; Simons and Sentürk, 1992). Sometimes <strong>the</strong><br />
change is abrupt, as was <strong>the</strong> case with <strong>the</strong> <strong>in</strong>troduction of<br />
<strong>the</strong> concepts surround<strong>in</strong>g hydraulic geometry that burst<br />
upon <strong>the</strong> fl uvial geomorphology scene, became widely<br />
accepted <strong>in</strong> less than a decade and cont<strong>in</strong>ued <strong>in</strong> common<br />
use for several decades (Leopold, 1994). The result of this<br />
constantly chang<strong>in</strong>g <strong>the</strong>ory is that <strong>the</strong> geomorphologist<br />
work<strong>in</strong>g <strong>in</strong> 2007 may perceive a very different system than<br />
one work<strong>in</strong>g just a few years before or later, even though<br />
<strong>the</strong> physical system <strong>in</strong> all cases would be <strong>the</strong> same. <strong>Uncerta<strong>in</strong>ty</strong>,<br />
<strong>the</strong>refore, is <strong>in</strong>cluded <strong>in</strong> <strong>the</strong> application of science<br />
<strong>in</strong> its broadest sense.<br />
Ano<strong>the</strong>r source of ambiguity <strong>in</strong> <strong>the</strong>ory for river restoration<br />
is <strong>the</strong> regional specifi city that is built <strong>in</strong>to much<br />
of fl uvial <strong>the</strong>ory. Much of what we <strong>the</strong>orize about<br />
s<strong>in</strong>gle-thread meander<strong>in</strong>g rivers comes from research<br />
experience <strong>in</strong> northwest Europe and eastern North America<br />
(Knighton, 1998), yet <strong>the</strong> global applicability of this work<br />
is largely untested. Most of <strong>the</strong> streams of northwest<br />
Europe and eastern North America that have been <strong>in</strong>tensively<br />
<strong>in</strong>vestigated are relatively small on a world-wide<br />
basis, and though some generalities certa<strong>in</strong>ly must apply<br />
<strong>in</strong> many locales, <strong>the</strong> details may differ. Until <strong>the</strong> late<br />
1990s, much of <strong>the</strong> <strong>the</strong>ory for dryland rivers came from<br />
experiences <strong>in</strong> <strong>the</strong> American Southwest (Graf, 1988), but<br />
more recent <strong>in</strong>vestigations <strong>in</strong> Australia by Gerald Nanson,<br />
Steven Tooth and o<strong>the</strong>rs, for example, have shown that <strong>the</strong><br />
American experience is not applicable <strong>in</strong> all drylands<br />
(Nanson and Knighton, 1996).
16 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
If it is true that we must <strong>the</strong>orize based on what we<br />
know best, it must also be true that we are still limited <strong>in</strong><br />
<strong>the</strong> range of our collective experience. As a result, when<br />
we apply exist<strong>in</strong>g <strong>the</strong>ory <strong>in</strong> new geographic sett<strong>in</strong>gs, <strong>the</strong>re<br />
is reasonable doubt about <strong>the</strong> applicability of that <strong>the</strong>ory,<br />
at least <strong>in</strong> its totality. Geomorphologists commonly recognize<br />
that it is unwise to extend statistical models beyond<br />
<strong>the</strong> numerical ranges of <strong>the</strong> data. It is equally risky to<br />
extend geomorphological models beyond <strong>the</strong> geographic<br />
ranges of <strong>the</strong>ir orig<strong>in</strong>. The extension of <strong>the</strong>ory also forces<br />
us to consider how much of <strong>the</strong> geomorphology and<br />
hydrology of a particular river is unique, regardless of its<br />
geographic location. Each reach (a few kilometers long)<br />
of a stream is likely to be unique but <strong>the</strong> overall operation<br />
and form of a river (hundreds of kilometers long) is likely<br />
to have many similarities with o<strong>the</strong>r systems of similar<br />
magnitude. At <strong>the</strong> more extensive end of this range of<br />
magnitudes, generalizations are possible, while at <strong>the</strong> local<br />
end of <strong>the</strong> range of magnitudes, uniqueness becomes more<br />
apparent.<br />
The <strong>in</strong>completeness of most fl uvial geomorphologic<br />
<strong>the</strong>ory is also a source of uncerta<strong>in</strong>ty. This <strong>in</strong>completeness<br />
is <strong>in</strong> part purely a function of <strong>the</strong> natural river system, for<br />
which <strong>in</strong>vestigators have n<strong>in</strong>e fundamental operat<strong>in</strong>g variables,<br />
but for which <strong>the</strong>re are only a very few connective<br />
ma<strong>the</strong>matical functions (Leopold, Wolman, and Miller,<br />
1964). But an equally important limitation of exist<strong>in</strong>g<br />
<strong>the</strong>ory is its lack of recognition of human effects. Throughout<br />
much of <strong>the</strong> twentieth century (with a few exceptions),<br />
geomorphology as a science pursued explanation for<br />
‘natural’ rivers and many <strong>in</strong>vestigators made a conscious<br />
effort to avoid <strong>the</strong> confound<strong>in</strong>g <strong>in</strong>fl uence of technological<br />
<strong>in</strong>fl uences. It has only been <strong>in</strong> <strong>the</strong> last twenty years that<br />
those human <strong>in</strong>fl uences, pervasive and signifi cant <strong>in</strong> many<br />
rivers of <strong>the</strong> world, have <strong>the</strong>mselves become <strong>the</strong> objects<br />
of study (Costa et al., 1995; Graf, 2001). By defi nition,<br />
rivers subject to restoration have undergone changes<br />
result<strong>in</strong>g from human management and technology, but<br />
exist<strong>in</strong>g <strong>the</strong>ory is remarkably weak with respect to <strong>the</strong>se<br />
issues.<br />
The Colorado <strong>River</strong> <strong>in</strong> <strong>the</strong> Grand Canyon <strong>in</strong> <strong>the</strong> USA<br />
provides an example of <strong>the</strong> issues related to uncerta<strong>in</strong>ty <strong>in</strong><br />
<strong>the</strong>ory. Glen Canyon Dam, several kilometers upstream<br />
from <strong>the</strong> Grand Canyon controls <strong>the</strong> fl ow of <strong>the</strong> river, and<br />
especially reduces annual fl ood peaks to less than half of<br />
<strong>the</strong>ir former magnitude. The dam also reduces <strong>the</strong> sediment<br />
supply to <strong>the</strong> downstream canyon by more than 80%.<br />
As a result, <strong>the</strong> river has eroded sandy beaches and bars<br />
that once were common <strong>in</strong> <strong>the</strong> canyon (National Research<br />
Council, 1996). <strong>River</strong> restoration for <strong>the</strong> canyon <strong>in</strong>cluded<br />
re<strong>in</strong>troduction of moderate fl oods to move <strong>the</strong> available<br />
sediment from <strong>the</strong> channel fl oor to elevated positions,<br />
restor<strong>in</strong>g <strong>the</strong>se ecological niches. Despite considerable<br />
research, <strong>the</strong>re were no established <strong>the</strong>ories to predict <strong>the</strong><br />
response of <strong>the</strong> river to <strong>the</strong> artifi cial fl oods and although<br />
<strong>the</strong>re have been several fl ood-simulat<strong>in</strong>g releases from <strong>the</strong><br />
dam, <strong>the</strong> restoration results are not yet apparent.<br />
2.3 UNCERTAINTY IN RESEARCH<br />
Research us<strong>in</strong>g admittedly limited <strong>the</strong>ory <strong>in</strong> support of<br />
restoration is subject to uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong> specifi cation<br />
of variables, assumptions, sampl<strong>in</strong>g, measurements and<br />
test<strong>in</strong>g of hypo<strong>the</strong>ses. The specifi cation or defi nition of<br />
variables, for example, is much entangled <strong>in</strong> <strong>the</strong> vagaries<br />
of science, law and personal perception of <strong>the</strong> researcher.<br />
Channel width provides an example. Most geomorphologists<br />
would agree that channel width is <strong>the</strong> distance across<br />
<strong>the</strong> active channel from one bank to <strong>the</strong> o<strong>the</strong>r, but <strong>the</strong><br />
application of this seem<strong>in</strong>gly simple proposition is devilishly<br />
diffi cult <strong>in</strong> many rivers. How should semi-permanent<br />
islands be taken <strong>in</strong>to account? What about ephemeral<br />
bars? How should width be determ<strong>in</strong>ed <strong>in</strong> <strong>the</strong> common<br />
circumstance where multiple sets of banks have resulted<br />
from episodic <strong>in</strong>cision or simply variable fl ows, which is<br />
often <strong>the</strong> case <strong>in</strong> arid, semi-arid, arctic or alp<strong>in</strong>e regions.<br />
Many legal systems also defi ne <strong>the</strong> channel as be<strong>in</strong>g<br />
‘between <strong>the</strong> banks,’ but do not specify which banks to use<br />
for <strong>the</strong> description (Graf, 1988).<br />
All geomorphological research <strong>in</strong>cludes assumptions<br />
which form ano<strong>the</strong>r source of uncerta<strong>in</strong>ty. The geographic<br />
and ecological complexity of rivers and <strong>the</strong>ir environments<br />
imply that when conduct<strong>in</strong>g <strong>in</strong>vestigations it is<br />
essential to focus on a few components and assume away<br />
<strong>the</strong> importance of variability <strong>in</strong> o<strong>the</strong>r factors that go<br />
unmeasured. In geomorphology, hydrology and eng<strong>in</strong>eer<strong>in</strong>g<br />
studies that support restoration, <strong>in</strong>vestigators often<br />
assume stationarity of <strong>the</strong> hydro–climatic processes rul<strong>in</strong>g<br />
<strong>the</strong> river. Stationarity means <strong>in</strong>vestigators assume that<br />
<strong>the</strong> underly<strong>in</strong>g statistical distributions describ<strong>in</strong>g climatic<br />
variables important to river processes are unchang<strong>in</strong>g.<br />
Standard magnitude/frequency analysis <strong>in</strong>cludes this<br />
assumption so that <strong>the</strong> researcher can address o<strong>the</strong>r variables<br />
of <strong>in</strong>terest to planners, <strong>in</strong>clud<strong>in</strong>g <strong>the</strong> return <strong>in</strong>tervals<br />
for various magnitudes of discharge. However, climate is<br />
anyth<strong>in</strong>g but stationary, and its variation is highly likely<br />
to <strong>in</strong>fl uence <strong>the</strong> statistical distributions upon which return<br />
<strong>in</strong>terval concepts depend. This variation is also likely to<br />
be signifi cant to <strong>the</strong> fl uvial system over time scales as<br />
short as decades, scales that encompass <strong>the</strong> likely project<br />
life of most restoration efforts. Predictions for <strong>the</strong> nearterm<br />
future of a few decades are <strong>the</strong>refore uncerta<strong>in</strong><br />
because <strong>the</strong> effects of expected climatic changes are not<br />
part of <strong>the</strong> analysis (see Chapter 13).
Sampl<strong>in</strong>g processes <strong>in</strong> fl uvial geomorphology also cast<br />
doubt on <strong>the</strong> confi dence users may have <strong>in</strong> <strong>the</strong> reliability<br />
of <strong>the</strong> result<strong>in</strong>g data. <strong>River</strong>s present <strong>the</strong> researcher with<br />
long l<strong>in</strong>es or corridors, and if <strong>the</strong> <strong>in</strong>vestigator uses cross<br />
sections or po<strong>in</strong>t samples, <strong>the</strong> selection of <strong>the</strong> locations of<br />
<strong>the</strong>se sample po<strong>in</strong>ts may strongly <strong>in</strong>fl uence <strong>the</strong> result<strong>in</strong>g<br />
data. The spac<strong>in</strong>g of meanders, riffl es and pools <strong>in</strong>troduces<br />
some regularity to <strong>the</strong> spac<strong>in</strong>g of important geomorphic<br />
and hydraulic features of rivers, so that if <strong>in</strong>vestigators use<br />
regular spac<strong>in</strong>g for sample cross sections or sites, that<br />
spac<strong>in</strong>g may co<strong>in</strong>cide with <strong>the</strong> spac<strong>in</strong>g of particular<br />
characteristics of <strong>the</strong> channel. For example, sample sites<br />
might occur only on riffl es, or only <strong>in</strong> pools, giv<strong>in</strong>g a false<br />
picture of <strong>the</strong> system. Truly random spac<strong>in</strong>g for samples<br />
may be statistically desirable, but a realistic view of geomorphic<br />
and hydrologic conditions demands that all <strong>the</strong><br />
various environments of <strong>the</strong> channel be <strong>in</strong>cluded <strong>in</strong> <strong>the</strong><br />
sample, someth<strong>in</strong>g that is not possible without some<br />
understand<strong>in</strong>g of <strong>the</strong> basic spatial framework of <strong>the</strong><br />
system.<br />
<strong>Uncerta<strong>in</strong>ty</strong> may be lessened if scale is part of <strong>the</strong> plann<strong>in</strong>g<br />
process for select<strong>in</strong>g sample sites. Sample schemes<br />
couched <strong>in</strong> <strong>the</strong> concept of river reaches permit bracket<strong>in</strong>g<br />
of relevant parts of <strong>the</strong> stream. A river reach is from one<br />
to a few kilometers <strong>in</strong> length and conta<strong>in</strong>s similar geomorphic<br />
conditions throughout its extent, sometimes with<br />
repetitive and alternat<strong>in</strong>g channel confi gurations such as<br />
pools and riffl es, pools and rapids, or meanders. Several<br />
reaches may make up a river segment. Geological boundaries,<br />
confl uences with major tributaries or human structures<br />
such as dams and diversion works create <strong>the</strong> upstream<br />
and downstream boundaries of each segment. Sampl<strong>in</strong>g<br />
schemes constructed with <strong>the</strong> reaches and segments as<br />
geographic frameworks are more likely to be <strong>in</strong>formative<br />
for restoration work than truly regular or truly random<br />
samples. Project design for restoration requires a clear<br />
assessment of <strong>the</strong> appropriate dimensions and spac<strong>in</strong>g of<br />
repetitions to <strong>in</strong>sure long term stability. If <strong>the</strong> dimensions<br />
and spac<strong>in</strong>g are not refl ective of <strong>the</strong> restored hydraulic<br />
conditions, <strong>the</strong> restored system will be unstable and may<br />
dis<strong>in</strong>tegrate.<br />
The example of restoration of <strong>the</strong> Platte <strong>River</strong> <strong>in</strong><br />
Nebraska <strong>in</strong> <strong>the</strong> USA illustrates <strong>the</strong> role of uncerta<strong>in</strong>ty <strong>in</strong><br />
research. Upstream fl ood control dams have brought about<br />
great changes <strong>in</strong> <strong>the</strong> hydrologic regime and geomorphology<br />
of <strong>the</strong> Platte, a serious issue because <strong>the</strong> river<br />
hosts several endangered bird species. The river orig<strong>in</strong>ally<br />
<strong>in</strong>cluded valuable habitats for birds, particularly <strong>the</strong><br />
whoop<strong>in</strong>g crane, but <strong>the</strong> hydrologic and geomorphic<br />
changes have reduced <strong>the</strong>ir habitat. <strong>River</strong> restoration<br />
<strong>in</strong>cludes return<strong>in</strong>g <strong>the</strong> river to conditions closer to those<br />
that prevailed before <strong>the</strong> dams were <strong>in</strong> place. Features such<br />
Sources of <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> Research 17<br />
as multiple channels, high and low islands, and complex<br />
bars are essential to <strong>the</strong> restoration. Research to support<br />
<strong>the</strong> restoration has been under way for more than a decade<br />
but has produced results that are diffi cult to <strong>in</strong>terpret<br />
(National Research Council, 2004). Cross-sectional<br />
surveys, for example, are numerous but not often conducted<br />
at <strong>the</strong> same places through time so that sampl<strong>in</strong>g<br />
is an issue. The fl ow parameters that are most important<br />
to <strong>the</strong> species may not be <strong>the</strong> same parameters that are<br />
important to <strong>the</strong> geomorphology. Stationarity is a particular<br />
problem <strong>in</strong> deal<strong>in</strong>g with <strong>the</strong> hydrologic records of <strong>the</strong><br />
Platte because <strong>the</strong> river is located on <strong>the</strong> boundary between<br />
sub-humid and sub-arid regions and is subject to climatic<br />
fl uctuations on decadal and century-long periods which<br />
casts doubts on shorter term records.<br />
2.4 UNCERTAINTY IN COMMUNICATION<br />
The successful use of science <strong>in</strong> formulat<strong>in</strong>g public policy<br />
for <strong>the</strong> restoration of rivers relies on accurate communication<br />
between researchers and decision makers, but this<br />
connection sometimes suffers from fail<strong>in</strong>gs by both participants.<br />
Scientists are usually accustomed to communicat<strong>in</strong>g<br />
with each o<strong>the</strong>r us<strong>in</strong>g a specifi c shared language of<br />
technical terms. Even terms shared with <strong>the</strong> general language<br />
may take on specifi c shad<strong>in</strong>gs of mean<strong>in</strong>g when<br />
used by specialists communicat<strong>in</strong>g with each o<strong>the</strong>r. A<br />
steep river gradient, for example, calls to m<strong>in</strong>d a general<br />
picture for most geomorphologists, perhaps of a step-pool<br />
sequence for mounta<strong>in</strong> streams or a braided channel <strong>in</strong><br />
o<strong>the</strong>r sett<strong>in</strong>gs. For <strong>the</strong> decision maker, a steep river may<br />
be one with water falls. More importantly, specialized<br />
terms and terms with nuances do not work well when <strong>the</strong><br />
listener or reader is not tra<strong>in</strong>ed or experienced <strong>in</strong> <strong>the</strong> scientifi<br />
c specialty. As a result, scientists have an obligation<br />
to use language without jargon and without assumptions<br />
when communicat<strong>in</strong>g with decision makers and <strong>the</strong><br />
<strong>in</strong>formed public, a task that seem<strong>in</strong>gly challenges many<br />
specialists.<br />
Decision makers, on <strong>the</strong> o<strong>the</strong>r hand, have poorly<br />
<strong>in</strong>formed expectations regard<strong>in</strong>g <strong>the</strong>ir scientifi c advisors.<br />
They expect clear, unambiguous answers to <strong>the</strong>ir questions<br />
and specifi c, robust predictions of future processes that<br />
might result from a variety of potential decisions. Science,<br />
however, rarely offers truly unambiguous conclusions, and<br />
caveats are many <strong>in</strong> applied geomorphology. Often, it is<br />
possible to predict <strong>the</strong> direction of change, but <strong>the</strong>re is<br />
greater trouble predict<strong>in</strong>g <strong>the</strong> magnitude of that change.<br />
To a public servant attempt<strong>in</strong>g to protect property values,<br />
such predictions may lack <strong>the</strong> required level of comfort.<br />
A common error <strong>in</strong> communication that results <strong>in</strong><br />
uncerta<strong>in</strong>ty occurs when <strong>the</strong> decision maker requests
18 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
specifi c numbers, such as ‘what will be <strong>the</strong> stable, long<br />
term width of <strong>the</strong> restored channel’, a reasonable question<br />
<strong>in</strong> <strong>the</strong> public policy arena. However, most scientifi c and<br />
eng<strong>in</strong>eer<strong>in</strong>g models predict some most likely value, enveloped<br />
<strong>in</strong> error bars. The most effective (and honest)<br />
report <strong>in</strong>cludes <strong>the</strong> most probable value but also <strong>in</strong>cludes<br />
some discussion of potential deviation from that expected<br />
value. Report<strong>in</strong>g of potential errors and ranges of possible<br />
values ra<strong>the</strong>r than s<strong>in</strong>gle numbers protects <strong>the</strong> <strong>in</strong>terest of<br />
everyone <strong>in</strong>volved and produces more reasonable expectations<br />
on <strong>the</strong> part of <strong>the</strong> general public.<br />
<strong>Restoration</strong> of <strong>the</strong> Elwha <strong>River</strong> <strong>in</strong> <strong>the</strong> state of Wash<strong>in</strong>gton<br />
exemplifi es <strong>the</strong> issues surround<strong>in</strong>g uncerta<strong>in</strong>ty <strong>in</strong> communication.<br />
National legislators directed that <strong>the</strong> river be<br />
restored to its ‘natural’ condition for <strong>the</strong> benefi t of endangered<br />
salmon, which use <strong>the</strong> river for spawn<strong>in</strong>g dur<strong>in</strong>g<br />
annual migrations (US House of Representatives, 1992).<br />
The method of restoration centers on <strong>the</strong> removal of two<br />
large dams. However, from a scientifi c standpo<strong>in</strong>t, it is<br />
impossible to meet <strong>the</strong> requirements of <strong>the</strong> law, because<br />
even with <strong>the</strong> removal of <strong>the</strong> dams, <strong>the</strong> river will not even<br />
approach truly natural conditions. Upstream land use,<br />
<strong>in</strong>clud<strong>in</strong>g logg<strong>in</strong>g, have altered <strong>the</strong> basic hydrologic and<br />
sediment regimes of <strong>the</strong> river, and <strong>in</strong> downstream areas<br />
levees and o<strong>the</strong>r structures prevent natural river processes.<br />
In fact, ‘natural’ conditions are a model to which restoration<br />
might aspire, but <strong>the</strong> pr<strong>in</strong>ciple as a true objective is<br />
unworkable.<br />
2.5 BIAS<br />
A fi nal source of uncerta<strong>in</strong>ty <strong>in</strong> science for river restoration<br />
lies with<strong>in</strong> <strong>the</strong> <strong>in</strong>tellect of <strong>the</strong> researchers <strong>the</strong>mselves.<br />
All geomorphologists are products of <strong>the</strong>ir cultural backgrounds,<br />
academic tra<strong>in</strong><strong>in</strong>g, experiences and personal<br />
characteristics (see Chapters 1 and 4). Researchers raised<br />
<strong>in</strong> <strong>in</strong>tellectually liberal surround<strong>in</strong>gs may be more question<strong>in</strong>g<br />
of established <strong>the</strong>ories than those raised <strong>in</strong> more<br />
rule-bound sett<strong>in</strong>gs, and each of <strong>the</strong>se types is likely to<br />
approach problems, evidence and methods with different<br />
biases and preferences. Gender may play a subtle, but as<br />
yet relatively unexplored role <strong>in</strong> our approaches to science.<br />
Academic tra<strong>in</strong><strong>in</strong>g for fl uvial geomorphology is highly<br />
variable from one group of teachers to ano<strong>the</strong>r, with some<br />
groups emphasiz<strong>in</strong>g a stochastic ra<strong>the</strong>r than a determ<strong>in</strong>istic<br />
approach. O<strong>the</strong>r contrast<strong>in</strong>g styles <strong>in</strong>clude greater<br />
emphasis on geographic approaches as opposed to those<br />
more strongly oriented toward eng<strong>in</strong>eer<strong>in</strong>g, research<br />
designs that emphasize small-scale (particle-sized) as<br />
opposed to ecosystem (or landscape) scale perspectives,<br />
or empirical ra<strong>the</strong>r than model-based approaches. Once<br />
tra<strong>in</strong>ed, <strong>the</strong> researcher is fur<strong>the</strong>r formed by personal fi eld<br />
experiences, with extensive backgrounds <strong>in</strong> dryland sett<strong>in</strong>gs<br />
produc<strong>in</strong>g a different perspective than experiences<br />
dom<strong>in</strong>ated by humid subtropical environments, tropical<br />
sett<strong>in</strong>gs or polar landscapes. F<strong>in</strong>ally, scientists are just like<br />
everyone else: some are stubborn, some are open m<strong>in</strong>ded;<br />
some are methodical while <strong>the</strong>ir colleagues make successful<br />
leaps of logic; and some are quick to reach judgements<br />
while o<strong>the</strong>rs seem never to reach closure on <strong>the</strong>ir<br />
conclusions.<br />
There is noth<strong>in</strong>g <strong>in</strong>herently wrong with biases, for to be<br />
biased is to be human. In river restoration, however, and<br />
especially when deal<strong>in</strong>g with decision makers, <strong>the</strong> wise<br />
scientists fi nd ways to communicate <strong>the</strong>ir biases to <strong>the</strong><br />
consumers of <strong>the</strong> scientifi c products. If <strong>the</strong> consumer<br />
(policy maker or <strong>in</strong>formed citizen) recognizes <strong>the</strong> biases<br />
and takes <strong>the</strong>m <strong>in</strong>to account, <strong>the</strong> <strong>in</strong>terpretation and application<br />
of <strong>the</strong> scientifi c products is confi dent by all parties<br />
<strong>in</strong>volved. Admission of biases by <strong>the</strong> researcher, usually<br />
<strong>in</strong> <strong>the</strong> form of a brief disclosure that clearly defi nes<br />
<strong>the</strong> schools of thought and experience of <strong>the</strong> researcher,<br />
enhances <strong>the</strong> professionalism of <strong>the</strong> <strong>in</strong>vestigator and fairly<br />
discloses to <strong>the</strong> consumer <strong>the</strong> background of <strong>the</strong> knowledge<br />
that forms <strong>the</strong> basis of decisions perta<strong>in</strong><strong>in</strong>g to public<br />
resources. In unusual cases, critics may contest decisions<br />
<strong>in</strong> adm<strong>in</strong>istrative hear<strong>in</strong>gs or <strong>in</strong> court proceed<strong>in</strong>gs, two<br />
venues where biases are likely to be revealed and explored<br />
<strong>in</strong> a combative environment. If biases have previously<br />
been revealed, <strong>the</strong>y lack s<strong>in</strong>ister overtones <strong>in</strong> a strategy<br />
that benefi ts both researcher and consumer.<br />
2.6 CONCLUSION<br />
<strong>Uncerta<strong>in</strong>ty</strong> is a fact of life as well as an <strong>in</strong>escapable<br />
feature of scientifi c research and decision mak<strong>in</strong>g for river<br />
restoration. Therefore, <strong>the</strong> researcher has two essential<br />
options: ei<strong>the</strong>r ignore <strong>the</strong> uncerta<strong>in</strong>ty and hope that it is<br />
not debilitat<strong>in</strong>g for <strong>the</strong> project at hand, or accept <strong>the</strong> uncerta<strong>in</strong>ty<br />
and use it as a feature of <strong>the</strong> research. The researcher<br />
can <strong>in</strong>vestigate <strong>the</strong> uncerta<strong>in</strong>ty, quantify it <strong>in</strong> some cases,<br />
and reveal it <strong>in</strong> explicit terms when report<strong>in</strong>g results. In<br />
this latter approach, uncerta<strong>in</strong>ty becomes an <strong>in</strong>tegral part<br />
of research for river restoration, a feature of <strong>the</strong> work that<br />
is a welcome challenge to be embraced and used to achieve<br />
a more effective end product. Project design may <strong>in</strong>clude<br />
a variety of channel dimensions and characteristics, for<br />
example, and avoid rely<strong>in</strong>g on a s<strong>in</strong>gle rigidly defi ned<br />
morphology, so that if orig<strong>in</strong>al understand<strong>in</strong>gs of <strong>the</strong><br />
system are not exactly correct <strong>the</strong> fi nal project will have<br />
some fl exibility. In o<strong>the</strong>r cases, it may be wise to simply<br />
allot more space for channel changes <strong>in</strong> <strong>the</strong> designed<br />
project to accommodate unforeseen adjustments. By<br />
deal<strong>in</strong>g directly with uncerta<strong>in</strong>ty, researcher and decision
maker <strong>in</strong>crease <strong>the</strong> probability <strong>in</strong> successfully restor<strong>in</strong>g a<br />
river with enhanced environmental and social benefi ts.<br />
REFERENCES<br />
Chang HH. 1998. Fluvial Processes <strong>in</strong> <strong>River</strong> Eng<strong>in</strong>eer<strong>in</strong>g.<br />
Krieger: Malabar, Florida.<br />
Collier MP, Webb RH, Andrews ED. 1997. Experimental fl ood<strong>in</strong>g<br />
<strong>in</strong> <strong>the</strong> Grand Canyon. Scientifi c American 276: 82–89.<br />
Costa JE, Miller AJ, Potter KW, Wilcock PR (Eds). 1995. Natural<br />
and Anthropogenic Infl uences <strong>in</strong> Fluvial Geomorphology: The<br />
Wolman Volume, Geophysical Monograph 89, American Geophysical<br />
Union: Wash<strong>in</strong>gton, DC.<br />
Graf WL. 1988. Defi nition of fl ood pla<strong>in</strong>s along arid-region<br />
rivers. In: Baker VR, Kochel RC, Patton PC (Eds), Flood<br />
Geomorphology, John Wiley & Sons, Inc.: New York, NY;<br />
231–242.<br />
Graf WL. 2001. Damage Control: Dams and <strong>the</strong> physical <strong>in</strong>tegrity<br />
of America’s rivers. Annals of <strong>the</strong> Association of American<br />
Geographers 91: 1–27.<br />
Sources of <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> Research 19<br />
Knighton D. 1998. Fluvial Forms and Processes: A New Perspective.<br />
Arnold: London.<br />
Leopold LB, Wolman MG, Miller JP. 1964. Fluvial Processes <strong>in</strong><br />
Geomorphology. WH Freeman: San Francisco, California.<br />
Leopold LB. 1994. A View of <strong>the</strong> <strong>River</strong>. Harvard University Press:<br />
Cambridge, Massachusetts.<br />
Nanson GC, Knighton AD. 1996. Anabranch<strong>in</strong>g rivers: Their<br />
cause, character and classifi cation. Earth Surface Processes<br />
and Landforms 21: 217–239.<br />
National Research Council. 1996. <strong>River</strong> Resource Management<br />
<strong>in</strong> <strong>the</strong> Grand Canyon. National Academy Press: Wash<strong>in</strong>gton,<br />
DC.<br />
National Research Council. 2004. Endangered and Threatened<br />
Species of <strong>the</strong> Platte <strong>River</strong>. National Academy Press: Wash<strong>in</strong>gton,<br />
D.C.<br />
Simons DB, Sentürk F. 1992. Sediment Transport Technology.<br />
Water Resources Publications: Littleton, Colorado.<br />
US House of Representatives. 1992. Jo<strong>in</strong>t Hear<strong>in</strong>gs on HR 4844,<br />
Elwha <strong>River</strong> Ecosystem and Fisheries <strong>Restoration</strong> Act, 102nd<br />
Congress, 2nd Session. Government Pr<strong>in</strong>t<strong>in</strong>g Offi ce: Wash<strong>in</strong>gton,<br />
DC.
3.1 INTRODUCTION<br />
<strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat Edited by Stephen Darby and David Sear<br />
© 2008 John Wiley & Sons, Ltd<br />
3<br />
The Scope of Uncerta<strong>in</strong>ties <strong>in</strong><br />
<strong>River</strong> <strong>Restoration</strong><br />
Joseph M. Wheaton 1 , Stephen E. Darby 2 and David A. Sear 2<br />
1 Institute of Geography and Earth Sciences, The University of Wales, UK<br />
2 School of Geography, University of Southampton, UK<br />
The science and practice of river restoration are both<br />
still very much <strong>in</strong> <strong>the</strong>ir adolescence (Palmer et al., 1997).<br />
Yet, both have been graced with fund<strong>in</strong>g and support<br />
from a diverse range of <strong>in</strong>terest groups (Malakoff, 2004).<br />
One of <strong>the</strong> premises of this book is that if fund<strong>in</strong>g is to<br />
cont<strong>in</strong>ue to be allocated to river restoration, it will have to<br />
be shown that river restoration is ‘work<strong>in</strong>g’ (see Preface;<br />
Wissmar and Bisson, 2003c). Defi nitions of ‘work<strong>in</strong>g’<br />
(often equated with success) are understandably subjective<br />
and vulnerable to uncerta<strong>in</strong>ties <strong>in</strong> <strong>the</strong> river restoration<br />
process, societal values, <strong>the</strong> fl uvial system and ecosystem<br />
response to restoration management activities. Davis and<br />
Slobodk<strong>in</strong> (2004) argued that defi n<strong>in</strong>g restoration goals<br />
and objectives is rightfully a value-based activity, as<br />
opposed to scientifi c activity. Each activity is <strong>in</strong>herently<br />
uncerta<strong>in</strong>. Paradoxically, <strong>the</strong> uncerta<strong>in</strong>ties <strong>in</strong>fl uenc<strong>in</strong>g<br />
river restoration projects are rarely recognised or quantifi<br />
ed, much less reported to stakeholders or <strong>the</strong> public<br />
(Walters, 1997).<br />
The topic of uncerta<strong>in</strong>ty <strong>in</strong> river restoration is riddled<br />
with complexity and confusion. Indeed, uncerta<strong>in</strong>ty manifests<br />
itself <strong>in</strong> many ways, as established <strong>in</strong> Chapters 1 and<br />
2. Lemons and Victor (see Chapter 1) have already illustrated<br />
how deep <strong>the</strong> value-laden dimensions of uncerta<strong>in</strong>ty<br />
lies, not just <strong>in</strong> decision mak<strong>in</strong>g, but <strong>in</strong> scientifi c research<br />
as well. Graf (see Chapter 2) expanded on this <strong>the</strong>me,<br />
cit<strong>in</strong>g uncerta<strong>in</strong>ties from <strong>the</strong>ories, <strong>the</strong> research itself, communication<br />
and biases among <strong>in</strong>vestigators. He concluded<br />
that <strong>the</strong> research can ei<strong>the</strong>r ‘ignore <strong>the</strong> uncerta<strong>in</strong>ty and<br />
hope that it is not debilitat<strong>in</strong>g for <strong>the</strong> project at hand,<br />
or accept <strong>the</strong> uncerta<strong>in</strong>ty and use it as a feature of <strong>the</strong><br />
research.’ In this chapter <strong>the</strong> rich topic of uncerta<strong>in</strong>ty is<br />
presented <strong>in</strong> a broader, more generic context. This foundation<br />
is <strong>in</strong>tended to help separate <strong>the</strong> sources and types of<br />
uncerta<strong>in</strong>ties that <strong>the</strong> various authors <strong>in</strong> this book present,<br />
and meanwhile unravel some of <strong>the</strong> ambiguities surround<strong>in</strong>g<br />
uncerta<strong>in</strong>ty <strong>in</strong> river restoration.<br />
As cautioned earlier, potentially signifi cant uncerta<strong>in</strong>ties<br />
are rarely recognised, much less explicitly dealt<br />
with <strong>in</strong> river restoration. Hence, a lexicon and typology<br />
for uncerta<strong>in</strong>ty is outl<strong>in</strong>ed fi rstly <strong>in</strong> this chapter. This is<br />
done to dispel <strong>the</strong> notion of a certa<strong>in</strong> world with certa<strong>in</strong><br />
outcomes with<strong>in</strong> <strong>the</strong> broader scope of types and sources<br />
of uncerta<strong>in</strong>ty. A return specifi cally to river restoration<br />
<strong>the</strong>n follows to identify types of uncerta<strong>in</strong>ties us<strong>in</strong>g <strong>the</strong><br />
above-mentioned typology. The tremendous diversity of<br />
river restoration <strong>in</strong> <strong>the</strong> context of uncerta<strong>in</strong>ties aris<strong>in</strong>g<br />
from restoration motives, notions and approaches are considered.<br />
A case will be made that a basic strategy for<br />
deal<strong>in</strong>g with uncerta<strong>in</strong>ty is needed by <strong>the</strong> river restoration<br />
community to allow both <strong>the</strong> community and <strong>in</strong>dividual<br />
<strong>in</strong>vestigators or practitioners to:<br />
explore <strong>the</strong> potential signifi cance (both <strong>in</strong> terms of<br />
unforeseen consequences and welcome surprises) or<br />
<strong>in</strong>signifi cance of uncerta<strong>in</strong>ties;<br />
effectively communicate uncerta<strong>in</strong>ties;<br />
eventually make adaptive, but transparent, decisions <strong>in</strong><br />
<strong>the</strong> face of uncerta<strong>in</strong>ty.
22 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
F<strong>in</strong>ally, it will be argued that amongst <strong>the</strong> various strategies<br />
for deal<strong>in</strong>g with uncerta<strong>in</strong>ty, <strong>the</strong> only strategy that might<br />
provide <strong>the</strong>se aims is one of embrac<strong>in</strong>g uncerta<strong>in</strong>ty.<br />
3.1.1 The Status Quo <strong>in</strong> <strong>River</strong> <strong>Restoration</strong><br />
The rapid rise and <strong>in</strong>ternational popularity of river restoration<br />
is both encourag<strong>in</strong>g and worrisome (Kondolf, 1996).<br />
Although sparse examples dat<strong>in</strong>g back to <strong>the</strong> 1930s exist1 ,<br />
river restoration has primarily developed on <strong>the</strong> coat tails<br />
of <strong>the</strong> environmental awareness movement of <strong>the</strong> late<br />
1970s (Graf, 1996; Sear, 1994). It is encourag<strong>in</strong>g that so<br />
much enthusiasm exists to restore rivers. Yet, it is <strong>in</strong>terest<strong>in</strong>g<br />
to note <strong>the</strong> societal choices between some mix of<br />
reactive restoration efforts <strong>in</strong> response to damage already<br />
done, as opposed to pro-active conservation actions to<br />
prevent fur<strong>the</strong>r damage (Boon, 1998). The <strong>in</strong>ternational<br />
popularity of river restoration is evident <strong>in</strong> <strong>the</strong> restoration<br />
literature (e.g. restoration <strong>in</strong> 21 different countries reported<br />
<strong>in</strong> Nijland and Cals, 2000), restoration databases (e.g. <strong>the</strong><br />
United K<strong>in</strong>gdom <strong>River</strong> <strong>Restoration</strong> Centre (RRC), United<br />
States Environmental Protection Agency (EPA)) 2 and an<br />
International <strong>River</strong> <strong>Restoration</strong> Survey3 launched by<br />
Wheaton et al. (2004c) with respondents from 34 different<br />
countries. In Denmark alone, 1068 restoration projects had<br />
been completed by Danish regional authorities by 1998<br />
(Hansen and Iversen, 1998); whereas <strong>in</strong> <strong>the</strong> United States,<br />
Malakoff (2004) reported that by 2004 more than $US10<br />
billion had been spent on a total of more than 30 000<br />
projects. The popularity of river restoration is apparent <strong>in</strong><br />
<strong>in</strong>ternational, national, regional and local public policy that<br />
actively promotes, requires and, <strong>in</strong> some cases, funds river<br />
restoration efforts (Jungwirth et al., 2002). However, <strong>the</strong>ir<br />
effectiveness is constra<strong>in</strong>ed by limited funds and scope to<br />
deal with closely related land use issues and o<strong>the</strong>r sociopolitical<br />
goals (Tockner and Stanford, 2002).<br />
Despite <strong>the</strong> popularity of river restoration <strong>in</strong> <strong>the</strong> developed<br />
nations of <strong>the</strong> world, <strong>the</strong> global decl<strong>in</strong>e of <strong>the</strong> physical<br />
and ecological <strong>in</strong>tegrity of rivers is diffi cult to overstate<br />
(Jungwirth et al., 2002; Vitousek et al., 1997). Indeed,<br />
1 The United States Department of Agriculture Forest Service<br />
started undertak<strong>in</strong>g ‘stream improvement’ <strong>in</strong> <strong>the</strong> 1930s with <strong>the</strong><br />
<strong>in</strong>tent of <strong>in</strong>creas<strong>in</strong>g salmonid production (Everest and Sedell,<br />
1984).<br />
2 RRC Database <strong>in</strong>cludes over 750 projects with<strong>in</strong> <strong>the</strong> United<br />
K<strong>in</strong>gdom: http://www.<strong>the</strong>rrc.co.uk; <strong>the</strong> USEPA <strong>River</strong> Corridor<br />
and Wetland <strong>Restoration</strong> Database <strong>in</strong>cludes over 600 projects<br />
throughout <strong>the</strong> United States: http://yosemite.epa.gov/water/<br />
restorat.nsf/rpd-2a.htm.<br />
3 Complete real time results, background <strong>in</strong>formation and forthcom<strong>in</strong>g<br />
<strong>in</strong>terpretations are available on <strong>the</strong> web: http://www.geog.<br />
soton.ac.uk/users/WheatonJ/<strong>Restoration</strong>Survey_Cover.asp.<br />
most restoration efforts still pale <strong>in</strong>to signifi cance relative<br />
to expand<strong>in</strong>g anthropogenic impacts on river<strong>in</strong>e landscapes<br />
(Tockner and Stanford, 2002). Even <strong>in</strong> parts of <strong>the</strong><br />
world where numerous river restoration efforts are already<br />
underway (i.e. Europe, North America and Australia), wetlands<br />
are actively be<strong>in</strong>g dra<strong>in</strong>ed and fi lled, rivers are still<br />
diverted and regulated, urban growth is encroach<strong>in</strong>g <strong>in</strong>to<br />
fl oodpla<strong>in</strong>s and headwaters, while we cont<strong>in</strong>ue to permanently<br />
alter bas<strong>in</strong> hydrology and fragment habitats (Coll<strong>in</strong>s<br />
et al., 2000; Moss, 2004; Mount, 1995). These problems<br />
pose even larger threats <strong>in</strong> <strong>the</strong> develop<strong>in</strong>g nations of<br />
<strong>the</strong> world (Marmulla, 2001). Over 250 new major dams<br />
become operational worldwide annually and 75 are<br />
planned for <strong>the</strong> Amazon Bas<strong>in</strong> alone (Rob<strong>in</strong>son et al.,<br />
2002). It seems logical that preservation should be easier<br />
to achieve than restoration (Frissell et al., 1993), but <strong>the</strong>re<br />
seems to be excessive confi dence <strong>in</strong> <strong>the</strong> ability to restore<br />
(Stewardson and Ru<strong>the</strong>rfurd, see Chapter 5), sometimes<br />
reduc<strong>in</strong>g restoration to a mitigation measure justify<strong>in</strong>g<br />
planned impacts or ma<strong>in</strong>ta<strong>in</strong><strong>in</strong>g <strong>the</strong> status quo. Both conservation<br />
and restoration are based on <strong>the</strong> transformation<br />
of uncerta<strong>in</strong> science and uncerta<strong>in</strong> notions of what is<br />
natural, ecosystem <strong>in</strong>tegrity and physical <strong>in</strong>tegrity <strong>in</strong>to<br />
societal goals (Graf, 2001; Lemons and Victor, see Chapter<br />
1). Additionally, <strong>the</strong> good <strong>in</strong>tentions of restoration projects<br />
may lead to un<strong>in</strong>tended but often foreseeable consequences.<br />
Even if society is will<strong>in</strong>g to make diffi cult sociopolitical<br />
decisions to support preservation and restoration<br />
of rivers, <strong>the</strong>re is no guarantee of desired outcomes<br />
follow<strong>in</strong>g.<br />
Given <strong>the</strong> dynamism of rivers, it seems obvious that <strong>the</strong><br />
outcomes of restoration projects are uncerta<strong>in</strong>. However,<br />
<strong>the</strong> restoration community seems hesitant to admit that <strong>the</strong><br />
goals and science that restoration are founded upon are<br />
uncerta<strong>in</strong> too (Stewardson and Ru<strong>the</strong>rfurd, see Chapter 5).<br />
Aside from <strong>in</strong>direct references to uncerta<strong>in</strong>ty <strong>in</strong> adaptive<br />
management programs, <strong>the</strong> river management community<br />
has largely brushed uncerta<strong>in</strong>ties aside (Clark, 2002;<br />
Wissmar and Bisson, 2003c). It is unclear whe<strong>the</strong>r this is<br />
a conscious or passive decision, though <strong>in</strong>dividual decisions<br />
to ignore uncerta<strong>in</strong>ty can be plausibly attributed to<br />
one or more of <strong>the</strong> follow<strong>in</strong>g:<br />
ignorance of uncerta<strong>in</strong>ty and/or its signifi cance;<br />
<strong>the</strong> hope that uncerta<strong>in</strong>ty is <strong>in</strong>signifi cant;<br />
an acknowledgement of uncerta<strong>in</strong>ty, but not know<strong>in</strong>g<br />
how to deal with it;<br />
be<strong>in</strong>g mis<strong>in</strong>formed about uncerta<strong>in</strong>ty, lead<strong>in</strong>g to <strong>the</strong><br />
assumption that it is <strong>in</strong>signifi cant;<br />
be<strong>in</strong>g knowledgeable about uncerta<strong>in</strong>ty, but hav<strong>in</strong>g<br />
established its <strong>in</strong>signifi cance.
Newson and Clark (see Chapter 14) attribute <strong>the</strong> river<br />
manager’s current treatment of uncerta<strong>in</strong>ty to a ‘riskaverse’<br />
management culture that prefers to entrench itself<br />
<strong>in</strong> ‘rituals of verifi cation’ aimed at m<strong>in</strong>imis<strong>in</strong>g liability<br />
(Power, 1999). <strong>Uncerta<strong>in</strong>ty</strong> is also frequently misunderstood<br />
by <strong>the</strong> general public (Pollack, 2003; Riebeek, 2002)<br />
as someth<strong>in</strong>g negative and undesirable (Newson and<br />
Clark, see Chapter 14). A widespread misconception that<br />
science embodies certa<strong>in</strong> knowledge persists <strong>in</strong> <strong>the</strong> reports<br />
of <strong>the</strong> ma<strong>in</strong>stream media and views of <strong>the</strong> general public<br />
(Clark, 2002; Riebeek, 2002). Such misconceptions fuel<br />
expectations that science-based approaches to river restoration<br />
will yield positive outcomes. Ironically, people confront<br />
uncerta<strong>in</strong>ties everyday without hostility and choose<br />
to rout<strong>in</strong>ely make decisions about <strong>the</strong> future (Pollack,<br />
2003).<br />
<strong>Restoration</strong> science and <strong>the</strong> restoration literature are not<br />
much fur<strong>the</strong>r along than practitioners and decision makers.<br />
Wissmar and Bisson (2003b) asserted that ‘a better understand<strong>in</strong>g<br />
of variability and uncerta<strong>in</strong>ty is critical to <strong>the</strong><br />
successful implementation of restoration programs for<br />
aquatic and riparian systems.’ Yet, buried with<strong>in</strong> a rich<br />
literature on restoration are only occasional pass<strong>in</strong>g<br />
mentions of uncerta<strong>in</strong>ty (Brookes and Shields, 1996) and<br />
a handful of explicit treatments (Johnson and Brown,<br />
2001; Johnson et al., 2002; Johnson and R<strong>in</strong>aldi, 1997;<br />
Johnson and R<strong>in</strong>aldi, 1998; Wissmar and Bisson, 2003c).<br />
These studies understandably tend to focus on a specifi c<br />
type of uncerta<strong>in</strong>ty that might be reasonably articulated<br />
with<strong>in</strong> a specifi ed page limit, so a more holistic treatment<br />
of uncerta<strong>in</strong>ty is necessary (Newson and Clark, see<br />
Chapter 14; Van Asselt, 2000). <strong>Restoration</strong> is established<br />
as one important component of environmental management.<br />
It would be a shame to lose what public support<br />
already exists for restoration if political scrut<strong>in</strong>y recasts<br />
unrealistic expectations of river restoration as a ‘failure’,<br />
as opposed to <strong>the</strong> <strong>in</strong>adequate consideration of uncerta<strong>in</strong>ty<br />
<strong>the</strong>y truly stem from.<br />
3.2 WHAT DO WE MEAN BY UNCERTAINTY?<br />
3.2.1 A Lexicon of <strong>Uncerta<strong>in</strong>ty</strong><br />
In <strong>the</strong> simplest sense, uncerta<strong>in</strong>ty is a lack of sureness<br />
about someth<strong>in</strong>g or someone (Merriam-Webster, 1994).<br />
However, uncerta<strong>in</strong>ty can be more than simply a lack<br />
of knowledge. It persists even <strong>in</strong> areas where knowledge<br />
is extensive; and knowledge does not necessarily equate<br />
to truth or certa<strong>in</strong>ty (Van Asselt and Rotmans, 2002).<br />
There are at least 24 potential synonyms for <strong>the</strong> noun<br />
uncerta<strong>in</strong>ty and 27 synonyms for <strong>the</strong> adjective uncerta<strong>in</strong><br />
The Scope of Uncerta<strong>in</strong>ties <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> 23<br />
(Table 3.1). There are a number of concepts related to and<br />
<strong>in</strong>fl uenced by uncerta<strong>in</strong>ty, but which differ from uncerta<strong>in</strong>ty<br />
itself. A selection of <strong>the</strong>se concepts is considered<br />
briefl y below.<br />
Accuracy: Accuracy refers to correctness or freedom from<br />
error. In measurement, accuracy refers to how close an<br />
<strong>in</strong>dividual measurement is to <strong>the</strong> ‘true’ or ‘correct’ value<br />
(Brown et al., 1994). The classic accuracy analogy is <strong>the</strong><br />
location of darts on a dart board – <strong>the</strong> closer <strong>the</strong> darts are<br />
to <strong>the</strong> <strong>in</strong>tended position (bull’s-eye) <strong>the</strong> more accurate. If<br />
one can be certa<strong>in</strong> about both <strong>the</strong> ‘true’ value (e.g. <strong>the</strong><br />
position of <strong>the</strong> bull’s-eye) and <strong>the</strong> value of <strong>the</strong> <strong>in</strong>dividual<br />
measurement (e.g. <strong>the</strong> position of <strong>the</strong> dart), <strong>the</strong>n <strong>the</strong> accuracy<br />
is actually a certa<strong>in</strong>ty. In practice, accuracy statements<br />
are uncerta<strong>in</strong> because ‘true’ values are often<br />
assumed and measurements have limited precision.<br />
Table 3.1 Potential synonyms of <strong>the</strong> noun ‘<strong>Uncerta<strong>in</strong>ty</strong>’ and<br />
<strong>the</strong> adjective ‘Uncerta<strong>in</strong>’<br />
Synonyms of <strong>Uncerta<strong>in</strong>ty</strong> Synonyms of Uncerta<strong>in</strong><br />
Ambiguity Ambiguous<br />
Indeterm<strong>in</strong>acy Causeless<br />
Capriciousness Capricious<br />
Chance Probabilistic<br />
– Deferred<br />
Danger Dangerous<br />
Disbelief Disbeliev<strong>in</strong>g<br />
Equivocation Equivocal<br />
Doubt Doubtful<br />
– Erratic<br />
Expectation –<br />
Future condition –<br />
Hesitation Hesitant<br />
Ignorance Ignorant<br />
Improbability Improbable<br />
Indecision Indecisive<br />
Indeterm<strong>in</strong>acy Indeterm<strong>in</strong>ant<br />
Insecurity Insecure<br />
Irresolution –<br />
Obscurity Obscure<br />
Surprise Surpris<strong>in</strong>g<br />
– Unau<strong>the</strong>ntic<br />
Un<strong>in</strong>telligibility Un<strong>in</strong>telligible<br />
– Unexpla<strong>in</strong>ed<br />
– Questionable<br />
Vacillation Vacillat<strong>in</strong>g<br />
Vagueness Vague<br />
– Undecided<br />
Unsureness Unsure<br />
Unpredictability Unpredictable
24 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Confi dence: Confi dence (e.g. <strong>in</strong> a statement, hypo<strong>the</strong>sis,<br />
measurement, feel<strong>in</strong>g or notion) relates to <strong>the</strong> degree of<br />
belief or level of certa<strong>in</strong>ty. Confi dence levels, for example,<br />
describe <strong>the</strong> probability that a given population parameter<br />
estimate falls with<strong>in</strong> a designated cont<strong>in</strong>uous statistical<br />
confi dence <strong>in</strong>terval.<br />
Divergence: Divergence describes a situation when similar<br />
causes produce dissimilar effects (Schumm, 1991). Divergence<br />
relates to uncerta<strong>in</strong>ty <strong>in</strong> situations where problems<br />
of cause and process are under consideration.<br />
Error: Error is <strong>the</strong> difference between a measured or calculated<br />
value and a ‘true’ value. In every day conversation,<br />
an error is a mistake. In science, error is <strong>the</strong> metric by<br />
which accuracy is reported and is not a synonym for uncerta<strong>in</strong>ty<br />
(Ellison et al., 2000). A ‘true’ value is certa<strong>in</strong> by<br />
defi nition. If <strong>the</strong> error between <strong>the</strong> ‘true’ value and a<br />
measured or calculated value is known <strong>the</strong>re is no uncerta<strong>in</strong>ty<br />
<strong>in</strong> pr<strong>in</strong>ciple. However, <strong>in</strong> practice ‘true’ values are<br />
often not known and <strong>in</strong>stead are assumed to be so, while<br />
<strong>the</strong> measured or calculated value may be uncerta<strong>in</strong>. Hence<br />
error becomes representative of uncerta<strong>in</strong>ty. Once errors<br />
are calculated, it can be helpful to consider whe<strong>the</strong>r <strong>the</strong><br />
error is systematic or random. Systematic errors stem from<br />
consistent mistakes and are often constant or predictable,<br />
affect<strong>in</strong>g <strong>the</strong> mean of a sample (i.e. bias, Trochim, 2000).<br />
Systematic errors can potentially be constra<strong>in</strong>ed as <strong>the</strong>ir<br />
source is identifi able. By contrast, random errors <strong>in</strong>fl uence<br />
<strong>the</strong> variability of a sample (not <strong>the</strong> mean) and are generally<br />
unpredictable or unconstra<strong>in</strong>able (Trochim, 2000).<br />
Exactness: Exactness is really a synonym for accuracy.<br />
However, it is worth po<strong>in</strong>t<strong>in</strong>g out that exactness has quite<br />
a different mean<strong>in</strong>g to exact. Exact statements or exact<br />
numbers, <strong>in</strong> pr<strong>in</strong>ciple, have no uncerta<strong>in</strong>ty about <strong>the</strong>m.<br />
They are statements of truth. By contrast, exactness is a<br />
relative measurement assigned to <strong>in</strong>exact statements or<br />
values (i.e. those with some uncerta<strong>in</strong>ty).<br />
Expectation: Expectation has to do with anticipation of<br />
probable or certa<strong>in</strong> events. <strong>Uncerta<strong>in</strong>ty</strong> fundamentally<br />
relates to expectations. When uncerta<strong>in</strong>ties are unknown,<br />
not fully considered or ignored, <strong>the</strong> degree that expectations<br />
may be unrealistic will generally <strong>in</strong>crease.<br />
Equifi nality: Equifi nality (also referred to as convergence),<br />
arises when different processes and causes produce<br />
similar effects (Schumm, 1991). In a modell<strong>in</strong>g context,<br />
Beven (1996a; 1996b) suggests that ‘<strong>the</strong> consequences of<br />
equifi nality are uncerta<strong>in</strong>ty <strong>in</strong> <strong>in</strong>ference and prediction.’ In<br />
a social context, a potentially limitless range of possibilities<br />
may lead to a s<strong>in</strong>gle event, such as <strong>the</strong> election or<br />
defeat of a politician.<br />
Precision: Precision is a measure of how closely <strong>in</strong>dividual<br />
measurements or calculations match one ano<strong>the</strong>r<br />
(Brown et al., 1994). Recall<strong>in</strong>g <strong>the</strong> dart board analogy<br />
from accuracy, a precisely thrown set of darts will cluster<br />
around one ano<strong>the</strong>r, but may be nowhere near <strong>the</strong> bull’seye.<br />
In measurement, <strong>the</strong> precision of an <strong>in</strong>strument refers<br />
to <strong>the</strong> fi nest scalar unit <strong>the</strong> <strong>in</strong>strument can resolve. Precision<br />
is related to uncerta<strong>in</strong>ty <strong>in</strong> that it defi nes a detection<br />
threshold, below which differences can not be discerned.<br />
Repeatability: Repeatability can be viewed as ei<strong>the</strong>r <strong>the</strong><br />
ability to reproduce <strong>the</strong> same measurement, result or<br />
calculation or <strong>the</strong> variability <strong>in</strong> repeated measurements,<br />
results or calculations. <strong>Uncerta<strong>in</strong>ty</strong> can simply limit<br />
repeatability or <strong>in</strong>crease variability.<br />
Risk: Risk is a measure of likelihood that an undesirable<br />
event or hazard will occur (Merriam-Webster, 1994). Ward<br />
(1998) credited Knight (1921) for mak<strong>in</strong>g <strong>the</strong> important<br />
clarifi cation between risk and <strong>the</strong> type of uncerta<strong>in</strong>ty for<br />
which <strong>the</strong>re exists ‘no valid basis of any k<strong>in</strong>d for classify<strong>in</strong>g<br />
<strong>in</strong>stances’:<br />
‘He used <strong>the</strong> term “risk” for situations <strong>in</strong> which an<br />
<strong>in</strong>dividual may not know <strong>the</strong> outcome of an event,<br />
but can form realistic expectations of <strong>the</strong> probabilities<br />
of <strong>the</strong> various possible outcomes based ei<strong>the</strong>r on<br />
ma<strong>the</strong>matical calculations or <strong>the</strong> history of previous<br />
occurrences.’<br />
Newson and Clark (see Chapter 14) contrast risk (with<br />
‘known’ impacts and probabilities) with uncerta<strong>in</strong>ty (with<br />
‘known’ impacts but ‘unknown’ probabilities) and ignorance<br />
(with ‘unknown’ impacts and probabilities).<br />
It is worth not<strong>in</strong>g that uncerta<strong>in</strong>ty itself and all <strong>the</strong><br />
related concepts outl<strong>in</strong>ed above are described <strong>in</strong> terms of<br />
<strong>the</strong>ir ‘degree’. That is, none of <strong>the</strong>se concepts are simple<br />
Aristotelian two-valued logic concepts (e.g. true–false).<br />
Each concept is measured along a cont<strong>in</strong>uum of values<br />
with end-members of total uncerta<strong>in</strong>ty (complete irreducible<br />
ignorance) and absolute certa<strong>in</strong>ty. Probabilistic uncerta<strong>in</strong>ty<br />
is an example of a quantifi cation of uncerta<strong>in</strong>ty, yet<br />
not all uncerta<strong>in</strong>ty is quantifi able. To quantify uncerta<strong>in</strong>ty<br />
it is necessary to estimate <strong>the</strong> degree of our limited knowledge.<br />
Yet, if a condition of irreducible ignorance is considered<br />
as one extreme of uncerta<strong>in</strong>ty, it is diffi cult at best<br />
to estimate <strong>the</strong> degree of someth<strong>in</strong>g we do not even know<br />
exists. With<strong>in</strong> this broad view of uncerta<strong>in</strong>ty, uncerta<strong>in</strong>ty<br />
might also be considered along a cont<strong>in</strong>uum that refl ects<br />
our ability to quantify it (Figure 3.1).<br />
In summary, when someone mentions uncerta<strong>in</strong>ty casually,<br />
it is diffi cult to discern whe<strong>the</strong>r <strong>the</strong>y are referr<strong>in</strong>g to
limited knowledge, a lack of knowledge altoge<strong>the</strong>r, or one<br />
of <strong>the</strong> above-mentioned concepts that are <strong>in</strong>fl uenced by<br />
uncerta<strong>in</strong>ty. Moreover, <strong>the</strong> lexicon provided here conta<strong>in</strong>s<br />
concepts that are highly <strong>in</strong>ter-related and easily confused.<br />
Similar to vague, pseudo-scientifi c buzzwords and catchall<br />
phrases like holistic and <strong>in</strong>tegrated, <strong>the</strong> term ‘uncerta<strong>in</strong>ty’<br />
alone evidently has little mean<strong>in</strong>g until its details<br />
are unravelled.<br />
3.2.2 A Typology for <strong>Uncerta<strong>in</strong>ty</strong><br />
S<strong>in</strong>ce uncerta<strong>in</strong>ty is so hard to defi ne, a classifi cation of<br />
uncerta<strong>in</strong>ty is often used (Van Asselt and Rotmans, 2002).<br />
The utility of any typology or classifi cation is ultimately<br />
dependent on its application (Kondolf, 1995b; Lew<strong>in</strong>,<br />
2001). Rotmans and van Asselt (2001) astutely po<strong>in</strong>ted<br />
out ‘<strong>the</strong>re is not one overall typology that satisfactorily<br />
covers all sorts of uncerta<strong>in</strong>ties, but that <strong>the</strong>re are many<br />
possible typologies’. In <strong>the</strong> context of this review, a<br />
typology was sought which considered sources of uncerta<strong>in</strong>ty<br />
and did not unnecessarily ignore any type of<br />
uncerta<strong>in</strong>ty. The exist<strong>in</strong>g van Asselt (2000) typology was<br />
chosen over o<strong>the</strong>rs because of its generic and <strong>in</strong>clusive<br />
consideration of uncerta<strong>in</strong>ty. The typology was fi rst <strong>in</strong>troduced<br />
<strong>in</strong> detail <strong>in</strong> van Asselt (2000) and concisely reviewed<br />
<strong>in</strong> Rotmans and van Asselt (2001) and van Asselt and<br />
Rotmans (2002).<br />
At <strong>the</strong> highest level, two sources of uncerta<strong>in</strong>ty exist:<br />
uncerta<strong>in</strong>ty due to variability and uncerta<strong>in</strong>ty due to<br />
limited knowledge (Figure 3.2). Van Asselt and Rotmans<br />
(2002) presented uncerta<strong>in</strong>ty due to variability fi rst as<br />
<strong>the</strong>se uncerta<strong>in</strong>ties ultimately comb<strong>in</strong>e to contribute to<br />
uncerta<strong>in</strong>ty due to limited knowledge. Environmental<br />
management is concerned with <strong>the</strong> management of <strong>in</strong>herently<br />
variable natural and managed systems. Knowledge<br />
The Scope of Uncerta<strong>in</strong>ties <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> 25<br />
Figure 3.1 The quantifi able cont<strong>in</strong>uum of uncerta<strong>in</strong>ty (Once uncerta<strong>in</strong>ties are acknowledge as unquantifi ed uncerta<strong>in</strong>ties, <strong>in</strong>creased<br />
knowledge about <strong>the</strong> uncerta<strong>in</strong>ties will determ<strong>in</strong>e <strong>the</strong>ir position on <strong>the</strong> cont<strong>in</strong>uum.)<br />
about natural change and variability <strong>in</strong> ecosystems, fl uvial<br />
systems and hydrologic systems is <strong>in</strong>complete and hence<br />
contributes to uncerta<strong>in</strong>ty due to limited knowledge<br />
(Wissmar and Bisson, 2003a). Five dist<strong>in</strong>ct subclasses<br />
of uncerta<strong>in</strong>ty due to variability are proposed: <strong>in</strong>herent<br />
natural randomness, value diversity (socio-political),<br />
behavioural diversity, societal randomness and technological<br />
surprise. Inherent natural randomness is attributed to<br />
‘<strong>the</strong> nonl<strong>in</strong>ear, chaotic and unpredictable nature of natural<br />
processes’. The natural variability of river systems should<br />
be a fundamental consideration <strong>in</strong> <strong>in</strong>tegrated river bas<strong>in</strong><br />
management and restoration; it is reviewed thoroughly <strong>in</strong><br />
Wissmar and Bisson (2003c). Value diversity, behavioural<br />
diversity and societal randomness each contribute to<br />
uncerta<strong>in</strong>ties <strong>in</strong> environmental management, particularly<br />
through stakeholder negotiations, public support, project<br />
fund<strong>in</strong>g, policy mak<strong>in</strong>g and <strong>in</strong>dividual perspectives. Technological<br />
surprises result from new breakthroughs, which<br />
may provide unforeseen benefi ts and/or br<strong>in</strong>g unforeseen<br />
consequences.<br />
Van Asselt and Rotmans (2002) separated seven types<br />
of uncerta<strong>in</strong>ty due to limited knowledge. Unlike uncerta<strong>in</strong>ties<br />
due to variability, <strong>the</strong>se are thought to map out<br />
along a cont<strong>in</strong>uum that refl ects <strong>the</strong> relative degree of<br />
uncerta<strong>in</strong>ty. At <strong>the</strong> highest degree of uncerta<strong>in</strong>ty are four<br />
‘structural uncerta<strong>in</strong>ties’ (van Asselt and Rotmans,<br />
2002):<br />
Irreducible ignorance: ‘We cannot know.’<br />
Indeterm<strong>in</strong>acy: ‘We will never know.’<br />
Reducible ignorance: ‘We do not know what we do not<br />
know.’<br />
Confl ict<strong>in</strong>g evidence: Knowledge is not fact but <strong>in</strong>terpretation,<br />
and <strong>in</strong>terpretations frequently contradict and<br />
challenge each o<strong>the</strong>r. ‘We don’t know what we know.’
26 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Figure 3.2 Typology for sources and degree of uncerta<strong>in</strong>ty (Adapted from Van Asselt’s (2000) proposed typology for uncerta<strong>in</strong>ties<br />
<strong>in</strong> <strong>in</strong>tegrated assessment.)<br />
Van Asselt and Rotmans (2002) <strong>the</strong>n proposed a transition<br />
<strong>in</strong>to ‘unreliability’ uncerta<strong>in</strong>ties of a relatively lesser<br />
degree:<br />
Practically immeasurable: A lack of data or <strong>in</strong>formation<br />
is always a reality <strong>in</strong> study<strong>in</strong>g natural systems.<br />
Not only are many natural phenomena <strong>in</strong>credibly diffi<br />
cult or impossible to measure, all are fundamentally<br />
limited by problems of temporal and spatial resolution,<br />
up-scal<strong>in</strong>g and averag<strong>in</strong>g (Kavvas, 1999). ‘We<br />
know what we don’t know’ (Van Asselt and Rotmans,<br />
2002).<br />
Lack of Observations and Measurements: Although<br />
<strong>in</strong> pr<strong>in</strong>ciple this is easy to identify and augment, <strong>in</strong><br />
practice this is always a factor. Borrow<strong>in</strong>g from van<br />
Asselt and Rotmans (2002): ‘could have, should<br />
have, would have, but didn’t.’<br />
Inexactness: Related to lack of precision, lack of<br />
accuracy, measurement and calculation errors. Under<br />
Klir and Yuan’s (1995) typology, <strong>the</strong>se are considered<br />
‘fuzz<strong>in</strong>ess’ or vagueness.<br />
The van Asselt (2000) typology is both more general<br />
and detailed than o<strong>the</strong>r typologies such as Klir and Yuan<br />
(1995). However, all provide a reasonable means to deal<br />
with <strong>the</strong> fi rst step to understand<strong>in</strong>g uncerta<strong>in</strong>ty. Namely,<br />
<strong>the</strong>y allow a systematic identifi cation of sources and types<br />
of uncerta<strong>in</strong>ties that could work <strong>in</strong> ei<strong>the</strong>r <strong>in</strong>dividual river<br />
restoration projects or <strong>in</strong>ternational policy mak<strong>in</strong>g on<br />
water and environmental management (see also Chapters<br />
1 and 2). In practice, it is recognised that <strong>the</strong> semantics of<br />
uncerta<strong>in</strong>ty will always be <strong>in</strong>terpreted differently <strong>in</strong> different<br />
professional contexts (Newson and Clark, see Chapter<br />
14). However, with<strong>in</strong> <strong>the</strong> context of this chapter, <strong>the</strong> van<br />
Asselt (2000) typology and associated mean<strong>in</strong>gs will be<br />
used consistently.<br />
3.2.3 How do Knowledge and <strong>Uncerta<strong>in</strong>ty</strong> Relate?<br />
The positivist view (Van Asselt and Rotmans, 2002) contends<br />
that as knowledge <strong>in</strong>creases, uncerta<strong>in</strong>ty decreases.<br />
Brookes et al. (1998) made <strong>the</strong> more restrictive but<br />
contradictory generalisation that ‘as knowledge relat<strong>in</strong>g<br />
to rivers and <strong>the</strong>ir fl oodpla<strong>in</strong>s <strong>in</strong>creases, uncerta<strong>in</strong>ty is<br />
<strong>in</strong>creased ra<strong>the</strong>r than decreased.’ So, which is it? In reality,<br />
<strong>the</strong>re is no unique relationship between uncerta<strong>in</strong>ty and<br />
knowledge (Van Asselt and Rotmans, 2002), nor is uncerta<strong>in</strong>ty<br />
a fi xed quantity that will always be reduced by scientifi<br />
c research (Jamieson, 1996). It is a highly contextual<br />
relationship dependent on <strong>the</strong> type of uncerta<strong>in</strong>ty (i.e.<br />
uncerta<strong>in</strong>ty due to lack of knowledge versus variability)<br />
and <strong>the</strong> specifi c circumstances under consideration. A few
examples of potential relationships between knowledge<br />
and uncerta<strong>in</strong>ty us<strong>in</strong>g <strong>the</strong> nomenclature of <strong>the</strong> van<br />
Asslet typology are illustrated <strong>in</strong> Figure 3.3. Hav<strong>in</strong>g<br />
established <strong>the</strong> basic term<strong>in</strong>ology of uncerta<strong>in</strong>ty, it is<br />
possible to discuss <strong>the</strong> sources of uncerta<strong>in</strong>ty with<strong>in</strong><br />
river restoration.<br />
The Scope of Uncerta<strong>in</strong>ties <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> 27<br />
3.3 REVISITING RIVER RESTORATION<br />
AND UNCERTAINTY<br />
It is diffi cult to generalise about <strong>the</strong> importance of uncerta<strong>in</strong>ty<br />
simply because restoration activities and <strong>the</strong> restoration<br />
community itself are so diverse. The stakeholders who<br />
Figure 3.3 Some potential relationships between knowledge and uncerta<strong>in</strong>ty through time (Contrary to <strong>the</strong> argument of <strong>the</strong> positivist,<br />
no unique <strong>in</strong>verse relationship between uncerta<strong>in</strong>ty and knowledge exists.)
28 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
<strong>in</strong>itiate river restoration projects <strong>in</strong>clude private <strong>in</strong>dividuals,<br />
non-governmental organisations (NGOs), governmental<br />
organisations and various collaborative comb<strong>in</strong>ations<br />
of <strong>the</strong> above. The restoration community is also comprised<br />
of practitioners, decision makers and scientists. No attempt<br />
is made here to list ‘all’ <strong>the</strong> uncerta<strong>in</strong>ties encountered<br />
throughout <strong>the</strong> restoration process as <strong>the</strong> daunt<strong>in</strong>g list<br />
would never be comprehensive, and is entirely perspective<br />
and project specifi c. For example, <strong>the</strong>re is little consensus<br />
over <strong>the</strong> mean<strong>in</strong>g of <strong>the</strong> term ‘river restoration’ with at<br />
least 30 different authors propos<strong>in</strong>g different defi nitions<br />
(Lemons and Victor, see Chapter 1; NAP, 2002; Newson,<br />
2002; Sear, 1994; Stockwell, 2000). Similar to Shields<br />
et al. (2003), ‘river restoration’ <strong>in</strong> this book is used as a<br />
catch-all term for a variety of management responses and<br />
activities used to address perceived problems with rivers<br />
(Kondolf, 1996). As a start<strong>in</strong>g po<strong>in</strong>t, a generic decision<br />
process, which most restoration projects loosely follow,<br />
highlight<strong>in</strong>g some of <strong>the</strong> common sources of uncerta<strong>in</strong>ty<br />
is mapped out <strong>in</strong> Table 3.2.<br />
3.3.1 Motives for <strong>Restoration</strong><br />
Once river restoration projects ga<strong>in</strong> momentum, it is easy<br />
to lose sight of why <strong>the</strong>y were orig<strong>in</strong>ally envisioned<br />
(Stewardson and Ru<strong>the</strong>rfurd, see Chapter 5). Here, <strong>the</strong><br />
motives for restoration are considered to represent more<br />
generalised aims than formalised and specifi c restoration<br />
objectives and activities (i.e. <strong>the</strong> ‘why’ <strong>in</strong>stead of <strong>the</strong><br />
‘what’). Eight common types of motives for river restoration<br />
(still o<strong>the</strong>rs exist) are listed below:<br />
1. Ecosystem <strong>Restoration</strong><br />
2. Habitat <strong>Restoration</strong><br />
3. Flood Control/Defence<br />
4. Floodpla<strong>in</strong> Reconnection<br />
5. Property and Infrastructure Protection (bank<br />
stability)<br />
6. Sediment Management<br />
7. Water Quality<br />
8. Aes<strong>the</strong>tic and Recreational.<br />
Considerable overlap exists between many of <strong>the</strong> above.<br />
For example, fl oodpla<strong>in</strong> reconnection can be a type of<br />
fl ood control. Habitat restoration and water quality restoration<br />
are sometimes considered forms of ecosystem<br />
restoration. In ano<strong>the</strong>r example, water quality restoration<br />
could be viewed by some as sediment management or<br />
by o<strong>the</strong>rs as aes<strong>the</strong>tic or recreational restoration. Thus, a<br />
hierarchical organisation of restoration motives would be<br />
highly subjective and dependent on <strong>in</strong>dividual values and<br />
perspectives. This <strong>in</strong> itself is not necessarily problematic.<br />
However, it represents a form of communication uncerta<strong>in</strong>ty<br />
aris<strong>in</strong>g out of value diversity which is often taken<br />
for granted. Once <strong>the</strong> motives (why to do it) for restoration<br />
are established, restoration aims fall <strong>in</strong>to place, but more<br />
specifi c objectives (what and how to do it) require careful<br />
consideration.<br />
Table 3.2 Sources of uncerta<strong>in</strong>ty <strong>in</strong> an environmental management decision process structure (Adapted from Chapman & Ward<br />
(2002))<br />
Stage <strong>in</strong> Decision Process <strong>Uncerta<strong>in</strong>ty</strong> About<br />
Monitor <strong>the</strong> environment and current operations Completeness, veracity and accuracy of <strong>in</strong>formation received, mean<strong>in</strong>g<br />
with<strong>in</strong> <strong>the</strong> organisation<br />
of <strong>in</strong>formation, <strong>in</strong>terpretation of implications<br />
Recognise an issue Signifi cance of issue, urgency, need for action<br />
Scope <strong>the</strong> Decision Appropriate frame of reference, scope of relevant organisation activities,<br />
who is <strong>in</strong>volved, who should be <strong>in</strong>volved, extent of separation from<br />
o<strong>the</strong>r decision issues<br />
Determ<strong>in</strong>e <strong>the</strong> performance criteria Relevant performance criteria, whose criteria, appropriate metrics,<br />
appropriate priorities and trade offs between different criteria<br />
Identify alternative courses of action† Nature of alternatives available (scope, tim<strong>in</strong>g, logistics <strong>in</strong>volved), what<br />
is possible, level of detail required, time available to identify<br />
alternatives<br />
Predict <strong>the</strong> outcomes of courses of action† Consequences, nature of <strong>in</strong>fl uenc<strong>in</strong>g factors, size of <strong>in</strong>fl uenc<strong>in</strong>g factors,<br />
effects and <strong>in</strong>teractions between <strong>in</strong>fl uenc<strong>in</strong>g factors (variability and<br />
tim<strong>in</strong>g), nature and signifi cance of assumptions made<br />
Choose a course of action How to weigh and compare predicted outcomes<br />
Implement <strong>the</strong> chosenalternative* How alternatives will work <strong>in</strong> practice<br />
Monitor and reviewperformance‡ What to monitor, how often to monitor, when to take fur<strong>the</strong>r action<br />
† = Most decision support systems only provide <strong>in</strong>put at <strong>the</strong>se levels; * = The precautionary pr<strong>in</strong>ciple is implemented here; ‡ = Adaptive management<br />
starts here and feeds back through <strong>the</strong> process as necessary.
Many have argued that uncerta<strong>in</strong>ty <strong>in</strong> assess<strong>in</strong>g restoration<br />
success arises from <strong>in</strong>adequate, vague and unclear<br />
restoration objectives (Jungwirth et al., 2002; Kondolf,<br />
1995a; Sk<strong>in</strong>ner et al., see Chapter 10). Motives may serve<br />
well as aims (not necessarily to be achieved by an <strong>in</strong>dividual<br />
project) but <strong>the</strong>y are <strong>in</strong>suffi cient to act as detailed<br />
project objectives, which <strong>in</strong> pr<strong>in</strong>ciple should be achievable.<br />
Us<strong>in</strong>g restoration motives carelessly as objectives<br />
produces unrealistic expectations. For example, <strong>in</strong> a recent<br />
request for proposals to fund community-based river restoration<br />
projects by American <strong>River</strong>s and <strong>the</strong> National<br />
Oceanic and Atmospheric Association, applicants were<br />
asked to demonstrate that <strong>the</strong>ir project: will successfully<br />
restore anadromous fi sh habitat, access to exist<strong>in</strong>g<br />
anadromous fi sh habitat, or natural river<strong>in</strong>e functions; is<br />
<strong>the</strong> correct approach, based on ecological, social, economic,<br />
and eng<strong>in</strong>eer<strong>in</strong>g considerations; will m<strong>in</strong>imise any<br />
identifi able short or long term negative impacts to <strong>the</strong> river<br />
system as a result of <strong>the</strong> project . . .’<br />
The problem with requir<strong>in</strong>g an applicant to make such<br />
bold statements about <strong>in</strong>dividual projects is that it asserts<br />
a level of confi dence <strong>in</strong> restoration simply not warranted<br />
by current science or practice and creates unrealistic<br />
expectations 4 (Stewardson and Ru<strong>the</strong>rfurd, see Chapter 5).<br />
Subtly reword<strong>in</strong>g such requirements to account for uncerta<strong>in</strong>ty<br />
could help recast river restoration <strong>in</strong> a tone commensurate<br />
with our abilities and uncerta<strong>in</strong>ties. Interest<strong>in</strong>gly,<br />
<strong>the</strong>se objectives are consistent with Clark’s (2002) critical<br />
synopsis of Predictive Management as opposed to<br />
adaptive management as <strong>the</strong> current model <strong>in</strong> river<br />
management.<br />
The restoration community has burdened itself with <strong>the</strong><br />
idea that restoration objectives should be scientifi cally<br />
based (Davis and Slobodk<strong>in</strong>, 2004). While science surely<br />
has an important role <strong>in</strong> restoration, Davis and Slobodk<strong>in</strong><br />
(2004) argued that determ<strong>in</strong><strong>in</strong>g restoration objectives<br />
is fundamentally a value-based and subjective process.<br />
Noth<strong>in</strong>g is seen as <strong>in</strong>herently wrong with this reality, so<br />
long as it is transparently recognised. From an uncerta<strong>in</strong>ty<br />
perspective, this means that restoration objectives are<br />
<strong>the</strong>refore sources of uncerta<strong>in</strong>ty due to variability; namely<br />
value diversity, behavioural diversity and societal randomness.<br />
For example, <strong>the</strong> fate of 81 000 hectares of forest<br />
land allocated for ecosystem restoration around <strong>the</strong> city of<br />
Chicago, Ill<strong>in</strong>ois has pitted two ‘environmental’ groups<br />
aga<strong>in</strong>st each o<strong>the</strong>r based on <strong>the</strong>ir contrast<strong>in</strong>g notions of<br />
‘what is natural’. The divergent environmental views are<br />
essentially split between preservationists, who wish to preserve<br />
<strong>the</strong> forest land planted <strong>in</strong> <strong>the</strong> 1800s, and restoration-<br />
4 This is fundamentally a communication uncerta<strong>in</strong>ty result<strong>in</strong>g<br />
from socio-political value diversity (see Figure 3.1).<br />
The Scope of Uncerta<strong>in</strong>ties <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> 29<br />
ists, who want to restore <strong>the</strong> pre-settlement (1830s) prairie<br />
and savannah (Alario and Brün, 2001). Both evoke emotional<br />
arguments, which can be supported on scientifi c<br />
grounds. ‘Which is right?’ is <strong>the</strong> wrong question to ask of<br />
science. Alario and Brün (2001) concluded that <strong>the</strong> appropriate<br />
arena to decide such an issue is a political decision<br />
mak<strong>in</strong>g process.<br />
3.3.2 Notions that Drive <strong>Restoration</strong><br />
Underly<strong>in</strong>g motives for river restoration and <strong>the</strong> eventual<br />
specifi c techniques tried to achieve <strong>the</strong>m are some very<br />
basic, yet highly uncerta<strong>in</strong> notions. S<strong>in</strong>ce <strong>the</strong>se basic<br />
notions are rarely questioned, it is important to highlight<br />
how <strong>the</strong>y <strong>in</strong>troduce uncerta<strong>in</strong>ty. Notions are also known<br />
as ‘Lietbilds’ – or target visions – and have ga<strong>in</strong>ed widespread<br />
acceptance <strong>in</strong> <strong>the</strong> restoration literature (Hughes,<br />
1995; Jungwirth et al., 2002; Kern, 1992). Notions, such<br />
as those <strong>in</strong> Table 3.3, that drive restoration strategies<br />
are frequently based on societal values and beliefs, or on<br />
popular, but by no means certa<strong>in</strong>, scientifi c paradigms<br />
(Davis and Slobodk<strong>in</strong>, 2004; McDonald et al., 2004;<br />
Rhoads et al., 1999).<br />
Falkenmark and Folke (2002) argued that susta<strong>in</strong>able<br />
catchment management must be based on ethical pr<strong>in</strong>ciples.<br />
They suggest that management based on scientifi c<br />
pr<strong>in</strong>ciples alone is primarily concerned with ‘do<strong>in</strong>g <strong>the</strong><br />
th<strong>in</strong>g right’, whereas notions that drive restoration strategies<br />
are actually driven by ‘do<strong>in</strong>g <strong>the</strong> right th<strong>in</strong>g.’ It is a<br />
presumption that good ethical practice generally translates<br />
<strong>in</strong>to good biological practice (Pister, 2001). Hence notions<br />
are vague ideas, perhaps based on scientifi c knowledge,<br />
but primarily supported by ethical beliefs and societal<br />
values. The restoration literature is rarely explicit <strong>in</strong> dist<strong>in</strong>guish<strong>in</strong>g<br />
<strong>the</strong> notions it advocates from <strong>the</strong> science used<br />
to support it. Phillip Williams (personal communication)<br />
asserts that ‘rigour’ <strong>in</strong> restoration plann<strong>in</strong>g should start<br />
with <strong>the</strong> development of an explicit conceptual model<br />
transparently describ<strong>in</strong>g our notions of how <strong>the</strong> river<br />
system functions5 . Such a conceptual model should identify<br />
both <strong>the</strong> historical context and <strong>the</strong> present day limitations<br />
(i.e. uncerta<strong>in</strong>ties). Wheaton et al. (2004a) argued<br />
that numerous conceptual models <strong>in</strong> <strong>the</strong> scientifi c literature<br />
already exist and can be borrowed or modifi ed to<br />
formulate a site or bas<strong>in</strong> specifi c conceptual model as <strong>the</strong><br />
basis for restoration. Yet, Stewardson and Ru<strong>the</strong>rfurd (see<br />
Chapter 5) describe three levels <strong>in</strong> restoration from which<br />
5 In pr<strong>in</strong>ciple, <strong>the</strong> process of ‘rigour’ <strong>in</strong> restoration plann<strong>in</strong>g still<br />
follows <strong>the</strong> generic environmental management decision process<br />
of Table 3.2. In essence what Phillip Williams, a seasoned practitioner,<br />
describes is an <strong>in</strong>formal Decision Support System<br />
(DSS).
30 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Table 3.3 Common motives that guide notions and drive river restoration efforts<br />
Notion Example(s)<br />
What is Natural?<br />
Nature is <strong>in</strong> Equilibrium ‘<strong>the</strong> equilibrium between sediment supply and available transport capacity.’ (Soar &<br />
Thorne 2001); ‘landforms can be considered as ei<strong>the</strong>r a stage <strong>in</strong> a cycle of erosion or as<br />
a system <strong>in</strong> dynamic equilibrium.’ (Schumm & Lichty 1965).<br />
Nature is <strong>in</strong> fl ux ‘Restored ecosystems are those <strong>in</strong> which <strong>the</strong> rates and types of disturbance do not exceed<br />
<strong>the</strong> capacity of <strong>the</strong> system to respond to <strong>the</strong>m.’ (Hruby 2003).<br />
Nature Constant ‘confi dence on global stability; <strong>the</strong>re are no limitations to development’ (Levy et al. 2000).<br />
Nature Balanced ‘<strong>the</strong> environment is forgiv<strong>in</strong>g of most shocks, but large perturbations can knock ecological<br />
variables <strong>in</strong>to new regions of <strong>the</strong> landscape.’ (Levy et al. 2000).<br />
Nature Ephemeral ‘<strong>the</strong> environment can not safely tolerate human modifi cations’ (Levy et al. 2000).<br />
Nature Resilient ‘ecosystems are adaptive, evolutionary, and self organis<strong>in</strong>g . . . ecological systems often<br />
thrive under conditions of high variability’ (Levy et al. 2000).<br />
Physical Integrity<br />
Physical Integrity ‘Physical Integrity for rivers refers to a set of active fl uvial processes and landforms<br />
where<strong>in</strong> channel, fl oodpla<strong>in</strong>s, sediments, and overall spatial confi guration ma<strong>in</strong>ta<strong>in</strong> a<br />
dynamic equilibrium, with adjustments not exceed<strong>in</strong>g limits of change defi ned by<br />
societal values. <strong>River</strong>s possess physical <strong>in</strong>tegrity when <strong>the</strong>ir processes and forms<br />
ma<strong>in</strong>ta<strong>in</strong> active connections with each o<strong>the</strong>r <strong>in</strong> <strong>the</strong> present hydrologic regime.’<br />
(Graf 2001).<br />
Alluvial <strong>River</strong> Attributes Several commonly known concepts that govern how alluvial channels work have been<br />
compiled <strong>in</strong>to a set of ‘attributes’ for alluvial river <strong>in</strong>tegrity (Trush et al. 2000).<br />
Ecological Integrity Ecological Integrity ‘ma<strong>in</strong>tenance of all <strong>in</strong>ternal and external processes and attributes<br />
<strong>in</strong>teract<strong>in</strong>g with <strong>the</strong> environment <strong>in</strong> such a way that <strong>the</strong> biotic community corresponds<br />
to <strong>the</strong> natural state of <strong>the</strong> type-specifi c aquatic habitat, accord<strong>in</strong>g to <strong>the</strong> pr<strong>in</strong>ciples of<br />
self-regulation, resilience and resistance.’ (Angermeier & Karr 1994).<br />
High Biodiversity = Ecological Natural systems foster biodiversity and artifi cial systems are homogenized and dom<strong>in</strong>ated<br />
Integrity<br />
by <strong>in</strong>vasive species (Ward et al. 2002, Lister 1998).<br />
Morphological Diversity =<br />
Biological Diversity<br />
epistemological uncerta<strong>in</strong>ties emerge: <strong>the</strong> validity of<br />
<strong>the</strong> conceptual model; whe<strong>the</strong>r <strong>the</strong> proposed <strong>in</strong>tervention<br />
results <strong>in</strong> <strong>the</strong> planned geomorphic change; and whe<strong>the</strong>r<br />
<strong>the</strong> change is susta<strong>in</strong>able. They <strong>the</strong>n caution that <strong>the</strong><br />
validity of <strong>the</strong> conceptual model is <strong>the</strong> source of <strong>the</strong> ‘most<br />
uncerta<strong>in</strong>ty.’ Return<strong>in</strong>g to Phillip William’s concept of<br />
rigour <strong>in</strong> plann<strong>in</strong>g, he argues restoration objectives should<br />
be based on an understand<strong>in</strong>g of how <strong>the</strong> conceptual<br />
model <strong>in</strong>teracts and responds to various societal motives<br />
Newson (2002) did not dispute <strong>the</strong> abundance of evidence support<strong>in</strong>g <strong>the</strong> l<strong>in</strong>kages between<br />
channel dynamics and biodiversity, but criticises <strong>the</strong> lack of direct collaboration<br />
between geomorphologists and ecologists to substantiate <strong>the</strong> l<strong>in</strong>ks <strong>in</strong> river management:<br />
‘<strong>the</strong> mantra “morphological diversity = biodiversity” currently rema<strong>in</strong>s an act of faith.’<br />
What is Susta<strong>in</strong>able?<br />
Susta<strong>in</strong>ability Accord<strong>in</strong>g to Cairns (2003), <strong>the</strong> notion of susta<strong>in</strong>ability is based on ‘<strong>the</strong> assumption that<br />
humank<strong>in</strong>d has <strong>the</strong> right to alter <strong>the</strong> planet so that human life can <strong>in</strong>habit Earth<br />
<strong>in</strong>defi nitely.’<br />
Geomorphic Susta<strong>in</strong>ability ‘susta<strong>in</strong>ability encompasses <strong>the</strong> notion of self-regulation of spontaneous functions (e.g.<br />
sediment deposition, colonisation and succession of vegetation) with m<strong>in</strong>imal<br />
<strong>in</strong>tervention and no adverse impact on <strong>the</strong> future aquatic environment whilst<br />
ma<strong>in</strong>ta<strong>in</strong><strong>in</strong>g <strong>the</strong> functions of <strong>the</strong> channel demanded by society (fl ood control,<br />
navigation etc.).’ (Sear 1996).<br />
(NRC, 1992). Based on specifi c objectives, a measurable<br />
set of <strong>in</strong>dicators and target levels can be selected (Doyle<br />
et al., 2000; Levy et al., 2000; Merkle and Kaupenjohann,<br />
2000; Smeets and Weter<strong>in</strong>gs, 1999). F<strong>in</strong>ally, a comparison<br />
of predicted <strong>in</strong>dicator responses to restoration <strong>in</strong>tervention<br />
versus <strong>in</strong>action should be used to decide whe<strong>the</strong>r restoration<br />
is appropriate. Although available science may be<br />
used to <strong>in</strong>form <strong>the</strong> steps lead<strong>in</strong>g up to this decision<br />
(Lemons and Victor, see Chapter 1), <strong>the</strong> decision whe<strong>the</strong>r
or not to proceed is ultimately a political one (Alario and<br />
Brun, 2001).<br />
3.3.3 Approaches to <strong>Restoration</strong><br />
Generally, river restoration projects consist of three<br />
components: plann<strong>in</strong>g, implementation and evaluation.<br />
The diversity of approaches available to implement <strong>the</strong>se<br />
components ra<strong>the</strong>r appropriately refl ects <strong>the</strong> varied types<br />
(motives) of restoration projects and physiographic sett<strong>in</strong>gs<br />
<strong>the</strong>y are applied <strong>in</strong>. Thus, historical and spatial cont<strong>in</strong>gencies<br />
are contribut<strong>in</strong>g to uncerta<strong>in</strong>ties due to natural<br />
variability (Phillips, 2001). Indeed, a plethora of restoration<br />
approaches and strategies has been formalised <strong>in</strong> both<br />
<strong>the</strong> peer-reviewed and grey literature (Wheaton et al.,<br />
2004a). Examples range from generalised approaches for<br />
stream restoration (e.g. FISRWG, 1998; Jungwirth et al.,<br />
2002; Koehn et al., 2001; NRC, 1992; RRC, 2002) to<br />
more specifi c strategies <strong>in</strong>corporat<strong>in</strong>g: fl uvial geomorphology<br />
(e.g. Brookes and Sear, 1996; Gilvear, 1999;<br />
Kondolf, 2000; Sear, 1994), ecosystem <strong>the</strong>ory (e.g.<br />
Richards et al., 2002; Stanford et al., 1996), hydraulic<br />
eng<strong>in</strong>eer<strong>in</strong>g (e.g. Shields, 1996) and detailed design procedures<br />
(Miller et al., 2001; Shields et al., 2003; Wheaton<br />
et al., 2004b). Most of <strong>the</strong> approaches have parallels <strong>in</strong><br />
structure and ideology (Wheaton et al., 2004a).<br />
Popular labels used to describe restoration approaches<br />
<strong>in</strong>clude holistic, science-based, <strong>in</strong>tegrated and multidiscipl<strong>in</strong>ary<br />
(Hildén, 2000; Jungwirth et al., 2002;<br />
Wissmar and Bisson, 2003a). S<strong>in</strong>ce most approaches<br />
purport or aim to be all of <strong>the</strong>se (Wheaton et al., 2004a),<br />
and <strong>the</strong> converse of each is perceived as negative, <strong>the</strong>re is<br />
little value <strong>in</strong> discrim<strong>in</strong>at<strong>in</strong>g approaches on <strong>the</strong>se grounds.<br />
However, <strong>the</strong>ir components (i.e. plann<strong>in</strong>g, implementation<br />
and monitor<strong>in</strong>g) can be differentiated us<strong>in</strong>g three descriptive<br />
metrics: <strong>the</strong> scale of restoration; form based versus<br />
process based; and active versus passive. These metrics<br />
can provide <strong>in</strong>sight <strong>in</strong>to <strong>the</strong> types of uncerta<strong>in</strong>ties encountered<br />
and expectations placed on restoration projects<br />
dur<strong>in</strong>g plann<strong>in</strong>g, implementation and monitor<strong>in</strong>g.<br />
S<strong>in</strong>ce <strong>the</strong> late 1990s, approaches almost unanimously<br />
call for catchment scale plann<strong>in</strong>g <strong>in</strong> restoration6 . However,<br />
confusion arises over whe<strong>the</strong>r this means: restore <strong>the</strong><br />
entire catchment; use watershed assessments to nest reach<br />
scale restoration <strong>in</strong> a catchment context (e.g. Bohn and<br />
Kershner, 2002; Brookes and Shields, 1996; Walker et al.,<br />
2002) or undertake a range of management and restoration<br />
activities across various spatial scales but nested with<strong>in</strong> a<br />
catchment context (e.g. Frissell et al., 1993; Roni et al.,<br />
2002). Ecosystem degradation has often taken place over<br />
6 See also Table 3.1.<br />
The Scope of Uncerta<strong>in</strong>ties <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> 31<br />
many decades or centuries and extends across landscape,<br />
catchment and regional scales (Palmer et al., 1997).<br />
However, restor<strong>in</strong>g an entire catchment is rarely viable<br />
(Brookes and Shields, 1996). Even those who call for<br />
ecological restoration of <strong>the</strong> entire catchment (e.g. Frissell<br />
et al., 1993) actually advocate achiev<strong>in</strong>g this through a<br />
range of targeted activities at various spatial and temporal<br />
scales.<br />
Most of <strong>the</strong> restoration literature also po<strong>in</strong>ts towards a<br />
consensus that a ‘process-based’ approach is superior to a<br />
‘form-based’ one (Wheaton et al., 2004a). Much of <strong>the</strong><br />
form versus process debate simplifi es down to <strong>the</strong> diffi -<br />
culty and/or appropriateness <strong>in</strong> select<strong>in</strong>g an analogue or<br />
reference condition. The frequently referenced ‘Lietbilds’<br />
or target visions (Kern, 1992) and <strong>the</strong> popular Rosgen<br />
approach to restoration (Malakoff, 2004; Rosgen, 1996)<br />
both rely heavily on analogues. Jungwirth et al. (2002)<br />
suggest that at least three methods for select<strong>in</strong>g analogue<br />
or reference conditions exist:<br />
Select an exist<strong>in</strong>g reference site with ‘desirable’ conditions<br />
(location substitution).<br />
Select a historical reference condition for <strong>the</strong> site of<br />
<strong>in</strong>terest on <strong>the</strong> basis of historical analysis (time for space<br />
substitution).<br />
Create a reference condition on <strong>the</strong> basis of <strong>the</strong>oretical<br />
models (ei<strong>the</strong>r conceptual or ma<strong>the</strong>matical).<br />
In referr<strong>in</strong>g to <strong>the</strong>se analogue conditions, is <strong>the</strong> desired<br />
form or <strong>the</strong> desired process <strong>the</strong>n mimicked? This seems<br />
to be <strong>the</strong> po<strong>in</strong>t of departure for op<strong>in</strong>ions with<strong>in</strong> <strong>the</strong> restoration<br />
literature. Some argue that any mimick<strong>in</strong>g of reference<br />
conditions is a form-based approach (McDonald et al.,<br />
2004). O<strong>the</strong>rs suggest that as long as ample consideration<br />
of susta<strong>in</strong><strong>in</strong>g processes and desired functions is made, <strong>the</strong><br />
use of analogue conditions can be process based (Palmer<br />
et al., 1997; Wheaton et al., 2004a). Although exact <strong>in</strong>terpretations<br />
are <strong>the</strong>mselves uncerta<strong>in</strong> and will cont<strong>in</strong>ue to<br />
spur debate over semantics (confl ict<strong>in</strong>g evidence uncerta<strong>in</strong>ties),<br />
most concur that consideration of susta<strong>in</strong><strong>in</strong>g<br />
processes is fundamental (Wheaton et al., 2004c).<br />
Fundamental methodological disagreements arise <strong>in</strong> <strong>the</strong><br />
restoration literature with respect to passive versus active<br />
approaches to river restoration (Edmonds et al., 2003;<br />
Wissmar and Beschta, 1998). Here, active approaches are<br />
referred to as those which <strong>in</strong>volve direct structural modifi<br />
cation to <strong>the</strong> river, its fl oodpla<strong>in</strong> or <strong>in</strong>frastructure <strong>the</strong>re<strong>in</strong><br />
(e.g. channel realignment, levee removal, <strong>in</strong>stream habitat<br />
structures). By contrast, passive approaches are those that<br />
‘rely on <strong>the</strong> river to do <strong>the</strong> work’ (e.g. fl ow augmentation,
32 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Figure 3.4 Five philosophical attitudes towards uncerta<strong>in</strong>ty (The Venn diagram is meant to illustrate <strong>the</strong> overlap between contemporary<br />
attitudes towards uncerta<strong>in</strong>ty. Note that ignor<strong>in</strong>g uncerta<strong>in</strong>ty shares no overlap with contemporary attitudes towards uncerta<strong>in</strong>ty.)<br />
change <strong>in</strong> landuse, manag<strong>in</strong>g nonpo<strong>in</strong>t sources of pollution,<br />
buffer strips) (Wissmar and Beschta, 1998). Us<strong>in</strong>g a<br />
‘process-based’ approach can make <strong>in</strong>tuitive sense for<br />
passive approaches to restoration. For example, provid<strong>in</strong>g<br />
fl ow releases from a reservoir to mimic a natural hydrograph<br />
and encourage mobilisation and reorganisation of<br />
sediments, may restore <strong>the</strong> processes that ‘allow <strong>the</strong> river<br />
to do <strong>the</strong> work’ (Stanford et al., 1996; Trush et al., 2000).<br />
However, active approaches are considered favourable<br />
when natural or passive recovery may take an unacceptably<br />
long time (Montgomery and Bolton, 2003). The<br />
choice of a passive versus active approach will depend<br />
very much on <strong>the</strong> specifi c social, political, economic<br />
and environmental cont<strong>in</strong>gencies of <strong>in</strong>dividual river bas<strong>in</strong>s<br />
(Wissmar et al., 2003), as well as <strong>the</strong> extent to which<br />
<strong>in</strong>itial conditions matter (Phillips, 2002). Wheaton et al.<br />
(2004b) suggested that <strong>in</strong> some spawn<strong>in</strong>g habitat rehabilitation<br />
contexts, it may be appropriate to employ passive<br />
approaches like gravel augmentation <strong>in</strong> conjunction with<br />
active approaches like spawn<strong>in</strong>g bed enhancement to kickstart<br />
recovery. Ultimately, all <strong>the</strong>se choices are fuelled by<br />
an uncerta<strong>in</strong> conceptual understand<strong>in</strong>g of <strong>the</strong> system and<br />
logical ideas about how best to proceed with restoration.<br />
Given <strong>the</strong>se <strong>in</strong>herent uncerta<strong>in</strong>ties, adaptive management<br />
is well suited to allow practitioners and decision makers<br />
to make a decision <strong>in</strong> <strong>the</strong> face of uncerta<strong>in</strong>ty, and to adjust<br />
that decision as time and new challenges unfold (Clark,<br />
2002; Lister, 1998).<br />
7<br />
See Section 3.3.1 for <strong>the</strong> relationship between expectation and<br />
uncerta<strong>in</strong>ty.<br />
3.4 PHILOSOPHIES OF UNCERTAINTY<br />
So, is all this uncerta<strong>in</strong>ty bad? By this po<strong>in</strong>t, it should be<br />
clear that uncerta<strong>in</strong>ty <strong>in</strong> river restoration is ubiquitous.<br />
However, different segments of society view uncerta<strong>in</strong>ty <strong>in</strong><br />
very different ways, depend<strong>in</strong>g on <strong>the</strong> context (Lemons<br />
and Victor, see Chapter 1). As already mentioned, ord<strong>in</strong>ary<br />
people are quite comfortable with <strong>the</strong> uncerta<strong>in</strong>ties of<br />
life <strong>in</strong> an <strong>in</strong>tuitive and nonexplicit sense (Anderson et al.,<br />
2003; Pollack, 2003). However, uncerta<strong>in</strong>ty <strong>in</strong> policy and<br />
science, especially as reported <strong>in</strong> <strong>the</strong> media (Riebeek,<br />
2002), are very different contexts. The choice of what to do<br />
about uncerta<strong>in</strong>ty is a philosophical question. Five potential<br />
philosophical treatments of uncerta<strong>in</strong>ty are proposed <strong>in</strong><br />
Figure 3.4. Each of <strong>the</strong>se philosophies is reviewed <strong>in</strong> <strong>the</strong><br />
rema<strong>in</strong><strong>in</strong>g sections and l<strong>in</strong>ked to current attitudes with<strong>in</strong><br />
different segments of <strong>the</strong> river restoration community.<br />
3.4.1 Ignore <strong>Uncerta<strong>in</strong>ty</strong><br />
It has already been argued here that <strong>the</strong> restoration community<br />
has tended to passively ignore uncerta<strong>in</strong>ty and<br />
possible explanations as to why this may be <strong>the</strong> case proposed.<br />
For example, managers, policy and decision makers<br />
are fearful of admitt<strong>in</strong>g uncerta<strong>in</strong>ties, as this may be seen<br />
as a sign of weakness (Clark, 2002; Levy et al., 2000).<br />
Now that public support exists for river restoration, so<br />
too does <strong>the</strong> expectation7 that <strong>the</strong> problems restoration<br />
addresses are well understood. Indeed, <strong>the</strong>se problems are<br />
reasonably well understood, but numerous uncerta<strong>in</strong>ties<br />
rema<strong>in</strong>. Aside from basic, and potentially reducible,
communication uncerta<strong>in</strong>ties <strong>the</strong> signifi cance of <strong>the</strong> vast<br />
majority of uncerta<strong>in</strong>ties associated with restoration are<br />
simply not known. Admittedly, specifi c examples of<br />
uncerta<strong>in</strong>ties <strong>in</strong> restoration may <strong>in</strong>deed be <strong>in</strong>signifi cant.<br />
However, to assume <strong>in</strong>signifi cance on both ethical and<br />
technical grounds without fi rst establish<strong>in</strong>g it might ultimately<br />
backfi re on <strong>the</strong> restoration community.<br />
3.4.2 Elim<strong>in</strong>ate <strong>Uncerta<strong>in</strong>ty</strong><br />
The positivist view of <strong>the</strong> world has fuelled much scientifi<br />
c progress on <strong>the</strong> notion that uncerta<strong>in</strong>ty is bad, absolute<br />
knowledge is good, and it is necessary to strive to<br />
elim<strong>in</strong>ate uncerta<strong>in</strong>ty (Klir and Yuan, 1995; Priddy, 1999;<br />
Van Asselt and Rotmans, 2002). This fosters an unnecessarily<br />
narrow view of uncerta<strong>in</strong>ty as subsumed entirely<br />
with<strong>in</strong> <strong>the</strong> realm of science. van Asselt and Rotmans<br />
(2002) argued this view grew out of <strong>the</strong> ‘Enlightenment<br />
Period’ or ‘Age of Reason’ of <strong>the</strong> 17th and 18th centuries<br />
where science was to be ‘<strong>the</strong> provider of certa<strong>in</strong>ty.’ Fur<strong>the</strong>r<br />
to this endeavour, many scientists assumed that unique<br />
causal laws exist for all natural phenomena and ignored<br />
<strong>the</strong> possibilities of <strong>in</strong>determ<strong>in</strong>acy and equifi nality (Wilson,<br />
2001). Many physical scientists still subscribe to a ‘positivist’<br />
view (Harman, 1998), implicitly associat<strong>in</strong>g uncerta<strong>in</strong>ty<br />
with an <strong>in</strong>ability to quantify <strong>the</strong> environment, ra<strong>the</strong>r<br />
than acknowledg<strong>in</strong>g a limited understand<strong>in</strong>g about <strong>the</strong><br />
environment itself (Klir and Yuan, 1995).<br />
Whe<strong>the</strong>r specifi c types of uncerta<strong>in</strong>ty can be elim<strong>in</strong>ated<br />
depends on an <strong>in</strong>dividual’s <strong>in</strong>terpretation of semantics.<br />
Under <strong>the</strong> holistic view of uncerta<strong>in</strong>ty advocated <strong>in</strong> this<br />
chapter uncerta<strong>in</strong>ty cannot be completely elim<strong>in</strong>ated.<br />
Pollack (2003) suggests that ‘uncerta<strong>in</strong>ty is always with<br />
us and can never be fully elim<strong>in</strong>ated’. O<strong>the</strong>r authors (e.g.<br />
Knight, 1921) suggest that some types of uncerta<strong>in</strong>ty can<br />
be transformed <strong>in</strong>to related concepts (e.g. error, expectation,<br />
reliability, risk) with <strong>the</strong> help of ma<strong>the</strong>matical constructs<br />
and knowledge ga<strong>in</strong>ed from historical <strong>in</strong>ference.<br />
Through this transformation, uncerta<strong>in</strong>ty of a specifi c type<br />
(i.e. uncerta<strong>in</strong>ty for which a valid basis for classifi cation<br />
exists) <strong>in</strong> a sense might be ‘elim<strong>in</strong>ated.’ Such a transformation<br />
represents an improved understand<strong>in</strong>g of uncerta<strong>in</strong>ty<br />
but does not truly ‘elim<strong>in</strong>ate’ it.<br />
With technological progress has come <strong>the</strong> expectation<br />
of greater predictive power. Priddy (1999) suggested, ‘<strong>the</strong><br />
strictest standard of truth <strong>in</strong> science is that of predictability.’<br />
Although <strong>in</strong>tuitively no one expects prediction to be<br />
completely free of uncerta<strong>in</strong>ty, <strong>the</strong> notion that uncerta<strong>in</strong>ty<br />
can be elim<strong>in</strong>ated is latent <strong>in</strong> <strong>the</strong> ma<strong>in</strong>stream media<br />
(Riebeek, 2002). Pollack (2003) argues that scientists are<br />
accustomed to deal<strong>in</strong>g with uncerta<strong>in</strong>ty explicitly, but <strong>the</strong><br />
general public’s familiarity with uncerta<strong>in</strong>ty is implicit<br />
The Scope of Uncerta<strong>in</strong>ties <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> 33<br />
and often confused. Jamieson (1996) suggests that, particularly<br />
with respect to decisions about <strong>in</strong>creased environmental<br />
protections, <strong>the</strong> ‘rhetorical role of uncerta<strong>in</strong>ty<br />
claims’ are used to suggest no action should be taken until<br />
uncerta<strong>in</strong>ty is elim<strong>in</strong>ated. Hence, it is concluded that<br />
attempts to elim<strong>in</strong>ate uncerta<strong>in</strong>ty are mislead<strong>in</strong>g and<br />
founded on ignorance of <strong>the</strong> pr<strong>in</strong>ciples of uncerta<strong>in</strong>ty.<br />
3.4.3 Reduce <strong>Uncerta<strong>in</strong>ty</strong><br />
A more pragmatic view of uncerta<strong>in</strong>ty seeks to reduce,<br />
ra<strong>the</strong>r than elim<strong>in</strong>ate, those specifi c elements that are perceived<br />
as problematic (Klir and Yuan, 1995). This approach<br />
to uncerta<strong>in</strong>ty is represented diagrammatically <strong>in</strong> Figure<br />
3.4. Notice that with regards to reduc<strong>in</strong>g uncerta<strong>in</strong>ty, <strong>the</strong><br />
key questions are, <strong>in</strong> order: can it be quantifi ed, is it signifi<br />
cant and can it be constra<strong>in</strong>ed? So long as <strong>the</strong> answer<br />
is ‘yes’ to all <strong>the</strong>se questions, uncerta<strong>in</strong>ty might be reduced.<br />
However, if <strong>the</strong> opposite is true, uncerta<strong>in</strong>ty is simply<br />
ignored. To move beyond uncerta<strong>in</strong>ty as an ambiguous<br />
buzzword that will forever plague scientists and decision<br />
makers, a broader view of uncerta<strong>in</strong>ty as <strong>in</strong>formation is<br />
appropriate (Newson and Clark, see Chapter 14).<br />
3.4.4 Cope with <strong>Uncerta<strong>in</strong>ty</strong><br />
Cop<strong>in</strong>g or liv<strong>in</strong>g with uncerta<strong>in</strong>ty represents a more proactive<br />
view of deal<strong>in</strong>g with uncerta<strong>in</strong>ty than elim<strong>in</strong>ation or<br />
reduction. This approach recognises that, regardless of<br />
<strong>the</strong> signifi cance of uncerta<strong>in</strong>ty and our ability/<strong>in</strong>ability to<br />
quantify or constra<strong>in</strong> it, we are always forced to cope with<br />
it. Especially with<strong>in</strong> <strong>the</strong> hydrologic and atmospheric<br />
modell<strong>in</strong>g literature, uncerta<strong>in</strong>ty is actively recognised<br />
and specifi c methods to cope with it are cont<strong>in</strong>ually be<strong>in</strong>g<br />
proposed (e.g. Beven, 1996a; Beven, 1996b; Osidele et al.,<br />
2003; Werritty, 2002).<br />
3.4.5 Embrace <strong>Uncerta<strong>in</strong>ty</strong><br />
Despite <strong>the</strong> advantages of efforts to cope with or reduce<br />
uncerta<strong>in</strong>ty over elim<strong>in</strong>at<strong>in</strong>g it, all <strong>the</strong> preced<strong>in</strong>g still fundamentally<br />
view uncerta<strong>in</strong>ty as negative. Several authors<br />
have departed from this view towards a more progressive<br />
view of embrac<strong>in</strong>g uncerta<strong>in</strong>ty (Johnson and Brown, 2001;<br />
Newson and Clark, see Chapter 14). One of <strong>the</strong> earlier<br />
proponents of this view appears to be Holl<strong>in</strong>g (1978), who<br />
argued:<br />
‘while efforts to reduce uncerta<strong>in</strong>ty are admirable . . . if<br />
not accompanied by an equal effort to design for<br />
uncerta<strong>in</strong>ty and obta<strong>in</strong> benefi ts from <strong>the</strong> unexpected,<br />
<strong>the</strong> best of predictive models will only lead to larger
34 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
problems aris<strong>in</strong>g more quickly and more often’ (<strong>in</strong>:<br />
Levy et al., 2000).<br />
Klir and Yuan (1995) considered uncerta<strong>in</strong>ty <strong>in</strong> modell<strong>in</strong>g<br />
as ‘an important commodity . . . , which can be traded<br />
for ga<strong>in</strong>s <strong>in</strong> <strong>the</strong> o<strong>the</strong>r essential characteristics of models.’<br />
O<strong>the</strong>rs have suggested that recognis<strong>in</strong>g that not all uncerta<strong>in</strong>ty<br />
is bad will be <strong>in</strong>creas<strong>in</strong>gly important to decision<br />
makers who are forced to make decisions <strong>in</strong> <strong>the</strong> face of<br />
uncerta<strong>in</strong>ty (Clark and Richards, 2002; Pollack, 2003).<br />
Especially <strong>in</strong> long term policy analysis (<strong>the</strong> next 20–100<br />
years) decision makers are faced with what Lempert et al.<br />
(2003) referred to as ‘deep uncerta<strong>in</strong>ty’. Johnson and<br />
Brown (2001) argued that <strong>in</strong>corporat<strong>in</strong>g uncerta<strong>in</strong>ty <strong>in</strong>to<br />
restoration design allows practitioners to consider multiple<br />
causes and hypo<strong>the</strong>sised fi xes; <strong>the</strong>reby reduc<strong>in</strong>g <strong>the</strong> potential<br />
for project failure. It has been argued here that uncerta<strong>in</strong>ty<br />
is not necessarily bad, but ignorance of it can foster<br />
unrealistic expectations. Chapman and Ward (2002) argued<br />
that uncerta<strong>in</strong>ty is not just as a risk, but also an opportunity.<br />
<strong>Uncerta<strong>in</strong>ty</strong> due to natural variability, <strong>in</strong> say fl ow regime,<br />
can be a particularly good th<strong>in</strong>g, for example by promot<strong>in</strong>g<br />
habitat heterogeneity and biodiversity (Clifford et al., see<br />
Chapter 7; Montgomery and Bolton, 2003).<br />
In Figure 3.5, <strong>the</strong> notions of embrac<strong>in</strong>g uncerta<strong>in</strong>ty are<br />
syn<strong>the</strong>sised <strong>in</strong> <strong>the</strong> context of <strong>the</strong> van Asselt (2000) typology.<br />
This approach embraces uncerta<strong>in</strong>ty as <strong>in</strong>formation<br />
and its potential for help<strong>in</strong>g avoid risks, or embrac<strong>in</strong>g<br />
unforeseen opportunities. Notice that <strong>the</strong> uncerta<strong>in</strong>ties are<br />
not treated uniformly, but <strong>in</strong>stead are segregated by source<br />
(i.e. due to limited knowledge or due to variability) and<br />
type. Anderson et al. (2003) note that environmental management<br />
problems are so diverse that a s<strong>in</strong>gle approach is<br />
unlikely to be appropriate for all. Thus, Chamberl<strong>in</strong>’s<br />
(1890) idea of multiple work<strong>in</strong>g hypo<strong>the</strong>ses is emerg<strong>in</strong>g<br />
<strong>in</strong> environmental management through advocat<strong>in</strong>g pluralistic<br />
approaches (e.g. Lempert et al., 2003; Van Asselt and<br />
Rotmans, 2002).<br />
The embrac<strong>in</strong>g uncerta<strong>in</strong>ty framework proposed here<br />
emphasises this po<strong>in</strong>t by structur<strong>in</strong>g a range of questions<br />
and possible management decisions based on <strong>the</strong> specifi c<br />
uncerta<strong>in</strong>ties at hand. In <strong>the</strong> spirit of ‘susta<strong>in</strong>able uncerta<strong>in</strong>ty’<br />
as proposed by Newson and Clark (see Chapter 14),<br />
this is not at all a rigid framework but <strong>in</strong>stead a loose and<br />
adaptive guide built around an uncerta<strong>in</strong>ty typology.<br />
Unlike <strong>the</strong> four o<strong>the</strong>r philosophical treatments of uncerta<strong>in</strong>ty,<br />
this allows <strong>the</strong> restoration scientist, practitioner or<br />
decision maker to:<br />
explore <strong>the</strong> potential signifi cance (both <strong>in</strong> terms of<br />
unforeseen consequences and welcome surprises) or<br />
<strong>in</strong>signifi cance of uncerta<strong>in</strong>ties;<br />
effectively communicate uncerta<strong>in</strong>ties;<br />
eventually make adaptive, but transparent, decisions <strong>in</strong><br />
<strong>the</strong> face of uncerta<strong>in</strong>ty.<br />
3.5 CONCLUSION<br />
In this chapter a very broad picture of uncerta<strong>in</strong>ty <strong>in</strong> river<br />
restoration and environmental management has been<br />
pa<strong>in</strong>ted. This was done to unravel <strong>the</strong> ambiguities around<br />
<strong>the</strong> notion of ‘certa<strong>in</strong>ty’ <strong>in</strong> restoration and recast uncerta<strong>in</strong>ty<br />
as useful <strong>in</strong>formation. In fact, <strong>the</strong> arguments and<br />
evidence presented challenge <strong>the</strong> view of scientifi c determ<strong>in</strong>istic<br />
‘certa<strong>in</strong>ty’ and societal beliefs that certa<strong>in</strong>ty is<br />
necessary <strong>in</strong> restoration. A typology for discrim<strong>in</strong>at<strong>in</strong>g<br />
uncerta<strong>in</strong>ty was reviewed that can be used to separate<br />
uncerta<strong>in</strong>ties that can lead to unforeseen and undesirable<br />
consequences from uncerta<strong>in</strong>ties that lead to potentially<br />
welcome surprises. Many of <strong>the</strong> uncerta<strong>in</strong>ties surround<strong>in</strong>g<br />
restoration motives, notions and approaches are most seriously<br />
manifested as communication uncerta<strong>in</strong>ties. That is,<br />
<strong>in</strong>stead of be<strong>in</strong>g expressed simply as uncerta<strong>in</strong>ties due<br />
to limited knowledge, <strong>the</strong>y are ignored and miscommunicated<br />
through <strong>the</strong> restoration process <strong>in</strong> a manner that<br />
prevents transparent decision mak<strong>in</strong>g. The signifi cance of<br />
<strong>the</strong> plethora of o<strong>the</strong>r uncerta<strong>in</strong>ties alluded to is largely<br />
situation-specifi c and, to date, unexplored.<br />
Five philosophical strategies for deal<strong>in</strong>g with uncerta<strong>in</strong>ty<br />
rang<strong>in</strong>g from <strong>the</strong> status quo of ignor<strong>in</strong>g uncerta<strong>in</strong>ty<br />
to <strong>the</strong> advocated embrac<strong>in</strong>g uncerta<strong>in</strong>ty were reviewed.<br />
Traditional scientifi c research has focused on a narrow<br />
class of uncerta<strong>in</strong>ties and adopted ‘elim<strong>in</strong>ate’ and ‘reduce’<br />
uncerta<strong>in</strong>ty philosophies. It is argued that it is unethical to<br />
assume that uncerta<strong>in</strong>ty is <strong>in</strong>signifi cant. There is an<br />
<strong>in</strong>creas<strong>in</strong>g recognition <strong>in</strong> environmental management that<br />
ethical and social dimensions are <strong>the</strong> primary drivers, with<br />
scientifi c and technical dimensions play<strong>in</strong>g a secondary<br />
role (Falkenmark and Folke, 2002; Lister, 1998) 8 . Thus, an<br />
emerg<strong>in</strong>g challenge which <strong>the</strong> restoration community is<br />
faced with is comb<strong>in</strong><strong>in</strong>g <strong>the</strong>se dimensions to ‘do <strong>the</strong> right<br />
th<strong>in</strong>g right.’ Out of <strong>the</strong> decision mak<strong>in</strong>g arena has emerged<br />
<strong>the</strong> pragmatic view of cop<strong>in</strong>g with uncerta<strong>in</strong>ty. However,<br />
from <strong>the</strong> suggestions and examples <strong>in</strong> <strong>the</strong> more general<br />
environmental management literature, it is concluded that<br />
embrac<strong>in</strong>g uncerta<strong>in</strong>ty could also help transcend <strong>the</strong> scientifi<br />
c research and decision mak<strong>in</strong>g boundaries <strong>in</strong> river<br />
restoration.<br />
8 Recall <strong>the</strong> development of notions <strong>in</strong> Section 3.3.2, and <strong>the</strong> dist<strong>in</strong>ction<br />
of Falkenmark and Folke (2002) between technical concerns<br />
(e.g. ‘do<strong>in</strong>g <strong>the</strong> th<strong>in</strong>g right’) and ethical concerns (e.g.<br />
‘do<strong>in</strong>g <strong>the</strong> right th<strong>in</strong>g’).
The Scope of Uncerta<strong>in</strong>ties <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> 35<br />
Figure 3.5 Framework for embrac<strong>in</strong>g uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong> decision mak<strong>in</strong>g process (This framework relies on <strong>the</strong> Van Asselt (2000)<br />
typology of uncerta<strong>in</strong>ty.)
36 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
REFERENCES<br />
Alario M, Brun M. 2001. <strong>Uncerta<strong>in</strong>ty</strong> and Controversy <strong>in</strong><br />
<strong>the</strong> Science and Ethics of Environmental Policy Mak<strong>in</strong>g.<br />
Theory and Science. 2 (1): Available on <strong>the</strong> web: \url{http://<br />
<strong>the</strong>oryandscience.icaap.org/content/vol002.001/02alariobrun.<br />
html}.<br />
Anderson JL, Hilborn RW, Lackey RT, Ludwig D. 2003. Watershed<br />
restoration: Adaptive decision mak<strong>in</strong>g <strong>in</strong> <strong>the</strong> face of<br />
uncerta<strong>in</strong>ty. In: Wissmar RC, Bisson PA, Duke M (Eds), Strategies<br />
for Restor<strong>in</strong>g <strong>River</strong> Ecosystems: Sources of Variability and<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Natural and Managed Systems. American<br />
Fisheries Society: Be<strong>the</strong>sda, Maryland; 185–201.<br />
Angermeier PL, Karr JR. 1994. Biological <strong>in</strong>tegrity versus biological<br />
diversity as policy directives: protect<strong>in</strong>g biotic resources.<br />
Bioscience 44: 690–697.<br />
Beven K. 1996a. Equifi nality and uncerta<strong>in</strong>ty <strong>in</strong> geomorphological<br />
modell<strong>in</strong>g. In: Rhoads BL, Thorn CE (Eds), The Scientifi c<br />
Nature of Geomorphology. John Wiley & Sons Ltd: Chichester;<br />
289–314.<br />
Beven K. 1996b. The limits of splitt<strong>in</strong>g: Hydrology. The Science<br />
of <strong>the</strong> Total Environment. 183: 89–97.<br />
Bohn BA, Kershner JL. 2002. Establish<strong>in</strong>g aquatic restoration<br />
priorities us<strong>in</strong>g a watershed approach. Journal of Environmental<br />
Management 64 (4): 355–363.<br />
Boon PJ. 1998. <strong>River</strong> restoration <strong>in</strong> fi ve dimensions. Aquatic Conservation:<br />
Mar<strong>in</strong>e and Freshwater Ecosystems 8 (1): 257–264.<br />
Brookes A, Downs P, Sk<strong>in</strong>ner K. 1998. <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>the</strong> eng<strong>in</strong>eer<strong>in</strong>g<br />
of wildlife habitats. Journal of <strong>the</strong> Chartered Institution<br />
of Water and Environmental Management 12 (1): 25–29.<br />
Brookes A, Sear DA. 1996. Geomorphological pr<strong>in</strong>ciples for<br />
restor<strong>in</strong>g channels. In: Brookes A, Shields FD (Eds), <strong>River</strong><br />
Channel <strong>Restoration</strong>: Guid<strong>in</strong>g Pr<strong>in</strong>ciples for Susta<strong>in</strong>able<br />
Projects. John Wiley & Sons Ltd: Chichester; 75–101.<br />
Brookes A, Shields FD. 1996. <strong>River</strong> Channel <strong>Restoration</strong>:<br />
Guid<strong>in</strong>g Pr<strong>in</strong>ciples for Susta<strong>in</strong>able Projects. John Wiley &<br />
Sons Ltd: Chichester.<br />
Brown TL, LeMay HE, Bursten BE. 1994. Chemistry: The Central<br />
Science (6th Edition). Prentice Hall: Englewood Cliffs, New<br />
Jersey.<br />
Cairns J. 2003. Numeracy and susta<strong>in</strong>ability. Ethics <strong>in</strong> Science<br />
and Environmental Politics 2003: 83–91.<br />
Chamberl<strong>in</strong> TC. 1890. The method of multiple work<strong>in</strong>g hypo<strong>the</strong>ses.<br />
Science. Repr<strong>in</strong>ted 1965: 148: 754–759.<br />
Chapman C, Ward S. 2002. <strong>Manag<strong>in</strong>g</strong> Project Risk and <strong>Uncerta<strong>in</strong>ty</strong>:<br />
A Constructively Simple Approach to Decision Mak<strong>in</strong>g.<br />
John Wiley & Sons Ltd: Chichester.<br />
Clark MJ. 2002. Deal<strong>in</strong>g with uncerta<strong>in</strong>ty: adaptive approaches<br />
to susta<strong>in</strong>able river management. Aquatic Conservation:<br />
Mar<strong>in</strong>e and Freshwater Ecosystems 12: 347–363.<br />
Clark MJ, Richards KJ. 2002. Support<strong>in</strong>g complex decisions for<br />
susta<strong>in</strong>able river management <strong>in</strong> England and Wales. Aquatic<br />
Conservation: Mar<strong>in</strong>e and Freshwater Ecosystems 12: 471–483.<br />
Coll<strong>in</strong>s JP, K<strong>in</strong>zig A, Grimm NB et al. 2000. A New Urban<br />
Ecology: Model<strong>in</strong>g human communities as <strong>in</strong>tegral parts of<br />
ecosystems poses special problems for <strong>the</strong> development and<br />
test<strong>in</strong>g of ecological <strong>the</strong>ory. American Scientist 88: 416–425.<br />
Davis MA, Slobodk<strong>in</strong> LB. 2004. The science and values of restoration<br />
ecology. <strong>Restoration</strong> Ecology 12 (1): 1–3.<br />
Doyle MW, Harbor JM, Rich CF, Spacie A. 2000. Exam<strong>in</strong><strong>in</strong>g <strong>the</strong><br />
effects of urbanization on streams us<strong>in</strong>g <strong>in</strong>dicators of geomorphic<br />
stability. Physical Geography 21 (2): 155–181.<br />
Edmonds RL, Francis RC, Mantua NJ, Petersen DL. 2003.<br />
Sources of climate variability <strong>in</strong> river ecosystems. In: Wissmar<br />
RC, Bisson PA, Duke M (Eds), Strategies for Restor<strong>in</strong>g <strong>River</strong><br />
Ecosystems: Sources of Variability and <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Natural<br />
and Managed Systems. American Fisheries Society: Be<strong>the</strong>sda,<br />
Maryland; 11–37.<br />
Ellison SLR, Rossle<strong>in</strong> M, Williams A (Eds). 2000. Quantify<strong>in</strong>g<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Analytical Measurement, Guide number 4.<br />
Eurachem/CITAC: Tedd<strong>in</strong>gton, UK.<br />
Everest FH, Sedell JR. 1984. Evaluat<strong>in</strong>g effectiveness of stream<br />
enhancement projects. In: Hassler TJ (Ed), Pacifi c Northwest<br />
Stream Habitat Management Workshop. American Fisheries<br />
Society: Humboldt State Universitiy, California; 156–164.<br />
Falkenmark M, Folke C. 2002. The ethics of socioecohydrological<br />
catchment management. Hydrology and Earth<br />
System Sciences 6 (1): 1–9.<br />
FISRWG. 1998. Stream Corridor <strong>Restoration</strong>: Pr<strong>in</strong>ciples, Processes,<br />
and Practices. GPO Item No. 0120-A: Federal Interagency<br />
Stream <strong>Restoration</strong> Work<strong>in</strong>g Group, Wash<strong>in</strong>gton, DC.<br />
Frissell CA, Liss WJ, Bayles D. 1993. An <strong>in</strong>tegrated, biophysical<br />
strategy for ecological restoration of large watersheds. In:<br />
Spangenborg NE, Potts DE (Eds), Chang<strong>in</strong>g Roles <strong>in</strong> Water<br />
Resources Management and Policy. American Water Resources<br />
Association: Be<strong>the</strong>sda, Maryland; 449–456.<br />
Gilvear DJ. 1999. Fluvial geomorphology and river eng<strong>in</strong>eer<strong>in</strong>g:<br />
future roles utiliz<strong>in</strong>g a fl uvial hydrosystems framework.<br />
Geomorphology 31: 229–245.<br />
Graf WL. 1996. Geomorphology and policy for restoration of<br />
impounded American rivers: What is ‘Natural?’ In: Rhoads BL,<br />
Thorn CE (Eds), The Scientifi c Nature of Geomorphology.<br />
John Wiley & Sons Ltd: Chichester; 221–254.<br />
Graf WL. 2001. Damage control: dams and <strong>the</strong> physical <strong>in</strong>tegrity<br />
of America’s rivers. Annals of <strong>the</strong> Association of American<br />
Geographers 91: 1–27.<br />
Hansen HO, Iversen TM. 1998. The European Centre for <strong>River</strong><br />
<strong>Restoration</strong> (ECRR). In: Hansen HO, Madsen BL (Eds), <strong>River</strong><br />
<strong>Restoration</strong> ’96 ECRR: Denmark; 73–79.<br />
Harman W. 1998. Global M<strong>in</strong>d Change: The Promise of <strong>the</strong> 21st<br />
Century (2nd Ed). Berrett–Koehler Publishers, Inc.: San<br />
Francisco, California.<br />
Hildén M. 2000. The role of <strong>in</strong>tegrat<strong>in</strong>g concepts <strong>in</strong> watershed<br />
rehabilitation. Ecosystem Health 6 (1): 39–50.<br />
Holl<strong>in</strong>g CS. 1978. Adaptive Environmental Assessment and Management.<br />
John Wiley & Sons Ltd: Chichester.<br />
Hruby T. 2003. Where is <strong>the</strong> ecology <strong>in</strong> wetland restoration?<br />
Society of Ecological <strong>Restoration</strong> Northwest and Society<br />
Wetland Scientists Jo<strong>in</strong>t Regional Conference. SERNW: Portland,<br />
Oregon.<br />
Hughes RM. 1995. Defi n<strong>in</strong>g acceptable biological status by compar<strong>in</strong>g<br />
with reference conditions. In: Davis WS, Simon RF<br />
(Eds), Biological Assessment and Decision Mak<strong>in</strong>g. Lewis<br />
Publishers: Boca Raton, Florida; 31–47.
Jamieson D. 1996. Scientifi c uncerta<strong>in</strong>ty: how do we know when<br />
to communicate research fi nd<strong>in</strong>gs to <strong>the</strong> public? The Science<br />
of <strong>the</strong> Total Environment 184: 103–107.<br />
Johnson PA, Brown ER. 2001. Incorporat<strong>in</strong>g uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong><br />
design of stream channel modifi cations. Journal of <strong>the</strong> American<br />
Water Resources Association 37 (5): 1225–1236.<br />
Johnson PA, Hey RD, Brown ER, Rosgen DL. 2002. Stream restoration<br />
<strong>in</strong> <strong>the</strong> vic<strong>in</strong>ity of bridges. Journal of <strong>the</strong> American<br />
Water Resources Association 38 (1): 55–67.<br />
Johnson PA, R<strong>in</strong>aldi M. 1997. <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>the</strong> design of stream<br />
channel restorations. In: Jr. FMH and Alsaffar A (Eds), Water<br />
for a Chang<strong>in</strong>g Global Community: The 27th Congress of <strong>the</strong><br />
International Association for Hydraulic Research. American<br />
Society of Civil Eng<strong>in</strong>eers Press: San Francisco, California;<br />
451–457.<br />
Johnson PA, R<strong>in</strong>aldi M. 1998. <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> stream channel<br />
restoration. In: Ayyub BM (Ed), <strong>Uncerta<strong>in</strong>ty</strong> Model<strong>in</strong>g and<br />
Analysis <strong>in</strong> Civil Eng<strong>in</strong>eer<strong>in</strong>g. CRC Press: London; 425–<br />
437.<br />
Jungwirth M, Muhar S, Schmutz S. 2002. Re-establish<strong>in</strong>g and<br />
assess<strong>in</strong>g ecological <strong>in</strong>tegrity <strong>in</strong> river<strong>in</strong>e landscapes. Freshwater<br />
Biology 47 (4): 867–887.<br />
Kavvas ML. 1999. On <strong>the</strong> coarse-gra<strong>in</strong><strong>in</strong>g of hydrologic processes<br />
with <strong>in</strong>creas<strong>in</strong>g scales. Journal of Hydrology 217 (3–4):<br />
191–202.<br />
Kern K. 1992. Rehabilitation of streams <strong>in</strong> South-West Germany.<br />
In: Boon PJ, Calow P, Petts G (Eds), <strong>River</strong> Conservation and<br />
Management. John Wiley & Sons Ltd: Chichester; 321–336.<br />
Klir GJ, Yuan B. 1995. Fuzzy Sets and Fuzzy Logic: Theory and<br />
Applications. Prentice Hall: Upper Saddle <strong>River</strong>, New Jersey.<br />
Knight F. 1921. Risk, <strong>Uncerta<strong>in</strong>ty</strong> and Profi t. Houghton Miffl <strong>in</strong>:<br />
New York.<br />
Koehn JD, Brierley GJ, Cant BL, Lucas AM. 2001. <strong>River</strong> <strong>Restoration</strong><br />
Framework, The National <strong>River</strong>s Consortium: Land and<br />
Water Australia, Canberra, ACT.<br />
Kondolf GM. 1995a. Five elements for effective evaluation of<br />
stream restoration. <strong>Restoration</strong> Ecology 3 (2): 133–136.<br />
Kondolf GM. 1995b. Geomorphological stream channel classifi -<br />
cation <strong>in</strong> aquatic habitat restoration – Uses and limitations.<br />
Aquatic Conservation: Mar<strong>in</strong>e and Freshwater Ecosystems<br />
5 (2): 127–141.<br />
Kondolf GM. 1996. A cross section of stream channel restoration.<br />
Journal of Soil and Water Conservation 51 (2): 119–125.<br />
Kondolf GM. 2000. Some suggested guidel<strong>in</strong>es for geomorphic<br />
aspects of anadromous salmonid habitat restoration proposals.<br />
<strong>Restoration</strong> Ecology 8 (1): 48–56.<br />
Lempert RJ, Popper SW, Bankes SC. 2003. Shap<strong>in</strong>g <strong>the</strong> Next<br />
One Hundred Years: New Methods for Quantitative, Long term<br />
Policy Analysis. The Rand Pardee Center: Santa Monica,<br />
California.<br />
Levy JK, Hipel KW, Kilgour DM. 2000. Us<strong>in</strong>g environmental<br />
<strong>in</strong>dicators to quantify <strong>the</strong> robustness of policy alternatives to<br />
uncerta<strong>in</strong>ty. Ecological Modell<strong>in</strong>g 130 (1–3): 79–86.<br />
Lew<strong>in</strong> J. 2001. Alluvial systematics. In: Maddy D, Mackl<strong>in</strong> MG,<br />
Woodard JC (Eds), <strong>River</strong> Bas<strong>in</strong> Sediment Systems: Archives of<br />
Environmental Change. A.A. Balkema Publishers: Steenwijik,<br />
The Ne<strong>the</strong>rlands; 19–41.<br />
The Scope of Uncerta<strong>in</strong>ties <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> 37<br />
Lister N-ME. 1998. A systems approach to biodiversity conservation<br />
plann<strong>in</strong>g. Environmental Monitor<strong>in</strong>g and Assessment 49:<br />
123–155.<br />
Malakoff D. 2004. The river Doctor: Profi le Dave Rosgen.<br />
Science. 305 (13 August): 937–939.<br />
Marmulla G. (Ed). 2001. Dams, Fish and Fisheries. Opportunities,<br />
Challenges and Confl ict Resolution, Technical Paper no.<br />
419. FAO Fisheries Department: Rome, Italy.<br />
McDonald A, Lane SN, Chalk EA, Haycock NE. 2004. <strong>River</strong>s of<br />
dreams: on <strong>the</strong> gulf between <strong>the</strong>oretical and practical aspects<br />
of an upland river restoration. Transactions of <strong>the</strong> Institute of<br />
British Geographers 29 (3): 257–281.<br />
Merkle A, Kaupenjohann M. 2000. Derivation of ecosystemic<br />
effect <strong>in</strong>dicators – method. Ecological Modell<strong>in</strong>g 130 (1–3):<br />
39–46.<br />
Merriam–Webster. 1994. Merriam Webster’s Collegiate Dictionary.<br />
Merriam–Webster, Inc.: Spr<strong>in</strong>gfi eld, Massachusetts.<br />
Miller DE, Skidmore PB, White DJ. 2001. Channel Design, Inter-<br />
Fluve, Inc., Prepared for Wash<strong>in</strong>gton Department of Fish and<br />
Wildlife: Boseman, Montana.<br />
Montgomery DR, Bolton SM. 2003. Hydrogemorphic variability<br />
and river restoration. In: Wissmar RC, Bisson PA, Duke M<br />
(Eds), Strategies for Restor<strong>in</strong>g <strong>River</strong> Ecosystems: Sources of<br />
Variability and <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Natural and Managed Systems.<br />
American Fisheries Society: Be<strong>the</strong>sda, Maryland; 39–80.<br />
Moss T. 2004. The governance of land use <strong>in</strong> river bas<strong>in</strong>s: prospects<br />
for overcom<strong>in</strong>g problems of <strong>in</strong>stitutional <strong>in</strong>terplay with<br />
<strong>the</strong> EU Water Framework Directive. Land Use Policy 21:<br />
85–94.<br />
Mount JF. 1995. California <strong>River</strong>s and Streams: The Confl ict<br />
Between Fluvial Process and Land Use. University of California<br />
Press: Berkeley, California.<br />
NAP. 2002. Riparian Areas: Functions and Strategies for Management.<br />
National Academy Press: Wash<strong>in</strong>gton, DC.<br />
Newson MD. 2002. Geomorphological concepts and tools for<br />
susta<strong>in</strong>able river ecosystem management. Aquatic Conservation:<br />
Mar<strong>in</strong>e and Freshwater Ecosystems 12 (4): 365–379.<br />
Nijland HJ, Cals MJR (Eds). 2000. <strong>River</strong> <strong>Restoration</strong> <strong>in</strong><br />
Europe: Practical Applications. ECRR: Wagen<strong>in</strong>gen, The<br />
Ne<strong>the</strong>rlands.<br />
NRC. 1992. <strong>Restoration</strong> of Aquatic Ecosystems: Science, Technology,<br />
and Public Policy. National Academy Press: Wash<strong>in</strong>gton,<br />
DC.<br />
Osidele OO, Zeng W, Beck MB. 2003. Cop<strong>in</strong>g with uncerta<strong>in</strong>ty:<br />
A case study <strong>in</strong> sediment transport and nutrient load analysis.<br />
Journal of Water Resources Plann<strong>in</strong>g and Management-Asce<br />
129 (4): 345–355.<br />
Palmer MA, Ambrose RF, Poff NL. 1997. Ecological <strong>the</strong>ory and<br />
community restoration ecology. <strong>Restoration</strong> Ecology 5 (4):<br />
291–300.<br />
Phillips JD. 2001. Cont<strong>in</strong>gency and generalization <strong>in</strong> pedology,<br />
as exemplifi ed by texture-contrast soils. Geoderma 102 (3–4):<br />
347–370.<br />
Phillips JD. 2002. Global and local factors <strong>in</strong> earth surface<br />
systems. Ecological Modell<strong>in</strong>g 149 (3): 257–272.<br />
Pister EP. 2001. Wilderness fi sh stock<strong>in</strong>g: History and perspective.<br />
Ecosystems 4: 279–286.
38 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Pollack HN. 2003. Uncerta<strong>in</strong> Science . . . Uncerta<strong>in</strong> World. Cambridge<br />
University Press: Cambridge, UK.<br />
Power M. 1999. The Audit Society: Rituals of Verifi cation. Oxford<br />
University Press: Oxford, UK.<br />
Priddy R. 1999. Science Limited. Available on <strong>the</strong> web: http://<br />
home.no.net/rrpriddy/<strong>in</strong>dexlim.html: Oslo, Norway.<br />
Rhoads BL, Wilson D, Urban M, Herricks EE. 1999. Interaction<br />
between scientists and nonscientists <strong>in</strong> community-based<br />
watershed management: Emergence of <strong>the</strong> concept of<br />
stream naturalization. Environmental Management 24 (3):<br />
297–308.<br />
Richards K, Bras<strong>in</strong>gton J, Hughes F. 2002. Geomorphic dynamics<br />
of fl oodpla<strong>in</strong>s: ecological implications and a potential modell<strong>in</strong>g<br />
strategy. Freshwater Biology 47 (4): 559–579.<br />
Riebeek H. 2002. The perception of scientifi c uncerta<strong>in</strong>ty <strong>in</strong><br />
science news writ<strong>in</strong>g. URL: http://tc.eserver.org/13650.html<br />
Accessed on: 19 September, 2003.<br />
Rob<strong>in</strong>son CT, Tockner K, Ward JV. 2002. The fauna of dynamic<br />
river<strong>in</strong>e landscapes. Freshwater Biology 47 (4): 661–677.<br />
Roni P, Beechie TJ, Bilby RE et al. 2002. A review of stream<br />
restoration and a hierarchal strategy for prioritiz<strong>in</strong>g restoration<br />
<strong>in</strong> Pacifi c Northwest watersheds. North American Journal of<br />
Fisheries Management 22: 1–20.<br />
Rosgen D. 1996. Applied <strong>River</strong> Morphology. Wildland Hydrology:<br />
Pagosa Spr<strong>in</strong>gs, Colorado.<br />
Rotmans J, Van Asselt MBA. 2001. <strong>Uncerta<strong>in</strong>ty</strong> management<br />
<strong>in</strong> Integrated Assessment Modell<strong>in</strong>g: Towards a pluralistic<br />
approach. Environmental Monitor<strong>in</strong>g and Assessment 69 (2):<br />
101–130.<br />
RRC. 2002. Manual of <strong>River</strong> <strong>Restoration</strong> Techniques. The <strong>River</strong><br />
<strong>Restoration</strong> Centre, Silsoe, UK.<br />
Schumm SA. 1991. To Interpret <strong>the</strong> Earth: Ten Ways to be Wrong.<br />
Cambridge University Press: Cambridge, UK.<br />
Schumm SA, Lichty RW. 1965. Time, space and causality <strong>in</strong><br />
geomorphology. American Journal of Science 263 (February):<br />
110–119.<br />
Sear DA. 1994. <strong>River</strong> restoration and geomorphology. Aquatic<br />
Conservation: Mar<strong>in</strong>e and Freshwater Ecosystems 4 (2):<br />
169–177.<br />
Sear DA. 1996. The Sediment System and Channel Stability. In:<br />
Brookes A, Shields FD (Eds), <strong>River</strong> Channel <strong>Restoration</strong>:<br />
Guid<strong>in</strong>g Pr<strong>in</strong>ciples for Susta<strong>in</strong>able Projects. John Wiley &<br />
Sons Ltd: Chichester, 149–177.<br />
Shields FD. 1996. Hydraulic and hydrologic stability. In: Brookes<br />
A, Shields FD (Eds), <strong>River</strong> Channel <strong>Restoration</strong>: Guid<strong>in</strong>g<br />
Pr<strong>in</strong>ciples for Susta<strong>in</strong>able Projects. John Wiley & Sons Ltd:<br />
Chichester; 103–126.<br />
Shields FD, Copeland RR, Kl<strong>in</strong>geman PC et al. 2003. Design for<br />
stream restoration. Journal of Hydraulic Eng<strong>in</strong>eer<strong>in</strong>g 129 (8):<br />
575–584.<br />
Smeets E, Weter<strong>in</strong>gs R. 1999. Environmental Indicators: Typology<br />
and Overview. Technical Report No 25, European Environment<br />
Agency: Copenhagen, Denmark.<br />
Soar PJ, Thorne CR. 2001. Channel <strong>Restoration</strong> Design for<br />
Meander<strong>in</strong>g <strong>River</strong>s. ERDC/CHL CR-01-1, U.S. Army Corps<br />
of Eng<strong>in</strong>eers, Eng<strong>in</strong>eer Research and Development Center,<br />
Vicksburg, Massachusetts.<br />
Stanford JA et al. 1996. A general protocol for restoration of<br />
regulated rivers. Regulated <strong>River</strong>s: Research and Management<br />
12: 391–413.<br />
Stockwell BR. 2000. The Mary <strong>River</strong> and Tributaries Rehabilitation<br />
Plan: A Review of <strong>River</strong> Processes, Fluvial Geomorphology<br />
and Ecological Concepts and Their Application to a <strong>River</strong><br />
Rehabilitation Plan for a Major South-East Queensland Catchment.<br />
Masters Dissertation, University of New England, New<br />
South Wales, Australia.<br />
Tockner K, Stanford JA. 2002. <strong>River</strong><strong>in</strong>e fl ood pla<strong>in</strong>s: present<br />
state and future trends. Environmental Conservation 29 (3):<br />
308–330.<br />
Trochim WM. 2000. The Research Methods Knowledge Base,<br />
2nd Edition. Atomic Dog Publish<strong>in</strong>g: C<strong>in</strong>c<strong>in</strong>nati, Ohio.<br />
Trush WJ, McBa<strong>in</strong> SM, Leopold LB. 2000. Attributes of an alluvial<br />
river and <strong>the</strong>ir relation to water policy and management.<br />
Proceed<strong>in</strong>gs of <strong>the</strong> National Academy of Sciences of <strong>the</strong> United<br />
States of America 97 (22): 11858–11863.<br />
Van Asselt MBA. 2000. Perspectives on <strong>Uncerta<strong>in</strong>ty</strong> and Risk:<br />
The PRIMA Approach to Decision Support. Kluwer Academic<br />
Publishers: Dordrecht, The Ne<strong>the</strong>rlands.<br />
Van Asselt MBA, Rotmans J. 2002. <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>in</strong>tegrated<br />
assessment modell<strong>in</strong>g – From positivism to pluralism. Climatic<br />
Change 54 (1–2): 75–105.<br />
Vitousek PM, Mooney HA, Lubchenco J, Melillo JM. 1997.<br />
Human dom<strong>in</strong>ation of Earth’s ecosystems. Science 277:<br />
494–499.<br />
Walker J, Diamond M, Naura M. 2002. The development of<br />
Physical Quality Objectives for rivers <strong>in</strong> England and Wales.<br />
Aquatic Conservation: Mar<strong>in</strong>e and Freshwater Ecosystems<br />
12 (4): 381–390.<br />
Walters CJ. 1997. Challenges <strong>in</strong> adaptive management of riparian<br />
and coastal ecosystems. Conservation Ecology [onl<strong>in</strong>e]: http://<br />
www.consecol.org/vol1/iss2/art1.<br />
Ward JV, Tockner K, Arscott DB, Claret C. 2002. <strong>River</strong><strong>in</strong>e landscape<br />
diversity. Freshwater Biology 47 (4): 517–539.<br />
Ward T. 1998. Risk and uncerta<strong>in</strong>ty <strong>in</strong> environmental policy<br />
evaluation. In: Dore MHI, Mount TD (Eds), Global Enviornmental<br />
Economics: Equity and <strong>the</strong> Limits of Markets. Blackwell<br />
Publishers: Oxford, UK; 116–135.<br />
Werritty A. 2002. Liv<strong>in</strong>g with uncerta<strong>in</strong>ty: climate change, river<br />
fl ows and water resource management <strong>in</strong> Scotland. The Science<br />
of The Total Environment 294 (1–3): 29–40.<br />
Wheaton JM, Pasternack GB, Merz JE. 2004a. Spawn<strong>in</strong>g habitat<br />
rehabilitation – II. Us<strong>in</strong>g hypo<strong>the</strong>sis test<strong>in</strong>g and development<br />
<strong>in</strong> design, Mokelumne <strong>River</strong>, California, USA. International<br />
Journal of <strong>River</strong> Bas<strong>in</strong> Management 2 (1): 21–37.<br />
Wheaton JM, Pasternack GB, Merz JE. 2004b. Spawn<strong>in</strong>g habitat<br />
rehabilitation – I. Conceptual approach and methods. International<br />
Journal of <strong>River</strong> Bas<strong>in</strong> Management 2 (1): 3–20.<br />
Wheaton JM, Sear DA, Darby SE, Milne JA. 2004c. The International<br />
<strong>River</strong> <strong>Restoration</strong> Survey. http://www.geog.soton.ac.<br />
uk/users/WheatonJ/<strong>Restoration</strong>Survey_Cover.asp. Accessed<br />
on: 15/06/04.<br />
Wilson DW. 2001. On <strong>the</strong> Problem of Indeterm<strong>in</strong>acy <strong>in</strong> Fluvial<br />
Geomorphology. PhD Thesis, University of Southampton,<br />
Southampton, UK.
Wissmar RC, Beschta RL. 1998. <strong>Restoration</strong> and management of<br />
riparian ecosystems: a catchment perspective. Freshwater<br />
Biology 40: 571–585.<br />
Wissmar RC, Bisson PA. 2003a. Strategies for restor<strong>in</strong>g river<br />
ecosystems: Sources of variability and uncerta<strong>in</strong>ty. In: Wissmar<br />
RC, Bisson PA, Duke M (Eds), Strategies for Restor<strong>in</strong>g <strong>River</strong><br />
Ecosystems: Sources of Variability and <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Natural<br />
and Managed Systems. American Fisheries Society: Be<strong>the</strong>sda,<br />
Maryland; 3–7.<br />
Wissmar RC, Bisson PA. 2003b. Strategies for restor<strong>in</strong>g rivers:<br />
Problems and opportunites. In: Wissmar RC, Bisson PA, Duke<br />
M (Eds), Strategies for Restor<strong>in</strong>g <strong>River</strong> Ecosystems: Sources of<br />
The Scope of Uncerta<strong>in</strong>ties <strong>in</strong> <strong>River</strong> <strong>Restoration</strong> 39<br />
Variability and <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Natural and Managed Systems.<br />
American Fisheries Society: Be<strong>the</strong>sda, Maryland; 245–262.<br />
Wissmar RC, Bisson PA (Eds). 2003c. Strategies for Restor<strong>in</strong>g<br />
<strong>River</strong> Ecosystems: Sources of Variability and <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong><br />
Natural and Managed Systems. American Fisheries Society:<br />
Be<strong>the</strong>sda, Maryland.<br />
Wissmar RC, Braatne JH, Beschta RL, Rood SB. 2003. Variability<br />
of riparian ecosystems: Implications for restoration. In:<br />
Wissmar RC, Bisson PA, Duke M (Eds), Strategies for Restor<strong>in</strong>g<br />
<strong>River</strong> Ecosystems: Sources of Variability and <strong>Uncerta<strong>in</strong>ty</strong><br />
<strong>in</strong> Natural and Managed Systems. American Fisheries Society:<br />
Be<strong>the</strong>sda, Maryland; 107–127.
SECTION II<br />
Plann<strong>in</strong>g and Design<strong>in</strong>g<br />
<strong>Restoration</strong> Projects
<strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat Edited by Stephen Darby and David Sear<br />
© 2008 John Wiley & Sons, Ltd<br />
4<br />
Plann<strong>in</strong>g <strong>River</strong> <strong>Restoration</strong> Projects:<br />
Social and Cultural Dimensions<br />
4.1 INTRODUCTION<br />
G. Mathias Kondolf 1 and Chia-N<strong>in</strong>g Yang 2<br />
1 Department of Landscape Architecture and Environmental Plann<strong>in</strong>g, University of California, USA<br />
2 Department of Landscape Architecture, California State Polytechnic University, USA<br />
Nationwide, s<strong>in</strong>ce 1990, at least US$17 billion has been<br />
spent on restoration projects, a fi gure that is an underestimate<br />
because most reported costs do not <strong>in</strong>clude staff<br />
time and many projects did not report <strong>the</strong>ir costs at all<br />
(Bernhardt et al., 2005). In many areas, river restoration<br />
has become an <strong>in</strong>dustry, with nonprofi t groups, government<br />
agencies and consult<strong>in</strong>g fi rms now depend<strong>in</strong>g upon<br />
river restoration funds to support large components of<br />
<strong>the</strong>ir budgets. For example, over four fi scal years from July<br />
2000 to June 2004, over US$100 million was disbursed<br />
by <strong>the</strong> California Department of Fish and Game to recipient<br />
groups and agencies <strong>in</strong> <strong>the</strong> Fishery <strong>Restoration</strong> Grants<br />
Program for restoration projects <strong>in</strong> coastal river bas<strong>in</strong>s<br />
(F. Sime, personal communication, December 2003),<br />
mostly to construct habitat enhancement structures <strong>in</strong><br />
salmon-bear<strong>in</strong>g rivers and streams. Elsewhere <strong>in</strong> California,<br />
stream restoration projects employ many <strong>in</strong> rural communities,<br />
<strong>in</strong>clud<strong>in</strong>g former timber cutters (Hamilton,<br />
1993). In <strong>the</strong> city of Bozeman, Montana, <strong>the</strong>re is suffi cient<br />
bus<strong>in</strong>ess to support six fi rms specializ<strong>in</strong>g <strong>in</strong> restor<strong>in</strong>g<br />
trout streams.<br />
<strong>River</strong> and stream restoration can be viewed as a contemporary<br />
phase of <strong>the</strong> environmental movement. Unlike<br />
early phases of <strong>the</strong> movement, which tended to document<br />
and draw attention to <strong>the</strong> nature and extent of environmental<br />
degradation (Carson, 1962; Ehrlich, 1968) and which<br />
<strong>the</strong>refore tended to be negative or pessimistic <strong>in</strong> tone,<br />
restor<strong>in</strong>g rivers and streams has a positive, pro-active connotation.<br />
This is especially true of streams <strong>in</strong> urban neigh-<br />
borhoods, where restoration projects can provide positive<br />
re<strong>in</strong>forcement and a sense of empowerment to local<br />
groups. In many respects, <strong>the</strong> greatest benefi ts to restor<strong>in</strong>g<br />
local urban creeks are probably <strong>the</strong> social benefi ts that<br />
accrue through <strong>the</strong> process of community build<strong>in</strong>g and <strong>the</strong><br />
public environmental education achieved.<br />
Technical specialists often assume implicitly that restoration<br />
is a technical problem, and a glance at articles<br />
published <strong>in</strong> restoration-related journals shows a preponderance<br />
of papers address<strong>in</strong>g <strong>the</strong> scientifi c aspects of<br />
project design and plann<strong>in</strong>g. However, restoration can be<br />
viewed as fundamentally a social phenomenon, as it results<br />
from a societal decision to restore some functions to a<br />
river (Eden et al., 2000). Its goals and implementation<br />
approaches can be <strong>in</strong>formed by science, but <strong>the</strong>y are<br />
essentially social <strong>in</strong> nature. The very fact that restoration<br />
has become such a widespread activity refl ects a change<br />
<strong>in</strong> public attitudes towards watercourses. The current attitudes<br />
are possible only thanks to past social <strong>in</strong>vestments<br />
<strong>in</strong> waste water treatment follow<strong>in</strong>g passage of <strong>the</strong> Clean<br />
Water Act <strong>in</strong> <strong>the</strong> United States (Wolman, 1971) and<br />
comparable legislation <strong>in</strong> o<strong>the</strong>r developed countries. The<br />
result<strong>in</strong>g improvements <strong>in</strong> water quality now make human<br />
contact with urban waters desirable and ecological restoration<br />
feasible, which was not <strong>the</strong> case <strong>in</strong> <strong>the</strong> past when <strong>the</strong>se<br />
channels were open sewers. As society and culture evolve,<br />
goals change and <strong>the</strong> realm of what is ‘feasible’ <strong>in</strong> river<br />
restoration can change dramatically, <strong>in</strong>troduc<strong>in</strong>g uncerta<strong>in</strong>ties<br />
for restoration plann<strong>in</strong>g and design – uncerta<strong>in</strong>ties<br />
that do not lend <strong>the</strong>mselves to technical eng<strong>in</strong>eer<strong>in</strong>g<br />
analysis.
44 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
This chapter considers social and cultural dimensions<br />
of river restoration, and related uncerta<strong>in</strong>ties for restoration<br />
plann<strong>in</strong>g and design. While recognized as important,<br />
<strong>the</strong>y are probably less well understood by <strong>the</strong> agencies<br />
<strong>in</strong>volved <strong>in</strong> design<strong>in</strong>g and implement<strong>in</strong>g restoration projects<br />
(FISCRWG, 1998). Systematic studies of public<br />
attitudes and expectations regard<strong>in</strong>g river restoration<br />
(Tunstall et al., 2000) have been rare <strong>in</strong> light of <strong>the</strong> enormous<br />
societal <strong>in</strong>vestment <strong>in</strong> restoration projects. There is<br />
no way this short chapter can do justice to this broad topic,<br />
but here<strong>in</strong> we attempt to raise some issues relevant to <strong>the</strong><br />
enterprise of river restoration, which we hope will be<br />
useful to scientists and practitioners <strong>in</strong> <strong>the</strong> fi eld. We fi rst<br />
briefl y review some important social aspects of river restoration:<br />
<strong>the</strong> land-use context of restoration projects,<br />
underly<strong>in</strong>g cultural preferences <strong>in</strong> river restoration design<br />
and <strong>the</strong> <strong>in</strong>creas<strong>in</strong>g importance of public participation<br />
<strong>in</strong> river management and restoration programs. We<br />
draw upon recent research to consider various human<br />
activities <strong>in</strong> urban streams and <strong>the</strong> confl icts among restoration<br />
goals of various professionals and stakeholder<br />
groups. F<strong>in</strong>ally, we present two brief case studies from<br />
nor<strong>the</strong>rn California, which illustrate social issues and<br />
attendant uncerta<strong>in</strong>ties <strong>in</strong> river restoration.<br />
4.2 OVERVIEW OF SOCIAL ASPECTS OF<br />
RIVER RESTORATION<br />
4.2.1 An Urban–Rural–Wilderness Cont<strong>in</strong>uum<br />
Appropriate goals and <strong>the</strong> solutions possible vary widely<br />
with context, from near wilderness to dense urban sett<strong>in</strong>gs<br />
(Figure 4.1). Where catchment processes are relatively<br />
unaltered and runoff and sediment load are virtually<br />
Attributes<br />
General approach<br />
Examples<br />
WILDERNESS<br />
Unaltered watershed.<br />
Channel may be altered.<br />
Can restore pre-disturbance,<br />
historical channel, by ei<strong>the</strong>r:<br />
1) lett<strong>in</strong>g river restore itself<br />
2) ‘carbon-copy’ approach<br />
Flow regime unchanged.<br />
Sediment load unchanged.<br />
No urban encroachment.<br />
Figure 4.1 An urban–wilderness cont<strong>in</strong>uum <strong>in</strong> river restoration<br />
unchanged, a restoration project can logically seek to<br />
restore pre-disturbance channel conditions, ei<strong>the</strong>r by<br />
giv<strong>in</strong>g fl oods and sediment transport <strong>the</strong> opportunity to<br />
recreate natural channel conditions (an approach often<br />
termed ‘passive restoration’) or by proactively reconstruct<strong>in</strong>g<br />
pre-disturbance channel form (<strong>the</strong> ‘carboncopy’<br />
approach of Brookes and Shields, 1996). An example<br />
would be a channel whose catchment land use has rema<strong>in</strong>ed<br />
constant, but whose form was altered by channel straighten<strong>in</strong>g<br />
or by removal of bank vegetation and consequent<br />
<strong>in</strong>stability. At this wilderness end of <strong>the</strong> cont<strong>in</strong>uum, it<br />
makes sense to ei<strong>the</strong>r let natural processes accomplish<br />
<strong>the</strong> restoration or to use <strong>the</strong> pre-disturbance channel as a<br />
template, because <strong>the</strong> processes that supported <strong>the</strong> predisturbance<br />
channel will tend to support <strong>the</strong> same channel<br />
form aga<strong>in</strong>. In ei<strong>the</strong>r case, to ma<strong>in</strong>ta<strong>in</strong> ecological values,<br />
<strong>the</strong> channel should be permitted to migrate freely and to<br />
fl ood overbank areas (Ward and Stanford, 1995).<br />
Mov<strong>in</strong>g towards <strong>the</strong> urbanized end of <strong>the</strong> cont<strong>in</strong>uum,<br />
land use change <strong>in</strong> <strong>the</strong> catchment has altered runoff<br />
and sediment load, so <strong>the</strong>re is no reason to expect predisturbance<br />
channel dimensions to be ma<strong>in</strong>ta<strong>in</strong>ed by<br />
current processes. At <strong>the</strong> extreme, urban development <strong>in</strong><br />
<strong>the</strong> catchment <strong>in</strong>creases peak fl ows such that <strong>the</strong> channel<br />
tends to <strong>in</strong>cise, which, if uncontrolled, may lead to bank<br />
collapse and channel widen<strong>in</strong>g. However, encroachment<br />
of urban development to <strong>the</strong> channel marg<strong>in</strong>s means that<br />
<strong>in</strong>cision and channel widen<strong>in</strong>g are socially unacceptable.<br />
At this urban extreme, restoration projects must be built<br />
to convey urban runoff without fl ood<strong>in</strong>g adjacent lands<br />
and to withstand <strong>in</strong>creased shear stresses of urban runoff<br />
without erosion. Here, restoration can be viewed as a form<br />
of garden<strong>in</strong>g, <strong>in</strong> which <strong>the</strong> elements are deliberately<br />
chosen and ma<strong>in</strong>ta<strong>in</strong>ed by human <strong>in</strong>put, albeit one that<br />
HIGHLY URBAN<br />
Transitional cases. Highly altered watershed and channel.<br />
Encroached banks.<br />
Can partly restore processes.<br />
Must decide what changes to accept as<br />
constra<strong>in</strong>ts, what to try to change/restore.<br />
Increase releases from dam?<br />
Reduce peak urban flow by detention?<br />
Add gravel below dams?<br />
Reduce erosion <strong>in</strong> watershed?<br />
Remove houses along bank/floodpla<strong>in</strong>?<br />
Cannot restore historical conditions.<br />
<strong>Restoration</strong><br />
as<br />
‘ garden<strong>in</strong>g:<br />
’<br />
choose elements to <strong>in</strong>clude but must account for<br />
erosive forces, altered hydrology.<br />
Social issues are important: potential to improve<br />
ecology/water quality are limited, so emphasis on<br />
community-build<strong>in</strong>g and environmental education.<br />
Flow regime altered.<br />
Sediment load altered.<br />
Urban encroachment to banks.<br />
Attributes<br />
General approach<br />
Examples
equires hard structures to resist erosive forces of urbanrunoff-augmented<br />
fl oods. In such cases, ‘naturalness’ <strong>in</strong><br />
restoration may be viewed as more an aes<strong>the</strong>tic choice<br />
than a real design approach. Such channels can be highly<br />
successful <strong>in</strong> provid<strong>in</strong>g recreational and aes<strong>the</strong>tic amenities,<br />
l<strong>in</strong>k<strong>in</strong>g communities through walk<strong>in</strong>g and bik<strong>in</strong>g<br />
trails, even provid<strong>in</strong>g for kayak<strong>in</strong>g and canoe<strong>in</strong>g. However,<br />
<strong>the</strong>y may be viewed as ‘water features’ (to borrow a term<br />
from <strong>the</strong> fi eld of landscape architecture) capable of convey<strong>in</strong>g<br />
fl oodwaters with<strong>in</strong> <strong>the</strong> channel and without erod<strong>in</strong>g<br />
banks, ra<strong>the</strong>r than large scale, dynamic ecosystems. They<br />
can still provide ecological benefi ts at a local scale, but<br />
without rebuild<strong>in</strong>g of <strong>the</strong> urban <strong>in</strong>frastructure <strong>the</strong>se waterways<br />
are unlikely to support sensitive target species at a<br />
large scale. In <strong>the</strong>se highly urban sett<strong>in</strong>gs, <strong>the</strong> ecological<br />
potential from a restoration project can rarely be comparable<br />
to that achieved <strong>in</strong> a less urban sett<strong>in</strong>g, and thus<br />
urban restoration projects may be best justifi ed by <strong>the</strong>ir<br />
potential social benefi ts as <strong>the</strong>y respond to human needs<br />
and uses.<br />
Many restoration projects can be seen to fall on a cont<strong>in</strong>uum<br />
between <strong>the</strong>se two extremes, with constra<strong>in</strong>ts, but<br />
with <strong>the</strong> potential to restore some natural processes and<br />
functions. It is <strong>in</strong> <strong>the</strong>se <strong>in</strong>termediate cases that <strong>the</strong> greatest<br />
uncerta<strong>in</strong>ties arise, as one person’s ‘constra<strong>in</strong>t’ may be<br />
ano<strong>the</strong>r’s opportunity to restore process. For example, if<br />
an upstream reservoir has elim<strong>in</strong>ated <strong>the</strong> natural magnitude<br />
and frequency of fl oods, should we accept this as<br />
a constra<strong>in</strong>t that effectively limits <strong>the</strong> degree to which<br />
natural ecosystem processes can be restored, or do we<br />
seek to alter <strong>the</strong> reservoir operation rules to more closely<br />
mimic natural fl ow patterns? Reservoir release patterns<br />
have been altered and aquatic ecological conditions<br />
improved on rivers such as <strong>the</strong> Green <strong>River</strong>, Kentucky<br />
(Postel and Ritcher, 2004), <strong>the</strong> St Mary <strong>River</strong>, Alberta<br />
(Rood and Mahoney, 2000) and Putah Creek, California<br />
(Marchetti and Moyle, 2001). Similarly, does <strong>the</strong> existence<br />
of human <strong>in</strong>frastructure or hous<strong>in</strong>g on a fl oodpla<strong>in</strong><br />
mean we cannot <strong>in</strong>undate this fl oodpla<strong>in</strong>? Or should <strong>the</strong><br />
restoration project <strong>in</strong>clude compensation for mov<strong>in</strong>g <strong>the</strong><br />
<strong>in</strong>appropriately-sited land use to higher ground, so overbank<br />
fl ood<strong>in</strong>g processes can be restored? These are social/<br />
political decisions, which can be <strong>in</strong>formed by science, but<br />
which cannot be predicted technically – add<strong>in</strong>g substantial<br />
uncerta<strong>in</strong>ty to restoration plann<strong>in</strong>g.<br />
4.2.2 Cultural Preferences <strong>in</strong> <strong>River</strong><br />
<strong>Restoration</strong> Design<br />
Unstated and often unacknowledged cultural preferences<br />
probably underlie many restoration design decisions. For<br />
example, grassy banks are preferred over shrubby or<br />
Plann<strong>in</strong>g <strong>River</strong> <strong>Restoration</strong> Projects: Social and Cultural Dimensions 45<br />
wooded banks along many restored streams <strong>in</strong> nor<strong>the</strong>rn<br />
Europe, refl ect<strong>in</strong>g <strong>the</strong> long history of pastoral land use.<br />
Similarly, open park-like landscape seems to be broadly<br />
preferred <strong>in</strong> western culture (Appleton, 1975), and residents<br />
near urban stream restoration projects <strong>in</strong> nor<strong>the</strong>rn<br />
California have compla<strong>in</strong>ed when restored streams become<br />
too ‘bushy’ and woody riparian vegetation blocks visual<br />
access to <strong>the</strong> stream bed (Purcell et al., 2002). Similarly,<br />
large woody debris <strong>in</strong> channels imparts a messy look, to<br />
which most people have a negative reaction (Piégay et al.,<br />
2005).<br />
In North America, restoration projects seek<strong>in</strong>g to create<br />
stable, symmetrically-meander<strong>in</strong>g channels have proliferated.<br />
In some cases, previously s<strong>in</strong>gle-thread channels<br />
have been reconstructed <strong>in</strong> attempts to create a more ideal,<br />
symmetrical meander<strong>in</strong>g form <strong>in</strong> <strong>the</strong> belief that <strong>the</strong>se<br />
would be more stable (Smith and Prestegaard, 2005). In<br />
o<strong>the</strong>r cases, <strong>the</strong> channels have been reconstructed with <strong>the</strong><br />
goal of convert<strong>in</strong>g braided rivers to s<strong>in</strong>gle-thread, meander<strong>in</strong>g<br />
rivers. In many cases, <strong>the</strong> streams so ‘restored’ were<br />
never s<strong>in</strong>gle-thread meander<strong>in</strong>g channels under natural<br />
conditions, and <strong>the</strong> projects can be viewed as essentially<br />
attempts to impose an idealized meander<strong>in</strong>g form onto <strong>the</strong><br />
river, as illustrated on Uvas Creek, California (Kondolf<br />
et al., 2001). Many of <strong>the</strong>se channel reconstructions have<br />
washed out with<strong>in</strong> months or years (Figure 4.2). Despite<br />
its mixed record of performance, <strong>the</strong> design approach<br />
underly<strong>in</strong>g most of <strong>the</strong>se projects – application of <strong>the</strong><br />
classifi cation scheme of Rosgen (1994) (NRC, 1992) –<br />
cont<strong>in</strong>ues to be popular among government agencies<br />
responsible for fund<strong>in</strong>g restoration projects. This is<br />
probably due to <strong>the</strong> ease with which <strong>the</strong> classifi cation<br />
scheme can be used and applied by those without academic<br />
tra<strong>in</strong><strong>in</strong>g <strong>in</strong> fl uvial geomorphology, <strong>the</strong> availability<br />
of commercial short courses teach<strong>in</strong>g users how to apply<br />
<strong>the</strong> scheme and – though largely unrealized and unacknowledged<br />
– <strong>the</strong> likelihood that <strong>the</strong> channel designs that<br />
result from apply<strong>in</strong>g <strong>the</strong> scheme satisfy a deep-seated<br />
cultural preference for stable, s<strong>in</strong>gle-thread meander<strong>in</strong>g<br />
channels.<br />
Research on human responses to landscape form suggest<br />
that subjects (at least <strong>in</strong> western culture) tend to prefer<br />
<strong>the</strong> ‘defl ected vistas’ <strong>in</strong> curved paths, rivers and valleys<br />
over straight l<strong>in</strong>es (Appleton, 1975), <strong>in</strong> part because <strong>the</strong>y<br />
elicited curiosity <strong>in</strong> subjects (Ulrich, 1983). Kaplan and<br />
Kaplan (1984) designated this landscape property as<br />
‘mystery’, convey<strong>in</strong>g <strong>the</strong> opportunity to explore and a<br />
promise to learn more with a chang<strong>in</strong>g vantage po<strong>in</strong>t as<br />
one moves more deeply <strong>in</strong>to <strong>the</strong> scene. What is probably<br />
a (near-) universal attraction to <strong>the</strong> form of meander<strong>in</strong>g<br />
channels was recognized <strong>in</strong> <strong>the</strong> 18th century by Hogarth<br />
(1753), who proposed that <strong>the</strong> ‘serpent<strong>in</strong>e’ l<strong>in</strong>e provided
46 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
a<br />
b<br />
Figure 4.2 (See also colour plate section) Uvas Creek viewed<br />
downstream from Santa Teresa Road bridge: (a) January 1996,<br />
two months after completion of <strong>the</strong> channel reconstruction<br />
project; (b) July 1997, after <strong>the</strong> constructed channel washed out<br />
<strong>in</strong> February 1996 dur<strong>in</strong>g an approximately six-year fl ow (Photo<br />
(a) courtesy of <strong>the</strong> City of Gilroy, (b) by Kondolf.)<br />
<strong>the</strong> greatest aes<strong>the</strong>tic pleasure, and more so when actively<br />
mov<strong>in</strong>g. A river mov<strong>in</strong>g through a meander<strong>in</strong>g channel<br />
thus has <strong>the</strong> elements needed for <strong>the</strong> experience of beauty<br />
under Hogarth’s <strong>the</strong>ory. This preference also found expression<br />
<strong>in</strong> <strong>the</strong> work of late 18th century English landscape<br />
designers such as Capability Brown, who built serpent<strong>in</strong>e<br />
channels on <strong>the</strong> estates of <strong>the</strong>ir wealthy clients.<br />
Despite <strong>the</strong> evidence for a landscape preference for <strong>the</strong><br />
meander<strong>in</strong>g channel form, <strong>the</strong> justifi cations for meander<strong>in</strong>g<br />
channels specifi ed <strong>in</strong> river restoration projects <strong>in</strong> North<br />
America are almost always stated <strong>in</strong> terms of bankfull<br />
discharge, width–depth ratios, meander wavelengths etc.<br />
The fundamental question as to whe<strong>the</strong>r a meander<strong>in</strong>g<br />
channel is appropriate at all is rarely addressed. Similarly,<br />
<strong>the</strong> notion that channels should be stable can be viewed<br />
as largely anthropocentric. Dynamic channels with variable<br />
fl ow regimes tend to support <strong>the</strong> greatest variety of<br />
habitats and best ecosystem function (Ward and Stanford,<br />
1995; Poff et al., 1997). Yet <strong>the</strong> meander<strong>in</strong>g channels<br />
constructed us<strong>in</strong>g <strong>the</strong> Rosgen approach have <strong>the</strong> outside<br />
of meander bends armored by root wads and boulders,<br />
with rock weirs at <strong>the</strong> crossovers to keep <strong>the</strong> ma<strong>in</strong> current<br />
away from <strong>the</strong> banks (e.g. Uvas Creek <strong>in</strong> Figure 4.2(a)).<br />
Indeed, <strong>the</strong> Rosgen scheme is used to select <strong>the</strong> ‘proper’<br />
geometry for a site, ‘proper’ mean<strong>in</strong>g it will be stable. We<br />
do not argue with <strong>the</strong> need to armor channels <strong>in</strong> dense<br />
urban areas or elsewhere when <strong>in</strong>frastructure is threatened<br />
by channel migration, but <strong>the</strong>se restoration projects typically<br />
<strong>in</strong>clude armored banks even at sites where channel<br />
migration would not threaten human works. The armor<strong>in</strong>g<br />
seems to be accepted <strong>in</strong> part because it consists of ‘natural’<br />
materials (i.e. it is not concrete) and because those <strong>in</strong>volved<br />
<strong>in</strong> fund<strong>in</strong>g and design<strong>in</strong>g <strong>the</strong>se projects hold a belief that<br />
a stable channel is preferable to an erod<strong>in</strong>g channel, even<br />
if <strong>in</strong> a rural or park sett<strong>in</strong>g.<br />
F<strong>in</strong>ally, stable, meander<strong>in</strong>g channels, fl anked by grassy<br />
banks, probably appeal to our aes<strong>the</strong>tic senses <strong>in</strong> large part<br />
because <strong>the</strong>y are ‘tidy’ landscapes. Natural riparian corridors<br />
are frequently <strong>in</strong>accessible thickets, which, while<br />
great habitat for wildlife, are unappeal<strong>in</strong>g to our western<br />
aes<strong>the</strong>tic sensibility. Nassauer (1995) demonstrated that<br />
for such ‘messy’ ecosystems to be widely accepted, we<br />
must set <strong>the</strong>m off with<strong>in</strong> a frame that conveys to <strong>the</strong><br />
viewer that <strong>the</strong> mess<strong>in</strong>ess is deliberate and not a sign of<br />
neglect. She demonstrated how ‘cues to care’ such as a<br />
neatly ma<strong>in</strong>ta<strong>in</strong>ed fence around a yard of native prairie<br />
could make <strong>the</strong> o<strong>the</strong>rwise messy bit of landscape acceptable<br />
with<strong>in</strong> <strong>the</strong> context of a suburban street.<br />
To <strong>the</strong> extent that public support for restoration is based<br />
on culturally-driven landscape preferences that are not<br />
recognized or articulated, this creates enormous uncerta<strong>in</strong>ty<br />
<strong>in</strong> river restoration projects, as public support cannot<br />
be predicted based on ‘logical’ analysis of how best to<br />
improve aquatic ecology or to manage fl oods. There is<br />
ano<strong>the</strong>r factor, which cannot be predicted by technical<br />
experts. The topic of human preference <strong>in</strong> landscape is an<br />
area of active research. Many of <strong>the</strong> fi nd<strong>in</strong>gs probably<br />
have relevance for river restoration, besides <strong>the</strong> few<br />
touched upon here.<br />
4.2.3 Public Participation and Active Stakeholders<br />
Today, public participation has become an <strong>in</strong>stitutionalized<br />
element <strong>in</strong> stream restoration. Public acceptance and<br />
support <strong>in</strong> many cases determ<strong>in</strong>es <strong>the</strong> ultimate success and<br />
susta<strong>in</strong>ability of a project. For example, provid<strong>in</strong>g public<br />
access to a restoration plan can substantially <strong>in</strong>crease<br />
public support for <strong>the</strong> plan (Bauer et al., 2002). Support<br />
for restoration is important not only <strong>in</strong> advocat<strong>in</strong>g for <strong>the</strong><br />
proposed project, but also <strong>in</strong> its stewardship after construction.<br />
Stewardship can be developed by encourag<strong>in</strong>g<br />
people to experience <strong>the</strong> restored natural areas (Ryan<br />
et al., 2002).
Increas<strong>in</strong>gly <strong>the</strong> role of stakeholders is not limited to<br />
provid<strong>in</strong>g review comments on draft documents, but<br />
to active participation <strong>in</strong> sett<strong>in</strong>g objectives and select<strong>in</strong>g<br />
implementation strategies. The success of such a collaborative<br />
plann<strong>in</strong>g process is often evaluated by whe<strong>the</strong>r or<br />
not agreement is reached among <strong>in</strong>terest groups. This<br />
approach implicitly assumes that <strong>the</strong>re is an optimal solution<br />
that satisfi es all <strong>in</strong>terests and is technically feasible.<br />
However, <strong>the</strong>re is no a priori reason to assume that this<br />
is <strong>the</strong> case, and <strong>in</strong> fact <strong>the</strong>re are good reasons to expect<br />
it frequently will not be. Accord<strong>in</strong>gly, confl icts can arise<br />
among different actors, such as between stakeholder<br />
groups, between professionals <strong>in</strong> different fi elds and<br />
between design professionals and residents, all creat<strong>in</strong>g<br />
uncerta<strong>in</strong>ties for restoration plann<strong>in</strong>g. These are discussed<br />
<strong>in</strong> more detail <strong>in</strong> <strong>the</strong> follow<strong>in</strong>g pages. Although water<br />
policy mak<strong>in</strong>g and plann<strong>in</strong>g rema<strong>in</strong>s a much contended<br />
arena <strong>in</strong> California, collaborative plann<strong>in</strong>g or policy<br />
mak<strong>in</strong>g has been documented as benefi cial not simply<br />
based on whe<strong>the</strong>r or not consent is reached among various<br />
stakeholder groups, but through <strong>the</strong> long term, <strong>in</strong>visible<br />
outcomes <strong>in</strong> terms of collective learn<strong>in</strong>g and accumulation<br />
of social, political and economic capitals (Connick and<br />
Innes, 2003).<br />
An important feature of river restoration today is <strong>the</strong><br />
proliferation of local creek groups, known <strong>in</strong> <strong>the</strong> United<br />
States as ‘Friends of’ <strong>the</strong> local creek, <strong>in</strong> <strong>the</strong> United<br />
K<strong>in</strong>gdom as river ‘Trusts’ (e.g. <strong>the</strong> Eden <strong>River</strong>s Trust). In<br />
<strong>the</strong> San Francisco Bay Area, <strong>the</strong>se groups have formed a<br />
signifi cant force <strong>in</strong> shap<strong>in</strong>g <strong>the</strong> fate of restoration projects.<br />
Friends groups not only voice <strong>the</strong>ir desires dur<strong>in</strong>g restoration<br />
plann<strong>in</strong>g, but <strong>in</strong> many cases <strong>the</strong>y have become<br />
<strong>the</strong> task force of implement<strong>in</strong>g plans and management<br />
regimes. To <strong>the</strong> restoration project designer, <strong>the</strong> potential<br />
role of local creek groups is a source of uncerta<strong>in</strong>ty. If a<br />
local group is active, it is important to work closely with<br />
it, both to improve <strong>the</strong> project design with respect to its<br />
social function<strong>in</strong>g and to improve <strong>the</strong> chances of successful<br />
implementation and susta<strong>in</strong>ability by virtue of <strong>the</strong><br />
public support a local group can often provide.<br />
There are also fundamental issues with representativeness<br />
<strong>in</strong> <strong>the</strong> stakeholder and public participation process.<br />
These processes can be drawn-out, and <strong>the</strong> long term<br />
active participants tend to be agency staff or <strong>in</strong>dustry<br />
representatives for whom participation is part of <strong>the</strong>ir job,<br />
or staff of NGOs who are often stretched th<strong>in</strong>ly amongst<br />
many such processes. To actively participate, members of<br />
<strong>the</strong> public at large must have <strong>the</strong> time and energy to devote<br />
to meet<strong>in</strong>gs over a long period (often exceed<strong>in</strong>g a year) at<br />
<strong>the</strong>ir own expense. Unless <strong>the</strong>y are strongly motivated –<br />
often by an imm<strong>in</strong>ent threat such as stopp<strong>in</strong>g a development<br />
<strong>in</strong> <strong>the</strong>ir neighborhood – few can fi nd <strong>the</strong> time to be<br />
Plann<strong>in</strong>g <strong>River</strong> <strong>Restoration</strong> Projects: Social and Cultural Dimensions 47<br />
active <strong>in</strong> <strong>the</strong> public participation process. This is refl ected<br />
by <strong>the</strong> survey results of <strong>the</strong> ‘befriended’ watersheds <strong>in</strong><br />
<strong>the</strong> San Francisco Bay Area. Neighborhoods with active<br />
‘Friends’ groups have a much higher average <strong>in</strong>come than<br />
areas that do not form creek groups (Moz<strong>in</strong>go, 2005),<br />
show<strong>in</strong>g urban stream stewardship <strong>in</strong> <strong>the</strong> United States<br />
still serves a clientele biased toward <strong>the</strong> upper and middle<br />
classes.<br />
4.3 HUMAN USES OF URBAN WATERWAYS<br />
While <strong>the</strong> habitat requirements of fi sh have been extensively<br />
studied (Reiser and Bjornn, 1979) and are used as<br />
a basis for design of restoration projects oriented towards<br />
salmon and trout (Flosi et al., 1998), <strong>the</strong> habitat requirements<br />
of humans <strong>in</strong> <strong>the</strong> stream environment, broadly<br />
construed, are less well understood. Recent research <strong>in</strong>to<br />
why certa<strong>in</strong> activities occur spontaneously at certa<strong>in</strong> parts<br />
of <strong>the</strong> stream suggests that <strong>the</strong>re are fundamental characteristics<br />
of streams that encourage, and can be designed<br />
for, recreational use. Here we review a range of human<br />
uses of stream corridors, emphasiz<strong>in</strong>g urban and suburban<br />
sett<strong>in</strong>gs.<br />
4.3.1 Camp<strong>in</strong>g by Homeless<br />
Riparian corridors have long been preferred sites for<br />
camp<strong>in</strong>g by homeless people. <strong>River</strong> corridors were sites of<br />
large camps of migratory workers and tramps <strong>in</strong> North<br />
America dur<strong>in</strong>g <strong>the</strong> depression of <strong>the</strong> 1930s, and homeless<br />
encampments are a common element along urban streams<br />
<strong>in</strong> California today, offer<strong>in</strong>g a relatively secluded refuge for<br />
<strong>the</strong> ‘down and out’. Homeless camps are often found under<br />
bridges, exploit<strong>in</strong>g <strong>the</strong> shelter from ra<strong>in</strong>, although <strong>the</strong>se<br />
sites are more accessible and thus more likely to be visited<br />
by o<strong>the</strong>rs and less private (Figure 4.3). Along Ledgewood<br />
Creek near Fairfi eld, California, a camp of fi fty residents<br />
had tents with carpeted fl oors, furniture, and batterypowered<br />
television; <strong>the</strong> residents reportedly left dur<strong>in</strong>g<br />
periodic police sweeps, only to return (Fagan, 2005).<br />
Migrants from <strong>the</strong> prov<strong>in</strong>ces of Cuba have settled along<br />
<strong>the</strong> banks of <strong>the</strong> Almendares <strong>River</strong> <strong>in</strong> Havana, form<strong>in</strong>g a<br />
squatter community known as ‘El Fangito’ (Figure 4.4).<br />
While <strong>the</strong> streets are mud, many of <strong>the</strong>se dwell<strong>in</strong>gs feature<br />
cement or tiled fl oors, furniture and television sets.<br />
Although <strong>the</strong> settlement was illegal, utilities have hooked<br />
up electrical power and water; sewage fl ows mostly<br />
through buried pipes directly to <strong>the</strong> river. The fl oodpla<strong>in</strong><br />
occupied by El Fangito is fl ooded every few years. Residents<br />
take <strong>the</strong>ir television sets and leave for higher ground<br />
when <strong>the</strong> river beg<strong>in</strong>s to rise. The management plan of <strong>the</strong><br />
Metropolitan Park of Havana (Fornes, 1994) calls for
48 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Figure 4.3 (See also colour plate section) Homeless campsite,<br />
San Pablo Creek, California (Photo by Kondolf, February<br />
2005.)<br />
Figure 4.4 (See also colour plate section) The squatter neighborhood<br />
El Fangito, Havana (Photo by Kondolf, March 2005.)<br />
mov<strong>in</strong>g <strong>the</strong>se residents from El Fangito to better, permanent<br />
hous<strong>in</strong>g and reforestation of <strong>the</strong> fl oodpla<strong>in</strong>, but an<br />
<strong>in</strong>ternational aid agency recently granted funds to build<br />
a levee around <strong>the</strong> settlement to protect it from<br />
fl oodwaters.<br />
The authors are aware of no studies focus<strong>in</strong>g on homeless<br />
use of river<strong>in</strong>e spaces, but our fi eld studies <strong>in</strong><br />
California suggest that this user group has little tolerance<br />
for o<strong>the</strong>r user groups and vice versa. We have observed<br />
that no o<strong>the</strong>r users were present near homeless camps<br />
(usually under bridges and beh<strong>in</strong>d thickets on fl oodpla<strong>in</strong>s).<br />
In Sonoma, California, children were often warned off by<br />
parents or scared away when <strong>the</strong>y accidentally <strong>in</strong>vaded<br />
homeless territory (Yang, 2004). In Japan, while homeless<br />
people also frequently reside under bridges of urban<br />
streams, <strong>the</strong> tension was not as high, as homelessness <strong>in</strong><br />
Japan is regarded ma<strong>in</strong>ly as a product of unjust <strong>in</strong>dustrial<br />
structure (Dohi, 1999), with little association with drug<br />
abuse and crime. In America, <strong>the</strong> fl ood of homelessness<br />
dur<strong>in</strong>g <strong>the</strong> past four decades is largely attributed to <strong>the</strong><br />
failure of ‘de<strong>in</strong>stitutionalization,’ a major <strong>in</strong>itiative under<br />
<strong>the</strong> Community Mental Health program that started <strong>in</strong><br />
1963 (CCHR, 2004).<br />
Occupation of river corridors by homeless can be an<br />
important source of uncerta<strong>in</strong>ty to <strong>the</strong> outcome of river<br />
restoration efforts. Use of stream corridors by homeless<br />
people has not (to <strong>the</strong> authors’ knowledge) been encouraged<br />
by designers. However, it is clearly one of <strong>the</strong> biggest<br />
uses along many urban rivers and streams. Because <strong>the</strong><br />
presence of homeless people could discourage use by<br />
o<strong>the</strong>r groups, <strong>the</strong> actual use of a restored stream corridor<br />
may be very different from that anticipated by project<br />
designers, <strong>in</strong>troduc<strong>in</strong>g uncerta<strong>in</strong>ties.<br />
4.3.2 Fish<strong>in</strong>g<br />
Fish<strong>in</strong>g is a traditional use of rivers and streams, rang<strong>in</strong>g<br />
from subsistence fi sh<strong>in</strong>g with traps and nets to purely sport<br />
fi sh<strong>in</strong>g <strong>in</strong> which <strong>the</strong> fi sh is released back to <strong>the</strong> stream.<br />
Fisheries <strong>in</strong> urban channels range widely from wild, anadromous<br />
salmonids <strong>in</strong> urban channels <strong>in</strong> <strong>the</strong> Pacifi c Northwest<br />
of North America to warm water species pulled from<br />
<strong>the</strong> polluted waters of Asian cities. Fish<strong>in</strong>g is usually well<br />
regulated by licens<strong>in</strong>g and many streams are artifi cially<br />
stocked. Fish<strong>in</strong>g is a well documented and well studied<br />
activity <strong>in</strong> rivers, a large subject well treated elsewhere<br />
and beyond <strong>the</strong> scope of this chapter. However, we po<strong>in</strong>t<br />
out that fi sh<strong>in</strong>g has long been an important activity draw<strong>in</strong>g<br />
people to rivers, similar to o<strong>the</strong>r activities we discuss<br />
below. Improv<strong>in</strong>g fi sh habitat is cited as a goal for many<br />
restoration projects and many fund<strong>in</strong>g sources are available<br />
to improve fi sheries.<br />
4.3.3 Water Sports<br />
Urban rivers (if not so polluted as to be unpleasant) have<br />
long been used for canoe<strong>in</strong>g and fl oat<strong>in</strong>g. On summer
weekends, <strong>the</strong> Chattahoochee <strong>River</strong> near Atlanta, Georgia,<br />
is packed with young people fl oat<strong>in</strong>g downstream on tire<br />
<strong>in</strong>ner tubes or rafts. Increas<strong>in</strong>gly, more active forms of<br />
kayak<strong>in</strong>g and canoe<strong>in</strong>g are be<strong>in</strong>g designed for <strong>in</strong> urban<br />
river restoration projects. For example, <strong>the</strong> steeper, upper<br />
reaches of <strong>the</strong> restored Boulder Creek <strong>in</strong> Boulder, Colorado,<br />
have been designed as a kayak course, and kayakers<br />
and canoers commonly cont<strong>in</strong>ue downstream through <strong>the</strong><br />
town.<br />
4.4 SPONTANEOUS USES OF<br />
URBAN WATERWAYS<br />
Although recreation is commonly cited as a goal of stream<br />
restoration projects, it is often treated perfunctorily compared<br />
to o<strong>the</strong>r goals such as fl ood control and habitat,<br />
except where its value can be expressed <strong>in</strong> monetary<br />
terms. In <strong>the</strong> context of cost and benefi t analysis, emphasis<br />
on recreation necessarily narrows down to <strong>the</strong> licensed,<br />
quantifi able activities such as fi sh<strong>in</strong>g and boat<strong>in</strong>g (NRC,<br />
1992). In contrast to such vacation-orientated uses, <strong>the</strong>re<br />
is a suite of more <strong>in</strong>tuitive and unplanned activities, hereby<br />
named ‘spontaneous uses,’ that <strong>in</strong>volve direct and active<br />
<strong>in</strong>teraction with <strong>the</strong> landscape, such as skipp<strong>in</strong>g rocks,<br />
catch<strong>in</strong>g frogs, collect<strong>in</strong>g nuts and swimm<strong>in</strong>g. When<br />
human uses are considered at all <strong>in</strong> urban stream restoration<br />
projects, <strong>the</strong> focus is typically on passive uses, such<br />
as trail walk<strong>in</strong>g and social ga<strong>the</strong>r<strong>in</strong>g. The orientation is<br />
also often towards adults uses only, whereas children may<br />
have very different (and strongly felt) attitudes towards<br />
stream environments (Tunstall et al., 2004; Yang, 2004).<br />
However, a grow<strong>in</strong>g literature suggests that <strong>the</strong> more<br />
<strong>in</strong>teractive activities are crucial to <strong>the</strong> form<strong>in</strong>g of environmental<br />
awareness (Chawla, 1988; Harvey, 1989; Orr,<br />
1992) and place attachment (Owens, 1988; Hester et al.,<br />
1988; Cooper-Marcus, 1992), and are benefi cial for<br />
healthy human development (Nicholson, 1971; Kaplan,<br />
1977; Cobb, 1977; Hart, 1979; Moore, 1986). Whe<strong>the</strong>r<br />
a restored channel encourages spontaneous use or not<br />
is a source of uncerta<strong>in</strong>ty to <strong>the</strong> ‘social’ success of a<br />
restoration project.<br />
To understand <strong>the</strong> specifi c habitat characteristics that<br />
permit and encourage spontaneous uses, Yang (2004)<br />
reviewed <strong>the</strong> literature to identify probable habitat characteristics<br />
encourag<strong>in</strong>g such uses and <strong>the</strong>n undertook systematic<br />
fi eld observations, especially of children, and<br />
<strong>in</strong>terviews of children and adults <strong>in</strong> fi eld areas <strong>in</strong> California<br />
and Japan. Although many of <strong>the</strong>se <strong>in</strong>teractions were<br />
engaged ma<strong>in</strong>ly by children, <strong>the</strong>y were not enjoyed by<br />
children exclusively. Adults accompany<strong>in</strong>g children, or<br />
even among a group of adults, appeared to fully enjoy<br />
such uses. From this research, we summarized <strong>the</strong> most<br />
Plann<strong>in</strong>g <strong>River</strong> <strong>Restoration</strong> Projects: Social and Cultural Dimensions 49<br />
common types of spontaneous <strong>in</strong>teraction and <strong>the</strong>ir habitat<br />
requirements.<br />
4.4.1 Quiet and Secluded Use<br />
Users who appreciate <strong>the</strong> stream environment <strong>in</strong> a transcendent<br />
way, go to <strong>the</strong> stream for a temporary escape,<br />
enjoy <strong>in</strong>timate relationships with signifi cant o<strong>the</strong>rs and<br />
those who pursue quiet read<strong>in</strong>g, th<strong>in</strong>k<strong>in</strong>g etc., are commonly<br />
attached to a specifi c base-po<strong>in</strong>t. Their territory<br />
may seem small, but <strong>the</strong> quality demands are high and<br />
specifi c. S<strong>in</strong>ce such users can stay for hours, a certa<strong>in</strong><br />
comfort level (dry seat<strong>in</strong>g, foothold and shade) is normally<br />
required. A rock, tree root, log, or a soft grassy spot<br />
by water are particularly appeal<strong>in</strong>g. Yet more than anyth<strong>in</strong>g<br />
else <strong>the</strong>y need privacy, or visual/auditory seclusion<br />
from supervision or o<strong>the</strong>r users. Lewis (1995) highlighted<br />
<strong>the</strong> value of San Leandro Creek, California, as a secret<br />
hid<strong>in</strong>g place and unsupervised play area, with many ‘fi rsttime’<br />
events of local youth. All his <strong>in</strong>terviewees who<br />
played <strong>the</strong>re appreciated this quality of nonsupervision.<br />
For this reason, <strong>the</strong>y preferred detoured or <strong>in</strong>conspicuous<br />
access and a back screen.<br />
The view toward dense foliage, open fi eld or expression<br />
of water surface, <strong>the</strong> sound of trickl<strong>in</strong>g water, <strong>the</strong> appearance<br />
of wildlife and easy access to water all tremendously<br />
enhance <strong>the</strong> value of quiet and secluded base-po<strong>in</strong>ts, as<br />
users cite <strong>the</strong>se elements as bestow<strong>in</strong>g <strong>the</strong> heal<strong>in</strong>g power<br />
of nature. Both adults and children have been found to<br />
seek out space for quiet and secluded use (Yang, 2004).<br />
Quiet and secluded base-po<strong>in</strong>ts are easily lost, not only<br />
because privacy is often lost with <strong>in</strong>creased urbanization,<br />
but also because planners and designers usually don’t<br />
design for such spots, operat<strong>in</strong>g <strong>in</strong>stead on a design model<br />
of a cheerful park for adult socializ<strong>in</strong>g and playgrounds<br />
where all children play toge<strong>the</strong>r.<br />
4.4.2 Adventures<br />
Adventure connects known to unknown parts <strong>in</strong> <strong>the</strong> landscape,<br />
expand<strong>in</strong>g cognitive and physical territory. Adventurers<br />
walk, bike, swim, leap, climb, creep and cross to<br />
‘conquer’ a new piece of landscape. A system of basepo<strong>in</strong>ts<br />
connected by diverse, usually three-dimensional<br />
paths, plays an important role <strong>in</strong> <strong>the</strong> process of expand<strong>in</strong>g<br />
territory. Dirt paths apparently possess special values to<br />
adventurers. On Marsh Creek, California, some adults<br />
favor dirt paths for aes<strong>the</strong>tic reasons, but children preferred<br />
dirt paths for <strong>the</strong> practical reasons that <strong>the</strong>y are<br />
usually avoided by adult bikers and runners, who are often<br />
impatient with children <strong>in</strong> <strong>the</strong>ir way, and <strong>the</strong> dirt path<br />
provides more <strong>in</strong>teractive features (Yang, 2004). On dirt
50 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
paths children can jump on mounds, leap <strong>in</strong>to muddy<br />
puddles, bend under low branches, or crouch to stare <strong>in</strong>to<br />
a gopher hole. The mounds, puddles, low branches or<br />
gopher holes would all be considered undesirable and<br />
elim<strong>in</strong>ated on a paved trail, but on a dirt path <strong>the</strong>y provide<br />
tempt<strong>in</strong>g <strong>in</strong>vitations for sensorial experiences.<br />
Adventurers are particularly keen to fi nd good stream<br />
cross<strong>in</strong>g po<strong>in</strong>ts. In smaller streams users search shallow<br />
and narrow spots with stepp<strong>in</strong>g rocks to set foot on or from<br />
which to build a bridge. In large rivers swimm<strong>in</strong>g across<br />
is a common game. If a rope and tree are available and <strong>the</strong><br />
stream is narrow enough, <strong>the</strong>y sw<strong>in</strong>g across <strong>the</strong> channel.<br />
Similarly, children <strong>in</strong> Turkey Brook, London, asked for<br />
ropes to sw<strong>in</strong>g on and logs to slide down (Tunstall et al.,<br />
2004). When <strong>the</strong> slope is right and <strong>the</strong> channel not too<br />
wide, bikers or skateboarders fl y across on wheels.<br />
Metal culverts are especially attractive to adventurers:<br />
it’s easy to make loud, eerie echoes <strong>in</strong> <strong>the</strong>m, <strong>the</strong>y are secret<br />
hide-outs, <strong>the</strong>y offer <strong>the</strong> allure of a connection to somewhere<br />
else and <strong>the</strong>y are perceived somehow off-limits.<br />
4.4.3 Wildlife Contact<br />
Wildlife contact can be by simple observation or active<br />
catch<strong>in</strong>g, two dist<strong>in</strong>ct modes of <strong>in</strong>teraction. Observers are<br />
usually <strong>in</strong>terested <strong>in</strong> all life forms <strong>the</strong>y see, from little bugs<br />
to big animals such as otters and raccoons. They <strong>in</strong>teract<br />
with <strong>the</strong> stream with a highly <strong>in</strong>tensive but un<strong>in</strong>trusive<br />
way. Wildlife sight<strong>in</strong>gs often occur <strong>in</strong> unexpected, uncalculated<br />
moments, produc<strong>in</strong>g a ‘wow’ experience.<br />
Catchers are more physically active and focus on certa<strong>in</strong><br />
target species, which are small enough to catch. Catch<strong>in</strong>g<br />
wildlife along creeks has traditionally provided subsistence,<br />
but <strong>in</strong> urban areas <strong>in</strong> developed nations today, catch<strong>in</strong>g<br />
is usually based upon affi nity toward <strong>the</strong> target and a<br />
sense of achievement. Fish, frogs, tadpoles, shrimps, crawdads,<br />
crabs and <strong>in</strong>sects are fasc<strong>in</strong>at<strong>in</strong>g creatures for users<br />
to match wits with. The habitats of catchers are as diverse<br />
as those of <strong>the</strong>ir target species and <strong>the</strong>ir spots correspond<br />
directly to those of <strong>the</strong>ir quarry. Children who actively<br />
catch wildlife tend to be agile and will<strong>in</strong>g to access diffi -<br />
cult sites, get wet or scratched and <strong>in</strong> general are highly<br />
adaptive to <strong>the</strong>ir environments (Figure 4.5). In Marsh<br />
Creek, crawdad hunters were often observed thriv<strong>in</strong>g at<br />
<strong>the</strong> least ‘user-friendly’ spots, such as among rugged<br />
riprap under road bridges or by grassy, muddy shores.<br />
Methods of catch<strong>in</strong>g are numerous, even for <strong>the</strong> same<br />
species. They range from <strong>the</strong> bare hand to highly elaborated<br />
means and tools. Catchers <strong>in</strong> various regions <strong>in</strong><br />
Japan and California often stored captured fi sh or crawdads<br />
temporarily <strong>in</strong> a conta<strong>in</strong>er or a little pond enclosed<br />
with sand or rocks. Most of <strong>the</strong> trapped creatures were set<br />
Figure 4.5 Crawdad from Marsh Creek caught by a child and<br />
draw<strong>in</strong>g of Marsh Creek wildlife by 4th grade child (both from<br />
Yang, 2004.)<br />
free after a short time, but some captures would become<br />
pets to be enjoyed at home until <strong>the</strong>y expired. Many catchers<br />
have learned from experiences which animals ‘work<br />
better’ as pets (Yang, 2004).<br />
Profi cient catchers and observers are often knowledgeable;<br />
<strong>the</strong>y can usually identify many species and know<br />
when and where to fi nd <strong>the</strong>m. Observers and catchers have<br />
similar habitat requirements: <strong>the</strong> environment needs to<br />
support a suffi ciently high density of wildlife and a mean<strong>in</strong>gful<br />
human/wildlife <strong>in</strong>terface. Though <strong>the</strong> former is a<br />
widely claimed goal <strong>in</strong> restoration and greenway projects,<br />
<strong>the</strong> latter is usually discouraged. For spontaneous users, a<br />
mean<strong>in</strong>gful wildlife/human <strong>in</strong>terface provides plenty of<br />
chances for close-up observation and hands-on catch<strong>in</strong>g,<br />
without <strong>the</strong> need of specialized equipment beyond that<br />
which can be made at home or obta<strong>in</strong>ed from a grocery<br />
store. Examples of such <strong>in</strong>terfaces are water edges framed<br />
by vegetation or porous structures where different species<br />
hide, or shallow water reaches adjacent to gravel bars<br />
where fry of amphibians and fi sh hatch. It is important that<br />
water edges designed to susta<strong>in</strong> a dense wildlife popula-
tion also rema<strong>in</strong> accessible to users, except <strong>in</strong> cases (rare<br />
<strong>in</strong> urban areas) where protected species need to be isolated<br />
from human harassment. When physical access is not feasible,<br />
visual access can be provided from <strong>the</strong> bank, bridges<br />
etc.<br />
4.4.4 Manipulat<strong>in</strong>g <strong>the</strong> Environment<br />
The value of creeks and rivers for spontaneous use depends<br />
largely on <strong>the</strong>ir provision of loose parts – elements that<br />
can be easily manipulated <strong>in</strong> <strong>the</strong> environment (Nicholson,<br />
1971). At least three categories of common uses rely on<br />
contact with rocks, plants, junk and o<strong>the</strong>r k<strong>in</strong>ds of loose<br />
parts <strong>in</strong> stream environments: collect<strong>in</strong>g, build<strong>in</strong>g and<br />
clever craft:<br />
• Collect<strong>in</strong>g allows one to discern treasures from <strong>the</strong> basically<br />
chaotic stream environment. Once purposely rummaged<br />
or fortuitously encountered, stones and o<strong>the</strong>r<br />
elements from <strong>the</strong> bed, plant parts and junk recovered<br />
from banks may be used <strong>in</strong> drama play, build<strong>in</strong>g, or <strong>in</strong><br />
displays <strong>in</strong> <strong>the</strong> collector’s yard or room. Gravel bars are<br />
prized sources of stones for collect<strong>in</strong>g and clay banks<br />
provide material for handcraft, mud ball fi ghts and gray<br />
make-up.<br />
• Build<strong>in</strong>g projects, whe<strong>the</strong>r big (e.g. tree houses, huts,<br />
bases, dams, bridges, ponds) or small (arrang<strong>in</strong>g rocks<br />
and sticks) are rooted <strong>in</strong> an <strong>in</strong>nate attempt to create an<br />
impact on <strong>the</strong> landscape (Figure 4.6). Through build<strong>in</strong>g,<br />
users claim <strong>the</strong>ir ownership and adapt <strong>the</strong> stream to<br />
<strong>the</strong>mselves. The result of spontaneous build<strong>in</strong>g usually<br />
is not durable enough to survive fl oods and o<strong>the</strong>r natural<br />
processes. Build<strong>in</strong>g may have practical purposes, but <strong>the</strong><br />
process is all-important: many children build, destroy<br />
and rebuild.<br />
Figure 4.6 12-year-old child’s stove <strong>in</strong> drama house by Marsh<br />
Creek (from Yang, 2004.)<br />
Plann<strong>in</strong>g <strong>River</strong> <strong>Restoration</strong> Projects: Social and Cultural Dimensions 51<br />
• Clever crafts, <strong>the</strong> skillful manipulation of materials<br />
found <strong>in</strong> stream environments (Yang, 2004), is usually<br />
quite precise <strong>in</strong> terms of materials and surround<strong>in</strong>gs.<br />
For example, to skip a rock (a trans-culturally popular<br />
trick), one needs a gravel bar conta<strong>in</strong><strong>in</strong>g platy stones<br />
with <strong>in</strong>termediate axes usually between about 30 and<br />
70 mm, and a fl at pool allow<strong>in</strong>g satisfactory skips. Along<br />
Sonoma Creek <strong>in</strong> California, adults applied red algae to<br />
sk<strong>in</strong> rash and children made fl utes with deer grass.<br />
Along Kure <strong>River</strong> <strong>in</strong> Japan, smashed mugwort was used<br />
to heal scratches and defog goggles, while foxtail stalks<br />
were made <strong>in</strong>to knots to trap frogs and shrimp.<br />
4.4.5 Wad<strong>in</strong>g and Paddl<strong>in</strong>g<br />
Small children and o<strong>the</strong>r users who don’t want to get very<br />
wet will wade <strong>in</strong> waters shallower than 0.5 m, with currents<br />
20 cm/s or less, such as shallow marg<strong>in</strong>s or backwaters.<br />
The range of paddl<strong>in</strong>g by children is usually only a<br />
few meters from <strong>the</strong> water edge and <strong>the</strong> dry spot. In large<br />
streams, some h<strong>in</strong>ts of boundary around a smaller space<br />
(e.g. a cover or re-entrant <strong>in</strong> <strong>the</strong> bank or offshore bar) are<br />
needed to overcome <strong>the</strong> uneas<strong>in</strong>ess <strong>in</strong>duced by an unlimited<br />
expanse of water. Paddlers prefer gently slop<strong>in</strong>g<br />
access to water (ra<strong>the</strong>r than grassy or upright banks) and<br />
sandy or clay bottoms (which provide comfortable footholds)<br />
(Yang, 2004).<br />
4.4.6 Swimm<strong>in</strong>g, Flush<strong>in</strong>g and Div<strong>in</strong>g<br />
Swimm<strong>in</strong>g occurs mostly <strong>in</strong> pool reaches more than 0.5 m<br />
deep, with velocities under 0.5 m/s and with gentle and<br />
gradual water edges at bars or ‘ledge’ banks protected<br />
by tree roots as entry po<strong>in</strong>ts. In large or swift rivers,<br />
swimmers also require ‘stopover bases’ (island bars,<br />
bridge piers etc.) at which to rest. Warm surfaces such as<br />
big rocks, pebble beach, concrete blocks, asphalt roads<br />
etc. are valuable dry spots. Flush<strong>in</strong>g makes clever use of<br />
locally concentrated fl ow (>0.5 m/s) and variations <strong>in</strong> bed<br />
form. Most commonly, fl ush<strong>in</strong>g is done <strong>in</strong> riffl es: <strong>the</strong><br />
fl usher would start at <strong>the</strong> end of <strong>the</strong> pool where <strong>the</strong> speed<br />
starts to pick up, allow<strong>in</strong>g his body to be carried by <strong>the</strong><br />
accelerat<strong>in</strong>g current downstream to be caught at <strong>the</strong> crest<br />
of riffl e (if shallow) or carried through <strong>the</strong> riffl e (if deeper)<br />
to <strong>the</strong> next pool. Hard structures <strong>in</strong> <strong>the</strong> streams such as<br />
bridge piers can also form concentrated currents for<br />
fl ush<strong>in</strong>g.<br />
Div<strong>in</strong>g <strong>in</strong> larger streams and rivers with deep pools<br />
is popular on hot days, offer<strong>in</strong>g <strong>the</strong> thrill of a free fall and<br />
<strong>the</strong> sudden imp<strong>in</strong>gement of cool water on <strong>the</strong> body. The
52 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
div<strong>in</strong>g height is limited by pool depth and <strong>the</strong> diver’s skill<br />
and nerve. We observed children div<strong>in</strong>g from a 15-m high<br />
treetop <strong>in</strong> <strong>the</strong> Kagami <strong>River</strong>, Japan (Figure 4.7). In stream<br />
environments, a pool deeper than two meters is rare and<br />
considered enough for moderate-height div<strong>in</strong>g. A good<br />
div<strong>in</strong>g spot also has a land<strong>in</strong>g spot with a gentler water<br />
edge, a path to connect land<strong>in</strong>g and launch spots <strong>in</strong>to a loop<br />
and, ideally, choices for different skill levels. Rock outcrops<br />
(adjacent to deep pools because <strong>the</strong>y <strong>in</strong>duce scour at<br />
high fl ow) provide steady footholds for <strong>the</strong> launch po<strong>in</strong>t<br />
and outcrops with complex form and multiple take-off<br />
po<strong>in</strong>ts for different dive heights allow divers to practice<br />
and build <strong>the</strong>ir courage and skills gradually. Div<strong>in</strong>g from<br />
<strong>the</strong> trees, one experiences <strong>the</strong> thrill of shak<strong>in</strong>g footholds;<br />
div<strong>in</strong>g from a rope sw<strong>in</strong>g, one challenges <strong>the</strong> arm strength,<br />
body balance and <strong>the</strong> tim<strong>in</strong>g to let go; div<strong>in</strong>g at concrete<br />
levees, one needs to leap forward to avoid <strong>the</strong> concrete<br />
foundation jett<strong>in</strong>g out beneath <strong>the</strong> mean water level (Yang,<br />
2004). For ‘thrill<strong>in</strong>g’ water contacts such as fl ush<strong>in</strong>g and<br />
div<strong>in</strong>g, routes that connect back to <strong>the</strong> set-<strong>in</strong> po<strong>in</strong>ts are<br />
<strong>in</strong>dispensable to support <strong>the</strong>ir repetitive characteristics.<br />
Figure 4.7 Children div<strong>in</strong>g <strong>in</strong>to <strong>the</strong> Kagami <strong>River</strong>, Japan (from<br />
Yang, 2004.)<br />
4.5 CONFLICTS AMONG MULTIPLE GOALS<br />
AND OBJECTIVES<br />
Everybody wants more nature but <strong>the</strong>re has been persistent<br />
confusion about <strong>the</strong> mean<strong>in</strong>g of ‘restoration’. The controversy<br />
over <strong>the</strong> term refl ects disagreements over goals, even<br />
when considered only with<strong>in</strong> <strong>the</strong> physical science realm.<br />
As causative agents, humans constantly change and control<br />
nature to ‘help’ it, lead<strong>in</strong>g to fundamental questions about<br />
goals. Given that nature is <strong>in</strong> constant fl ux and <strong>the</strong>re is no<br />
s<strong>in</strong>gle correct condition (Hull and Robertson, 2000), <strong>the</strong><br />
choice of desired end state for restoration will perforce<br />
<strong>in</strong>volve societal priorities. Because river restoration projects<br />
are now commonly undertaken with <strong>the</strong> <strong>in</strong>volvement<br />
of multiple professionals and stakeholders, and because<br />
all <strong>the</strong> goals and objectives are essentially value-driven,<br />
three types of confl icts often arise <strong>in</strong> project implementation,<br />
creat<strong>in</strong>g uncerta<strong>in</strong>ties for <strong>the</strong> course of river restoration<br />
plann<strong>in</strong>g and implementation.<br />
4.5.1 Confl icts among Professionals<br />
Eng<strong>in</strong>eers, fl uvial geomorphologists, ecologists and landscape<br />
architects are tra<strong>in</strong>ed to see <strong>the</strong> stream differently<br />
(Figure 4.8). In <strong>the</strong> past, <strong>the</strong>y have all shaped or reshaped<br />
streams with <strong>the</strong>ir particular value systems and discipl<strong>in</strong>ary<br />
tools. Eng<strong>in</strong>eer<strong>in</strong>g has been <strong>the</strong> s<strong>in</strong>gle most powerful<br />
profession <strong>in</strong> past stream transformation, alter<strong>in</strong>g rivers<br />
for fl ood control, water supply and navigation. Hydraulic<br />
eng<strong>in</strong>eers model fl ows under conditions where <strong>the</strong> variables<br />
are controlled, usually approximat<strong>in</strong>g channel shapes<br />
as simpler geometric entities. Deviations from clear water<br />
and Euclidean channel shapes are treated with adjustments<br />
<strong>in</strong> formulas. Fluvial geomorphologists tend to approach<br />
problems at larger scales and over longer periods. The<br />
eng<strong>in</strong>eer or manager may pose a question such as, ‘What<br />
k<strong>in</strong>d of bank protection should we use along this reach<br />
of stream?’ The fl uvial geomorphologist will tend to ask<br />
why <strong>the</strong> bank is erod<strong>in</strong>g <strong>in</strong> <strong>the</strong> fi rst place, whe<strong>the</strong>r it<br />
is simply part of <strong>the</strong> natural channel migration process or<br />
a result of changes <strong>in</strong> <strong>the</strong> catchment upstream. Especially<br />
<strong>in</strong> <strong>the</strong> latter case, it is likely that plac<strong>in</strong>g bank protection<br />
will not ‘solve’ <strong>the</strong> problem but will <strong>in</strong>duce problems<br />
elsewhere.<br />
Ecologists tend to view streams as organic compounds<br />
of habitats. They perceive fi ne details of leaf litter and its<br />
decompostion, moss on boulders, food cha<strong>in</strong>s, and cycles<br />
of nitrogen, carbon etc. Traditional biologists may see<br />
unspoiled natural process as <strong>the</strong> reference condition<br />
aga<strong>in</strong>st which to measure degradation and a return to that<br />
condition as a restoration goal. Human activities are<br />
viewed as ‘impact.’ Add<strong>in</strong>g a spatial structural perspective,
<strong>the</strong>y view streams as ‘corridors,’ a crucial element <strong>in</strong> landscape<br />
to allow movement of species and <strong>the</strong>refore to ma<strong>in</strong>ta<strong>in</strong><br />
biodiversity and long term genetic diversity. These<br />
professionals have long realized that to ma<strong>in</strong>ta<strong>in</strong> a healthy<br />
ecosystem, a river needs fl oods (Poff et al., 1997; Junk<br />
et al., 1989). Developers and fl ood control agencies often<br />
resist los<strong>in</strong>g developable lands to fl ood <strong>in</strong>undation and,<br />
through <strong>the</strong>ir <strong>in</strong>fl uence on <strong>the</strong> political process, typically<br />
succeed <strong>in</strong> implement<strong>in</strong>g fl ood control measures such as<br />
dams or levees that permit <strong>the</strong>m to build on fl oodpla<strong>in</strong>s.<br />
Similarly, natural channel migration is an important<br />
process to create diverse habitats, but riverside developments<br />
are threatened by bank erosion, result<strong>in</strong>g <strong>in</strong> pressure<br />
to stabilize <strong>the</strong> river bank with hard structures. As a result,<br />
although ecologists and environmental scientists have<br />
Plann<strong>in</strong>g <strong>River</strong> <strong>Restoration</strong> Projects: Social and Cultural Dimensions 53<br />
Figure 4.8 Different perceptions and attitudes towards rivers by eng<strong>in</strong>eers, biologists etc. (Source: Hough (1990), adapted from<br />
draw<strong>in</strong>gs orig<strong>in</strong>ally prepared by Newbury (unpublished data)).<br />
been <strong>in</strong>stitutionalized <strong>in</strong>to <strong>the</strong> plann<strong>in</strong>g process s<strong>in</strong>ce <strong>the</strong><br />
1960s, <strong>the</strong>y often rema<strong>in</strong> ‘second-class citizens’ <strong>in</strong> affect<strong>in</strong>g<br />
<strong>the</strong> design of urban stream channels (Riley, 1998).<br />
A traditional tenet of landscape architects is to view<br />
landscape <strong>in</strong> abstract, formal, aes<strong>the</strong>tic terms: forms,<br />
l<strong>in</strong>es, colors, textures and <strong>the</strong>ir <strong>in</strong>ter-relationships (Daniel<br />
and V<strong>in</strong><strong>in</strong>g, 1983). Although visual aes<strong>the</strong>tics are usually<br />
<strong>the</strong> paramount ‘public’ goal, designers also emphasize <strong>the</strong><br />
cultural and historical signifi cance of urban streams, as<br />
well as <strong>the</strong> user’s experiences. Some landscape architects<br />
are well tra<strong>in</strong>ed ecologically and effectively <strong>in</strong>tegrate ecological<br />
considerations <strong>in</strong> <strong>the</strong>ir designs; some are <strong>in</strong>volved<br />
<strong>in</strong> successful efforts to redevelop urban waterfronts to<br />
revitalize downtown economies, attract tourists and<br />
provide recreation opportunities for urban residents (Otto
54 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
et al., 2004). These projects provide chances to enhance<br />
physical and visual connection with streams by plac<strong>in</strong>g<br />
walkways along <strong>the</strong>m, or by promot<strong>in</strong>g vistas and fac<strong>in</strong>g<br />
commercial fronts to <strong>the</strong> streams (Jones and Battaglia,<br />
1989). Landscape architects also transform fl oodpla<strong>in</strong>s to<br />
open spaces to accommodate civic activities such as<br />
exhibits, concerts, fairs or sports. The design of <strong>the</strong>se open<br />
spaces, however, often takes its model from pastoral parks<br />
or architectural plazas and while <strong>the</strong>y may provide effective<br />
urban spaces, <strong>the</strong> purported uses and <strong>the</strong> design<br />
schemes are often <strong>in</strong> confl ict with <strong>the</strong> chaotic character of<br />
fl oods and organic quality of riparian habitats.<br />
4.5.2 Confl icts Among Stakeholder Groups<br />
Humans use rivers for many different purposes, so it<br />
should come as no surprise that human expectations and<br />
<strong>the</strong> demands of rivers often confl ict. The same confl icts<br />
that manifest <strong>the</strong>mselves <strong>in</strong> management of exist<strong>in</strong>g river<br />
channels tend to emerge when river restoration projects<br />
are conceived and scoped.<br />
Some of <strong>the</strong> best documented such confl icts are <strong>in</strong> <strong>the</strong><br />
Colorado <strong>River</strong> below Glen Canyon Dam, where cold<br />
water releases have allowed a ra<strong>in</strong>bow trout (Oncorhynchus<br />
mykiss) fi shery, highly valued by anglers, to become<br />
established. The ra<strong>in</strong>bow trout are exotic to <strong>the</strong> river and<br />
would not have survived <strong>the</strong> high temperatures and high<br />
suspended sediment loads characteristic of <strong>the</strong> pre-dam<br />
river. The post-dam river is now unfavorable to native fi sh,<br />
such as humpback chub (Cila cypha) and razorback sucker<br />
(Xyrauchen texanus). The native fi sh are ugly and undesirable<br />
as sport fi sh, but <strong>the</strong>y are native to <strong>the</strong> river and <strong>the</strong>ir<br />
numbers have dw<strong>in</strong>dled such that several species are now<br />
listed as threatened or endangered (Schmidt et al., 1998).<br />
Where <strong>the</strong> exotic trout and native fi sh coexist, <strong>the</strong> trout<br />
may prey on <strong>the</strong> natives. Proposed actions to improve<br />
conditions for <strong>the</strong> native species have encountered resistance<br />
from trout fi sh<strong>in</strong>g groups. It is well established that<br />
<strong>the</strong> reduction <strong>in</strong> high fl ows effected by Glen Canyon Dam<br />
has had numerous ecological effects on <strong>the</strong> reach downstream<br />
and thus deliberate high fl ow releases are planned<br />
<strong>in</strong> efforts to restore <strong>the</strong> reach. The fi rst such release, a<br />
much-publicized fl ow of 1300 m 3 s −1 <strong>in</strong> 1996, was only<br />
about one-third of <strong>the</strong> average annual pre-dam high fl ow.<br />
The fl ow was limited to avoid <strong>in</strong>undat<strong>in</strong>g a rare snail that<br />
had extended its range down <strong>the</strong> canyon walls dur<strong>in</strong>g <strong>the</strong><br />
post-dam period (Marzolf et al., 1998). Thus, restoration<br />
of a dynamic fl ow regime (with attendant benefi ts for <strong>the</strong><br />
river ecosystem) was perceived to confl ict with protection<br />
of <strong>the</strong> rare snail. A similar confl ict among user groups is<br />
on <strong>the</strong> North Fork Fea<strong>the</strong>r <strong>River</strong>, California, where high<br />
fl ows released periodically to provide fl ows for rafters<br />
have scoured benthic macro<strong>in</strong>vertebrates, wash<strong>in</strong>g <strong>the</strong>se<br />
and o<strong>the</strong>r organisms downstream (Garcia and Associates,<br />
2005).<br />
4.5.3 Confl icts Between Professionals and<br />
Local Groups<br />
Perhaps <strong>the</strong> best documented example of a professional–<br />
local group confl ict <strong>in</strong>volves terrestrial habitat restoration,<br />
<strong>the</strong> Chicago prairie restoration controversy. Efforts to<br />
restore 7000 acres of <strong>the</strong> DuPage County forest reserves<br />
<strong>in</strong> <strong>the</strong> Chicago metropolitan area back to <strong>the</strong> historical oak<br />
savanna and tallgrass prairie condition were attacked by<br />
local groups and residents who opposed remov<strong>in</strong>g trees<br />
and brush. Ryan (2000) concluded that <strong>the</strong> discrepancy<br />
between restoration planners and neighborhood users<br />
stemmed from differences <strong>in</strong> attachment. While both<br />
groups were attached to nature, <strong>the</strong>ir attachment can be<br />
diverse and contradictory. Scientists and volunteers are<br />
attached to a particular type of orig<strong>in</strong>al landscape, which<br />
is established through environmental criteria such as biodiversity<br />
and system <strong>in</strong>tegrity. Such attachment is not<br />
bound to a place – <strong>the</strong> same habitat image can be reproduced<br />
elsewhere and still be satisfactory. On <strong>the</strong> o<strong>the</strong>r<br />
hand, <strong>the</strong> attachment of local residents is <strong>in</strong>tertw<strong>in</strong>ed <strong>in</strong><br />
locale and context. Individual trees, albeit non-native, bear<br />
an identity <strong>in</strong> terms of furnish<strong>in</strong>g <strong>the</strong> spot for children to<br />
play tag or a seat for quietness or fram<strong>in</strong>g a magnifi cent<br />
view toward sundown. In o<strong>the</strong>r words, <strong>the</strong> attachment of<br />
local residents is composed of life memories.<br />
A different confl ict between professionals and a local<br />
group occurred <strong>in</strong> <strong>the</strong> nor<strong>the</strong>rn Sierra Nevada of California<br />
<strong>in</strong> <strong>the</strong> early 1990s. In plann<strong>in</strong>g a restoration project on<br />
Jamison Creek <strong>in</strong> Plumas-Eureka State Park (<strong>the</strong> Park), a<br />
local nonprofi t group active <strong>in</strong> implement<strong>in</strong>g stream restoration<br />
projects (but without expertise <strong>in</strong> fl uvial geomorphology)<br />
challenged <strong>the</strong> effective discharge analysis<br />
conducted by a university team, contend<strong>in</strong>g that <strong>the</strong> bankfull<br />
discharge was only about one-third that computed by<br />
<strong>the</strong> university team. Despite a thorough and well documented<br />
scientifi c report support<strong>in</strong>g <strong>the</strong> university team’s<br />
analysis, <strong>the</strong> Park rejected <strong>the</strong> analysis and sided with <strong>the</strong><br />
local nonprofi t group, stat<strong>in</strong>g that it preferred <strong>the</strong> smaller<br />
design discharge because ‘a smaller channel is better for<br />
fi sh habitat’. The Park also cited that fact that <strong>the</strong> local<br />
Coord<strong>in</strong>ated Resource Management Program group (composed<br />
of local agency staff, landowners and staff of <strong>the</strong><br />
nonprofi t group (none of whom possessed expertise <strong>in</strong><br />
fl uvial geomorphology) had voted <strong>in</strong> favor of <strong>the</strong> lower<br />
design discharge.<br />
The notion that one can arbitrarily choose a design discharge<br />
and build a stream channel to smaller dimensions
is a fasc<strong>in</strong>at<strong>in</strong>g one, but not one supported by geomorphic<br />
science. Likewise, <strong>the</strong> notion that scientifi c questions<br />
should be put to a vote by a group without expertise <strong>in</strong> <strong>the</strong><br />
fi eld raises questions about <strong>the</strong> role of science <strong>in</strong> such a<br />
restoration design process. Ultimately, <strong>the</strong> Park had <strong>the</strong><br />
authority and responsibility to design and construct <strong>the</strong><br />
channel as it saw fi t. The university team withdrew from<br />
<strong>the</strong> project and <strong>the</strong> local nonprofi t group proceeded to<br />
design and build a channel reconstruction <strong>in</strong> 1995. The<br />
project was damaged by <strong>the</strong> high fl ows of 1996, repaired,<br />
and <strong>the</strong>n completely washed out by high fl ows <strong>in</strong> 1997. In<br />
2000, <strong>the</strong> Park sent out a call for proposals to reconstruct<br />
<strong>the</strong> channel once aga<strong>in</strong>.<br />
4.6 CASE STUDIES<br />
4.6.1 Baxter Creek, El Cerrito<br />
Baxter Creek dra<strong>in</strong>s an 11-km2 urban area of El Cerrito,<br />
California, debouch<strong>in</strong>g <strong>in</strong>to San Francisco Bay at Richmond<br />
(Figure 4.9). In 1997, <strong>the</strong> City of El Cerrito replaced<br />
a 70-m reach of fail<strong>in</strong>g culvert (<strong>in</strong> a small neighborhood<br />
park) with an open channel (Figure 4.10). The open<br />
channel was stabilized with a series of boulder weirs<br />
Pacific<br />
Ocean<br />
San<br />
Francisco<br />
San<br />
Pablo<br />
Bay<br />
San<br />
Francisco<br />
Bay<br />
Plann<strong>in</strong>g <strong>River</strong> <strong>Restoration</strong> Projects: Social and Cultural Dimensions 55<br />
Suisun Bay<br />
Richmond<br />
a b<br />
Brentwood<br />
Berkeley<br />
San Jose<br />
Figure 4.9 Location map, Baxter (a) and Marsh Creeks (b),<br />
Contra Costa County.<br />
(which dissipated energy from <strong>the</strong> 10% gradient) and <strong>the</strong><br />
banks were planted with willow (salix spp). Post-project<br />
appraisals <strong>in</strong> 1999 and 2004 (Purcell et al., 2002; Purcell,<br />
2004) showed that <strong>the</strong> biotic condition of <strong>the</strong> restored<br />
reach was measurably better than an unrestored control<br />
section upstream and that <strong>the</strong> biotic condition did not<br />
improve fur<strong>the</strong>r between 1999 (two years post-project)<br />
and 2004 (seven years post-project), <strong>in</strong>dicat<strong>in</strong>g <strong>the</strong> stream<br />
may have reached its biotic potential with<strong>in</strong> two years.<br />
Purcell et al. (2002) conducted an attitud<strong>in</strong>al survey of<br />
<strong>the</strong> residents with<strong>in</strong> one block of <strong>the</strong> daylighted section of<br />
Baxter Creek. Of <strong>the</strong> 45 responses received, most were<br />
positive overall about <strong>the</strong> restoration, but many expressed<br />
concerns that <strong>the</strong> willow trees, some of which had grown<br />
to over 6 m <strong>in</strong> height, blocked <strong>the</strong> view across <strong>the</strong> park<br />
and potentially provided hid<strong>in</strong>g places for burglars. In a<br />
repeat survey of <strong>the</strong> neighborhood <strong>in</strong> 2004 (n = 45),<br />
Purcell found that about half of those who had moved to<br />
<strong>the</strong> neighborhood after <strong>the</strong> completion of <strong>the</strong> restoration<br />
did not realize <strong>the</strong> creek had formerly been <strong>in</strong> an underground<br />
culvert. Nearly all respondents reported <strong>the</strong>y<br />
enjoyed liv<strong>in</strong>g near <strong>the</strong> creek, many cit<strong>in</strong>g <strong>the</strong> sounds of<br />
<strong>the</strong> water, aes<strong>the</strong>tics, or accessibility for children or dogs.<br />
69% perceived an improvement s<strong>in</strong>ce <strong>the</strong> restoration was<br />
completed; 31% said conditions had worsened. Overall,<br />
<strong>the</strong> project was successful <strong>in</strong> creat<strong>in</strong>g a vibrant stream<br />
corridor where formerly <strong>the</strong>re had been only a relatively<br />
sterile strip of lawn. The success of <strong>the</strong> project led to <strong>the</strong><br />
formation of <strong>the</strong> ‘Friends of Baxter Creek’, a group which<br />
subsequently supported two o<strong>the</strong>r restoration projects <strong>in</strong><br />
downstream reaches of Baxter Creek (Lisa Owens-Viani,<br />
personal communication, 2006).<br />
Figure 4.10 Baxter Creek <strong>in</strong> Po<strong>in</strong>sett Park, El Cerrito, California.<br />
Photo by Alison Purcell, April 2007, about 10 years after<br />
construction. Note height of willows, some exceed<strong>in</strong>g 6 m.
56 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
There was some negative reaction to ‘overgrown’<br />
vegetation (Purcell, 2004), which is not unusual <strong>in</strong> urban<br />
stream restorations <strong>in</strong> nor<strong>the</strong>rn California. Blackberry<br />
Creek <strong>in</strong> <strong>the</strong> Thousand Oaks School <strong>in</strong> Berkeley was<br />
removed from an underground culvert and replaced with<br />
an open channel <strong>in</strong> 1995. Some residents had negative<br />
reactions to <strong>the</strong> density of willow growth, which precluded<br />
access or see<strong>in</strong>g <strong>in</strong>to <strong>the</strong> stream channel. In plann<strong>in</strong>g<br />
<strong>the</strong> 1990 restoration of Cortland Creek <strong>in</strong> Oakland,<br />
<strong>the</strong> plans by creek activists to extensively plant willows<br />
(for habitat) met with resistance from local residents, who<br />
did not want to create a place where crim<strong>in</strong>als could hide<br />
(Walter Hood, personal communication, 1995). Though<br />
not as well documented as <strong>the</strong> Baxter Creek case, many<br />
o<strong>the</strong>r urban stream restoration projects have been marked<br />
by similar confl icts between plant<strong>in</strong>g willows to enhance<br />
habitat and <strong>the</strong> desire by residents and police to see <strong>in</strong>to<br />
<strong>the</strong> channel to discourage crim<strong>in</strong>als. This has been true<br />
especially <strong>in</strong> low-<strong>in</strong>come neighborhoods, where concern<br />
about crime may be greater.<br />
4.6.2 Marsh Creek, Brentwood, California<br />
Marsh Creek dra<strong>in</strong>s 332 km2 , with its upper bas<strong>in</strong> mostly<br />
woodland, rangeland and farmland, and its lower 15 km<br />
travers<strong>in</strong>g a broad alluvial fan, which now supports <strong>the</strong><br />
urban areas of Brentwood and Oakley, about 60 km nor<strong>the</strong>ast<br />
of San Francisco (Figure 4.9). As typical of Mediterranean-climate<br />
streams, runoff from <strong>the</strong> catchment was<br />
naturally <strong>in</strong>termittent <strong>in</strong> all but wet years. There is little<br />
record of <strong>the</strong> historical channel conditions <strong>in</strong> Marsh Creek<br />
<strong>in</strong> Brentwood, but historical maps from <strong>the</strong> late 1800s and<br />
early 1900s show multiple, s<strong>in</strong>uous channels and active<br />
channel migration (Rob<strong>in</strong>s and Ca<strong>in</strong>, 2002) (Figure 4.11).<br />
As agriculture expanded onto <strong>the</strong> fertile soils <strong>in</strong> <strong>the</strong> early–<br />
mid 20th century, and <strong>in</strong> response to fl ood<strong>in</strong>g of downtown<br />
Brentwood <strong>in</strong> <strong>the</strong> 1950s, <strong>the</strong> channel of Marsh Creek was<br />
straightened, <strong>the</strong> riparian corridor largely cleared and a<br />
fl ood-control reservoir constructed about 3 km upstream<br />
of Brentwood (Figure 4.12).<br />
Brentwood has grown rapidly, <strong>in</strong>creas<strong>in</strong>g <strong>in</strong> population<br />
from 7500 <strong>in</strong> 1990, to 23 000 <strong>in</strong> 2000, to 33 000 <strong>in</strong> 2003<br />
(Ca<strong>in</strong> et al., 2003). Many of <strong>the</strong>se residents commute (oneway<br />
travel times of over an hour) to jobs <strong>in</strong> <strong>the</strong> San Francisco<br />
Bay Region. As Brentwood has grown, <strong>in</strong>terest has<br />
grown <strong>in</strong> enhanc<strong>in</strong>g <strong>the</strong> creek corridor for human uses,<br />
remov<strong>in</strong>g barriers to salmonid migration and improv<strong>in</strong>g<br />
stormwater detention. A watershed study (Ca<strong>in</strong> et al., 2003)<br />
documented historical changes <strong>in</strong> physical and biological<br />
conditions, identify<strong>in</strong>g signifi cant effects of straighten<strong>in</strong>g<br />
on channel form and <strong>in</strong>stream habitat, effects of <strong>the</strong> altered<br />
fl ow regime on habitat, effects of former mercury m<strong>in</strong><strong>in</strong>g<br />
upstream and urban/agricultural runoff on water quality<br />
and <strong>the</strong> loss of native plant and animal species.<br />
To better understand <strong>the</strong> perceptions and preferences of<br />
local residents, Yang (2004) surveyed 1800 residents<br />
liv<strong>in</strong>g with<strong>in</strong> 400 m of Marsh Creek <strong>in</strong> Brentwood to<br />
assess <strong>the</strong>ir perceptions (and ideal images) of <strong>the</strong> creek.<br />
The residents consistently presented an ideal image of <strong>the</strong><br />
creek, identify<strong>in</strong>g luxuriant woods, year-round runn<strong>in</strong>g<br />
water, bountiful wildlife and easy access as features of <strong>the</strong><br />
‘natural’ or ‘orig<strong>in</strong>al’ Marsh Creek. However, this idyllic<br />
image of <strong>the</strong> creek is largely <strong>in</strong>consistent with <strong>the</strong> character<br />
of <strong>the</strong> creek as documented by historical evidence, with<br />
its Mediterranean-climate runoff regime. Similarly, many<br />
residents delighted <strong>in</strong> contact with wildlife, but did not<br />
realize that <strong>the</strong> most contacted species, i.e. crayfi sh, bullfrogs,<br />
bluegill and largemouth bass, are not native, but<br />
exotic generalists. Likewise, with vegetation, most subjects<br />
did not dist<strong>in</strong>guish native from <strong>in</strong>troduced plant<br />
species and even those who could tended to prefer vegetation<br />
that ‘looks natural without be<strong>in</strong>g overgrown’ regardless<br />
of orig<strong>in</strong> (Yang, 2004).<br />
The problems identifi ed by <strong>the</strong> residents contrasted<br />
sharply with those presented by <strong>the</strong> professionals <strong>in</strong> <strong>the</strong><br />
watershed report. By far <strong>the</strong> lead<strong>in</strong>g concern of surveyed<br />
residents was garbage and dump<strong>in</strong>g <strong>in</strong> <strong>the</strong> creek. Many<br />
residents considered <strong>the</strong> summer water levels too low,<br />
evidently without understand<strong>in</strong>g <strong>the</strong> highly seasonal nature<br />
of fl ow <strong>in</strong> Mediterranean-climate streams. Residents also<br />
regarded ‘mosquitoes/pests’ as more serious than ‘monotonous<br />
channel form’ and ‘poor habitat value,’ both major<br />
issues identifi ed <strong>in</strong> <strong>the</strong> watershed report (Ca<strong>in</strong> et al.,<br />
2003). Only ‘not enough shade’ was identifi ed as a concern<br />
both by <strong>the</strong> surveyed residents and <strong>in</strong> <strong>the</strong> watershed report<br />
(Yang, 2004).<br />
The substantial differences <strong>in</strong> perception and landscape<br />
preference between restoration scientists and local residents<br />
will be a source of uncerta<strong>in</strong>ty <strong>in</strong> sett<strong>in</strong>g restoration<br />
priorities and garner<strong>in</strong>g public support for restoration<br />
projects. As fund<strong>in</strong>g becomes available to plan restoration<br />
projects <strong>in</strong> Brentwood, <strong>the</strong>se gaps will need to be addressed<br />
<strong>in</strong> a participatory context so that confl icts <strong>in</strong> restoration<br />
goals can be reduced.<br />
4.7 CONCLUSIONS<br />
Uncerta<strong>in</strong>ties on <strong>the</strong> social and cultural fronts of stream<br />
restoration can be viewed as signify<strong>in</strong>g forward progress<br />
<strong>in</strong> <strong>the</strong> fi eld, ra<strong>the</strong>r than simply fur<strong>the</strong>r impediments<br />
<strong>in</strong> implement<strong>in</strong>g projects. Twenty years ago, when <strong>the</strong><br />
notions of creek restoration fi rst became widespread <strong>in</strong> <strong>the</strong><br />
United States, many eng<strong>in</strong>eers regarded ecological concerns<br />
as obstacles to be overcome <strong>in</strong> <strong>the</strong> s<strong>in</strong>gle-m<strong>in</strong>ded
a 1 0.5 0<br />
1 km b<br />
Plann<strong>in</strong>g <strong>River</strong> <strong>Restoration</strong> Projects: Social and Cultural Dimensions 57<br />
Figure 4.11 Topographic map details of marsh Creek <strong>in</strong> Brentwood: (a) 1914 and (b) 1978. (Source: US Geological Survey topographic<br />
maps.)<br />
Figure 4.12 (See also colour plate section) View of Marsh Creek<br />
channel <strong>in</strong> Brentwood (Photo by Kondolf, September 1991.)<br />
pursuit of dik<strong>in</strong>g, channeliz<strong>in</strong>g, straighten<strong>in</strong>g and culvert<strong>in</strong>g<br />
streams. It was only when eng<strong>in</strong>eers started to confront<br />
o<strong>the</strong>r viewpo<strong>in</strong>ts that ‘uncerta<strong>in</strong>ties’ were <strong>in</strong>troduced<br />
<strong>in</strong> <strong>the</strong>ir modus operandi. While confl icts between eng<strong>in</strong>eers<br />
and ecologists persist <strong>in</strong> restoration projects, by and<br />
large <strong>the</strong> eng<strong>in</strong>eer<strong>in</strong>g profession has embraced <strong>the</strong> need to<br />
work effectively with geomorphologists and biologists<br />
to achieve effective ecosystem restoration. Now we see<br />
<strong>in</strong>creas<strong>in</strong>gly that <strong>the</strong> ecological eng<strong>in</strong>eer<strong>in</strong>g approach is<br />
perturbed by <strong>the</strong> uncerta<strong>in</strong>ties <strong>in</strong>troduced by social and<br />
cultural concerns. In o<strong>the</strong>r words, <strong>the</strong> current phenomenon<br />
of restoration professionals experienc<strong>in</strong>g uncerta<strong>in</strong>ties on<br />
all fronts may be simply an <strong>in</strong>dicator of a rapidly broaden<strong>in</strong>g<br />
viewpo<strong>in</strong>t and recognition of problems without commensurate<br />
solutions.
58 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Although <strong>the</strong> possible range of goals, values, perception,<br />
aes<strong>the</strong>tic taste, use and mean<strong>in</strong>gs of a population for<br />
a stream can be overwhelm<strong>in</strong>g, and confl icts sometimes<br />
unavoidable, <strong>the</strong> goal of susta<strong>in</strong>able stream restoration<br />
<strong>in</strong>creas<strong>in</strong>gly requires <strong>in</strong>tegration of diverse po<strong>in</strong>ts of view.<br />
Institutionally, citizen groups have now become an <strong>in</strong>tegral<br />
part of many restoration efforts. Professionals have<br />
learnt that collaborative plann<strong>in</strong>g processes may not br<strong>in</strong>g<br />
about a fast solution, but it may provide substantial long<br />
term benefi ts <strong>in</strong> terms of political and social capital, and<br />
may yield more susta<strong>in</strong>able restoration projects.<br />
To be successful on a susta<strong>in</strong>able basis, stream restorations<br />
must be both technically sound and enjoy strong<br />
public support. Although decisions <strong>in</strong> stream restoration<br />
are essentially value driven, sound science is fundamental<br />
to constra<strong>in</strong> <strong>the</strong> range of possible solutions and evaluate<br />
possible alternatives. Without it, a Jamison Creek situation<br />
can result, <strong>in</strong> which <strong>the</strong> responsible agency selects a<br />
scientifi cally unsound option and <strong>the</strong> project fails. On <strong>the</strong><br />
o<strong>the</strong>r hand, a technically sound restoration plan is unlikely<br />
to be funded and implemented without strong public<br />
support, and unlikely to be susta<strong>in</strong>able if built without<br />
local buy-<strong>in</strong>.<br />
Where <strong>the</strong>re are signifi cant uncerta<strong>in</strong>ties on social and<br />
cultural aspects, <strong>the</strong>se should probably be settled before<br />
proceed<strong>in</strong>g to settle technical uncerta<strong>in</strong>ties. For example,<br />
until <strong>the</strong> values of large woody debris for fi sh or boaters<br />
are established, <strong>the</strong>re may be little po<strong>in</strong>t <strong>in</strong> quantify<strong>in</strong>g<br />
its catchment production and morphological qualities. In<br />
cities, we fi nd that <strong>the</strong> recreational potential of spontaneous<br />
uses is often conspicuous <strong>in</strong> its absence from <strong>the</strong><br />
agenda of stream restoration. Once <strong>the</strong>ir importance is<br />
recognized and spontaneous uses and <strong>the</strong>ir implied societal<br />
values are added to <strong>the</strong> restoration agenda, more<br />
precise research may be needed to assess <strong>the</strong>m.<br />
Cultural preferences (commonly unacknowledged)<br />
largely shape restoration goals. Build<strong>in</strong>g a culturally preferred<br />
form (such as a stable, meander<strong>in</strong>g channel) is<br />
perfectly reasonable as a restoration goal, but we suspect<br />
that <strong>the</strong> fi eld would benefi t from an explicit recognition of<br />
this as motivation, ra<strong>the</strong>r than cloak<strong>in</strong>g such projects <strong>in</strong><br />
seem<strong>in</strong>gly scientifi c details of channel morphology and<br />
(commonly vague) references to improved fi sh habitat. To<br />
<strong>the</strong> degree that cultural preferences rema<strong>in</strong> unacknowledged,<br />
<strong>the</strong>y <strong>in</strong>troduce greater uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong> trajectory<br />
of restoration projects. Cultural preference for tidy landscapes<br />
over messy landscapes (Piégay et al., 2005) should<br />
be acknowledged, so that ‘overgrown’ riparian zones can<br />
ei<strong>the</strong>r be ‘framed’ (Nassauer, 1995) or simply avoided <strong>in</strong><br />
urban areas. For ecological design to be truly successful<br />
and widely accepted, designers will need to fi nd ways to<br />
make stream restoration compell<strong>in</strong>g as designs (Moz<strong>in</strong>go,<br />
1997). The concept of ‘eco-revelatory design’ suggests<br />
that by accept<strong>in</strong>g humans <strong>in</strong>to <strong>the</strong> restored ecosystem and<br />
design<strong>in</strong>g <strong>the</strong> project to reveal ecological processes, we<br />
may achieve ecosystem restoration (to <strong>the</strong> extent possible<br />
<strong>in</strong> urban areas) while still ga<strong>in</strong><strong>in</strong>g public acceptance<br />
(Galatowitsch, 1998).<br />
4.8 ACKNOWLEDGEMENTS<br />
The research on which this chapter is based was partially<br />
supported by a grant from <strong>the</strong> University of California,<br />
Berkeley, Department of Landscape Architecture Beatrix<br />
Farrand Fund. Shannah Anderson contributed substantially<br />
with support<strong>in</strong>g research, fi gure and manuscript<br />
preparation, and review comments. Louise Moz<strong>in</strong>go contributed<br />
valuable ideas and references. The chapter was<br />
improved through comments by anonymous reviewers<br />
and <strong>the</strong> volume’s editors, Dave Sear and Steve Darby.<br />
REFERENCES<br />
Appleton J. 1975. The Experience of Landscape. John Wiley &<br />
Sons Ltd: Chichester.<br />
Bauer DM, Cyr NE, Swallow SK. 2002. Public preferences for<br />
compensatory mitigation of salt marsh losses: a cont<strong>in</strong>gent<br />
choice of alternatives. Conservation Biology 18: 401–411.<br />
Bernhardt ES et al. 2005. Syn<strong>the</strong>siz<strong>in</strong>g US river restoration<br />
efforts. Science 308: 636–637.<br />
Brookes A, Shields FD. 1996. <strong>River</strong> Channel <strong>Restoration</strong>:<br />
Guid<strong>in</strong>g Pr<strong>in</strong>ciples For Susta<strong>in</strong>able Projects. John Wiley &<br />
Sons Ltd: Chichester.<br />
Ca<strong>in</strong> JR, Rob<strong>in</strong>s JD, Beamish SS. 2003. The Past and Present<br />
Condition of <strong>the</strong> Marsh Creek Watershed. Natural Heritage<br />
Institute: San Francisco, California.<br />
Carson R. 1962. Silent Spr<strong>in</strong>g. Houghton Miffl <strong>in</strong> Company: New<br />
York, New York.<br />
CCHR (Citizens Commission on Human Rights). 2004. The<br />
real crisis on mental health today. Available at Community<br />
Ru<strong>in</strong>: Psychiatry’s Coercive Care, CCHR: Los Angeles,<br />
California.<br />
Chawla L. 1988. Children’s concern for <strong>the</strong> natural environment.<br />
Children’s Environment Quarterly 5 (3):13–20.<br />
Cobb E. 1977. The Ecology of Imag<strong>in</strong>ation <strong>in</strong> Childhood. Columbia<br />
University Press: New York, New York.<br />
Connick S, Innes J. 2003. Outcomes of collaborative water policy<br />
mak<strong>in</strong>g: apply<strong>in</strong>g complexity th<strong>in</strong>k<strong>in</strong>g to evaluation. Journal<br />
of Environmental Plann<strong>in</strong>g and Management 46: 177–197.<br />
Cooper-Marcus C. 1992. Environmental memories. In: Low SM,<br />
Altman I (Eds), Place Attachment, Human Behavior and Environment:<br />
Advances <strong>in</strong> Theory and Research, Vol. 12. Plenum<br />
Press: New York, New York; 87–112.<br />
Daniel TC, V<strong>in</strong><strong>in</strong>g J. 1983. Methodological issues <strong>in</strong> <strong>the</strong> assessment<br />
of landscape quality. In: Altman I, Wohlwill JF (Eds),<br />
Behavior and Natural Environment, Plenum Press: New York,<br />
New York; 39–84.
Dohi M. 1999. The community design process at Kamagasaki,<br />
Osaka, Japan. In: Hester RT, Kwesk<strong>in</strong> C (Eds), Democratic<br />
Design <strong>in</strong> <strong>the</strong> Pacifi c Rim, Ridge Times Press: Mendoc<strong>in</strong>o,<br />
California; 228–241.<br />
Eden SS, Tunstall S, Tapsell S. 2000. Translat<strong>in</strong>g nature: river<br />
restoration as nature-culture. Environment and Plann<strong>in</strong>g D:<br />
Society and Space 18: 257–273.<br />
Ehrlich PR. 1968. The Population Bomb. Buccaneer Books: New<br />
York, New York.<br />
Fagan K. 2005. Solano’s hidden homeless. San Francisco Chronicle.<br />
13 November 2005, pp. 1, A10–11.<br />
FISCRWG (Federal Interagency Stream Corridor <strong>Restoration</strong><br />
Work<strong>in</strong>g Group). 1998. Stream Corridor <strong>Restoration</strong> Handbook.<br />
Natural Resources Conservation Service: Wash<strong>in</strong>gton,<br />
DC.<br />
Flosi G et al. 1998. California Salmonid Stream Habitat <strong>Restoration</strong><br />
Manual (3rd edition). California Department of Fish and<br />
Game: Sacramento, California.<br />
Fornes J. 1994. Plan for <strong>the</strong> Metropolitan Park of Havana.<br />
Parque Metropolitano de la Habana: Havana, Cuba.<br />
Galatowitsch SM. 1998. Ecological design for environmental<br />
problem solv<strong>in</strong>g. Landscape Journal 17: 99–107.<br />
Garcia and Associates. 2005. Rock Creek-Cresta recreational<br />
streamfl ow monitor<strong>in</strong>g year one report: 2002 macro<strong>in</strong>vertebrate<br />
drift sampl<strong>in</strong>g North Fork Fea<strong>the</strong>r <strong>River</strong>, Plumas County<br />
CA. Prepared for Pacifi c Gas and Electric Company (FERC<br />
No. 1962).<br />
Hamilton J. 1993. Streams of hope. Sierra September/October<br />
1993: 98–122.<br />
Hart R. 1979. Children’s Experience of Place. Irv<strong>in</strong>gton Publishers:<br />
New York, New York.<br />
Harvey MR. 1989. Children’s experiences with vegetation.<br />
Children’s Environment Quarterly 6 (1): 36–43.<br />
Hester RT et al. 1988. ‘We’d like to tell you . . .’: Children’s<br />
views of life <strong>in</strong> Westport, California. Small Town 18 (4):<br />
19–24.<br />
Hogarth W. 1753. The Analysis of Beauty. Repr<strong>in</strong>ted <strong>in</strong> 1997 by<br />
Yale University Press: New Haven.<br />
Hough M. 1990. Out of Place: Restor<strong>in</strong>g Identity to <strong>the</strong> Regional<br />
Landscape. Yale University Press, New Haven.<br />
Hull RB, Robertson DP. 2000. The language of nature matters:<br />
we need a more public ecology. In: Gobster PH, Hull RB<br />
(Eds), Restor<strong>in</strong>g Nature, Island Press: Wash<strong>in</strong>gton, DC;<br />
7–118.<br />
Jones DR, Battaglia AM. 1989. Ma<strong>in</strong> Street <strong>River</strong>s: Mak<strong>in</strong>g Connections<br />
Between <strong>River</strong>s and Towns. Pennsylvania State University:<br />
State College, Pennsylvania.<br />
Junk WJ, Bayley PB, Sparks RE. 1989. The fl ood pulse concept<br />
<strong>in</strong> river-fl oodpla<strong>in</strong> system. Canadian Journal of Fisheries and<br />
Aquatic Sciences. 106: 110–127.<br />
Kaplan R. 1977. Preference and everyday nature: method and<br />
application. In: Stokols D (Ed), Perspectives on Environment<br />
and Behavior: Theory, Research, and Applications, Plenum<br />
Press: New York, New York; 235–250.<br />
Kaplan R, Kaplan S. 1989. The Experience of Nature: A Psychological<br />
Perspective. Cambridge University Press: New York,<br />
New York.<br />
Plann<strong>in</strong>g <strong>River</strong> <strong>Restoration</strong> Projects: Social and Cultural Dimensions 59<br />
Kondolf GM, Smeltzer MW, Railsback S. 2001. Design and<br />
performance of a channel reconstruction project <strong>in</strong> a coastal<br />
California gravel-bed stream. Environmental Management 28<br />
(6): 761–776.<br />
Lewis SA. 1995. Design and plann<strong>in</strong>g implications of uses, perceptions<br />
and attitudes of San Leandro Creek. Master’s <strong>the</strong>sis.<br />
Univeristy of California, Berkeley.<br />
Marchetti MP, Moyle PB. 2001. Effects of fl ow regime on fi sh<br />
assemblages <strong>in</strong> a regulated California stream. Ecological<br />
Applications 11 (2): 530–539.<br />
Marzolf GR, Valdez RA, Schmidt JC, Webb RH. 1998. Perspectives<br />
on river restoration <strong>in</strong> <strong>the</strong> Grand Canyon. Bullet<strong>in</strong> of <strong>the</strong><br />
Ecological Society of America 79 (4): 250–254.<br />
Moore RC. 1986. Childhood’s Doma<strong>in</strong>. Croom Helm: London.<br />
Moz<strong>in</strong>go LA. 1997. The aes<strong>the</strong>tics of ecological design: see<strong>in</strong>g<br />
science as culture. Landcape Journal 16: 46–59.<br />
Moz<strong>in</strong>go LA. 2005. Community participation and creek restoration<br />
<strong>in</strong> <strong>the</strong> East Bay of San Francisco, California. In: Hou J,<br />
Francis M, Brightbill N (Eds), (Re)construct<strong>in</strong>g Communities,<br />
<strong>the</strong> 5th Pacifi c Rim Conference on Participatory Community<br />
Design, Center for Design Research: University of California,<br />
Davis; 249–251.<br />
Nassauer JI. 1995. Messy ecosystems, orderly frames. Landscape<br />
Journal 14: 161–170.<br />
Nicholson S. 1971. The <strong>the</strong>ory of loose parts. Landscape Architecture<br />
62 (1): 30–34.<br />
NRC (National Research Council). 1992. <strong>Restoration</strong> of Aquatic<br />
Ecosystems. National Academy Press: Wash<strong>in</strong>gton, DC.<br />
Orr D. 1992. Ecological Literacy: Education and <strong>the</strong> Transition<br />
to a Postmodern World. State University of New York Press:<br />
Albany, New York.<br />
Otto B, McCormick K, Leccese M. 2004. Ecological riverfront<br />
design: restor<strong>in</strong>g rivers, connect<strong>in</strong>g communities. American<br />
Plann<strong>in</strong>g Association Plann<strong>in</strong>g Advisory Service Report<br />
Number 518–519, American Plann<strong>in</strong>g Association: Chicago,<br />
Ill<strong>in</strong>ois.<br />
Owens PE. 1988. Natural landscapes, ga<strong>the</strong>r<strong>in</strong>g places, and prospect<br />
refuges: characteristics of outdoor places valued by teens.<br />
Children’s Environments Quarterly 5 (2): 17–24.<br />
Piégay H, Mutz M, Gregory KJ. 2005. Public perception as a<br />
barrier to <strong>in</strong>troduc<strong>in</strong>g wood <strong>in</strong> rivers for restoration purposes.<br />
Environmental Management 36 (5): 665–674.<br />
Poff NL et al. 1997. The natural fl ow regime. BioScience 47:<br />
769–784.<br />
Postel S, Richter B. 2004. <strong>River</strong>s for Life. Island Press: Covelo,<br />
California.<br />
Purcell AH, Friedrich C, Resh VH. 2002. An assessment of a<br />
small urban stream restoration project <strong>in</strong> nor<strong>the</strong>rn California.<br />
<strong>Restoration</strong> Ecology 10: 685–694.<br />
Purcell AH. 2004. A long term post-project evaluation of an<br />
urban stream restoration project (Baxter Creek, el Cerrito, California).<br />
Term project <strong>in</strong> <strong>River</strong> <strong>Restoration</strong>, LA227, University of<br />
California, Berkeley. Available through <strong>the</strong> University of California<br />
library website (http://repositories.cdlib.org/wrca/<br />
restoration/)<br />
Reiser DW, Bjornn TC. 1979. Habitat requirements of anadromous<br />
salmonids. USDA General Technical Report PNW-96.
60 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Riley AL. 1998. Restor<strong>in</strong>g Streams <strong>in</strong> Cities: A Guide for Planners,<br />
Policymakers, and Citizens. Island Press: Wash<strong>in</strong>gton,<br />
DC.<br />
Rob<strong>in</strong>s JD, Ca<strong>in</strong> JR. 2002. The Past and Present Condition of <strong>the</strong><br />
Marsh Creek Watershed. Report by <strong>the</strong> Natural Heritage Institute<br />
and <strong>the</strong> Delta Science Center to <strong>the</strong> California Coastal<br />
Conservancy: Oakland, California.<br />
Rood SB, Mahoney JM. 2000. Revised <strong>in</strong>stream fl ow regulation<br />
enables cottonwood recruitment along <strong>the</strong> St Mary <strong>River</strong>,<br />
Alberta, Canada. <strong>River</strong>s 7: 109–125.<br />
Rosgen DL. 1994. A classifi cation of natural rivers. Catena 22:<br />
169–199.<br />
Ryan RL. 2000. A people-centered approach to restoration projects:<br />
<strong>in</strong>sights from understand<strong>in</strong>g attachment to urban natural<br />
areas. In: Gobster PH, Hull RB (Eds), Restor<strong>in</strong>g Nature, Island<br />
Press: Wash<strong>in</strong>gton, DC; 209–228.<br />
Ryan RL, Erickson DL, DeYoung R. 2002. Farmers’ motivations<br />
for adopt<strong>in</strong>g conservation practices along riparian zones <strong>in</strong> a<br />
Midwestern agricultural watershed. Journal of Environmental<br />
Plann<strong>in</strong>g and Management 46 (1): 19–37.<br />
Schmidt JC et al. 1998. Science and values <strong>in</strong> river restoration <strong>in</strong><br />
<strong>the</strong> Grand Canyon. BioScience 48 (9): 735–747.<br />
Smith SM, Prestegaard KL. 2005. Hydraulic performance of a<br />
morphology-based stream channel design. Water Resources<br />
Research 41 (W11413): 1–17.<br />
Tunstall SM, Penn<strong>in</strong>g-Rowsell EC, Tapsell SM, Eden SE. 2000.<br />
<strong>River</strong> restoration: public attitudes and expectations. Journal of<br />
<strong>the</strong> Institution of Water and Environmental Management 14<br />
(5): 363–370.<br />
Tunstall SM, Tapsell SM, House M. 2004. Children’s perceptions<br />
of river landscapes and play: what children’s photographs<br />
reveal. Landscape Research 29 (2): 181–204.<br />
Ulrich RS. 1983. Aes<strong>the</strong>tic and effective response to<br />
natural environment. In: Altman I, Wohlwill J (Eds), Behavior<br />
and <strong>the</strong> Natural Environment, Plenum: New York, New<br />
York.<br />
Ward JV, Stanford JA. 1995. Ecological connectivity <strong>in</strong> alluvial<br />
river ecosystems and its disruption by fl ow regulation. Regulated<br />
rivers: Research and Management 11: 105–119.<br />
Wolman MG. 1971. The nation’s rivers. Science 174: 905–918.<br />
Yang C-N. 2004. Invit<strong>in</strong>g Spontaneous Use <strong>in</strong>to Urban Streams.<br />
Doctoral dissertation. University of California, Berkeley.<br />
Available onl<strong>in</strong>e at http://www.lib.berkeley.edu/WRCA/<br />
restoration/<strong>the</strong>ses.html.
<strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat Edited by Stephen Darby and David Sear<br />
© 2008 John Wiley & Sons, Ltd<br />
5<br />
Conceptual and Ma<strong>the</strong>matical Modell<strong>in</strong>g <strong>in</strong><br />
<strong>River</strong> <strong>Restoration</strong>: Do We Have<br />
Unreasonable Confi dence?<br />
Michael Stewardson 1 and Ian Ru<strong>the</strong>rfurd 2<br />
1 Department of Civil and Environmental Eng<strong>in</strong>eer<strong>in</strong>g and eWater CRC, The University of Melbourne, Australia<br />
2 Geography Program, School of Resource Management, The University of Melbourne, Australia<br />
5.1 INTRODUCTION<br />
5.1.1 Geomorphic Modell<strong>in</strong>g <strong>in</strong><br />
<strong>Restoration</strong> Plann<strong>in</strong>g<br />
From <strong>the</strong> 1960s to <strong>the</strong> 1980s, applied fl uvial geomorphology<br />
described <strong>the</strong> degradation of streams <strong>in</strong> response to<br />
human disturbance. This endeavor was part of a larger<br />
quest to expla<strong>in</strong> <strong>the</strong> controls on stream form. From <strong>the</strong><br />
1990s, <strong>the</strong> discipl<strong>in</strong>e has found new vigor as skills have<br />
been turned to <strong>the</strong> restoration of those degraded streams.<br />
However, this change does not simply represent a new job<br />
opportunity; it represents a fundamental test of our knowledge<br />
of processes controll<strong>in</strong>g stream form.<br />
It is relatively easy to describe <strong>the</strong> degradation of fl uvial<br />
systems, and much of this work has identifi ed associations<br />
with human disturbances ra<strong>the</strong>r than strict causations.<br />
Recall <strong>the</strong> protracted debates about whe<strong>the</strong>r arroyo <strong>in</strong>cision<br />
was caused by clear<strong>in</strong>g, channelisation or climate<br />
change (Cooke and Reeves, 1976). It is much more challeng<strong>in</strong>g,<br />
fi rstly to recommend priorities for rehabilitat<strong>in</strong>g<br />
streams and, secondly, to implement <strong>the</strong>se changes. Geomorphologists<br />
have moved from be<strong>in</strong>g observers of human<br />
impact, to advisors and managers, actively <strong>in</strong>terven<strong>in</strong>g <strong>in</strong><br />
stream channels for environmental outcomes. Most of this<br />
<strong>in</strong>tervention has been <strong>in</strong> <strong>the</strong> areas of: channelised streams<br />
(Larson and Goldsmith, 1997), sediment slugs (Ru<strong>the</strong>rfurd,<br />
2001) and mitigat<strong>in</strong>g <strong>the</strong> effects of dams and fl ow<br />
regulation. Numerous articles describe <strong>the</strong> contribution<br />
that geomorphologists can make to stream restoration<br />
endeavors and <strong>the</strong>re is no shortage of admonitions to<br />
‘<strong>in</strong>clude a geomorphologist on every restoration project’<br />
(Sear, 1994; Brookes, 1995; Brierley et al., 1996; Brookes<br />
and Sear, 1996) and on every stream eng<strong>in</strong>eer<strong>in</strong>g project<br />
(Gilvear, 1999).<br />
Fluvial geomorphologists produce conceptual and<br />
ma<strong>the</strong>matical models that are central to many stream restoration<br />
projects. In this chapter it is argued that managers,<br />
ecologists, and even geomorphologists <strong>the</strong>mselves, can<br />
have a false sense of confi dence <strong>in</strong> <strong>the</strong>se models. As communities<br />
around <strong>the</strong> world <strong>in</strong>vest <strong>in</strong> stream restoration<br />
projects, false confi dence can have a huge direct and<br />
opportunity cost. We believe it is better to be frank about<br />
model uncerta<strong>in</strong>ties from <strong>the</strong> outset to promote realistic<br />
expectations of project success and a balanced and welltargeted<br />
<strong>in</strong>vestment <strong>in</strong> <strong>in</strong>vestigations to reduce <strong>the</strong>se<br />
uncerta<strong>in</strong>ties. Fluvial geomorphologists contribute to<br />
stream restoration projects <strong>in</strong> three ma<strong>in</strong> ways:<br />
1. The most basic work of a fl uvial geomorphologist <strong>in</strong><br />
a restoration project is to describe how a stream has<br />
changed its form over time, and to identify <strong>the</strong> factors<br />
that are responsible for <strong>the</strong>se changes. Such reconstructions<br />
have now passed from be<strong>in</strong>g a curiosity-driven<br />
activity for scientists, to basic consult<strong>in</strong>g practice. The<br />
result of <strong>the</strong>se <strong>in</strong>vestigations is usually a conceptual<br />
model [or a perceptual model <strong>in</strong> <strong>the</strong> parlance of Beven<br />
(2001)] that may be based on space-for-time substitution<br />
or historical co<strong>in</strong>cidence between geomorphic and
62 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
o<strong>the</strong>r catchment changes. Perhaps <strong>the</strong> best established<br />
geomorphic conceptual models are <strong>the</strong> <strong>in</strong>cised stream<br />
models (Schumm et al., 1984; Simon, 1989) where<br />
<strong>the</strong>re may be some empirical guidance for <strong>the</strong> magnitude<br />
of change (e.g. a threshold width–depth ratio for<br />
<strong>in</strong>cised streams Schumm et al., 1984). The conceptual<br />
model serves to describe <strong>the</strong> ‘orig<strong>in</strong>al’ condition of a<br />
system, which often becomes <strong>the</strong> ‘target’ for <strong>the</strong> restoration<br />
project. Fur<strong>the</strong>r, <strong>the</strong> conceptual model is used to<br />
predict <strong>the</strong> trajectory of channel change if <strong>the</strong>re is no<br />
<strong>in</strong>tervention.<br />
2. Geomorphologists recommend specifi c actions (or<br />
<strong>in</strong>terventions) that will alter process and form, to move<br />
<strong>the</strong> stream toward <strong>the</strong> ‘target’ state, identifi ed from <strong>the</strong><br />
conceptual model. These actions usually <strong>in</strong>volve chang<strong>in</strong>g<br />
<strong>the</strong> fl ow or sediment regime, or modify<strong>in</strong>g <strong>the</strong><br />
boundary of <strong>the</strong> channel to cause a change <strong>in</strong> some<br />
geomorphic variable (e.g. width, scour frequency,<br />
erosion rate). Design of project specifi cs is often based<br />
on a ma<strong>the</strong>matical model of geomorphic response<br />
related to <strong>the</strong> specifi c action proposed.<br />
3. They also predict <strong>the</strong> ‘secondary’ (perhaps un<strong>in</strong>tended)<br />
consequences of stream restoration projects. This<br />
could also be described as <strong>the</strong> ‘susta<strong>in</strong>ability’ of <strong>the</strong><br />
<strong>in</strong>tervention.<br />
Thus, uncerta<strong>in</strong>ty emerges at three levels: <strong>the</strong> validity<br />
of <strong>the</strong> conceptual model; whe<strong>the</strong>r <strong>the</strong> proposed <strong>in</strong>tervention<br />
results <strong>in</strong> <strong>the</strong> planned geomorphic change; and, fi nally,<br />
whe<strong>the</strong>r <strong>the</strong> change is susta<strong>in</strong>able. Typical examples of<br />
<strong>in</strong>-stream geomorphic actions <strong>in</strong> restoration projects are:<br />
• Incised, channelised streams are <strong>the</strong> classical example of<br />
stream restoration. A geomorphologist reconstructs<br />
<strong>the</strong> orig<strong>in</strong>al dimensions of a channelised stream <strong>in</strong><br />
Denmark by superimpos<strong>in</strong>g old maps and photos, and by<br />
mak<strong>in</strong>g pa<strong>in</strong>stak<strong>in</strong>g fi eld observations (Neilsen, 1996).<br />
Work<strong>in</strong>g with an eng<strong>in</strong>eer, a ‘stable’ re-meandered<br />
stream path that l<strong>in</strong>ks <strong>the</strong> palaeochannels for a rehabilitated<br />
stream is <strong>the</strong>n designed. The dimensions of <strong>the</strong><br />
channel and <strong>the</strong> variations <strong>in</strong> depth (<strong>the</strong> pool-riffl e<br />
sequence) are designed to be scaled to catchment area<br />
(Newbury and Gaboury, 1993). F<strong>in</strong>ally, <strong>the</strong> restoration<br />
that will lead to <strong>in</strong>creased erosion downstream of <strong>the</strong><br />
reach is predicted.<br />
• Gully<strong>in</strong>g has dumped a large pulse of sand <strong>in</strong>to a stream.<br />
A geomorphologist applies <strong>the</strong> classical ‘wave’ model<br />
of sand slug migration (Gilbert, 1917; Neilsen, 1996)<br />
and, on <strong>the</strong> basis of historical movement of <strong>the</strong> sand<br />
front, concludes that it will take over 50 years for <strong>the</strong><br />
sand to move through <strong>the</strong> reach. Extract<strong>in</strong>g <strong>the</strong> sand<br />
at a defi ned rate will protect <strong>the</strong> downstream reaches<br />
and accelerate recovery (Ru<strong>the</strong>rfurd, 2001). Build<strong>in</strong>g<br />
artifi cial spur dikes will also create habitat pools<br />
(Kuhnle et al., 2002).<br />
5.1.2 Why do we Care about <strong>Uncerta<strong>in</strong>ty</strong>?<br />
Regan et al. (2002) defi ne epistemic uncerta<strong>in</strong>ty as uncerta<strong>in</strong>ty<br />
associated with knowledge of <strong>the</strong> state of a system;<br />
it can be classifi ed <strong>in</strong>to six ma<strong>in</strong> types <strong>in</strong>clud<strong>in</strong>g random<br />
measurement error, natural variation and model uncerta<strong>in</strong>ty<br />
(compare to <strong>the</strong> classifi cation presented <strong>in</strong> Chapter<br />
3). This chapter is specifi cally concerned with <strong>the</strong>se three<br />
sources of uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong> context of conceptual and<br />
ma<strong>the</strong>matical geomorphic models used <strong>in</strong> restoration projects.<br />
Measurements of various sorts are used to establish<br />
<strong>the</strong> state of a system. Measurements may be used directly<br />
(e.g. <strong>the</strong> diameter of an <strong>in</strong>dividual gra<strong>in</strong> of sediment), but<br />
can also be transformed us<strong>in</strong>g some k<strong>in</strong>d of calibration<br />
(e.g. stream discharge, which is often estimated from an<br />
observation of stage and transformed to discharge us<strong>in</strong>g a<br />
rat<strong>in</strong>g curve). Measurement error is <strong>the</strong> result of errors <strong>in</strong><br />
<strong>the</strong> direct measurement and subsequent transformations.<br />
Natural variation presents a challenge for observ<strong>in</strong>g <strong>the</strong><br />
true state of a system. To address uncerta<strong>in</strong>ty, assumptions<br />
are often made about <strong>the</strong> statistical properties of environmental<br />
variations. Model uncerta<strong>in</strong>ty can arise both from<br />
<strong>the</strong> choice of variables and processes to be represented <strong>in</strong><br />
<strong>the</strong> model and from <strong>the</strong> method used to represent <strong>the</strong> relations<br />
between variables.<br />
<strong>Uncerta<strong>in</strong>ty</strong> is defi ned here as <strong>the</strong> range of possible<br />
values for a model variable (<strong>in</strong>put or response). One part<br />
of uncerta<strong>in</strong>ty is accuracy, which is <strong>the</strong> difference between<br />
a measurement and <strong>the</strong> ‘true value’. Uncerta<strong>in</strong>ties <strong>in</strong><br />
response variables are <strong>the</strong> consequence of <strong>the</strong> need to<br />
choose a conceptual and ma<strong>the</strong>matical representation of<br />
river processes, and uncerta<strong>in</strong>ties <strong>in</strong> <strong>in</strong>put parameters for<br />
<strong>the</strong> model estimated from fi eld measurements, previous<br />
studies or by calibration. Uncerta<strong>in</strong>ties can also exist as a<br />
consequence of unknown future environmental conditions<br />
<strong>in</strong> particular climatic conditions. It is possible that restoration<br />
works are destroyed by an extreme fl ood event not<br />
considered <strong>in</strong> <strong>the</strong> plann<strong>in</strong>g process. Design models often<br />
consider a range of expected conditions based on conditions<br />
experienced <strong>in</strong> <strong>the</strong> past, but uncerta<strong>in</strong>ties <strong>in</strong> <strong>the</strong><br />
actual conditions over <strong>the</strong> life <strong>the</strong> project create fur<strong>the</strong>r<br />
uncerta<strong>in</strong>ty <strong>in</strong> geomorphic design.<br />
Discussion of uncerta<strong>in</strong>ty, <strong>in</strong> <strong>the</strong> realm of management,<br />
can quickly degenerate <strong>in</strong>to an unfocused quest for greater<br />
accuracy, greater precision, more samples and more effort.<br />
It is usually scientists who write about certa<strong>in</strong>ty and <strong>the</strong>y<br />
may have a completely different perspective on <strong>the</strong> issue
Conceptual and Ma<strong>the</strong>matical Modell<strong>in</strong>g <strong>in</strong> <strong>River</strong> <strong>Restoration</strong>: Do We Have Unreasonable Confi dence? 63<br />
to managers. The notion of ‘certa<strong>in</strong>ty’ is not simply a<br />
statistical/scientifi c issue; it is at <strong>the</strong> heart of <strong>the</strong> social<br />
process of decision mak<strong>in</strong>g. The implications of estimat<strong>in</strong>g<br />
geomorphic uncerta<strong>in</strong>ty can be unexpected to those<br />
focused primarily on improved confi dence <strong>in</strong> model prediction.<br />
Whilst more certa<strong>in</strong> predictions should improve<br />
<strong>the</strong> chance of project success, consideration of uncerta<strong>in</strong>ties<br />
does not <strong>in</strong> itself improve model performance and can<br />
reduce public support for a project.<br />
It is necessary to be clear about why we care about<br />
geomorphic uncerta<strong>in</strong>ty <strong>in</strong> stream restoration. In reality,<br />
most scientifi c geomorphologists revel <strong>in</strong> <strong>the</strong> uncerta<strong>in</strong>ties<br />
of river process and form. Most of our endeavour is<br />
directed at situations where accepted models do not work,<br />
where we are uncerta<strong>in</strong>. If it was all certa<strong>in</strong>, we would do<br />
someth<strong>in</strong>g else. But <strong>in</strong> <strong>the</strong> realm of management, uncerta<strong>in</strong>ty<br />
is seldom welcome. Whilst <strong>the</strong> enthusiasm to rehabilitate<br />
streams will not disappear, <strong>the</strong> <strong>in</strong>itial fl ush of<br />
community and government support for <strong>the</strong>se endeavors<br />
could be lost if geomorphologists and o<strong>the</strong>r scientists<br />
foster unrealistic expectations of <strong>the</strong>se projects. As geomorphologists<br />
engage with eng<strong>in</strong>eers and managers, <strong>the</strong>y<br />
must decide how to deal with uncerta<strong>in</strong>ty (Volkman,<br />
1999), when admitt<strong>in</strong>g to too much uncerta<strong>in</strong>ty can weaken<br />
support for a project and possibly stall it. However, underplay<strong>in</strong>g<br />
uncerta<strong>in</strong>ties will underm<strong>in</strong>e confi dence <strong>in</strong> geomorphic<br />
advice when some projects <strong>in</strong>evitably ‘fail’.<br />
Ano<strong>the</strong>r reason to care about uncerta<strong>in</strong>ty is that it provides<br />
<strong>the</strong> justifi cation for improved geomorphic <strong>in</strong>vestigations<br />
prior to completion of a restoration plan. As will be<br />
demonstrated, <strong>in</strong> some cases it may be relatively cheap and<br />
easy to reduce uncerta<strong>in</strong>ty, but it can also be very expensive<br />
to do so. A strong conclusion of this chapter is that,<br />
given <strong>the</strong> large cost of stream restoration projects, sound<br />
geomorphic advice has often tended to be undervalued.<br />
Large projects are sometimes launched on <strong>the</strong> basis of<br />
fl imsy conceptual models, possibly because <strong>the</strong>ir uncerta<strong>in</strong>ty<br />
has not been properly considered.<br />
F<strong>in</strong>ally, adaptive management is often proposed as <strong>the</strong><br />
only reasonable way forward <strong>in</strong> <strong>the</strong> face of uncerta<strong>in</strong> restoration<br />
outcomes (see Chapter 14). The adaptive approach<br />
is to treat <strong>the</strong> restoration project as an experiment designed<br />
to <strong>in</strong>form our knowledge of how rivers respond to restoration.<br />
This knowledge is used to improve river restoration<br />
decisions for <strong>the</strong> particular river and presumably elsewhere.<br />
Despite <strong>the</strong> frequent calls for adaptive management<br />
of river restoration, actual examples of success are rare for<br />
a number of reasons (Walters, 1997; Ladson and Argent,<br />
2002). We argue that systematic consideration of uncerta<strong>in</strong>ties<br />
<strong>in</strong> geomorphic modell<strong>in</strong>g is essential to <strong>the</strong><br />
application of adaptive approaches <strong>in</strong> river restoration.<br />
Management experiments and associated monitor<strong>in</strong>g need<br />
be targeted to reduc<strong>in</strong>g <strong>the</strong>se uncerta<strong>in</strong>ties if <strong>the</strong>y are to<br />
feed back effectively <strong>in</strong>to future restoration practice.<br />
5.1.3 Introduction to Case Studies<br />
Geomorphologists normally acknowledge <strong>the</strong> uncerta<strong>in</strong>ties<br />
<strong>in</strong> <strong>the</strong>ir models but it is rare for <strong>the</strong>se uncerta<strong>in</strong>ties to<br />
be quantifi ed or systematically exam<strong>in</strong>ed. There could be<br />
a perception that <strong>the</strong>se uncerta<strong>in</strong>ties are relatively small<br />
and have limited signifi cance <strong>in</strong> restoration decisions.<br />
There may be limited experience amongst geomorphologists<br />
<strong>in</strong> <strong>the</strong> quantitative aspects of uncerta<strong>in</strong>ty analysis<br />
which could discourage attempts to handle <strong>the</strong>se explicitly.<br />
There may also be a concern that quantify<strong>in</strong>g uncerta<strong>in</strong>ties<br />
will underm<strong>in</strong>e support for a project. In any case,<br />
<strong>the</strong>re is currently very little published <strong>in</strong>formation on <strong>the</strong><br />
scale of uncerta<strong>in</strong>ties <strong>in</strong> geomorphic studies for river<br />
restoration. <strong>Uncerta<strong>in</strong>ty</strong> analyses <strong>in</strong> case study projects are<br />
needed to <strong>in</strong>form discussion of how best to handle <strong>the</strong>se<br />
uncerta<strong>in</strong>ties <strong>in</strong> <strong>the</strong> future.<br />
This chapter exam<strong>in</strong>es uncerta<strong>in</strong>ties <strong>in</strong> geomorphic<br />
modell<strong>in</strong>g for two river restoration projects. The two case<br />
studies <strong>in</strong>volve, respectively, conceptual and design models<br />
for restoration plann<strong>in</strong>g. The scale of uncerta<strong>in</strong>ties and<br />
also <strong>the</strong> potential benefi ts of a systematic analysis of<br />
uncerta<strong>in</strong>ties <strong>in</strong> <strong>the</strong>se projects are exam<strong>in</strong>ed. Based on<br />
<strong>the</strong>se case studies, it is suggested how uncerta<strong>in</strong>ties might<br />
best be handled <strong>in</strong> <strong>the</strong> future by geomorphologists develop<strong>in</strong>g<br />
conceptual and design models for river restoration<br />
plann<strong>in</strong>g.<br />
The fi rst case study concerns <strong>the</strong> development of a<br />
conceptual model for plann<strong>in</strong>g Australia’s largest stream<br />
restoration project, on <strong>the</strong> Snowy <strong>River</strong>. The <strong>in</strong>fl uence of<br />
<strong>the</strong> conceptual model on proposed plans is exam<strong>in</strong>ed and,<br />
<strong>in</strong> particular, <strong>the</strong> implication of subsequent changes <strong>in</strong> <strong>the</strong><br />
geomorphic model. The second case study concerns <strong>the</strong><br />
design of a fl ush<strong>in</strong>g fl ow <strong>in</strong> <strong>the</strong> Goulburn <strong>River</strong>, Victoria,<br />
us<strong>in</strong>g a one-dimensional hydraulic modell<strong>in</strong>g approach. In<br />
this design problem, multiple sources of uncerta<strong>in</strong>ties are<br />
quantifi ed and <strong>the</strong> key uncerta<strong>in</strong>ties <strong>in</strong> design<strong>in</strong>g <strong>the</strong> fl ush<strong>in</strong>g<br />
fl ow identifi ed. A two-dimensional numerical model<br />
may have had some advantages over a one-dimensional<br />
model approach <strong>in</strong> this problem. However, a onedimensional<br />
model (similar to HEC RAS) is used because<br />
it is currently <strong>the</strong> standard approach used throughout <strong>the</strong><br />
<strong>in</strong>dustry. Thus, we are concerned with <strong>the</strong> uncerta<strong>in</strong>ty associated<br />
with current practice <strong>in</strong> stream restoration projects<br />
ra<strong>the</strong>r than with <strong>the</strong> level of certa<strong>in</strong>ty that is <strong>the</strong>oretically<br />
possible with detailed scientifi c study. It is often<br />
assumed that uncerta<strong>in</strong>ties associated with modell<strong>in</strong>g <strong>the</strong><br />
physical response of channels are small (c.f. biological<br />
responses, see Chapter 8). The case studies demonstrate
64 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
that uncerta<strong>in</strong>ties associated with <strong>the</strong> ‘accepted’ approaches<br />
to geomorphic modell<strong>in</strong>g can be substantial.<br />
5.2 CASE STUDY OF A GEOMORPHIC<br />
CONCEPTUAL MODEL<br />
5.2.1 Introduction<br />
The follow<strong>in</strong>g section describes <strong>the</strong> chronological development<br />
of <strong>the</strong> geomorphic conceptual model for <strong>the</strong><br />
restoration of <strong>the</strong> lower 32 km of <strong>the</strong> Snowy <strong>River</strong> <strong>in</strong><br />
south-eastern Australia. The Snowy <strong>River</strong> dra<strong>in</strong>s a catchment<br />
of 15 800 km2 situated <strong>in</strong> New South Wales and Victoria<br />
<strong>in</strong> south-east Australia, before discharg<strong>in</strong>g <strong>in</strong>to Bass<br />
Strait (Brizga and F<strong>in</strong>layson, 1994). For most of its course,<br />
<strong>the</strong> river fl ows <strong>in</strong> a narrow valley, that only widens around<br />
Bete Belong (Figure 5.1). The river below Bete Belong is<br />
a perched, sand-bed, channel between 70 m and 170 m<br />
wide. The downstream end of <strong>the</strong> Snowy <strong>River</strong> is estuar<strong>in</strong>e,<br />
with tidal <strong>in</strong>fl uences persist<strong>in</strong>g upstream to Orbost.<br />
Figure 5.1 The lower Snowy <strong>River</strong> <strong>in</strong> Victoria (from F<strong>in</strong>layson and Bird, 1989)<br />
When <strong>the</strong> fi rst European settlers arrived <strong>in</strong> <strong>the</strong> late<br />
1840s, <strong>the</strong> Snowy <strong>River</strong> fl oodpla<strong>in</strong> was swampy wetlands<br />
away from <strong>the</strong> channel, with <strong>the</strong> higher levees along <strong>the</strong><br />
channel covered <strong>in</strong> warm temperate ra<strong>in</strong>forest (described<br />
as ‘jungle’) (Owen, 1997). By <strong>the</strong> 1880s, <strong>the</strong> stream banks<br />
had been cleared and <strong>the</strong> clear<strong>in</strong>g and dra<strong>in</strong><strong>in</strong>g of <strong>the</strong><br />
wetlands had begun. By <strong>the</strong> 1930s, artifi cial fl ood levees<br />
had been built along much of <strong>the</strong> river and much of <strong>the</strong><br />
swampy wetlands had been dra<strong>in</strong>ed for graz<strong>in</strong>g. Over<br />
<strong>the</strong> same period, large woody debris was removed from<br />
<strong>the</strong> stream. This began <strong>in</strong> <strong>the</strong> 1880s to aid navigation<br />
(Seddon, 1994) and reached a peak <strong>in</strong> <strong>the</strong> 1950s to combat<br />
perceived aggradation of <strong>the</strong> bed and to reduce fl ood peaks<br />
(F<strong>in</strong>layson and Bird, 1989).<br />
A second phase of disturbance was <strong>in</strong>itiated by construction<br />
<strong>in</strong> <strong>the</strong> river’s headwaters under <strong>the</strong> Snowy Mounta<strong>in</strong><br />
Scheme. Completed <strong>in</strong> 1967, <strong>the</strong> Snowy Mounta<strong>in</strong> Scheme<br />
diverts water out of <strong>the</strong> Snowy <strong>River</strong> catchment for hydroelectric<br />
power generation and for <strong>the</strong> supply of water for<br />
irrigation <strong>in</strong> <strong>the</strong> neighbor<strong>in</strong>g Murray and Murrumbidgee
Conceptual and Ma<strong>the</strong>matical Modell<strong>in</strong>g <strong>in</strong> <strong>River</strong> <strong>Restoration</strong>: Do We Have Unreasonable Confi dence? 65<br />
catchments. Ersk<strong>in</strong>e et al. (1999) concluded that <strong>the</strong> Snowy<br />
Mounta<strong>in</strong> Scheme had reduced median and low fl ows by<br />
60–70%, as far downstream as Jarrahmond.<br />
5.2.2 The Geomorphic Model<br />
The follow<strong>in</strong>g is a chronological summary of <strong>the</strong> history of<br />
stream channel change recorded by various workers on <strong>the</strong><br />
lower Snowy <strong>River</strong> and explanations for <strong>the</strong>se changes.<br />
1. There are anecdotal suggestions that <strong>the</strong> lower Snowy<br />
<strong>River</strong> <strong>in</strong>creased its width at <strong>the</strong> end of <strong>the</strong> n<strong>in</strong>eteenth<br />
century. For example, <strong>the</strong> Snowy <strong>River</strong> Mail (July 1,<br />
1884) wrote:<br />
‘The river which, when settlement fi rst appeared<br />
here, was not half its present width, is <strong>in</strong>creas<strong>in</strong>g at<br />
every fl ood, widen<strong>in</strong>g and shallow<strong>in</strong>g <strong>the</strong> channel<br />
and precipitat<strong>in</strong>g huge trees <strong>in</strong>to its bed . . .’ (cited <strong>in</strong><br />
Seddon, 1994).<br />
In review<strong>in</strong>g <strong>the</strong>se claims, F<strong>in</strong>layson and Bird (1989)<br />
concluded that <strong>the</strong>re was not suffi cient evidence to substantiate<br />
this catastrophic <strong>in</strong>crease <strong>in</strong> width.<br />
2. S<strong>in</strong>ce <strong>the</strong> 1930s, people liv<strong>in</strong>g along <strong>the</strong> Snowy <strong>River</strong><br />
have been conv<strong>in</strong>ced that <strong>the</strong> river was becom<strong>in</strong>g<br />
shallower and fi ll<strong>in</strong>g with sand (Strom, 1936). Flush<strong>in</strong>g<br />
out this sand was probably <strong>the</strong> earliest ‘restoration’<br />
target for <strong>the</strong> river, lead<strong>in</strong>g to desnagg<strong>in</strong>g and o<strong>the</strong>r<br />
works <strong>in</strong> <strong>the</strong> stream <strong>in</strong> <strong>the</strong> 1950s. Despite <strong>the</strong> local<br />
conviction that <strong>the</strong> bed has aggraded, comparisons of<br />
16 repeat cross-sections, dat<strong>in</strong>g from <strong>the</strong> 1920s, along<br />
<strong>the</strong> Jarrahmond reach, do not support this view (Gippel,<br />
2002). Instead <strong>the</strong> bed level fl uctuates over a range of<br />
±2 m. Brizga and F<strong>in</strong>layson (1994) suggested that <strong>the</strong><br />
reduced fl ows of <strong>the</strong> river s<strong>in</strong>ce regulation mean that<br />
more of <strong>the</strong> bed is now visible, lead<strong>in</strong>g to <strong>the</strong> illusion<br />
of aggradation. This is a controversial suggestion<br />
amongst <strong>the</strong> local people, who still see ameliorat<strong>in</strong>g <strong>the</strong><br />
effects of sedimentation as <strong>the</strong> major restoration target<br />
for <strong>the</strong> river. This perception may come, <strong>in</strong> part, from<br />
<strong>the</strong> fact that some deep pools <strong>in</strong> <strong>the</strong> river have certa<strong>in</strong>ly<br />
fi lled-<strong>in</strong> s<strong>in</strong>ce <strong>the</strong> 1970s, particularly around Bete<br />
Bolong (Gippel, 2002).<br />
3. In <strong>the</strong> 1990s, attention turned to <strong>the</strong> effect of <strong>the</strong> Snowy<br />
Mounta<strong>in</strong> Scheme on <strong>the</strong> geomorphology of <strong>the</strong> river.<br />
Brizga and F<strong>in</strong>layson (1992) mentioned <strong>the</strong> loss of<br />
lateral bars, and <strong>the</strong>ir associated pools <strong>in</strong> <strong>the</strong> lower<br />
Snowy <strong>River</strong>, and speculated that <strong>the</strong> cause could be<br />
fl ow regulation (Figure 5.2). Ersk<strong>in</strong>e and Tilleard<br />
(1997) described <strong>the</strong> loss of <strong>the</strong> bars and related it more<br />
Figure 5.2 1940 aerial photograph of <strong>the</strong> Lower Snowy <strong>River</strong><br />
upstream of Lynn’s Gulch, show<strong>in</strong>g well developed lateral bars<br />
and associated pools<br />
strongly to <strong>the</strong> loss of some formative discharges after<br />
regulation <strong>in</strong> 1967. ‘. . . rhythmically spaced, bankattached,<br />
alternate side bars with well defi ned poolriffl<br />
e sequence were present above <strong>the</strong> estuary at Bete<br />
Bolong before 1967 and <strong>the</strong>y have never reformed<br />
s<strong>in</strong>ce <strong>the</strong>n . . .’ (Ersk<strong>in</strong>e et al., 1999). Stewardson<br />
(1998) concluded that, not only has <strong>the</strong> frequency of<br />
fl ows that are thought to form bars and pools fallen<br />
dramatically, but <strong>the</strong> <strong>in</strong>cidence of low fl ows that can<br />
potentially <strong>in</strong>-fi ll <strong>the</strong> pools with sediment has <strong>in</strong>creased.<br />
Replac<strong>in</strong>g deep pools with a ‘plane bed’ of sand is<br />
thought to provide poor habitat, particularly for migrat<strong>in</strong>g<br />
fi sh (Raadick and O’Connor, 1997). Twelve of <strong>the</strong><br />
seventeen fi sh species found <strong>in</strong> <strong>the</strong> river are migratory.<br />
Return<strong>in</strong>g <strong>the</strong> pools to <strong>the</strong> lower Snowy <strong>River</strong> presented<br />
an elegant and achievable goal for stream restoration<br />
and was <strong>the</strong> ma<strong>in</strong> recommendation of <strong>the</strong> fi rst<br />
restoration plan (ID&A, 1998). The recommendation
66 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
was also supported by <strong>the</strong> Snowy Water Inquiry<br />
(1998a,b). Follow<strong>in</strong>g <strong>the</strong>se reports, <strong>the</strong> effort <strong>in</strong> <strong>the</strong><br />
project swung to design<strong>in</strong>g timber pile fi elds (retards)<br />
at <strong>the</strong> historical locations of <strong>the</strong> bank attached side<br />
bars. It was hoped that <strong>the</strong>se retards would encourage<br />
sand deposition and eventually scour pools (ID&A,<br />
1998; Gippel et al., 2002).<br />
4. Gippel (2002) has comprehensively reviewed <strong>the</strong> evidence<br />
for channel change <strong>in</strong> <strong>the</strong> lower Snowy <strong>River</strong>,<br />
add<strong>in</strong>g <strong>the</strong> evidence from two student <strong>the</strong>ses. Gippel<br />
concludes that it is likely that <strong>the</strong> Snowy <strong>River</strong> did<br />
dramatically <strong>in</strong>crease <strong>in</strong> width after <strong>the</strong> 1870 fl ood, <strong>in</strong><br />
common with o<strong>the</strong>r rivers <strong>in</strong> Gippsland (Brooks and<br />
Brierley, 1997; Brooks et al., 2003). This suggests that<br />
<strong>the</strong> pre-1967 un-regulated river (that has up till now<br />
formed <strong>the</strong> ‘reference’ for <strong>the</strong> restoration strategy) was<br />
<strong>in</strong> far from ‘natural condition’. Gippel concludes that<br />
<strong>the</strong> evidence that regulation removed <strong>the</strong> naturally<br />
occurr<strong>in</strong>g lateral bars, is weak. An alternative possibility<br />
is that bars did not form until <strong>the</strong> river widened after<br />
disturbance, dur<strong>in</strong>g <strong>the</strong> early period of European settlement<br />
and, even <strong>the</strong>n, <strong>the</strong> bars occurred only when<br />
hydrological conditions were ideal. Regulation caused<br />
<strong>the</strong>se ideal hydrological conditions to be less likely. In<br />
addition, Gippel concluded that:<br />
After 1940, <strong>the</strong> alternate bar and pool morphology<br />
only ever existed <strong>in</strong> <strong>the</strong> straight Jarrahmond/Bete<br />
Bolong reach (10.2 km of <strong>the</strong> 32 km lowland section<br />
of <strong>the</strong> river).<br />
It has been <strong>in</strong>correctly assumed that <strong>the</strong> alternate bars<br />
always occurred <strong>in</strong> <strong>the</strong> same location, with <strong>the</strong> same<br />
wavelength, when <strong>in</strong> fact <strong>the</strong>y moved and changed<br />
form.<br />
Even more fundamentally, <strong>the</strong> l<strong>in</strong>k between pools<br />
and fi sh diversity has never been well established.<br />
Even <strong>in</strong> <strong>the</strong> orig<strong>in</strong>al report by Raadik and O’Connor<br />
(1997), <strong>the</strong> abundance and diversity of fi sh <strong>in</strong> <strong>the</strong><br />
reaches without pools were found to be higher than<br />
<strong>in</strong> <strong>the</strong> reference reaches with pools.<br />
There is some question about whe<strong>the</strong>r <strong>the</strong> pools can<br />
be susta<strong>in</strong>ed. The persistent low fl ows that characterise<br />
<strong>the</strong> regulated regime could quickly <strong>in</strong>-fi ll any<br />
pools that are scoured by favourable fl ow events.<br />
At present, lateral bars and pools rema<strong>in</strong> <strong>the</strong> major<br />
focus of <strong>the</strong> restoration plan on <strong>the</strong> Snowy <strong>River</strong>. The<br />
Victorian Government is develop<strong>in</strong>g a major trial (US$ 1.5<br />
million) of retard structures (<strong>in</strong> <strong>the</strong> fi eld and <strong>in</strong> fl umes), as<br />
a pilot project before build<strong>in</strong>g <strong>the</strong> major pile-fi elds. The<br />
trial has been held up because <strong>the</strong> local community would<br />
not accept <strong>the</strong> assurances of <strong>the</strong> eng<strong>in</strong>eers and geomor-<br />
phologists that <strong>the</strong> proposed log structures would not<br />
cause any change <strong>in</strong> fl ood stage or duration. Ironically, <strong>the</strong><br />
project has been stopped, not because of scientifi c uncerta<strong>in</strong>ty<br />
about <strong>the</strong> goals of <strong>the</strong> project, but because <strong>the</strong><br />
community rema<strong>in</strong>s uncerta<strong>in</strong> about someth<strong>in</strong>g that <strong>the</strong><br />
eng<strong>in</strong>eers and scientists are certa<strong>in</strong> of: <strong>the</strong> m<strong>in</strong>or hydraulic<br />
effect of <strong>the</strong> works on fl oods.<br />
5.2.3 Analysis of Uncerta<strong>in</strong>ties<br />
The foundation for stream restoration projects are conceptual<br />
models of physical and biological change. This case<br />
study demonstrates that <strong>the</strong> problem with much of <strong>the</strong><br />
proposed restoration comes from a false sense of certa<strong>in</strong>ty<br />
<strong>in</strong> <strong>the</strong>se models. Several geomorphologists have contributed<br />
to a conceptual model of channel change. The restoration<br />
plan has proceeded on <strong>the</strong> basis of a conceptual<br />
model that dismissed aggradation by sand as a geomorphic<br />
change, but did identify a clear co<strong>in</strong>cidence between<br />
fl ow regulation and <strong>the</strong> loss of lateral bars. This neat conceptual<br />
model <strong>the</strong>n formed <strong>the</strong> basis for a major restoration<br />
project <strong>in</strong>volv<strong>in</strong>g eng<strong>in</strong>eer<strong>in</strong>g works to artifi cially<br />
recreate lateral bars and pools. Over <strong>the</strong> last fi ve years, <strong>the</strong><br />
eng<strong>in</strong>eer<strong>in</strong>g aspects of <strong>the</strong> project have taken hold. The<br />
key question for managers now is not whe<strong>the</strong>r pools are<br />
an appropriate goal but which eng<strong>in</strong>eer<strong>in</strong>g design will<br />
develop pools most effi ciently.<br />
It is reasonable to imag<strong>in</strong>e that managers beg<strong>in</strong> by be<strong>in</strong>g<br />
uncerta<strong>in</strong> about how a geomorphic system functions. They<br />
<strong>the</strong>n commission <strong>in</strong>vestigations that lead to progressively<br />
greater certa<strong>in</strong>ty, until restoration decisions can be made<br />
with confi dence. In <strong>the</strong> turbulent boundary between<br />
science and management (Cullen, 1989), however, certa<strong>in</strong>ty<br />
is a fi ckle commodity. Consider <strong>the</strong> chronology of<br />
certa<strong>in</strong>ty <strong>in</strong> <strong>the</strong> Snowy <strong>River</strong> project:<br />
1. The local community and <strong>the</strong> river managers were<br />
completely certa<strong>in</strong> that <strong>the</strong> river was aggrad<strong>in</strong>g and that<br />
<strong>the</strong> appropriate restoration strategy was to somehow<br />
remove <strong>the</strong> sand. Theories of aggradation were subsequently<br />
dismissed follow<strong>in</strong>g <strong>the</strong> geomorphic <strong>in</strong>vestigation<br />
of Brizga and F<strong>in</strong>layson (1994).<br />
2. Early geomorphic assessments concluded that regulation<br />
had led to <strong>the</strong> loss of alternate bars and restoration<br />
plans were developed to restore <strong>the</strong>m. Gippel’s (2002)<br />
subsequent review revealed strong evidence of channel<br />
widen<strong>in</strong>g <strong>in</strong> <strong>the</strong> late 1800s. S<strong>in</strong>ce alternate bars have<br />
only been found <strong>in</strong> wider reaches of <strong>the</strong> lower Snowy<br />
<strong>River</strong>, and only dur<strong>in</strong>g ideal hydrological periods,<br />
alternate bars were probably not a natural feature of <strong>the</strong><br />
channel prior to European settlement.
Conceptual and Ma<strong>the</strong>matical Modell<strong>in</strong>g <strong>in</strong> <strong>River</strong> <strong>Restoration</strong>: Do We Have Unreasonable Confi dence? 67<br />
3. F<strong>in</strong>ally, <strong>the</strong> restoration trial is be<strong>in</strong>g delayed because<br />
scientists are not able to conv<strong>in</strong>ce local stakeholders<br />
about <strong>the</strong> m<strong>in</strong>imal hydraulic effect of <strong>the</strong> proposed<br />
works.<br />
There are many examples where geomorphic conceptual<br />
models have been found want<strong>in</strong>g when used as a basis<br />
for management. Much of <strong>the</strong> criticism <strong>in</strong> <strong>the</strong> literature<br />
<strong>in</strong>volves situations where ‘real geomorphology’ has been<br />
replaced by ‘cookbook’ approaches (Sear, 1994). Much of<br />
this criticism has been directed at restoration projects<br />
based on classifi cation systems, such as that of Rosgen<br />
(1996). In one example, Kondolf et al. (2001) described<br />
a meander<strong>in</strong>g channel <strong>in</strong>appropriately constructed <strong>in</strong> a<br />
naturally braided stream. Miller and Ritter (1996) provide<br />
a more general review of <strong>the</strong> Rosgen method. The Snowy<br />
<strong>River</strong> Project, by contrast, is not superfi cial geomorphology,<br />
but uses <strong>the</strong> earnest pr<strong>in</strong>ciples recommended by<br />
geomorphologists (Kondolf, 2000). Even ‘real geomorphology’<br />
can be uncerta<strong>in</strong>. Kondolf and Micheli (1995)<br />
were correct when <strong>the</strong>y argued that every stream restoration<br />
project must be considered an experiment. This sits<br />
well with <strong>the</strong> sometimes stark uncerta<strong>in</strong>ties that surround<br />
diagnosis of geomorphic ‘problems’. However, it does not<br />
sit well with multi-million dollar <strong>in</strong>tervention projects.<br />
The lower Snowy <strong>River</strong> restoration project, for example,<br />
is planned to cost US$20 million, and this is before <strong>the</strong><br />
environmental fl ow component is <strong>in</strong>cluded.<br />
5.2.4 Discussion<br />
Process based conceptual models of stream channel<br />
change (with or without biological conceptual models) are<br />
one of <strong>the</strong> key contributions of geomorphology to stream<br />
restoration. However, <strong>the</strong>y are also a major source of<br />
uncerta<strong>in</strong>ty. A restoration project will proceed on <strong>the</strong> basis<br />
of a geomorphic conceptual model about which everybody<br />
is <strong>in</strong>itially confi dent. However, <strong>the</strong> Snowy <strong>River</strong> case<br />
study shows that this confi dence can be mislead<strong>in</strong>g.<br />
Fur<strong>the</strong>r <strong>in</strong>vestigation can reduce confi dence <strong>in</strong> a model,<br />
but by <strong>the</strong> time geomorphologists have settled on a conceptual<br />
model that deserves confi dence, <strong>the</strong> restoration<br />
process has moved on. Chang<strong>in</strong>g <strong>the</strong> model becomes<br />
<strong>in</strong>creas<strong>in</strong>gly diffi cult <strong>in</strong> <strong>the</strong> political and management<br />
process. Multi-million dollar restoration projects can be<br />
launched on <strong>the</strong> basis of uncerta<strong>in</strong> conceptual models. Not<br />
surpris<strong>in</strong>gly, managers and eng<strong>in</strong>eers are impatient to<br />
proceed to <strong>the</strong> ‘real’ bus<strong>in</strong>ess of restoration, which is<br />
build<strong>in</strong>g th<strong>in</strong>gs and chang<strong>in</strong>g th<strong>in</strong>gs. At least here <strong>the</strong><br />
uncerta<strong>in</strong>ty can be quantifi ed.<br />
Geomorphologists have argued strongly <strong>the</strong> central<br />
importance of understand<strong>in</strong>g <strong>the</strong> geomorphic context of<br />
restoration projects to ensure <strong>the</strong>ir susta<strong>in</strong>ability. If this<br />
is <strong>the</strong> case, a restoration plan will be underm<strong>in</strong>ed if <strong>the</strong><br />
conceptual model from which it was developed is wrong.<br />
In <strong>the</strong> case of <strong>the</strong> Snowy <strong>River</strong>, <strong>the</strong> <strong>in</strong>itial model of aggradation<br />
led to calls for artifi cial sand extraction to restore<br />
<strong>the</strong> river. However, such efforts would have been unsusta<strong>in</strong>able<br />
s<strong>in</strong>ce <strong>the</strong> sediment <strong>in</strong> <strong>the</strong> Snowy <strong>River</strong> bed was<br />
not a discrete slug of sand. Any pool formed by sand<br />
extraction would have been <strong>in</strong>fi lled dur<strong>in</strong>g subsequent<br />
storms.<br />
The accuracy of a conceptual model can be critical to<br />
<strong>the</strong> success of a project. However, <strong>the</strong>re is often only a<br />
poor basis for judg<strong>in</strong>g <strong>the</strong> levels of uncerta<strong>in</strong>ty <strong>in</strong> those<br />
models. Some conceptual models are reasonably simple<br />
and have been tested <strong>in</strong> numerous situations. The channelised<br />
stream and sand-slug models are examples of welltested<br />
models <strong>in</strong> general terms, although apply<strong>in</strong>g <strong>the</strong>m <strong>in</strong><br />
specifi c cases is challeng<strong>in</strong>g because <strong>the</strong>y provide a direction<br />
of change ra<strong>the</strong>r than ei<strong>the</strong>r rates or magnitudes.<br />
So, given that each stream requires a variant of a conceptual<br />
model, how is <strong>the</strong> certa<strong>in</strong>ty of that model judged?<br />
At present <strong>the</strong> model will be presented, with more or less<br />
confi dence, after a geomorphic study. That confi dence is<br />
based on ei<strong>the</strong>r <strong>the</strong> weight of reconstructed evidence (‘all<br />
12 of <strong>the</strong> cross-sections changed <strong>in</strong> <strong>the</strong> same way’), on<br />
precedents provided by analogous cases (‘a very similar<br />
model has been described on three nearby rivers’), or on<br />
<strong>the</strong> reputation and forcefulness of <strong>the</strong> <strong>in</strong>vestigator. Given<br />
that most of <strong>the</strong>se restoration projects cannot wait for<br />
models to be published <strong>in</strong> <strong>the</strong> peer-reviewed literature,<br />
how can managers judge <strong>the</strong> uncerta<strong>in</strong>ty of <strong>the</strong>se conceptual<br />
models? Here are fi ve proposals:<br />
1. The most obvious way to improve confi dence is to<br />
subject a conceptual model to anonymous, external<br />
review, by dis<strong>in</strong>terested ‘experts’. Whilst this sounds<br />
easy, <strong>the</strong> pool of appropriate reviewers can be small<br />
and, from a manager’s perspective, plans can get bogged<br />
down <strong>in</strong> what appear to be petty, academic debates.<br />
2. Simple guidel<strong>in</strong>es could be established for evaluat<strong>in</strong>g<br />
<strong>the</strong> uncerta<strong>in</strong>ty of a conceptual model based on <strong>the</strong> type<br />
and strength of evidence provided. Evidence <strong>in</strong> <strong>the</strong><br />
form of a ma<strong>the</strong>matical expression of <strong>the</strong> conceptual<br />
model tested on long term geomorphic data should<br />
provide more confi dence <strong>in</strong> <strong>the</strong> model than a model<br />
based on anecdotal accounts. The guidel<strong>in</strong>es could be<br />
applied by <strong>the</strong> geomorphologists develop<strong>in</strong>g <strong>the</strong> model<br />
to advise managers on model uncerta<strong>in</strong>ty. It would<br />
also be possible to use <strong>the</strong> guidel<strong>in</strong>es to recommend<br />
additional <strong>in</strong>vestigations that could reduce model<br />
uncerta<strong>in</strong>ty. Alternate conceptual models could also be<br />
compared us<strong>in</strong>g <strong>the</strong>se guidel<strong>in</strong>es.
68 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
3. A rule-of-thumb <strong>in</strong> eng<strong>in</strong>eer<strong>in</strong>g projects is that about<br />
10% of <strong>the</strong> total cost of a project should be spent on<br />
design. Given that stream restoration projects tend to<br />
cost AUS$10 5 –$10 6 , perhaps a good guide is to argue<br />
that all of <strong>the</strong> design would be cost<strong>in</strong>g around $US<br />
10 000 to $US 100 000, with <strong>the</strong> conceptual model<br />
absorb<strong>in</strong>g from 20–40% of that amount. This might<br />
avoid <strong>the</strong> situation described <strong>in</strong> <strong>the</strong> Snowy <strong>River</strong><br />
Project, where a simple/elegant conceptual model<br />
based on a modest study sets <strong>in</strong> tra<strong>in</strong> a management<br />
response that it is hard to modify. A problem with this<br />
approach is that <strong>the</strong> geomorphic <strong>in</strong>vestigation that<br />
establishes <strong>the</strong> conceptual model may be commissioned<br />
before <strong>the</strong> full costs of <strong>the</strong> restoration project<br />
are known. In this case, additional geomorphic <strong>in</strong>vestigations<br />
may be commissioned later <strong>in</strong> <strong>the</strong> plann<strong>in</strong>g of<br />
<strong>the</strong> project after prelim<strong>in</strong>ary estimates of project costs.<br />
The challenge is to avoid advanc<strong>in</strong>g too far with <strong>the</strong><br />
design before <strong>the</strong> conceptual model is fi nalised.<br />
4. Modellers often develop <strong>the</strong> model with one set of data<br />
and test it on ano<strong>the</strong>r set. Thus, some data are held back<br />
for verifi cation. A similar approach could be used with<br />
conceptual models. For example, <strong>in</strong> <strong>the</strong> Snowy <strong>River</strong><br />
project, a proportion of aerial photographs could have<br />
been held back and used to verify <strong>the</strong> model later<br />
(given that <strong>the</strong>re are eight sets of photos, this may be<br />
possible).<br />
5. All of <strong>the</strong> data and <strong>in</strong>formation could be collected and<br />
collated and <strong>the</strong>n passed onto an appropriate third party<br />
(or more than one), without <strong>in</strong>terpretation, so that an<br />
<strong>in</strong>dependent <strong>in</strong>terpretation of <strong>the</strong> data could be developed.<br />
The diffi culty with this is that <strong>the</strong> conceptual<br />
model developed by <strong>the</strong> third party may require a different<br />
type of data for test<strong>in</strong>g, which would require<br />
fur<strong>the</strong>r data collation, add<strong>in</strong>g substantially to project<br />
costs.<br />
5.3 CASE STUDY OF A GEOMORPHIC<br />
DESIGN MODEL<br />
5.3.1 Introduction<br />
So far, <strong>the</strong> role of geomorphologists <strong>in</strong> us<strong>in</strong>g conceptual<br />
models to identify ‘reference’ states and actions that can<br />
move a stream toward <strong>the</strong> ‘reference’ state have been discussed.<br />
The conceptual model will help to identify a restoration<br />
action, but a ma<strong>the</strong>matical model can be required<br />
to design specifi c aspects of <strong>the</strong> <strong>in</strong>tervention. The most<br />
common restoration <strong>in</strong>terventions recommended <strong>in</strong> <strong>the</strong><br />
nor<strong>the</strong>rn hemisphere relate to structural and fl ow changes<br />
that will improve <strong>the</strong> success of fi sh populations. Hav<strong>in</strong>g<br />
been <strong>in</strong>volved <strong>in</strong> many such projects <strong>in</strong> Australia, we have<br />
noted <strong>the</strong> tendency for managers (and ecologists) to<br />
assume that <strong>the</strong> ‘physical stuff’ – <strong>the</strong> hydrology, hydraulics<br />
and geomorphology – is well understood, with low errors<br />
and low uncerta<strong>in</strong>ty. They assume that most uncerta<strong>in</strong>ty<br />
lies <strong>in</strong> <strong>the</strong> biological responses to <strong>in</strong>tervention. The follow<strong>in</strong>g<br />
section suggests that <strong>the</strong>y may be giv<strong>in</strong>g <strong>the</strong> geomorphological<br />
studies we exam<strong>in</strong>e too much credit!<br />
In this section we exam<strong>in</strong>e <strong>the</strong> uncerta<strong>in</strong>ties <strong>in</strong> a ma<strong>the</strong>matical<br />
model used to design a restoration action. The<br />
case study concerns <strong>the</strong> design of fl ush<strong>in</strong>g-fl ows for a<br />
gravel-bed stream. Numerous authors have identifi ed <strong>the</strong><br />
problem of fi ne sediment <strong>in</strong>fi ltration <strong>in</strong>to a stream bed<br />
known as colmation (Sear, 1993; O’Neill and Kuhns,<br />
1994; Milhous, 1995; Kondolf and Wilcock, 1996). The<br />
loss of competent fl ows below dams means that coarse<br />
beds are <strong>in</strong>fi ltrated by fi ne sediment, damag<strong>in</strong>g habitat for<br />
macro<strong>in</strong>vertebrates and fi sh. The only way to clean out <strong>the</strong><br />
bed is to <strong>in</strong>itiate movement of <strong>the</strong> coarse fraction. A<br />
common geomorphic goal of fl ow management is to ‘turnover<br />
<strong>the</strong> bed’. The uncerta<strong>in</strong>ty <strong>in</strong> this question is whe<strong>the</strong>r<br />
<strong>the</strong> predicted fl ow will turn-over <strong>the</strong> bed. The cost of <strong>the</strong><br />
uncerta<strong>in</strong>ty is <strong>the</strong> chance that <strong>the</strong> bed does not move when<br />
<strong>the</strong> target amount of water is released. If <strong>the</strong> bed is not<br />
turned over, <strong>the</strong>n that volume of water has been wasted.<br />
Similarly, if <strong>the</strong> bed moves at a discharge below <strong>the</strong> target<br />
discharge, <strong>the</strong>n <strong>the</strong> extra volume of water released has<br />
been wasted. Thus, <strong>in</strong> this example, <strong>the</strong> uncerta<strong>in</strong>ty can<br />
be expressed <strong>in</strong> terms of <strong>the</strong> extra water that has to be<br />
released to be confi dent that <strong>the</strong> bed will fl ush. This also<br />
allows <strong>the</strong> relative sav<strong>in</strong>g <strong>in</strong> water to be estimated if managers<br />
do various th<strong>in</strong>gs to reduce <strong>the</strong> uncerta<strong>in</strong>ty.<br />
The case study considered is that of fl ush<strong>in</strong>g fl ows for<br />
<strong>the</strong> mid-Goulburn <strong>River</strong>, downstream of <strong>the</strong> Eildon Dam,<br />
<strong>in</strong> nor<strong>the</strong>rn Victoria, Australia. The Goulburn <strong>River</strong> is <strong>the</strong><br />
largest stream <strong>in</strong> Victoria (catchment of 20 000 km 2 ) and<br />
<strong>the</strong> largest Victorian tributary to <strong>the</strong> Murray <strong>River</strong>. It is a<br />
meander<strong>in</strong>g, anabranch<strong>in</strong>g river, that is regulated for irrigation<br />
by <strong>the</strong> Eildon Dam. The bankfull discharge is not<br />
clearly defi ned but fl ow along anabranch channels beg<strong>in</strong>s<br />
at about 120 m 3 /s <strong>in</strong> <strong>the</strong> reach below <strong>the</strong> Eildon Dam.<br />
Abundant data is available for <strong>the</strong> stream, compiled for a<br />
recent environmental fl ow study (DNRE, 2002).<br />
5.3.2 The Geomorphic Model<br />
The volume of water required to deliver <strong>the</strong> fl ush<strong>in</strong>g fl ow<br />
is estimated <strong>in</strong> <strong>the</strong> follow<strong>in</strong>g three stages. It must be<br />
emphasised that this is an approach used for design<strong>in</strong>g a<br />
fl ush<strong>in</strong>g fl ow. This approach is not advocated for use <strong>in</strong><br />
o<strong>the</strong>r studies. Alternate methods for calculat<strong>in</strong>g fl ush<strong>in</strong>g<br />
fl ows might be considered, particulary given <strong>the</strong> magni-
Conceptual and Ma<strong>the</strong>matical Modell<strong>in</strong>g <strong>in</strong> <strong>River</strong> <strong>Restoration</strong>: Do We Have Unreasonable Confi dence? 69<br />
tude of uncerta<strong>in</strong>ties <strong>in</strong> this analysis described later. The<br />
three stages are:<br />
1. Estimat<strong>in</strong>g <strong>the</strong> critical bed shear stress for <strong>in</strong>cipient<br />
motion of <strong>the</strong> bed sediments us<strong>in</strong>g Shields’ entra<strong>in</strong>ment<br />
function; <strong>the</strong> critical shear stress is given by:<br />
τ crit = (γs − γw)d50τ* (5.1)<br />
where γ s and γw are <strong>the</strong> specifi c weights of <strong>the</strong> bed sediment<br />
and water respectively, d 50 is <strong>the</strong> median gra<strong>in</strong> size<br />
of <strong>the</strong> bed sediments and τ* is Shields’ dimensionless<br />
shear stress, often estimated as 0.045 for gravel bed<br />
rivers (Buffi ngton and Montgomery, 1997).<br />
2. Estimat<strong>in</strong>g <strong>the</strong> discharge at which <strong>the</strong> mean shear stress<br />
for <strong>the</strong> reach is equal to <strong>the</strong> critical shear stress for<br />
<strong>in</strong>cipient motion, us<strong>in</strong>g a one-dimensional hydraulic<br />
model (applied over a 2 km reach of <strong>the</strong> Goulburn<br />
<strong>River</strong>); and<br />
3. Estimat<strong>in</strong>g <strong>the</strong> volume of water required to mimic <strong>the</strong><br />
natural duration and frequency of fl ow spells dur<strong>in</strong>g<br />
which bed sediments are mobilised based on an analysis<br />
of a modelled natural fl ow series.<br />
Importantly, a one-dimensional hydraulic model is<br />
used ra<strong>the</strong>r than <strong>the</strong> more sophisticated two- or threedimensional<br />
models now available because it is <strong>the</strong><br />
approach widely used <strong>in</strong> practice for restoration projects.<br />
The uncerta<strong>in</strong>ties associated with a two-dimensional<br />
model will be different, but it is not possible to suggest<br />
whe<strong>the</strong>r <strong>the</strong>y would be smaller or larger than those us<strong>in</strong>g<br />
a one-dimensional analysis without a proper <strong>in</strong>vestigation.<br />
The modell<strong>in</strong>g procedure also <strong>in</strong>cludes <strong>the</strong> preparation of<br />
<strong>in</strong>put data which <strong>in</strong>cludes fi eld surveys and selection of<br />
an appropriate value for Shields’ dimensionless shear<br />
stress (Table 5.1). 0.045 is used as <strong>the</strong> value for Shields’<br />
dimensionless shear stress <strong>in</strong> <strong>the</strong> <strong>in</strong>itial model, but <strong>the</strong><br />
implications of uncerta<strong>in</strong>ty <strong>in</strong> this parameter are also considered.<br />
In this case study <strong>the</strong> concern is not so much with<br />
<strong>the</strong> correct value of this parameter as <strong>the</strong> effect of parameter<br />
uncerta<strong>in</strong>ty.<br />
Us<strong>in</strong>g a standard Wolman count (sample size = 100),<br />
<strong>the</strong> median bed material sediment size is estimated as<br />
30 mm. Us<strong>in</strong>g Shields’ entra<strong>in</strong>ment function, <strong>in</strong>cipient<br />
motion for <strong>the</strong> gra<strong>in</strong> size occurs at 22 N/m 2 . Seventeen<br />
cross-sections are surveyed along a 2 km reach of <strong>the</strong> mid-<br />
Goulburn river. A one-dimensional hydraulic model is<br />
calibrated for <strong>the</strong>se cross-sections us<strong>in</strong>g water levels surveyed<br />
at a fl ow close to <strong>the</strong> mean. Calibration was achieved<br />
by adjust<strong>in</strong>g <strong>the</strong> Mann<strong>in</strong>g n roughness parameter. Assum<strong>in</strong>g<br />
Mann<strong>in</strong>g n is <strong>in</strong>variant with discharge, <strong>the</strong> hydraulic<br />
model estimates that <strong>the</strong> mean shear stress for <strong>the</strong> reach<br />
is equal to 22 N/m 2 at a fl ow of 360 m 3 /s.<br />
To estimate <strong>the</strong> natural frequency of bed mobilisation<br />
events <strong>the</strong> natural fl ow regime is needed, but <strong>the</strong>re is no<br />
record of <strong>the</strong> natural fl ow regime <strong>in</strong> <strong>the</strong> mid-Goulburn<br />
<strong>River</strong> because it has been regulated s<strong>in</strong>ce fl ow gaug<strong>in</strong>g<br />
commenced. Instead a natural fl ow series is modelled<br />
us<strong>in</strong>g available streamfl ow data upstream of Eildon Dam.<br />
This requires estimation of fl ows from ungauged portions<br />
of <strong>the</strong> catchment by scal<strong>in</strong>g fl ows <strong>in</strong> <strong>the</strong> gauged catchments,<br />
comb<strong>in</strong><strong>in</strong>g fl ows from <strong>the</strong> various sub-catchments<br />
and rout<strong>in</strong>g fl ows to <strong>the</strong> study reach. Us<strong>in</strong>g this modelled<br />
natural fl ow series, it is estimated that 157 GL/year is<br />
required to mimic <strong>the</strong> natural frequency and duration of<br />
fl ows exceed<strong>in</strong>g 360 m 3 /s.<br />
5.3.3 Analysis of Uncerta<strong>in</strong>ties<br />
There are a number of sources of uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong> estimated<br />
channel fl ush<strong>in</strong>g discharge and volume of water<br />
required to mimic <strong>the</strong> natural frequency and duration of<br />
channel fl ush<strong>in</strong>g events (Table 5.1). In this study, uncerta<strong>in</strong>ty<br />
<strong>in</strong> <strong>the</strong> estimated threshold discharge for sediment<br />
fl ush<strong>in</strong>g is assessed based on a consideration of:<br />
Table 5.1 Summary of errors and uncerta<strong>in</strong>ties considered <strong>in</strong> this analysis (<strong>the</strong> text <strong>in</strong> italics shows <strong>the</strong> values used <strong>in</strong> <strong>the</strong><br />
analysis)<br />
Source of Error Channel Hydraulics Critical Shear Flow Regime<br />
Sample uncerta<strong>in</strong>ty Cross-section sampl<strong>in</strong>g Number of particles <strong>in</strong> sample Number of years <strong>in</strong> <strong>the</strong> record<br />
(used n = 17)<br />
(used n = 100)<br />
(used n = 25)<br />
Measurement error (Neglected here) Measurement of gra<strong>in</strong><br />
diameter (± 5 mm)<br />
Error <strong>in</strong> rat<strong>in</strong>g curve (r2 = 0.95)<br />
Model error Mann<strong>in</strong>g n (Based on data Shields entra<strong>in</strong>ment function* Estimat<strong>in</strong>g fl ow from ungauged<br />
presented <strong>in</strong> Hicks and<br />
catchments (20% of catchment<br />
Mason, 1998)<br />
area) and fl ow rout<strong>in</strong>g*<br />
* see text for explanation.
70 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
• errors <strong>in</strong> measurement of gra<strong>in</strong> sizes;<br />
• sampl<strong>in</strong>g bed particles to estimate median gra<strong>in</strong> size;<br />
• estimat<strong>in</strong>g Shields’ dimensionless shear stress (τ*).<br />
• assum<strong>in</strong>g <strong>the</strong> channel roughness parameter (i.e. Mann<strong>in</strong>g<br />
n) is constant with discharge;<br />
• sampl<strong>in</strong>g of cross-sections from <strong>the</strong> 2 km reach.<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>the</strong> volume of water to mimic <strong>the</strong> duration<br />
and frequency of sediment fl ush<strong>in</strong>g fl ows is assessed based<br />
on a consideration of <strong>the</strong>se same factors <strong>in</strong> addition to:<br />
• errors <strong>in</strong> measur<strong>in</strong>g discharge us<strong>in</strong>g a rat<strong>in</strong>g curve at<br />
each gauge;<br />
• us<strong>in</strong>g a sample of <strong>the</strong> fl ow record to represent <strong>the</strong> long<br />
term fl ow regime;<br />
• errors <strong>in</strong> modell<strong>in</strong>g natural fl ows at <strong>the</strong> survey site.<br />
In this section <strong>the</strong> effect of <strong>the</strong>se uncerta<strong>in</strong>ties is quantifi<br />
ed us<strong>in</strong>g a Monte Carlo Analysis, which <strong>in</strong>volves runn<strong>in</strong>g<br />
<strong>the</strong> fl ush<strong>in</strong>g fl ow model many times (1000 <strong>in</strong> this case)<br />
with different, but equally plausible sets of <strong>in</strong>put parameters<br />
to generate a range of plausible model outputs (Manly,<br />
1997). For each of <strong>the</strong> 1000 replicates values were randomly<br />
chosen for each of <strong>the</strong> <strong>in</strong>put parameters from <strong>the</strong><br />
range of possible values. These were selected from a probability<br />
distribution centred on <strong>the</strong> best estimate used <strong>in</strong> <strong>the</strong><br />
orig<strong>in</strong>al analysis (described <strong>in</strong> <strong>the</strong> previous section). The<br />
comb<strong>in</strong>ed uncerta<strong>in</strong>ty is estimated by comb<strong>in</strong><strong>in</strong>g random<br />
selections of values for each <strong>in</strong>put variable.<br />
It is necessary to evaluate <strong>the</strong> uncerta<strong>in</strong>ty for each<br />
<strong>in</strong>put parameter so that it is known how it should be<br />
varied <strong>in</strong> <strong>the</strong> replicate model runs. A highly uncerta<strong>in</strong><br />
parameter should be varied over a bigger range than a more<br />
accurately known parameter. The methods used to evaluate<br />
parameter uncerta<strong>in</strong>ties are described below. In some<br />
cases, <strong>the</strong>se methods are obvious. For <strong>the</strong> case of <strong>the</strong><br />
median particle size, it is relatively straight forward to<br />
express <strong>the</strong> estimate of <strong>the</strong> median gra<strong>in</strong> size as a distribution<br />
of possible values, based on <strong>the</strong> size of <strong>the</strong> sample of<br />
bed gra<strong>in</strong>s. It is not so straightforward to evaluate uncerta<strong>in</strong>ties<br />
<strong>in</strong> recorded discharge series or calibrated values<br />
of <strong>the</strong> Mann<strong>in</strong>g channel roughness parameter (n). A best<br />
effort is made to quantify <strong>the</strong>se uncerta<strong>in</strong>ties, but <strong>the</strong>se are<br />
best described as models of uncerta<strong>in</strong>ty, albeit models<br />
which cannot be truly tested. The result is a stochastic<br />
model where some components are determ<strong>in</strong>istic and some<br />
are random.<br />
Median Particle Size<br />
The median gra<strong>in</strong> size is estimated from a sample of 100<br />
bed particles (a standard Wolman count). To estimate <strong>the</strong><br />
uncerta<strong>in</strong>ty associated with sampl<strong>in</strong>g gra<strong>in</strong> sizes, 1000<br />
replicate samples (each with 100 gra<strong>in</strong> sizes) are syn<strong>the</strong>tically<br />
generated by assum<strong>in</strong>g that <strong>the</strong> natural log of<br />
gra<strong>in</strong> sizes (<strong>in</strong> millimetres) are distributed normally with<br />
a mean of 3.4 and standard deviation of 0.51. This gives<br />
a true median particle size of 30 mm, although sample<br />
medians for each replicate will vary about this value.<br />
Errors <strong>in</strong> measurement of gra<strong>in</strong> size diameter are modelled<br />
as normally distributed with a mean of zero and standard<br />
deviation of 3 mm. This gives 90% confi dence <strong>in</strong>tervals<br />
on gra<strong>in</strong> size measurements of ±5 mm. To represent<br />
measurement errors, each gra<strong>in</strong> size <strong>in</strong> each of <strong>the</strong> 1000<br />
replicate samples is perturbed by add<strong>in</strong>g this random<br />
component.<br />
Entra<strong>in</strong>ment Threshold<br />
Buffi ngton and Montgomery (1997) collated values provided<br />
<strong>in</strong> <strong>the</strong> literature for <strong>the</strong> Shields’ dimensionless shear<br />
stress value for <strong>in</strong>cipient motion. A value of 0.045 is often<br />
used as <strong>the</strong> best estimate of τ* for gravel-bed rivers. The<br />
uncerta<strong>in</strong>ty <strong>in</strong> τ* is estimated from <strong>the</strong> range of values<br />
compiled by Buffi ngton and Montgomery that are likely<br />
to occur <strong>in</strong> gravel-bed streams with Reynolds roughness<br />
number <strong>in</strong> <strong>the</strong> range 25 to 1000 (Figure 5.3). The distribution<br />
of residuals for <strong>the</strong> regression <strong>in</strong> Figure 5.3 was replicated<br />
by assum<strong>in</strong>g that <strong>the</strong> distribution was log-normally<br />
distributed with a standard deviation <strong>in</strong> <strong>the</strong> log-error of<br />
0.03, and randomly generat<strong>in</strong>g 1000 replicate values us<strong>in</strong>g<br />
this error model.<br />
Dimensionless shear stress<br />
0.1<br />
y = 0.017x 0.145<br />
R 2 = 0.150<br />
n = 98<br />
0.01<br />
10 100 1000<br />
Reynolds roughness number<br />
Figure 5.3 Values of Reynolds roughness number (for <strong>the</strong> range<br />
25 to 1000) and dimensionless shear stress obta<strong>in</strong>ed from published<br />
<strong>in</strong>cipient motion studies and collated by Buffi ngton and<br />
Montgomery (1997)
Conceptual and Ma<strong>the</strong>matical Modell<strong>in</strong>g <strong>in</strong> <strong>River</strong> <strong>Restoration</strong>: Do We Have Unreasonable Confi dence? 71<br />
Survey Errors<br />
<strong>Uncerta<strong>in</strong>ty</strong> associated with cross-section survey errors<br />
were neglected <strong>in</strong> this study and likely to be small given<br />
<strong>the</strong> high accuracy of survey equipment relative to o<strong>the</strong>r<br />
sources of uncerta<strong>in</strong>ty. The survey of <strong>the</strong> Goulburn <strong>River</strong><br />
reach <strong>in</strong>cluded a sample of 17 evenly-spaced crosssections<br />
over a reach length of 2 km. There is uncerta<strong>in</strong>ty<br />
<strong>in</strong> <strong>the</strong> estimate of reach-averaged bed shear stress depend<strong>in</strong>g<br />
on how well <strong>the</strong> sampled cross-sections represent <strong>the</strong><br />
variability of <strong>the</strong> reach. To estimate this uncerta<strong>in</strong>ty, replicate<br />
samples of 17 cross-sections were generated us<strong>in</strong>g<br />
a bootstrap procedure (Manly, 1997) <strong>in</strong> which 17 crosssections<br />
were randomly selected from <strong>the</strong> 17 available,<br />
‘with replacement’. This bootstrap procedure was used to<br />
generate 1000 replicate samples.<br />
Errors <strong>in</strong> <strong>the</strong> Hydraulic Model<br />
The critical shear stress was converted <strong>in</strong>to a discharge<br />
us<strong>in</strong>g a one-dimensional hydraulic model. The major<br />
uncerta<strong>in</strong>ty <strong>in</strong> this method is <strong>the</strong> roughness coeffi cient. In<br />
<strong>the</strong> model, Mann<strong>in</strong>g n was calculated from a s<strong>in</strong>gle stage<br />
discharge measurement <strong>in</strong> <strong>the</strong> reach. The hydraulic model<strong>in</strong>g<br />
required <strong>the</strong> assumption that Mann<strong>in</strong>g n was <strong>the</strong>n<br />
constant over <strong>the</strong> range of discharges considered (up to<br />
bankfull). It is well known that Mann<strong>in</strong>g n varies with<br />
discharge, sometimes <strong>in</strong>creas<strong>in</strong>g and sometimes decreas<strong>in</strong>g<br />
with <strong>in</strong>creas<strong>in</strong>g fl ow. The error associated with this<br />
assumption was estimated us<strong>in</strong>g <strong>the</strong> large range of roughness<br />
estimates for gravel-bed New Zealand rivers (Hicks<br />
and Mason, 1998). These data reveal that <strong>the</strong> <strong>in</strong>verse of<br />
Mann<strong>in</strong>g roughness parameter ( 1 – n) is approximately proportional<br />
to <strong>the</strong> log of discharge <strong>in</strong> most cases (Stewardson<br />
and Anderson, 2002). A log regression was fi tted to data<br />
for each of 72 sites [provided by Hicks and Mason, 1998]<br />
to provide a constant (c) and coeffi cient (k) <strong>in</strong> this regression<br />
equation for each site:<br />
1<br />
c 1 k<br />
n<br />
Q ⎛ ⎞<br />
= ⎜ + ln − ⎟<br />
(5.2)<br />
⎝ Q ⎠<br />
where Q – is <strong>the</strong> mean daily fl ow at <strong>the</strong> site. About one<br />
quarter of <strong>the</strong> coeffi cients (k) were negative, <strong>in</strong>dicat<strong>in</strong>g<br />
<strong>in</strong>creas<strong>in</strong>g Mann<strong>in</strong>g n with discharge at <strong>the</strong>se sites. The<br />
parameter k defi nes variation <strong>in</strong> Mann<strong>in</strong>g n with discharge.<br />
For our site, c was determ<strong>in</strong>ed by calibration to <strong>the</strong><br />
observed water surface profi le (us<strong>in</strong>g k = 0). The change<br />
<strong>in</strong> shear stress with discharge was <strong>the</strong>n modelled us<strong>in</strong>g <strong>the</strong><br />
one-dimensional model, and run 1000 times, each time<br />
us<strong>in</strong>g a different value of k selected at random from <strong>the</strong><br />
72 values for <strong>the</strong> New Zealand rivers.<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>the</strong> Hydrology<br />
Flows at <strong>the</strong> Goulburn <strong>River</strong> site are regulated by operation<br />
of Lake Eildon, a large water supply reservoir,<br />
upstream of <strong>the</strong> site. Flow data from gaug<strong>in</strong>g stations<br />
located on n<strong>in</strong>e unregulated tributaries upstream of <strong>the</strong><br />
Goulburn <strong>River</strong> site were used to model natural daily fl ows<br />
at this site (i.e. fl ows that would occur <strong>in</strong> <strong>the</strong> absence of<br />
Lake Eildon, neglect<strong>in</strong>g rout<strong>in</strong>g effects). A 25 year series<br />
of daily fl ows was generated for <strong>the</strong> survey site from <strong>the</strong><br />
25 year fl ow records at n<strong>in</strong>e streamfl ow gauges. These<br />
25 year series are a sample of <strong>the</strong> long term fl ow regime.<br />
A bootstrap procedure was used to generate 1000 replicate<br />
25-year series by sampl<strong>in</strong>g random years from <strong>the</strong> 25 year<br />
sequences. Variability <strong>in</strong> <strong>the</strong> replicates represents uncerta<strong>in</strong>ty<br />
<strong>in</strong> <strong>the</strong> long term fl ow regime. This approach assumes<br />
fl ow <strong>in</strong>dependence between years.<br />
Flow modell<strong>in</strong>g for <strong>the</strong> study was based on records of<br />
discharge at n<strong>in</strong>e streamfl ow gauges on unregulated tributaries<br />
of <strong>the</strong> Goulburn <strong>River</strong>. Discharge is estimated at<br />
<strong>the</strong> gauges from rat<strong>in</strong>g curves. Rat<strong>in</strong>g curves are fi tted to<br />
periodic measurements of discharge and stage. There are<br />
errors <strong>in</strong> discharges provided by <strong>the</strong> rat<strong>in</strong>g curve as a<br />
consequence of errors <strong>in</strong> <strong>the</strong>se periodic measurements of<br />
stage and discharge. Rat<strong>in</strong>g curves are often fi tted by a<br />
log–log regression and an r2 of 0.95 is generally regarded<br />
as a good fi t (Clarke, 1999). Error <strong>in</strong> discharges derived<br />
from <strong>the</strong> rat<strong>in</strong>g curves are represented by a random component<br />
added to <strong>the</strong> log-discharge where <strong>the</strong> error is distributed<br />
normally with standard deviation of 0.05 times<br />
<strong>the</strong> mean of log-discharges recorded at <strong>the</strong> gauge. This can<br />
be represented by:<br />
Q′= e<br />
ln Q+ N(<br />
0005) ∑ ln Q<br />
n<br />
1<br />
, .<br />
(5.3)<br />
where Q is <strong>the</strong> recorded daily discharge, N(0,0.05) denotes<br />
a normally distributed random variable with mean of zero<br />
and standard deviation of 0.05, and n is <strong>the</strong> number of days<br />
of recorded fl ow. This error would result <strong>in</strong> an r 2 of 0.95<br />
for a rat<strong>in</strong>g curve fi tted by log–log regression.<br />
The 25 year natural daily fl ow series at <strong>the</strong> study site<br />
was modeled <strong>in</strong> two parts: (i) estimat<strong>in</strong>g fl ows from <strong>the</strong><br />
ungauged portion of <strong>the</strong> catchment by scal<strong>in</strong>g gauged<br />
fl ows <strong>in</strong> unregulated tributaries; and (ii) summ<strong>in</strong>g fl ows<br />
recorded at <strong>the</strong> streamfl ow gauges and estimated for <strong>the</strong><br />
ungauged portions of <strong>the</strong> catchment. Rout<strong>in</strong>g effects,<br />
which generally result <strong>in</strong> attenuation and delay of fl ood<br />
pulses fur<strong>the</strong>r downstream, were ignored because no data<br />
were available to calibrate a rout<strong>in</strong>g model. <strong>Uncerta<strong>in</strong>ty</strong><br />
<strong>in</strong> <strong>the</strong> scal<strong>in</strong>g parameter for <strong>the</strong> ungauged catchment<br />
was evaluated us<strong>in</strong>g a comparison of fl ows recorded at<br />
seven different gauges (see Appendix 5.1). There is some
72 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong> modelled natural fl ows as a consequence<br />
of ignor<strong>in</strong>g rout<strong>in</strong>g effects. This uncerta<strong>in</strong>ty was represented<br />
us<strong>in</strong>g a rout<strong>in</strong>g model (described by Stewardson<br />
and Cott<strong>in</strong>gham, 2002), where <strong>the</strong> rout<strong>in</strong>g parameters<br />
were randomly perturbed accord<strong>in</strong>g to an assumed distribution<br />
of possible values (see Appendix 5.1).<br />
Results of Monte Carlo Analysis<br />
The critical shear stress for <strong>in</strong>cipient motion for this reach<br />
has wide confi dence limits vary<strong>in</strong>g from 18 N/m2 , up to<br />
26 N/m2 (Figure 5.4). A larger uncerta<strong>in</strong>ty applies to <strong>the</strong><br />
Shear stress (N/m 2 )<br />
Flow volume (GL/year)<br />
70<br />
60<br />
50<br />
40<br />
30<br />
20<br />
10<br />
0<br />
800<br />
600<br />
400<br />
200<br />
reach shear stress<br />
critical shear stress<br />
0 200 400 600 800<br />
Flow (m 3 /s)<br />
(a)<br />
0<br />
0 200 400 600 800<br />
Flow threshold (m3 /s)<br />
(b)<br />
Figure 5.4 (a) reach-averaged shear stress estimated us<strong>in</strong>g a<br />
one-dimensional hydraulic model and <strong>the</strong> critical shear stress<br />
estimated us<strong>in</strong>g Shields, entra<strong>in</strong>ment function; and (b) <strong>the</strong> volume<br />
of artifi cial fl ow spells required to mimic <strong>the</strong> frequency and duration<br />
of natural fl ow spells for vary<strong>in</strong>g fl ow threshold. (Dashed<br />
l<strong>in</strong>es <strong>in</strong>dicate 90% confi dence <strong>in</strong>tervals based on consideration of<br />
all errors described <strong>in</strong> Table 5.1.)<br />
possible range of shear stresses that occur <strong>in</strong> <strong>the</strong> reach for<br />
a given discharge. The fl ush<strong>in</strong>g fl ow is estimated to be<br />
360 m 3 /s, but this has 90% confi dence limits of between<br />
250 m 3 /s and well over 500 m 3 /s.<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>the</strong> critical discharge for bed fl ush<strong>in</strong>g is<br />
reported us<strong>in</strong>g <strong>the</strong> magnitude of <strong>the</strong> 90% confi dence <strong>in</strong>terval<br />
(Table 5.2). Comb<strong>in</strong><strong>in</strong>g all sources of uncerta<strong>in</strong>ty, <strong>the</strong><br />
magnitude of this confi dence <strong>in</strong>terval is 560 m 3 /s. The<br />
major source of uncerta<strong>in</strong>ty <strong>in</strong> estimat<strong>in</strong>g <strong>the</strong> critical discharge<br />
comes from error <strong>in</strong> <strong>the</strong> estimate of Mann<strong>in</strong>g n <strong>in</strong><br />
<strong>the</strong> hydraulic model. This is because n is calibrated to a<br />
s<strong>in</strong>gle discharge (effectively <strong>the</strong> mean discharge of 60 m 3 /<br />
s) and becomes <strong>in</strong>creas<strong>in</strong>gly uncerta<strong>in</strong> as discharge<br />
<strong>in</strong>creases. This provides a massive 430 m 3 /s range of discharge<br />
(Table 5.2). By comparison, if <strong>the</strong> only source of<br />
uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong> estimate was error <strong>in</strong> measur<strong>in</strong>g <strong>the</strong><br />
particle size, <strong>the</strong>n <strong>the</strong> range <strong>in</strong> <strong>the</strong> estimate would only be<br />
50 m 3 /s. The implication of Table 5.2 is that ever more<br />
detailed estimates of <strong>the</strong> bed-particle size distribution, or<br />
even <strong>in</strong>creas<strong>in</strong>gly sophisticated bed-load transport threshold<br />
models, will not remove <strong>the</strong> major uncerta<strong>in</strong>ty associated<br />
with <strong>the</strong> channel hydraulics.<br />
In a good environmental fl ow project, a manager would<br />
not simply specify a fl ow magnitude, <strong>the</strong> frequency and<br />
duration of that critical fl ow would also be specifi ed. A<br />
popular approach used to defi ne <strong>the</strong> frequency and duration<br />
is <strong>the</strong> ‘natural fl ow paradigm’ (Poff et al., 1997;<br />
Richter et al., 1997), which suggests that an environmental<br />
fl ow should mimic <strong>the</strong> natural regime. To estimate this <strong>the</strong><br />
average duration and frequency of critical transport events<br />
<strong>in</strong> <strong>the</strong> natural regime is calculated. The volume of water<br />
required to mimic <strong>the</strong>se events is <strong>the</strong> average duration<br />
multiplied by <strong>the</strong> frequency, multiplied by <strong>the</strong> threshold<br />
discharge. The curve <strong>in</strong> Figure 5.4(b) shows that, as <strong>the</strong><br />
fl ow threshold <strong>in</strong>creases, <strong>the</strong> total fl ow required decreases<br />
substantially. If, for example, <strong>the</strong> threshold fl ush<strong>in</strong>g fl ow<br />
Table 5.2 Size of 90% confi dence <strong>in</strong>terval (expressed <strong>in</strong> m 3 /s)<br />
for threshold discharge for <strong>in</strong>cipient motion for different<br />
sources of error (best estimate of threshold discharge is<br />
360 m 3 /s)<br />
Source of error Channel Hydraulics Critical Shear<br />
Measurement error Neglected 50<br />
Model error 430 180<br />
Sample error 200 120<br />
Comb<strong>in</strong>ed 540 220<br />
All sources comb<strong>in</strong>ed 560<br />
Note: The numbers are not additive. The comb<strong>in</strong>ed uncerta<strong>in</strong>ty comes<br />
from MonteCarlo simulations us<strong>in</strong>g <strong>the</strong> full distributions. This table<br />
should be compared with Table 5.1.
Conceptual and Ma<strong>the</strong>matical Modell<strong>in</strong>g <strong>in</strong> <strong>River</strong> <strong>Restoration</strong>: Do We Have Unreasonable Confi dence? 73<br />
Table 5.3 Size of 90% confi dence <strong>in</strong>terval (<strong>in</strong> GL/year) for volume of discharge required to mimic natural frequency and duration<br />
of bed scour<strong>in</strong>g events for different sources of error (best estimate of volume is 157 GL/year)<br />
Source of Error Channel Hydraulics Critical Shear Flow Regime<br />
Measurement error neglected 37 20<br />
Model error 234 167 13<br />
Sample error 164 104 107<br />
Comb<strong>in</strong>ed 267 209 125<br />
All sources comb<strong>in</strong>ed 350<br />
is 260 m 3 /s, <strong>the</strong>n to generate <strong>the</strong> natural duration and frequency<br />
of <strong>the</strong>se fl ow pulses would require about 160 GL/<br />
yr, but it could range between 10 GL/year and 360 GL/yr<br />
(i.e. <strong>the</strong> 90th percentile confi dence <strong>in</strong>terval).<br />
The numbers <strong>in</strong> Table 5.3 are <strong>the</strong> confi dence <strong>in</strong>tervals on<br />
this volume of <strong>the</strong> fl ush<strong>in</strong>g fl ow. For example, consider<strong>in</strong>g<br />
all sources of error, <strong>the</strong> 90% confi dence limits are 10 GL/<br />
year and 360 GL/year so <strong>the</strong> magnitude of <strong>the</strong> <strong>in</strong>terval is<br />
350 GL/year which is close to 10% of <strong>the</strong> mean annual<br />
fl ow! Expressed ano<strong>the</strong>r way, 360 GL have to be allocated<br />
for <strong>the</strong> fl ow component <strong>in</strong> order to be 95% confi dent that<br />
<strong>the</strong>re is suffi cient water that <strong>the</strong> bed will be turned over at<br />
a natural frequency and duration, given all of <strong>the</strong> sources of<br />
uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong> method. How much is this water worth?<br />
One way to estimate this is to consider <strong>the</strong> cost of <strong>in</strong>frastructure<br />
projects that would be required to achieve comparable<br />
water sav<strong>in</strong>gs (pipel<strong>in</strong>es etc.). The Snowy Water<br />
Inquiry (1998a,b) estimated that environmental fl ows,<br />
costed <strong>in</strong> this way, were worth $US 0.45 million per GL/<br />
year. Clearly this uncerta<strong>in</strong>ty has economic signifi cance.<br />
5.3.4 Discussion<br />
In this simple analysis three o<strong>the</strong>r sources of uncerta<strong>in</strong>ty<br />
have not been considered. The fi rst is <strong>the</strong> simplifi cation<br />
of <strong>the</strong> fl ush<strong>in</strong>g fl ow problem to a s<strong>in</strong>gle threshold shear<br />
stress. The range of τ* used accounts for many of <strong>the</strong><br />
well-known problems of hid<strong>in</strong>g, pack<strong>in</strong>g and o<strong>the</strong>r particle<br />
<strong>in</strong>teractions that affect bed material transport, but <strong>the</strong>re is<br />
still <strong>the</strong> issue of hysteresis. The same discharge on <strong>the</strong><br />
ris<strong>in</strong>g and fall<strong>in</strong>g limb of <strong>the</strong> hydrograph has very different<br />
transport<strong>in</strong>g capacity (Gomez, 1991). It is also possible<br />
that a fl ow exceed<strong>in</strong>g <strong>the</strong> threshold for <strong>in</strong>cipient motion<br />
will be required to mobilize bulk quantities of <strong>the</strong> bed<br />
sediments.<br />
The second uncerta<strong>in</strong>ty not considered is <strong>the</strong> longitud<strong>in</strong>al<br />
variation <strong>in</strong> <strong>the</strong> effect of <strong>the</strong> fl ush<strong>in</strong>g fl ow. In this case<br />
a threshold fl ow <strong>in</strong> a s<strong>in</strong>gle reach has been estimated. The<br />
target fl ow may well turn <strong>the</strong> bed over at <strong>the</strong> sample site,<br />
but what will <strong>the</strong> fl ow do up and downstream of that site?<br />
Given storage effects, a larger fl ow might have to be<br />
released to turn <strong>the</strong> bed over fur<strong>the</strong>r downstream. The<br />
result might be that <strong>the</strong> upstream site is not just ‘turnedover’<br />
but it is progressively scoured away. This <strong>in</strong>troduces<br />
<strong>the</strong> third uncerta<strong>in</strong>ty which is <strong>the</strong> susta<strong>in</strong>ability of <strong>the</strong><br />
process target.<br />
The fl ush<strong>in</strong>g fl ow may work <strong>the</strong> fi rst time, but <strong>the</strong> bed<br />
will <strong>the</strong>n presumably progressively armour, as <strong>the</strong>re may<br />
be no source of coarse sediment below <strong>the</strong> dam. Ei<strong>the</strong>r <strong>the</strong><br />
fl ush<strong>in</strong>g fl ow will have to be progressively <strong>in</strong>creased over<br />
time, or <strong>the</strong> idea of a fl ush<strong>in</strong>g fl ow is simply not susta<strong>in</strong>able<br />
<strong>in</strong> this situation. Such analyses are completed as<br />
steady-state models when <strong>in</strong> reality <strong>the</strong> release of environmental<br />
fl ows will lead to changes <strong>in</strong> <strong>the</strong> <strong>in</strong>put variables.<br />
An example of this problem is <strong>the</strong> major fl ush<strong>in</strong>g-fl ow<br />
experiment carried out on <strong>the</strong> Colorado <strong>River</strong> (Collier<br />
et al., 1997). The fl ood was successful <strong>in</strong> scour<strong>in</strong>g and<br />
rejuvenat<strong>in</strong>g <strong>the</strong> po<strong>in</strong>t-bars of <strong>the</strong> river below <strong>the</strong> Glen<br />
Canyon Dam but this does not mean that it can be successful<br />
<strong>in</strong>defi nitely.<br />
Major sources of uncerta<strong>in</strong>ty have been considered <strong>in</strong> a<br />
reasonably simple geomorphic problem: a fl ush<strong>in</strong>g fl ow.<br />
Most discussion of uncerta<strong>in</strong>ty <strong>in</strong> this type of analysis is<br />
associated with <strong>the</strong> entra<strong>in</strong>ment function, or <strong>in</strong> <strong>the</strong> measurement<br />
of <strong>the</strong> particle size (Kondolf and Wilcock, 1996).<br />
It is <strong>in</strong>terest<strong>in</strong>g to note <strong>the</strong> many sources of uncerta<strong>in</strong>ty<br />
(and error) <strong>in</strong> such analyses. In fact, for this example, <strong>the</strong><br />
ma<strong>in</strong> uncerta<strong>in</strong>ty comes from <strong>the</strong> hydraulics, particularly<br />
<strong>the</strong> estimation of roughness.<br />
It is pert<strong>in</strong>ent to ask what options are available to reduce<br />
<strong>the</strong> uncerta<strong>in</strong>ty? For a geomorphologist <strong>the</strong> response may<br />
be to argue for <strong>the</strong> use of a more sophisticated bed-load<br />
threshold function, but this will not reduce <strong>the</strong> uncerta<strong>in</strong>ty<br />
associated with <strong>the</strong> hydraulics and hydrology. The geomorphologist<br />
could also argue that more research would<br />
reduce uncerta<strong>in</strong>ty. This may be true, but it is also fair to<br />
say that 150 years of research <strong>in</strong>to bed-load transport<br />
re<strong>in</strong>forces <strong>the</strong> view that <strong>the</strong>re is unlikely to be a simple<br />
pr<strong>in</strong>ciple that will dramatically change our capacity to<br />
predict this process. Instead <strong>the</strong> best option is to make<br />
extra measurements at <strong>the</strong> site. The geomorphologists<br />
could ei<strong>the</strong>r measure variables that reduce <strong>the</strong> range of
74 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
uncerta<strong>in</strong>ty <strong>in</strong> each of <strong>the</strong> modelled variables discussed<br />
above, or <strong>the</strong>y could directly measure bed-load movement<br />
to observe <strong>the</strong> discharge at which <strong>the</strong> bed turns over.<br />
Reduc<strong>in</strong>g Model <strong>Uncerta<strong>in</strong>ty</strong><br />
Extra fi eld measurements can reduce <strong>the</strong> uncerta<strong>in</strong>ty and<br />
so <strong>the</strong> amount of water required for <strong>the</strong> fl ush<strong>in</strong>g fl ow. The<br />
options <strong>in</strong>clude:<br />
• extra cross-section surveys;<br />
• record<strong>in</strong>g stages along <strong>the</strong> reach over <strong>the</strong> range of fl ows<br />
be<strong>in</strong>g considered to avoid <strong>the</strong> need to extrapolate<br />
Mann<strong>in</strong>g n from calibration at a s<strong>in</strong>gle discharge; or<br />
• larger bed material samples and more careful measurement<br />
of particle diameter.<br />
Most of <strong>the</strong> error <strong>in</strong> this example lies <strong>in</strong> estimates<br />
of <strong>the</strong> channel hydraulics (Table 5.4). The most effective<br />
way to reduce uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong>se modelled estimates<br />
is to survey water levels over a range of discharges to<br />
allow calibration of Mann<strong>in</strong>g n for a range of fl ows. Some<br />
uncerta<strong>in</strong>ty would rema<strong>in</strong> for extrapolations outside this<br />
range of discharges. Similarly, survey<strong>in</strong>g more cross-sections<br />
will produce a greater decrease <strong>in</strong> error than will a<br />
more accurate estimate of <strong>the</strong> entra<strong>in</strong>ment function.<br />
Overall, <strong>the</strong> estimates are <strong>in</strong>sensitive to uncerta<strong>in</strong>ties <strong>in</strong><br />
<strong>the</strong> fl ow regime, such as an <strong>in</strong>crease <strong>in</strong> <strong>the</strong> length of fl ow<br />
record, or improvements <strong>in</strong> <strong>the</strong> accuracy of <strong>the</strong> rat<strong>in</strong>g<br />
curves.<br />
Ano<strong>the</strong>r way to reduce uncerta<strong>in</strong>ty (and <strong>the</strong> waste of<br />
water) is not to mimic a ‘natural’ fl ow regime. The uncerta<strong>in</strong>ty<br />
(and cost) that this can <strong>in</strong>troduce must be appreciated.<br />
A reductionist approach is to understand <strong>the</strong> target<br />
processes, and release fl ows when required, to achieve this<br />
goal. In <strong>the</strong> case of a fl ush<strong>in</strong>g fl ow, it would be better to<br />
monitor <strong>the</strong> bed processes so that managers knew when<br />
<strong>the</strong> bed needed to be fl ushed (say after a particular series<br />
of lower fl ows). Thus, <strong>the</strong> cost of lost fl ow would be<br />
replaced by <strong>the</strong> cost of monitor<strong>in</strong>g <strong>the</strong> bed.<br />
A clear conclusion from this analysis is that a few extra<br />
measurements can dramatically reduce uncerta<strong>in</strong>ty. In this<br />
particular case, it may be cheaper and more accurate to<br />
simply observe <strong>the</strong> processes directly, ra<strong>the</strong>r than rely on<br />
modell<strong>in</strong>g. The most effi cient way to achieve this is by<br />
trial releases of water from dams (<strong>in</strong> <strong>the</strong> fl ush<strong>in</strong>g-fl ow<br />
case). Dam managers may argue that <strong>the</strong> water is too<br />
valuable to ‘waste’ on such exercises. However, at least<br />
<strong>in</strong> our example, far more water (and hence money) could<br />
be wasted as a consequence of <strong>the</strong> uncerta<strong>in</strong>ty.<br />
However, river restoration projects are often carried out<br />
under tight time constra<strong>in</strong>ts for political or economic<br />
reasons. There is rarely time to monitor <strong>the</strong> processes that<br />
underp<strong>in</strong> <strong>the</strong> modell<strong>in</strong>g or to check <strong>the</strong> predictions. In a<br />
typical stream rehabilitation project, a consultant is<br />
engaged and given perhaps a month or two to report. Any<br />
fi eld <strong>in</strong>vestigations are expected to be brief <strong>in</strong>spections to<br />
develop a conceptual model.<br />
5.4 CONCLUDING DISCUSSION<br />
A central contribution of geomorphology to <strong>the</strong> new practice<br />
of stream restoration is <strong>in</strong> develop<strong>in</strong>g conceptual<br />
models that describe <strong>the</strong> change of stream form and<br />
process over time. This exercise also provides a ‘reference’<br />
condition that can be <strong>the</strong> target for restoration<br />
actions. If <strong>the</strong> conceptual model is wrong, or is too simplistic,<br />
<strong>the</strong>n it does not matter how well executed <strong>the</strong><br />
management actions are, <strong>the</strong> project outcomes rema<strong>in</strong><br />
highly uncerta<strong>in</strong>.<br />
In this chapter <strong>the</strong> development of <strong>the</strong> conceptual geomorphic<br />
model for <strong>the</strong> restoration of <strong>the</strong> lower Snowy<br />
<strong>River</strong> has been described. This is an <strong>in</strong>terest<strong>in</strong>g example<br />
because it illustrates <strong>the</strong> tension between <strong>the</strong> typically<br />
iterative process of geomorphic discovery and <strong>the</strong> need<br />
for action by managers. It also shows <strong>the</strong> dangers of too<br />
much confi dence <strong>in</strong> early <strong>in</strong>terpretations, when <strong>the</strong>se can<br />
trigger expensive management actions that are diffi cult to<br />
reverse even with adaptive management (Stankey et al.,<br />
2003).<br />
<strong>Restoration</strong> can be a very expensive exercise but <strong>the</strong>re<br />
is often little basis for managers to assess how much confi<br />
dence <strong>the</strong>y should have <strong>in</strong> conceptual models that are<br />
presented to <strong>the</strong>m. Five methods have been proposed that<br />
Table 5.4 Size of 90% confi dence <strong>in</strong>terval (expressed <strong>in</strong> GL/year) for volume of discharge required to mimic natural frequency<br />
and duration of bed scour<strong>in</strong>g events with one source of uncerta<strong>in</strong>ty removed (e.g. if completely certa<strong>in</strong> about <strong>the</strong> hydraulic<br />
roughness of <strong>the</strong> channel, <strong>the</strong>n <strong>the</strong> confi dence <strong>in</strong>terval would be reduced from 350 GL/yr down to 280 GL/year, as shown <strong>in</strong> <strong>the</strong><br />
bottom left cell of <strong>the</strong> table)<br />
Source of Error Omitted from Analysis Channel Hydraulics Critical Shear Flow Regime<br />
Sample error Cross-sections 280 No. of particles 330 Years of record 340<br />
Measurement error – Gra<strong>in</strong> size 350 Rat<strong>in</strong>g curve 350<br />
Model error Mann<strong>in</strong>g n 270 Shields entra<strong>in</strong>ment 320 Extrapolat<strong>in</strong>g from gauged<br />
catchments 350
Conceptual and Ma<strong>the</strong>matical Modell<strong>in</strong>g <strong>in</strong> <strong>River</strong> <strong>Restoration</strong>: Do We Have Unreasonable Confi dence? 75<br />
can be used to test <strong>the</strong> validity of conceptual models that<br />
often fall outside <strong>the</strong> area of peer review: <strong>in</strong>dependent<br />
review of <strong>the</strong> model; establish<strong>in</strong>g simple guidel<strong>in</strong>es for<br />
evaluat<strong>in</strong>g <strong>the</strong> uncerta<strong>in</strong>ty of a conceptual model based on<br />
<strong>the</strong> type of strength of evidence provided; ensur<strong>in</strong>g that<br />
an appropriate percentage of <strong>the</strong> total cost of <strong>the</strong> project<br />
is committed to <strong>the</strong> development of <strong>the</strong> conceptual model;<br />
hold<strong>in</strong>g back some of <strong>the</strong> <strong>in</strong>formation and data that underp<strong>in</strong><br />
<strong>the</strong> model, and us<strong>in</strong>g <strong>the</strong>se data to verify <strong>the</strong> hypo<strong>the</strong>tical<br />
model; and pass<strong>in</strong>g all of <strong>the</strong> data and <strong>in</strong>formation to<br />
a third party, without <strong>in</strong>terpretation, so that a compet<strong>in</strong>g<br />
model can be developed.<br />
The feasibility has been demonstrated of quantify<strong>in</strong>g<br />
uncerta<strong>in</strong>ty <strong>in</strong> a geomorphic design model, us<strong>in</strong>g <strong>the</strong> case<br />
study of a fl ow to fl ush <strong>the</strong> fi ne sediments from <strong>the</strong> bed of<br />
<strong>the</strong> Goulburn <strong>River</strong>. In this case, <strong>the</strong> uncerta<strong>in</strong>ty was very<br />
large. Surpris<strong>in</strong>gly, <strong>the</strong> ma<strong>in</strong> source of uncerta<strong>in</strong>ty came<br />
from <strong>the</strong> hydraulic modell<strong>in</strong>g and relatively little from<br />
ei<strong>the</strong>r <strong>the</strong> bed-load entra<strong>in</strong>ment function or <strong>the</strong> hydrological<br />
modell<strong>in</strong>g of <strong>the</strong> fl ow series. It is also strik<strong>in</strong>g how<br />
add<strong>in</strong>g extra fi eld measurements can reduce <strong>the</strong> uncerta<strong>in</strong>ty<br />
<strong>in</strong> <strong>the</strong> model estimates, particularly <strong>in</strong> estimat<strong>in</strong>g<br />
Mann<strong>in</strong>g n. In fact, <strong>the</strong> key to uncerta<strong>in</strong>ty often comes<br />
down to modell<strong>in</strong>g versus monitor<strong>in</strong>g.<br />
It is necessary to <strong>in</strong>clude some form of data ga<strong>the</strong>r<strong>in</strong>g<br />
with any geomorphic modell<strong>in</strong>g exercises, <strong>in</strong>clud<strong>in</strong>g<br />
modell<strong>in</strong>g geomorphic response to river restoration. It has<br />
been shown how an uncerta<strong>in</strong>ty analysis might be used to<br />
direct this data ga<strong>the</strong>r<strong>in</strong>g effort to most effectively reduce<br />
uncerta<strong>in</strong>ty <strong>in</strong> model predictions. It seems logical that<br />
modell<strong>in</strong>g and data ga<strong>the</strong>r<strong>in</strong>g should be <strong>in</strong>tegrated to m<strong>in</strong>imise<br />
uncerta<strong>in</strong>ties <strong>in</strong> restoration. In many cases, some<br />
form of monitor<strong>in</strong>g is carried out as part of <strong>the</strong> restoration<br />
implementation to check that <strong>the</strong> project achieves what<br />
was <strong>in</strong>tended. However, <strong>the</strong>se monitor<strong>in</strong>g programs are<br />
rarely designed to verify or improve <strong>the</strong> models used <strong>in</strong><br />
plann<strong>in</strong>g <strong>the</strong> restoration. <strong>Uncerta<strong>in</strong>ty</strong> analysis appears to<br />
be useful for optimis<strong>in</strong>g such monitor<strong>in</strong>g programs to<br />
provide improved models for subsequent restoration<br />
decisions.<br />
In some cases <strong>the</strong> most effective way to reduce uncerta<strong>in</strong>ties<br />
is to run trial programs <strong>in</strong> <strong>the</strong> actual systems be<strong>in</strong>g<br />
restored. Even <strong>in</strong> <strong>the</strong>se cases, <strong>the</strong> model and uncerta<strong>in</strong>ty<br />
analysis can provide <strong>the</strong> basis for design<strong>in</strong>g <strong>the</strong> trial to<br />
ensure that it tackles <strong>the</strong> key sources of uncerta<strong>in</strong>ty <strong>in</strong><br />
model predictions. In <strong>the</strong> case of <strong>the</strong> lower Snowy <strong>River</strong>,<br />
<strong>the</strong> state government has recognised <strong>the</strong> value of additional<br />
geomorphic measurements and analysis, and<br />
<strong>in</strong>vested <strong>in</strong> a restoration trial which <strong>in</strong>cludes fi eld measurements<br />
and physical modell<strong>in</strong>g to improve <strong>the</strong> geomorphic<br />
design. It is possible to design monitor<strong>in</strong>g activities<br />
without regard to uncerta<strong>in</strong>ties dur<strong>in</strong>g <strong>the</strong> plann<strong>in</strong>g phase<br />
of a restoration project. However, it is proposed that<br />
stronger <strong>in</strong>tegration of monitor<strong>in</strong>g and modell<strong>in</strong>g activities<br />
will lead to greater improvements <strong>in</strong> <strong>the</strong> knowledge<br />
underp<strong>in</strong>n<strong>in</strong>g river restoration.<br />
Geomorphic models, like models developed <strong>in</strong> any<br />
science, are subject to some uncerta<strong>in</strong>ty. Although this<br />
uncerta<strong>in</strong>ty is widely acknowledged it is rarely evaluated.<br />
We conclude that <strong>the</strong>re is unreasonable confi dence <strong>in</strong><br />
geomorphic models used for river restoration. Certa<strong>in</strong>ly<br />
we were surprised by <strong>the</strong> magnitude of uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong><br />
fl ush<strong>in</strong>g fl ow example and we have been <strong>in</strong>volved <strong>in</strong> many<br />
of <strong>the</strong>se modell<strong>in</strong>g studies <strong>in</strong> Australia. There is rarely any<br />
thought given to <strong>the</strong> best approach to modell<strong>in</strong>g, <strong>in</strong>clud<strong>in</strong>g<br />
<strong>the</strong> calculation of <strong>in</strong>put parameters to m<strong>in</strong>imise uncerta<strong>in</strong>ties<br />
<strong>in</strong> restoration decisions. There is a need to reth<strong>in</strong>k our<br />
approach to geomorphic modell<strong>in</strong>g <strong>in</strong> <strong>the</strong> context of river<br />
restoration, <strong>in</strong> particular <strong>the</strong> benefi ts of evaluat<strong>in</strong>g uncerta<strong>in</strong>ties<br />
<strong>in</strong> both our conceptual and ma<strong>the</strong>matical models.<br />
This requires more expertise, <strong>in</strong>formation and fund<strong>in</strong>g<br />
but <strong>the</strong> result will be more realistic expectations by those<br />
<strong>in</strong>volved <strong>in</strong> <strong>the</strong> restoration project and a more careful<br />
approach to optimis<strong>in</strong>g data ga<strong>the</strong>r<strong>in</strong>g for calculat<strong>in</strong>g<br />
model parameters and verify<strong>in</strong>g model structure.<br />
REFERENCES<br />
Beven K. 2001. Ra<strong>in</strong>fall–Runoff Modell<strong>in</strong>g: The Primer. John<br />
Wiley & Sons Ltd: Chichester.<br />
Brizga SO, F<strong>in</strong>layson BL. 1992. The Snowy <strong>River</strong> Sediment<br />
Study: Investigation <strong>in</strong>to <strong>the</strong> Distribution, Transport and<br />
Sources of Sand <strong>in</strong> <strong>the</strong> Snowy <strong>River</strong> between Lake J<strong>in</strong>dabyne<br />
and Jarrahmond. Department of Water Resources: Melbourne,<br />
Victoria, Australia.<br />
Brizga SO, F<strong>in</strong>layson BL. 1994. Interactions between upland<br />
catchment and lowland rivers: an applied Australian case<br />
study. Geomorphology 9: 189–201.<br />
Brookes A. 1995. Challenges and objectives for geomorphology<br />
<strong>in</strong> UK river management. Earth Surface Processes and Landforms<br />
20: 593–610.<br />
Brookes A, Sear DA. 1996. Geomorphological pr<strong>in</strong>ciples for<br />
restor<strong>in</strong>g channels. In: Brookes A, Shields FD (Eds), <strong>River</strong><br />
Channel <strong>Restoration</strong>: Guid<strong>in</strong>g Pr<strong>in</strong>ciples for Susta<strong>in</strong>able<br />
Projects. John Wiley & Sons Ltd: Chichester; 75–101.<br />
Brooks AP, Brierley GJ, Millar RG. 2003. The long-term control<br />
of vegetation and woody debris on channel and fl oodpla<strong>in</strong><br />
evolution: <strong>in</strong>sights from a paired catchment study <strong>in</strong> sou<strong>the</strong>astern<br />
Australia. Geomorphology 51: 7–29.<br />
Brooks AP, Brierley GJ. 1997. Geomorphic responses of Lower<br />
Bega <strong>River</strong> to catchment disturbance, 1851–1926. Geomorphology<br />
18: 291–304.<br />
Buffi ngton JM, Montgomery DR. 1997. A systematic analysis of<br />
eight decades of <strong>in</strong>cipient motion studies, with special reference<br />
to gravel-bedded rivers. Water Resources Research 33<br />
(8): 1993–2029.<br />
Clarke RT. 1999. <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> <strong>the</strong> estimation of mean annual<br />
fl ood due to rat<strong>in</strong>g curve <strong>in</strong>defi nition. Journal of Hydrology<br />
222: 185–190.
76 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Collier MP, Webb RH, Andrews ED. 1997. Experimental fl ood<strong>in</strong>g<br />
<strong>in</strong> Grand Canyon. Scientifi c American 276: 82–89.<br />
Cooke RV, Reeves RW. 1976. Arroyos and Environmental<br />
Change <strong>in</strong> <strong>the</strong> American Southwest. Oxford Research Studies<br />
<strong>in</strong> Geography, Oxford.<br />
Cullen P. 1989. The Turbulent Boundary Between Science and<br />
Management. Canberra College of Advanced Education:<br />
Canberra, Australia.<br />
DNRE 2002. The FLOWS method: A method for determ<strong>in</strong><strong>in</strong>g<br />
environmental water requirements <strong>in</strong> Victoria Victorian<br />
Department of Natural Resources and Environment,<br />
Australia.<br />
Ersk<strong>in</strong>e WD, Terrazzolo N, Warner RF. 1999. <strong>River</strong> restoration<br />
from <strong>the</strong> hydrogeomorphic impacts of a large hydro-electric<br />
power project: Snowy <strong>River</strong>, Australia. Regulated <strong>River</strong>s 15:<br />
3–24.<br />
Ersk<strong>in</strong>e WD, Tilleard JW. 1997. Formative processes of alternat<strong>in</strong>g,<br />
bank-attached side bars and associated pool-riffl e<br />
sequences on sand-bed streams similar to <strong>the</strong> Snowy <strong>River</strong> at<br />
Jarrahmond, Victoria. Department of Natural Resources and<br />
Environment: Melbourne, Victoria, Australia.<br />
F<strong>in</strong>layson BL, Bird JF. 1989. Initial <strong>in</strong>vestigation <strong>in</strong>to <strong>the</strong> extent<br />
and nature of <strong>the</strong> current sedimentation problem on <strong>the</strong> lower<br />
Snowy <strong>River</strong>. Centre for Environmental Applied Hydrology,<br />
The University of Melbourne: Melbourne, Australia.<br />
Gilbert GK. 1917. Hydraulic m<strong>in</strong><strong>in</strong>g debris <strong>in</strong> <strong>the</strong> Sierra Nevada.<br />
United States Geological Survey Professional Paper 105:<br />
1–154.<br />
Gilvear DJ. 1999. Fluvial geomorphology and river eng<strong>in</strong>eer<strong>in</strong>g:<br />
future roles utiliz<strong>in</strong>g a fl uvial hydrosystems framework. Geomorphology<br />
31: 229–245.<br />
Gippel CJ. 2002. The Victorian Snowy <strong>River</strong>: Review of Historical<br />
Environmental Change and Proposed <strong>Restoration</strong> Options.<br />
East Gippsland Catchment Management Authority: Victoria,<br />
Australia.<br />
Gippel CJ, Anderson BA, Marsh N. 2002. Trial of Snowy <strong>River</strong><br />
<strong>Restoration</strong> Concept Plan: Scop<strong>in</strong>g Study to Review Struc -<br />
tural <strong>Restoration</strong> Options. Report by Fluvial Systems<br />
Pty Ltd, Stockton, to Snowy <strong>River</strong> <strong>Restoration</strong> Project, East<br />
Gippsland Catchment Management Authority: Victoria,<br />
Australia.<br />
Gomez B. 1991. Bedload transport. Earth-Science Reviews 31:<br />
89–132.<br />
Hicks DM, Mason PD. 1998. Roughness characteristics of New<br />
Zealand <strong>River</strong>s, National Institute of Water and Atmospheric<br />
Research, Water Resource Publications: Englewood, New<br />
Zealand.<br />
ID&A 1998. <strong>River</strong> restoration concept plan for <strong>the</strong> Snowy <strong>River</strong><br />
<strong>in</strong> Victoria. Wangaratta, Victoria, Consultants report to East<br />
Gippsland Catchment Management Authority: Victoria,<br />
Australia.<br />
Kondolf GM. 2000. Some suggested guidel<strong>in</strong>es for geomorphic<br />
aspects of anadromous salmonid habitat restoration proposals.<br />
<strong>Restoration</strong> Ecology 8 (1): 48–56.<br />
Kondolf GM, Micheli ER. 1995. Evaluat<strong>in</strong>g stream restoration<br />
projects. Environmental Management 19 (1): 1–15.<br />
Kondolf GM, Smeltzer MW, Railsback SF. 2001. Design and<br />
performance of a channel reconstruction project <strong>in</strong> a coastal<br />
California gravel-bed stream. Environmental Management 28<br />
(6): 761–776.<br />
Kondolf GM, Wilcock PR. 1996. The fl ush<strong>in</strong>g fl ow problem:<br />
defi n<strong>in</strong>g and evaluat<strong>in</strong>g objectives. Water Resources Research<br />
32 (8): 2589–2599.<br />
Kuhnle RA, Alonso CV, Shields FD. 2002. Local scour associated<br />
with angled spur dikes. Journal of Hydraulic Eng<strong>in</strong>eer<strong>in</strong>g<br />
128 (12): 1087–1093.<br />
Ladson AR, Argent RM. 2002. Adaptive management of environmental<br />
fl ows: lessons for <strong>the</strong> Murray-Darl<strong>in</strong>g Bas<strong>in</strong> from<br />
three large North American <strong>River</strong>s. Australian Journal of<br />
Water Resources 5 (1): 89–102<br />
Larson M, Goldsmith W. 1997. Incised channel stabilization and<br />
enhancement <strong>in</strong>tegrat<strong>in</strong>g geomorphology and bioeng<strong>in</strong>eer<strong>in</strong>g.<br />
In: Wang SSY, Langendoen EJ, Shields FD (Eds), Management<br />
of Landscapes Disturbed by Channel Incision, University<br />
of Mississippi: Oxford, Mississippi; 458–465.<br />
Manly BFJ. 1997. Randomization, Bootstrap and Monte Carlo<br />
Methods <strong>in</strong> Biology. Chapman and Hall: London.<br />
Milhous RT. 1995. Flush<strong>in</strong>g fl ows for habitat restoration. In:<br />
Espey W (Ed) Water Resources Eng<strong>in</strong>eer<strong>in</strong>g, American<br />
Society of Civil Eng<strong>in</strong>eers: San Antonio, Texas; 663–667.<br />
Miller JR, Ritter JB. 1996. An exam<strong>in</strong>ation of <strong>the</strong> Rosgen classifi<br />
cation of natural rivers. Catena 27: 295–299.<br />
Neilsen MB. 1996. Lowland stream restoration <strong>in</strong> Denmark. In:<br />
Brookes A, Shields FD (Eds), <strong>River</strong> Channel <strong>Restoration</strong>:<br />
Guid<strong>in</strong>g Pr<strong>in</strong>ciples for Susta<strong>in</strong>able Projects. John Wiley &<br />
Sons Ltd: Chichester; 269–290.<br />
Newbury R, Gaboury M. 1993. Stream analysis and fi sh habitat<br />
design – a fi eld manual. Newbury Hydraulics Ltd: British<br />
Columbia, Canada.<br />
O’Neill MP, Kuhns MR. 1994. Stream bank erosion and fl ush<strong>in</strong>g<br />
fl ows. Stream Notes (July), USDA Forest Service.<br />
Owen R. 1997. The Lower Snowy <strong>River</strong> Revegetation Strategy.<br />
Snowy <strong>River</strong> Improvement Trust: Nungurner Hills Nursery,<br />
Victoria, Australia.<br />
Poff NL, Allan JD, Ba<strong>in</strong> MB et al. 1997. The natural fl ow regime,<br />
a paradigm for river conservation and restoration. BioScience<br />
47: 769–84.<br />
Raadick TA, O’Connor JP. 1997. Fish and Decapod Crustacean<br />
Survey, and Habitat Assessment, of <strong>the</strong> Lower Snowy <strong>River</strong>,<br />
Victoria. Report to <strong>the</strong> Snowy <strong>River</strong> Improvement Trust by <strong>the</strong><br />
Freshwater Ecology Division of <strong>the</strong> Department of Natural<br />
Resources.<br />
Regan HM, Colyvan M, Burgman M. 2002. A taxonomy of<br />
uncerta<strong>in</strong>ty for ecology and conservation biology. Ecological<br />
Applications 12 (2): 618–628.<br />
Richter BD, Baumgartner JV, Wig<strong>in</strong>gton R, Braun DP. 1997.<br />
How much water does a river need? Freshwater Biology 37:<br />
231–249.<br />
Rosgen DL. 1996. Applied <strong>River</strong> Morphology. Wildland Hydrology:<br />
Pagosa Spr<strong>in</strong>gs, Colorado.<br />
Ru<strong>the</strong>rfurd ID. 2001. Storage and movement of slugs of sand <strong>in</strong><br />
a large catchment: develop<strong>in</strong>g a plan to rehabilitate <strong>the</strong> Glenelg
Conceptual and Ma<strong>the</strong>matical Modell<strong>in</strong>g <strong>in</strong> <strong>River</strong> <strong>Restoration</strong>: Do We Have Unreasonable Confi dence? 77<br />
<strong>River</strong>, SE Australia. In: Anthony DJ, Harvey MD, Laronne JB,<br />
Mosley MP (Eds), Apply<strong>in</strong>g Geomorphology to Environmental<br />
Management. Water Resources Publications: Denver, Colorado;<br />
309–332.<br />
Schumm SA, Harvey MD, Watson CC. 1984. Incised Channels:<br />
Morphology, Dynamics and Control. Water Resources Publications:<br />
Littleton, Colorado.<br />
Sear DA. 1993. F<strong>in</strong>e sediment <strong>in</strong>fi ltration <strong>in</strong>to gravel spawn<strong>in</strong>g<br />
beds with<strong>in</strong> a regulated river experienc<strong>in</strong>g fl oods: Ecological<br />
implications for salmonids. Regulated <strong>River</strong>s 8: 373–390.<br />
Sear DA. 1994. <strong>River</strong> restoration and geomorphology. Aquatic<br />
Conservation: Mar<strong>in</strong>e and Freshwater Ecosystems 4:<br />
169–177.<br />
Seddon GS. 1994. Search<strong>in</strong>g for <strong>the</strong> Snowy, an Environmental<br />
History. Allen and Unw<strong>in</strong>: St Leonards, New South Wales.<br />
Simon A. 1989. A model of channel response <strong>in</strong> disturbed<br />
alluvial channels. Earth Surface Processes and Landforms 14:<br />
11–26.<br />
Snowy Water Inquiry 1998a. Snowy Water Inquiry: Draft Options<br />
for Discussion. Snowy Water Inquiry: Sydney, NSW,<br />
Australia.<br />
Snowy Water Inquiry1998b. Appendix of Resource Materials<br />
(Part 2). Snowy Water Inquiry: Sydney, NSW, Australia.<br />
Stankey GH, Bormann, BT, Ryan C. 2003. Adaptive management<br />
and <strong>the</strong> northwest forest plan: Rhetoric and reality.<br />
Journal of Forestry 40: 40–46.<br />
Stewardson MJ. 1998. Pool formation – fl uvial processes, Section<br />
6 <strong>in</strong> ID&A, <strong>River</strong> restoration concept plan for <strong>the</strong> Snowy <strong>River</strong><br />
<strong>in</strong> Victoria. Report to East Gippsland Catchment Management<br />
Authority by ID&A: Wangaratta, Victoria, Australia.<br />
Stewardson MJ, Anderson B. 2002. Variations <strong>in</strong> <strong>the</strong> fl ow resistance<br />
of natural channels with discharge. Proceed<strong>in</strong>gs of <strong>the</strong><br />
Hydrology and Water Resources Symposium. Institution of<br />
Eng<strong>in</strong>eers: Melbourne, Australia.<br />
Stewardson MJ, Cott<strong>in</strong>gham P. 2002. A demonstration of <strong>the</strong><br />
fl ow events method: Environmental fl ow requirements of <strong>the</strong><br />
Broken <strong>River</strong>. Australian Journal of Water Resources 5 (1):<br />
35–48.<br />
Strom HG. 1936. The Flood Problem <strong>in</strong> Gippsland, 2. The Snowy<br />
<strong>River</strong>. State <strong>River</strong>s and Water Supply Commission: Melbourne,<br />
Victoria, Australia.<br />
Volkman JM. 1999. How do you learn from a river? <strong>Manag<strong>in</strong>g</strong><br />
uncerta<strong>in</strong>ty <strong>in</strong> species conservation policy. Wash<strong>in</strong>gton Law<br />
Review 74: 719–762.<br />
Walters C. 1997. Challeges <strong>in</strong> adaptive management of riparian<br />
and coastal ecosystems. Conservation Ecology [onl<strong>in</strong>e] URL<br />
http://www.consecol.org/vol1/iss2/art1.<br />
APPENDIX 5.1<br />
Flows at <strong>the</strong> Goulburn <strong>River</strong> site are regulated by operation<br />
of Lake Eildon, a large water supply reservoir, upstream of<br />
<strong>the</strong> study site. Flow data for gaug<strong>in</strong>g stations located on<br />
n<strong>in</strong>e unregulated tributaries upstream of <strong>the</strong> Goulburn<br />
<strong>River</strong> site were used to model natural daily fl ows at this site<br />
(i.e. fl ows that would occur <strong>in</strong> <strong>the</strong> absence of Lake Eildon).<br />
Seven of <strong>the</strong>se tributaries are upstream of Lake Eildon and<br />
two o<strong>the</strong>r tributaries have confl uences with <strong>the</strong> Goulburn<br />
<strong>River</strong> between Lake Eildon and <strong>the</strong> survey site (Table 5.5).<br />
In total 80% of <strong>the</strong> 4225 km2 catchment at <strong>the</strong> survey site<br />
is upstream of <strong>the</strong>se fl ow gauges (510 km2 upstream Eildon<br />
and 314 km2 between Eildon and <strong>the</strong> study site). Flows <strong>in</strong><br />
<strong>the</strong> ungauged portion of <strong>the</strong> catchments upstream and<br />
downstream of Lake Eildon are estimated by scal<strong>in</strong>g fl ow<br />
<strong>in</strong> <strong>the</strong> gauged portion of <strong>the</strong> catchments upstream and<br />
downstream of Lake Eildon respectively. The scal<strong>in</strong>g<br />
factor was estimated us<strong>in</strong>g a l<strong>in</strong>ear function of <strong>the</strong> ratio of<br />
gauged and ungauged catchment areas, derived from <strong>the</strong><br />
available streamfl ow data (Figure 5.5). Daily fl ows at<br />
<strong>the</strong> survey site are estimated as <strong>the</strong> sum of daily fl ows (for<br />
<strong>the</strong> same day) at <strong>the</strong> upstream gauges and estimated <strong>in</strong> <strong>the</strong><br />
ungauged catchments. Rout<strong>in</strong>g effects (i.e. travel times<br />
and attenuation of fl ood peaks) are neglected because <strong>the</strong>re<br />
Table 5.5 Gauged tributaries of <strong>the</strong> Goulburn <strong>River</strong> upstream of <strong>the</strong> Goulburn <strong>River</strong><br />
survey site<br />
Tributary Catchment area (km 2 )<br />
Upstream of Lake Eildon<br />
Delatite <strong>River</strong> 368<br />
Howqua <strong>River</strong> 368<br />
Jamison <strong>River</strong> 368<br />
Goulburn <strong>River</strong> at Dohertys Infl ow 694<br />
Big <strong>River</strong> 619<br />
Ford Creek 115<br />
Brankeet Creek 121<br />
Between Lake Eildon and <strong>the</strong> surveys site<br />
Rubicon <strong>River</strong> 129<br />
Acheron <strong>River</strong> 619
78 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
is no <strong>in</strong>formation with which to establish a rout<strong>in</strong>g<br />
model.<br />
The regression equation <strong>in</strong> Figure 5.5 is used to estimate<br />
a scal<strong>in</strong>g factor for estimat<strong>in</strong>g fl ows <strong>in</strong> <strong>the</strong> ungauged catchments.<br />
Data po<strong>in</strong>ts <strong>in</strong> this plot show <strong>the</strong> mean ration of fl ow<br />
Statistic of daily flow ratio<br />
0.5<br />
0.45<br />
0.4<br />
0.35<br />
0.3<br />
0.25<br />
0.2<br />
0.15<br />
0.1<br />
0.05<br />
0<br />
y = 1.33x - 0.05<br />
R 2 = 0.87<br />
Table 5.6 Characteristics of <strong>the</strong> ratio of daily fl ows at each of seven streamfl ow gauges to <strong>the</strong> sum of daily fl ows at <strong>the</strong> o<strong>the</strong>r six<br />
gauges (Gauges are all located upstream of Lake Eildon)<br />
Gauge Catchment<br />
Area (km 2 )<br />
y = 0.35x - 0.0037<br />
R 2 = 0.83<br />
0 0.1 0.2 0.3 0.4<br />
Ratio catchment areas<br />
mean of daily flow ratio<br />
standard dev. of daily flow ratio<br />
Figure 5.5 Mean and standard deviation of daily fl ows ratio and<br />
catchment area ratios for seven streamfl ow gauges upstream of<br />
Lake Eildon. Ratios are calculated at each gauge <strong>in</strong> turn by divid<strong>in</strong>g<br />
daily fl ows and catchment area at <strong>the</strong> gauge by <strong>the</strong> sum of<br />
daily fl ows and catchment areas at <strong>the</strong> o<strong>the</strong>r six gauges<br />
Mean Flow<br />
(ML/day)<br />
Catchment<br />
Area Ratio<br />
Mean of<br />
Daily Flow<br />
Ratio<br />
Standard<br />
Deviation of Daily<br />
Flow Ratio<br />
Characteristics for <strong>the</strong> streamfl ow gauges upstream of Lake Eildon<br />
Delatite <strong>River</strong> 368 297 0.16 0.10 0.037 0.66<br />
Howqua <strong>River</strong> 368 477 0.16 0.18 0.051 0.94<br />
Jamison <strong>River</strong> 368 560 0.16 0.21 0.056 0.91<br />
Goulburn <strong>River</strong> 694 882 0.35 0.35 0.097 0.93<br />
Big <strong>River</strong> 619 843 0.30 0.46 0.14 0.93<br />
Ford Creek 115 34 0.045 0.0049 0.011 0.44<br />
Brankeet Creek 121 48 0.048 0.022 0.018 0.56<br />
Characteristics estimated for <strong>the</strong> ungauged catchments us<strong>in</strong>g regression equation <strong>in</strong> Figure 5.5<br />
Upstream of Eildon 510 – 0.19 0.21 0.063 0.9*<br />
Eildon to study site 314 – 0.42 0.51 0.14 0.9*<br />
* selected based on serial correlations derived from streamfl ow gauge data<br />
at one of <strong>the</strong> gauges to sum of fl ow at all o<strong>the</strong>r gauges. In<br />
reality this ratio varies from day to day. The mean, standard<br />
deviation and lag-1 serial correlation for this ratio<br />
were calculated for each of <strong>the</strong> seven sites upstream of<br />
Lake Eildon (Table 5.6). These characteristics were estimated<br />
for <strong>the</strong> ungauged catchments us<strong>in</strong>g <strong>the</strong> regression<br />
equation <strong>in</strong> Figure 5.5. A stochastic model is used to represent<br />
uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong> fl ow estimated <strong>in</strong> <strong>the</strong> ungauged<br />
catchment. This model gives fl ow from <strong>the</strong> ungauged<br />
catchment on <strong>the</strong> ith day of <strong>the</strong> 25 year record as:<br />
Q = [ mr . + ( 1−m) r 1] Q and r = N(<br />
ms , ) (5.4)<br />
ui , i i− ∑<br />
j<br />
ji ,<br />
i<br />
where Qj,i is <strong>the</strong> fl ow at <strong>the</strong> jth gauge site on <strong>the</strong> ith day.<br />
Values of m were calibrated as 0.1 to provide a lag-1 serial<br />
correlation of 0.9 for <strong>the</strong> fl ow ratios <strong>in</strong> ungauged catchments<br />
upstream and downstream of Lake Eildon. The<br />
daily parameter r i is a normally distributed random variable.<br />
The mean (µ) of r i was set equal to <strong>the</strong> mean of daily<br />
fl ow ratios estimated from <strong>the</strong> regression <strong>in</strong> Figure 5.5 The<br />
standard deviation (σ) was adjusted to replicate <strong>the</strong> standard<br />
deviations calculated from <strong>the</strong> regressions <strong>in</strong> Figure<br />
5.5. Standard deviations chosen for upstream and downstream<br />
of Lake Eildon were 0.27 and 0.61 respectively.<br />
Note that <strong>the</strong>se are higher than <strong>the</strong> values estimated<br />
directly from <strong>the</strong> regression equations to account for <strong>the</strong><br />
effect of lag correlation <strong>in</strong> <strong>the</strong> stochastic model. This stochastic<br />
model is used to generate replicate 25-year timeseries<br />
of fl ow from <strong>the</strong> ungauged catchments.<br />
Lag-1 Serial Correlation<br />
of Daily Flow Ratio
6.1 INTRODUCTION: THE CASE<br />
FOR RESTORATION<br />
6<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Riparian and<br />
Floodpla<strong>in</strong> <strong>Restoration</strong><br />
Franc<strong>in</strong>e M.R. Hughes 1 , Timothy Moss 2 and Keith S. Richards 3<br />
1 Department of Life Sciences, Anglia Rusk<strong>in</strong> University, UK<br />
2 Institute for Regional Development and Structural Plann<strong>in</strong>g (IRS), Germany<br />
3 Department of Geography, University of Cambridge, UK<br />
The riparian and fl oodpla<strong>in</strong> zones of rivers are physically<br />
dynamic places, subject to <strong>the</strong> delivery and removal of<br />
water and sediments dur<strong>in</strong>g fl ood events. Ecosystems that<br />
occupy <strong>the</strong>se places have evolved a tolerance to <strong>the</strong>se<br />
natural disturbance processes and many of <strong>the</strong>ir component<br />
species have become dependent on <strong>the</strong>m for completion<br />
of <strong>the</strong>ir life cycles. In many river valleys, riparian and<br />
fl oodpla<strong>in</strong> zones would once have been occupied by forested<br />
ecosystems, composed of a dynamic mosaic of<br />
forest types <strong>in</strong> different successional stages, <strong>in</strong>terspersed<br />
with more open wetland communities with emergent vegetation.<br />
Such fl oodpla<strong>in</strong> forests have high levels of biodiversity<br />
because <strong>the</strong>y are at <strong>the</strong> <strong>in</strong>terface between terrestrial<br />
and lotic ecosystems (Petts, 1990). In addition, <strong>the</strong>y experience<br />
frequent disturbance from fl oods which create <strong>the</strong><br />
conditions for a heterogeneous mosaic of habitats across<br />
<strong>the</strong> fl oodpla<strong>in</strong>, each support<strong>in</strong>g a varied mix of species<br />
(Nilsson et al., 1991a; Hughes et al., 2005). A high plant<br />
species diversity has been recorded on fl oodpla<strong>in</strong>s from<br />
rivers <strong>in</strong> many different bioclimatic zones. For example,<br />
fl oodpla<strong>in</strong>s <strong>in</strong> <strong>the</strong> Amazon bas<strong>in</strong> account for 20% of tree<br />
species diversity (Junk et al., 1989). In <strong>the</strong> Tana <strong>River</strong><br />
fl oodpla<strong>in</strong> forests <strong>in</strong> Kenya, which stretch for only 200 km<br />
of river length and average only 1 km <strong>in</strong> width, 175 woody<br />
plant species, over 250 species of birds and at least 57<br />
species of mammals have been found, <strong>in</strong>clud<strong>in</strong>g two<br />
endemic primates (Medley and Hughes, 1996).<br />
Riparian and fl oodpla<strong>in</strong> ecosystems not only have<br />
<strong>in</strong>tr<strong>in</strong>sic and <strong>in</strong>tangible values associated with <strong>the</strong>se high<br />
levels of biodiversity, and with <strong>the</strong>ir diverse landscape<br />
character, but <strong>the</strong>y can also be valued because <strong>the</strong>y contribute<br />
to timber production and carbon sequestration,<br />
fl oodwater storage, groundwater recharge, pollution<br />
control and even recreation. The importance of riparian<br />
and fl oodpla<strong>in</strong> zones can thus lie <strong>in</strong> this wide range of<br />
natural functions and services that <strong>the</strong>y provide, although<br />
<strong>the</strong>re is uncerta<strong>in</strong>ty about how <strong>the</strong>se can be valued over<br />
time, given <strong>the</strong>ir dynamic and vary<strong>in</strong>g nature. Such valuation<br />
is today <strong>in</strong>creas<strong>in</strong>gly necessary, s<strong>in</strong>ce hydrological<br />
pathways <strong>in</strong> river bas<strong>in</strong>s have often been altered <strong>in</strong>directly<br />
through land-use change and directly through management<br />
of <strong>the</strong> fl ow regime. In downstream fl oodpla<strong>in</strong> zones<br />
<strong>in</strong> particular <strong>the</strong>re have been many eng<strong>in</strong>eered changes to<br />
river channels, lead<strong>in</strong>g to isolation of fl oodpla<strong>in</strong>s from<br />
<strong>the</strong>ir channels and to severe damage or eradication of<br />
fl oodpla<strong>in</strong> ecosystems. All <strong>the</strong>se changes have severely<br />
limited or even destroyed <strong>the</strong> capacity of fl oodpla<strong>in</strong>s to<br />
deliver <strong>the</strong>ir natural functions and services.<br />
The disappearance of <strong>the</strong>se ecosystems from <strong>the</strong> landscape<br />
is poignantly illustrated by <strong>the</strong> case of fl oodpla<strong>in</strong><br />
forests <strong>in</strong> Europe. 90% of <strong>the</strong>se forested ecosystems have<br />
disappeared and rema<strong>in</strong><strong>in</strong>g patches are often <strong>in</strong> critical<br />
condition. They are listed <strong>in</strong> Annexe I of <strong>the</strong> European<br />
Habitats Directive (92/43/EEC, 1992) as a priority habitat<br />
type and are <strong>in</strong>cluded <strong>in</strong> <strong>the</strong> Natura 2000 network of nature<br />
reserves. In western Europe <strong>the</strong>y are more reduced <strong>in</strong><br />
extent than <strong>in</strong> eastern and central Europe where some<br />
<strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat Edited by Stephen Darby and David Sear<br />
© 2008 John Wiley & Sons, Ltd
80 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Figure 6.1 Map of rema<strong>in</strong><strong>in</strong>g European fl oodpla<strong>in</strong> forests (based on data from UNEP – World Conservation Monitor<strong>in</strong>g Centre <strong>in</strong><br />
UNEP–WCMC, 2000 and Girel et al., 2003)<br />
impressive patches rema<strong>in</strong> (Figure 6.1). Even here, <strong>the</strong>ir<br />
extent is not easy to gauge as many of <strong>the</strong> areas marked<br />
as fl oodpla<strong>in</strong> forest have <strong>in</strong> fact been converted to areas<br />
of forestry on fl oodpla<strong>in</strong>s, without any of <strong>the</strong> characteristic<br />
dynamic features of a naturally function<strong>in</strong>g fl oodpla<strong>in</strong><br />
forest and frequently dom<strong>in</strong>ated by non-native species.<br />
For example, <strong>in</strong> Hungary, where fl ood control works have<br />
reduced <strong>the</strong> fl oodpla<strong>in</strong> area across all river systems from<br />
2.3 billion km 2 to only 1500 km 2 , 40% of <strong>the</strong> rema<strong>in</strong><strong>in</strong>g<br />
areas of fl oodpla<strong>in</strong> forests have been converted to forestry<br />
plantations (Haraszthy, 2001).<br />
6.2 POLICY-RELATED WINDOWS<br />
OF OPPORTUNITY<br />
However, s<strong>in</strong>ce <strong>the</strong> mid-1990s, shifts <strong>in</strong> policy content and<br />
style <strong>in</strong> <strong>the</strong> fi elds of fl ood protection, nature conservation<br />
and agriculture at European Union (EU) and national<br />
levels are creat<strong>in</strong>g ‘w<strong>in</strong>dows of opportunity’ for fl oodpla<strong>in</strong><br />
restoration (Table 6.1). In <strong>the</strong> fi eld of fl ood protection,<br />
recent major fl ood events <strong>in</strong> France, Germany and <strong>the</strong><br />
United K<strong>in</strong>gdom, for <strong>in</strong>stance, have accelerated <strong>the</strong> will<strong>in</strong>gness<br />
of authorities to enterta<strong>in</strong> catchment-oriented<br />
approaches and soft-eng<strong>in</strong>eer<strong>in</strong>g techniques of fl ood pro-<br />
tection, creat<strong>in</strong>g new opportunities for fl oodpla<strong>in</strong> restoration.<br />
The sheer cost of improv<strong>in</strong>g and ma<strong>in</strong>ta<strong>in</strong><strong>in</strong>g physical<br />
fl ood defences, <strong>in</strong> particular <strong>in</strong> rural areas, is rais<strong>in</strong>g <strong>in</strong>terest<br />
<strong>in</strong> alternative strategies. These alternative, <strong>in</strong>tegrated<br />
fl ood management strategies developed at <strong>the</strong> catchment<br />
scale are now considered viable to provide appropriate<br />
‘standards of service’ <strong>in</strong> areas of particular risk, while also<br />
ensur<strong>in</strong>g no net loss of ecosystem status, and even allow<strong>in</strong>g<br />
enhancement of aquatic, riparian and fl oodpla<strong>in</strong> environments.<br />
The evaluation of such strategies refl ects a<br />
policy shift from fl ood defence to fl ood risk management,<br />
and <strong>in</strong>cludes <strong>the</strong> possibility of <strong>in</strong>creas<strong>in</strong>g <strong>the</strong> frequency of<br />
fl ood<strong>in</strong>g and reduc<strong>in</strong>g <strong>the</strong> standard of service <strong>in</strong> some<br />
fl oodpla<strong>in</strong> locations where land is of relatively low value.<br />
This can have <strong>the</strong> effect of stor<strong>in</strong>g fl oodwater and attenuat<strong>in</strong>g<br />
hydrograph peaks, reduc<strong>in</strong>g fl ood potential <strong>in</strong> downstream<br />
high-value urban fl oodpla<strong>in</strong>s that are o<strong>the</strong>rwise at<br />
risk.<br />
Water protection agencies, concerned at water shortages<br />
and motivated by <strong>the</strong> EU Water Framework Directive,<br />
are also show<strong>in</strong>g <strong>in</strong>creased <strong>in</strong>terest <strong>in</strong> water fl ow regimes<br />
across whole catchments and <strong>in</strong> <strong>the</strong> potential of fl oodpla<strong>in</strong>s<br />
to improve water quality as part of a policy shift<br />
from downstream protection to upstream river bas<strong>in</strong> man-
Table 6.1 Recent policy shifts conducive to fl oodpla<strong>in</strong> restoration<br />
Policy fi eld Forces for change Policy issues<br />
agement. For nature conservationists restored fl oodpla<strong>in</strong>s<br />
represent important habitats that can contribute to meet<strong>in</strong>g<br />
biodiversity targets <strong>in</strong> accordance with <strong>the</strong> EU Habitats<br />
and Birds Directives; here <strong>the</strong> policy shift is <strong>in</strong> part from<br />
species protection to habitat enhancement. Political pressure<br />
is grow<strong>in</strong>g for more environmentally-sensitive forms<br />
of agriculture and forestry, creat<strong>in</strong>g new fund<strong>in</strong>g opportunities<br />
for extensive practices more suited to fl oodpla<strong>in</strong><br />
restoration, <strong>in</strong> a policy shift from agricultural support to<br />
<strong>in</strong>tegrated rural development. Agricultural policy focused<br />
on agri-environmental management means that fl oodpla<strong>in</strong>s<br />
orig<strong>in</strong>ally expensively dra<strong>in</strong>ed and protected can<br />
now be considered sites for re<strong>in</strong>stat<strong>in</strong>g o<strong>the</strong>r functions<br />
whose relative values are perceived to have <strong>in</strong>creased.<br />
F<strong>in</strong>ally, land-use plann<strong>in</strong>g regulations are be<strong>in</strong>g modifi ed<br />
to offer more effective protection of exist<strong>in</strong>g fl oodpla<strong>in</strong>s<br />
and, <strong>in</strong> some <strong>in</strong>stances, earmark<strong>in</strong>g land for <strong>the</strong> future<br />
restoration of fl oodpla<strong>in</strong>s. Spann<strong>in</strong>g <strong>the</strong>se sectoral policy<br />
shifts is a trend towards greater policy <strong>in</strong>tegration and<br />
stakeholder participation over schemes of this k<strong>in</strong>d (Pahl-<br />
Wostl, 2002, 2004), <strong>in</strong>formed <strong>in</strong> part by debates on susta<strong>in</strong>able<br />
development and new forms of governance<br />
(Bressers and Kuks, 2003).<br />
<strong>Manag<strong>in</strong>g</strong> <strong>the</strong>se changes is a challenge for <strong>the</strong> responsible<br />
agencies, which must grapple with questions of<br />
multi-functionality, multiple and nested scales, crosssectoral<br />
activity, policy <strong>in</strong>terplay, actor collaboration and<br />
issues of considerable complexity. One <strong>in</strong>stitutional framework<br />
for manag<strong>in</strong>g <strong>the</strong> changes <strong>in</strong> values, and <strong>the</strong> associated<br />
restoration of natural functions <strong>in</strong> fl oodpla<strong>in</strong>s, is<br />
provided by <strong>the</strong> Water Framework Directive <strong>in</strong> Europe,<br />
and by <strong>the</strong> tools developed for its implementation – <strong>River</strong><br />
Bas<strong>in</strong> Management Plans. In <strong>the</strong> United K<strong>in</strong>gdom, exist<strong>in</strong>g<br />
tools for strategic fl ood management (such as Catchment<br />
Flood Management Plans) now need adaptation to<br />
allow for ecosystem enhancement, comb<strong>in</strong><strong>in</strong>g <strong>the</strong> Envi-<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Riparian and Floodpla<strong>in</strong> <strong>Restoration</strong> 81<br />
Flood protection Flood<strong>in</strong>g events; climate change;<br />
Risk management, soft eng<strong>in</strong>eer<strong>in</strong>g techniques,<br />
<strong>in</strong>frastructure costs; environmental quality natural fl ood storage<br />
Water protection EU Water Framework Directive; water<br />
Catchment-oriented approaches, fl ow regimes,<br />
quality/quantity problems<br />
wetlands, geomorphology<br />
Nature conservation EU Habitats Directive; concerns for<br />
biodiversity<br />
Functional fl oodpla<strong>in</strong> ecosystems<br />
Land-use plann<strong>in</strong>g L<strong>in</strong>kage of fl ood<strong>in</strong>g events to land use Plann<strong>in</strong>g mechanisms for protect<strong>in</strong>g and creat<strong>in</strong>g<br />
areas for fl ood retention<br />
Rural development EU Rural Development Regulation; spatial Integrated approaches to rural economic<br />
disparities<br />
development<br />
Agriculture Agenda 2000; public health concerns;<br />
environmental degradation<br />
Improved agri-environmental schemes<br />
ronment Agency’s roles <strong>in</strong> relation to fl ood management<br />
and conservation (and implementation of <strong>the</strong> EU Habitats<br />
Directive). These tools will thus necessarily focus on<br />
multi-functional management across a catchment-reachhabitat<br />
scale hierarchy, which will imply a variation <strong>in</strong> <strong>the</strong><br />
level of detailed plann<strong>in</strong>g at different scales. At <strong>the</strong> catchment<br />
scale, <strong>the</strong> plann<strong>in</strong>g is essentially a strategic assessment<br />
for manag<strong>in</strong>g priorities <strong>in</strong> relation to budgetary<br />
provision, while at <strong>the</strong> reach scale plann<strong>in</strong>g is focused on<br />
<strong>the</strong> implementation of specifi c policies. This has implications<br />
for <strong>the</strong> mean<strong>in</strong>g of uncerta<strong>in</strong>ty, s<strong>in</strong>ce qualitative<br />
general goals set at <strong>the</strong> catchment strategic scale may<br />
allow greater fl exibility, while more restrictive quantitative<br />
goals may be set at <strong>the</strong> reach scale. However, underly<strong>in</strong>g<br />
this is <strong>the</strong> governance requirement for stakeholder <strong>in</strong>volvement,<br />
expos<strong>in</strong>g <strong>the</strong> question of <strong>the</strong> accountability, <strong>in</strong> democratic<br />
societies, of those <strong>in</strong>stitutions responsible for<br />
goal-sett<strong>in</strong>g <strong>in</strong> relation to restoration <strong>in</strong>itiatives. This <strong>in</strong>troduces<br />
ano<strong>the</strong>r level of uncerta<strong>in</strong>ty.<br />
Decision makers must thus manage complexity and<br />
uncerta<strong>in</strong>ty, and cope <strong>in</strong> an adaptive manner with un<strong>in</strong>tended<br />
negative effects which emerge alongside <strong>the</strong> advantages<br />
derived from <strong>the</strong> policies <strong>the</strong>y enact and implement.<br />
In <strong>the</strong> context of fl oodpla<strong>in</strong> restoration, <strong>the</strong>y may seek to<br />
identify ‘reference’ conditions towards which <strong>the</strong> Water<br />
Framework Directive encourages a shift; however, <strong>the</strong>re is<br />
considerable uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong> defi nition of such reference,<br />
or target, conditions. This ‘goal’ of a restoration<br />
<strong>in</strong>itiative may suggest identifi cation of a set of performance<br />
<strong>in</strong>dicators, but here aga<strong>in</strong> <strong>the</strong>re is uncerta<strong>in</strong>ty. If <strong>the</strong><br />
strategic goal of restoration is to recover <strong>the</strong> dynamics of<br />
natural process and function (by lett<strong>in</strong>g <strong>the</strong> river do <strong>the</strong><br />
work aga<strong>in</strong>), <strong>the</strong> performance criteria are very different,<br />
and necessarily less rigid given <strong>the</strong> time-variability of<br />
fl ows, than if <strong>the</strong> goal is set as a restoration of a specifi c<br />
(static) form.
82 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
There has thus been a grow<strong>in</strong>g <strong>in</strong>terest <strong>in</strong> restoration of<br />
riparian and fl oodpla<strong>in</strong> ecosystems, driven by <strong>in</strong>creas<strong>in</strong>g<br />
knowledge of <strong>the</strong> biophysical l<strong>in</strong>kages between parts of<br />
river bas<strong>in</strong>s and, <strong>in</strong> Europe, by shift<strong>in</strong>g policy directions<br />
and an <strong>in</strong>creased frequency and severity of fl ood events.<br />
Awareness of <strong>the</strong> potential impacts of global climate<br />
change and of <strong>the</strong> impacts of river management activities<br />
such as dam build<strong>in</strong>g on hydrological patterns <strong>in</strong> river<br />
bas<strong>in</strong>s (Montgomery and Boulton, 2003; World Commission<br />
on Dams, 2000) have led to a broaden<strong>in</strong>g of approaches<br />
to fl ood management. Coupled with a wider acceptance of<br />
multi-functional fl oodpla<strong>in</strong>s and more <strong>in</strong>tegrated policy<br />
contexts for <strong>the</strong>se, this has led to <strong>the</strong> proliferation of river<br />
restoration projects <strong>in</strong> many countries (Palmer et al., 2005).<br />
However, river restoration that <strong>in</strong>volves <strong>the</strong> planned return<br />
of fl ood<strong>in</strong>g and its associated geomorphological processes<br />
also reduces <strong>the</strong> predictability of river behaviour for river<br />
managers. The uncerta<strong>in</strong>ty associated with this development<br />
is <strong>the</strong> subject of this chapter, which cont<strong>in</strong>ues with a<br />
consideration of <strong>the</strong> spatial and temporal dynamics that<br />
underp<strong>in</strong> <strong>the</strong> natural function of rivers and <strong>the</strong>ir riparian<br />
and fl oodpla<strong>in</strong> zones.<br />
6.3 THE NATURAL FUNCTIONAL DYNAMICS<br />
OF RIPARIAN ENVIRONMENTS<br />
To achieve <strong>the</strong> goal of <strong>in</strong>tegrated management of fl oodpla<strong>in</strong>s,<br />
<strong>in</strong>clud<strong>in</strong>g restoration of <strong>the</strong>ir physical and ecological<br />
functions, it is necessary to understand some of <strong>the</strong><br />
key relationships that determ<strong>in</strong>e <strong>the</strong> health of <strong>the</strong> aquatic,<br />
riparian and fl oodpla<strong>in</strong> ecosystems. This section accord<strong>in</strong>gly<br />
briefl y reviews this understand<strong>in</strong>g, focus<strong>in</strong>g particularly<br />
on scale and spatial relationships with<strong>in</strong> <strong>the</strong> dra<strong>in</strong>age<br />
bas<strong>in</strong>, and on <strong>the</strong> dynamics of <strong>in</strong>ter-related processes that<br />
defi ne <strong>the</strong> functional status of rivers and <strong>the</strong>ir fl oodpla<strong>in</strong>s<br />
(Malanson, 1993).<br />
6.3.1 Scale and Spatial Relationships: Longitud<strong>in</strong>al<br />
and Lateral<br />
It is fi rst necessary to recognise that a river responds to<br />
conditions with<strong>in</strong> <strong>the</strong> upstream catchment dra<strong>in</strong><strong>in</strong>g to it,<br />
and that <strong>the</strong> river corridor is <strong>the</strong>refore <strong>the</strong> terrestrial low<br />
po<strong>in</strong>t which receives water fl ows, sediments, nutrients and<br />
plant propagules from its contribut<strong>in</strong>g area (Brierly and<br />
Fryirs, 2000). Accord<strong>in</strong>gly, <strong>the</strong> status of a river reach is<br />
dependent on its catchment environment. In a natural river<br />
reach, <strong>the</strong>re will be a diversity of habitats – pools, riffl es,<br />
gravel bars, po<strong>in</strong>t bars, steep banks, levees, side channels<br />
and chutes, abandoned channels, ox-bow lakes, back-<br />
swamps and fl oodpla<strong>in</strong>s. These habitats will be preferentially<br />
occupied by particular species of fauna and fl ora<br />
(see, for example, Marston et al., 1995; Richards et al.,<br />
2002). It is often <strong>the</strong> case that conservation focuses on<br />
particular species, but a key to general ecological health<br />
is <strong>the</strong> ma<strong>in</strong>tenance of habitats (Ward and Tockner, 2001).<br />
However, <strong>the</strong> habitats refl ect <strong>the</strong> behaviour of <strong>the</strong> river at<br />
<strong>the</strong> reach scale. For example, if a river meanders without<br />
constra<strong>in</strong>t, and <strong>in</strong>undates its fl oodpla<strong>in</strong> roughly once every<br />
1–5 years, it is likely that it will create and ma<strong>in</strong>ta<strong>in</strong> a<br />
natural diversity of habitat. Whe<strong>the</strong>r this behaviour occurs<br />
will depend on <strong>the</strong> way <strong>in</strong> which <strong>the</strong> catchment is occupied<br />
and managed, and delivers water and sediment to <strong>the</strong><br />
reach <strong>in</strong> question. Thus <strong>the</strong> crucial connections to understand<br />
are those between <strong>the</strong> hydrology, sediment supply<br />
and ecology of <strong>the</strong> contribut<strong>in</strong>g catchment area, and <strong>the</strong><br />
character of <strong>the</strong> river reach to which it dra<strong>in</strong>s. This results<br />
<strong>in</strong> a structure of longitud<strong>in</strong>al relationships and upstream–<br />
downstream connectivity, while lateral connections are<br />
also critical <strong>in</strong> l<strong>in</strong>k<strong>in</strong>g <strong>the</strong> aquatic (river) and terrestrial<br />
(fl oodpla<strong>in</strong>) environments across <strong>the</strong> riparian zone.<br />
The ecological status of a river reach is strongly dependent<br />
on longitud<strong>in</strong>al (downstream) connectivity, and this<br />
is refl ected <strong>in</strong> <strong>the</strong> river cont<strong>in</strong>uum concept (Vannote et al.,<br />
1980; Petts et al., 2000). This is based on <strong>the</strong> idea that<br />
rivers transport water and sediment downstream through<br />
a systematic cont<strong>in</strong>uum of conditions from steep, headwater<br />
reaches with coarse bed material, and shallow,<br />
tumbl<strong>in</strong>g fl ow, <strong>in</strong> narrow valleys, to more gently-slop<strong>in</strong>g<br />
lowland reaches with fi ne silty, sandy beds, deeper, slow<br />
fl ow and wide fl oodpla<strong>in</strong>s. Related to this, <strong>the</strong> ecology<br />
changes systematically from upstream to downstream<br />
reaches, and a cont<strong>in</strong>uum here refl ects <strong>the</strong> chang<strong>in</strong>g<br />
nature of biological processes. Upstream, woody debris is<br />
<strong>in</strong>troduced from <strong>the</strong> riparian vegetation and is ‘processed’<br />
by ‘shredders’ (<strong>in</strong>vertebrates that consume leaves and<br />
woody material) <strong>in</strong> <strong>the</strong> stream, while also supply<strong>in</strong>g nutrients<br />
and dissolved organic carbon. Fur<strong>the</strong>r downstream,<br />
<strong>the</strong>se products of upstream processes are used by o<strong>the</strong>r<br />
species, as when fi ne particulate organic matter is ei<strong>the</strong>r<br />
collected or fi ltered by species which are ‘ga<strong>the</strong>rers’. This<br />
implies that management which prevents upstream biological<br />
functions will have downstream effects because of<br />
<strong>the</strong> cont<strong>in</strong>uous relationships exist<strong>in</strong>g along <strong>the</strong> river<br />
course. Similarly, manag<strong>in</strong>g <strong>the</strong> river by <strong>in</strong>troduc<strong>in</strong>g dams<br />
and weirs will <strong>in</strong>terrupt <strong>the</strong> cont<strong>in</strong>uity of migration of<br />
species by <strong>in</strong>hibit<strong>in</strong>g upstream fi sh migration to spawn<strong>in</strong>g<br />
sites, and <strong>the</strong> downstream fl ow of both seeds and plant<br />
material from which vegetative reproduction may take<br />
place.<br />
There may be some debate about <strong>the</strong> details of <strong>the</strong><br />
‘river cont<strong>in</strong>uum’; an alternative is <strong>the</strong> process doma<strong>in</strong>s
concept (Montgomery, 1999), which argues that localscale<br />
geomorphological and biological processes determ<strong>in</strong>e<br />
<strong>the</strong> stream habitat, <strong>the</strong> disturbance regimes and <strong>the</strong><br />
species <strong>in</strong>teractions that <strong>in</strong>fl uence stream communities<br />
and biodiversity. However, for practical purposes, whe<strong>the</strong>r<br />
<strong>the</strong>re are longitud<strong>in</strong>al zones, or a longitud<strong>in</strong>al cont<strong>in</strong>uum,<br />
is less important than that <strong>the</strong>re are strong upstream–<br />
downstream relationships. Aquatic organisms have<br />
evolved to be adjusted to <strong>the</strong> most probable set of physical<br />
conditions aris<strong>in</strong>g from <strong>the</strong> fl uvial geomorphology and<br />
hydrology. Downstream reaches rely on carbon <strong>in</strong>puts<br />
from upstream, while middle reaches are sites of primary<br />
production (with clear water and less shad<strong>in</strong>g by<br />
vegetation).<br />
Longitud<strong>in</strong>al relationsips are supplemented by lateral<br />
connections (Ward and Stanford, 1995). The high biodiversity<br />
of river corridors is, <strong>in</strong> part, a refl ection of <strong>the</strong><br />
diversity of habitat at <strong>the</strong> marg<strong>in</strong> between two dist<strong>in</strong>ct<br />
ecosystems, those of <strong>the</strong> river and <strong>the</strong> fl oodpla<strong>in</strong>. The<br />
riparian zone is an aquatic–terrestrial ecotone, and is a<br />
zone of transition and exchange which benefi ts from, and<br />
regulates, <strong>the</strong> processes and functions of <strong>the</strong> ecosystems<br />
it connects. Interference with one <strong>in</strong>evitably affects <strong>the</strong><br />
o<strong>the</strong>r, and <strong>the</strong> ecotone is itself a high priority for conservation.<br />
When <strong>the</strong> channel is deepened and dredged, or when<br />
embankments are created, <strong>the</strong> lateral connectivity between<br />
river and fl oodpla<strong>in</strong> is disrupted. In some cases this<br />
restricts access to lateral channels, side-arms and ox-bow<br />
lakes, which convert to terrestrial status. The bi-directional<br />
connection between <strong>the</strong> river and groundwater is<br />
<strong>in</strong>hibited and fl oodpla<strong>in</strong> recharge is restricted. The potential<br />
for fi sh species to use fl oodpla<strong>in</strong> woodland as a fl oodperiod<br />
refuge is prevented and <strong>the</strong> fi sh population suffers.<br />
Bank-side vegetation provides shade, which regulates<br />
water temperature, and tree roots a diversity of microhabitats,<br />
which benefi t <strong>the</strong> aquatic fauna; it also supplies<br />
organic material to <strong>the</strong> stream. Removal of <strong>the</strong> vegetation<br />
removes <strong>the</strong>se functions and causes a loss of productivity,<br />
habitat and biodiversity.<br />
6.3.2 Dynamics: Variable Flows, Sediment Delivery<br />
and Channel Migration<br />
Aquatic and riparian species are adapted to and require<br />
variable fl ows (<strong>in</strong>clud<strong>in</strong>g overbank fl ows), and are sensitive<br />
to <strong>the</strong> tim<strong>in</strong>g of that variation. They also require different<br />
fl ows at various stages of <strong>the</strong>ir life cycles. For<br />
example, salmonidae require a stable gravel substrate<br />
when spawn<strong>in</strong>g, but at o<strong>the</strong>r times depend on fl ows<br />
which dilate <strong>the</strong> gravel and fl ush out fi ne sediment that<br />
has accumulated <strong>in</strong> <strong>the</strong> pore spaces. The preferred water<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Riparian and Floodpla<strong>in</strong> <strong>Restoration</strong> 83<br />
depth, velocity and substrate for a given species will also<br />
change between <strong>the</strong> fry, juvenile and adult life stages, and<br />
this requires that <strong>the</strong>re is access to spatial variation of<br />
habitat, which provides suitable refuge locations dur<strong>in</strong>g<br />
high fl ows for <strong>in</strong>dividuals which cannot survive extreme<br />
conditions.<br />
Some riparian tree species, such as black poplar<br />
(Populus nigra) and willows, require occasional large<br />
fl oods to create new sedimentary surfaces for colonisation<br />
and regeneration, but also need a gradual recession<br />
dur<strong>in</strong>g <strong>the</strong> period of seedl<strong>in</strong>g establishment for <strong>the</strong>ir survival.<br />
The tim<strong>in</strong>g of high fl ow needs to match <strong>the</strong> release<br />
of seeds <strong>in</strong> late spr<strong>in</strong>g and early summer; this is an issue<br />
<strong>in</strong> river bas<strong>in</strong>s with reservoirs or water transfers that<br />
reduce fl ows and change <strong>the</strong> natural tim<strong>in</strong>g of high fl ows.<br />
The role of high fl ows is represented <strong>in</strong> <strong>the</strong> fl ood pulse<br />
concept (Junk et al., 1989; Middleton, 1999), which<br />
emphasises that fl oods cause channel migration and create<br />
new habitat, and supply nutrients and genetic material<br />
(seeds and plant material) to <strong>the</strong> riparian environment.<br />
However, <strong>the</strong>se processes can occur across a range of<br />
fl ows, so a general idea of a range of fl ow pulses (Tockner<br />
et al., 2000) is also found <strong>in</strong> <strong>the</strong> literature on riparian<br />
ecology.<br />
Of course, fl ows that are too high may both<br />
cause damage and create anaerobic conditions which<br />
prevent germ<strong>in</strong>ation and cause mortality <strong>in</strong> early seedl<strong>in</strong>gs.<br />
This illustrates that <strong>the</strong> process of manag<strong>in</strong>g<br />
fl ows for ecological benefi ts requires a subtle understand<strong>in</strong>g<br />
of <strong>the</strong> impacts of fl ows on <strong>the</strong> life histories of a<br />
range of species. This has given rise to <strong>the</strong> concept of<br />
‘environmental fl ows’, which <strong>in</strong>clude a wide range of fl ow<br />
levels at different times, and with different recurrences,<br />
each of which contributes to regeneration and to different<br />
life-cycle stages of different aspects of <strong>the</strong> aquatic and<br />
riparian ecology (Hill and Platts, 1991; Mahoney and<br />
Rood, 1998). The management of such variable fl ows,<br />
seasonally and <strong>in</strong>ter-annually, is a complex and uncerta<strong>in</strong><br />
process demand<strong>in</strong>g an adaptive response to <strong>the</strong> unfold<strong>in</strong>g<br />
of <strong>the</strong> unpredictable climatically-driven fl ow regime.<br />
However, it is a process which is now occurr<strong>in</strong>g <strong>in</strong> many<br />
climatic regions and is considered fur<strong>the</strong>r <strong>in</strong> <strong>the</strong> follow<strong>in</strong>g<br />
section.<br />
An important characteristic of natural, unmanaged<br />
rivers is that <strong>the</strong>y migrate across <strong>the</strong>ir fl oodpla<strong>in</strong>s at rates<br />
that depend on <strong>the</strong> energy of <strong>the</strong>ir fl ows and <strong>the</strong> resistance<br />
of <strong>the</strong>ir perimeter sediments to entra<strong>in</strong>ment, erosion and<br />
transport. This means that <strong>the</strong>y gradually turn over <strong>the</strong>ir<br />
fl oodpla<strong>in</strong> sediments, and that <strong>the</strong> fl oodpla<strong>in</strong> consists of a<br />
mosaic of surfaces of vary<strong>in</strong>g age and sedimentology<br />
giv<strong>in</strong>g rise to a mosaic of diverse habitats (Salo et al.,<br />
1986; Nilsson et al., 1991b).
84 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
The most well-known form of such migration is that of<br />
river meander<strong>in</strong>g, which arises because of bank erosion<br />
on <strong>the</strong> outside of bends and <strong>the</strong> deposition of po<strong>in</strong>t bars<br />
on <strong>the</strong> <strong>in</strong>sides. This process is what creates <strong>the</strong> diversity<br />
of habitat and is <strong>the</strong>refore very important for biodiversity.<br />
It is commonly understood that disturbance of ecosystems<br />
is critical for <strong>the</strong> structur<strong>in</strong>g and ma<strong>in</strong>tenance of biodiversity,<br />
as it stops a plant succession from progress<strong>in</strong>g to a<br />
uniform mature state and cont<strong>in</strong>ually resets <strong>the</strong> succession<br />
locally and re-<strong>in</strong>troduces pioneer species. In some ecosystems,<br />
disturbance may be caused by w<strong>in</strong>d, fi re or human<br />
<strong>in</strong>fl uences (<strong>in</strong> shift<strong>in</strong>g cultivation, for example). In river<br />
corridors, it is caused by erosion, deposition and river<br />
migration. The <strong>in</strong>termediate disturbance hypo<strong>the</strong>sis<br />
(Connell, 1978; Ward et al., 1999) argues that maximum<br />
biodiversity occurs under conditions of <strong>in</strong>termediate disturbance.<br />
Too much disturbance (as <strong>in</strong> very dynamic<br />
braided rivers) results <strong>in</strong> a preponderance of pioneer<br />
species, while too little disturbance allows a succession to<br />
progress towards a mature and relatively uniform plant<br />
community. Both have lower biodiversity than occurs<br />
with rates of disturbance that are <strong>in</strong>termediate. If <strong>in</strong> river<br />
corridors it is high fl ows, erosion, deposition and channel<br />
migration (<strong>the</strong> river dynamics) that promote biodiversity,<br />
it follows that practices of river management designed to<br />
<strong>in</strong>hibit movement of <strong>the</strong> river (such as embankment and<br />
bank protection) are likely to reduce biodiversity. Ecological<br />
status will improve <strong>the</strong>refore when rivers are given <strong>the</strong><br />
freedom to move. This is a concept which is now enshr<strong>in</strong>ed<br />
<strong>in</strong> policy for <strong>the</strong> management of <strong>the</strong> tributaries of <strong>the</strong><br />
Rh<strong>in</strong>e and <strong>the</strong> Meuse <strong>in</strong> The Ne<strong>the</strong>rlands, and <strong>in</strong>volves<br />
relocation of embankments, <strong>in</strong>creased off-channel fl ood<br />
storage and restoration of <strong>the</strong> functions of abandoned<br />
channels and side-arms.<br />
From this outl<strong>in</strong>e it is evident that <strong>the</strong> characteristics of<br />
<strong>the</strong> natural biophysical system of <strong>the</strong> river-fl oodpla<strong>in</strong><br />
environment are connectivity, spatial variability and temporal<br />
<strong>in</strong>stability (dynamics). It is generally acknowledged<br />
that restoration of <strong>the</strong>se characteristics is a necessary component<br />
of any restoration <strong>in</strong>itiative although <strong>the</strong> scale at<br />
which <strong>the</strong>y can be restored is very variable. The uncerta<strong>in</strong>ty<br />
posed by restor<strong>in</strong>g <strong>the</strong> very characteristics that were<br />
removed <strong>in</strong> order to make rivers more manageable and by<br />
extension less uncerta<strong>in</strong> <strong>in</strong> <strong>the</strong>ir behaviour also needs to<br />
be accommodated (see also Gregory and Downs, Chapter<br />
13). This is as much a question of chang<strong>in</strong>g attitudes<br />
(personal, public and <strong>in</strong>stitutional) through education as it<br />
is a question of understand<strong>in</strong>g more about <strong>the</strong> biophysical<br />
processes, although it rema<strong>in</strong>s <strong>the</strong> case that such understand<strong>in</strong>g<br />
is highly uncerta<strong>in</strong> when it requires prediction of<br />
future behaviour at a specifi c location with<strong>in</strong> a river<br />
system (e.g. Chapter 5).<br />
6.4 THE SCALE AND PRACTICE<br />
OF RESTORATION<br />
The practice of river restoration is <strong>in</strong>formed by <strong>the</strong> cont<strong>in</strong>u<strong>in</strong>g<br />
scientifi c <strong>in</strong>vestigation of <strong>the</strong> vital l<strong>in</strong>ks, at different<br />
spatial and temporal scales, between biotic and abiotic<br />
components across a catchment. However, it is also limited<br />
by compet<strong>in</strong>g needs for resources with<strong>in</strong> a catchment.<br />
The net result <strong>in</strong> many river bas<strong>in</strong>s is that river restoration<br />
takes place at relatively small, confi ned sites where re<strong>in</strong>statement<br />
of <strong>the</strong> l<strong>in</strong>ks between <strong>the</strong> river and its fl oodpla<strong>in</strong><br />
is limited <strong>in</strong> degree and extent and primarily focussed<br />
on re-establish<strong>in</strong>g lateral connectivity. It is important to<br />
dist<strong>in</strong>guish between river restoration that takes place at<br />
this spatially limited scale of <strong>the</strong> reach or section of a<br />
reach and river restoration that takes <strong>in</strong>to consideration <strong>the</strong><br />
longitud<strong>in</strong>al l<strong>in</strong>kages with<strong>in</strong> a river bas<strong>in</strong> and manages<br />
<strong>the</strong> primary <strong>in</strong>puts of water and sediment across <strong>the</strong> whole<br />
catchment. There are many thousands of river restoration<br />
projects worldwide that have been implemented at <strong>the</strong> fi rst<br />
scale but far fewer examples of resource management<br />
tak<strong>in</strong>g place at <strong>the</strong> scale of <strong>the</strong> catchment (see Chapter 13).<br />
At both scales <strong>the</strong>re is a considerable challenge for scientists<br />
to defi ne ecosystem needs <strong>in</strong> a way that can guide<br />
policy formulation and management action (Poff et al.,<br />
2003). However, <strong>the</strong> challenge is considerably greater at<br />
<strong>the</strong> scale of <strong>the</strong> catchment because levels of uncerta<strong>in</strong>ty<br />
about <strong>the</strong> ecological and o<strong>the</strong>r outcomes and <strong>the</strong> number<br />
and range of stakeholders that need to engage with <strong>the</strong><br />
process <strong>in</strong>crease rapidly as spatial scale <strong>in</strong>creases. Fur<strong>the</strong>rmore,<br />
until quite recently, <strong>the</strong> water needs of humans<br />
and those of ecosystems have been seen <strong>in</strong> competition<br />
(Richter et al., 2003) and <strong>the</strong> quantum leap <strong>in</strong> human<br />
perception from this to view<strong>in</strong>g ecosystems as legitimate<br />
users of water (K<strong>in</strong>g and Louwe, 1998; Naiman et al.,<br />
2002) has largely still to be made.<br />
To illustrate <strong>the</strong> issues and practice of work<strong>in</strong>g at<br />
different scales, <strong>the</strong> example of <strong>the</strong> needs of fl oodpla<strong>in</strong><br />
forest ecosystems and <strong>the</strong> different scales at which those<br />
needs can be provided is considered here. There are many<br />
<strong>in</strong>ter-related variables operat<strong>in</strong>g <strong>in</strong> a fully functional fl oodpla<strong>in</strong><br />
forest ecosystem but, never<strong>the</strong>less, <strong>the</strong>ir diverse,<br />
mobile vegetation mosaics can be said to have four essential<br />
requirements to be self-regenerat<strong>in</strong>g (Table 6.2). To<br />
get all of <strong>the</strong>se it is necessary to manage <strong>the</strong> disturbance<br />
processes that arrive at a fl oodpla<strong>in</strong> and this can be done<br />
<strong>in</strong> a number of ways at both <strong>the</strong> reach and catchment<br />
scales.<br />
6.4.1 Catchment-Scale Management<br />
At <strong>the</strong> scale of <strong>the</strong> catchment, restoration <strong>in</strong>itiatives are<br />
likely to <strong>in</strong>volve management of physical processes <strong>in</strong> one
or more places <strong>in</strong> <strong>the</strong> catchment upstream of <strong>the</strong> fl oodpla<strong>in</strong>,<br />
so that <strong>the</strong>y eventually have an effect on <strong>the</strong> disturbances<br />
arriv<strong>in</strong>g <strong>in</strong> <strong>the</strong> fl oodpla<strong>in</strong> zone. This type of<br />
disturbance management is ‘<strong>in</strong>direct’ but it is also <strong>the</strong><br />
most desirable for long term successful and self-susta<strong>in</strong><strong>in</strong>g<br />
restoration or management of fl oodpla<strong>in</strong> forests. It allows<br />
<strong>the</strong> river to fl ood and <strong>the</strong> channel to move and create its<br />
own sites for <strong>the</strong> regeneration of trees. It is, however, quite<br />
diffi cult to achieve for a variety of reasons:<br />
• It is not easy to predict where disturbances will have an<br />
<strong>in</strong>fl uence <strong>in</strong> <strong>the</strong> fl oodpla<strong>in</strong> zone and it is <strong>the</strong>refore an<br />
uncerta<strong>in</strong> form of management.<br />
• Unpredictability and reduced control have not been<br />
desirable features for river managers and as such <strong>the</strong>y<br />
tend to be avoided.<br />
• In highly managed and fragmented river systems <strong>the</strong>re<br />
may be too many human <strong>in</strong>terventions for it to work.<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Riparian and Floodpla<strong>in</strong> <strong>Restoration</strong> 85<br />
Table 6.2 The four essential requirements for a self-regenerat<strong>in</strong>g fl oodpla<strong>in</strong> forest (from Hughes and Muller, 2003; Hughes et al.,<br />
2005)<br />
Requirement Rationale<br />
Flows needed by fl oodpla<strong>in</strong> forests • Regular fl ows which replenish and ma<strong>in</strong>ta<strong>in</strong> fl oodpla<strong>in</strong> water tables.<br />
These fl ows allow established trees to grow.<br />
• Periodic high fl ows which cause channel movement and sediment<br />
deposition. These provide potential regeneration sites and should be<br />
variable between years.<br />
• Well-timed fl ows through <strong>the</strong> fi rst grow<strong>in</strong>g season which allow<br />
delivery of seeds to <strong>the</strong> fl oodpla<strong>in</strong> and establishment of seedl<strong>in</strong>gs.<br />
Unseasonal high fl ows can cause high mortality to seedl<strong>in</strong>gs <strong>in</strong> <strong>the</strong>ir<br />
fi rst grow<strong>in</strong>g season.<br />
Regeneration sites needed by fl oodpla<strong>in</strong> forests • Open sites as many pioneer tree species typical of fl oodpla<strong>in</strong> forests<br />
cannot tolerate competition.<br />
• Sites that are moist through <strong>the</strong> fi rst grow<strong>in</strong>g season to facilitate<br />
regeneration.<br />
• Sites near <strong>the</strong> water’s edge because <strong>the</strong>se tend to be moister and catch<br />
organic debris. However, sites right on <strong>the</strong> water’s edge tend to suffer<br />
from fl ow disturbance and waterlogg<strong>in</strong>g.<br />
• A variety of sediment types to provide regeneration niches for a<br />
variety of species.<br />
Water table conditions needed by fl oodpla<strong>in</strong> forests • Water tables accessible to <strong>the</strong> roots of seedl<strong>in</strong>gs through <strong>the</strong>ir fi rst<br />
grow<strong>in</strong>g season.<br />
• Gradual recession of water tables follow<strong>in</strong>g a fl ood.<br />
• Limited waterlogg<strong>in</strong>g.<br />
Propagation materials needed by fl oodpla<strong>in</strong> forests • Seeds which are carried by <strong>the</strong> river and deposited dur<strong>in</strong>g fl oods. The<br />
phenology of seed release and <strong>the</strong> tim<strong>in</strong>g of fl ood peaks are critical <strong>in</strong><br />
any year for successful establishment of seedl<strong>in</strong>gs.<br />
• Vegetative material which arrives by fl ood or is deposited locally.<br />
• Seeds that are carried <strong>in</strong> <strong>the</strong> w<strong>in</strong>d. Whereas seeds carried <strong>in</strong> <strong>the</strong> river<br />
always move from upstream areas to downstream areas, seeds carried<br />
<strong>in</strong> <strong>the</strong> w<strong>in</strong>d tend to move <strong>in</strong> <strong>the</strong> direction of prevail<strong>in</strong>g w<strong>in</strong>ds.<br />
• It requires consensus among a huge number of<br />
stakeholders.<br />
The ways <strong>in</strong> which this catchment-scale management<br />
of water and sediment resources takes place is varied<br />
but <strong>the</strong>re are now a number of both established and<br />
emerg<strong>in</strong>g methodologies which have been put <strong>in</strong>to practice,<br />
many of which are reviewed more fully elsewhere<br />
(Arth<strong>in</strong>gton, 1998; Hughes and Rood, 2003; Postel and<br />
Richter, 2003):<br />
• Managed releases downstream from impoundments.<br />
This methodology is relatively simple <strong>in</strong> concept and<br />
<strong>in</strong>volves plann<strong>in</strong>g fl ow releases from structures such as<br />
dams so that <strong>the</strong>y provide maximum benefi t to downstream<br />
aquatic and riparian ecosystems as well as to<br />
o<strong>the</strong>r users. In practice this can only be contemplated<br />
where <strong>the</strong> eng<strong>in</strong>eer<strong>in</strong>g design of <strong>the</strong> dam structure gives
86 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
suffi cient control to release fl ows. The approach has<br />
been well-tried <strong>in</strong> North America and particularly good<br />
examples can be found <strong>in</strong> <strong>the</strong> St Mary <strong>River</strong> <strong>in</strong> Alberta,<br />
Canada (Rood and Mahoney, 2000) (Case Study 6.1)<br />
and <strong>in</strong> <strong>the</strong> Truckee <strong>River</strong> of Nevada, USA (Rood et al.,<br />
2003). Both examples use <strong>the</strong> Recruitment Box Model<br />
of Mahoney and Rood (1998) which allows quantitative<br />
description of <strong>the</strong> comb<strong>in</strong>ed requirement for appropriately<br />
timed high fl ows to create and saturate suitable<br />
fl oodpla<strong>in</strong> sites downstream and subsequent gradual<br />
fl ow recession (ramp<strong>in</strong>g rates) to permit seedl<strong>in</strong>g survival.<br />
Detailed knowledge of <strong>the</strong> requirements of <strong>the</strong><br />
germ<strong>in</strong>ation and seedl<strong>in</strong>g establishment phases of <strong>the</strong><br />
life-cycles of target tree species have to be known and<br />
ecologically relevant fl ows have to be characterised.<br />
There is considerable debate on how best to characterise<br />
<strong>the</strong> most ecologically critical aspects of fl ow regimes<br />
(Olden and Poff, 2003) such that an optimal balance is<br />
struck between a m<strong>in</strong>imum of hydrological <strong>in</strong>dices and<br />
maximum explanation of ecosystem function by <strong>the</strong>se<br />
<strong>in</strong>dices. The process becomes more diffi cult as projects<br />
go beyond prescrib<strong>in</strong>g fl ows for a s<strong>in</strong>gle ecosystem, like<br />
Case Study 6.1<br />
The ‘Recruitment Box’ model developed by Mahoney<br />
and Rood (1998) del<strong>in</strong>eates a zone on a fl oodpla<strong>in</strong>,<br />
defi ned by elevation and time, <strong>in</strong> which riparian cottonwood<br />
seedl<strong>in</strong>gs are likely to become successfully<br />
established if streamfl ow patterns are favourable<br />
(Figure 6.2). Along <strong>the</strong> St Mary <strong>River</strong> <strong>in</strong> Alberta,<br />
Canada, fl ow regulation from a headwater dam built <strong>in</strong><br />
1953 led to high mortality of established cottonwood<br />
(Populus deltoides) trees and no recruitment of new<br />
trees <strong>in</strong> <strong>the</strong> fl oodpla<strong>in</strong> downstream due to <strong>in</strong>suffi cient<br />
fl ows at critical times <strong>in</strong> <strong>the</strong> grow<strong>in</strong>g season (Rood<br />
et al., 1995) (Figure 6.3). Dur<strong>in</strong>g <strong>the</strong> 1990s, after identifi<br />
cation of <strong>the</strong> cause of <strong>the</strong> problem, regional water<br />
resource managers implemented changes <strong>in</strong> <strong>the</strong> operation<br />
of <strong>the</strong> St Mary Dam (Rood and Mahoney, 2000)<br />
(Figure 6.4). In particular, fl ows were designed to<br />
provide a gradual reduction <strong>in</strong> <strong>the</strong> fall<strong>in</strong>g limb of <strong>the</strong><br />
hydrograph after <strong>the</strong> spr<strong>in</strong>g snow-melt fl ood ra<strong>the</strong>r<br />
than <strong>the</strong> abrupt fall that occurred dur<strong>in</strong>g <strong>the</strong> post-dam<br />
period. This occurred because <strong>the</strong> dam operators shut<br />
<strong>the</strong> spillway gates at <strong>the</strong> dam to divert water <strong>in</strong>to irrigation<br />
channels. The gradual recession of fl oods was<br />
considered vital for replenishment of water tables <strong>in</strong><br />
<strong>the</strong> fl oodpla<strong>in</strong> zone and for <strong>the</strong> establishment of cottonwood<br />
seedl<strong>in</strong>gs whose roots are unable to ma<strong>in</strong>ta<strong>in</strong><br />
contact with <strong>the</strong> water table if it falls too rapidly.<br />
Figure 6.2 The recruitment Box model applied to <strong>the</strong> lower St<br />
Mary <strong>River</strong> (modifi ed from Mahoney and Rood, 1998)<br />
(a)<br />
(b)<br />
Figure 6.3 (a) A photograph of <strong>the</strong> St Mary <strong>River</strong> fl oodpla<strong>in</strong><br />
taken <strong>in</strong> 1991 shows dead cottonwoodtrees before successful<br />
fl ood releases were implemented (Photograph by Franc<strong>in</strong>e<br />
Hughes). (b) (See also colour plate section) By 2002 <strong>the</strong>re is<br />
signifi cant growth of young cottonwood trees on <strong>the</strong> fl oodpla<strong>in</strong><br />
follow<strong>in</strong>g planned releases (Photograph by Stewart Rood)
Figure 6.4 The two pairs of graphs depict a stage hydrograph<br />
with superimposed recruitment box at <strong>the</strong> ‘ideal’ time and elevation<br />
and with ideal drawdown rates. (a) In 1964, a post-dam fl ood<br />
was managed for maximal cut-back ra<strong>the</strong>r than naturalised recession.<br />
Regeneration of trees did not occur <strong>in</strong> that year. (b) In 1995,<br />
a managed fl ow, us<strong>in</strong>g <strong>the</strong> recruitment box as a guide, successfully<br />
promoted regeneration with a well-timed fl ood peak and<br />
suitable fl ood recession rates through <strong>the</strong> fi rst grow<strong>in</strong>g season<br />
(from Rood and Mahoney, 2000)<br />
a fl oodpla<strong>in</strong> forest, <strong>in</strong>to satisfy<strong>in</strong>g multiple ecosystem<br />
needs.<br />
• Flow allocation methodologies. <strong>River</strong> fl ows are altered<br />
as soon as water is used for human purposes. Richter<br />
et al. (2003) state that ‘<strong>the</strong> ultimate challenge of ecologically<br />
susta<strong>in</strong>able water management is to design and<br />
implement a water management program that stores and<br />
diverts water for human purposes <strong>in</strong> a manner that does<br />
Case Study 6.2<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Riparian and Floodpla<strong>in</strong> <strong>Restoration</strong> 87<br />
not cause affected ecosystems to degrade or simplify’.<br />
There is necessarily a limit to how much water can be<br />
taken out of a river, and at what times of <strong>the</strong> year, if an<br />
ecosystem like a fl oodpla<strong>in</strong> forest is to rema<strong>in</strong> selfsusta<strong>in</strong>able.<br />
The question of how to allocate water to<br />
achieve susta<strong>in</strong>able water use for a range of ecosystems<br />
and human purposes is addressed by a series of holistic<br />
fl ow allocation methodologies, many of which have<br />
been developed <strong>in</strong> semi-arid areas <strong>in</strong> Australia, South<br />
Africa and North America. Exampes of <strong>the</strong>se models<br />
<strong>in</strong>clude <strong>the</strong> Bench Mark<strong>in</strong>g Methodology (Arth<strong>in</strong>gton,<br />
1998; Brizga, 2000) and <strong>the</strong> Flow <strong>Restoration</strong> Methodology<br />
(Arth<strong>in</strong>gton et al., 2000) <strong>in</strong> Australia; The Build<strong>in</strong>g<br />
Block Methodology (K<strong>in</strong>g and Louwe, 1998) and<br />
<strong>the</strong> DRIFT methodology (K<strong>in</strong>g et al., 2003; Brown and<br />
Joubert, 2003) <strong>in</strong> South Africa. Closely allied to <strong>the</strong>se<br />
models is <strong>the</strong> Adaptive Water Management Framework<br />
for <strong>in</strong>itiat<strong>in</strong>g ecologically susta<strong>in</strong>able water management<br />
programmes <strong>in</strong> <strong>the</strong> United States and elsewhere<br />
<strong>in</strong> <strong>the</strong> world by Richter et al. (2003). The general<br />
approach of <strong>the</strong>se methodologies is to determ<strong>in</strong>e <strong>the</strong><br />
‘environmental fl ows’ necessary to susta<strong>in</strong> aquatic and<br />
riparian ecosystems and <strong>the</strong>n to <strong>in</strong>tegrate <strong>the</strong> identifi ed<br />
fl ow needs with o<strong>the</strong>r fl ow requirements with<strong>in</strong> <strong>the</strong> river<br />
bas<strong>in</strong> on seasonal, <strong>in</strong>ter-annual and even decaded timescales.<br />
They are multi-stage, consensual or round-table<br />
approaches requir<strong>in</strong>g <strong>the</strong> <strong>in</strong>put of many experts and data<br />
on a range of aquatic and riparian ecosystems, on hydrological<br />
and geomorphological aspects, on modell<strong>in</strong>g of<br />
relationships between hydrological and biological attributes<br />
and eventually on all <strong>the</strong> o<strong>the</strong>r water uses <strong>in</strong> <strong>the</strong><br />
river bas<strong>in</strong>. It is at <strong>the</strong> stage when environmental fl ows<br />
are determ<strong>in</strong>ed that fl oodpla<strong>in</strong> forest requirements are<br />
<strong>in</strong>cluded <strong>in</strong> <strong>the</strong> process. The applicability of <strong>the</strong>se methodologies<br />
to <strong>the</strong> restoration of fl oodpla<strong>in</strong> forests <strong>in</strong><br />
Europe is discussed by Hughes and Rood (2003).<br />
Ano<strong>the</strong>r holistic approach to <strong>the</strong> management of water<br />
resources is <strong>the</strong> ‘alternative futures’ approach. This<br />
approach <strong>in</strong>cludes consideration of land uses across a<br />
catchment as well as fl ow allocation and among o<strong>the</strong>r<br />
places has been applied <strong>in</strong> <strong>the</strong> Willamette <strong>River</strong> Bas<strong>in</strong> <strong>in</strong><br />
Oregon, USA (Case Study 6.2).<br />
In <strong>the</strong> Willamette Bas<strong>in</strong>, <strong>in</strong> western Oregon, USA, an alternative futures analysis has been carried out to <strong>in</strong>form community<br />
decision mak<strong>in</strong>g regard<strong>in</strong>g land and water use <strong>in</strong> <strong>the</strong> river bas<strong>in</strong>. This is a participatory approach to river bas<strong>in</strong><br />
plann<strong>in</strong>g that presents stakeholders with a range of alternative scenarios <strong>in</strong>volv<strong>in</strong>g higher or lower levels of land and<br />
water use <strong>in</strong> <strong>the</strong> bas<strong>in</strong> and <strong>the</strong>ir resultant environmental impacts. In <strong>the</strong> Willamette Bas<strong>in</strong>, <strong>the</strong> current and historical<br />
landscapes were analysed and <strong>the</strong>n three future scenarios were generated, refl ect<strong>in</strong>g vary<strong>in</strong>g assumptions about land
88 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
and water use. Historical data on land and water use and population levels dat<strong>in</strong>g from 1850 were used and scenarios<br />
were projected to 2050 (Baker et al., 2004). Scenarios were evaluated on four areas of resource endpo<strong>in</strong>ts that were<br />
considered to be of value to stakeholders: water availability; river attributes such as channel structure, riparian and<br />
<strong>in</strong>stream ecosystem richness; ecological condition of streams (us<strong>in</strong>g fi sh and benthic <strong>in</strong>vertebrates as <strong>in</strong>dicators); and<br />
terrestrial wildlife <strong>in</strong> <strong>the</strong> bas<strong>in</strong> (Dole and Niemi, 2004) (Figure 6.5).<br />
The fl oodpla<strong>in</strong> of <strong>the</strong> Willamette <strong>River</strong> has been studied specifi cally with regard to prioritis<strong>in</strong>g parts of <strong>the</strong> historical<br />
fl oodpla<strong>in</strong> that might be particularly suitable for restoration us<strong>in</strong>g a landscape modell<strong>in</strong>g approach (Hulse<br />
and Gregory, 2004). This categorises <strong>the</strong> fl oodpla<strong>in</strong> <strong>in</strong>to areas of high and low potential for a range of both biophysical<br />
properties and socio-economic constra<strong>in</strong>ts considered at three spatial scales: <strong>the</strong> river network, <strong>the</strong> reach and<br />
<strong>the</strong> focal area. Scale was considered important because <strong>in</strong>teractions between bio-physical and socio-economic<br />
factors and with <strong>the</strong>m priorities, change at different spatial scales. Biophysical factors <strong>in</strong>cluded characteristics like<br />
channel complexity and hydrology and fl oodpla<strong>in</strong> vegetation type. Socio-economic constra<strong>in</strong>ts <strong>in</strong>cluded factors such<br />
as private or state ownership of fl oodpla<strong>in</strong> land and population density. The units used for apply<strong>in</strong>g <strong>the</strong>se categories<br />
are ‘slices’ of fl oodpla<strong>in</strong> at right angles to fl ow, each one kilometre long. Sections of <strong>the</strong> river with low constra<strong>in</strong>ts<br />
and high opportunity were identifi ed us<strong>in</strong>g this process and enabled prioritisation of candidate river and fl oodpla<strong>in</strong><br />
restoration sites. The results are presented as maps with marked areas that <strong>in</strong>tegrate representation of processes with<br />
patterns (Hulse and Gregory, 2004) (Figure 6.6). This methodology was used <strong>in</strong> <strong>the</strong> generation of scenarios for <strong>the</strong><br />
alternative futures analysis <strong>in</strong> <strong>the</strong> Willamette Bas<strong>in</strong> and particularly <strong>in</strong> assessment of <strong>the</strong> sensitivity of endpo<strong>in</strong>ts <strong>in</strong><br />
<strong>the</strong> river valley (Baker et al., 2004).<br />
Figure 6.5 Percentage change <strong>in</strong> <strong>the</strong> ‘forested riparian’ <strong>in</strong>dicator of natural resource condition <strong>in</strong> <strong>the</strong> Willamette <strong>River</strong> Bas<strong>in</strong>, <strong>in</strong> <strong>the</strong><br />
historical and three future scenarios. The <strong>in</strong>dicator is <strong>the</strong> percentage of a 120-metre wide riparian buffer strip with forest vegetation<br />
along all streams <strong>in</strong> <strong>the</strong> valley ecoregion (Hulse et al., 2002). The future scenario labels are described as: Conservative – a low level<br />
of development <strong>in</strong>volv<strong>in</strong>g a high degree of natural resource protection; Plan Trend – a level of development consistent with current<br />
policies and trends; Development – refl ects a loosen<strong>in</strong>g of current policies to allow a freer re<strong>in</strong> to market forces across all landscape<br />
components but still with<strong>in</strong> <strong>the</strong> range of what stakeholders considered plausible (From Ecological Application (2004) EA14-2,<br />
pp. 313–324, fi gure 4. Repr<strong>in</strong>ted with permission from The Ecological Society of America.)
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Riparian and Floodpla<strong>in</strong> <strong>Restoration</strong> 89<br />
Figure 6.6 Graphical example of river reaches (1, 2 and 3) with co<strong>in</strong>cident low constra<strong>in</strong>t and high opportunity to restore channel<br />
complexity and native fl oodpla<strong>in</strong> forest. The units of (a) Population density, are people per square kilometre circa 1990 with<strong>in</strong> each<br />
one kilometre slice of <strong>the</strong> fl oodpla<strong>in</strong>. The units of (b), number of structures, are rural build<strong>in</strong>gs per square kilometre circa 1990 with<strong>in</strong><br />
each one kilometre slice of <strong>the</strong> fl oodpla<strong>in</strong>. The units of (c), loss of channel complexity, are net <strong>in</strong>crease or decrease <strong>in</strong> channel length<br />
with<strong>in</strong> each one kilometre slice of <strong>the</strong> fl oodpla<strong>in</strong> between 1850 and 1995. The units of (d), loss of fl oodpla<strong>in</strong> forest, are net decrease<br />
<strong>in</strong> area of fl oodpla<strong>in</strong> forest with<strong>in</strong> each one kilometre slice of <strong>the</strong> fl oodpla<strong>in</strong> between 1850 and 1990. (From Hulse & Gregory (2004)<br />
Urban Ecosystems 7 (3): 295–314. Repr<strong>in</strong>ted with k<strong>in</strong>d permission from Spr<strong>in</strong>ger Science and Bus<strong>in</strong>ess Media.)
90 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Although <strong>the</strong> number of river bas<strong>in</strong>s to which <strong>the</strong>se<br />
emerg<strong>in</strong>g holistic methodologies have been applied is<br />
<strong>in</strong>creas<strong>in</strong>g, <strong>the</strong>y rema<strong>in</strong> <strong>the</strong> exception and not <strong>the</strong> rule.<br />
Richter and Postel (2004) give 350 as <strong>the</strong> number of river<br />
bas<strong>in</strong>s worldwide that are listed on a data base held by <strong>the</strong><br />
United States’ Nature Conservancy and are optimistic that<br />
obstacles to implement<strong>in</strong>g <strong>the</strong>se approaches are <strong>in</strong>creas<strong>in</strong>gly<br />
be<strong>in</strong>g overcome. Very positive steps have been taken<br />
<strong>in</strong> Europe to apply more holistic approaches to river bas<strong>in</strong><br />
management through <strong>the</strong> European Union Water Framework<br />
Directive and its required <strong>River</strong> Bas<strong>in</strong> Management<br />
Plans. However, management of total water volume and<br />
its seasonal distribution (currently carried out by most<br />
national river management agencies <strong>in</strong> Europe) tend to be<br />
limited to <strong>the</strong> management of water quantities <strong>in</strong> rivers to<br />
satisfy requirements for pollution dilution and <strong>the</strong> ma<strong>in</strong>tenance<br />
of specifi ed m<strong>in</strong>imum fl ows. This management is<br />
largely effected through control of water abstraction<br />
licences (Figure 6.7). It is usual for management of low<br />
fl ows to be separated from management of fl oods and this<br />
is an area that needs to be addressed to achieve a more<br />
holistic approach.<br />
• Sediment management. While a huge amount of literature<br />
has appeared on <strong>the</strong> management of fl ows, far less<br />
research has been carried out on <strong>the</strong> management of<br />
sediments and <strong>the</strong>re are also few examples of projects<br />
that <strong>in</strong>volve restoration of sediment loads <strong>in</strong> rivers. In<br />
<strong>the</strong> European context, sediment loads have drastically<br />
changed over <strong>the</strong> last 200 years <strong>in</strong> response to changes<br />
<strong>in</strong> mass movements <strong>in</strong> upper catchments, to sand and<br />
gravel extraction <strong>in</strong> fl oodpla<strong>in</strong> zones, to <strong>the</strong> <strong>in</strong>stallation<br />
of upstream impoundments and to <strong>the</strong> armour<strong>in</strong>g of<br />
river banks with artifi cial dykes. In <strong>the</strong> Drôme <strong>River</strong> <strong>in</strong><br />
France measures are proposed to restore sediment loads<br />
to <strong>the</strong> river. These <strong>in</strong>clude a moratorium on gravel<br />
extraction and clearance of river bank vegetation to remobilise<br />
sediment through bank erosion (Michelot,<br />
1995). The aim is to improve <strong>the</strong> delivery of sediment<br />
to <strong>the</strong> Ramieres du Val de Drôme Nature Reserve, which<br />
is designated for its high quality fl oodpla<strong>in</strong> forest <strong>in</strong> two<br />
active, braided river reaches.<br />
6.4.2 Reach-Scale Management<br />
At <strong>the</strong> scale of <strong>the</strong> reach, <strong>the</strong>re are two ma<strong>in</strong> approaches<br />
to carry<strong>in</strong>g out river and riparian restoration. The fi rst<br />
<strong>in</strong>volves manag<strong>in</strong>g physical processes locally <strong>in</strong> <strong>the</strong> fl oodpla<strong>in</strong><br />
zone, for example by <strong>in</strong>troduc<strong>in</strong>g sluices <strong>in</strong>to side<br />
channels to control water levels <strong>in</strong> selected parts of a<br />
fl oodpla<strong>in</strong>. The second <strong>in</strong>volves manag<strong>in</strong>g <strong>the</strong> landforms<br />
<strong>in</strong> <strong>the</strong> fl oodpla<strong>in</strong> so that <strong>the</strong> physical disturbances act differently<br />
on each part of <strong>the</strong> fl oodpla<strong>in</strong>. Both types of disturbance<br />
management are more directed to specifi ed<br />
reaches of a river than manag<strong>in</strong>g fl ows of water and sediment<br />
at a catchment scale. However, <strong>the</strong>y have <strong>the</strong> disadvantage<br />
that <strong>the</strong>y can only have a relatively local effect. In<br />
most parts of Europe it is <strong>the</strong> type of management that can<br />
most readily be promoted and <strong>in</strong> many river bas<strong>in</strong>s it has<br />
already taken place through a number of river restoration<br />
projects. <strong>Restoration</strong> at this scale is easier to achieve than<br />
catchment-scale management for a variety of reasons:<br />
Figure 6.7 A typical environmental allocation for a UK catchment is shown <strong>in</strong> this graph (from Environment Agency, 2002). It<br />
varies considerably through <strong>the</strong> seasons and is determ<strong>in</strong>ed through a technical approach called <strong>the</strong> Resource Assessment and Management<br />
(RAM) framework. First each river reach <strong>in</strong> <strong>the</strong> catchment is assessed on <strong>the</strong> basis of its physical characteristics, fi sh populations,<br />
macrophytes and macro-<strong>in</strong>vertebrates to produce an environmental weight<strong>in</strong>g. Five environmental weight<strong>in</strong>g bands are used to classsify<br />
<strong>the</strong> sensitivity of each reach to <strong>the</strong> effects of water abstraction. The environmental weight<strong>in</strong>g is used with long term fl ow duration<br />
data to derive an Ecological Flow Objective and <strong>the</strong> percentage left for abstraction. The Ecological Flow Objective seeks to protect<br />
low fl ows and fl ow variability by allow<strong>in</strong>g percentages of fl ow bands to be abstracted. The impacts of groundwater abstraction (both<br />
seasonally and spatially) on river fl ows is also <strong>in</strong>corporated <strong>in</strong>to <strong>the</strong> process. To preserve water levels <strong>in</strong> fl oodpla<strong>in</strong> sites adjacent to<br />
rivers, a site-scale approach has been used <strong>in</strong> parts of <strong>the</strong> UK us<strong>in</strong>g Water Level Management Plans. These are usually applied to sites<br />
designated as wetland nature reserves whose water tables are related to those of adjacent river courses <strong>in</strong> ei<strong>the</strong>r a direct or <strong>in</strong>direct<br />
way
• It is more predictable than catchment-scale management<br />
of disturbances and as such is relatively easy to control<br />
and poses less uncerta<strong>in</strong>ties.<br />
• <strong>River</strong> managers are already used to manag<strong>in</strong>g rivers at<br />
a local scale through <strong>the</strong> eng<strong>in</strong>eer<strong>in</strong>g of fl ood defences.<br />
• It can neatly be fi tted <strong>in</strong>to chosen river sections between<br />
reaches where o<strong>the</strong>r human <strong>in</strong>terventions dom<strong>in</strong>ate.<br />
• It requires consensus among a much smaller group of<br />
stakeholders.<br />
Exampes of restoration at this scale are now common<br />
and can be categorised <strong>in</strong>to a series of activities which may<br />
be carried out s<strong>in</strong>gly or <strong>in</strong> comb<strong>in</strong>ation at any <strong>in</strong>dividual<br />
site (Hughes and Muller, 2003):<br />
• <strong>Restoration</strong> of <strong>the</strong> channel wetted perimeter. These projects<br />
have as a ma<strong>in</strong> aim <strong>the</strong> improvement of <strong>in</strong>stream<br />
habitats, often for fi sh populations. They emphasise<br />
<strong>in</strong>creas<strong>in</strong>g <strong>the</strong> heterogeneity of physical habitat. For<br />
example, <strong>the</strong> Ecological recovery of <strong>the</strong> V<strong>in</strong>del and Pitte<br />
<strong>River</strong>s (EVP) project, <strong>in</strong> nor<strong>the</strong>rn Sweden <strong>in</strong>volves<br />
removal of <strong>in</strong>stream structures <strong>in</strong>stalled dur<strong>in</strong>g <strong>the</strong> n<strong>in</strong>eteenth<br />
century for driv<strong>in</strong>g logs for <strong>the</strong> timber <strong>in</strong>dustry.<br />
In addition, large boulders that were removed for smooth<br />
fl oat<strong>in</strong>g of log rafts are be<strong>in</strong>g put back to diversify<br />
<strong>in</strong>stream habitats (Nilsson et al., 2005).<br />
• Re-connection of side arms <strong>in</strong> rivers. Most projects <strong>in</strong><br />
this category aim to remove one or more sections of<br />
artifi cial embankments to allow fl ows to penetrate fl oodpla<strong>in</strong><br />
zones that have been cut off. There are a number<br />
of examples of this approach that are well documented<br />
<strong>in</strong> <strong>the</strong> literature, such as <strong>the</strong> Regelsbrunner Au project<br />
on <strong>the</strong> <strong>River</strong> Danube <strong>in</strong> Austria (Scheimer et al., 1999)<br />
or <strong>the</strong> L’Ile de la Platière and Rosillon Channel sites on<br />
<strong>the</strong> <strong>River</strong> Rhône <strong>in</strong> France (Michelot, 1995; Downs<br />
et al., 2002). These projects had as a major objective <strong>the</strong><br />
restoration of both <strong>the</strong> quality of fl oodpla<strong>in</strong> forests and<br />
improved opportunites for forest regeneration by<br />
<strong>in</strong>creas<strong>in</strong>g lateral connectivities between <strong>the</strong> channel<br />
and <strong>the</strong> fl oodpla<strong>in</strong>.<br />
• Increase <strong>in</strong> fl oodpla<strong>in</strong> storage capacity through sett<strong>in</strong>g<br />
back defences, lower<strong>in</strong>g fl ood defences or lower<strong>in</strong>g <strong>the</strong><br />
fl oodpla<strong>in</strong>. There are major and well-documented plans<br />
for a range of <strong>the</strong>se activities <strong>in</strong> <strong>the</strong> distributaries of <strong>the</strong><br />
<strong>River</strong> Rh<strong>in</strong>e <strong>in</strong> The Ne<strong>the</strong>rlands and also <strong>in</strong> <strong>the</strong> <strong>River</strong><br />
Meuse. They form part of a master plan to <strong>in</strong>crease fl ood<br />
storage capacity <strong>in</strong> <strong>the</strong>se rivers <strong>in</strong> anticipation of sealevel<br />
rise and <strong>in</strong>creas<strong>in</strong>g frequency of fl oods from<br />
upstream (Middelkoop and van Haselen, 1999). These<br />
activities are currently restricted to a series of discrete<br />
sites, such as at <strong>the</strong> Mill<strong>in</strong>gerwaard Nature Reserve <strong>in</strong><br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Riparian and Floodpla<strong>in</strong> <strong>Restoration</strong> 91<br />
The Ne<strong>the</strong>rlands where fl ood defences have been<br />
lowered. The general approach <strong>in</strong> The Ne<strong>the</strong>rlands is<br />
shown <strong>in</strong> Figure 6.8. In <strong>the</strong> United K<strong>in</strong>gdom fl ood<br />
storage washlands are proposed along some rivers but<br />
<strong>in</strong> all <strong>the</strong>se projects <strong>the</strong> design of dra<strong>in</strong>age dykes and<br />
control structures will be very important <strong>in</strong> determ<strong>in</strong>g<br />
<strong>the</strong> amount of control on <strong>the</strong> length of time and depth<br />
that water stays <strong>in</strong> <strong>the</strong> washland. Creative confi guration<br />
of <strong>the</strong> fl oodpla<strong>in</strong> surface can mimic natural fl oodpla<strong>in</strong><br />
habitat heterogeneity to give conservation ga<strong>in</strong>s as well<br />
as fl ood storage ga<strong>in</strong>s if water control is fl exible enough<br />
to operate for <strong>the</strong> good of wetland ecosystems as well<br />
as for fl ood control.<br />
• Management of <strong>the</strong> river’s sediment load. Sedimentation<br />
is an essential process along <strong>the</strong> marg<strong>in</strong>s of river channels<br />
as newly created alluvial bars are prime regeneration<br />
sites for many species of fl oodpla<strong>in</strong> vegetation.<br />
Groynes can be used to create artifi cial ‘beaches’<br />
although <strong>the</strong>ir primary purpose is usually to ma<strong>in</strong>ta<strong>in</strong> a<br />
channel for navigation. Re-activation of erosion <strong>in</strong> sites<br />
where embankments have been removed will alter <strong>the</strong><br />
sediment loads downstream.<br />
• Management of riparian vegetation. This can take <strong>the</strong><br />
form of plant<strong>in</strong>g fl oodpla<strong>in</strong> forests or wait<strong>in</strong>g for natural<br />
regeneration to take place. In ei<strong>the</strong>r case, management<br />
of <strong>the</strong> vegetation that grows may be considered necessary,<br />
though this is not <strong>in</strong> <strong>the</strong> spirit of self-susta<strong>in</strong>ability.<br />
In many river bas<strong>in</strong>s, graz<strong>in</strong>g by domestic animals <strong>in</strong><br />
riparian zones prevents natural regeneration and fenc<strong>in</strong>g<br />
Figure 6.8 In The Ne<strong>the</strong>rlands, large parts of <strong>the</strong> Rh<strong>in</strong>e fl oodpla<strong>in</strong><br />
will be lowered as part of <strong>the</strong> Flood Action Plans drawn up<br />
by <strong>the</strong> International Rh<strong>in</strong>e Committee. The aim is to <strong>in</strong>crease<br />
fl ood storage capacity of <strong>the</strong> Rh<strong>in</strong>e fl oodpla<strong>in</strong>s. The proposed<br />
works are shown schematically <strong>in</strong> this diagram (from Middelkoop<br />
and Van Haselen, 1999)
92 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
off <strong>the</strong> riparian zone becomes an important form of<br />
management. Management of <strong>the</strong> vegetation also has<br />
implications for <strong>the</strong> volume of woody debris that arrives<br />
<strong>in</strong> a river. MacNally et al. (2002) suggest that <strong>in</strong> <strong>the</strong><br />
Murray-Darl<strong>in</strong>g <strong>River</strong> system <strong>in</strong> Australia only 15% of<br />
<strong>the</strong> former woody debris load is now present. There are<br />
projects where woody debris has been put back <strong>in</strong>to<br />
rivers and some where artifi cial log jams have been built<br />
to test <strong>the</strong>ir restoration effect (Dewberry et al., 1998).<br />
6.4.3 The Use of Reference Systems <strong>in</strong> <strong>River</strong><br />
<strong>Restoration</strong> Projects<br />
It is common to set objectives for river restoration projects<br />
by us<strong>in</strong>g reference systems. In <strong>the</strong> EU Water Framework<br />
Directive (2000/60/EC) it is a major and perhaps unrealistic<br />
aim to use a network of near-natural reference systems<br />
so that <strong>the</strong> ecological status of rivers can be measured<br />
aga<strong>in</strong>st <strong>the</strong>m. In essence this <strong>in</strong>volves fi nd<strong>in</strong>g a river<br />
system that has <strong>the</strong> attributes considered desirable <strong>in</strong> <strong>the</strong><br />
restored system and us<strong>in</strong>g it as a template on which to base<br />
<strong>the</strong> restoration activities. Criteria specifi cally for establish<strong>in</strong>g<br />
riparian reference conditions are proposed by Harris<br />
(1999) us<strong>in</strong>g multivariate analyses of aspects of vegetation<br />
community composition and structure. Typically a reference<br />
system will be part of or <strong>the</strong> whole of a less-damaged<br />
river system, preferably located <strong>in</strong> a similar bioclimatic<br />
zone and <strong>in</strong> a river bas<strong>in</strong> exhibit<strong>in</strong>g similar physiographic<br />
characteristics. Alternatively it can be an historic system,<br />
whose attributes are known from maps, old photographs<br />
or written accounts. A critique of <strong>the</strong> use of reference<br />
systems to defi ne objectives <strong>in</strong> river restoration is given <strong>in</strong><br />
Hughes et al. (2005), where <strong>the</strong> follow<strong>in</strong>g categories of<br />
problems encountered with <strong>the</strong> use of reference systems<br />
are discussed:<br />
• There are often no appropriate reference systems to<br />
use.<br />
• Many catchment parameters have changed s<strong>in</strong>ce <strong>the</strong><br />
times of chosen historic reference systems.<br />
• Climate change has been cont<strong>in</strong>uous through <strong>the</strong><br />
Holocene.<br />
• Projected climate change is of uncerta<strong>in</strong> magnitude.<br />
• Alien species have become common <strong>in</strong> <strong>the</strong> landscape<br />
and cannot be avoided.<br />
• Landscape context changes through time.<br />
Us<strong>in</strong>g reference systems can give river managers a misplaced<br />
confi dence <strong>in</strong> <strong>the</strong> predictability of ecological outcomes<br />
<strong>in</strong> river restoration projects (Hughes et al., 2005),<br />
although <strong>the</strong> degree to which <strong>the</strong> project outcomes and <strong>the</strong><br />
reference system co<strong>in</strong>cide can largely be decided at <strong>the</strong><br />
outset when objectives are set (Simons and Boeters, 1998).<br />
Evaluation of restoration projects is also usually carried<br />
out aga<strong>in</strong>st <strong>the</strong> reference system, although with highly<br />
variable levels of rigour (Anderson and Dugger, 1998;<br />
Stream Corridor Work<strong>in</strong>g Group, 1998).<br />
Never<strong>the</strong>less, reference systems can provide useful,<br />
broad guid<strong>in</strong>g images for restoration. In <strong>the</strong> United<br />
K<strong>in</strong>gdom, a number of sources of <strong>in</strong>formation are used.<br />
At <strong>the</strong> catchment scale, <strong>the</strong> National Vegetation Classifi cation<br />
(Rodwell, 1991a, 1991b, 1992, 1995, 2000), The<br />
National Biodiversity Network (NBN), <strong>the</strong> Multi-Agency<br />
Geographic Information for <strong>the</strong> Countryside (MAGIC)<br />
databases on Habitats and Sites of Special Scientifi c <strong>in</strong>terest<br />
(SSSI’s), provide <strong>in</strong>formation on <strong>the</strong> distribution of<br />
habitats, communities and species across <strong>the</strong> United<br />
K<strong>in</strong>gdom. For river reaches, <strong>the</strong>re may be <strong>River</strong> Morphology,<br />
<strong>River</strong> Habitat and <strong>River</strong> Corridor Surveys, and<br />
Hydro-Morphology Quality Assessments that have already<br />
been undertaken <strong>in</strong> some relatively undisturbed reaches<br />
(RSPB, NRA, RSNC, 1994). Us<strong>in</strong>g data on <strong>the</strong> catchment<br />
area, slope and river corridor width of <strong>the</strong> reaches, it may<br />
be possible to extrapolate to o<strong>the</strong>r river reaches with<br />
similar characteristics. Overall, <strong>the</strong> aim should be to<br />
develop a sense of how <strong>the</strong> spatial structure of <strong>the</strong> catchment<br />
hydrology and river morphology and ecology are<br />
related, and how <strong>the</strong> catchment and river function under<br />
‘natural’ conditions, bear<strong>in</strong>g <strong>in</strong> m<strong>in</strong>d that any system is <strong>in</strong><br />
a transient state over time.<br />
6.5 MONITORING AND EVALUATING<br />
RESTORATION<br />
6.5.1 What is Ecological Success and How Do We<br />
Evaluate It?<br />
A widely applicable scheme for evaluat<strong>in</strong>g <strong>the</strong> ecological<br />
success of river restoration projects does not currently<br />
exist, though many authors have emphasised <strong>the</strong> need for<br />
such a scheme (NRC, 1992; Downs et al., 2002). Part of<br />
<strong>the</strong> problem is <strong>in</strong> defi n<strong>in</strong>g exactly what is meant by ecological<br />
success and recently an attempt has been made to<br />
identify fi ve criteria that can be used for its measurement<br />
(Palmer et al., 2005):<br />
1. The design of an ecological river restoration project<br />
should be based on a specifi ed guid<strong>in</strong>g image of a more<br />
dynamic, healthy river that could exist at <strong>the</strong> site.<br />
2. The river’s ecological condition must be measurably<br />
improved.<br />
3. The river system must be more self-susta<strong>in</strong><strong>in</strong>g and<br />
resilient to external perturbations so that only m<strong>in</strong>imal<br />
follow-up ma<strong>in</strong>tenance is needed.
4. Dur<strong>in</strong>g <strong>the</strong> construction phase, no last<strong>in</strong>g harm is<br />
<strong>in</strong>fl icted on <strong>the</strong> ecosystem.<br />
5. Both pre- and post-construction ecological assessment<br />
is carried out and <strong>the</strong> <strong>in</strong>formation made available.<br />
Refi nements to this scheme have been proposed by<br />
Jansson et al. (2005) and Gillilan et al. (2005), <strong>in</strong>clud<strong>in</strong>g<br />
<strong>the</strong> need to <strong>in</strong>clude stated ecological mechanisms by<br />
which any <strong>in</strong>tended restoration will reach its goal. A major<br />
diffi culty with this scheme is giv<strong>in</strong>g practical mean<strong>in</strong>g to<br />
<strong>the</strong> concepts of resiliency and self-susta<strong>in</strong>ability, pr<strong>in</strong>cipally<br />
because <strong>the</strong>se are entirely relative terms that are also<br />
closely l<strong>in</strong>ked to <strong>the</strong> concept of dynamic equilibrium.<br />
The term ‘resilience’ was fi rst co<strong>in</strong>ed <strong>in</strong> <strong>the</strong> 1970s <strong>in</strong><br />
<strong>the</strong> fi eld of population ecology and used to characterise<br />
<strong>the</strong> magnitude of population perturbations a system could<br />
tolerate before chang<strong>in</strong>g <strong>in</strong>to some qualitatively different<br />
dynamic state (Holl<strong>in</strong>g, 1973; May, 1976a). More recently<br />
it has been used to describe <strong>the</strong> ability of an ecosystem to<br />
rega<strong>in</strong> a functional state follow<strong>in</strong>g disturbance and <strong>the</strong><br />
rapidity of this process is a measure of its resilience<br />
(War<strong>in</strong>g, 1989). The disturbance is usually temporary and<br />
if it is suffi ciently regular, component species of <strong>the</strong> ecosystem<br />
often evolve a dependence on it to complete <strong>the</strong>ir<br />
life cycles. Both ecological and evolutionary time scales<br />
must be considered <strong>in</strong> assess<strong>in</strong>g <strong>the</strong> signifi cance of disturbances<br />
of different magnitudes and frequencies <strong>in</strong> river<br />
and riparian ecosystems (Poff, 1992; Hughes, 1994; Dodds<br />
et al., 2004). As described earlier <strong>in</strong> this chapter, <strong>in</strong> <strong>the</strong><br />
case of riparian and fl oodpla<strong>in</strong> ecosystems, <strong>the</strong> disturbance<br />
is provided by fl oods and many riparian species<br />
have <strong>in</strong>deed evolved a dependence on <strong>the</strong>se fl oods for <strong>the</strong><br />
regeneration phase of <strong>the</strong>ir life cycles. ‘Dynamic equilibrium’<br />
encompasses <strong>the</strong> notion of chang<strong>in</strong>g parameters<br />
with<strong>in</strong> a stable framework and has often been used <strong>in</strong><br />
descriptions of whole ecosystems. If <strong>the</strong> stable framework<br />
changes <strong>the</strong>n a qualitatively different dynamic state aga<strong>in</strong><br />
prevails. The defi nition of a stable framework is usually<br />
determ<strong>in</strong>ed by <strong>the</strong> temporal and spatial scales of an <strong>in</strong>vestigation<br />
and often limited by an <strong>in</strong>vestigator’s consideration<br />
of change through time and over space. However, as<br />
stated by May (1976b), <strong>in</strong> <strong>the</strong> real world <strong>the</strong>re are no fi xed<br />
parameter values; environmental parameters and <strong>in</strong>teractions<br />
between organisms and between organisms and <strong>the</strong>ir<br />
environment are constantly fl uctuat<strong>in</strong>g. It follows that<br />
stable frameworks do not exist <strong>in</strong> <strong>the</strong> real world unless<br />
<strong>the</strong>y are artifi cially delimited spatially, temporally or<br />
both.<br />
Measur<strong>in</strong>g resilience to evaluate how successful we<br />
have been <strong>in</strong> return<strong>in</strong>g it to a riparian or fl oodpla<strong>in</strong> ecosystem<br />
dur<strong>in</strong>g a river restoration project is very diffi cult<br />
to do. If a framework has been identifi ed to describe a<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Riparian and Floodpla<strong>in</strong> <strong>Restoration</strong> 93<br />
dynamic equilibrium, <strong>the</strong>n with<strong>in</strong> <strong>the</strong> defi ned framework<br />
it is possible to predict and measure recovery of a disturbed<br />
ecosystem. However, it is a huge challenge to<br />
predict <strong>the</strong> functions of ecosystems when <strong>the</strong>y beg<strong>in</strong> to<br />
fl uctuate along new trajectories caused by environmental<br />
change or <strong>in</strong>clude new species arrivals. S<strong>in</strong>ce all ecosystems,<br />
<strong>in</strong>clud<strong>in</strong>g restored ecosystems, are mov<strong>in</strong>g along<br />
fl uctuat<strong>in</strong>g trajectories, it follows that, conceptually, it is<br />
impossible to measure resilience and <strong>the</strong>refore to evaluate<br />
ecological success us<strong>in</strong>g this criterion (see Chapters 8 and<br />
11 for fur<strong>the</strong>r discussion). In practical terms it is possible<br />
to del<strong>in</strong>eate a framework for measur<strong>in</strong>g resilience though<br />
different types of ecologists might defi ne very different<br />
frameworks for <strong>the</strong> same restoration project. A fl oodpla<strong>in</strong><br />
forest ecologist might defi ne a framework where selfsusta<strong>in</strong>ability<br />
is measured <strong>in</strong> terms of turnover rates<br />
for fl oodpla<strong>in</strong> habitats (10 2 –10 3 years) related to fl ood<br />
return periods (Mahoney and Rood, 1998; Hughes and<br />
Rood, 2001). A fi sheries ecologist might assess selfsusta<strong>in</strong>ability<br />
<strong>in</strong> terms of provision of <strong>in</strong>stream habitats<br />
that permit completion of fi sh life cycles over annual or<br />
10 1 year time frames (e.g. Frissel and Nawa, 1992).<br />
The evaluation of success us<strong>in</strong>g such measurements<br />
also changes with scale and is discussed with respect to<br />
fl oodpla<strong>in</strong> forests by Hughes et al. (2005). When viewed<br />
at a small spatial scale (10 2 metres), habitat patches <strong>in</strong> <strong>the</strong><br />
forest will change from year to year as a response to<br />
channel movement and species arrival, death or migration.<br />
At a broader spatial scale of a reach or whole fl oodpla<strong>in</strong>,<br />
variability becomes less pronounced because <strong>the</strong> balance<br />
of different habitat patches rema<strong>in</strong>s more constant (Figures<br />
6.9(a) and 6.9(b)), particularly over short time frames of<br />
10 1 to 10 2 years (see Ward et al., 2002). However, over<br />
longer periods (10 3 years), this balance at <strong>the</strong> reach scale<br />
might shift <strong>in</strong> response to climate change, sea level change,<br />
isostatic uplift, change <strong>in</strong> availability of propagules or<br />
changes <strong>in</strong> <strong>the</strong> biophysical attributes of <strong>the</strong> catchment<br />
(Figure 6.9(c)). In this scenario, shift<strong>in</strong>g geomorphological<br />
processes (for example from aggradation to downcutt<strong>in</strong>g<br />
of <strong>the</strong> river valley) and changed channel patterns<br />
cause major changes to ecological vectors and patterns <strong>in</strong><br />
<strong>the</strong> fl oodpla<strong>in</strong>. Evaluation of success can only be relative<br />
to <strong>the</strong>se chang<strong>in</strong>g frameworks or else with<strong>in</strong> <strong>the</strong> context<br />
of a framework that has been held stable.<br />
6.5.2 What is Ecological Quality and How Does it<br />
Relate to Ecological Success?<br />
Closely related to <strong>the</strong> consideration of how to measure<br />
ecological success is <strong>the</strong> need to give mean<strong>in</strong>g to <strong>the</strong><br />
concept of ecological quality. Both concepts are dependent<br />
on <strong>the</strong> monitor<strong>in</strong>g or surveillance of a series of physical
94 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Figure 6.9 At <strong>the</strong> scale of a whole fl oodpla<strong>in</strong>, progressive or rapid changes can take place <strong>in</strong> <strong>the</strong> distribution of fl oodpla<strong>in</strong> vegetation<br />
communities. Thus <strong>in</strong> 6.9(a), t1 is biodiversity at an <strong>in</strong>itial time period and consists of vegetation communities a and b, present<br />
<strong>in</strong> <strong>the</strong> proportion of 4a to 3b. In 6.9(b), t2 is biodiversity at a later time. Vegetation communities a and b are still present <strong>in</strong> <strong>the</strong> same<br />
balance but <strong>the</strong>y are all <strong>in</strong> different locations follow<strong>in</strong>g shifts <strong>in</strong> channel location. Over much longer time frames (or over rapid time<br />
frames follow<strong>in</strong>g extreme events or human <strong>in</strong>tervention), <strong>the</strong>re may be changes <strong>in</strong> catchment parameters that alter <strong>the</strong> geomorphological<br />
patterns and hydrological activity of <strong>the</strong> river. In 6.9(c), <strong>the</strong> meander<strong>in</strong>g river has become braided and a new vegetation community<br />
c has arrived. The biodiversity (t3) of fl oodpla<strong>in</strong> vegetation communities has now changed (from Hughes et al., 2005)<br />
and biological parameters that are representative of ecosystem<br />
function. Whereas <strong>the</strong> fi rst is tied to <strong>the</strong> evaluation<br />
of a restoration project, <strong>the</strong> second is descriptive of <strong>the</strong><br />
state of an ecosystem. However, <strong>the</strong>y are <strong>in</strong>terrelated <strong>in</strong><br />
functional terms and decisions on what to monitor to<br />
measure ecological quality will have a signifi cant impact<br />
on our ability to evaluate ecological success <strong>in</strong> river restoration,<br />
s<strong>in</strong>ce this will depend on pre- and post-project<br />
measurements of ecosystem-relevant parameters. In<br />
Europe, <strong>the</strong> Water Framework Directive (WFD) (2000/60/<br />
EC) provides a legislative framework for water policy and<br />
among o<strong>the</strong>r th<strong>in</strong>gs aims to ma<strong>in</strong>ta<strong>in</strong> and improve <strong>the</strong><br />
aquatic environment through attention to both water<br />
quality and quantity. It requires that member states ensure<br />
good ‘chemical status’ and ‘ecological status’ of surface<br />
waters. Groundwaters must meet ‘good groundwater<br />
status’. There is no objective for riparian and fl oodpla<strong>in</strong><br />
ecosystems; however, <strong>the</strong> health of groundwaterdependent<br />
wetlands is an <strong>in</strong>dicator of ‘good groundwater<br />
status’. ‘Ecological status’ is described as ‘an expression<br />
of <strong>the</strong> quality of <strong>the</strong> structure and function<strong>in</strong>g of aquatic<br />
ecosystems associated with surface waters’ and member<br />
states are required to monitor this status. To do this <strong>the</strong>re<br />
has to be clear understand<strong>in</strong>g of what ‘ecological status’<br />
means, and much debate about this has followed publication<br />
of <strong>the</strong> Water Framework Directive (WFD) both at<br />
European Union and national levels.<br />
The ma<strong>in</strong> descriptors of ecological quality for rivers are<br />
grouped <strong>in</strong>to biological, hydrogeomorphological and<br />
physico-chemical elements (Table 6.3). Soon after adoption<br />
of <strong>the</strong> WFD (22nd December, 2000), <strong>the</strong> fi rst volume<br />
of Common Implementation Strategy (CIS) guidance was<br />
produced. Its ma<strong>in</strong> aim is to help member states share<br />
expertise on implementation of <strong>the</strong> technical aspects of <strong>the</strong><br />
WFD, <strong>in</strong>clud<strong>in</strong>g achievement of good ecological status for<br />
surface waters. Annex V of <strong>the</strong> WFD <strong>in</strong>cludes normative<br />
defi nitions of <strong>the</strong> three ecological status categories (High,<br />
Good and Moderate). The levels at which <strong>the</strong>se three categories<br />
will be set is <strong>the</strong> subject of a major EU <strong>in</strong>tercalibration<br />
process that should have been completed <strong>in</strong>
2006, but which rema<strong>in</strong>s contentious and partial. This<br />
process is be<strong>in</strong>g carried out by <strong>the</strong> ‘European Centre for<br />
Ecological Water Quality and Intercalibration’ (EEWAI),<br />
which aims to compare <strong>the</strong> different national classifi cation<br />
systems for ecological status assessment. Two key parts of<br />
<strong>the</strong> process are identifi cation of reference conditions and<br />
establishment of monitor<strong>in</strong>g protocols. It is already known<br />
that river hydromorphology will have to be restored<br />
(unless <strong>the</strong> river stretch is designated ‘heavily modifi ed’,<br />
i.e. restoration would adversely affect its function for<br />
navigation, fl ood protection or power generation) if it is<br />
prevent<strong>in</strong>g biological elements from achiev<strong>in</strong>g ‘good ecological<br />
status’ (Withr<strong>in</strong>gton, Natural England, personal<br />
communication).<br />
Individual member states have to <strong>in</strong>tegrate implementation<br />
of <strong>the</strong> WFD with <strong>the</strong>ir already established protocols<br />
for monitor<strong>in</strong>g and evaluat<strong>in</strong>g <strong>the</strong> status of <strong>the</strong>ir environment,<br />
species and habitats. In <strong>the</strong> United K<strong>in</strong>gdom, <strong>the</strong><br />
statutory conservation agencies have recently <strong>in</strong>troduced<br />
a system of ‘Common Standards Monitor<strong>in</strong>g’ (CSM) for<br />
sites designated under national legislation and European<br />
directives. Site-based conservation is a signifi cant part of<br />
biodiversity and earth science conservation <strong>in</strong> <strong>the</strong> United<br />
K<strong>in</strong>gdom and evaluation of <strong>the</strong> effectiveness of measures<br />
put <strong>in</strong>to place to achieve biodiversity conservation is <strong>the</strong><br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Riparian and Floodpla<strong>in</strong> <strong>Restoration</strong> 95<br />
Table 6.3 Quality elements for <strong>the</strong> classifi cation of Ecological Status-<strong>River</strong>s (EU Water Framework Directive, 2000/60/EC,<br />
Annex V)<br />
Quality Element Characteristics<br />
Biological elements • Composition and abundance of aquatic fl ora<br />
• Composition and abundance of benthic <strong>in</strong>vertebrate fauna<br />
• Composition, abundance and age structure of fi sh fauna<br />
Hydrogeomorphological elements support<strong>in</strong>g<br />
<strong>the</strong> biological elements<br />
Chemical and physico-chemical elements<br />
support<strong>in</strong>g <strong>the</strong> biological elements<br />
• Hydrological regime<br />
– quantity and dynamics of water fl ow<br />
– connection to groundwater bodies<br />
• <strong>River</strong> cont<strong>in</strong>uity<br />
• Morphological conditions<br />
– river depth and width variation<br />
– structure and substrate of <strong>the</strong> river bed<br />
– structure of <strong>the</strong> riparian zone<br />
• General<br />
– <strong>the</strong>rmal conditions<br />
– oxygenation conditions<br />
– sal<strong>in</strong>ity<br />
– acidifi cation status<br />
– nutrient conditions<br />
• Specifi c pollutants<br />
– pollution by all priority substances identifi ed as be<strong>in</strong>g<br />
discharged <strong>in</strong>to <strong>the</strong> body of water<br />
– pollution by o<strong>the</strong>r substances identifi ed as be<strong>in</strong>g discharged <strong>in</strong><br />
signifi cant quantities <strong>in</strong>to <strong>the</strong> body of water<br />
ma<strong>in</strong> aim of <strong>the</strong> CSM process. The CSM guidance for<br />
monitor<strong>in</strong>g rivers designated as important by national legislation<br />
(Sites of Special Scientifi c Interest (SSSIs) under<br />
<strong>the</strong> Wildlife and Countryside Act, 1981) recommends <strong>the</strong><br />
use of fl uvial geomorphological audit to defi ne modifi cations<br />
to rivers and identify options for river restoration<br />
(JNCC, 2004). Riparian zones are <strong>in</strong>cluded <strong>in</strong> <strong>the</strong> CSM<br />
guidance for rivers, both <strong>in</strong> <strong>the</strong>ir own right as woodland,<br />
grassland or swamp communities and for <strong>the</strong> contribution<br />
<strong>the</strong>y make to <strong>in</strong>-channel river communities. Flow levels<br />
and patterns are mostly assessed <strong>in</strong> terms of <strong>the</strong>ir importance<br />
to <strong>in</strong>stream species and habitats ra<strong>the</strong>r than to adjacent<br />
terrestrial ecosystems. The functional attributes of<br />
riparian zones is considered through measurement of<br />
factors such as production of woody debris and leaf litter<br />
as is <strong>the</strong>ir <strong>in</strong>tr<strong>in</strong>sic habitat value, for example bird, mammal<br />
or <strong>in</strong>vertebrate habitat. Such monitor<strong>in</strong>g schemes for designated<br />
sites will be <strong>in</strong>tegrated with<strong>in</strong> broader-scale, river<br />
bas<strong>in</strong> wide monitor<strong>in</strong>g of ecological quality, under <strong>the</strong><br />
WFD, of both aquatic and groundwater-dependent terrestrial<br />
ecosystems.<br />
The design and density of monitor<strong>in</strong>g networks eventually<br />
established on European rivers will determ<strong>in</strong>e<br />
how useful <strong>the</strong>y are for also evaluat<strong>in</strong>g river restoration<br />
success. Numerous research projects funded by <strong>the</strong>
96 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
European Commission are <strong>in</strong> <strong>the</strong> process of explor<strong>in</strong>g<br />
monitor<strong>in</strong>g protocols for particular groups of species, such<br />
as macro<strong>in</strong>vertebrates and fi sh, with a particular emphasis<br />
on <strong>the</strong>ir transferability, comparability and ability to <strong>in</strong>dicate<br />
ecological quality. For example, <strong>the</strong> LIFE <strong>in</strong> UK<br />
<strong>River</strong>s Project (1998–2003) produced monitor<strong>in</strong>g protocols<br />
for rivers with Callitricho-Batrachion vegetation and<br />
for 10 river-dependent species such as <strong>the</strong> otter (Lutra<br />
lutra). O<strong>the</strong>rs are concerned with defi nition of reference<br />
conditions, tak<strong>in</strong>g <strong>in</strong>to account climate change (e.g.<br />
EUROLIMPACTS).<br />
6.5.3 How Do We Value Biodiversity?<br />
Most commonly, use of <strong>the</strong> term biodiversity <strong>in</strong>dicates a<br />
measure of species numbers, although habitat diversity<br />
and genetic diversity are also vitally important (Heywood,<br />
1995; Gopal and Junk, 2001). It is relatively straightforward<br />
to measure species diversity compared with ecological<br />
and genetic diversity, although <strong>the</strong> signifi cance of<br />
chang<strong>in</strong>g biodiversity rema<strong>in</strong>s elusive <strong>in</strong> all cases. In<br />
riparian and fl oodpla<strong>in</strong> habitats, levels of species diversity<br />
are related to both <strong>the</strong> ecotonal situation of <strong>the</strong>se habitats<br />
and to <strong>the</strong>ir physical heterogeneity. Lateral connectivity<br />
between riparian habitats and channels and longitud<strong>in</strong>al<br />
connectivities between upper and lower parts of river<br />
bas<strong>in</strong>s all contribute to high levels of biodiversity (Tabacchi<br />
et al., 1996). Less well understood are <strong>the</strong> vertical<br />
connectivities between groundwater zones and riparian<br />
zones but <strong>the</strong>y contribute <strong>in</strong> complex ways to habitat<br />
quality <strong>in</strong> <strong>in</strong>terstices of fl oodpla<strong>in</strong> sediments (Lambs,<br />
2004) and to diversity of organisms <strong>in</strong> <strong>the</strong> hyporheic zone<br />
(Gibert et al., 1997; Hancock et al., 2005).<br />
Species composition on fl oodpla<strong>in</strong>s has changed over<br />
long and short time scales as shown by chang<strong>in</strong>g pollen<br />
profi les from peat deposits <strong>in</strong> fl oodpla<strong>in</strong> zones (Godw<strong>in</strong>,<br />
1941), respond<strong>in</strong>g to climate change and <strong>the</strong> progression<br />
of species through <strong>the</strong> landscape. In <strong>the</strong> last few centuries,<br />
many new or alien species have arrived on fl oodpla<strong>in</strong>s as<br />
a result of <strong>in</strong>troductions and garden escapes. Because<br />
riparian and fl oodpla<strong>in</strong> zones are physically highly<br />
dynamic, and are populated by species able to cope with<br />
habitat mobility, <strong>the</strong>y provide excellent habitats for ecological<br />
pioneers, which many <strong>in</strong>vasive species can be classifi<br />
ed as (Planty-Tabacchi et al., 1996). These species have<br />
often out-competed native species because hydrological<br />
and geomorphological factors have changed and <strong>the</strong><br />
species that have evolved <strong>the</strong>ir life cycles to fi t <strong>in</strong> with<br />
natural hydrological cycles are no longer favoured. In river<br />
restoration projects today, <strong>the</strong> presence of alien species<br />
can be viewed <strong>in</strong> several ways. They can ei<strong>the</strong>r be eradicated<br />
by active processes, such as clearance, or <strong>the</strong>y can<br />
have conditions for <strong>the</strong>ir regeneration made diffi cult by<br />
alter<strong>in</strong>g physical <strong>in</strong>puts, such as fl ood tim<strong>in</strong>g to reduce<br />
<strong>the</strong>ir competitive edge. On <strong>the</strong> o<strong>the</strong>r hand, <strong>the</strong> number of<br />
alien species present <strong>in</strong> many river systems makes <strong>the</strong>ir<br />
complete eradication impossible, and acceptance of some<br />
of <strong>the</strong>m as components of <strong>the</strong> ecosystems found <strong>in</strong> riparian<br />
zones is ano<strong>the</strong>r approach. It is easier to justify this<br />
approach <strong>in</strong> view of predicted climate change, which will<br />
signifi cantly alter <strong>the</strong> range that many species can occupy<br />
(Jensen, 2004) and <strong>the</strong> hydrological patterns <strong>in</strong> river bas<strong>in</strong>s<br />
(Montgomery and Boulton, 2003) but <strong>in</strong> ways that are<br />
currently uncerta<strong>in</strong>. It is certa<strong>in</strong> that species assemblages<br />
<strong>in</strong> <strong>the</strong> landscape will change and perhaps have no presentday<br />
analogues and that an open-m<strong>in</strong>ded approach to<br />
acceptable community composition will have to be taken<br />
(Hughes et al., 2005).<br />
6.6 INTERCONNECTIVITY AND BOUNDARY<br />
CROSSING: INSTITUTIONAL COMPLEXITIES<br />
OF RESTORING FLOODPLAINS<br />
The task of restor<strong>in</strong>g fl oodpla<strong>in</strong>s poses multi-dimensional<br />
challenges to policy makers and project managers alike.<br />
Involv<strong>in</strong>g essentially a reconfi guration of <strong>the</strong> <strong>in</strong>teraction<br />
between a river and adjacent low-ly<strong>in</strong>g land, fl oodpla<strong>in</strong><br />
restoration has far-reach<strong>in</strong>g implications for exist<strong>in</strong>g<br />
forms of water and land use. Floodpla<strong>in</strong>s provide multiple<br />
functions and services for humans as well as <strong>the</strong> natural<br />
environment. These can range from valuable artefacts for<br />
socio-economic reproduction, such as crops, timber, water<br />
or prime land for development, to less tangible but equally<br />
valuable functions, such as protection from fl ood<strong>in</strong>g,<br />
attractive landscapes or opportunities for recreational pursuits.<br />
The restoration of functional fl oodpla<strong>in</strong>s requires<br />
changes to exist<strong>in</strong>g activities on <strong>the</strong> site of <strong>the</strong> fl oodpla<strong>in</strong><br />
itself, but also – particularly <strong>in</strong> <strong>the</strong> case of larger schemes<br />
– along whole reaches of a river and even a whole catchment.<br />
On this wider scale it can signifi cantly <strong>in</strong>fl uence, for<br />
<strong>in</strong>stance, levels of fl ood protection, <strong>the</strong> navigability of a<br />
river reach or <strong>the</strong> viability of current farm<strong>in</strong>g practices. In<br />
this way, fl oodpla<strong>in</strong> restoration affects a wide range of<br />
stakeholders and <strong>in</strong>terests (Adams and Perrow, 1999;<br />
Adams et al., 2004; Turner et al., 2000; Adger and Luttrell,<br />
2000), mak<strong>in</strong>g it potentially highly controversial.<br />
Beh<strong>in</strong>d <strong>the</strong>se stakeholders and <strong>in</strong>terests lie <strong>in</strong>stitutions<br />
– understood here as rule systems – designed to<br />
protect and provide a variety of private and public goods,<br />
rang<strong>in</strong>g from commercial products to rights of access. For<br />
each of <strong>the</strong> policy fi elds affected by fl oodpla<strong>in</strong> restoration<br />
– primarily water protection, fl ood defence, nature conservation,<br />
recreation, navigation, urban and rural development<br />
– complex <strong>in</strong>stitutional arrangements have been
designed and adapted over <strong>the</strong> years. Each <strong>in</strong>stitutional<br />
arrangement comprises a set of codifi ed norms (such as<br />
laws, regulations and contractual obligations), plann<strong>in</strong>g<br />
<strong>in</strong>struments and fund<strong>in</strong>g mechanisms, as well as standardised<br />
procedures of operation, values and accepted<br />
practices of <strong>the</strong> relevant organised and <strong>in</strong>dividual actors.<br />
A scheme to restore a fl oodpla<strong>in</strong> requires <strong>the</strong> successful<br />
enrolment of <strong>the</strong>se <strong>in</strong>stitutions and organisations <strong>in</strong> such<br />
a way as to create a result acceptable to <strong>the</strong> pr<strong>in</strong>cipal<br />
stakeholders. This is a highly complex process, <strong>in</strong>volv<strong>in</strong>g<br />
multiple uncerta<strong>in</strong>ties (on <strong>in</strong>stitutional constra<strong>in</strong>ts, see<br />
WWF, 2000).<br />
<strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>in</strong>terdependence of multiple functions,<br />
actors and <strong>in</strong>stitutions is, however, not <strong>the</strong> only major<br />
socio-political challenge of fl oodpla<strong>in</strong> restoration. Considerable<br />
uncerta<strong>in</strong>ty is generated by <strong>the</strong> diverse spatial scales<br />
and time frames <strong>in</strong>volved. As described above, restor<strong>in</strong>g a<br />
fl oodpla<strong>in</strong> requires consideration of <strong>the</strong> longitud<strong>in</strong>al connectivity<br />
of a fl oodpla<strong>in</strong> to river uses both up- and downstream<br />
as well as of <strong>the</strong> lateral connectivity to ways <strong>in</strong><br />
which adjacent land is used (Adams and Perrow, 1999).<br />
Even on site, <strong>in</strong>terventions generally cut across several<br />
functional and adm<strong>in</strong>istrative boundaries. These can relate<br />
to <strong>the</strong> spatial remit of local landowners and farmers, plann<strong>in</strong>g<br />
authorities, government agencies, protected areas or<br />
<strong>in</strong>frastructure networks (e.g. rail, roads, canals). Temporally,<br />
functional fl oodpla<strong>in</strong>s are characterised by <strong>the</strong>ir<br />
dependence on fl ood<strong>in</strong>g events which are, by <strong>the</strong>ir nature,<br />
periodic and unpredictable, and by signifi cant time lags<br />
between changes <strong>in</strong> biotic and abiotic systems (see above;<br />
Adams and Perrow, 1999). Socio-economically, too, <strong>the</strong><br />
process of restor<strong>in</strong>g a fl oodpla<strong>in</strong> is framed by diverse time<br />
scales, rang<strong>in</strong>g from <strong>the</strong> payback periods for <strong>in</strong>vestments<br />
<strong>in</strong> altered practices of agriculture and forestry to <strong>the</strong> electoral<br />
periods of key public authorities. Floodpla<strong>in</strong> restoration<br />
is, <strong>the</strong>refore, not only highly complex but also highly<br />
unpredictable. The <strong>in</strong>stitutional challenge is fur<strong>the</strong>r complicated<br />
by <strong>the</strong> fact that many <strong>in</strong>stitutional arrangements<br />
– <strong>in</strong> particular for nature conservation – exhibit a strong<br />
tendency to protect exist<strong>in</strong>g conditions ra<strong>the</strong>r than encourage<br />
change, and those which do pursue change are generally<br />
geared towards achiev<strong>in</strong>g specifi c targets ra<strong>the</strong>r than<br />
creat<strong>in</strong>g suitable frameworks for open-ended processes, as<br />
is required for functional fl oodpla<strong>in</strong>s.<br />
6.6.1 Effective Institutions: The Search for Optimal<br />
Fit, Interplay and Scale<br />
Our knowledge of <strong>in</strong>stitutions which can support – or<br />
obstruct – <strong>the</strong> protection of public goods such as water,<br />
fl ood defence and biodiversity has been develop<strong>in</strong>g rapidly<br />
over <strong>the</strong> past decade (cf. Breit et al., 2003). However, rela-<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Riparian and Floodpla<strong>in</strong> <strong>Restoration</strong> 97<br />
tively little is known about <strong>the</strong> <strong>in</strong>stitutional dimensions of<br />
fl oodpla<strong>in</strong> restoration itself. Exceptions <strong>in</strong>clude studies<br />
of <strong>in</strong>stitutional constra<strong>in</strong>ts and complexities (Adams and<br />
Perrow, 1999), compet<strong>in</strong>g discourses of fl oodpla<strong>in</strong> restoration<br />
(Adams et al., 2004), relevant European Union policies<br />
(WWF, 2000, 2004) and European case studies (e.g. Zöckler,<br />
2000) and economic valuations of <strong>the</strong> functions and services<br />
provided by fl oodpla<strong>in</strong>s or wetlands (Gren et al.,<br />
1995; Turner et al., 2000; Adger and Luttrell, 2000).<br />
The Science Plan of <strong>the</strong> Institutional Dimensions of<br />
Global Environmental Change Project (IDGEC) of <strong>the</strong><br />
International Human Dimensions Programme (IHDP)<br />
offers useful analytical frameworks for conceptualis<strong>in</strong>g<br />
some of <strong>the</strong> essential <strong>in</strong>stitutional challenges of resource<br />
management <strong>in</strong> general and fl oodpla<strong>in</strong> restoration <strong>in</strong> particular<br />
(Young, 1999, 2002). It identifi es three generic<br />
factors <strong>in</strong>fl uenc<strong>in</strong>g <strong>the</strong> effectiveness of environmental<br />
<strong>in</strong>stitutions: problems of fi t, problems of <strong>in</strong>terplay and<br />
problems of scale.<br />
The issue of fi t addresses <strong>the</strong> need to develop <strong>in</strong>stitutional<br />
arrangements that match <strong>the</strong> properties of <strong>the</strong> biogeophysical<br />
systems <strong>the</strong>y are designed to regulate. Fit can<br />
relate to a variety of ecosystem properties. The follow<strong>in</strong>g<br />
are identifi ed <strong>in</strong> <strong>the</strong> IDGEC Science Plan: closed vs open<br />
systems; heterogeneity/homogeneity; <strong>in</strong>terdependencies<br />
among subsystems; simplicity/complexity; productivity/<br />
metabolism; cyclicity/periodicity; resilience; equilibria;<br />
dynamics (Young, 1999, p. 47). Problems of spatial misfi t<br />
are a particularly common cause of <strong>in</strong>stitutional <strong>in</strong>effectiveness.<br />
The territories covered by <strong>in</strong>stitutions rarely<br />
match those of biogeophysical systems, result<strong>in</strong>g <strong>in</strong> an<br />
<strong>in</strong>ability of <strong>the</strong> <strong>in</strong>stitutions to <strong>in</strong>ternalise external effects<br />
(both positive and negative) effectively. The management<br />
of fl oodpla<strong>in</strong>s is fraught with boundary problems of this<br />
k<strong>in</strong>d. Floodpla<strong>in</strong> restoration not only works across a<br />
variety of physical spaces along <strong>the</strong> river and across <strong>the</strong><br />
catchment, but also <strong>in</strong>volves <strong>in</strong>stitutions and organisations<br />
from multiple policy fi elds – from nature conservation and<br />
fl ood defence to agriculture – each with <strong>the</strong>ir own spatial<br />
remits.<br />
Interplay relates, by contrast, to <strong>in</strong>terdependencies<br />
between different <strong>in</strong>stitutions. The assumption here is that<br />
<strong>the</strong> effectiveness of an <strong>in</strong>stitution depends not only on its<br />
<strong>in</strong>herent qualities but also on how well it builds on, and is<br />
connected to, <strong>the</strong> broader <strong>in</strong>stitutional context. Institutional<br />
<strong>in</strong>terplay can be horizontal, between different policy<br />
fi elds, and vertical, between different levels of social<br />
organisation. A fur<strong>the</strong>r dist<strong>in</strong>ction is made between functional<br />
l<strong>in</strong>kages emanat<strong>in</strong>g from <strong>the</strong> properties of <strong>the</strong> <strong>in</strong>stitutions<br />
<strong>in</strong>volved and political l<strong>in</strong>kages as expressions of<br />
deliberation (Young, 2002). Problems of <strong>in</strong>terplay are very<br />
familiar to efforts to restore fl oodpla<strong>in</strong>s, which are often
98 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
confounded by <strong>the</strong> <strong>in</strong>ability to bridge differences <strong>in</strong> <strong>the</strong><br />
objectives, power structures and modes of action of <strong>the</strong><br />
various key organisations. Horizontal <strong>in</strong>terplay is complicated<br />
by <strong>the</strong> number of policy fi elds affected and vertical<br />
<strong>in</strong>terplay by <strong>the</strong> <strong>in</strong>creas<strong>in</strong>g role of <strong>the</strong> European Union and<br />
catchment-scale approaches to fl oodpla<strong>in</strong> management.<br />
Problems of scale can be of a spatial and a temporal<br />
nature. Spatially <strong>the</strong> effectiveness of an <strong>in</strong>stitution depends<br />
on fi nd<strong>in</strong>g <strong>the</strong> appropriate level of social organisation for<br />
specifi c <strong>in</strong>struments and measures, tak<strong>in</strong>g consideration of<br />
<strong>the</strong> connectivity between scales and <strong>the</strong> needs this creates<br />
for multi-level and multi-directional forms of governance.<br />
Temporally <strong>the</strong> issue of scale is about manag<strong>in</strong>g <strong>the</strong><br />
diverse time frames with<strong>in</strong> which actors operate, <strong>in</strong>stitutions<br />
work, projects are implemented, ideas are generated<br />
etc. Here, too, <strong>the</strong> relevance to fl oodpla<strong>in</strong> restoration is<br />
self-evident. Identify<strong>in</strong>g <strong>the</strong> appropriate spatial scale for<br />
policy development, strategic guidance, operational management,<br />
public participation and so on is of paramount<br />
importance. Similarly, actors operate accord<strong>in</strong>g to very<br />
different time scales, some of which are rigid and predictable,<br />
o<strong>the</strong>rs much less so.<br />
Research on problems of fi t, <strong>in</strong>terplay and scale suggests<br />
that solutions are rarely straightforward. For <strong>in</strong>stance,<br />
efforts to overcome problems of spatial fi t by <strong>in</strong>stitutionalis<strong>in</strong>g<br />
river bas<strong>in</strong> management can create new misfi ts and<br />
disturb exist<strong>in</strong>g modes of <strong>in</strong>terplay (Moss, 2003). Success<br />
would appear to be dependent less on attempt<strong>in</strong>g to reduce<br />
<strong>the</strong> given complexities and more on fi nd<strong>in</strong>g ways of<br />
accommodat<strong>in</strong>g complexity. It is argued here that a similar<br />
approach is required when deal<strong>in</strong>g with uncerta<strong>in</strong>ty. Given<br />
that substantial uncerta<strong>in</strong>ties will cont<strong>in</strong>ue to surround<br />
fl oodpla<strong>in</strong> restoration despite advances <strong>in</strong> our knowledge<br />
of <strong>the</strong> physical, biological and socio-political systems, it<br />
makes sense to consider possible cop<strong>in</strong>g strategies. The<br />
rema<strong>in</strong>der of this section addresses ways of cop<strong>in</strong>g with<br />
<strong>in</strong>stitutional uncerta<strong>in</strong>ty <strong>in</strong> fl oodpla<strong>in</strong> restoration with an<br />
empirically based study of three different approaches.<br />
6.6.2 Cop<strong>in</strong>g with <strong>Uncerta<strong>in</strong>ty</strong> I:<br />
Keep<strong>in</strong>g <strong>Restoration</strong> Simple<br />
The fi rst option for cop<strong>in</strong>g with uncerta<strong>in</strong>ty is to limit <strong>the</strong><br />
scope and scale of restoration. This was particularly<br />
common of <strong>the</strong> earlier schemes to restore fl oodpla<strong>in</strong>s <strong>in</strong><br />
Europe. Up until <strong>the</strong> late 1990s most fl oodpla<strong>in</strong> restoration<br />
schemes were small-scale and site-based. They were<br />
typically s<strong>in</strong>gle-issue projects, target<strong>in</strong>g environmental<br />
improvements as a rule. They <strong>in</strong>volved only a small<br />
number of actors and policy <strong>in</strong>struments, often rely<strong>in</strong>g on<br />
a s<strong>in</strong>gle source of fund<strong>in</strong>g for <strong>the</strong> physical <strong>in</strong>terventions.<br />
It is generally true to say that <strong>the</strong>se early generation res-<br />
toration schemes were conducted largely <strong>in</strong> isolation from<br />
national or regional policy <strong>in</strong>itiatives, whe<strong>the</strong>r for fl ood<br />
protection, biodiversity enhancement or rural development.<br />
Examples <strong>in</strong>clude <strong>the</strong> Rhe<strong>in</strong>vorland-Süd project <strong>in</strong><br />
Germany, a scheme to improve hydrological and ecological<br />
conditions by widen<strong>in</strong>g ducts and remov<strong>in</strong>g structures<br />
<strong>in</strong> a section of <strong>the</strong> Rh<strong>in</strong>e fl oodpla<strong>in</strong> near Rastatt, <strong>the</strong> Long<br />
Eau project <strong>in</strong> England, <strong>in</strong> which fl ood banks were set back<br />
for primarily environmental benefi ts, and <strong>the</strong> Bourret<br />
project <strong>in</strong> France, a scheme to reconnect an old arm to <strong>the</strong><br />
<strong>River</strong> Garonne and restore an alluvial forest – aga<strong>in</strong> primarily<br />
for environmental benefi ts.<br />
Be<strong>in</strong>g relatively unambitious and straightforward,<br />
schemes of this k<strong>in</strong>d tend to avoid <strong>the</strong> most press<strong>in</strong>g problems<br />
associated with high levels of uncerta<strong>in</strong>ty and complexity.<br />
They have succeeded <strong>in</strong> restor<strong>in</strong>g fl oodpla<strong>in</strong><br />
habitats with limited resources and, <strong>in</strong> some cases, with<strong>in</strong><br />
a short period. With <strong>the</strong> benefi t of relatively straightforward<br />
adm<strong>in</strong>istrative procedures, organisational structures<br />
and fund<strong>in</strong>g mechanisms it has been demonstrated how<br />
fl oodpla<strong>in</strong> ecosystems can be restored on a small scale.<br />
These schemes do, however, have several critical limitations.<br />
They rarely <strong>in</strong>corporate a catchment perspective on<br />
restoration, but concentrate on <strong>the</strong> site itself. Be<strong>in</strong>g predom<strong>in</strong>antly<br />
s<strong>in</strong>gle-issue schemes, <strong>the</strong>y regularly overlook<br />
potential benefi ts for o<strong>the</strong>r policy areas, such as fl ood<br />
protection, recreation or rural development. Little attention<br />
is generally paid to cultivat<strong>in</strong>g support for <strong>the</strong> project<br />
<strong>in</strong> <strong>the</strong> wider policy mak<strong>in</strong>g doma<strong>in</strong>, scientifi c communities<br />
or even <strong>in</strong> <strong>the</strong> local community. The performance of<br />
many such projects is rarely monitored or evaluated<br />
systematically.<br />
In terms of resolv<strong>in</strong>g problems of fi t, <strong>in</strong>terplay and scale<br />
it is possible to observe how schemes of this k<strong>in</strong>d are illequipped<br />
to meet <strong>the</strong> pr<strong>in</strong>cipal <strong>in</strong>stitutional challenges to<br />
fl oodpla<strong>in</strong> restoration set out above. Spatially, <strong>the</strong> small<br />
scale and site focus of <strong>the</strong> projects offer little opportunity<br />
to consider <strong>the</strong> catchment dimensions of fl ow regimes and<br />
biophysical connectivity beyond <strong>the</strong> immediate reach.<br />
Institutional <strong>in</strong>terplay may be less diffi cult but only<br />
because <strong>the</strong> number of policy fi elds and organisations<br />
<strong>in</strong>volved is kept small. Integrat<strong>in</strong>g <strong>the</strong> schemes <strong>in</strong>to <strong>the</strong><br />
development of <strong>the</strong> locality or region may well prove diffi<br />
cult at a later date for this reason. Problems of scale are<br />
reduced to site-based perspectives, miss<strong>in</strong>g valuable<br />
opportunities to explore multi-scalar solutions.<br />
6.6.3 Cop<strong>in</strong>g with <strong>Uncerta<strong>in</strong>ty</strong> II: Embrac<strong>in</strong>g <strong>the</strong><br />
Challenge of Open Outcomes<br />
Follow<strong>in</strong>g recent shifts <strong>in</strong> policies towards fl ood risk management,<br />
<strong>in</strong>tegrated water resources management and
ural development, more favourable framework conditions<br />
are creat<strong>in</strong>g w<strong>in</strong>dows of opportunity for ambitious forms<br />
of fl oodpla<strong>in</strong> restoration (see previously). In response to<br />
<strong>the</strong>se <strong>in</strong>stitutional drivers, and also to a grow<strong>in</strong>g recognition<br />
of <strong>the</strong> <strong>in</strong>adequacies of current ways of manag<strong>in</strong>g<br />
rivers, a new generation of fl oodpla<strong>in</strong> restoration schemes<br />
is emerg<strong>in</strong>g which are of a quite different scope to those<br />
of <strong>the</strong> early to mid-1990s. These schemes deliberately set<br />
out to address some of <strong>the</strong> complex challenges to largescale,<br />
<strong>in</strong>tegrated fl oodpla<strong>in</strong> restoration. Dist<strong>in</strong>ctive features<br />
of <strong>the</strong> new generation schemes are <strong>the</strong>ir multiple<br />
objectives (cover<strong>in</strong>g, for <strong>in</strong>stance, fl ood defence, biodiversity,<br />
rural development and water quality management),<br />
<strong>the</strong>ir wide actor engagement (<strong>in</strong>clud<strong>in</strong>g <strong>the</strong> relevant policy<br />
fi elds, local authorities, non-government organisations<br />
and <strong>the</strong> general public), <strong>the</strong>ir use of various <strong>in</strong>struments<br />
from different policy fi elds (e.g. jo<strong>in</strong>t fund<strong>in</strong>g from fl ood<br />
defence and agri-environment budgets) and <strong>the</strong>ir <strong>in</strong>teraction<br />
with policy mak<strong>in</strong>g processes, serv<strong>in</strong>g for <strong>in</strong>stance as<br />
pilot projects for national policy development. In addition,<br />
<strong>the</strong>y often have a long term vision for <strong>the</strong> measures envisaged<br />
and take a catchment – or at least large-scale – perspective<br />
on <strong>the</strong> fl oodpla<strong>in</strong>. <strong>Restoration</strong> sites are selected<br />
accord<strong>in</strong>g to <strong>the</strong>ir suitability for <strong>the</strong> catchment and not<br />
primarily because <strong>the</strong>y are available.<br />
Examples <strong>in</strong>clude: <strong>the</strong> Lenzen project <strong>in</strong> Germany, a<br />
major scheme to set back fl ood banks along a length of<br />
<strong>the</strong> Elbe river allow<strong>in</strong>g fl ood<strong>in</strong>g primarily for nature conservation,<br />
but also fl ood defence and regional development,<br />
benefi ts; <strong>the</strong> Parrett Catchment Project <strong>in</strong> England,<br />
an ongo<strong>in</strong>g project to promote more susta<strong>in</strong>able techniques<br />
of fl ood management <strong>in</strong> <strong>the</strong> whole catchment<br />
serv<strong>in</strong>g multiple purposes (fl ood protection, water level<br />
regulation, biodiversity targets, rural development); and<br />
<strong>the</strong> La Bassée project <strong>in</strong> France, a planned, large-scale<br />
fl ood retention scheme on <strong>the</strong> Se<strong>in</strong>e upstream of Paris with<br />
multiple benefi ts (fl ood protection, biodiversity, regional<br />
development). S<strong>in</strong>ce <strong>the</strong>se new generation schemes were<br />
only launched from <strong>the</strong> late 1990s onwards and are all at<br />
very early stages of implementation it is at present impossible<br />
to judge <strong>the</strong>ir effectiveness. They would at least<br />
appear to have <strong>the</strong> potential to overcome some of <strong>the</strong><br />
pr<strong>in</strong>cipal <strong>in</strong>stitutional constra<strong>in</strong>ts to fl oodpla<strong>in</strong> restoration<br />
which have thwarted or curtailed efforts <strong>in</strong> <strong>the</strong> past.<br />
Our research fi nd<strong>in</strong>gs, however, caution aga<strong>in</strong>st overoptimistic<br />
expectations from <strong>the</strong> new generation of projects.<br />
Early signs suggest that <strong>the</strong> sheer complexity of <strong>the</strong><br />
tasks <strong>the</strong>y are tackl<strong>in</strong>g and <strong>the</strong> uncerta<strong>in</strong>ties <strong>the</strong>y are<br />
expos<strong>in</strong>g are pos<strong>in</strong>g a major problem for project management.<br />
Build<strong>in</strong>g and ma<strong>in</strong>ta<strong>in</strong><strong>in</strong>g <strong>the</strong> large partnerships<br />
takes time and care. Strik<strong>in</strong>g an acceptable balance and<br />
negotiat<strong>in</strong>g trade-offs between diverse policy objectives<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Riparian and Floodpla<strong>in</strong> <strong>Restoration</strong> 99<br />
is very demand<strong>in</strong>g. Access<strong>in</strong>g multiple fund<strong>in</strong>g sources<br />
requires a high degree of fl exibility to satisfy different<br />
fund<strong>in</strong>g agencies. Attempts to enroll <strong>in</strong>struments from<br />
different policy fi elds can reveal serious <strong>in</strong>compatibilities<br />
and <strong>in</strong>consistencies. As a result, project design and<br />
implementation has become more complex, more timeconsum<strong>in</strong>g<br />
and more expensive, endanger<strong>in</strong>g effective<br />
implementation.<br />
Floodpla<strong>in</strong> restoration schemes of this k<strong>in</strong>d are mak<strong>in</strong>g<br />
substantial steps towards address<strong>in</strong>g problems of fi t, <strong>in</strong>terplay<br />
and scale. Their catchment orientation and long term<br />
perspective create a better fi t between ecosystem properties<br />
of <strong>the</strong> fl oodpla<strong>in</strong> and <strong>the</strong> <strong>in</strong>stitutional arrangements<br />
for its restoration, both <strong>in</strong> spatial and temporal terms.<br />
Build<strong>in</strong>g on better <strong>in</strong>terplay between <strong>in</strong>stitutions is central<br />
to <strong>the</strong> new generation projects, as is exploit<strong>in</strong>g different<br />
scales of action – from national pilots to local management<br />
teams – for different purposes. What <strong>the</strong> schemes<br />
are reveal<strong>in</strong>g, though, are serious secondary problems<br />
associated with this more ambitious and <strong>in</strong>tegrated<br />
approach to fl oodpla<strong>in</strong> restoration. Efforts to take on problems<br />
of fi t, <strong>in</strong>terplay and scale are prov<strong>in</strong>g, <strong>in</strong> many cases,<br />
too demand<strong>in</strong>g for project management. This does not<br />
query <strong>the</strong> desirability of address<strong>in</strong>g <strong>the</strong>se core <strong>in</strong>stitutional<br />
problems but, ra<strong>the</strong>r, raises questions about how project<br />
managers can be assisted <strong>in</strong> do<strong>in</strong>g so.<br />
6.6.4 Cop<strong>in</strong>g with <strong>Uncerta<strong>in</strong>ty</strong> III: Tighten<strong>in</strong>g<br />
Controls to Secure Better Policy Delivery<br />
A third way of cop<strong>in</strong>g with uncerta<strong>in</strong>ty can be identifi ed<br />
not at <strong>the</strong> level of <strong>in</strong>dividual schemes of fl oodpla<strong>in</strong> restoration<br />
but <strong>in</strong> <strong>the</strong> development and pursuit of policy. Here <strong>the</strong><br />
uncerta<strong>in</strong>ty addressed relates to <strong>the</strong> outcomes of new policies<br />
and <strong>the</strong> strategy is to attempt to m<strong>in</strong>imise uncerta<strong>in</strong>ties<br />
of policy delivery by tighten<strong>in</strong>g controls over those<br />
entrusted with implementation. As described earlier, <strong>in</strong><br />
several policy fi elds of direct relevance to fl oodpla<strong>in</strong><br />
restoration a more <strong>in</strong>tegrated and holistic approach to<br />
problem solv<strong>in</strong>g can be detected. This applies particularly<br />
to fl ood protection, follow<strong>in</strong>g recent fl ood<strong>in</strong>g events, water<br />
resources management, <strong>in</strong> response to <strong>the</strong> Water Framework<br />
Directive, and nature conservation. Changes <strong>in</strong><br />
policy content potentially conducive to fl oodpla<strong>in</strong> restoration<br />
can also be observed – if to a lesser degree – across<br />
Europe <strong>in</strong> spatial plann<strong>in</strong>g, agriculture, forestry and<br />
rural development. Many of <strong>the</strong>se supranational and<br />
national policy <strong>in</strong>itiatives are characterised by a more<br />
comprehensive problem analysis, longer term visions for<br />
improvements, better cross-sectoral policy <strong>in</strong>tegration,<br />
more strategic guidance and stronger and broader local<br />
partnerships. The guid<strong>in</strong>g pr<strong>in</strong>ciples underp<strong>in</strong>n<strong>in</strong>g
100 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Integrated Water Resources Management (IWRM) on a<br />
river bas<strong>in</strong> scale are a case <strong>in</strong> po<strong>in</strong>t (WWF, 2004, p. 24).<br />
Ironically, whilst policy content is, generally speak<strong>in</strong>g,<br />
tak<strong>in</strong>g a broader perspective, modes of policy implementation<br />
are to some extent becom<strong>in</strong>g more restrictive. In<br />
recent years a grow<strong>in</strong>g array of <strong>in</strong>struments have been<br />
<strong>in</strong>troduced which set out <strong>in</strong> detail not only what policy is<br />
to be pursued but also how this is to be done at <strong>the</strong> operational<br />
level. Targets are set to measure progress, strict<br />
consultation procedures must be followed to ga<strong>in</strong> plann<strong>in</strong>g<br />
approval, match-fund<strong>in</strong>g is required to demonstrate multifunctionality<br />
and audits are conducted to assess <strong>the</strong> performance<br />
of projects and programmes alike. These controls<br />
are widely justifi ed by governments <strong>in</strong> terms of <strong>the</strong>ir value<br />
<strong>in</strong> improv<strong>in</strong>g accountability, policy <strong>in</strong>tegration and, above<br />
all, cost effectiveness.<br />
Our research suggests, however, that measures of this<br />
k<strong>in</strong>d are hav<strong>in</strong>g important (un<strong>in</strong>tended) negative effects on<br />
<strong>the</strong> ability of project managers to implement fl oodpla<strong>in</strong><br />
restoration schemes. The fi rst problem relates to <strong>the</strong> cumulative<br />
effect of <strong>the</strong> new policy <strong>in</strong>itiatives. Each of <strong>the</strong> new<br />
requirements – whe<strong>the</strong>r on policy content or style – may<br />
<strong>in</strong>dividually make a lot of sense. Experiences of policy<br />
implementation show that <strong>the</strong> comb<strong>in</strong>ed effect of multiple<br />
new requirements can be to create a degree of management<br />
complexity that can severely delay <strong>the</strong> progress of<br />
some restoration projects and cause o<strong>the</strong>rs to be shelved.<br />
Ironically, <strong>the</strong>refore, effective policy delivery is be<strong>in</strong>g<br />
jeopardised by <strong>the</strong> sheer extent of policy reform.<br />
The second problem is more fundamental, hav<strong>in</strong>g to<br />
do with an emergent culture of control <strong>in</strong> policy mak<strong>in</strong>g<br />
circles. The measures to <strong>in</strong>crease accountability, policy<br />
<strong>in</strong>tegration and cost effectiveness refl ect not only very<br />
justifi able concerns about effective policy implementation<br />
and effi cient use of public funds but also <strong>the</strong> concerns of<br />
senior management <strong>in</strong> many government agencies that <strong>the</strong><br />
policy reth<strong>in</strong>k<strong>in</strong>g described above is not fi lter<strong>in</strong>g down<br />
effectively to <strong>the</strong> operational level. This argument is used<br />
by senior offi cers to justify tighter control – or ‘guidance’<br />
– to assure more effective implementation. Whilst <strong>the</strong> need<br />
for greater strategic guidance over such complex issues as<br />
a catchment-scale approach to fl ood protection is undisputed<br />
by all those <strong>in</strong>volved, one of <strong>the</strong> effects has been to<br />
restrict <strong>the</strong> freedom of action of project and programme<br />
managers. In <strong>the</strong> past <strong>the</strong>ir judgement – for example on<br />
whe<strong>the</strong>r to fund a restoration project – was based on <strong>the</strong>ir<br />
<strong>in</strong>dividual expertise, local knowledge and professional<br />
experience; today it is framed much more by targets<br />
devised at regional, national or even supranational levels.<br />
Consequently, <strong>the</strong> nature of <strong>the</strong>ir work is adapt<strong>in</strong>g <strong>in</strong> order<br />
to meet what Michael Power has termed <strong>the</strong> ‘rituals of<br />
verifi cation’ required by audit<strong>in</strong>g processes (Power, 1997).<br />
For <strong>the</strong> task of restor<strong>in</strong>g fl oodpla<strong>in</strong>s this poses a particular<br />
dilemma: whilst recent policy shifts and new generation<br />
schemes are encourag<strong>in</strong>g fl oodpla<strong>in</strong> restoration to enterta<strong>in</strong><br />
greater risks and uncerta<strong>in</strong>ties, adm<strong>in</strong>istrative procedures<br />
to assure policy implementation are becom<strong>in</strong>g<br />
<strong>in</strong>creas<strong>in</strong>gly risk-averse.<br />
In terms of fi t, <strong>in</strong>terplay and scale <strong>the</strong> picture here is<br />
more differentiated. Recent policy <strong>in</strong>itiatives relat<strong>in</strong>g to<br />
fl oodpla<strong>in</strong> restoration are certa<strong>in</strong>ly address<strong>in</strong>g very clearly<br />
problems of spatial and temporal fi t, tak<strong>in</strong>g a more catchment-oriented<br />
and long term perspective on <strong>the</strong> river and<br />
land management. Inter-sectoral <strong>in</strong>terplay is also strong.<br />
Vertical <strong>in</strong>terplay and issues of scale appear more problematic,<br />
however. The rhetoric of policy documents tend<br />
to be very supportive of multi-level and multi-direction<br />
governance. The reality – whe<strong>the</strong>r <strong>in</strong>tentional or not – is<br />
often very different, with control mechanisms of central<br />
government agencies reach<strong>in</strong>g down <strong>in</strong>to <strong>the</strong> operational<br />
level of project management to an unprecedented extent.<br />
This, it appears, is underm<strong>in</strong><strong>in</strong>g both project implementation<br />
and – ultimately – policy delivery.<br />
6.6.5 Mak<strong>in</strong>g Policy More Sensitive to <strong>the</strong> Challenges<br />
of Project Management<br />
It has been observed here how policy makers, <strong>in</strong> <strong>the</strong>ir<br />
efforts to encourage <strong>in</strong>tegrated, cross-sectoral and multiagency<br />
action, often overlook <strong>the</strong> implications for implementation<br />
at <strong>the</strong> operational level. In future, more<br />
consideration needs to be given to how <strong>in</strong>dividual policy<br />
<strong>in</strong>centives work <strong>in</strong> conjunction with o<strong>the</strong>rs; that is, how<br />
<strong>the</strong>y alter <strong>the</strong> exist<strong>in</strong>g <strong>in</strong>stitutional sett<strong>in</strong>g. In addition,<br />
policy makers need to be more sensitive to <strong>the</strong> contexts of<br />
action <strong>in</strong> which <strong>the</strong>ir <strong>in</strong>struments operate. What makes a<br />
policy <strong>in</strong>strument effective is not <strong>the</strong> assumed preferences<br />
of <strong>in</strong>dividuals act<strong>in</strong>g accord<strong>in</strong>g to a rational choice logic<br />
but <strong>the</strong> real scope and will<strong>in</strong>gness of stakeholders to alter<br />
<strong>the</strong>ir practices. More feedback <strong>in</strong>to policy mak<strong>in</strong>g processes<br />
is needed about <strong>the</strong> real-life experiences of project<br />
management and stakeholder <strong>in</strong>volvement at <strong>the</strong> operational<br />
level. This will require more monitor<strong>in</strong>g of how and<br />
why policy <strong>in</strong>struments do or do not work <strong>in</strong> practice. In<br />
addition, if uncerta<strong>in</strong>ty and complexity are unavoidable<br />
features of ambitious schemes of fl oodpla<strong>in</strong> restoration, as<br />
it appears, <strong>the</strong>n project managers and o<strong>the</strong>r stakeholders<br />
need assistance <strong>in</strong> assess<strong>in</strong>g <strong>the</strong> situation and <strong>the</strong>ir own<br />
ability to meet <strong>the</strong> challenge. Policy needs to provide not<br />
only targets for orientation but also frameworks for develop<strong>in</strong>g<br />
<strong>the</strong> necessary economic, social and <strong>in</strong>stitutional<br />
capital at local and regional level, and <strong>in</strong>struments capable<br />
of adapt<strong>in</strong>g to <strong>the</strong> dynamics of a fl oodpla<strong>in</strong> restoration<br />
process. On this basis policy makers and project managers
alike should be better equipped to identify and exploit<br />
w<strong>in</strong>dows of opportunity for <strong>the</strong> restoration of fl oodpla<strong>in</strong>s<br />
<strong>in</strong> <strong>the</strong> future.<br />
6.7 CONCLUSION<br />
A case has been made for <strong>the</strong> re-<strong>in</strong>troduction of diversity<br />
and variability <strong>in</strong> <strong>the</strong> riparian zone. It is advocated that<br />
this is done through <strong>the</strong> management of physical processes<br />
and with<strong>in</strong> a policy context that is sensitive to <strong>the</strong> complexities<br />
of successfully implement<strong>in</strong>g river restoration<br />
<strong>in</strong>itiatives. Inevitably <strong>the</strong> re-<strong>in</strong>troduction of diversity and<br />
variability will also <strong>in</strong>troduce a signifi cant level of uncerta<strong>in</strong>ty<br />
for river managers. The challenge is to decide what<br />
are acceptable levels of uncerta<strong>in</strong>ty for different stakeholders,<br />
for different scales of <strong>in</strong>volvement (such as catchment<br />
strategies or plann<strong>in</strong>g for a particular reach) and for<br />
different purposes (such as biodiversity targets or fl ood<br />
management). In <strong>the</strong> light of projected climate changes it<br />
is suggested that <strong>the</strong> key is to fi nd ways of adapt<strong>in</strong>g to<br />
uncerta<strong>in</strong>ty (see Chapter 14), ra<strong>the</strong>r than aim<strong>in</strong>g to reduce<br />
uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong> scientifi c sense of <strong>the</strong> word.<br />
6.8 ACKNOWLEDGEMENTS<br />
Much of <strong>the</strong> th<strong>in</strong>k<strong>in</strong>g that has gone <strong>in</strong>to this chapter has<br />
arisen from <strong>the</strong> work of <strong>the</strong> EC-funded project FLOBAR2<br />
(EVK1-CT-1999-00031). We thank all our colleagues on<br />
that project for stimulat<strong>in</strong>g and enjoyable discussions over<br />
many years of work<strong>in</strong>g toge<strong>the</strong>r. We would also like to<br />
thank Stewart Rood, John Mahoney, David Hulse and Joan<br />
Baker for permission to use <strong>the</strong>ir work <strong>in</strong> <strong>the</strong> two case<br />
study boxes; David Withr<strong>in</strong>gton of Natural England for<br />
read<strong>in</strong>g and comment<strong>in</strong>g on this manuscript; and Ian<br />
Agnew of <strong>the</strong> Department of Geography at <strong>the</strong> University<br />
of Cambridge for draw<strong>in</strong>g <strong>the</strong> fi gures.<br />
REFERENCES<br />
Adams WM, Perrow M. 1999. Scientifi c and <strong>in</strong>stitutional constra<strong>in</strong>ts<br />
on <strong>the</strong> restoration of European fl oodpla<strong>in</strong>s. In: Marriott<br />
S, Alexander J, Hey, R (Eds), Floodpla<strong>in</strong>s: Interdiscipl<strong>in</strong>ary<br />
Approaches. Geological Society of London Special Publ. 163,<br />
Geological Society: London; 89–97.<br />
Adams WM, Perrow MR, Carpenter A. 2004. Conservatives and<br />
champions: river managers and river restoration discourses <strong>in</strong><br />
<strong>the</strong> United K<strong>in</strong>gdom. Environment and Plann<strong>in</strong>g A 36: 1929–<br />
1942.<br />
Adger WN, Luttrell C. 2000. Property rights and <strong>the</strong> utilisation<br />
of wetlands. Ecological Economics 35: 75–89.<br />
Anderson DH, Dugger BD. 1998. A conceptual basis for evaluat<strong>in</strong>g<br />
restoration success. In: Wadsworth KG (Ed.), Chang<strong>in</strong>g<br />
Resource Values <strong>in</strong> Challeng<strong>in</strong>g Times. Transactions, 63rd<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Riparian and Floodpla<strong>in</strong> <strong>Restoration</strong> 101<br />
North American Wildlife and Natural Resources Conference,<br />
Wildlife Management Institute: Wash<strong>in</strong>gton, DC; 111–120.<br />
Arth<strong>in</strong>gton AH. 1998. Comparative Evaluation of Environmental<br />
Flow Assessment Techniques:Review of Holistic methodologies.<br />
LWRRDC Occasional Paper 26/98.<br />
Arth<strong>in</strong>gton AH, Brizga SO, Choy SC et al. 2000. Environmental<br />
Flow Requirements of <strong>the</strong> Brisbane <strong>River</strong> Downstream from<br />
Wivenhoe Dam. South East Queensland Water Corporation<br />
Ltd. and Centre for Catchment and In-stream Research, Griffi th<br />
University: Brisbane, Queensland, Australia.<br />
Baker JP, Hulse DW, Gregory SV et al. 2004. Alternative futures<br />
for <strong>the</strong> Willamette <strong>River</strong> Bas<strong>in</strong>, Oregon. Ecological Applications<br />
14 (2): 313–324.<br />
Breit H, Engels A, Moss T, Troja M. 2003. How Institutions<br />
Change. Perspectives on Social Learn<strong>in</strong>g <strong>in</strong> Global and Local<br />
Environmental Contexts. Leske and Budrich: Opladen,<br />
Germany.<br />
Bressers H, Kuks S. 2003. What does governance mean? From<br />
conception to elaboration. In: Bressers H, Rosenbaum W (Eds),<br />
Achiev<strong>in</strong>g Susta<strong>in</strong>able Development: <strong>the</strong> Challenge of Governance<br />
Across Spatial Scales. Praeger: New York.<br />
Brierley GJ, Fryirs K. 2000. <strong>River</strong> styles, a geomorphic approach<br />
to catchment characterization: implications for river rehabilitation<br />
<strong>in</strong> Bega Catchment, New South Wales, Australia. Environmental<br />
Management 25: 661–679.<br />
Brizga SO. 2000. Burnett <strong>River</strong> water allocation management<br />
plan:Proposed environmental fl ow performance measures.<br />
Department of Natural Resources: Brisbane, Australia.<br />
Brown CA, Joubert A. 2003. Us<strong>in</strong>g multicriteria analysis to<br />
develop environmental fl ow scenarios for rivers targeted for<br />
water resource management. Water South Africa 29 (4):<br />
365–378.<br />
Connell JH. 1978. Diversity <strong>in</strong> tropical ra<strong>in</strong> forests and coral<br />
reefs. Science 199: 1302–1310.<br />
Dewberry C, Burns P, Hood L. 1998. After <strong>the</strong> fl ood: <strong>the</strong> effects<br />
of <strong>the</strong> storms of 1996 on a creek restoration. <strong>Restoration</strong> and<br />
Management Notes 16: 174–182.<br />
Dodds WK, Gido K, Whiles MR et al. 2004. Life on <strong>the</strong> edge:<br />
The ecology of Great Pla<strong>in</strong>s prairie streams. BioScience 54:<br />
205–216.<br />
Dole D, Niemi E. 2004. Future water allocation and <strong>in</strong>-stream<br />
values <strong>in</strong> <strong>the</strong> Willamette <strong>River</strong> Bas<strong>in</strong>: a bas<strong>in</strong>-wide analysis.<br />
Ecological Applications 14: 355–367.<br />
Downs PW, Sk<strong>in</strong>ner KS, Kondolf GM. 2002. <strong>River</strong>s and streams.<br />
In: Perrow MR, Davy AJ (Eds), Handbook of Ecological <strong>Restoration</strong>,<br />
Volume 2 <strong>Restoration</strong> <strong>in</strong> Practice. Cambridge University<br />
Press: Cambridge; 267–296.<br />
EEC. 1992. Directive 92/43/EEC of <strong>the</strong> European Parliament and<br />
of <strong>the</strong> Council of 21 May 1992 on <strong>the</strong> conservation of natural<br />
habitats and of wild fauna and fl ora. OJ L 206, 22 July 1992,<br />
The European Parliament: Brussels, Belgium.<br />
Environment Agency. 2002. A Framework for Catchment Abstraction<br />
Management Strategies. R&D Technical Report W6-066/<br />
TR, RAM version 3 Summary Report. Environment Agency:<br />
Bristol.<br />
European Union. 2000. Directive 2000/60/EC of <strong>the</strong> European<br />
Parliament and of <strong>the</strong> Council of 23 October 2000 Establish<strong>in</strong>g
102 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
a Framework for Community Action <strong>in</strong> <strong>the</strong> Field of Water<br />
Policy. OJ L 327, 23 October 2000, The European Parliament:<br />
Brussels, Belgium.<br />
Frissell CA, Nawa R. 1992. Incidences and causes of physical<br />
failure of artifi cial habit structures <strong>in</strong> streams of Western<br />
Oregon and Wash<strong>in</strong>gton. North American Journal of Fisheries<br />
Management 12: 187–197.<br />
Gibert J, Fournier F, Mathieu J. 1997. The groundwater/surface<br />
waters ecotone perspective: State of <strong>the</strong> art. In: Gibert J,<br />
Mathieu J, Fournier F (Eds), Groundwater/Surface Water Ecotones:<br />
Biological and Hydrological Interactions and Management<br />
Options. International Hydrology Series, Cambridge<br />
University Press: Cambridge; 3–8.<br />
Gillilan S, Boyd K, Hoitsma T, Kauffman M. 2005. Challenges<br />
<strong>in</strong> develop<strong>in</strong>g and implement<strong>in</strong>g ecological standards for geomorphic<br />
river restoration projects: A comment on Palmer et al.<br />
(2004). Journal of Applied Ecology 42 (2): 223–227.<br />
Girel J, Hughes FMR, Moss T et al. 2003. A case for fl oodpla<strong>in</strong><br />
forests? In: Hughes FMR. (Ed) The Flooded Forest: Guidance<br />
for Policy Makers and <strong>River</strong> Managers <strong>in</strong> Europe on <strong>the</strong> <strong>Restoration</strong><br />
of Floodpla<strong>in</strong> Forests. FLOBAR2, Department of<br />
Geography, University of Cambridge: Cambridge; 6–23.<br />
Godw<strong>in</strong> H. 1941. Studies of <strong>the</strong> post-glacial history of British<br />
Vegetation. II Fenland Pollen Diagrams. Philosophical Transactions<br />
of <strong>the</strong> Royal Society Series B 230: 239–284.<br />
Gopal B, Junk WJ. 2001. Assessment, determ<strong>in</strong>ants, function and<br />
conservation of biodiversity <strong>in</strong> wetlands: present status and<br />
future needs. In: Gopal B, Junk WJ, Davis JA (Eds), Biodiversity<br />
<strong>in</strong> Wetlands: Assessment, Conservation and Function,<br />
Volume 2. Backhuys Publishers: Leiden, The Ne<strong>the</strong>rlands;<br />
277–302.<br />
Gren I-M, Groth K-H, Sylvén M. 1995. Economic Values of<br />
Danube Floodpla<strong>in</strong>s. Journal of Environmental Management<br />
45: 333–345.<br />
Hancock PJ, Boulton AJ, Humphreys WF. 2005. Aquifers and<br />
hyporheic zones: Towards an ecological understand<strong>in</strong>g of<br />
groundwater. Hydrogeology Journal 13: 98–111.<br />
Haraszthy L. 2001. The fl oodpla<strong>in</strong> forests <strong>in</strong> Hungary. In: Klimo<br />
E, Hager H (Eds), The Floodpla<strong>in</strong> Forests <strong>in</strong> Europe: Current<br />
Situation and Perspectives. Kon<strong>in</strong>klijke Brill NV, Leiden and<br />
European Forest Institute Research Report No. 10: Leiden, The<br />
Ne<strong>the</strong>rlands; 17–24.<br />
Harris RR. 1999. Defi n<strong>in</strong>g reference conditions for restoration of<br />
riparian plant communities: examples from California, USA.<br />
Environmental Management 24: 55–63.<br />
Heywood VH (Ed). 1995. Global Biodiversity Assessment. Cambridge<br />
University Press: Cambridge.<br />
Hill M, Platts WS. 1991. Ecological and geomorphological concepts<br />
for <strong>in</strong>stream and out-of-channel fl ow requirements.<br />
<strong>River</strong>s 2: 319–343.<br />
Holl<strong>in</strong>g CS. 1973. Resilience and stability of ecological systems.<br />
Annual Review of Ecology and Systematics 5: 25–37.<br />
Hughes FMR. 1994. Environmental change, disturbance and<br />
regeneration <strong>in</strong> semi-arid fl oodpla<strong>in</strong> forests. In: Mill<strong>in</strong>gton AC,<br />
Pye K (Eds), Environmental Change <strong>in</strong> Drylands: Biogeographical<br />
and Geomorphological Perspectives. John Wiley &<br />
Sons Ltd: Chichester; 321–346.<br />
Hughes FMR, Muller E. 2003. How can fl oodpla<strong>in</strong> forests be<br />
restored. In: Hughes FMR (Ed), The Flooded Forest: Guidance<br />
for Policy Makers and <strong>River</strong> Managers <strong>in</strong> Europe on <strong>the</strong> <strong>Restoration</strong><br />
of Floodpla<strong>in</strong> Forests. FLOBAR2, Department of<br />
Geography, University of Cambridge: Cambridge; 49–70.<br />
Hughes FMR, Rood SB. 2001. Floodpla<strong>in</strong>s. In: Warren A, French<br />
JR (Eds), Habitat Conservation: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> Physical Environment<br />
John Wiley & Sons Ltd: Chichester; 105–121.<br />
Hughes FMR, Rood SB. 2003. The allocation of river fl ows for<br />
<strong>the</strong> restoration of woody riparian and fl oodpla<strong>in</strong> forest ecosystems:<br />
a review of approaches and <strong>the</strong>ir application <strong>in</strong> Europe.<br />
Environmental Management 32: 12–33.<br />
Hughes FMR, Colston A, Mountford JO. 2005. Restor<strong>in</strong>g riparian<br />
ecosystems: <strong>the</strong> challenge of accommodat<strong>in</strong>g variability and<br />
design<strong>in</strong>g restoration trajectories. Ecology and Society 10(1),<br />
article 12 Onl<strong>in</strong>e URL: http://www.ecologyandsociety.org/<br />
vol10/iss1/art12/<br />
Hulse DW, Gregory SV. 2004. Integrat<strong>in</strong>g resilience <strong>in</strong>to fl oodpla<strong>in</strong><br />
restoration. Urban Ecosystems 7: 295–314.<br />
Hulse DW, Gregory SV, Baker JP (Eds), 2002. Willamette <strong>River</strong><br />
Bas<strong>in</strong>: Trajectories of Environmental and Ecological Change.<br />
Oregon State University Press: Corvallis, Oregon.<br />
Jansson R, Backx H, Boulton AJ et al. 2005. Stat<strong>in</strong>g mechanisms<br />
and refi n<strong>in</strong>g criteria for ecologically successful river restoration:<br />
A comment on Palmer et al. (2005). Journal of Applied<br />
Ecology 42 (2): 218–222.<br />
Jensen MN. 2004. Climate warm<strong>in</strong>g shakes up species. BioScience<br />
54: 722–729.<br />
JNCC. 2004. Common Standard Monitor<strong>in</strong>g Guidance for <strong>River</strong>s.<br />
Jo<strong>in</strong>t Nature conservation Committee: Peterborough, UK.<br />
Johnson WC. 2000. Tree recruitment and survival <strong>in</strong> rivers: <strong>in</strong>fl uence<br />
of hydrological processes. Hydrological Processes 14:<br />
3051–3074.<br />
Junk WJ, Bayley PB, Sparks RE. 1989. The fl ood-pulse concept<br />
<strong>in</strong> river-fl oodpla<strong>in</strong> systems. Canadian Special Publication of<br />
Fisheries and Aquatic Sciences 106: 110–127.<br />
K<strong>in</strong>g J, Louwe D. 1998. Instream fl ow requirements for regulated<br />
rivers <strong>in</strong> South Africa us<strong>in</strong>g <strong>the</strong> build<strong>in</strong>g block methodology.<br />
Aquatic Ecosystem Health and Management 1: 109–124.<br />
K<strong>in</strong>g J, Brown C, Sabat H. 2003. A scenario-based holistic<br />
approach to environmental fl ow assessments for rivers. <strong>River</strong><br />
Research and Applications 19: 619–639.<br />
Lambs L. 2004. Interactions between groundwater and surface<br />
water at river banks and <strong>the</strong> confl uence of rivers. Journal of<br />
Hydrology 288: 312–326.<br />
MacNally R, Park<strong>in</strong>son A, Horrocks G, Young M. 2002. Current<br />
loads of coarse woody debris on sou<strong>the</strong>astern Australian fl oodpla<strong>in</strong>s:<br />
Evaluation and change and implications for restoration.<br />
Ecological <strong>Restoration</strong> 10 (4): 627–635.<br />
Mahoney JM, Rood SB. 1998. Streamfl ow requirements for cottonwood<br />
seedl<strong>in</strong>g recruitment-An <strong>in</strong>tegrative Model. Wetlands<br />
18: 634–645.<br />
Malanson CP. 1993. Riparian Landscapes Cambridge University<br />
Press: Cambridge.<br />
Marston RA, Girel J, Pautou G et al. 1995. Channel metamorphosis,<br />
fl oodpla<strong>in</strong> disturbance, and vegetation development –<br />
A<strong>in</strong> <strong>River</strong>. Geomorphology 13: 121–131.
May RM. 1976a. Models for two <strong>in</strong>teract<strong>in</strong>g populations. In: May<br />
RM (Ed.), Theoretical Ecology: Pr<strong>in</strong>ciples and Applications.<br />
Blackwell Scientifi c Publications: Oxford; 49–70.<br />
May RM. 1976b. Patterns <strong>in</strong> multi-species communities. In: May<br />
RM (Ed.), Theoretical Ecology: Pr<strong>in</strong>ciples and Applications.<br />
Blackwell Scientifi c Publications: Oxford; 142–162.<br />
Medley K, Hughes FMR. 1996. <strong>River</strong><strong>in</strong>e forests. In: McClanahan<br />
TR (Ed.), Ecosystems and <strong>the</strong>ir Conservation <strong>in</strong> East Africa.<br />
Oxford University Press: Oxford; 361–384.<br />
Michelot J-L. 1995. Gestion Patrimoniale des Milieux Naturels<br />
Fluviaux: Guide Technique. Reserves Naturelles de France.<br />
Middelkoop H, van Haselen COG (Eds), 1999. Twice a <strong>River</strong> –<br />
Rh<strong>in</strong>e and Meuse <strong>in</strong> <strong>the</strong> Ne<strong>the</strong>rlands. RIZA Report no. 99.003,<br />
Institute for Inland Water Management and Waste Water Treatment/RIZA:<br />
Arnhem, The Ne<strong>the</strong>rlands.<br />
Middleton B. 1999. Flood Puls<strong>in</strong>g and Disturbance Dynamics.<br />
John Wiley & Sons Inc.: New York.<br />
Montgomery DR. 1999. Process doma<strong>in</strong>s and <strong>the</strong> river cont<strong>in</strong>uum.<br />
Journal of <strong>the</strong> American Water Resources Association 35:<br />
397–410.<br />
Montgomery DR, Boulton SM. 2003. In: Wissmar RC, Bisson<br />
PA (Eds), Strategies for Restor<strong>in</strong>g <strong>River</strong> Ecosystems: Sources<br />
of Variability and <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Natural and Managed Systems.<br />
American Fisheries Society: Be<strong>the</strong>sda, Maryland; 39–80.<br />
Moss T. 2003. Solv<strong>in</strong>g problems of ‘fi t’ at <strong>the</strong> expense of problems<br />
of ‘<strong>in</strong>terplay’? The spatial reorganisation of water management<br />
follow<strong>in</strong>g <strong>the</strong> EU Water Framework Directive. In:<br />
Breit H, Engels A, Moss T, Troja M (Eds), How Institutions<br />
Change. Perspectives on Social Learn<strong>in</strong>g <strong>in</strong> Global and Local<br />
Environmental Contexts. Leske and Budrich: Opladen,<br />
Germany; 85–121.<br />
Naiman RJ, Bunn SE, Nilsson C et al. 2002. Legitimiz<strong>in</strong>g fl uvial<br />
ecosystems as users of water: An overview. Environmental<br />
Management 30: 455–467.<br />
Nilsson C, Grelsson G, Johansson M, Sperens U. 1991a. Small<br />
rivers behave like large rivers: effects of postglacial history on<br />
plant species richness along river banks. Journal of Biogeography<br />
18: 533–541.<br />
Nilsson C, Ekblad A, Gardfjell M, Carlberg B. 1991b. Long term<br />
effects of river regulation on river marg<strong>in</strong> vegetation. Journal<br />
of Applied Ecology 28: 963–987.<br />
Nilsson C, Lepori F, Malmqvist B et al. 2005. Forecast<strong>in</strong>g<br />
environmental responses to restoration of rivers used as log<br />
fl oatways: An <strong>in</strong>terdiscipl<strong>in</strong>ary challenge. Ecosystems 8:<br />
779–800.<br />
NRC (National Research Council). 1992. New Strategies for<br />
America’s Watersheds. National Academic Press: Wash<strong>in</strong>gton,<br />
DC.<br />
Olden JD, Poff NL. 2003. Redundancy and <strong>the</strong> choice of hydrologic<br />
<strong>in</strong>dices for characteriz<strong>in</strong>g streamfl ow regimes. <strong>River</strong><br />
Research and Applications 19: 101–121.<br />
Pahl-Wostl C. 2002. Participative and stakeholder-based policy<br />
design, analysis and evaluation processes. Integrated Assessment<br />
3: 3–14.<br />
Pahl-Wostl C. 2004. Information, public empowerment, and <strong>the</strong><br />
management of urban watersheds. Environmental Modell<strong>in</strong>g<br />
and Software 20: 457–467.<br />
<strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Riparian and Floodpla<strong>in</strong> <strong>Restoration</strong> 103<br />
Palmer MA, Bernhardt ES, Allen JD et al. 2005. Standards for<br />
ecologically successful river restoration. Journal of Applied<br />
Ecology 42 (2): 208–217.<br />
Petts GE. 1990. The role of ecotones <strong>in</strong> aquatic landscape management.<br />
In: Naiman RJ, Decamps H (Eds), The Ecology and<br />
Management of Aquatic-Terrestrial Ecotones. UNESCO/MAB<br />
Series 4: Par<strong>the</strong>non, New Jersey; 227–261.<br />
Petts GE, Gurnell AM, Gerrard AJ et al. 2000. Longitud<strong>in</strong>al<br />
variations <strong>in</strong> exposed river<strong>in</strong>e sediments; a context for <strong>the</strong><br />
ecology of <strong>the</strong> Fiume Tagliamento, Italy. Aquatic Conservation:<br />
Mar<strong>in</strong>e and Freshwater Ecosystems 10: 249–266.<br />
Planty-Tabacchi A-M, Tabacchi E, Naiman RJ et al. 1996. Invasibility<br />
of species-rich communities <strong>in</strong> riparian zones. Conservation<br />
Biology 10: 598–607.<br />
Poff NL. 1992. Why disturbance can be predictable: A perspective<br />
on <strong>the</strong> defi nition of disturbance <strong>in</strong> streams. Journal of <strong>the</strong> North<br />
American Benthological Society 11: 86–92<br />
Poff NL, Allan JD, Palmer MA et al. 2003. <strong>River</strong> fl ows and water<br />
wars: emerg<strong>in</strong>g science for environmental decision mak<strong>in</strong>g.<br />
Frontiers <strong>in</strong> Ecology and Environment 1 (6): 298–306.<br />
Postel S, Richter B. 2003. <strong>River</strong>s for Life: <strong>Manag<strong>in</strong>g</strong> Water for<br />
People and Nature. Island Press: Wash<strong>in</strong>gton, DC.<br />
Power M. 1997. The Audit Society. Rituals of Verifi cation. Oxford<br />
University Press: Oxford.<br />
Richards KS, Bras<strong>in</strong>gton J, Hughes FMR. 2002. Geomorphic<br />
dynamics of fl oodpla<strong>in</strong>s: Ecological implications and a potential<br />
modell<strong>in</strong>g strategy. Freshwater Biology 47: 1–22.<br />
Richter B, Postel S. 2004. Sav<strong>in</strong>g Earth’s rivers. Issues <strong>in</strong> Science<br />
and Technology Spr<strong>in</strong>g issue: 31–36.<br />
Richter BD, Ma<strong>the</strong>ws R, Harrison DL, Wig<strong>in</strong>gton R. 2003. Ecologically<br />
susta<strong>in</strong>able water management: manag<strong>in</strong>g river fl ows<br />
for ecological <strong>in</strong>tegrity. Ecological Applications 13 (1):<br />
206–224.<br />
Rodwell JS (Ed.). 1991a. British Plant Communities, Volume 1:<br />
Woodlands and Scrub. Cambridge University Press:<br />
Cambridge.<br />
Rodwell JS (Ed.). 2000. British Plant Communities, Volume 2:<br />
Mires and Heath. Cambridge University Press: Cambridge.<br />
Rodwell JS (Ed.). 1991b. British Plant Communities, Volume 3:<br />
Grasslands and Montane Communities. Cambridge University<br />
Press: Cambridge.<br />
Rodwell JS (Ed.). 1992. British Plant Communities, Volume 4:<br />
Aquatic Communities, Swamps and Tall-Herb Fens. Cambridge<br />
University Press: Cambridge.<br />
Rodwell JS (Ed.). 1995. British Plant Communities, Volume 5:<br />
Maritime Communities and Vegetation of Open Habitats. Cambridge<br />
University Press: Cambridge.<br />
Rood SB, Mahoney JM. 2000. Revised <strong>in</strong>stream fl ow regulation<br />
enables cottonwood recruitment along <strong>the</strong> St Mary <strong>River</strong>,<br />
Alberta, Canada. <strong>River</strong>s 7: 109–125.<br />
Rood SB, Mahoney JM, Reid D, Zilm L. 1995. Instream<br />
fl ows and <strong>the</strong> decl<strong>in</strong>e of riparian cottonwoods along <strong>the</strong><br />
St. Mary <strong>River</strong>, Alberta. Canadian Journal of Botany 73:<br />
1250–1260.<br />
Rood SB, Gourley C, Ammon EM et al. 2003. Flows for fl oodpla<strong>in</strong><br />
forests: successful riparian restoration along <strong>the</strong> lower<br />
Truckee <strong>River</strong>, Nevada, U.S.A. BioScience 53: 647–656.
104 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
RSPB, NRA and RSNC. 1994. The New <strong>River</strong>s and Wildlife<br />
Handbook. RSPB: Sandy, Bedfordshire.<br />
Salo J, Kalliola R, Häkk<strong>in</strong>en I et al. 1986. <strong>River</strong> dynamics and<br />
<strong>the</strong> diversity of Amazon lowland forests. Nature 322:<br />
254–258.<br />
Scheimer F, Baumgartner C, Tockner K. 1999. <strong>Restoration</strong> of<br />
fl oodpla<strong>in</strong> rivers: <strong>the</strong> Danube restoration project. Regulated<br />
<strong>River</strong>s, Research and Management 15: 231–244.<br />
Simons J, Boeters R. 1998. A systematic approach to ecologically<br />
sound river bank management. In: de Waal LC, Large ARG,<br />
Wade PM (Eds), Rehabilitation of <strong>River</strong>s: Pr<strong>in</strong>ciples and<br />
Implementation. John Wiley & Sons Ltd: Chichester; 57–85.<br />
Stream Corridor Work<strong>in</strong>g Group. 1998. Stream Corridor <strong>Restoration</strong>:<br />
Pr<strong>in</strong>ciples, Processes and Practices. Federal Stream<br />
Interagency Stream <strong>Restoration</strong> Work<strong>in</strong>g Group, USA.<br />
Tabacchi E, Planty-Tabacchi AM, Sal<strong>in</strong>as MJ, Decamps H. 1996.<br />
Landscape structure and diversity <strong>in</strong> riparian plant communities:<br />
A longitud<strong>in</strong>al comparative study. Regulated <strong>River</strong>s:<br />
Research and Management 12: 367–390.<br />
Tockner K, Malard F, Ward JV. 2000. An extension of <strong>the</strong> Flood<br />
Pulse concept. Hydrological Processes 14: 2861–2883.<br />
Turner RK, van den Bergh JCJM, Söderqvist T et al. 2000. Ecological-economic<br />
analysis of wetlands: scientifi c <strong>in</strong>tegration<br />
for management and policy. Ecological Economics 35: 7–23.<br />
UNEP – World Conservation Monitor<strong>in</strong>g Centre. 2000. European<br />
Forests and Protected Areas: Gap Analysis. UNEP – WCMC:<br />
Cambridge, UK.<br />
Vannote RL, M<strong>in</strong>shall GW, Cumm<strong>in</strong>s KW et al. 1980. The river<br />
cont<strong>in</strong>uum concept. Canadian Journal of Fisheries and Aquatic<br />
Sciences 37: 130–137.<br />
Ward JV, Stanford JA. 1995. Ecological connectivity <strong>in</strong><br />
alluvial river ecosystems and its disruption by fl ow regulation.<br />
Regulated <strong>River</strong>s: Research and Management 11: 105–<br />
119.<br />
Ward JV, Tockner K. 2001. Biodiversity: towards a unify<strong>in</strong>g<br />
<strong>the</strong>me for river ecology. Freshwater Biology 46: 807–819.<br />
Ward JV, Tockner K, Schmeier F. 1999. Biodiversity of fl oodpla<strong>in</strong><br />
ecosystems: ecotones and connectivity. Regulated <strong>River</strong>s:<br />
Research & Management 15: 125–139.<br />
Ward JV, Tockner K, Arscott DB, Claret C. 2002. <strong>River</strong><strong>in</strong>e landscape<br />
diversity. Freshwater Biology 47: 517–539.<br />
War<strong>in</strong>g RH. 1989. Ecosystems: fl uxes of matter and energy. In:<br />
Cherrett JM (Ed.), Ecological Concepts: The Contribution of<br />
Ecology to an Understand<strong>in</strong>g of <strong>the</strong> Natural World. British<br />
Ecological Society and Blackwell Scientifi c Publications:<br />
Oxford; 17–42.<br />
World Commission on Dams 2000. Dams and Development: A<br />
New Framework for Decision Mak<strong>in</strong>g. Earthscan Publications<br />
Ltd: London.<br />
WWF 2000. Wise Use of Floodpla<strong>in</strong>s. Policy and Economic<br />
Analysis of Floodpla<strong>in</strong> <strong>Restoration</strong> <strong>in</strong> Europe. Opportunities<br />
and Obstacles. Report, November 2000.<br />
WWF 2004. Liv<strong>in</strong>g with Floods: Achiev<strong>in</strong>g Ecologically Susta<strong>in</strong>able<br />
Flood Management <strong>in</strong> Europe. Report, June 2004.<br />
Young O. 1999. Institutional Dimensions of Global Environmental<br />
Change. Science Plan. IHDP Report No. 9, IHDP: Bonn,<br />
Germany.<br />
Young O. 2002. The Institutional Dimensions of Environmental<br />
Change. Fit, Interplay, and Scale. MIT Press: Cambridge,<br />
Massachusetts.<br />
Zöckler C. 2000. Wise Use of Floodpla<strong>in</strong>s – LIFE Environment<br />
Project: A Review of 12 WWF <strong>River</strong> <strong>Restoration</strong> Projects<br />
Across Europe. WWF European Freshwater Programme:<br />
Copenhagen, Denmark.
<strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat Edited by Stephen Darby and David Sear<br />
© 2008 John Wiley & Sons, Ltd<br />
7<br />
Hydrological and Hydraulic Aspects of<br />
<strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for<br />
Ecological Purposes<br />
7.1 INTRODUCTION<br />
N.J. Clifford 1 , M.C. Acreman 2 and D.J. Booker 2,3<br />
1 School of Geography, University of Nott<strong>in</strong>gham, UK<br />
2 Centre for Ecology and Hydrology, UK<br />
3 (now at) National Institute of Water and Atmospheric Research, New Zealand<br />
This chapter presents river restoration as a hybrid activity,<br />
<strong>in</strong>volv<strong>in</strong>g hydrological, geomorphological and ecological<br />
expertise. It <strong>in</strong>troduces a range of techniques and methods<br />
that are appropriate to each of <strong>the</strong>se areas, gives examples<br />
of <strong>the</strong>ir application and reviews some of <strong>the</strong> fundamental<br />
opportunities and limitations of current restoration practice.<br />
<strong>River</strong> restoration is not only a hybrid activity but also<br />
an emerg<strong>in</strong>g one (both scientifi cally and practically).<br />
‘Uncerta<strong>in</strong>ties’ are present at every stage of restoration<br />
<strong>in</strong>tervention. These span such basic issues as: <strong>the</strong> ability<br />
to support ideas of catchment hydrology or fl ow regime<br />
with<strong>in</strong> which to frame restoration designs; provid<strong>in</strong>g fi eld<br />
evidence for conceptual and numerical simulation models;<br />
captur<strong>in</strong>g <strong>the</strong> natural range of variability <strong>in</strong>herent <strong>in</strong><br />
complex and dynamic physical–ecological systems; and<br />
<strong>in</strong>corporat<strong>in</strong>g all of <strong>the</strong>se uncerta<strong>in</strong>ties <strong>the</strong>mselves <strong>in</strong>to<br />
restoration design and appraisal practice. The chapter concludes<br />
that fi ve basic sources of uncerta<strong>in</strong>ty (also see<br />
Chapters 1 to 3) underp<strong>in</strong> contemporary river restoration<br />
<strong>the</strong>ory and practice: data (type, quality and quantity);<br />
characterisation (of those physical and ecological phenomena<br />
<strong>in</strong>volved <strong>in</strong> <strong>the</strong> restoration attempt); coupl<strong>in</strong>g (of<br />
physical environmental and ecological dynamics); awareness<br />
(of opportunities and limitations) and fl exibility (<strong>in</strong><br />
<strong>the</strong> approach to design and evaluation). Evaluat<strong>in</strong>g each<br />
of <strong>the</strong>se sources at every stage of <strong>the</strong> restoration is,<br />
perhaps, <strong>the</strong> best way of manag<strong>in</strong>g such uncerta<strong>in</strong>ties and,<br />
too, of improv<strong>in</strong>g prospects for <strong>the</strong> <strong>in</strong>corporation of more<br />
sophisticated simulation approaches to river restoration<br />
design and appraisal <strong>in</strong> <strong>the</strong> future.<br />
7.2 THE RIVER AND CATCHMENT AS AN<br />
UNCERTAIN SYSTEM<br />
<strong>Restoration</strong> of rivers may be viewed as <strong>the</strong> jo<strong>in</strong>t product<br />
of <strong>the</strong> physical structure of <strong>the</strong> channel and its fl oodpla<strong>in</strong>,<br />
<strong>the</strong>ir hydrological <strong>in</strong>tegration and <strong>the</strong>ir ecological value<br />
and function. Channels and fl oodpla<strong>in</strong>s are characterised<br />
by <strong>the</strong> tim<strong>in</strong>g and quantity of fl ows and sediments which<br />
<strong>the</strong> channel conveys, and which <strong>the</strong> fl oodpla<strong>in</strong> stores. As<br />
a result, channel restoration <strong>in</strong> its widest sense encompasses<br />
structural modifi cations of channel form, <strong>the</strong> reestablishment<br />
of natural fl ow regime and <strong>the</strong> reconnection<br />
(or even recreation) of channel and fl oodpla<strong>in</strong> areas (see<br />
Chapter 6). Both <strong>the</strong> problems of, and solution to, hydrological<br />
and hydraulic uncerta<strong>in</strong>ties <strong>in</strong> river restoration thus<br />
arise from consider<strong>in</strong>g <strong>the</strong> channel as an embedded part<br />
of <strong>the</strong> wider fl uvial hydrosystem (Petts and Amoros,<br />
1996). Incorporat<strong>in</strong>g hydrological connections or disturbances<br />
requires an holistic approach to streamfl ow management<br />
(Hill et al., 1991) <strong>in</strong> recognition of <strong>the</strong> many<br />
scales of connection or ‘dimensions’ of <strong>the</strong> fl uvial system<br />
(Boon, 1998; and for general review see Clifford 2001).<br />
Approaches to cost effective and multifunction river<br />
rehabilitation works have <strong>in</strong>creas<strong>in</strong>gly emphasised <strong>the</strong><br />
need to <strong>in</strong>clude an ecological perspective. Attention<br />
has focused on restor<strong>in</strong>g susta<strong>in</strong>able hydrological and
106 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
hydraulic habitats us<strong>in</strong>g pr<strong>in</strong>ciples of fl uvial behaviour<br />
(Newbury and Gaboury, 1993) as shown <strong>in</strong> Figure 7.1.<br />
This fi gure illustrates <strong>the</strong> hierarchy of feedbacks (or <strong>in</strong>terconnections)<br />
between physical structure, hydrological<br />
function and ecological responses <strong>in</strong> an alluvial river–<br />
fl oodpla<strong>in</strong> system which are relevant when design<strong>in</strong>g and<br />
assess<strong>in</strong>g river restoration or rehabilitation schemes. Flow<br />
and sediment transport depend upon catchment <strong>in</strong>puts of<br />
material and energy. These <strong>in</strong> turn ma<strong>in</strong>ta<strong>in</strong> and confi gure<br />
<strong>the</strong> channel shape, determ<strong>in</strong>e <strong>the</strong> sediment and bedform<br />
environments and create diversity of fl ow patterns and<br />
structures with vary<strong>in</strong>g degrees of coherence and spatial<br />
coverage. Ecological function of <strong>the</strong> channel is primarily<br />
a response to local (imposed) conditions of velocity, depth<br />
and substrate (<strong>the</strong> hydraulic variables), whose spatial<br />
(cross-section, reach-scale) and temporal (event-specifi c<br />
Human Land<br />
Use and Flow<br />
Regulation<br />
• water<br />
• sediment<br />
• nutrients<br />
Watershed Inputs<br />
Fluvial Geomorphic Processes<br />
• sediment transport/deposition/scour<br />
• channel migration and bank erosion<br />
• floodpla<strong>in</strong> construction and <strong>in</strong>undation<br />
• surface and groundwater <strong>in</strong>teractions<br />
Geomorphic Attributes<br />
• channel morphology (size, slope, shape,<br />
bed and bank composition)<br />
• floodpla<strong>in</strong> morphology<br />
• water turbidity and temperature<br />
Habitat Structure, Complexity, and Connectivity<br />
• <strong>in</strong>stream aquatic habitat<br />
• shaded riparian aquatic habitat<br />
• riparian woodlands<br />
• seasonally <strong>in</strong>undated floodpla<strong>in</strong> wetlands<br />
Biotic Responses<br />
(Aquatic, Riparian, and Terrestrial Plants and Animals)<br />
• abundance and distribution of native and exotic species<br />
• community composition and structure<br />
• food web structure<br />
and seasonal) characteristics refl ect wider reach-scale,<br />
<strong>in</strong>ter-reach and catchment controls (which are primarily<br />
hydrologically determ<strong>in</strong>ed).<br />
While <strong>the</strong> nature of <strong>the</strong> l<strong>in</strong>kages <strong>in</strong> Figure 7.1 is well<br />
understood, giv<strong>in</strong>g precise values to quantities and tim<strong>in</strong>gs<br />
of material and energy transfers, and account<strong>in</strong>g for <strong>the</strong><br />
feedbacks between <strong>the</strong>m, gives rise to uncerta<strong>in</strong>ties at all<br />
scales. These uncerta<strong>in</strong>ties are compounded by <strong>the</strong> recognition<br />
that climate and land use, which determ<strong>in</strong>e catchment<br />
ra<strong>in</strong>fall–run-off response, are nonstationary (that is,<br />
<strong>the</strong>y are chang<strong>in</strong>g <strong>in</strong> <strong>the</strong>ir mean level and variance) through<br />
time. In <strong>the</strong> United K<strong>in</strong>gdom, for example, Prudhomme<br />
et al. (2003) exam<strong>in</strong>e <strong>the</strong> implications of no less than<br />
25 000 climate scenarios for four typical fl ood events as<br />
applied to fi ve catchments. Most scenarios show an<br />
<strong>in</strong>crease <strong>in</strong> both <strong>the</strong> magnitude and frequency of fl ood<br />
• energy<br />
• large woody debris<br />
• chemical pollutants<br />
Natural<br />
Disturbance<br />
Figure 7.1 The nested hierarchy of <strong>the</strong> channel system and its associated habitat potential and biotic response (Tuolumne <strong>River</strong><br />
restoration program summary report, summary of studies, conceptual models, restoration projects, and ongo<strong>in</strong>g monitor<strong>in</strong>g. Prepared<br />
for <strong>the</strong> CALFED/AFRP Adaptive Management Forum, with assistance from <strong>the</strong> Tuolun<strong>in</strong>e <strong>River</strong> Technical Advisory Committee,<br />
(2001). Reproduced with permission from Stillwater Sciences.)
Hydrological and Hydraulic Aspects of <strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 107<br />
events, but <strong>the</strong> largest uncerta<strong>in</strong>ty arises from <strong>the</strong> type<br />
of Global Climate Model used: <strong>the</strong> magnitude of modelled<br />
change varies by a factor of n<strong>in</strong>e <strong>in</strong> Nor<strong>the</strong>rn England<br />
and Scotland! At scales below <strong>the</strong> level of climatic<br />
<strong>in</strong>put, major uncerta<strong>in</strong>ties rema<strong>in</strong> <strong>in</strong> <strong>the</strong> modell<strong>in</strong>g and<br />
prediction of ra<strong>in</strong>fall–run-off relationships, upon which<br />
channel fl ow ultimately depends. Catchment run-off modell<strong>in</strong>g<br />
is <strong>in</strong>creas<strong>in</strong>gly able to <strong>in</strong>corporate greater physical<br />
parameterisation, and to distribute parameter comb<strong>in</strong>ations<br />
around <strong>the</strong> catchment and between surface and subsurface<br />
fl ows. Similar predictions may, however, be<br />
obta<strong>in</strong>ed us<strong>in</strong>g models that <strong>in</strong>corporate differ<strong>in</strong>g levels of<br />
sophistication and differ<strong>in</strong>g parameter values, giv<strong>in</strong>g rise<br />
to model ‘equifi nality’. This raises concerns about <strong>the</strong><br />
degree to which fundamental understand<strong>in</strong>g of process<br />
can, <strong>in</strong> fact, lead to better <strong>in</strong>formation on future (or<br />
designed) outcomes (Beven, 2001). Ano<strong>the</strong>r major source<br />
of concern is <strong>the</strong> degree to which model predictions may<br />
be appropriately up- or down-scaled (for a comprehensive<br />
review of <strong>the</strong>se issues, see Beven, 2000 and Bierkens<br />
et al., 2000).<br />
Gilvear et al. (2002) and Hendry et al. (2003) po<strong>in</strong>t out<br />
that <strong>the</strong> hydrological basis for determ<strong>in</strong><strong>in</strong>g future fi sheries<br />
stocks is complicated by issues of water and sediment<br />
quality, as well as quantity. Water quality necessitates<br />
consideration of longer- and shorter-term land use histories,<br />
run-off behaviour and <strong>the</strong> monitor<strong>in</strong>g or modell<strong>in</strong>g<br />
of diffuse as well as po<strong>in</strong>t-source pollutants. ‘Hydrology’<br />
itself, <strong>the</strong>refore, has many uncerta<strong>in</strong> components,<br />
and Clarke SL et al. (2003) call for <strong>the</strong> development<br />
of ‘hydrogeomorphological knowledge’ of catchments<br />
support<strong>in</strong>g ‘tools’ for water resource management. The<br />
need to service ever-more complex models and to comply<br />
with <strong>in</strong>creas<strong>in</strong>gly complex (but frequently compet<strong>in</strong>g) legislative<br />
requirements will also place grow<strong>in</strong>g pressures for<br />
<strong>the</strong> provision of data (both quantity and quality). This is<br />
likely to demand changes to long-stand<strong>in</strong>g methods of<br />
data acquisition and process<strong>in</strong>g (Marsh, 2002). All of<br />
<strong>the</strong>se essentially hydrological issues are encountered<br />
before <strong>the</strong> channel-scale is addressed from a hydraulic<br />
standpo<strong>in</strong>t!<br />
With<strong>in</strong> <strong>the</strong> channel at reach and subreach scales, it is<br />
hydraulics which underp<strong>in</strong> connections between <strong>the</strong> physical<br />
and ecological environments. Determ<strong>in</strong><strong>in</strong>g an appropriate<br />
channel morphology and <strong>the</strong> characteristics of fl ow<br />
behaviour <strong>in</strong> response to chang<strong>in</strong>g discharge and sediment<br />
transport are key factors <strong>in</strong> design<strong>in</strong>g restoration schemes.<br />
Such schemes must be both susta<strong>in</strong>able <strong>in</strong> a physical<br />
sense, and functional <strong>in</strong> an ecological sense. Both considerations<br />
require some allowance for natural dynamics and<br />
post-design evolution. Over <strong>the</strong> last two decades, a new<br />
subject of ‘eco-hydraulics’ has developed (Leclerc, 2002),<br />
<strong>in</strong> which a range of monitor<strong>in</strong>g and modell<strong>in</strong>g strategies<br />
l<strong>in</strong>k <strong>the</strong> expertise of eng<strong>in</strong>eers, geomorphologies and<br />
ecologists to design and assess restoration or rehabilitation<br />
schemes. Yet, from very basic stages of fl ow ‘characterisation’<br />
through to fl ow modell<strong>in</strong>g, habitat simulation and<br />
post-project appraisal, numerous uncerta<strong>in</strong>ties exist. This<br />
chapter reviews <strong>the</strong> sources and implications of <strong>the</strong>se<br />
uncerta<strong>in</strong>ties, and provides case study examples of current<br />
and prospective restoration practice motivated by ecohydraulic<br />
considerations.<br />
7.3 HYDROLOGICAL ASPECTS AND<br />
UNCERTAINTY IN CHANNEL RESTORATION<br />
7.3.1 Connectivity, Disturbance and <strong>the</strong> Ecological<br />
Flow Regime <strong>in</strong> Channels<br />
A river ecosystem and its associated benefi ts to humank<strong>in</strong>d<br />
are strongly conditioned by <strong>the</strong> pattern of fl ows<br />
between days, seasons and years (<strong>in</strong>clud<strong>in</strong>g fl oods and<br />
droughts) that occur with<strong>in</strong> <strong>the</strong> dra<strong>in</strong>age bas<strong>in</strong>. For<br />
example, fl oods ma<strong>in</strong>ta<strong>in</strong> river structure and sediment<br />
distribution (Hill et al., 1991), medium fl ows trigger<br />
fi sh migration (Junk et al., 1989) and low fl ows ma<strong>in</strong>ta<strong>in</strong><br />
species diversity (Everard, 1996). The pattern of fl ows<br />
required to support a river ecosystem is called <strong>the</strong> environmental<br />
fl ow requirements (Dyson et al., 2003). If <strong>the</strong><br />
river fl ow pattern is altered from its natural regime, <strong>the</strong>n<br />
<strong>the</strong> river ecosystem will change from its natural state. Too<br />
much fl ow at <strong>the</strong> wrong time can be as damag<strong>in</strong>g as too<br />
little fl ow. The fl ow regime of a river describes <strong>the</strong> temporal<br />
variability of run-off with<strong>in</strong> a s<strong>in</strong>gle hydrological<br />
year (<strong>in</strong>ter-annual, such as between w<strong>in</strong>ter fl oods, spr<strong>in</strong>g<br />
snowmelt run-off, summer basefl ow), and from year-toyear<br />
(<strong>in</strong>tra-annual, such as dry years, wet years or alternat<strong>in</strong>g<br />
periods or cycles of drought and higher fl ows).<br />
Recognition of <strong>the</strong> importance of dynamism and adjustment<br />
<strong>in</strong> fl ow regime for <strong>the</strong> physical and biotic system is<br />
someth<strong>in</strong>g of a recent paradigm shift <strong>in</strong> river management<br />
and restoration (Bergen et al., 2001; Newson, 2002), but<br />
poses areas of additional design and management uncerta<strong>in</strong>ty.<br />
Traditional eng<strong>in</strong>eer<strong>in</strong>g <strong>in</strong>tervention was guided by<br />
pr<strong>in</strong>ciples of control and stability, which led to ‘certa<strong>in</strong>’<br />
or fi xed design and performance criteria. Design<strong>in</strong>g-<strong>in</strong><br />
variability, or allow<strong>in</strong>g for a degree of post-modifi cation<br />
‘naturalisation’ to channel works is not only a novel<br />
concept but is also largely untried or tested. It requires <strong>the</strong><br />
sett<strong>in</strong>g of generous tolerances on design functional requirements,<br />
to refl ect <strong>the</strong> degree of ignorance and paucity of<br />
research (Bergen et al., 2001). In this respect, ‘uncerta<strong>in</strong>ty’<br />
is itself someth<strong>in</strong>g of an essential design criterion<br />
(also see Chapter 14)!<br />
The most extreme changes to a fl ow regime are often<br />
associated with construction of a dam, where <strong>the</strong> entire
108 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
pattern of fl ows may be replaced by a constant low fl ow.<br />
This can have serious negative impacts on <strong>the</strong> river ecosystem<br />
downstream (Petts, 1984). A key part of river restoration<br />
is often <strong>the</strong> reduction of abstractions (Barker and<br />
Kirmond, 1998) or release of water from dams (Acreman,<br />
2002) to restore river fl ows to a more natural pattern.<br />
While geomorphologists and hydrologists have tended to<br />
emphasise connectivity and transfer of water and materials<br />
through <strong>the</strong> fl uvial system, <strong>the</strong> ‘real’ state of catchments<br />
<strong>in</strong> <strong>the</strong> developed world is of highly fragmented, and<br />
largely modifi ed transfers. Graf (2001) estimates that only<br />
2% of <strong>the</strong> 3.5 million miles of streams <strong>in</strong> <strong>the</strong> United States<br />
are unaffected by dams, and that 18% of this is actually<br />
under <strong>the</strong> waters of reservoirs. The picture of fragmentation,<br />
threat and change is similar <strong>in</strong> <strong>the</strong> United K<strong>in</strong>gdom,<br />
where less than 10% of rivers are free from structural<br />
modifi cation of channel and banks and 53% of rivers have<br />
fl ow regimes altered by more than 20% (Acreman, 2000).<br />
Indeed, such is <strong>the</strong> scale of <strong>in</strong>terference with natural conditions<br />
of catchments <strong>in</strong> developed countries, that <strong>the</strong> most<br />
conv<strong>in</strong>c<strong>in</strong>g strategy for environmental enhancement <strong>in</strong> <strong>the</strong><br />
short to medium term may be to use <strong>the</strong> capacity for<br />
regulation to remediate fl ows. The alternative, to remove<br />
expensive, fi xed <strong>in</strong>frastructure at <strong>the</strong> larger scale (i.e.<br />
restoration of fl ows), may require unrealistic cultural,<br />
political and economic shifts (Graf, 2001).<br />
Remediation of fl ows might be accomplished by changes<br />
to <strong>the</strong> operat<strong>in</strong>g rules of dams, by redesign and by physical<br />
renovation of structures. A high profi le case of relax<strong>in</strong>g<br />
impoundment occurred on <strong>the</strong> Colorado <strong>River</strong> <strong>in</strong> <strong>the</strong> United<br />
States <strong>in</strong> 1998 (Schmidt et al., 1998) and illustrates<br />
Pr<strong>in</strong>ciple 1<br />
variation <strong>in</strong> flow regime<br />
Habitat complexity biotic diversity<br />
discharge<br />
Pr<strong>in</strong>ciple 4<br />
time<br />
natural regime discourages <strong>in</strong>vasions<br />
<strong>the</strong> importance of unit<strong>in</strong>g o<strong>the</strong>rwise confl ict<strong>in</strong>g policies<br />
to achieve environmental benefi ts. Acreman (2002)<br />
addresses <strong>the</strong> implications of modifi ed natural systems for<br />
<strong>the</strong> practice of scientifi c hydrology, argu<strong>in</strong>g that <strong>the</strong> realities<br />
of modifi ed systems must be <strong>in</strong>corporated <strong>in</strong>to traditional<br />
hydrological tra<strong>in</strong><strong>in</strong>g and practice (for review see Clifford,<br />
2002).<br />
The natural fl ow regime of a river is a function of <strong>the</strong><br />
magnitude, duration, frequency and tim<strong>in</strong>g of precipitation;<br />
<strong>the</strong> form of <strong>the</strong> precipitation (ra<strong>in</strong> or snow) and <strong>the</strong><br />
characteristics of <strong>the</strong> dra<strong>in</strong>age bas<strong>in</strong> (which determ<strong>in</strong>es<br />
how precipitation translates <strong>in</strong>to streamfl ow via surface<br />
and subsurface run-off). Each stream has a unique fl ow<br />
regime, characterised by <strong>the</strong> stream fl ow hydrograph. Four<br />
pr<strong>in</strong>ciples which might be used to guide restoration or<br />
enhancement of fl ow for ecologically motivated restoration<br />
schemes based upon common hydrograph characteristics<br />
are described <strong>in</strong> Figure 7.2 (Bunn and Arthr<strong>in</strong>gton,<br />
2002).<br />
From this it can be seen that almost all aspects of <strong>the</strong><br />
fl ow hydrograph have some importance ei<strong>the</strong>r to <strong>in</strong>dividual<br />
species at various stages of <strong>the</strong>ir life cycle, or to <strong>the</strong><br />
determ<strong>in</strong>ation of species assemblages and hence biotic<br />
diversity and abundance more generally. These hydrograph<br />
components may become more or less important <strong>in</strong><br />
<strong>the</strong> context of both <strong>in</strong>ter- and <strong>in</strong>tra-annual streamfl ow<br />
variation. Hypo<strong>the</strong>sised relationships between water<br />
years, hydrograph components and ecological processes<br />
developed for <strong>the</strong> restoration of gravel-bed stream<br />
environments downstream of <strong>the</strong> Sierra Nevada foothills,<br />
California, USA, are illustrated <strong>in</strong> Table 7.1.<br />
Pr<strong>in</strong>ciple 3<br />
access to floodpla<strong>in</strong>s (lateral<br />
connectivity; longitud<strong>in</strong>al<br />
connectivity) essential<br />
Pr<strong>in</strong>ciple 2<br />
regular variation and stable<br />
baseflows determ<strong>in</strong>e life<br />
history patterns<br />
(spawn<strong>in</strong>g/recruitment)<br />
Figure 7.2 Basic pr<strong>in</strong>ciples govern<strong>in</strong>g ecological response to <strong>the</strong> stream hydrograph (Source: modifi ed after Bunn and Arthr<strong>in</strong>gton,<br />
2002)
Hydrological and Hydraulic Aspects of <strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 109<br />
Table 7.1 General hypo<strong>the</strong>sised relationships between hydrograph components and ecosystem processes for gravel-bed streams<br />
downstream of <strong>the</strong> Sierra Nevada foothills (Modifi ed from San Joaqu<strong>in</strong> <strong>River</strong> <strong>Restoration</strong> Study Background Report, prepared for<br />
Friant Water Users Authority, L<strong>in</strong>dsay, CA, and Natural Resources Defense Council, San Francisco, CA, (2002). Reproduced with<br />
permission from McBa<strong>in</strong> & Trush, Inc.)<br />
Hydrograph<br />
component<br />
Geomorphic-hydrologic<br />
processes<br />
Snowmelt peak Wetter years: bed mobility,<br />
long duration fl oodpla<strong>in</strong><br />
<strong>in</strong>undation, moderate<br />
channel migration,<br />
groundwater recharge<br />
Normal years: bed<br />
mobility, short duration<br />
fl oodpla<strong>in</strong> <strong>in</strong>undation<br />
Seasonal recession<br />
limb<br />
Summer–fall<br />
basefl ow<br />
W<strong>in</strong>ter–spr<strong>in</strong>g<br />
ra<strong>in</strong>fall run-off<br />
Gradual decrease <strong>in</strong> water<br />
stage, ma<strong>in</strong>ta<strong>in</strong><br />
fl oodpla<strong>in</strong> soil moisture<br />
Wetter years: channel<br />
avulsion, signifi cant<br />
channel migration, bed<br />
scour and deposition,<br />
bed mobility, fl oodpla<strong>in</strong><br />
scour, fl oodpla<strong>in</strong><br />
<strong>in</strong>undation, fi ne<br />
sediment deposition on<br />
fl oodpla<strong>in</strong>s, large<br />
woody debris<br />
recruitment<br />
Normal years: Some<br />
channel migration,<br />
m<strong>in</strong>or bed scour, bed<br />
mobility, fl oodpla<strong>in</strong><br />
<strong>in</strong>undation, some fi ne<br />
sediment deposition on<br />
fl oodpla<strong>in</strong>s<br />
Riparian processes Salmonid life-history processes<br />
Wetter years: riparian<br />
seedl<strong>in</strong>g scour with<strong>in</strong><br />
bankfull channel, riparian<br />
seedl<strong>in</strong>g <strong>in</strong>itiation on<br />
fl oodpla<strong>in</strong>s, discourages<br />
riparian seedl<strong>in</strong>g <strong>in</strong>itiation<br />
with<strong>in</strong> bankfull channel<br />
Normal years: periodic<br />
riparian seedl<strong>in</strong>g <strong>in</strong>itiation<br />
on fl oodpla<strong>in</strong>s<br />
Wetter years: Allow riparian<br />
seedl<strong>in</strong>g establishment on<br />
fl oodpla<strong>in</strong>s<br />
Normal and drier years:<br />
Discourages riparian<br />
seedl<strong>in</strong>g establishment on<br />
fl oodpla<strong>in</strong>s by desiccat<strong>in</strong>g<br />
<strong>the</strong>m, encourage seedl<strong>in</strong>g<br />
establishment with<strong>in</strong><br />
bankfull channel<br />
Encourages late seed<strong>in</strong>g<br />
riparian vegetation<br />
<strong>in</strong>itiation and<br />
establishment with<strong>in</strong><br />
bankfull channel<br />
Wetter years: mature<br />
riparian removal with<strong>in</strong><br />
bankfull channel and<br />
portions of fl oodpla<strong>in</strong>,<br />
scour of seedl<strong>in</strong>gs with<strong>in</strong><br />
bankfull channel, seedbed<br />
creation on fl oodpla<strong>in</strong>s<br />
for new cohort <strong>in</strong>itiation,<br />
microtopography from<br />
fl oodpla<strong>in</strong> scour and fi ne<br />
sediment deposition<br />
Normal years: scour of<br />
seedl<strong>in</strong>gs with<strong>in</strong> bankfull<br />
channel, some fi ne<br />
sediment deposition on<br />
fl oodpla<strong>in</strong>s<br />
Wetter years: Increase juvenile growth rates by<br />
long-term fl oodpla<strong>in</strong> <strong>in</strong>undation, <strong>in</strong>crease<br />
strand<strong>in</strong>g by <strong>in</strong>undat<strong>in</strong>g fl oodpla<strong>in</strong>s,<br />
stimulate outmigration, reduce predation<br />
mortality by reduc<strong>in</strong>g smolt density and<br />
<strong>in</strong>creas<strong>in</strong>g turbidity<br />
Normal years: Increase juvenile growth rates<br />
by short-term fl oodpla<strong>in</strong> <strong>in</strong>undation,<br />
<strong>in</strong>crease strand<strong>in</strong>g by short-term fl oodpla<strong>in</strong><br />
<strong>in</strong>undation, stimulate outmigration, reduce<br />
predation mortality by reduc<strong>in</strong>g smolt<br />
density and <strong>in</strong>creas<strong>in</strong>g turbidity<br />
Drier years: Increase outmigration predation<br />
mortality by <strong>in</strong>creas<strong>in</strong>g density and reduc<strong>in</strong>g<br />
turbidity<br />
Wetter years: Increase outmigration success by<br />
reduc<strong>in</strong>g water temperatures and extend<strong>in</strong>g<br />
outmigration period<br />
Normal years: Increase outmigration success<br />
by reduc<strong>in</strong>g water temperatures and<br />
extend<strong>in</strong>g outmigration period<br />
Drier years: Increase outmigration mortality by<br />
<strong>in</strong>creas<strong>in</strong>g water temperatures and<br />
shorten<strong>in</strong>g outmigration period<br />
Water temperature for over-summer<strong>in</strong>g<br />
juveniles and spr<strong>in</strong>g-run adults, immigration<br />
for fall-run adults<br />
Wetter years: partial loss of cohort due to redd<br />
scour or entombment from deposition,<br />
improve spawn<strong>in</strong>g gravel quality by<br />
scour<strong>in</strong>g/redeposit<strong>in</strong>g bed and transport<strong>in</strong>g<br />
fi ne sediment, mortality by fl ush<strong>in</strong>g fry and<br />
juveniles, mortality by strand<strong>in</strong>g fry and<br />
juveniles on fl oodpla<strong>in</strong>s, reduce growth<br />
dur<strong>in</strong>g periods of high turbidity, reduce<br />
predation dur<strong>in</strong>g periods of high turbidity,<br />
creation and ma<strong>in</strong>tenance of high quality<br />
aquatic habitat<br />
Normal years: improve spawn<strong>in</strong>g gravel<br />
quality by mobilis<strong>in</strong>g bed and transport<strong>in</strong>g<br />
fi ne sediment, low mortality by fl ush<strong>in</strong>g fry<br />
and juveniles, low mortality by strand<strong>in</strong>g fry<br />
and juveniles on fl oodpla<strong>in</strong>s, reduce growth<br />
dur<strong>in</strong>g periods of high turbidity, reduce<br />
predation dur<strong>in</strong>g periods of high turbidity,<br />
ma<strong>in</strong>tenance of high quality aquatic habitat<br />
W<strong>in</strong>ter basefl ows F<strong>in</strong>e sediment transport Increase habitat area <strong>in</strong> natural channel<br />
morphology
110 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
The ‘Hydrograph Component Analysis’ illustrated <strong>in</strong><br />
Table 7.1 enables <strong>the</strong> user to connect <strong>the</strong> geomorphic and<br />
ecological consequences (or responses) to <strong>the</strong> results of<br />
a hydrological analysis. For example, snowmelt peaks<br />
and w<strong>in</strong>ter run-off occurr<strong>in</strong>g <strong>in</strong> wetter years may have<br />
signifi cantly different consequences <strong>in</strong> terms of enhanced<br />
channel change than those occurr<strong>in</strong>g <strong>in</strong> drier years.<br />
Channel changes have associated ecological detriments <strong>in</strong><br />
<strong>the</strong> shorter term, but provide ecological benefi ts <strong>in</strong> <strong>the</strong><br />
longer term. Reduc<strong>in</strong>g hydrological uncerta<strong>in</strong>ties before<br />
restoration (for design purposes) and after restoration (for<br />
monitor<strong>in</strong>g and appraisal purposes) thus requires <strong>the</strong><br />
deployment of techniques that span <strong>the</strong> determ<strong>in</strong>ation of<br />
low fl ows, which characterise <strong>the</strong> occurrence and magnitudes<br />
of high fl ows, and which identify any temporal<br />
changes or trends with respect to <strong>the</strong>se. Subsequently,<br />
l<strong>in</strong>kage of <strong>the</strong>se hydrological parameters to physical (sediment<br />
transport, channel morphological) and ecological<br />
responses may be undertaken. Whereas ‘classical’ hydrology<br />
has focused on determ<strong>in</strong><strong>in</strong>g catchment ‘signatures’<br />
with respect to <strong>the</strong> unit hydrograph (<strong>the</strong> average shape of<br />
storm hydrographs with equal distributions of ra<strong>in</strong>fall),<br />
m<strong>in</strong>imis<strong>in</strong>g uncerta<strong>in</strong>ty <strong>in</strong> restoration applications entails<br />
a more sophisticated paradigm for application. This<br />
requires consideration of hydrograph variability, <strong>the</strong><br />
potential l<strong>in</strong>ks between this and o<strong>the</strong>r properties of <strong>the</strong><br />
catchment network, as well as knowledge of land use<br />
management history. Collectively, it is <strong>the</strong>se which defi ne<br />
a ‘natural fl ow’ as <strong>the</strong> alternative paradigm for hydrological<br />
assessment and river restoration schemes.<br />
While that natural fl ow is now accepted as fundamental<br />
to improved management and study of rivers (Richter et<br />
al., 1996; Poff et al., 1997), it is also becom<strong>in</strong>g clear that<br />
<strong>the</strong> response of organisms to fl ood and drought may be<br />
evidenced over different time scales, and may be fundamentally<br />
different <strong>in</strong> character. Human modifi cations to<br />
<strong>the</strong> fl ow regime may also alter <strong>the</strong> course and viability of<br />
o<strong>the</strong>rwise natural adaptive strategies (Lytle and Poff,<br />
2004). Thus, while general pr<strong>in</strong>ciples or paradigms can be<br />
identifi ed, application to specifi c regions, case studies or<br />
species, may require considerable fi eld calibration, and/or<br />
engender considerable uncerta<strong>in</strong>ty <strong>in</strong> assess<strong>in</strong>g <strong>the</strong> output<br />
of models and scenarios. The follow<strong>in</strong>g sections describe<br />
some of <strong>the</strong> techniques used, and identify areas of uncerta<strong>in</strong>ty<br />
associated with <strong>the</strong>ir application and <strong>in</strong>terpretation<br />
<strong>in</strong> channel restoration designs.<br />
7.3.2 Determ<strong>in</strong><strong>in</strong>g <strong>the</strong> Environmental Flow Regime<br />
for <strong>River</strong> <strong>Restoration</strong>: Some Alternative Frameworks<br />
There is no simple formula or fi gure that can be given for<br />
<strong>the</strong> environmental fl ow requirements or <strong>in</strong>stream fl ow<br />
requirements of rivers. Much depends on <strong>the</strong> desired<br />
future character of <strong>the</strong> river ecosystem under study, which<br />
may be set by legislation or negotiated as a trade-off<br />
between water users. The fl ow allocated to a river may thus<br />
be primarily a matter of social choice, with science provid<strong>in</strong>g<br />
technical support to help determ<strong>in</strong>e <strong>the</strong> river ecosystem<br />
response under various fl ow regimes. Most scientifi c<br />
efforts have been directed to <strong>the</strong> determ<strong>in</strong>ation of m<strong>in</strong>imum<br />
fl ows <strong>in</strong> rivers, particularly dur<strong>in</strong>g dry periods of <strong>the</strong> year<br />
(or <strong>in</strong> drier years). These have been thought to be <strong>the</strong> most<br />
important limit<strong>in</strong>g factor <strong>in</strong> <strong>the</strong> ecological status or health<br />
of <strong>the</strong> river environment. Dur<strong>in</strong>g <strong>the</strong> past 20 years, a range<br />
of methods has been developed to help set environmental<br />
fl ows, each with advantages and disadvantages <strong>in</strong> particular<br />
circumstances of application.<br />
Where fl ow data exist, methods for <strong>the</strong> determ<strong>in</strong>ation<br />
of m<strong>in</strong>imum fl ows have been primarily statistical. Criteria<br />
for method selection <strong>in</strong>clude: <strong>the</strong> type of issue (abstraction,<br />
dam, run-of-river scheme); expertise, time and money<br />
available; and <strong>the</strong> legislative framework with<strong>in</strong> which <strong>the</strong><br />
fl ows must be set. Ungauged catchments give rise to particular<br />
uncerta<strong>in</strong>ties, which may be reduced by construct<strong>in</strong>g<br />
regional regression curves from rivers where data do<br />
exist, supplemented by statistical or geomorphologcal<br />
models of response (see Smakht<strong>in</strong>, 2001 for a comprehensive<br />
review of most methods and applications). Most<br />
recently, means of connect<strong>in</strong>g fl ow to modelled or observed<br />
ecological response have been developed to <strong>in</strong>form hydrological<br />
analyses. There are thus four broad categories of<br />
methods (Acreman and K<strong>in</strong>g, 2003) which may be used<br />
to reduce uncerta<strong>in</strong>ty <strong>in</strong> assess<strong>in</strong>g <strong>the</strong> hydrological basis<br />
for channel restoration designs: look-up tables; desk top<br />
analysis; functional analyses; and habitat modell<strong>in</strong>g. Each<br />
of <strong>the</strong>se methods may <strong>in</strong>volve different degrees of ‘expert’<br />
<strong>in</strong>volvement and may address all or just parts of <strong>the</strong> river<br />
system. Consequently, <strong>the</strong> use of experts and <strong>the</strong> degree<br />
to which methods are holistic are considered as crosscutt<strong>in</strong>g<br />
issues. The basic pr<strong>in</strong>ciples, strengths and weaknesses<br />
of each of <strong>the</strong>se methods are outl<strong>in</strong>ed below.<br />
Look-up Tables<br />
Worldwide, <strong>the</strong> most commonly applied methods to defi ne<br />
target river fl ows are rules-of-thumb based on simple<br />
<strong>in</strong>dices given <strong>in</strong> look-up tables. The most widely employed<br />
<strong>in</strong>dices are purely hydrological but those employ<strong>in</strong>g ecological<br />
data were also developed <strong>in</strong> <strong>the</strong> 1970s. Many early<br />
applications of environmental fl ow sett<strong>in</strong>g focused on<br />
s<strong>in</strong>gle species or s<strong>in</strong>gle issues. For example, much of <strong>the</strong><br />
demand for environmental fl ows <strong>in</strong> North America and<br />
nor<strong>the</strong>rn Europe was from sport fi shermen concerned<br />
about <strong>the</strong> decl<strong>in</strong>e <strong>in</strong> numbers of trout and salmon follow-
Hydrological and Hydraulic Aspects of <strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 111<br />
<strong>in</strong>g abstractions and dam operations. Environmental fl ows<br />
were <strong>the</strong>n set to ma<strong>in</strong>ta<strong>in</strong> critical levels of habitat (<strong>in</strong>clud<strong>in</strong>g<br />
sediment, fl ow velocity, depth) for <strong>the</strong>se species. Part<br />
of <strong>the</strong> justifi cation was that <strong>the</strong>se species are very sensitive<br />
to fl ow and if <strong>the</strong> fl ow is appropriate for <strong>the</strong>m, it will be<br />
for o<strong>the</strong>r parts of <strong>the</strong> ecosystem.<br />
Eng<strong>in</strong>eers have traditionally used hydrologicallydefi<br />
ned <strong>in</strong>dices for water management rules to set compensation<br />
fl ows below reservoirs and weirs. Examples are<br />
percentages of <strong>the</strong> mean fl ow or exceedence percentiles<br />
from a fl ow duration curve. (The fl ow duration curve is a<br />
water resources tool that defi nes <strong>the</strong> proportion of time<br />
that a given fl ow is equalled or exceeded). This approach<br />
has been adopted for environmental fl ow sett<strong>in</strong>g to determ<strong>in</strong>e<br />
simple operat<strong>in</strong>g rules for dams or off-take structures<br />
where few or no local ecological data are available.<br />
A hydrological <strong>in</strong>dex is used <strong>in</strong> France, where <strong>the</strong> Freshwater<br />
Fish<strong>in</strong>g Law of 1984 required that residual fl ows <strong>in</strong><br />
bypassed sections of river must be a m<strong>in</strong>imum of 1 – 40 of <strong>the</strong><br />
mean fl ow for exist<strong>in</strong>g schemes and 1 – 10 of <strong>the</strong> mean fl ow<br />
for new schemes (Souchon and Keith, 2001). In Brazil,<br />
fl ows below dams must be at least 80% of m<strong>in</strong>imum<br />
monthly average fl ow (Benetti et al., 2002). In regulat<strong>in</strong>g<br />
abstractions <strong>in</strong> <strong>the</strong> United K<strong>in</strong>gdom, an <strong>in</strong>dex of natural<br />
low fl ow has been employed to defi ne <strong>the</strong> environmental<br />
fl ow. Q 95 (i.e. that fl ow which is equalled or exceeded for<br />
95% of <strong>the</strong> time) is often used. The fi gure of Q 95 was<br />
chosen purely on hydrological grounds. However, <strong>the</strong><br />
implementation of this approach often <strong>in</strong>cludes ecological<br />
<strong>in</strong>formation (Barker and Kirmond, 1998).<br />
The Tennant Method (Tennant, 1976) was developed to<br />
specify m<strong>in</strong>imum fl ows to protect a healthy river environment.<br />
This employed calibration data from hundreds of<br />
rivers <strong>in</strong> <strong>the</strong> mid-Western states of <strong>the</strong> United States. Percentages<br />
of <strong>the</strong> mean annual fl ow are specifi ed that provide<br />
different quality habitat for fi sh, e.g. 10% for poor quality<br />
(survival), 30% for moderate habitat (satisfactory) and<br />
60% for excellent habitat. This approach can be used elsewhere<br />
but <strong>the</strong> exact <strong>in</strong>dices need to be re-calculated for<br />
each new region. The <strong>in</strong>dices are modifi ed where run-off<br />
<strong>in</strong> <strong>the</strong> spr<strong>in</strong>g is important and are widely used <strong>in</strong> plann<strong>in</strong>g<br />
at <strong>the</strong> river bas<strong>in</strong> level.<br />
Mat<strong>the</strong>ws and Bao (1991) concluded that methods<br />
based on proportions of mean fl ow were not suitable for<br />
<strong>the</strong> fl ow regimes of Texan rivers as <strong>the</strong>y often resulted <strong>in</strong><br />
an unrealistically high fl ow. Instead, <strong>the</strong>y devised a method<br />
us<strong>in</strong>g variable percentages of <strong>the</strong> monthly median fl ow,<br />
based on fi sh <strong>in</strong>ventories and known life history requirements,<br />
fl ow frequency distributions and conditions for<br />
special periods and processes (e.g. migration).<br />
The advantage of all look-up approaches is that once<br />
developed, application requires relatively few resources.<br />
However, simple hydrological <strong>in</strong>dices are not readily<br />
transferable between regions without re-calibration. Even<br />
<strong>the</strong>n, <strong>the</strong>y do not take account of site-specifi c conditions.<br />
In particular, adjust<strong>in</strong>g hydrological <strong>in</strong>dices does not<br />
ensure ecological validity without a correspond<strong>in</strong>g<br />
ecological re-assessment, but ecological data may be<br />
much more costly and time consum<strong>in</strong>g to collect. In<br />
general, look-up tables are thus particularly appropriate<br />
for low controversy situations. They also tend to be<br />
precautionary.<br />
Desk Top Analysis<br />
Desk top analyses tend to focus on analysis of exist<strong>in</strong>g<br />
rout<strong>in</strong>e data (such as river fl ows from gaug<strong>in</strong>g stations<br />
and/or fi sh data from regular surveys) although data may<br />
be collected as part of a specifi c project at a particular site<br />
or sites on a river. Some desk top methods are purely<br />
hydrological. For example, Richter et al., (1996, 1997)<br />
developed a hydrological method <strong>in</strong>tended for sett<strong>in</strong>g<br />
benchmark fl ows on rivers, where a natural ecosystem is<br />
<strong>the</strong> primary objective. Development of <strong>the</strong> method relies<br />
upon identifi cation of <strong>the</strong> components of a natural fl ow<br />
regime, <strong>in</strong>dexed by magnitude (of both high and low<br />
fl ows), tim<strong>in</strong>g (<strong>in</strong>dexed by monthly statistics), frequency<br />
(number of events) and duration (<strong>in</strong>dexed by mov<strong>in</strong>g<br />
average m<strong>in</strong>ima and maxima). The method uses gauged or<br />
modelled daily fl ows and a set of 32 <strong>in</strong>dices. A range of<br />
variation of <strong>the</strong> <strong>in</strong>dices may <strong>the</strong>n be set, based upon ±1<br />
standard deviation from <strong>the</strong> mean or between <strong>the</strong> 25th and<br />
75th percentiles. Variability <strong>in</strong> stream fl ow is essential <strong>in</strong><br />
susta<strong>in</strong><strong>in</strong>g ecosystem <strong>in</strong>tegrity (long term ma<strong>in</strong>tenance of<br />
biodiversity and productivity) and resiliency (<strong>the</strong> capacity<br />
to endure natural and human disturbances – Stanford<br />
et al., 1996). This method is <strong>in</strong>tended to defi ne <strong>in</strong>terim<br />
standards, which can be monitored and revised. However,<br />
so far, <strong>the</strong>re has not been enough research to relate <strong>the</strong><br />
fl ow statistics to specifi c elements of <strong>the</strong> ecosystem.<br />
A particular aspect of desk top analysis relates to <strong>the</strong><br />
characterisation of high fl ows <strong>in</strong> rivers, especially <strong>the</strong><br />
identifi cation of group<strong>in</strong>gs of higher fl ow events, ei<strong>the</strong>r<br />
with<strong>in</strong> or between years. The emphasis on higher fl ows is<br />
someth<strong>in</strong>g of a counter to <strong>the</strong> previous concentration of<br />
hydrological analyses on low fl ow requirements. It refl ects<br />
<strong>the</strong> importance of higher magnitude fl ows <strong>in</strong> connect<strong>in</strong>g<br />
<strong>the</strong> various parts of <strong>the</strong> catchment–fl oodpla<strong>in</strong>–channel<br />
system, <strong>in</strong> susta<strong>in</strong><strong>in</strong>g sediment transport through <strong>the</strong><br />
fl uvial system to ma<strong>in</strong>ta<strong>in</strong> channel and fl oodpla<strong>in</strong> structure<br />
(<strong>in</strong>clud<strong>in</strong>g larger-scale riparian woodlands – Gregory et<br />
al., 2003) and <strong>in</strong> servic<strong>in</strong>g particular stages of species’ life<br />
cycles. In this context, fl ow variability or hydrological<br />
disturbance is thus both a potential <strong>in</strong>dicator of land use
112 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
change and an important control on river ecology. Knowledge<br />
of <strong>the</strong> chang<strong>in</strong>g attributes of such disturbance is<br />
important <strong>in</strong> sett<strong>in</strong>g <strong>the</strong> appropriate historical and contemporary<br />
context for stream fl ow restoration (through consideration<br />
of land use and climatic histories), and for<br />
design<strong>in</strong>g-<strong>in</strong> crucial aspects of fl ow variability.<br />
By way of illustration, Archer and Newson (2002) focus<br />
on <strong>the</strong> number and frequency of rises and falls above<br />
selected threshold levels (pulses) expressed as multiples<br />
of <strong>the</strong> median fl ow, toge<strong>the</strong>r with <strong>the</strong> duration of <strong>the</strong> events<br />
obta<strong>in</strong>ed when analys<strong>in</strong>g <strong>the</strong> discharge history of <strong>the</strong> Coalburn<br />
catchment, UK (Figure 7.3). The pulse duration is<br />
<strong>the</strong> time from ris<strong>in</strong>g above <strong>the</strong> threshold to fall<strong>in</strong>g below<br />
<strong>the</strong> same threshold, and <strong>the</strong> number of pulses refl ects <strong>the</strong><br />
magnitude of <strong>the</strong> chosen threshold fl ow. Provided that <strong>the</strong><br />
data series are extensive enough, of suffi cient resolution<br />
(for example, daily mean fl ows are unlikely to be an adequate<br />
basis for analysis when catchment lag is much less<br />
than one day) and are accompanied by o<strong>the</strong>r appropriate<br />
documentary/archival <strong>in</strong>formation, <strong>the</strong> analysis may be<br />
extended to consider periods of land use or climatic<br />
change, as shown <strong>in</strong> Figure 7.4.<br />
The Coalburn is a small upland catchment with an area<br />
of 1.5 km 2 and an altitud<strong>in</strong>al range from 270–330 m. The<br />
natural surface material comprises a cover of blanket peat<br />
(0.5–3 m thick) overly<strong>in</strong>g glacial till up to 5 m <strong>in</strong> thickness.<br />
This was ploughed <strong>in</strong> 1972, result<strong>in</strong>g <strong>in</strong> a dra<strong>in</strong>age density<br />
60 times greater than <strong>the</strong> orig<strong>in</strong>al stream network. In <strong>the</strong><br />
spr<strong>in</strong>g of 1973, 90% of <strong>the</strong> catchment was planted with<br />
Sitka spruce and, s<strong>in</strong>ce <strong>the</strong>n, growth rates have been variable,<br />
reach<strong>in</strong>g 1 m height <strong>in</strong> 1978 and 7–12 m <strong>in</strong> 1996, by<br />
which time some 60% of <strong>the</strong> catchment had reached<br />
canopy closure. With this knowledge of land use management,<br />
<strong>the</strong> hydrological data can be analysed as shown <strong>in</strong><br />
Discharge<br />
1 2<br />
1<br />
1<br />
2<br />
Time<br />
Figure 7.4 to reveal phases of hydrological response to <strong>the</strong><br />
land use change. Thus, immediately follow<strong>in</strong>g plough<strong>in</strong>g<br />
and plant<strong>in</strong>g, pulse numbers and total pulse duration<br />
<strong>in</strong>crease, and subsequently decl<strong>in</strong>e with vegetation maturity,<br />
whereas <strong>the</strong> average duration of <strong>in</strong>dividual events<br />
<strong>in</strong>creases as vegetation matures. These patterns <strong>in</strong> hydrological<br />
response are most clearly shown when <strong>the</strong> analysis<br />
is applied to events defi ned by thresholds of 2–6 times <strong>the</strong><br />
median fl ow. Importantly, <strong>the</strong> method fur<strong>the</strong>r demonstrates<br />
that, even where peak and time-to-peak of <strong>the</strong> unit hydrograph<br />
are similar <strong>in</strong> pre- and post-disturbance situations,<br />
<strong>the</strong>re are differences <strong>in</strong> pulse numbers and pulse magnitudes,<br />
which are thought to help determ<strong>in</strong>e habitat potential<br />
and ecological response. Not surpris<strong>in</strong>gly given<br />
its emphasis on higher magnitude threshold events, <strong>the</strong><br />
method is least satisfactory with respect to disclos<strong>in</strong>g<br />
behaviour of low fl ows, and this limitation is important <strong>in</strong><br />
relation to use of <strong>the</strong> technique to ‘set’ regulated low fl ows<br />
as described above. In such cases, it may, however, be<br />
supplemented by o<strong>the</strong>r procedures and <strong>the</strong> method should<br />
be thought of as complement<strong>in</strong>g traditional approaches so<br />
as to <strong>in</strong>clude hydrological variability as a key ecological<br />
determ<strong>in</strong>ant.<br />
Ano<strong>the</strong>r area of research that may be considered under<br />
<strong>the</strong> head<strong>in</strong>g of desk top analysis (but which underp<strong>in</strong>s<br />
o<strong>the</strong>r methodologies described below) is <strong>the</strong> prediction of<br />
hydrograph characteristics from catchment channel<br />
network properties. Here, <strong>the</strong> focus of attention has been<br />
on <strong>the</strong> effects of network scale and <strong>the</strong> relative tim<strong>in</strong>g<br />
of <strong>the</strong> hillslope hydrograph and channel water rout<strong>in</strong>g.<br />
Because <strong>the</strong> density of channels <strong>in</strong> a river network refl ects<br />
climate and land use, <strong>the</strong>re should be a network response<br />
(transformation) to any changes <strong>in</strong> <strong>the</strong>se driv<strong>in</strong>g variables<br />
(Kirkby, 1993). The goal has been to <strong>in</strong>corporate network<br />
2 3<br />
3 × Median<br />
2 × Median<br />
Median<br />
3<br />
4 5<br />
Figure 7.3 Defi nition diagram of pulses above selected thresholds and pulse duration (between arrows) <strong>in</strong> hydrological time series<br />
(Repr<strong>in</strong>ted from D. Archer et al. (2002), Journal of Hydrology 268, 244–258, with permission from Elsevier.)
Hydrological and Hydraulic Aspects of <strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 113<br />
Annual number of pulses<br />
Total duration (hrs) above threshold<br />
Average duration of flow above<br />
80<br />
70<br />
60<br />
50<br />
40<br />
20<br />
10<br />
0<br />
5000<br />
4500<br />
4000<br />
3500<br />
3000<br />
2500<br />
2000<br />
1500<br />
1000<br />
500<br />
0<br />
threshold (hrs) 30<br />
35<br />
30<br />
25<br />
20<br />
15<br />
10<br />
5<br />
0<br />
>0.5M<br />
>M<br />
>2M<br />
>3M<br />
>4M<br />
>0.5M<br />
>M<br />
>2M<br />
>3M<br />
>4M<br />
>0.5M<br />
>M<br />
>2M<br />
>3M<br />
>4M<br />
>5M<br />
>5M<br />
>5M<br />
>6M<br />
>7M<br />
>8M<br />
>10M<br />
>15M<br />
(a)<br />
(b)<br />
(c)<br />
>20M<br />
Multiples of median flow<br />
>6M<br />
>7M<br />
>8M<br />
>10M<br />
>15M<br />
>20M<br />
>30M<br />
>40M<br />
>30M<br />
Multiples of median flow<br />
>6M<br />
>7M<br />
>8M<br />
>10M<br />
>15M<br />
>20M<br />
>30M<br />
Multiples of median flow<br />
>50M<br />
>60M<br />
1967-99<br />
1967-71<br />
1974-82<br />
1983-90<br />
1992-99<br />
>80M<br />
>40M<br />
>50M<br />
>60M<br />
>80M<br />
>40M<br />
>50M<br />
>60M<br />
>80M<br />
>100M<br />
1967-99<br />
1967-71<br />
1974-82<br />
1983-90<br />
1992-99<br />
>100M<br />
1967-99<br />
1967-71<br />
1974-82<br />
1983-90<br />
1992-99<br />
Figure 7.4 Pulse number (a), total pulse duration (b) and average pulse duration (c) for <strong>the</strong> Coalburn catchment over <strong>the</strong> full range<br />
of fl ows <strong>in</strong> pre- and post-dra<strong>in</strong>age and plant<strong>in</strong>g periods (Repr<strong>in</strong>ted from D. Archer et al. (2002), Journal of Hydrology 268, 244–258,<br />
with permission from Elsevier.)<br />
>100M
114 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
geometry <strong>in</strong>to a catchment hydrograph forecast<strong>in</strong>g model<br />
and this has given rise to <strong>the</strong> Geomorphological Unit<br />
Hydrograph (GUH). This defi nes <strong>the</strong> average shape of<br />
storm hydrographs with equal distributions of ra<strong>in</strong>fall as<br />
a function of bifurcation ratio, bas<strong>in</strong> area ratio, stream<br />
length ratio, watershed order and a mean wait<strong>in</strong>g time for<br />
each order (Rodriguez-Iturbe and Valdes, 1979):<br />
and<br />
peak discharge = 1.31/ L ΩRL 0.43<br />
time to peak = 0.44 L Ω RB 0.55 RA −0.55 RL −0.38<br />
where RB = bifurcation ratio; RA = bas<strong>in</strong> area ratio; RL =<br />
stream length ratio; and bas<strong>in</strong> order = Ω.<br />
The GUH may be criticised because of its dependence<br />
on network properties that are apparently <strong>in</strong>sensitive to<br />
changes and its neglect of some aspects describ<strong>in</strong>g network<br />
shape, topology and topography. More recent developments<br />
of this approach are directed to <strong>in</strong>corporat<strong>in</strong>g<br />
measures of channel network width and network l<strong>in</strong>k concentration<br />
functions (for detailed review, see Mousouridis,<br />
2001).<br />
O<strong>the</strong>r research has also related river fl ow directly to<br />
ecological data, such as population numbers or <strong>in</strong>dices<br />
of community structure calculated from species lists.<br />
However, it is diffi cult to derive biotic <strong>in</strong>dices that are only<br />
sensitive to fl ow and not to o<strong>the</strong>r factors (e.g. habitat<br />
structure or water quality). At <strong>the</strong> m<strong>in</strong>imum, biotic <strong>in</strong>dices<br />
designed for water quality monitor<strong>in</strong>g purposes should be<br />
used with extreme caution (Armitage and Petts, 1992).<br />
Generally, such data are scarce, and <strong>in</strong>terpretation is complicated<br />
where methods of collection encompass po<strong>in</strong>tspecifi<br />
c measurements obta<strong>in</strong>ed at different times, and<br />
spatially-distributed methods pick<strong>in</strong>g-out more persistent<br />
patterns refl ect<strong>in</strong>g longer term aspects of land use and<br />
water quality (Dakova et al., 2000). In addition, it is not<br />
always clear which fl ow variables to choose to represent<br />
different aspects of <strong>the</strong> fl ow regime which are of<br />
most ecological relevance. Studies which directly relate<br />
‘responses’ such as macro<strong>in</strong>vertebrate abundance to<br />
hydro-climatological and sediment variables confi rm <strong>the</strong><br />
importance of <strong>in</strong>termediate levels of disturbance as underp<strong>in</strong>n<strong>in</strong>g<br />
greatest habitat diversity, and thus imply that<br />
modell<strong>in</strong>g must <strong>in</strong>corporate more than one aspect of<br />
hydrological diversity (Reice et al., 1990).<br />
The lack of complementary hydrological and biological<br />
data is often a limit<strong>in</strong>g factor, and sometimes rout<strong>in</strong>ely<br />
collected data ga<strong>the</strong>red for o<strong>the</strong>r purposes turn out to be<br />
unsuitable. In addition, time series may not be <strong>in</strong>dependent<br />
(which can violate assumptions of classical statistical<br />
techniques) while <strong>the</strong> representation and characterisation<br />
of disturbances (‘extremes’) necessitates longer, higherquality<br />
time series. For example, Clausen and Biggs (2000)<br />
exam<strong>in</strong>ed 35 fl ow variables us<strong>in</strong>g daily mean fl ows for a<br />
seven-year record common to 62 perennial rivers <strong>in</strong> New<br />
Zealand. Based upon a covariance analysis among <strong>the</strong> sites<br />
through a pr<strong>in</strong>cipal components analysis, <strong>the</strong> 35 variables<br />
could be collapsed <strong>in</strong>to four variable group<strong>in</strong>gs relat<strong>in</strong>g<br />
to: size of river; overall variability of <strong>the</strong> fl ow; volume of<br />
high fl ow; and frequency of high fl ow. Signifi cantly, <strong>the</strong><br />
statistical properties (particularly <strong>in</strong>tra-annual variability)<br />
varied between groups, necessitat<strong>in</strong>g <strong>the</strong> use of a suite of<br />
different variables from each group to adequately represent<br />
<strong>the</strong> facets of fl ow of most ecological relevance. This illustrates<br />
<strong>the</strong> requirements for high quality data and, too, <strong>the</strong><br />
uncerta<strong>in</strong>ties of <strong>in</strong>terpretation and application.<br />
A method recently developed <strong>in</strong> <strong>the</strong> United K<strong>in</strong>gdom<br />
<strong>in</strong>volv<strong>in</strong>g ecological data is <strong>the</strong> Lotic Invertebrate Index<br />
for Flow Evaluation (LIFE; Extence et al., 1999). The<br />
LIFE score is based on <strong>the</strong> abundance and sensitivity to<br />
water velocity of different taxa, collected <strong>in</strong> rout<strong>in</strong>e macro<strong>in</strong>vertebrate<br />
monitor<strong>in</strong>g data. Mov<strong>in</strong>g averages of preced<strong>in</strong>g<br />
fl ow have shown good relationships with LIFE scores<br />
over a range of sites (Figure 7.5), but it is as yet uncerta<strong>in</strong><br />
as to how <strong>the</strong> approach can be used to manage river fl ows.<br />
Never<strong>the</strong>less, <strong>the</strong> pr<strong>in</strong>ciple is believed to be sound and<br />
LIFE has <strong>the</strong> major advantage of utilis<strong>in</strong>g <strong>the</strong> data collected<br />
by exist<strong>in</strong>g bio-monitor<strong>in</strong>g programmes.<br />
Functional Analysis Methods – <strong>the</strong> Build<strong>in</strong>g<br />
Block Methodology<br />
The third group of methods to determ<strong>in</strong>e fl ow requirements<br />
for restoration and ecological purposes <strong>in</strong>cludes<br />
those that build an understand<strong>in</strong>g of <strong>the</strong> functional l<strong>in</strong>ks<br />
between hydrology and ecology <strong>in</strong> <strong>the</strong> entire river system.<br />
These methods take a broad view and comb<strong>in</strong>e hydrological<br />
analysis, hydraulic rat<strong>in</strong>g <strong>in</strong>formation (to estimate<br />
sediment transport and channel capacity) and many aspects<br />
of <strong>the</strong> river ecosystem. Perhaps <strong>the</strong> best known is <strong>the</strong><br />
Build<strong>in</strong>g Block Methodology (BBM) developed <strong>in</strong> South<br />
Africa (Rowntree and Wadeson, 1998; K<strong>in</strong>g et al., 2000).<br />
The basic premise of <strong>the</strong> BBM is that river<strong>in</strong>e species are<br />
reliant on three basic elements (build<strong>in</strong>g blocks) of <strong>the</strong><br />
fl ow regime, each of which has a particular ecological and<br />
geomorphological signifi cance: low fl ows, freshets and<br />
fl oods. An acceptable fl ow regime for ecosystem ma<strong>in</strong>tenance<br />
can thus be constructed by comb<strong>in</strong><strong>in</strong>g <strong>the</strong>se build<strong>in</strong>g<br />
blocks. In a typical exercise to determ<strong>in</strong>e <strong>in</strong>stream fl ow<br />
requirements through <strong>the</strong> build<strong>in</strong>g block method, fi ve<br />
stages may be <strong>in</strong>volved (Rowntree and Wadeson, 1998): a<br />
general assessment of catchment condition to determ<strong>in</strong>e
Hydrological and Hydraulic Aspects of <strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 115<br />
Discharge (m³/s)<br />
10.00<br />
1.00<br />
0.10<br />
0.01<br />
0.00<br />
6.0<br />
86 87 88 88 89 90 91 92 93 94 95 96 97 98 99 00<br />
Year<br />
Flow<br />
Figure 7.5 Example <strong>River</strong> Flow (logarithmic scale) and LIFE Score time series<br />
potential for geomorphological change <strong>in</strong> a river; an evaluation<br />
of river network characteristics to help identify representative<br />
river reaches to determ<strong>in</strong>e <strong>the</strong> <strong>in</strong>stream fl ow<br />
requirements; an assessment of <strong>the</strong> hydraulic diversity at<br />
various stages for each of <strong>the</strong> morphological units present<br />
<strong>in</strong> <strong>the</strong>se reaches (see below for discussion of biotopes); an<br />
assessment of <strong>the</strong> magnitude and importance of freshet<br />
and fl oods; and an assessment of any <strong>in</strong>ter-bas<strong>in</strong> water<br />
transfers.<br />
The BBM thus revolves around a team of physical scientists<br />
(e.g. hydrologists and geomorphologists) and biological<br />
scientists (e.g. botanists and fi sh biologists) who<br />
follow a series of structured analyses to come to a consensus<br />
on <strong>the</strong> build<strong>in</strong>g blocks of <strong>the</strong> fl ow regime. The BBM<br />
has a detailed manual for implementation (K<strong>in</strong>g et al.,<br />
2000); is rout<strong>in</strong>ely used <strong>in</strong> South Africa to comply with<br />
<strong>the</strong> 1998 Water Act; has been applied <strong>in</strong> Australia (Arth<strong>in</strong>gton<br />
and Long 1997, Arth<strong>in</strong>gton and Lloyd, 1998); and is<br />
be<strong>in</strong>g trialled <strong>in</strong> <strong>the</strong> United States.<br />
To overcome <strong>the</strong> diffi culties <strong>in</strong> relat<strong>in</strong>g changes <strong>in</strong> <strong>the</strong><br />
fl ow regime directly to <strong>the</strong> response of multiple species<br />
and communities, approaches have been developed that<br />
use habitat for target species as an <strong>in</strong>termediate step. Of<br />
those environmental conditions required by an <strong>in</strong>dividual<br />
animal liv<strong>in</strong>g <strong>in</strong> a river, it is <strong>the</strong> physical aspects that are<br />
most heavily impacted by changes to <strong>the</strong> fl ow regime.<br />
Habitat models seek to l<strong>in</strong>k data on <strong>the</strong> physical conditions<br />
(such as water depths and velocities) <strong>in</strong> rivers at different<br />
fl ows (ei<strong>the</strong>r measured data or derived from computer<br />
models) with data on those physical conditions which key<br />
animal or plant species (or <strong>the</strong>ir <strong>in</strong>dividual developmental<br />
stages) require. Once functional relationships between<br />
physical habitat and ecology have been defi ned, <strong>the</strong>y are<br />
l<strong>in</strong>ked to fl ow scenarios <strong>in</strong> river restoration and design (see<br />
Section 7.3 for fur<strong>the</strong>r details on habitat modell<strong>in</strong>g).<br />
Life Score<br />
8.0<br />
7.5<br />
7.0<br />
6.5<br />
LIFE Score<br />
The Holistic Approach<br />
Where <strong>in</strong>tervention seeks fl ow remediation or restoration,<br />
<strong>the</strong>re is an implicitly holistic approach <strong>in</strong>sofar as all elements<br />
of <strong>the</strong> river ecosystem are likely to be supported.<br />
However, more and more methods now adopt an overtly<br />
holistic assessment of <strong>the</strong> whole ecosystem, such as<br />
associated wetlands, groundwater and estuaries; all<br />
species that are sensitive to fl ow (<strong>in</strong>vertebrates, plants and<br />
animals); <strong>the</strong> human context; and all aspects of <strong>the</strong> hydrological<br />
regime, <strong>in</strong>clud<strong>in</strong>g fl oods, droughts, and water<br />
quality. A fundamental pr<strong>in</strong>ciple is to ma<strong>in</strong>ta<strong>in</strong> natural<br />
variability of fl ows and to allocate fl ow based upon likely<br />
habitat impact. This ecosystem approach is especially<br />
important <strong>in</strong> <strong>the</strong> management and restoration of dryland<br />
environments, where <strong>in</strong>herent variability of hydrological<br />
‘extremes’ is a limit<strong>in</strong>g factor ecologically (Thoms and<br />
Sheldon, 2002). Ecosystem approaches <strong>in</strong>herently allow<br />
for complexity and diversity, and may also capitalise on<br />
features such as persistence and evolution – ecological<br />
systems may have self-organis<strong>in</strong>g or self-design<strong>in</strong>g<br />
attributes (Bergen et al., 2001).<br />
Generally, holistic approaches make use of teams of<br />
experts and may also <strong>in</strong>volve <strong>the</strong> participation of stakeholders,<br />
thus extend<strong>in</strong>g holism beyond scientifi c issues.<br />
Where methods have <strong>the</strong> characteristic of be<strong>in</strong>g holistic<br />
<strong>the</strong>y clearly have <strong>the</strong> advantage of cover<strong>in</strong>g <strong>the</strong> whole<br />
hydrological–ecological–stakeholder system. The disadvantage<br />
is that it is diffi cult to identify and to collect <strong>the</strong><br />
relevant data which capture such diverse perspectives.<br />
7.3.3 Choice of Method<br />
Selection of <strong>the</strong> most appropriate method for l<strong>in</strong>k<strong>in</strong>g<br />
hydrological criteria to river restoration designs, where a
116 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Table 7.2 Some advantages and disadvantages of different methods and characteristics of sett<strong>in</strong>g environmental fl ows<br />
Method type Sub-type Advantages Disadvantages and Uncerta<strong>in</strong>ties<br />
Look-up table Hydrological Cheap, rapid to use once calculated Not site-specifi c. Hydrological <strong>in</strong>dices not<br />
valid ecologically<br />
Ecological Ecological <strong>in</strong>dices need region-specifi c data<br />
to be calculated<br />
Desk top Hydrological Site specifi c<br />
Limited new data collection<br />
Long time series required<br />
Hydraulic No explicit use of ecological data<br />
Ecological Ecological data time consum<strong>in</strong>g to collect<br />
Functional analysis Flexible, robust, more focused on<br />
whole ecosystem<br />
range (or set) of methods is available, is considered briefl y<br />
here. Some of <strong>the</strong> advantages and disadvantages which<br />
underlie <strong>the</strong> uncerta<strong>in</strong>ties <strong>in</strong> us<strong>in</strong>g different approaches are<br />
summarised <strong>in</strong> Table 7.2.<br />
Broadly, Table 7.2 illustrates two trade-offs: that between<br />
local sensitivity or regional coverage, and that between<br />
data availability and complexity or simplicity of data <strong>in</strong>terpretation.<br />
Mov<strong>in</strong>g through <strong>the</strong> table, <strong>the</strong> more <strong>in</strong>tegrated<br />
and hence more complex methods are characterised by a<br />
greater degree of connection between physical and biological<br />
parameters. The key element of this <strong>in</strong>tegration is to<br />
describe or model <strong>the</strong> effects of fl ows on habitat structure<br />
and function. In turn, this necessitates <strong>the</strong> addition of<br />
hydraulic and hydrodynamic considerations to supplement<br />
<strong>the</strong> hydrological foundations described above.<br />
7.4 HYDRAULIC ASPECTS AND UNCERTAINTY<br />
IN CHANNEL RESTORATION<br />
7.4.1 Connect<strong>in</strong>g Flows, Sediments and Ecological<br />
Response: <strong>the</strong> Physical Habitat and Eco-Hydraulics<br />
Over <strong>the</strong> past decade, <strong>the</strong>re has been a trend towards<br />
approach<strong>in</strong>g river habitat assessment and rehabilitation<br />
design us<strong>in</strong>g comb<strong>in</strong>ations of fi eld survey and predictive<br />
hydraulic models (Chapter 5). ‘Eco-hydraulics’ (Leclerc,<br />
2002) is now a commonplace term <strong>in</strong> both <strong>the</strong> academic<br />
and practitioner literature, and <strong>the</strong> l<strong>in</strong>kage between physical,<br />
chemical and biotic components of <strong>the</strong> river environment<br />
is central <strong>in</strong> efforts to restore and ma<strong>in</strong>ta<strong>in</strong> ecological<br />
habitat and function (Kemp et al., 2000). To realise <strong>the</strong>se<br />
benefi ts, however, cost-effective (but scientifi cally sound)<br />
means of comb<strong>in</strong><strong>in</strong>g traditional fi eld monitor<strong>in</strong>g and<br />
Expensive to collect all relevant data and to<br />
employ wide range of experts. Consensus<br />
of experts may not be achieved.<br />
Habitat modell<strong>in</strong>g Replicable, predictive Expensive to collect hydraulic and ecological<br />
data<br />
survey with emerg<strong>in</strong>g modell<strong>in</strong>g and design-support<br />
approaches are required. There rema<strong>in</strong>s much work to be<br />
done both to encourage multi-discipl<strong>in</strong>ary co-operation<br />
and to ‘<strong>in</strong>vent’ new trans-discipl<strong>in</strong>ary areas of research<br />
and practical expertise (Janauer, 2000).<br />
Physical habitat can be defi ned as a set of physical<br />
conditions that can be measured and compared to <strong>the</strong><br />
conditions that may be suitable for specifi c species or<br />
<strong>in</strong>dividuals at a particular stage of <strong>the</strong>ir life cycle. The<br />
fundamental aspects of <strong>the</strong> eco-hydraulic approach to river<br />
characterisation and restoration are summarised <strong>in</strong> Figure<br />
7.6. This is based upon <strong>the</strong> idea that species assemblages<br />
and/or abundances are organised to refl ect physical conditions<br />
at a range of scales: catchment conditions (such as<br />
slope, geology and ra<strong>in</strong>fall regime) determ<strong>in</strong>e at <strong>the</strong><br />
most fundamental level <strong>the</strong> k<strong>in</strong>d of species which may be<br />
present; at <strong>the</strong> reach scale, assemblages or abundances<br />
from with<strong>in</strong> this broader range are most likely to be<br />
observed depend<strong>in</strong>g on fl ow regime; at <strong>the</strong> sub-reach<br />
scale, particular species or <strong>in</strong>dividuals at particular stages<br />
of <strong>the</strong>ir life cycle, are found <strong>in</strong> niches with particular fl ow<br />
and sedimentological conditions.<br />
While <strong>the</strong> catchment scale conditions <strong>the</strong> range of<br />
habitat potential, at <strong>the</strong> smaller scales physical habitat is<br />
commonly defi ned by comb<strong>in</strong>ations of depth, velocity,<br />
substrate and distance to cover. In this respect, physical<br />
habitat quality is normally considered <strong>in</strong>dependently of<br />
water quality issues. It may be approached (at its simplest)<br />
from visual survey, through statistical models based upon<br />
measured associations, and via complex determ<strong>in</strong>istic<br />
modell<strong>in</strong>g, <strong>in</strong> which <strong>the</strong> physical habitat is fi rst modelled<br />
and <strong>the</strong>n l<strong>in</strong>ked to ecological response. Although <strong>the</strong><br />
potential of eco-hydraulic modell<strong>in</strong>g has been generally
DRAINAGE<br />
BASIN<br />
10 6 -10 5 years<br />
Hydrological and Hydraulic Aspects of <strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 117<br />
10 4 -10 3 m<br />
10 3 -10 2 m<br />
FLOODPLAIN<br />
10 4 -10 3 years<br />
10 2 -10 1 m<br />
REACH<br />
10 2 -10 1 years<br />
HABITAT<br />
10 1 -10 0 years<br />
10 –1 m<br />
Sand silt<br />
over cobbles<br />
Gravel<br />
MICROHABITAT<br />
10 0 -10 –1 years<br />
Aquatic and<br />
semi-aquatic<br />
vegetation<br />
Leaf and stick<br />
detritus <strong>in</strong><br />
marg<strong>in</strong><br />
Figure 7.6 Spatial scale and distribution of stream habitats based upon fl ow–sediment <strong>in</strong>teractions (Source: Naiman et al., 1992<br />
after Frissell et al., 1986)<br />
Fall<br />
Rapid<br />
Marg<strong>in</strong>al<br />
Deadwater<br />
Chute<br />
acknowledged positively, some signifi cant areas of uncerta<strong>in</strong>ty<br />
rema<strong>in</strong>. A fundamental requirement is <strong>the</strong> ability<br />
to demonstrate (ra<strong>the</strong>r than assume) a l<strong>in</strong>kage between<br />
measured parameters delimit<strong>in</strong>g physical habitat and<br />
<strong>the</strong> species occurrence and abundances (‘responses’ or<br />
‘tolerances’) which <strong>the</strong>se are supposed to condition.<br />
7.4.2 Survey and Visual Methods: Biotopes and<br />
Functional Habitats<br />
At present, hydraulic performance and habitat are essentially<br />
approached separately. Two alternatives are available:<br />
‘bottom up’, <strong>in</strong> which <strong>in</strong>-stream habitat units are<br />
PHYSICAL BIOTOPES FUNCTIONAL (MESO-) HABITATS<br />
Riffle<br />
Boll<br />
Run<br />
glide<br />
Pool<br />
Float<strong>in</strong>g<br />
leaves<br />
Submerged<br />
plants<br />
(slow)<br />
Cobbles/<br />
pobbles<br />
Rock Woody<br />
debris<br />
Tree roots/<br />
overhang<strong>in</strong>g<br />
vegetation<br />
(fast)<br />
Gravel Emergent<br />
plants<br />
Sand<br />
Silt<br />
Marg<strong>in</strong>al<br />
plants<br />
Figure 7.7 Representation of physical biotopes and functional habitats <strong>in</strong> a stream sub-reach (Reproduced from M. D. Newson and<br />
C. L. Newson (2000) ‘Geomorphology, ecology and river channel habitat: mesoscale approaches to bas<strong>in</strong>-scale challenges’ Progress<br />
<strong>in</strong> Physical Geography 24, 195–217, with <strong>the</strong> permission of Edward Arnold (Publishers) Ltd.)<br />
<strong>in</strong>ferred from a knowledge of hydraulic conditions that<br />
defi ne physical biotopes (channel features characterised<br />
through hydraulic measurement or visual survey of surface<br />
fl ow type; Padmore, 1997); and ‘top down’, <strong>in</strong> which<br />
functional or meso- habitats are <strong>in</strong>ferred from analysis of<br />
biological communities associated with substrate and vegetation<br />
characteristics (Kemp et al., 1999). A common<br />
strategy is to visually classify velocity–depth comb<strong>in</strong>ations<br />
(i.e. <strong>the</strong> biotopes) which might <strong>the</strong>n be associated<br />
with functional habitats, or discrete species assemblages<br />
as <strong>in</strong>dicated <strong>in</strong> Figure 7.7 (Newson and Newson, 2000).<br />
Biotopes are easily recognised from fi eld survey<br />
without costly fi eld measurement and, when l<strong>in</strong>ked with
118 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
knowledge of habitat suitability, allow some <strong>in</strong>sight <strong>in</strong>to<br />
likely biological function. Despite early optimism with<br />
respect to this approach, relat<strong>in</strong>g biotopes to mean<strong>in</strong>gful<br />
biotic function appears more complicated than fi rst<br />
assumed (Kemp et al., 2000; Leclerc, 2002). Demonstrable<br />
relationships between biotopes and functional habitats<br />
rema<strong>in</strong> scarce. While <strong>the</strong> survey or biotope method is<br />
meant to circumvent issues relat<strong>in</strong>g to s<strong>in</strong>gle-species preference<br />
curves or s<strong>in</strong>gle measures of channel environment<br />
(s<strong>in</strong>ce <strong>the</strong> biotope is an ‘<strong>in</strong>tegrated’ habitat), major uncerta<strong>in</strong>ties<br />
<strong>in</strong> improv<strong>in</strong>g l<strong>in</strong>ks between biotopes (from simple<br />
survey or modell<strong>in</strong>g) and functional habitats rema<strong>in</strong><br />
(Newson and Newson, 2000). In part, <strong>the</strong> scarcity of data<br />
refl ects <strong>the</strong> cost and practical limitations associated with<br />
fi ne-scale fi eld survey (Newson et al., 1998) and diffi culties<br />
<strong>in</strong> monitor<strong>in</strong>g highly dynamic, stage-dependent fl ow<br />
characteristics. There is also <strong>the</strong> question of <strong>the</strong> appropriate<br />
range of supposed habitat determ<strong>in</strong>ants to associate<br />
with particular species or species clusters, at differ<strong>in</strong>g<br />
stages of <strong>the</strong>ir life cycles, and which may vary greatly<br />
between species found even <strong>in</strong> close proximity. Key questions<br />
which belie <strong>the</strong> uncerta<strong>in</strong>ty <strong>in</strong> an o<strong>the</strong>rwise appeal<strong>in</strong>g<br />
approach are: how consistent are our observations?<br />
How stable are biotopes with stage? How do biotopes<br />
relate to channel morphology and biology?<br />
Biotope mapp<strong>in</strong>g has been supported by ecological<br />
observations that species patch<strong>in</strong>ess (superimposed on a<br />
general downstream cont<strong>in</strong>uum) does occur, and that this<br />
may fur<strong>the</strong>r be related to discharge exceedence values<br />
(Newson and Newson, 2000). This suggests that <strong>the</strong>re<br />
should be a l<strong>in</strong>k with similar patch<strong>in</strong>ess of functional<br />
habitat, if this, too, can be defi ned. Functional habits may<br />
be discrim<strong>in</strong>ated on <strong>the</strong> basis of comb<strong>in</strong>ed fl ow <strong>in</strong>dices<br />
such as <strong>the</strong> Froude number. (The Froude number is <strong>the</strong><br />
ratio of depth to <strong>the</strong> square root of depth multiplied by<br />
gravitational acceleration). From a survey of 32 sites <strong>in</strong><br />
eastern England, Kemp et al. (2000) suggest that functional<br />
habits fall <strong>in</strong>to two groups: lowest Froude number<br />
classes, and those with Froude number above 0.5. Insofar<br />
as <strong>the</strong> Froude number itself is assumed to refl ect biotopes<br />
derived from measurement or observation, <strong>the</strong>n a connection<br />
between biotopes and functional habitats should exist,<br />
but <strong>the</strong> ecological signifi cance of this l<strong>in</strong>k is uncerta<strong>in</strong> <strong>in</strong><br />
<strong>the</strong> absence of simultaneous, po<strong>in</strong>t-by-po<strong>in</strong>t measures of<br />
physical and ecological <strong>in</strong>dicators, and because of feedbacks<br />
or <strong>in</strong>terdependence between <strong>the</strong> apparently <strong>in</strong>dependent<br />
classifi cation variables. For example, vegetation<br />
growth affects velocities and sedimentation, and hence<br />
helps determ<strong>in</strong>e <strong>the</strong> local Froude number, which is supposedly<br />
<strong>in</strong>dependent. A recent <strong>in</strong>vestigation by Clifford<br />
et al. (2006) identifi es <strong>the</strong> need for more consistent use of<br />
bitope and habitat defi nitions, and improved experimental<br />
design when assess<strong>in</strong>g <strong>the</strong> l<strong>in</strong>kage between fl ows, habitats<br />
and ecology. Use of <strong>the</strong> Froude number may also be questioned<br />
because it may obscure differences between very<br />
different fl ow–depth comb<strong>in</strong>ations.<br />
Figure 7.8 illustrates ano<strong>the</strong>r basic area of uncerta<strong>in</strong>ty<br />
<strong>in</strong> <strong>the</strong> biotope and physical habitat approach, but also<br />
perhaps one way <strong>in</strong> which this uncerta<strong>in</strong>ty might be<br />
managed or, <strong>in</strong> future, reduced. The fi gures show reaches<br />
of <strong>the</strong> <strong>River</strong> Cole, near Birm<strong>in</strong>gham, UK, and <strong>the</strong> <strong>River</strong><br />
Tern, Shropshire, UK, which have been ‘classifi ed’ <strong>in</strong>to<br />
zones of statistically similar velocity and/or depth behaviour<br />
at various fl ow stages (Emery et al., 2003). Both<br />
rivers have a well-marked riffl e-pool bedform sequence (a<br />
common biotope), but <strong>the</strong> Cole is a less s<strong>in</strong>uous channel,<br />
with more regular bank morphology and less bankside tree<br />
growth. The approach uses cluster analysis (a hierarchical<br />
statistical method of association) as a reproducible means<br />
of assess<strong>in</strong>g both <strong>the</strong> degree, location and persistence of<br />
cluster<strong>in</strong>g or patch<strong>in</strong>ess of physical habitat, which might<br />
support more objective identifi cation of biotopes.<br />
Cluster analysis is an agglomerative process to identify<br />
homogeneous group<strong>in</strong>gs (patches) based upon selected<br />
characteristics. In this case, standardised fi eld velocity<br />
(measured velocity scaled accord<strong>in</strong>g to <strong>the</strong> mean and variance<br />
of <strong>the</strong> total measured distribution) was used. Initially,<br />
all velocity observations were considered as separate, but<br />
were <strong>the</strong>n successively comb<strong>in</strong>ed accord<strong>in</strong>g to ‘distance’<br />
(i.e. differences) from neighbours and from emerg<strong>in</strong>g<br />
clusters. The next stage was to assess cluster number <strong>in</strong><br />
relation to morphology, spatial coverage and stability.<br />
ANOVA (analysis of variance test<strong>in</strong>g) was used to test<br />
differ<strong>in</strong>g cluster numbers for statistically signifi cant differences<br />
at each fl ow stage and to assess <strong>the</strong>ir coverage<br />
and coherence as fl ow stage varied. The cluster analysis is<br />
<strong>the</strong>refore a means of defi n<strong>in</strong>g biotopes on <strong>the</strong> basis of fi eld<br />
measurements, ra<strong>the</strong>r than on <strong>the</strong> basis of visual survey.<br />
Figures 7.8(a) and (b) show <strong>the</strong> spatial coverage and location<br />
of <strong>the</strong> six dist<strong>in</strong>ct velocity clusters at higher fl ow stage<br />
result<strong>in</strong>g from <strong>the</strong> analysis, while Figures 7.8(c) and (d)<br />
detail <strong>the</strong> velocity characteristics associated with each<br />
cluster over a range of fl ow changes from low fl ow to a<br />
high <strong>in</strong>-bank fl ow.<br />
Several po<strong>in</strong>ts emerge from this analysis which <strong>in</strong>dicate<br />
that <strong>the</strong> method might help both understand and, <strong>in</strong> time,<br />
reduce uncerta<strong>in</strong>ties <strong>in</strong> habitat recognition and appraisal.<br />
The method is clearly sensitive to bedform amplitude and<br />
planform complexity. Even where <strong>the</strong> basic biotope structure<br />
of rivers is similar, <strong>the</strong> differ<strong>in</strong>g patterns of coverage<br />
are suffi cient to <strong>in</strong>dicate a degree of ‘biotope subtlety’ –<br />
contrast Figures 7.8(a), where fl ow organisation at high<br />
stage is marked by l<strong>in</strong>earity, and 7.8(b), where patch<strong>in</strong>ess<br />
is dom<strong>in</strong>ant. This is supported by <strong>the</strong> stage-dependent
North<strong>in</strong>g (m)<br />
Velocity (m/s)<br />
1060<br />
1040<br />
1020<br />
1000<br />
980<br />
960<br />
940<br />
920<br />
Hydrological and Hydraulic Aspects of <strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 119<br />
Cluster1<br />
Cluster2<br />
Cluster3<br />
Cluster4<br />
Cluster5<br />
Cluster6<br />
North<strong>in</strong>g (m)<br />
1070<br />
1060<br />
1050<br />
1040<br />
1030<br />
1020<br />
Cluster1<br />
Cluster2<br />
Cluster3<br />
Cluster4<br />
Cluster5<br />
Cluster6<br />
900<br />
4990 5000 5010 5020<br />
East<strong>in</strong>g (m)<br />
5030 5040<br />
1010<br />
4985 4990 4995<br />
East<strong>in</strong>g (m)<br />
5000 5005<br />
(a) (b)<br />
1<br />
0.8<br />
0.9<br />
0.8<br />
0.7<br />
0.7<br />
0.6<br />
0.6<br />
0.5<br />
0.5<br />
0.4<br />
0.4<br />
0.3<br />
0.3<br />
0.2<br />
0.2<br />
0.1<br />
0.1<br />
0<br />
Low ‘x’ V mean Med ‘x’ V mean High ‘x’ V mean<br />
0<br />
Low ‘x’ V mean Med ‘x’ V mean High ‘x’ V mean<br />
Stage class Stage class<br />
(c) (d)<br />
Velocity (m/s)<br />
Figure 7.8 Results of cluster analysis to determ<strong>in</strong>e coverage and behaviour of velocity ‘patches’ <strong>in</strong> <strong>the</strong> <strong>River</strong>s Cole, Birm<strong>in</strong>gham (a<br />
and c) and Tern, Shropshire (b and d) (Source: Modifi ed from Emery, 2003)<br />
velocity behaviour seen <strong>in</strong> Figures 7.8(c) and (d). In <strong>the</strong><br />
<strong>River</strong> Cole, <strong>the</strong> straighter, simpler channel, patch coherence<br />
<strong>in</strong>creases and cluster number decreases (coverage<br />
<strong>in</strong>creases) as fl ow stage rises. Thus, <strong>the</strong> river ‘simplifi es’<br />
with <strong>in</strong>creas<strong>in</strong>g fl ow stage <strong>in</strong>to three basic group<strong>in</strong>gs:<br />
channel marg<strong>in</strong>s, channel centrel<strong>in</strong>e, and areas above<br />
bedform crests. However, <strong>in</strong> <strong>the</strong> <strong>River</strong> Tern, <strong>the</strong> lower<br />
stage complexity is largely ma<strong>in</strong>ta<strong>in</strong>ed as fl ow stage rises,<br />
possibly because <strong>the</strong> fl ow now <strong>in</strong>teracts with large bankside<br />
tree roots creat<strong>in</strong>g new or ma<strong>in</strong>ta<strong>in</strong><strong>in</strong>g old, biotopes<br />
(for fur<strong>the</strong>r discussion, see Emery et al., 2003).<br />
The o<strong>the</strong>r key po<strong>in</strong>ts emerg<strong>in</strong>g from this analysis are<br />
equally important and draw attention to some pitfalls <strong>in</strong><br />
identifi cation and representation, which might improve<br />
survey methods <strong>in</strong> <strong>the</strong> future. First, <strong>in</strong> assess<strong>in</strong>g biotopes<br />
and functional habitats, <strong>the</strong> entire fl ow regime should be<br />
considered – what exists at low fl ow may or may not<br />
change <strong>in</strong> character as stage rises. Second, <strong>the</strong>re is an<br />
obvious role for hydraulic and hydrodynamic fl ow modell<strong>in</strong>g.<br />
This might support fi eld observations (which may be<br />
limited to s<strong>in</strong>gle fl ows) and assist <strong>in</strong> restoration design and<br />
appraisal, where <strong>the</strong> ability to predict and to visualise
120 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
chang<strong>in</strong>g aspects of fl ow and habitat performance as fl ow<br />
stage varies is crucial.<br />
7.4.3 Hydraulic Habitat Simulation and Modell<strong>in</strong>g<br />
Assessment of river fl ow management options, <strong>in</strong>clud<strong>in</strong>g<br />
river restoration and changes <strong>in</strong> fl ow regime, often <strong>in</strong>volves<br />
assess<strong>in</strong>g scenarios that fall outside <strong>the</strong> range of observed<br />
conditions. This negates <strong>in</strong>vestigation us<strong>in</strong>g direct analysis<br />
of fi eld observations and necessitates <strong>the</strong> use of predictive<br />
models. Models employ a variety of hydrological, morphological<br />
and hydraulic parameters to predict variations,<br />
which might <strong>the</strong>n be related to <strong>the</strong> abundance and distribution<br />
of aquatic organisms. The parameters vary accord<strong>in</strong>g<br />
to <strong>the</strong> spatial scale of <strong>the</strong> simulation. Habitat modell<strong>in</strong>g<br />
frequently <strong>in</strong>volves two steps: <strong>the</strong> physical analysis and<br />
evaluation of <strong>the</strong> hydraulic and morphological aspects of<br />
<strong>the</strong> problem, and <strong>the</strong> subsequent l<strong>in</strong>kage to ecological<br />
response.<br />
Uncerta<strong>in</strong>ties <strong>in</strong> <strong>the</strong> Hydraulic and Hydrodynamic<br />
Modell<strong>in</strong>g Process: <strong>the</strong> Fundamentals<br />
All hydraulic and hydrodynamic fl ow simulation models<br />
beg<strong>in</strong> with equations for <strong>the</strong> conservation of mass and<br />
momentum, which are modifi ed for <strong>the</strong> bed and surface<br />
boundary conditions found <strong>in</strong> shallow, open channel fl ow.<br />
As a result of limitations imposed by <strong>the</strong> scale of <strong>the</strong><br />
smaller turbulent motions, <strong>the</strong> govern<strong>in</strong>g equations are not<br />
tractable <strong>in</strong> <strong>the</strong>ir orig<strong>in</strong>al form and require additional<br />
modifi cations or ‘closure models’ to provide computational<br />
outputs. These assumptions <strong>in</strong> turn give rise to new<br />
terms <strong>in</strong> <strong>the</strong> govern<strong>in</strong>g equations, which are <strong>the</strong>mselves<br />
subsequently manipulated to satisfy <strong>the</strong> requirements of<br />
<strong>the</strong> solution, which may be represented as a one-, two- or<br />
three-dimensional numerical scheme (for full review, see<br />
Lane, 1998). Generally, <strong>in</strong> eco-hydraulic and river restoration<br />
applications, <strong>the</strong> uncerta<strong>in</strong>ties <strong>in</strong> model derivation are<br />
secondary to those of model application to <strong>the</strong> particular<br />
case study. In effect, modell<strong>in</strong>g is really a recursive<br />
process, <strong>in</strong>volv<strong>in</strong>g uncerta<strong>in</strong>ties at all stages from premodel<br />
data collection, through model verifi cation dur<strong>in</strong>g<br />
<strong>the</strong> numerical calculation, to post-modell<strong>in</strong>g validation,<br />
where results are compared with fi eld measurements and<br />
o<strong>the</strong>r sources of evidence/expertise to assess, or appraise<br />
<strong>the</strong> model performance.<br />
Traditionally, hydraulic modell<strong>in</strong>g has been focused on<br />
one-dimensional representation of open channel fl ow<br />
(Chow, 1973), <strong>in</strong> which <strong>the</strong> cross-sectionally averaged<br />
depth and velocity are obta<strong>in</strong>ed. These models may be<br />
used to simulate <strong>the</strong> passage of fl oods through reaches and<br />
channel systems, or for <strong>the</strong> assessment of channel stability<br />
given particular boundary characteristics (see Chapter 5<br />
for more details). In ecological applications, <strong>the</strong>y may be<br />
of use to determ<strong>in</strong>e depth and tim<strong>in</strong>g of <strong>in</strong>undation associated<br />
with particular events. Hydrographs are routed<br />
through cross-sections ei<strong>the</strong>r by runn<strong>in</strong>g a series of steadystate<br />
solutions for different discharge ‘b<strong>in</strong>s’ <strong>in</strong> <strong>the</strong> hydrograph<br />
or us<strong>in</strong>g unsteady methods to model <strong>the</strong> hydrograph.<br />
Common one-dimensional implementations of this<br />
sort are iSIS, HEC–RAS and MIKE11, while <strong>the</strong> most<br />
popular hydraulically-coupled habitat suitability model –<br />
PHABSIM (Spence and Hickley, 2000, and below) – uses<br />
one-dimenstional approaches to predict velocity and depth<br />
<strong>in</strong> channel cells or slices, defi ned between adjacent measured<br />
cross-sections. In <strong>the</strong>ir simplest forms, <strong>the</strong>se models<br />
rely on <strong>the</strong> ability to specify <strong>the</strong> relationship between<br />
depth, water surface slope and velocity via an empirical<br />
‘constant’ known as Mann<strong>in</strong>g’s n, which is normally<br />
assumed to represent <strong>the</strong> fl ow resistance aris<strong>in</strong>g from<br />
boundary friction of <strong>the</strong> channel. The value of n is usually<br />
determ<strong>in</strong>ed from look-up tables (e.g. Chow, 1973). Coeffi<br />
cients of channel expansion and contraction may be used<br />
<strong>in</strong> addition to channel geometry to account for additional<br />
components of energy loss aris<strong>in</strong>g from fl ow acceleration<br />
and deceleration. Away from channels of very simple<br />
geometry and boundary characteristics, <strong>the</strong> appropriate<br />
determ<strong>in</strong>ation of n is more subtle: to ma<strong>in</strong>ta<strong>in</strong> predictive<br />
success, for example, n is varied <strong>in</strong>versely with depth and,<br />
<strong>in</strong> many situations, n is really a compound calibration<br />
factor express<strong>in</strong>g (with more-or-less success) <strong>the</strong> various<br />
contributions to fl ow resistance, and hence energy loss <strong>in</strong><br />
channels: gra<strong>in</strong>, bedform, and spill (channel curvature).<br />
Some of <strong>the</strong> issues <strong>in</strong> more complex modell<strong>in</strong>g are outl<strong>in</strong>ed<br />
below.<br />
Two- and three-dimensional Numerical Schemes <strong>in</strong><br />
Ecological <strong>Restoration</strong> Applications<br />
Ideally, two- and three-dimensional approaches are required<br />
for realistic simulation of fl ow behaviour and to better represent<br />
<strong>the</strong> diversity and variability of physical habitats<br />
(Crowder and Diplas, 2000), thus overcom<strong>in</strong>g <strong>the</strong> limitations<br />
of popular habitat simulation models. In particular,<br />
account<strong>in</strong>g for small scale stage-dependent variations <strong>in</strong><br />
fl ow is important <strong>in</strong> <strong>the</strong> ma<strong>in</strong>tenance and survival of <strong>the</strong><br />
biotic community. In this way, more complex numerical<br />
fl ow simulation offers potential as a ‘support<strong>in</strong>g’ function<br />
for river rehabilitation schemes. Two- and threedimensional<br />
codes are now widely accessible and can be<br />
run over a wide range of discharge conditions, <strong>in</strong>clud<strong>in</strong>g<br />
over-bank fl ows, but <strong>the</strong>re are numerous sources of uncerta<strong>in</strong>ty<br />
at all stages of <strong>the</strong> modell<strong>in</strong>g process associated with<br />
<strong>the</strong>se simulations. Models <strong>in</strong>volve many underly<strong>in</strong>g
Hydrological and Hydraulic Aspects of <strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 121<br />
assumptions at <strong>the</strong>ir <strong>in</strong>put stage, <strong>the</strong>y are sensitive to <strong>the</strong><br />
numerical structures <strong>in</strong>volved <strong>in</strong> obta<strong>in</strong><strong>in</strong>g <strong>the</strong> fl ow solution<br />
and outputs may be diffi cult to observe or measure <strong>in</strong><br />
<strong>the</strong> fi eld. Some of <strong>the</strong> more common models, model applications<br />
and model uncerta<strong>in</strong>ties are listed <strong>in</strong> Table 7.3.<br />
Perhaps <strong>the</strong> most basic but much neglected area of<br />
uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong> application of more complex models<br />
rests with <strong>the</strong> fi eld data requirements for model calibration<br />
and validation. This is most onerous where modell<strong>in</strong>g is<br />
used at a spatial resolution commensurate with sub-reach<br />
detail <strong>in</strong> habitat <strong>in</strong> conditions of complex topography<br />
(Crowder and Diplas, 2000). These have led some to argue<br />
that <strong>the</strong> ecological and hydraulic/hydrodynamic approaches<br />
should proceed <strong>in</strong>dependently, because river environments<br />
are ‘too complex’ for a coupled approach (Kondolf et al.,<br />
2000). Field observations <strong>in</strong>dicate that physical habitat<br />
may be structured at an extremely small scale, such as<br />
around boulders, or extremely close to channel banks<br />
(Railsback et al., 1999), which necessitate huge <strong>in</strong>creases<br />
<strong>in</strong> model effort and computational time, and whose characteristics<br />
may not, <strong>in</strong> any case, be picked up by conventional<br />
fi eld survey. Figure 7.9 for example, illustrates how<br />
our knowledge of <strong>the</strong> character of <strong>the</strong> river changes as <strong>the</strong><br />
number and density of fi eld measurements at various fl ow<br />
stages is <strong>in</strong>creased or decreased.<br />
Here, <strong>the</strong> results of an orig<strong>in</strong>al fi eld survey of more than<br />
350 data po<strong>in</strong>ts obta<strong>in</strong>ed by survey through a 150 m river<br />
reach of <strong>the</strong> <strong>River</strong> Cole (represent<strong>in</strong>g both longitud<strong>in</strong>al<br />
and cross section variation with an average po<strong>in</strong>t-to-po<strong>in</strong>t<br />
spac<strong>in</strong>g of 1–2 m) have been plotted as a frequency histogram<br />
of velocities, and <strong>the</strong>n subsequently re-plotted <strong>in</strong> <strong>the</strong><br />
same way but with <strong>the</strong> orig<strong>in</strong>al dataset systematically<br />
degraded by a factor of one half or one third. The resultant<br />
changes thus give an idea of <strong>the</strong> effects of vary<strong>in</strong>g <strong>the</strong> fi eld<br />
data collection efforts or schemes <strong>in</strong> characteris<strong>in</strong>g ‘true’<br />
fi eld velocities, both prior to modell<strong>in</strong>g simulation and <strong>in</strong><br />
post-modell<strong>in</strong>g assessment. Several po<strong>in</strong>ts are worthy of<br />
note from this analysis. Firstly, channel velocity characteristics<br />
differ depend<strong>in</strong>g upon fl ow stage: at lower fl ow<br />
stage, results are more skewed than at higher stage, but at<br />
<strong>the</strong> higher stage <strong>the</strong>re is <strong>the</strong> suggestion of multiple modes<br />
<strong>in</strong> an o<strong>the</strong>rwise more even velocity distribution. This<br />
change <strong>in</strong> velocity distribution should be <strong>in</strong>terpreted <strong>in</strong><br />
conjunction with <strong>the</strong> results of <strong>the</strong> biotope analysis <strong>in</strong><br />
Section 7.4.2, s<strong>in</strong>ce <strong>the</strong> fi eld data are common to both. In<br />
this case, it seems that, as biotopes ‘clarify’, <strong>the</strong> distribution<br />
of velocities also becomes organised about dom<strong>in</strong>ant<br />
peaks, refl ect<strong>in</strong>g <strong>the</strong> <strong>in</strong>fl uence of channel marg<strong>in</strong>s, channel<br />
centrel<strong>in</strong>es and shallows over bedform crests. The second<br />
po<strong>in</strong>t to note is that sub-sampl<strong>in</strong>g from <strong>the</strong> orig<strong>in</strong>al data<br />
Table 7.3 Common numerical fl ow models, <strong>the</strong>ir pr<strong>in</strong>cipal applications <strong>in</strong> river restoration and eco-hydraulics, and some issues<br />
relat<strong>in</strong>g to parameterisation and model <strong>in</strong>terpretation<br />
Class of model<br />
and common<br />
implementation<br />
One-dimensionnal<br />
HEC-RAS<br />
ISIS<br />
backwater<br />
Two-dimensionnal<br />
Telemac<br />
RMA<br />
Three-dimensionnal<br />
SSIIM<br />
Fluent<br />
CFX<br />
Pr<strong>in</strong>cipal application <strong>in</strong><br />
channel restoration<br />
fl ood rout<strong>in</strong>g and channel<br />
conveyance<br />
<strong>in</strong>itial water surface<br />
elevation (<strong>in</strong> higher order<br />
models); velocity and<br />
depth between cross<br />
sections <strong>in</strong> PHABSIM<br />
dynamic simulation of<br />
channel fl ow and<br />
fl oodpla<strong>in</strong> <strong>in</strong>undation;<br />
cross-sectional patterns<br />
of habitat suitability<br />
detailed sub-reach<br />
modell<strong>in</strong>g of habitat<br />
suitability<br />
Required parameters Pr<strong>in</strong>cipal aspects of uncerta<strong>in</strong>ty<br />
downstream discharge;<br />
upstream water level;<br />
Mann<strong>in</strong>g’s n; channel<br />
expansion and contraction<br />
coeffi cients; basic channel<br />
cross section morphology<br />
boundary roughness (ks);<br />
fl ow <strong>in</strong>formation as above;<br />
bed topography; turbulence<br />
closure model; choice of<br />
numerical solver and<br />
relaxation coeffi cients;<br />
correct estimation of n (particularly with<br />
depth and vegetation changes) and<br />
channel expansion and contraction<br />
coeffi cients; loss of <strong>in</strong>formation on<br />
cross-sectional fl ow distribution; no<br />
account of channel curvature<br />
loss of <strong>in</strong>formation on depth-related fl ow<br />
properties; time-stepp<strong>in</strong>g <strong>in</strong> unsteady<br />
solutions; poor representation of<br />
secondary fl ow from channel<br />
discont<strong>in</strong>uities; mesh<strong>in</strong>g issues;<br />
representation of wett<strong>in</strong>g and dry<strong>in</strong>g<br />
as above most of <strong>the</strong> above, plus: meshdependence<br />
<strong>in</strong> <strong>the</strong> vertical as well as<br />
cross-section; determ<strong>in</strong>ation of <strong>the</strong> free<br />
water surface
122 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Frequency (%)<br />
Frequency (%)<br />
30<br />
25<br />
20<br />
15<br />
10<br />
5<br />
0<br />
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4<br />
30<br />
25<br />
20<br />
15<br />
10<br />
5<br />
Resultant velocity (m 3 s –1 )<br />
(a)<br />
full data set<br />
half data set<br />
third data set<br />
0<br />
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4<br />
Resultant velocity (m 3 s –1 )<br />
(b)<br />
full data set<br />
half data set<br />
third data set<br />
Figure 7.9 Effect of degrad<strong>in</strong>g <strong>the</strong> velocity sample on <strong>the</strong> distribution of depth average velocities: a) low fl ow; b) high <strong>in</strong>-bank<br />
fl ow
Hydrological and Hydraulic Aspects of <strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 123<br />
does not have an even effect <strong>in</strong> all velocity classes. This<br />
may or may not be of signifi cance but is likely to impart<br />
uncerta<strong>in</strong>ty <strong>in</strong>to <strong>the</strong> assessment of model performance if<br />
<strong>the</strong> affected velocity classes are those judged to be of most<br />
relevance for habitat suitability. This arises because chang<strong>in</strong>g<br />
<strong>the</strong> frequency of particular velocity class observations<br />
is equivalent to under- or over-represent<strong>in</strong>g <strong>the</strong> spatial<br />
coverage associated with that velocity. For fur<strong>the</strong>r discussion,<br />
see Clifford et al. (2002a, 2002b).<br />
Once acceptable fi eld velocity data have been obta<strong>in</strong>ed,<br />
<strong>the</strong>se must be coupled with representations of boundary<br />
topography and channel geometry. This should be suffi -<br />
ciently detailed to <strong>in</strong>corporate both small-scale irregularities<br />
<strong>in</strong> bed and banks, <strong>the</strong> major breaks of slope associated<br />
with channel marg<strong>in</strong>s and more subtle fl oodpla<strong>in</strong> topography.<br />
Breakl<strong>in</strong>es are notoriously diffi cult to handle, and if<br />
fi ne-scale survey is unavailable <strong>the</strong>n topography has to be<br />
<strong>in</strong>terpolated, <strong>in</strong>troduc<strong>in</strong>g a fur<strong>the</strong>r source of uncerta<strong>in</strong>ty,<br />
which also occurs <strong>in</strong> areas of low relief, too. Simple l<strong>in</strong>ear<br />
<strong>in</strong>terpolation, for example, may be preferable to avoid<br />
spurious high- and low-po<strong>in</strong>ts (French and Clifford, 2000).<br />
Topography forms <strong>the</strong> template of <strong>the</strong> modell<strong>in</strong>g grid or<br />
mesh which is draped over <strong>the</strong> boundary surface and which<br />
forms <strong>the</strong> basis for <strong>the</strong> numerical solution to <strong>the</strong> govern<strong>in</strong>g<br />
fl ow equations. The form of <strong>the</strong> mesh depends upon <strong>the</strong><br />
k<strong>in</strong>d of model scheme adopted. Meshes might exploit triangular<br />
(<strong>in</strong> fi nite element schemes) or rectangular elements<br />
(<strong>in</strong> fi nite difference and fi nite volume schemes), of<br />
ei<strong>the</strong>r constant or variable size. Triangular meshes are<br />
more versatile, but most ecological applications use rectangular<br />
grid elements and structured grid schemes. While<br />
allow<strong>in</strong>g elements to vary <strong>in</strong> size, <strong>the</strong>se always require <strong>the</strong><br />
same number of elements <strong>in</strong> each cross plane of solution.<br />
In practice, <strong>the</strong> constra<strong>in</strong>ts of gridd<strong>in</strong>g normally require<br />
some degradation <strong>in</strong> <strong>the</strong> quality of <strong>the</strong> topographic representation,<br />
s<strong>in</strong>ce multiply<strong>in</strong>g <strong>the</strong> number of very small elements<br />
to represent more detailed topographic variation<br />
imparts a disproportionately large <strong>in</strong>crement <strong>in</strong> computer<br />
time required to solve <strong>the</strong> problem and may result <strong>in</strong><br />
<strong>in</strong>stabilities <strong>in</strong> <strong>the</strong> solution. Many of <strong>the</strong>se issues receive<br />
a thorough exploration <strong>in</strong> Anderson et al. (1996). As computer<br />
power has <strong>in</strong>creased, more detailed modell<strong>in</strong>g is<br />
possible, but approach<strong>in</strong>g closer representations of reality<br />
does not simply depend upon enhanced computer speed<br />
or memory. A key issue is that, with smaller elements, bed<br />
gra<strong>in</strong>s and micro-topography may approach or exceed <strong>the</strong><br />
size of model elements, render<strong>in</strong>g solution impossible.<br />
This is especially pert<strong>in</strong>ent <strong>in</strong> two-dimensional modell<strong>in</strong>g,<br />
where <strong>the</strong> number and size of elements <strong>in</strong> <strong>the</strong> vertical may<br />
also be varied, and <strong>the</strong> issue is related to <strong>the</strong> wider question<br />
of <strong>the</strong> appropriate parameterisation of multiple scales of<br />
fl ow resistance (Nicholas, 2001).<br />
With respect to fl ow resistance, most two-dimensional<br />
models often rema<strong>in</strong> reliant on a boundary resistance<br />
value specifi ed by n or by Chezy’s conveyance coeffi cient,<br />
C. Three-dimensional models employ a boundary roughness<br />
specifi ed <strong>in</strong> terms of an equivalent roughness length,<br />
k s, which propagates through <strong>the</strong> boundary cells. While ks<br />
can be determ<strong>in</strong>ed <strong>in</strong> relation to <strong>the</strong> boundary gra<strong>in</strong> size,<br />
<strong>the</strong> numerical value required to obta<strong>in</strong> plausible solutions<br />
is generally much larger than gra<strong>in</strong> size alone would imply,<br />
giv<strong>in</strong>g rise to uncerta<strong>in</strong>ty <strong>in</strong> its <strong>in</strong>terpretation. Flow resistance<br />
effects beyond sk<strong>in</strong> friction may be <strong>in</strong>corporated<br />
witt<strong>in</strong>gly or unwitt<strong>in</strong>gly <strong>in</strong>to <strong>the</strong> mesh<strong>in</strong>g strategy – for<br />
example, by vary<strong>in</strong>g <strong>the</strong> size of cells around <strong>the</strong> boundary<br />
as compared to <strong>the</strong> ma<strong>in</strong> body of <strong>the</strong> fl ow (Clifford et al.,<br />
2002a). O<strong>the</strong>r areas of uncerta<strong>in</strong>ty <strong>in</strong> both model application<br />
and result <strong>in</strong>terpretation <strong>in</strong>clude: whe<strong>the</strong>r <strong>the</strong> model<br />
uses a fi xed water surface (lid) or allows <strong>the</strong> water<br />
surface to freely adjust; <strong>the</strong> scheme used to allow <strong>the</strong><br />
simulation of turbulence effects (<strong>the</strong> closure assumptions);<br />
and <strong>the</strong> complexity of <strong>the</strong> solver (which determ<strong>in</strong>es how<br />
many and which neighbour<strong>in</strong>g po<strong>in</strong>ts are considered <strong>in</strong> <strong>the</strong><br />
propagation of <strong>the</strong> numerical solution through <strong>the</strong> mesh).<br />
Many of <strong>the</strong>se issues are common to both two- and threedimensional<br />
schemes, although similar strategies and<br />
choices will not necessarily have <strong>the</strong> same effects <strong>in</strong> <strong>the</strong><br />
differ<strong>in</strong>g model implementations. Nicholas (2001) provides<br />
a comprehensive start<strong>in</strong>g po<strong>in</strong>t for assess<strong>in</strong>g and<br />
manag<strong>in</strong>g <strong>the</strong>se uncerta<strong>in</strong>ties. The particular limitation of<br />
two-dimentional models is <strong>the</strong> parameterisation of secondary<br />
circulation effects on momentum transport, which<br />
has concentrated on <strong>the</strong> effects of channel curvature, but<br />
not on topographic discont<strong>in</strong>uities and river confl uences<br />
(Lane, 1998).<br />
Rarely do <strong>the</strong> uncerta<strong>in</strong>ties aris<strong>in</strong>g with respect to each<br />
aspect of <strong>the</strong> modell<strong>in</strong>g occur <strong>in</strong>dependently: Figure 7.10<br />
for example, illustrates <strong>the</strong> chang<strong>in</strong>g fl ow representations<br />
aris<strong>in</strong>g from coarser- and fi ner-scale element sizes us<strong>in</strong>g<br />
<strong>the</strong> SSIMM model of <strong>the</strong> <strong>River</strong> Cole, near Birm<strong>in</strong>gham,<br />
UK. Apart from <strong>the</strong> element size, all o<strong>the</strong>r model parameters<br />
are held constant. As <strong>the</strong> results demonstrate, modelled<br />
velocities vary <strong>in</strong> both planform and <strong>in</strong> cross-section<br />
and vertical distributions as grid element size is varied<br />
from approximately 2–0.5 m. The coarse grid distorts<br />
<strong>the</strong> channel shape from generally rectangular to a trapezoidal<br />
cross section and this contributes to bias<strong>in</strong>g <strong>the</strong> faster<br />
fl ow towards <strong>the</strong> channel centre, rais<strong>in</strong>g velocities above<br />
those measured <strong>in</strong> <strong>the</strong> fi eld. This effect is also enhanced<br />
by <strong>the</strong> <strong>in</strong>teraction of grid element size and boundary<br />
roughness, which <strong>in</strong> <strong>the</strong> model extends to 20% of <strong>the</strong> near<br />
boundary grid element. When elements are large, a larger<br />
proportion of <strong>the</strong> channel is thus affected by <strong>the</strong> boundary<br />
roughness, even though <strong>the</strong> numerical value of k s is <strong>the</strong>
124 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
HIGH flow<br />
coarse grid<br />
Pool<br />
Pool<br />
1.4<br />
1.3<br />
1.2<br />
1.1<br />
1<br />
0.9<br />
0.8<br />
0.7<br />
0.6<br />
0.5<br />
0.4<br />
0.3<br />
0.2<br />
0.1<br />
1.4<br />
1.3<br />
1.2<br />
1.1<br />
1<br />
0.9<br />
0.8<br />
0.7<br />
0.6<br />
0.5<br />
0.4<br />
0.3<br />
0.2<br />
0.1<br />
Modelled<br />
velocities<br />
Modelled<br />
velocities<br />
Pool<br />
HIGH flow<br />
f<strong>in</strong>e grid<br />
Pool<br />
Riffle crest<br />
Riffle crest<br />
(m)<br />
(m)<br />
1060<br />
1040<br />
1020<br />
1000<br />
980<br />
960<br />
940<br />
1060<br />
1040<br />
1020<br />
1000<br />
980<br />
960<br />
940<br />
N<br />
920<br />
4980 5000<br />
(m)<br />
5020 5040<br />
N<br />
PT 4<br />
Flow<br />
PT 4<br />
Flow<br />
PT 2<br />
PT 2<br />
RC 4<br />
PT 3<br />
RC 4<br />
RC 2<br />
PT 3<br />
RC 2<br />
RC 3<br />
RC 3<br />
PT 1<br />
RC 1<br />
PT 1<br />
RC 1<br />
0.3<br />
0.25<br />
0.2<br />
0.15<br />
0.1<br />
0.05<br />
0<br />
-0.05<br />
-0.1<br />
-0.15<br />
-0.2<br />
-0.25<br />
-0.3<br />
0.3<br />
0.25<br />
0.2<br />
0.15<br />
0.1<br />
0.05<br />
0<br />
-0.05<br />
-0.1<br />
-0.15<br />
-0.2<br />
-0.25<br />
-0.3<br />
920<br />
4980 5000<br />
(m)<br />
5020 5040<br />
Modelledfield<br />
velocities<br />
Modelledfield<br />
velocities<br />
Figure 7.10 (See also colour plate section) Chang<strong>in</strong>g SSIIM model output and model ‘fi t’ with as a function of mesh element size<br />
<strong>in</strong> a reach of <strong>the</strong> <strong>River</strong> Cole, near Birm<strong>in</strong>gham, UK
Hydrological and Hydraulic Aspects of <strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 125<br />
same. This effective <strong>in</strong>crease <strong>in</strong> friction comb<strong>in</strong>ed with <strong>the</strong><br />
reduced area of <strong>the</strong> channel <strong>in</strong> <strong>the</strong> coarser grid scheme<br />
‘push’ <strong>the</strong> fl ow to <strong>the</strong> centrel<strong>in</strong>e. In <strong>the</strong> streamwise direction,<br />
<strong>the</strong> coarser grid resolution is also <strong>in</strong>suffi cient to represent<br />
<strong>the</strong> ‘patch<strong>in</strong>ess’ of velocity seen <strong>in</strong> <strong>the</strong> fi ner grid<br />
case and, hence, both misses and distorts <strong>the</strong> spatial distribution<br />
of potential physical habitat. By contrast, use of<br />
<strong>the</strong> fi ner mesh over-emphasises velocity patch<strong>in</strong>ess and<br />
gives rise to problems of near-bank velocity representation.<br />
Here, apparently high velocity values cannot be corroborated<br />
with fi eld data which were collected at coarser<br />
resolution.<br />
Some models allow adaptive grids which vary <strong>in</strong> complexity<br />
as <strong>the</strong> problem evolves. The most fundamental<br />
illustration of this is <strong>the</strong> ability (or o<strong>the</strong>rwise) to cope with<br />
wett<strong>in</strong>g and dry<strong>in</strong>g, which will occur as fl ow stage rises<br />
and <strong>in</strong>undates fl oodpla<strong>in</strong> or channel marg<strong>in</strong>s, before reced<strong>in</strong>g.<br />
Full simulation of this requires an unsteady solution<br />
coupled to one of several schemes to ei<strong>the</strong>r <strong>in</strong>clude or<br />
exclude fully or partially dry elements. Dynamic solutions<br />
are generally confi ned to two-dimensional schemes. A<br />
simpler solution is to produce a number of grids, each of<br />
which is fully wet for <strong>the</strong> particular fl ow stage under consideration.<br />
In analogy with one-dimensional models, <strong>the</strong><br />
hydrograph may <strong>the</strong>n be simulated from a series of steady<br />
state solutions.<br />
Much of <strong>the</strong> uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong> modell<strong>in</strong>g process may<br />
be summarised by <strong>the</strong> terms verifi cation and validation.<br />
These are fundamental considerations <strong>in</strong> numerical fl ow<br />
modell<strong>in</strong>g but have recently been re-exam<strong>in</strong>ed as <strong>the</strong><br />
number of models and range of model application has<br />
burgeoned. Verifi cation relates to <strong>the</strong> ability to correctly<br />
solve <strong>the</strong> appropriate equations of <strong>the</strong> model problem at<br />
hand (i.e. whe<strong>the</strong>r <strong>the</strong> solution is an accurate solution of<br />
<strong>the</strong> particular choice of model); whereas validation relates<br />
to <strong>the</strong> plausibility of <strong>the</strong> model as a whole and <strong>the</strong> test<strong>in</strong>g<br />
of parameters predicted by <strong>the</strong> model (thus <strong>in</strong>volv<strong>in</strong>g<br />
<strong>the</strong> degree of fi t between prediction and measurement).<br />
As model applications have <strong>in</strong>creased, <strong>the</strong> dist<strong>in</strong>ctions<br />
between <strong>the</strong> two have become more blurred, and <strong>the</strong> term<br />
model assessment or appraisal has been used (Lane and<br />
Richards, 2001). Hardy et al. (2003) follow conventional<br />
eng<strong>in</strong>eer<strong>in</strong>g practice, argu<strong>in</strong>g that verifi cation is <strong>the</strong> essential<br />
element <strong>in</strong> <strong>the</strong> assessment process, which must precede<br />
attempts at validation. However, relatively little attention<br />
has been given to develop<strong>in</strong>g means of model assessment<br />
that are tailored to <strong>the</strong> application, particularly with respect<br />
to eco-hydraulics. In environmental systems, for example,<br />
<strong>the</strong> degree of experimental closure is much less than <strong>in</strong><br />
laboratory conditions or eng<strong>in</strong>eer<strong>in</strong>g problems, and <strong>the</strong><br />
application of models is as much (if not more) to produce<br />
data to support or <strong>in</strong>form new fi eld strategies or explana-<br />
tory <strong>in</strong>terpretations/<strong>in</strong>sights (Oreskes et al., 1994) as is it<br />
is to judge <strong>the</strong> correspondence between measured and<br />
modelled values. Thus, <strong>in</strong> ecological applications, where<br />
<strong>the</strong> uncerta<strong>in</strong>ties of measurements (both physical and biological)<br />
are coupled with an <strong>in</strong>tr<strong>in</strong>sic dynamism and variability,<br />
Clifford et al. (2005) have argued that assessment<br />
of models necessarily requires a more ‘relaxed’ approach.<br />
This is illustrated <strong>in</strong> Figure 7.11, where planform representations<br />
of model:fi eld correspondence of velocities for<br />
<strong>the</strong> <strong>River</strong> Cole are shown.<br />
In Figure 7.11(a), <strong>the</strong> fi t of <strong>the</strong> model is shown for a<br />
higher fl ow stage, purely as <strong>the</strong> absolute value of <strong>the</strong> po<strong>in</strong>tby-po<strong>in</strong>t<br />
modelled-fi eld velocity. On this basis, approximately<br />
80% of <strong>the</strong> channel area is modelled to with<strong>in</strong><br />
±0.2 ms −1 , and approximately 40% to with<strong>in</strong> ±0.1 ms −1<br />
(Figure 7.11(c)). Figure 7.11(b), however, shows <strong>the</strong><br />
improvement <strong>in</strong> fi t when a degree of ‘relaxation’ is applied.<br />
In this case, velocity ‘errors’ were recoded to zero<br />
provided that two criteria were met: fi rst, <strong>the</strong> modelled<br />
velocity must lie with<strong>in</strong> 0.1 ms −1 of <strong>the</strong> fi eld velocity and,<br />
second, this 0.1 ms −1 difference must also lie with<strong>in</strong> a 1 m<br />
radius of <strong>the</strong> actual fi eld measurement. By allow<strong>in</strong>g this<br />
relaxation, <strong>the</strong> fi t of <strong>the</strong> model is dramatically improved:<br />
reference to Figure 7.11(c) shows that only approximately<br />
5% of <strong>the</strong> channel fails to fulfi l <strong>the</strong>se jo<strong>in</strong>t criteria. The<br />
justifi cation for <strong>the</strong> relaxation lies <strong>in</strong> <strong>the</strong> fact that: (a) fi eld<br />
measurements are rarely precise enough to warrant absolute<br />
comparison to modelled results, particularly where<br />
both fi eld and modelled velocities have been <strong>in</strong>terpolated<br />
to <strong>the</strong> same grid for comparison; (b) <strong>in</strong> ecological applications,<br />
<strong>the</strong> uncerta<strong>in</strong>ties <strong>in</strong> relat<strong>in</strong>g physical parameters to<br />
ecological response are large; and (c) <strong>in</strong> any case plant and<br />
animal communities are mobile, dynamic communities,<br />
ra<strong>the</strong>r than static and adapt to exploit <strong>the</strong> most favourable<br />
environments. Ecological simulation conta<strong>in</strong>s, <strong>the</strong>n, an<br />
essential uncerta<strong>in</strong>ty, which might usefully be <strong>in</strong>corporated<br />
<strong>in</strong>to numerical model assessment so as to improve<br />
<strong>in</strong>terpretation!<br />
Ecological Uncerta<strong>in</strong>ties <strong>in</strong> Eco-hydraulic Modell<strong>in</strong>g<br />
Ecological uncerta<strong>in</strong>ties relate ma<strong>in</strong>ly to <strong>the</strong> simplifi cation,<br />
observation and test<strong>in</strong>g of habitat suitability criteria.<br />
Most frequently, <strong>the</strong>se have been characterised <strong>in</strong> terms of<br />
species’ preferences or abundances, result<strong>in</strong>g <strong>in</strong> Habitat<br />
Suitability Indices (HSIs), although more recent approaches<br />
seek l<strong>in</strong>kages to particular aspects of life cycle and growth<br />
of organisms or communities (see <strong>the</strong> later section on<br />
bioenergetics <strong>in</strong> this chapter). Preference criteria may be<br />
simple (univariate) or complex (multivariate) and may be<br />
related to <strong>the</strong> physically-determ<strong>in</strong>ed environment through<br />
correspondence (association) or through <strong>the</strong> application of
126 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
(m)<br />
Cumulative Area (%)<br />
100<br />
90<br />
80<br />
70<br />
60<br />
50<br />
40<br />
30<br />
20<br />
10<br />
0<br />
-1<br />
-0.8<br />
(a)<br />
PT1<br />
RC 1<br />
1.2<br />
1.1<br />
1<br />
0.9<br />
0.8<br />
0.7<br />
0.6<br />
0.5<br />
0.4<br />
0.3<br />
0.2<br />
0.1<br />
20 5040<br />
relaxed model<br />
orig<strong>in</strong>al model<br />
-0.6<br />
-0.4<br />
-0.2<br />
(m)<br />
1060<br />
1040<br />
1020<br />
1000<br />
980<br />
960<br />
940<br />
N<br />
PT4<br />
Flow<br />
PT 2<br />
0.3<br />
0.2<br />
0.1<br />
0<br />
-0.1<br />
-0.2<br />
-0.3<br />
920<br />
4980 5000 5020 5040<br />
(m)<br />
Modelled - Field Resultant Velocities (m s –1 )<br />
(c)<br />
orig<strong>in</strong>al model<br />
relaxed model<br />
0<br />
0.2<br />
0.4<br />
0.6<br />
(b)<br />
RC 4<br />
PT 3<br />
Model<br />
RC 2<br />
RC 3<br />
PT1<br />
Relaxed model<br />
Figure 7.11 Spatial (a and b) and areal (c) representations of model fi t to fi eld measurements for <strong>the</strong> <strong>River</strong> Cole, Birm<strong>in</strong>gham<br />
(Source: Based upon Figures 5 and 6 of Clifford et al., 2005)<br />
0.8<br />
RC 1<br />
1
Hydrological and Hydraulic Aspects of <strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 127<br />
‘rules’ which allow a degree of uncerta<strong>in</strong>ty as an essential<br />
element of <strong>the</strong> procedure. Once physical habitat and ecological<br />
response are obta<strong>in</strong>ed, results of <strong>the</strong> simulation are<br />
expressed <strong>in</strong> graphical or tabular form as values of <strong>the</strong><br />
output between limits or above thresholds, often with a<br />
spatial component, expressed as Weighted Useable Area<br />
(WUA). WUA is an aggregate measure of physical habitat<br />
quality and quantity and will be specifi c to a particular<br />
discharge and species/life stage. However, it is also clear<br />
that patterns or responses <strong>in</strong> ecology frequently result<br />
from persistent variability and renewal of opportunities,<br />
ra<strong>the</strong>r than <strong>the</strong> ‘static’ presence/absence of clear habitat<br />
delimiters. There have, <strong>the</strong>refore, been calls to explore <strong>the</strong><br />
capacity of eco-hydraulic modell<strong>in</strong>g to identify capacity,<br />
opportunity and variability ra<strong>the</strong>r than <strong>the</strong> prediction of<br />
exclusion via <strong>in</strong>dividual tolerance boundaries (Bergen<br />
et al., 2001, and Reynolds, 2002).<br />
In Europe, <strong>the</strong> Water Framework Directive will require<br />
river management at <strong>the</strong> catchment scale (Logan & Furse,<br />
2002) and this is be<strong>in</strong>g mirrored <strong>in</strong> national legislation<br />
and subsequent water management activity by organisations,<br />
such as <strong>the</strong> Environment Agency of England and<br />
Wales. Therefore, a major research question for habitat<br />
modell<strong>in</strong>g is whe<strong>the</strong>r exist<strong>in</strong>g reached-based methods<br />
can be scaled up, or whe<strong>the</strong>r an entirely new approach<br />
needs to be developed. Physical habitat modell<strong>in</strong>g <strong>in</strong>vestigations<br />
are typically confi ned to short lengths of river,<br />
approximately 50–200 m. A representative reach approach<br />
may be taken <strong>in</strong> which a reach is subjectively chosen<br />
to represent a longer length of river, <strong>in</strong>clud<strong>in</strong>g <strong>the</strong><br />
direct proportions of habitats with<strong>in</strong> that reach. A habitat<br />
mapp<strong>in</strong>g approach entails classify<strong>in</strong>g and record<strong>in</strong>g habitat<br />
types (pools, riffl es, glides etc) over long lengths of<br />
river and <strong>the</strong>n choos<strong>in</strong>g cross-sections to represent <strong>the</strong><br />
identifi ed habitat types. The results from each crosssection<br />
are <strong>the</strong>n weighted accord<strong>in</strong>g to <strong>the</strong> proportions<br />
of <strong>the</strong> identifi ed habitat types (Morhardt et al., 1983;<br />
Maddock, 1999).<br />
Progress towards catchment-scale modell<strong>in</strong>g has been<br />
made <strong>in</strong> associated fi elds such as fl ood modell<strong>in</strong>g, which<br />
provides broad-scale hydraulic predictions. This provides<br />
a potential basis for catchment-scale physical habitat<br />
assessment us<strong>in</strong>g <strong>in</strong>puts of discharge, water levels and<br />
channel geometry supplied from one-dimensional hydraulic<br />
models. Velocity variations across each cross-section<br />
can <strong>the</strong>n be predicted us<strong>in</strong>g disaggregation algorithms<br />
such as that used <strong>in</strong> <strong>the</strong> HEC–RAS model (US Army<br />
Corps of Eng<strong>in</strong>eers, 2002). Velocity and depth are used to<br />
assess habitat quality for target species. A major issue<br />
when apply<strong>in</strong>g this method is <strong>the</strong> loss of local detail when<br />
scal<strong>in</strong>g up. Booker et al. (2004a) describe <strong>the</strong> application<br />
of a method to <strong>the</strong> <strong>River</strong> Itchen, UK. This was able to<br />
predict physical habitat for an entire catchment under different<br />
water use strategy scenarios compared with present<br />
and naturalised situations.<br />
Some of <strong>the</strong> many (and complex) uncerta<strong>in</strong>ties <strong>in</strong> ecohydraulic<br />
modell<strong>in</strong>g are explored below <strong>in</strong> <strong>the</strong> context<br />
of <strong>the</strong> most commonly applied numerical scheme, <strong>the</strong><br />
PHABSIM model.<br />
7.4.4 The PHABSIM Model<br />
The fi rst step <strong>in</strong> formulat<strong>in</strong>g a habitat modell<strong>in</strong>g approach<br />
for rivers was published by Waters (1976). This led to <strong>the</strong><br />
more formal description of a computer model called<br />
PHABSIM (Physical Habitat Simulation) by <strong>the</strong> US Fish<br />
and Wildlife Service (Bovee, 1982; Bovee et al., 1998;<br />
Milhous, 1999). The PHABSIM system (Bovee, 1982;<br />
Bovee et al., 1998) is a suite of numerical models that<br />
allows quantifi cation of physical habitat for a given site,<br />
defi ned <strong>in</strong> terms of <strong>the</strong> comb<strong>in</strong>ation of depth, velocity<br />
and substrate/cover at a particular discharge (e.g. Johnson<br />
et al., 1993; Elliott et al., 1996). This system is most<br />
commonly used to assess <strong>the</strong> availability of suitable habitat<br />
for fi sh, although macro<strong>in</strong>vertebrates (Gore et al., 1998)<br />
and macrophytes (Hearne et al., 1994), which have measurable<br />
physical habitat requirements, have also been<br />
<strong>the</strong> focus of PHABSIM studies. In <strong>the</strong> United K<strong>in</strong>gdom<br />
<strong>the</strong> method has been used to assess changes <strong>in</strong> physical<br />
habitat associated with alterations <strong>in</strong> fl ow regime; for<br />
example, application to <strong>the</strong> <strong>River</strong>s Allen (Johnson et al.,<br />
1995), Piddle (Strevens, 1999) and Kennet (McPherson,<br />
1997) to aid management decisions based on <strong>the</strong> effects<br />
of abstraction by groundwater pump<strong>in</strong>g on habitat availability.<br />
The model has also been used to assess <strong>the</strong> impact<br />
of channel restoration (Acreman and Elliott, 1996) and<br />
difference <strong>in</strong> habitat caused by different levels of channel<br />
modifi cation (Booker & Dunbar, 2004; Booker et al.,<br />
2003). Over <strong>the</strong> years, <strong>the</strong> methodology used has been<br />
adapted by various researchers and <strong>in</strong>stitutions (Table 7.4).<br />
This has lead to <strong>the</strong> development of o<strong>the</strong>r models that<br />
follow basically <strong>the</strong> same approach (Parasiewicz and<br />
Dunbar, 2001).<br />
The approach adopted <strong>in</strong> many PHABSIM studies has<br />
been outl<strong>in</strong>ed by Elliott et al. (1999) and Johnson et al.<br />
(1995). This approach <strong>in</strong>cludes identifi cation of river<br />
sectors and species of <strong>in</strong>terest, identifi cation of habitats that<br />
exist with<strong>in</strong> <strong>the</strong> river length of <strong>in</strong>terest, selection of crosssections<br />
which represent replicates of each habitat type and<br />
collection of model calibration data (water surface elevation,<br />
depth and velocity). The distribution of depths and<br />
velocities are <strong>the</strong>n predicted for <strong>the</strong> range of discharge<br />
required. Predicted depths and velocities and measured<br />
substrate classes are <strong>the</strong>n compared with HSIs. This allows
128 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Table 7.4 Examples of adaptations of PHABSIM <strong>in</strong> different countries<br />
Country <strong>River</strong> Software name Adaptation Reference<br />
France Sa<strong>in</strong>t Sauveur brook EVHA French version of PHABSIM Rousel et al. (1999)<br />
Taiwan Chou-Shui Creek PHABSIM Incorporates affects of simulated<br />
substrate changes<br />
Wu & Wang (2002)<br />
Italy Adda PHABSIM Utilises bivariate HSIs Vismara et al. (2001)<br />
USA North Fork Middle PHABSIM Incorporates an <strong>in</strong>dividual based Van W<strong>in</strong>kle et al. (1998)<br />
Fork Tule <strong>River</strong><br />
model<br />
Canada Waterton PHABSIM Uses hydraulic output from a 2D<br />
model<br />
Ghanem et al. (1996)<br />
F<strong>in</strong>land <strong>River</strong> Oulujoki FISU Uses hydraulic output from a 2D<br />
model<br />
Yrjänä et al. (1999)<br />
France Rhone STATHAB Uses a statistical hydraulics model. Lamouroux et al. (1999)<br />
Norway Mandal SSIIM Uses hydraulic output from a 3D<br />
model<br />
Fjeldstad (2001)<br />
Switzerland Brenno CASIMIR Incorporates fuzzy rules for fi sh<br />
and <strong>in</strong>vertebrate HSIs<br />
http://www.greenhydro.ch<br />
USA Qu<strong>in</strong>nebaug Meso-HABSIM Broader scale, uses more habitat<br />
variables<br />
Parasiewicz (2001)<br />
Figure 7.12 Habitat suitability <strong>in</strong>dices (HSIs) calculated for juvenile salmon preference us<strong>in</strong>g pooled results from several rivers<br />
(Source: Data from Dunbar et al., 2001)<br />
prediction of usable physical habitat for <strong>the</strong> species/life<br />
stage of <strong>in</strong>terest, as WUA <strong>in</strong> m 2 per 1000 m of river<br />
channel.<br />
The form of HSIs used <strong>in</strong> PHABSIM applications will<br />
affect <strong>the</strong> results (Booker and Dunbar, 2004) and is <strong>the</strong>refore<br />
one source of uncerta<strong>in</strong>ty affect<strong>in</strong>g results. HSIs have<br />
been categorised <strong>in</strong>to several groups. These are:<br />
• Type I: HSIs that are derived from expert op<strong>in</strong>ion or<br />
through <strong>in</strong>formation published <strong>in</strong> literature.<br />
• Type II: HSIs that have been derived through frequency<br />
analysis of physical habitat conditions used by<br />
different species or life stages as identifi ed <strong>in</strong> fi eld<br />
observations.<br />
• Type III: Type II HSIs that have been corrected for<br />
habitat availability. These HSIs are referred to as preference<br />
curves.<br />
• Type IV: Multi-variant HSIs that weight habitat suitability<br />
based on depth and velocity toge<strong>the</strong>r.<br />
Site-specifi c HSIs can be developed for particular rivers.<br />
However, this is costly and time consum<strong>in</strong>g. HSIs that<br />
used pooled results from several rivers <strong>in</strong> an attempt to<br />
create generalised curves that can be transferred between<br />
rivers are shown <strong>in</strong> Figure 7.12. Belaud et al. (1989)<br />
reported similarities between four site-specifi c HSIs and<br />
generalised HSIs, suggest<strong>in</strong>g that generalised HSIs may<br />
be more useful. Roussel et al. (1999) added to <strong>the</strong> debate
Hydrological and Hydraulic Aspects of <strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 129<br />
on HSIs by suggested that different HSIs should be used<br />
for rest<strong>in</strong>g and feed<strong>in</strong>g fi sh. Vismara et al. (2001) compared<br />
univariate and bivariate HSIs, conclud<strong>in</strong>g that depth<br />
is much more important than velocity <strong>in</strong> defi n<strong>in</strong>g habitat<br />
suitability requirements when us<strong>in</strong>g bivariate models.<br />
Evaluation of PHABSIM<br />
The PHABSIM modell<strong>in</strong>g methodology may be applied<br />
before and after a restoration scheme to <strong>in</strong>fer changes <strong>in</strong><br />
habitat us<strong>in</strong>g a methodology such as Elliott et al. (1996).<br />
Alternatively, predicted results may be compared with<br />
changes <strong>in</strong> abundance of fi sh derived from surveys before<br />
and after <strong>the</strong> scheme. In both <strong>in</strong>stances, it is important to<br />
remember <strong>the</strong> simplicity of <strong>the</strong> PHABSIM approaches,<br />
both hydraulic and habitat modell<strong>in</strong>g. Successful application<br />
still requires considerable calibration and refi nement.<br />
Calibration may be based on empirical (regression) fi ts <strong>in</strong><br />
an effort to capture stage-dependent as well as spatial<br />
variations between vertical cross-section slices (Milhous<br />
et al., 1989). Alternatively, local and stage-dependent<br />
variations may be estimated from sparse velocity measurements<br />
at a s<strong>in</strong>gle stage by back-calculat<strong>in</strong>g of Mann<strong>in</strong>g’s<br />
n on a slice-by-slice basis, hav<strong>in</strong>g fur<strong>the</strong>r adjusted to<br />
match <strong>the</strong> modelled fl ow. In this case, however, <strong>the</strong> physical<br />
<strong>in</strong>tegrity of <strong>the</strong> model is <strong>the</strong>n lost and cells are not truly<br />
associated by sound hydraulic pr<strong>in</strong>ciples (Ghanem et al.,<br />
1996). In nei<strong>the</strong>r case, <strong>the</strong>refore, is <strong>the</strong> model a real substitute<br />
for fur<strong>the</strong>r fi eld measurement. Frequently, <strong>the</strong>re<br />
is an <strong>in</strong>teraction between <strong>the</strong> uncerta<strong>in</strong>ties <strong>in</strong>herent <strong>in</strong><br />
a one-dimensional model and those aris<strong>in</strong>g <strong>in</strong> higherorder<br />
model applications, s<strong>in</strong>ce <strong>the</strong> output of <strong>the</strong> onedimensional<br />
model is sometimes required to determ<strong>in</strong>e <strong>the</strong><br />
<strong>in</strong>itial water surface profi le for <strong>the</strong> more complex model.<br />
While PHABSIM rema<strong>in</strong>s someth<strong>in</strong>g of an ‘<strong>in</strong>dustry<br />
standard’ tool, <strong>the</strong> biological representation is also often<br />
highly simplifi ed and empirical (Pusey, 1998). Wood et al.<br />
(2001) found that <strong>the</strong> strength of ecological relationships<br />
<strong>in</strong>creased and model errors are reduced with <strong>in</strong>creas<strong>in</strong>g<br />
spatial resolution, <strong>in</strong>dicat<strong>in</strong>g that most potential is still at<br />
<strong>the</strong> site-scale but that all modell<strong>in</strong>g efforts are limited by<br />
<strong>the</strong> lack of longer-term data sets, too. However, one advantage<br />
of this habitat modell<strong>in</strong>g approach is that <strong>the</strong>re are<br />
clear manuals that defi ne step-by-step procedures, which<br />
allow replication of results by different researchers. The<br />
disadvantage of this is that it has led to poor applications<br />
by practitioners with little experience. Best results are<br />
obta<strong>in</strong>ed where teams <strong>in</strong>clud<strong>in</strong>g hydraulic eng<strong>in</strong>eers,<br />
hydrologists and ecologists work toge<strong>the</strong>r, us<strong>in</strong>g habitat<br />
modell<strong>in</strong>g as a basis for <strong>the</strong>ir river-specifi c studies. Major<br />
strengths and weaknesses of <strong>the</strong> PHABSIM methodology<br />
are summarised <strong>in</strong> Table 7.5.<br />
7.4.5 More Complex Hydrodynamic Modell<strong>in</strong>g and<br />
Habitat Simulation<br />
Numerous specifi c modell<strong>in</strong>g applications have been<br />
described that demonstrate some k<strong>in</strong>d of improvement on<br />
simpler schemes. However, <strong>the</strong>se have not given rise to<br />
any s<strong>in</strong>gle package that is <strong>the</strong> logical replacement to<br />
PHABSIM. Greater hydraulic process representation may<br />
be achieved us<strong>in</strong>g two- and three-dimensional computational<br />
fl uid dynamics models (Alfredsen et al., 1997;<br />
Booker, 2003). New approaches to quantify<strong>in</strong>g hydraulic<br />
habitat have been published (Peters et al., 1995; Nestler<br />
and Sutton, 2000). New habitat models have <strong>in</strong>cluded<br />
additional variables and have been expanded to <strong>the</strong> community<br />
level (Ba<strong>in</strong> et al., 1988; Ba<strong>in</strong>, 1995; Lamoroux<br />
et al., 1998). All of <strong>the</strong>se improvements currently come at<br />
a cost of <strong>in</strong>creased complexity, although <strong>in</strong> <strong>the</strong> future it is<br />
hoped that <strong>the</strong>y can be used to derive general rules with<br />
which to develop improved look-up methods and to defi ne<br />
<strong>the</strong> impacts of river fl ow regulation on populations ra<strong>the</strong>r<br />
than habitats (Hardy, 1998).<br />
Two-dimensional PHABSIM<br />
The PHABSIM method may also be applied us<strong>in</strong>g spatially<br />
cont<strong>in</strong>uous patterns of depth and velocity as predicted<br />
by hydrodynamic models (e.g. Ghanem et al.,<br />
1996). Spatially cont<strong>in</strong>uous <strong>in</strong>formation negates any problems<br />
caused by sampl<strong>in</strong>g frequencies applied <strong>in</strong> fi eld<br />
based studies. This approach also allows visualisation of<br />
areas of suitable habitat. A comparison of habitat suitability<br />
predicted us<strong>in</strong>g two different HSIs for a 50 m reach<br />
of <strong>the</strong> Bere stream, Dorset, UK, is shown <strong>in</strong> Figure 7.13.<br />
These maps show how truncation of <strong>the</strong> depth HSI can<br />
lead to changes <strong>in</strong> patterns of habitat suitability. An<br />
accompany<strong>in</strong>g sensitivity analysis dur<strong>in</strong>g <strong>the</strong> construction<br />
of Figure 7.13 also demonstrates how uncerta<strong>in</strong>ty <strong>in</strong><br />
habitat modell<strong>in</strong>g will depend on <strong>the</strong> <strong>in</strong>teractions between<br />
<strong>the</strong> sampl<strong>in</strong>g density and <strong>the</strong> HSI used. In analogy with<br />
<strong>the</strong> analysis of velocity data <strong>in</strong> Figure 7.9, spatially cont<strong>in</strong>uous<br />
simulations can be sub-sampled to simulate how<br />
<strong>the</strong> number of cross-sections and po<strong>in</strong>t measurements<br />
might affect results had <strong>the</strong> <strong>in</strong>formation been collected <strong>in</strong><br />
<strong>the</strong> fi eld. In general, predicted habitat decreases as more<br />
po<strong>in</strong>ts are added across each section. This occurs because<br />
poor habitat at <strong>the</strong> marg<strong>in</strong>s of <strong>the</strong> channel is more likely<br />
to be sampled when more po<strong>in</strong>ts are measured along each<br />
section. When <strong>the</strong> truncated HSIs were used, <strong>the</strong> positions<br />
of <strong>the</strong> cross-sections <strong>in</strong> relation to a deep pool <strong>in</strong> <strong>the</strong><br />
middle of <strong>the</strong> reach had a strong <strong>in</strong>fl uence on predicted<br />
habitat. When small numbers of cross-sections were used,<br />
<strong>the</strong> likelihood of <strong>the</strong> pool be<strong>in</strong>g over- or under-sampled
130 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Table 7.5 Major strengths and weaknesses of <strong>the</strong> application of <strong>the</strong> PHABSIM methodology<br />
Strengths Weaknesses<br />
The method used is replicable and <strong>the</strong>refore<br />
does not rely on expert op<strong>in</strong>ion<br />
The method can be applied us<strong>in</strong>g hydraulic<br />
predictions calculated us<strong>in</strong>g a range of<br />
techniques from crude stage-discharge<br />
relationships to spatially cont<strong>in</strong>uous<br />
computational fl uid dynamics modell<strong>in</strong>g.<br />
Channel change can be <strong>in</strong>corporated through<br />
repeated survey<br />
Multi-variate habitat suitability <strong>in</strong>dices have<br />
been developed (Vismara et al., 2001).<br />
Application of habitat mapp<strong>in</strong>g has allowed<br />
results to be up-scaled to represent <strong>the</strong><br />
length of river under <strong>in</strong>vestigation<br />
(Maddock and Bird, 1996; Maddock, 1999).<br />
Recent studies have confi rmed that physical<br />
habitat is an important factor controll<strong>in</strong>g<br />
fi sh populations <strong>in</strong> <strong>the</strong> long term (Sabaton<br />
et al., 2003; Souchon and Capra, 2003).<br />
No readily applicable alternative methods<br />
have been developed which enable<br />
prediction of population changes as <strong>the</strong><br />
result of water resource impacts.<br />
was higher. Results also suggest that, as <strong>the</strong> number of<br />
cross-section <strong>in</strong>creases, <strong>the</strong>re is less variation <strong>in</strong> predicted<br />
habitat above six cross-sections when <strong>the</strong> non-truncated<br />
HSIs were used and eight cross-sections when <strong>the</strong> truncated<br />
HSIs were used.<br />
Swimm<strong>in</strong>g Speeds<br />
A more sophisticated example of habitat modell<strong>in</strong>g us<strong>in</strong>g<br />
hydrodynamic calculations is to compare patterns of<br />
velocities predicted by three-dimensional models with <strong>the</strong><br />
swimm<strong>in</strong>g speeds of fi sh to assess <strong>the</strong> effects of channel<br />
design on fi sh habitat. The swimm<strong>in</strong>g performance of fi sh<br />
can also be analysed to assess <strong>the</strong> health of <strong>the</strong> fi sh when<br />
exposed to sub-lethal toxic chemicals (e.g. Alsop et al.,<br />
1999). One frequently used measurement is <strong>the</strong> Maximum<br />
Results may be sensitive to <strong>the</strong> location of cross-sections and<br />
<strong>the</strong> position of measurements along <strong>the</strong>se cross-sections<br />
Inaccuracies <strong>in</strong> hydraulic modell<strong>in</strong>g, especially <strong>in</strong> steep boulder<br />
rivers (Azzell<strong>in</strong>o & Vismara, 2001) and where vegetation<br />
growth causes changes <strong>in</strong> stage-discharge relationships<br />
throughout <strong>the</strong> year (Hearne et al. 1994) will affect habitat<br />
predictions.<br />
Requires re-calibration to assess changes <strong>in</strong> channel morphology<br />
over time. Incorporation of channel change would require<br />
expensive repeated surveys or sediment transport predictions,<br />
which have <strong>the</strong>ir own sources of uncerta<strong>in</strong>ty.<br />
Rules for weight<strong>in</strong>g depth, velocity and substrate are not readily<br />
available.<br />
Typically restricted to depth, velocity and substrate, although<br />
variables such as water quality, vegetation cover can be<br />
<strong>in</strong>corporated.<br />
Reached-based results (cover<strong>in</strong>g approximately 100 m of river<br />
length) must be scaled up.<br />
Calculates habitat suitability and not population numbers or<br />
presence/absence.<br />
In most studies physical habitat availability at low fl ow has been<br />
<strong>the</strong> ma<strong>in</strong> area of <strong>in</strong>terest, with PHABSIM used primarily to<br />
compare <strong>the</strong> implications of alternative fl ow regulation<br />
scenarios on habitat. This has led to an emphasis on <strong>the</strong><br />
relationship between discharge and usable habitat given a<br />
distribution of relatively shallow depths and slow velocities.<br />
Susta<strong>in</strong>ed Swimm<strong>in</strong>g Speed (MSSS). This is defi ned<br />
as <strong>the</strong> maximum velocity at which a fi sh can swim for<br />
a period of more than 200 m<strong>in</strong>utes (Turnpenny et al.,<br />
2001).<br />
Booker (2003) compared simulated velocity patterns to<br />
MSSS of roach, dace and chub to assess habitat suitability<br />
at high fl ows <strong>in</strong> two reaches of <strong>the</strong> <strong>River</strong> Tame, an urban<br />
river <strong>in</strong> Birm<strong>in</strong>gham, UK. The percentage of <strong>in</strong>-stream<br />
area that was less than <strong>the</strong> MSSS was used as an <strong>in</strong>dicator<br />
of habitat quality. This was <strong>the</strong> area of river, at a specifi ed<br />
distance from <strong>the</strong> bed, <strong>in</strong> which fi sh of a certa<strong>in</strong> size and<br />
species could <strong>the</strong>oretically susta<strong>in</strong> a position for at least<br />
200 m<strong>in</strong>utes. Due to uncerta<strong>in</strong>ty as to <strong>the</strong> exact distance<br />
between <strong>the</strong> fi sh and <strong>the</strong> river bed dur<strong>in</strong>g high fl ows, and<br />
to simplify <strong>the</strong> three-dimensional nature of <strong>the</strong> analysis,<br />
different heights above <strong>the</strong> bed were considered separately.
Hydrological and Hydraulic Aspects of <strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 131<br />
Figure 7.13 Maps of habitat suitability derived us<strong>in</strong>g <strong>the</strong> same hydraulic patterns but different suitability curves: (a) HSIs given <strong>in</strong><br />
Dunbar et al. (2001), (b) <strong>the</strong> same HSIs truncated for depth
132 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
The percentage of survivable area for chub, dace and<br />
roach at bankfull discharge was calculated at two sites<br />
(Figure 7.14). The ‘Highly Modifi ed’ reach was a 91 m<br />
straightened reach with an average width of 11.9 m and<br />
artifi cially streng<strong>the</strong>ned banks conta<strong>in</strong>ed with<strong>in</strong> a larger<br />
two-stage channel. This reach had no dist<strong>in</strong>ct geomorphological<br />
features and relatively uniform bed topography. In<br />
contrast, <strong>the</strong> ‘Less Modifi ed’ reach was 139 m <strong>in</strong> length<br />
with an average width of 10.0 m. This reach was also<br />
conta<strong>in</strong>ed with<strong>in</strong> a two-stage channel but had one bank<br />
consist<strong>in</strong>g of natural material, a slightly s<strong>in</strong>uous path and<br />
an undulat<strong>in</strong>g bed profi le. Figure 7.14(b) shows that as <strong>the</strong><br />
body length <strong>in</strong>creases <strong>the</strong>re is an <strong>in</strong>crease <strong>in</strong> <strong>the</strong> area of<br />
habitat <strong>in</strong> which a fi sh is likely to be able to susta<strong>in</strong> a stationary<br />
position. This is because bigger fi sh can susta<strong>in</strong><br />
faster swimm<strong>in</strong>g speeds. Similarly, as distance from <strong>the</strong><br />
bed <strong>in</strong>creases <strong>the</strong> percentage of survivable habitat<br />
decreases due to faster velocities. Results showed that at<br />
<strong>the</strong> Highly Modifi ed site <strong>the</strong> uniform channel structure<br />
supported no discrete areas of slower velocity which could<br />
be used as refugia. At <strong>the</strong> Less Modifi ed site a deep pool<br />
approximately 40 m from <strong>the</strong> upstream boundary provided<br />
a fl ow refuge that could act as a niche of suitable habitat.<br />
There was also a separate discrete area of slower velocity<br />
created by <strong>the</strong> shelter<strong>in</strong>g effect of a change <strong>in</strong> direction of<br />
<strong>the</strong> river planform.<br />
Bioenergetics<br />
Bioenergetic models attempt to quantify <strong>the</strong> trade-off<br />
between energy ga<strong>in</strong>ed through feed<strong>in</strong>g and that lost<br />
through swimm<strong>in</strong>g ‘costs’. Costs arise from hold<strong>in</strong>g position<br />
<strong>in</strong> a mov<strong>in</strong>g fl ow, <strong>in</strong> <strong>the</strong> capture of food, as well as<br />
through digestion, faeces and ur<strong>in</strong>e (Hayes et al., 2000).<br />
Fish activity may be directly observed and related to fl ow<br />
conditions or estimated on <strong>the</strong> basis of models that use<br />
results of physiological experiments undertaken on fi sh<br />
which are forced to swim aga<strong>in</strong>st fl ows of constant<br />
velocity. For example, Booker et al. (2004b) used a threedimensional<br />
Computational Fluid Dynamics (3D-CFD)<br />
model to simulate hydraulic patterns <strong>in</strong> a 50 m reach of<br />
<strong>the</strong> Bere stream, Dorset, UK. This <strong>in</strong>formation was <strong>the</strong>n<br />
comb<strong>in</strong>ed with a bioenergetic model that used behavioural<br />
and physiological relationships to quantify <strong>the</strong> spatial<br />
pattern of energy ga<strong>in</strong> when feed<strong>in</strong>g on <strong>in</strong>vertebrates drift<strong>in</strong>g<br />
<strong>in</strong> <strong>the</strong> river. The model was tested by compar<strong>in</strong>g patterns<br />
of predicted energy <strong>in</strong>take with observed habitat use<br />
by juvenile salmonids at different times of day. A map of<br />
energy <strong>in</strong>take predicted by <strong>the</strong> model compared with<br />
observed locations of feed<strong>in</strong>g and rest<strong>in</strong>g fi sh at <strong>the</strong> site<br />
is shown <strong>in</strong> Figure 7.15.<br />
There are many uncerta<strong>in</strong>ties that must be considered<br />
when us<strong>in</strong>g this bioenergetic modell<strong>in</strong>g approach as a tool<br />
to assist river restoration. The approach relies on <strong>the</strong> accuracy<br />
of both <strong>the</strong> hydrodynamic predictions and <strong>the</strong> accuracy<br />
of <strong>the</strong> physiological algorithms and parameters used.<br />
Enders et al. (2003) demonstrate that models of fi sh activity<br />
cost based upon constant velocity experiments may<br />
underestimate actual costs by up to a factor of 4.2, because<br />
turbulent fl ow fl uctuations are neglected. Direct observations<br />
of feed<strong>in</strong>g behaviour illustrate close relationships<br />
to fi sh activity and characteristics of turbulent fl ow (Enders<br />
Figure 7.14 (a) Mean ‘maximum susta<strong>in</strong>able swimm<strong>in</strong>g speed’ and ‘burst swimm<strong>in</strong>g speed’ for roach, dace and chub at 8 °C (based<br />
on data from Clough & Turnpenny, 2001), (b) Percentage volume of habitat less than <strong>the</strong> mean ‘maximum susta<strong>in</strong>able swimm<strong>in</strong>g<br />
speed’ at different distances from <strong>the</strong> bed <strong>in</strong> each reach at 18 m 3 s −1 (Source: Figure 11 of Booker, 2003)
Hydrological and Hydraulic Aspects of <strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 133<br />
Figure 7.15 Predicted net energy <strong>in</strong>take for a 0.1 m fi sh and observed fi sh locations<br />
et al., 2005). Models may also be hard to test aga<strong>in</strong>st fi eld<br />
observations of habitat uses because factors o<strong>the</strong>r than<br />
energy <strong>in</strong>take can <strong>in</strong>fl uence microhabitat selection. For<br />
example, <strong>the</strong> proximity to o<strong>the</strong>r fi sh (Valdimarsson and<br />
Metcalfe, 2001) and predation risk or distance to cover<br />
(Mesick, 1988). Fish do not spend 100% of <strong>the</strong>ir time driftfeed<strong>in</strong>g.<br />
They may feed very effi ciently for short periods<br />
and <strong>the</strong>n retreat to more sheltered locations (Gries and<br />
Juanes, 1998). Fur<strong>the</strong>rmore, <strong>the</strong> decision to select a certa<strong>in</strong><br />
position may not result from conditions (e.g. drift density)<br />
at that time but conditions at a previous time or conditions<br />
over a longer period. Also, a fi sh may not have perfect<br />
knowledge of its habitat. This means that fi sh feed<strong>in</strong>g <strong>in</strong> an<br />
energetically poor area may not be aware that <strong>the</strong>re are<br />
more favourable alternative positions elsewhere.<br />
7.4.6 Alternatives to Physical Habitat Models<br />
Several alternatives to complex eco-hydraulic simulation<br />
have emerged. A fi eld approach known as ‘Expert Habitat<br />
Mapp<strong>in</strong>g’ (EHM) is used <strong>in</strong> streams that are hydraulically<br />
complex, particularly where <strong>the</strong> irregularity and variability<br />
of boundary conditions makes hydraulic modell<strong>in</strong>g<br />
<strong>in</strong>appropriate (for example, <strong>in</strong> steep bedrock rivers with<br />
large woody debris). EHM uses habitat suitability criteria<br />
but <strong>the</strong>n relies on fi sh biologists mapp<strong>in</strong>g habitat based on<br />
those criteria at a range of fl ows to develop fl ow-vs-habitat<br />
curves. The method works well provided that <strong>the</strong> HSI cri-<br />
teria are used and verifi ed with spot measurements, and<br />
<strong>the</strong> biologists are experienced.<br />
Alternatively, physically-parameterised models may be<br />
replaced by stochastic or hybrid approaches such as cellular<br />
autometa (Chen et al., 2002) or neural network<br />
models (Werner and Oback, 2001; Reyjol et al., 2001;<br />
Gevrey et al., 2003) as witnessed <strong>in</strong> hydrology and fl oodpla<strong>in</strong><br />
<strong>in</strong>undation studies. As yet, however, <strong>the</strong>se approaches<br />
have focused on simulat<strong>in</strong>g ecological characteristics:<br />
appropriate scales for applications, for <strong>the</strong> development of<br />
rules and tra<strong>in</strong><strong>in</strong>g data sets rema<strong>in</strong> to be explored.<br />
7.5 CONCLUSIONS<br />
Restor<strong>in</strong>g and rehabilitat<strong>in</strong>g rivers for ecological purposes<br />
is an essentially multi-discipl<strong>in</strong>ary and multi-stage activity.<br />
There is grow<strong>in</strong>g awareness that to be successful (that<br />
is, to produce schemes which are susta<strong>in</strong>able <strong>in</strong> <strong>the</strong><br />
medium- to longer-term) truly functional environments<br />
are required, where catchment hydrology provides <strong>the</strong><br />
background, or context, for reach and <strong>in</strong>ter-reach fl ows<br />
and sediment transport. In turn, <strong>the</strong>se hydraulic variables<br />
structure physical habitats, which <strong>the</strong>mselves support a<br />
diverse ecological response. Monitor<strong>in</strong>g of fl ows and<br />
aquatic ecology have been long-stand<strong>in</strong>g areas of research<br />
and cont<strong>in</strong>ue to be vitally important but, <strong>in</strong>creas<strong>in</strong>gly,<br />
emphasis has been upon <strong>the</strong> modell<strong>in</strong>g or simulation of<br />
fl ows and <strong>the</strong> habitats which <strong>the</strong>se determ<strong>in</strong>e. Modell<strong>in</strong>g
134 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
is useful <strong>in</strong> extend<strong>in</strong>g <strong>the</strong> range of fi eld observation, <strong>in</strong><br />
scenario-type exercises to evaluate restoration schemes<br />
before <strong>the</strong>ir implementation and also <strong>in</strong> provid<strong>in</strong>g outputs<br />
aga<strong>in</strong>st which schemes might be appraised after implementation.<br />
Ideally, modell<strong>in</strong>g efforts should proceed with<br />
<strong>the</strong> closest possible coupl<strong>in</strong>g between <strong>the</strong>ir physical<br />
(hydraulic and hydrodynamic) and biological elements,<br />
and <strong>in</strong> close association with improved strategies for fi eld<br />
monitor<strong>in</strong>g, too. Runn<strong>in</strong>g through this chapter are some<br />
very basic <strong>the</strong>mes. These refl ect <strong>the</strong> pr<strong>in</strong>cipal sources of<br />
uncerta<strong>in</strong>ty <strong>in</strong> restoration schemes for ecological purposes<br />
and must be addressed to improve both fi eld and modell<strong>in</strong>g<br />
aspects of restoration design:<br />
data requirements – <strong>the</strong> appropriate amount and k<strong>in</strong>d of<br />
fi eld data required to adequately specify pre- and post<br />
design aspects of <strong>the</strong> river;<br />
characterisation – of fl ow coherence and species assemblages<br />
or behaviour, which partly follows from data<br />
considerations as above, but which also <strong>in</strong>cludes design<br />
criteria relat<strong>in</strong>g to channel form, as well as parameterisation<br />
of models;<br />
coupl<strong>in</strong>g – of key physical and ecological aspects of river<br />
system function <strong>in</strong> models;<br />
awareness – of model application, limitations and sensitivity,<br />
and of o<strong>the</strong>r limitations;<br />
development – of newer, more fl exible means of model<br />
evaluation and of restoration appraisal more generally.<br />
While each of <strong>the</strong> above rema<strong>in</strong> sources of uncerta<strong>in</strong>ty,<br />
<strong>the</strong>y represent, too, areas of opportunity for <strong>the</strong> rapid<br />
development of research and <strong>in</strong>tervention protocols. For<br />
<strong>the</strong> practitioner, ask<strong>in</strong>g questions from each of <strong>the</strong>se areas<br />
– and questions which cut across or l<strong>in</strong>k between <strong>the</strong>m – is<br />
perhaps <strong>the</strong> best form of guidance for recognis<strong>in</strong>g, reduc<strong>in</strong>g<br />
and even <strong>in</strong>corporat<strong>in</strong>g uncerta<strong>in</strong>ty <strong>in</strong>to river restoration<br />
activity from an ecological standpo<strong>in</strong>t.<br />
ACKNOWLEDGEMENTS<br />
This chapter was produced as part of research projects<br />
NER/A/S/1998/00009, ‘Identifi cation of physically-based<br />
design criteria for riffl e-pool sequences <strong>in</strong> river rehabilitation’,<br />
NER/D/S/2000/01422, ‘Formation of a new river<br />
channel: fl ow, sediment and vegetation dynamics’, and<br />
NER/T/S/2001/01250, ‘Vegetation <strong>in</strong>fl uences on fi ne sediment<br />
and propagule dynamics <strong>in</strong> groundwater-fed rivers:<br />
implications for river management, restoration and riparian<br />
biodiversity’. These were funded <strong>in</strong> <strong>the</strong> United<br />
K<strong>in</strong>gdom by <strong>the</strong> Natural Environment Research Council.<br />
SSIIM is produced and made available by Nils R. B.<br />
Olsen, Department of Hydraulic and Environmental<br />
Eng<strong>in</strong>eer<strong>in</strong>g, The Norwegian University of Science and<br />
Technology. Scott McBa<strong>in</strong> k<strong>in</strong>dly provided references to<br />
<strong>the</strong> fi gures and tabular <strong>in</strong>formation used <strong>in</strong> <strong>the</strong> text, and<br />
advised on <strong>in</strong>itial chapter layout and content. Nigel Wright<br />
made additional suggestions <strong>in</strong> respect of hydraulic<br />
modell<strong>in</strong>g.<br />
REFERENCES<br />
Acreman MC. 2000. The chang<strong>in</strong>g hydrology of <strong>the</strong> UK. In:<br />
Acreman MC (Ed), The Hydrology of <strong>the</strong> UK: A Study of<br />
Change. Routledge: London, UK.<br />
Acreman MC. 2002. Case Studies of Managed Flood Releases.<br />
Environmental Flow Assessment Part III. World Bank Water<br />
Resources and Environmental Management Best Practice Brief<br />
No 8, World Bank: Wash<strong>in</strong>gton, DC.<br />
Acreman MC, Elliott CRN. 1996. Evaluation of <strong>the</strong> river Wey<br />
restoration project us<strong>in</strong>g <strong>the</strong> Physical HABitat SIMuation<br />
(PHABSIM) model. Proceed<strong>in</strong>gs of <strong>the</strong> 31st MAFF Conference<br />
of <strong>River</strong> and Coastal Eng<strong>in</strong>eers, Keele University, 3–5<br />
July 1996; 1.2.1–1.2.14.<br />
Acreman MC, K<strong>in</strong>g J. 2003. Defi n<strong>in</strong>g water requirements. In:<br />
Dyson M, Bergkamp G, Scanlon H (Eds), Flow: The Essentials<br />
of Environmental Flows. IUCN: Gland, Switzerland and Cambridge,<br />
UK.<br />
Alfredsen K, Marchand W, Bakken TH, Harby A. 1997. Application<br />
and comparison of computer models quantify<strong>in</strong>g impacts<br />
of river regulation on fi sh habitat. In: Broch E, Lysne DK,<br />
Flatabo N, Helland-Hansen E (Eds), Proceed<strong>in</strong>gs of <strong>the</strong> 3rd<br />
International conference on hydropower Hydropower ’97, A.A.<br />
Balkema Publishers: Rotterdam/Brookfi eld.<br />
Alsop DH, McGeer JC, McDonald DG, Wood CM. 1999. Costs<br />
of chronic waterborne z<strong>in</strong>c exposure and <strong>the</strong> consequences of<br />
z<strong>in</strong>c acclimation on <strong>the</strong> gill/z<strong>in</strong>c <strong>in</strong>teractions of ra<strong>in</strong>bow trout<br />
<strong>in</strong> hard and soft water. Environmental Toxicology and Chemistry<br />
18: 1014–1025.<br />
Anderson MG, Wall<strong>in</strong>g DE, Bates PD (Eds). 1996. Floodpla<strong>in</strong><br />
Processes. John Wiley & Sons Ltd: Chichester, UK.<br />
Archer D, Newson, M. 2002. The use of <strong>in</strong>dices of fl ow variability<br />
<strong>in</strong> assess<strong>in</strong>g <strong>the</strong> hydrological and <strong>in</strong>stream habitat impacts of<br />
upland afforestation and dra<strong>in</strong>age. Journal of Hydrology 268:<br />
244–258.<br />
Armitage PD, Petts GE. 1992. Biotic score and prediction to<br />
assess <strong>the</strong> effects of water abstraction on river macro<strong>in</strong>vertebrates<br />
for conservation purposes. Aquatic Conservation:<br />
Mar<strong>in</strong>e and Freshwater Ecosystems 2: 1–17.<br />
Arth<strong>in</strong>gton AH, Long GC (Eds). 1997. Logan <strong>River</strong> Trial of <strong>the</strong><br />
Build<strong>in</strong>g Block Methodology for Assess<strong>in</strong>g Environmental<br />
Flow Requirements: Background Papers. Centre for Catchment<br />
and In-Stream Research and Department of Natural Resoruces:<br />
Brisbane, Queensland.<br />
Arth<strong>in</strong>gton AH, Lloyd R (Eds). 1998. Logan <strong>River</strong> Trial of <strong>the</strong><br />
Build<strong>in</strong>g Block Methodology for Assess<strong>in</strong>g Environmental<br />
Flow Requirements: Workshop Report. Centre for Catchment<br />
and In-Stream Research and Dept Natural Resources: Brisbane,<br />
Queensland, Australia.
Hydrological and Hydraulic Aspects of <strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 135<br />
Azzell<strong>in</strong>o A, Vismara R. 2001. Pool quality <strong>in</strong>dex: New method<br />
to defi ne m<strong>in</strong>imum fl ow requirements of high-gradient, loworder<br />
streams. Journal of Environmental Eng<strong>in</strong>eer<strong>in</strong>g 127 (11):<br />
1003–1013.<br />
Barker I, Kirmond A. 1998. <strong>Manag<strong>in</strong>g</strong> surface water abstraction.<br />
In: Wheater H, Kirby C (Eds), Hydrology <strong>in</strong> a Chang<strong>in</strong>g Environment<br />
British Hydrological Society, 249–258.<br />
Ba<strong>in</strong> MB. 1995. Habitat at <strong>the</strong> local scale: multivariate patterns for<br />
stream fi shes. Bull. Fr. Peche Pisic. 337/338/339: 165–177.<br />
Ba<strong>in</strong> MB, F<strong>in</strong>n JT, Booke HE. 1988. Streamfl ow regulation and<br />
fi sh community structure. Ecology 69: 382–392.<br />
Belaud A, Chaveroche P, Lim P, Sabaton C. 1989. Probability-ofuse<br />
curves applied to brown trout (Salmo trutta fario L.) <strong>in</strong><br />
rivers of sou<strong>the</strong>rn France. Regulated <strong>River</strong>s: Research and<br />
Management 3: 321–336.<br />
Benetti AD, Lanna AE, Cobalch<strong>in</strong>i MS. 2002. Current practice<br />
for establish<strong>in</strong>g environmental fl ows <strong>in</strong> Brazil. Proceed<strong>in</strong>gs of<br />
<strong>the</strong> Fourth International Ecohydraulics Symposium, Sou<strong>the</strong>rn<br />
Waters, Cape Town.<br />
Bergen SD, Bolton SM, Fridley JL. 2001. Design Pr<strong>in</strong>ciples for<br />
ecological eng<strong>in</strong>eer<strong>in</strong>g. Ecological Eng<strong>in</strong>eer<strong>in</strong>g 18: 201–210.<br />
Beven K. 2000. Ra<strong>in</strong>fall–runoff Modell<strong>in</strong>g. John Wiley & Sons<br />
Ltd: Chichester, UK.<br />
Beven K. 2001. On modell<strong>in</strong>g as collective <strong>in</strong>telligence. Hydrological<br />
Processes 15: 2205–2207.<br />
Bierkens MFP, F<strong>in</strong>ke PA, de Willigen P. 2000. Upscal<strong>in</strong>g and<br />
Downscal<strong>in</strong>g Methods for Environmental Research. Kluwer<br />
Academic Publishers: Dordrecht, The Ne<strong>the</strong>rlands.<br />
Booker DJ. 2003. Hydraulic modell<strong>in</strong>g of fi sh habitat <strong>in</strong> urban<br />
rivers dur<strong>in</strong>g high fl ows. Hydrological Processes 17: 577–599.<br />
Booker DJ, Dunbar MJ. 2004. Application of Physical HAbitat<br />
SIMulation (PHABSIM) modell<strong>in</strong>g to modifi ed urban river<br />
channels. <strong>River</strong> Research and Applications 20: 167–183.<br />
Booker DJ, Dunbar MJ, Acreman MC et al. 2004a. Habitat assessment<br />
at <strong>the</strong> catchment scale; application to <strong>the</strong> <strong>River</strong> Itchen,<br />
UK. British Hydrological Society International Conference:<br />
Hydrology: Science & Practice for <strong>the</strong> 21st Century, July 2004,<br />
volume II, 10–18, British Hydrological Society, London, UK.<br />
Booker DJ, Dunbar MJ, Ibbotson AT. 2004b Predict<strong>in</strong>g juvenile<br />
salmonid drift-feed<strong>in</strong>g habitat quality us<strong>in</strong>g a three-dimensional<br />
hydraulic-bioenergetic model. Ecological Modell<strong>in</strong>g<br />
177: 157–177.<br />
Booker DJ, Dunbar MJ, Shamseld<strong>in</strong> A et al. 2003. Physical<br />
habitat assessment <strong>in</strong> urban rivers under future fl ow scenarios.<br />
CIWEM Journal 17: 251–257.<br />
Boon PJ. 1998. <strong>River</strong> restoration <strong>in</strong> fi ve dimensions. Aquatic Conservation:<br />
Mar<strong>in</strong>e and Freshwater Ecosystems 8: 257–264.<br />
Bovee KD. 1982. A Guide to Stream Habitat Analysis us<strong>in</strong>g <strong>the</strong><br />
IFIM. US Fish and Wildlife Service Report FWS/OBS-82/26:<br />
Fort Coll<strong>in</strong>s, Colorado.<br />
Bovee KD, Lamb BL, Bartholow JM et al. 1998. Stream habitat<br />
analysis us<strong>in</strong>g <strong>the</strong> Instream Flow Incremental Methodology. US<br />
Geological Survey, Biological Resources Division, Information<br />
and Technology Report USGS/BRD/ITR–1998–004.<br />
Bunn SE, Arth<strong>in</strong>gton AH. 2002. Basic pr<strong>in</strong>ciples and ecological<br />
consequences of altered fl ow regimes for aquatic biodiversity.<br />
Environmental Management 50: 492–507.<br />
Chen Q, Mynett AE, M<strong>in</strong>ns AW. 2002. Application of cellular<br />
automata to modell<strong>in</strong>g competitive growths of two underwater<br />
specieis Chara aspera and Potamogeton pect<strong>in</strong>atus <strong>in</strong> Lake<br />
Veluwe. Ecological Modell<strong>in</strong>g 147: 253–265.<br />
Chow VT. 1973. Open Channel Hydraulics. McGraw-Hill: Tokyo,<br />
Japan.<br />
Clarke SJ, Bruce-Burgess L, Wharton G. 2003. L<strong>in</strong>k<strong>in</strong>g form and<br />
function: towards an eco-hydromorphic approach to susta<strong>in</strong>able<br />
rive restoration. Aquatic Conservation: Mar<strong>in</strong>e and Freshwater<br />
Ecosystems 13: 439–450.<br />
Clausen B, Biggs BJF. 2000. Flow variables for ecological studies<br />
<strong>in</strong> temperate streams: group<strong>in</strong>gs based on covariance. Journal<br />
of Hydrology 237: 184–197.<br />
Clifford NJ. 2001. Conservation and <strong>the</strong> river channel environment.<br />
In: Warren A, French JR (Eds), Habitat Conservation.<br />
John Wiley & Sons Ltd: Chichester, UK; 67–104.<br />
Clifford NJ. 2002. Hydrology: <strong>the</strong> chang<strong>in</strong>g paradigm. Progress<br />
<strong>in</strong> Physical Geography 26: 290–301.<br />
Clifford NJ, Soar PJ, Emery JC et al. 2002a. Numerical fl ow modell<strong>in</strong>g<br />
for eco-hydraulic and river rehabilitation applications:<br />
a case study of <strong>the</strong> <strong>River</strong> Cole, Birm<strong>in</strong>gham, UK. In: Bousmar<br />
D, Zech Y (Eds), <strong>River</strong> Flow 2002. Lisse: Swets & Zeitl<strong>in</strong>ger/<br />
Balkema: Rotterdam, The Ne<strong>the</strong>rlands; 1195–1204.<br />
Clifford NJ, Soar PJ, Emery JC et al. 2002b. Susta<strong>in</strong><strong>in</strong>g water<br />
related ecosystems – <strong>the</strong> role of <strong>in</strong>-stream bedform design <strong>in</strong><br />
river channel rehabilitation. In: Friend 2002 – Regional hydrology:<br />
bridg<strong>in</strong>g <strong>the</strong> gap between research and practice. IAHS<br />
Publication 274: 407–413, IAHS, Wall<strong>in</strong>gford, UK.<br />
Clifford NJ, Soar PJ, Harmar OP et al. 2005. Assessment of<br />
hydrodynamic simulation results for eco-hydrological and ecohydraulic<br />
applications: a spatial semivariance approach.<br />
Hydrological Processes 19: 3631–3648.<br />
Clifford NJ, Harmar OP, Harvey G, Petts GE. 2006. Physical<br />
habitat, eco-hydraulics and river design: a review and reevaluation<br />
of some popular concepts and methods. Aquatic Conservation:<br />
Mar<strong>in</strong>e and Freshwater Ecosystems 16: 389–408.<br />
Clough SC, Turnpenny AWH. 2001. Swimm<strong>in</strong>g Speeds <strong>in</strong> Fish:<br />
Phase 1. Environment Agency R&D Technical Report W2–<br />
026/TR1, Environment Agency: Bristol, UK.<br />
Crowder DW, Diplas P. 2000. Us<strong>in</strong>g two-dimensional hydrodynamic<br />
models at scales of ecological importance. Journal of<br />
Hydrology 230: 172–191.<br />
Dakova S, Uzunov Y, Mandadjiev D. 2000. Low fl ow – <strong>the</strong> river’s<br />
ecosystem limit<strong>in</strong>g factor. Ecological Eng<strong>in</strong>eer<strong>in</strong>g 16: 167–<br />
174.<br />
Dunbar MJ, Ibbotson AT, Gow<strong>in</strong>g IM et al. 2001. Ecologically<br />
Acceptable Flows Phase III: Fur<strong>the</strong>r Validation of PHABSIM<br />
for <strong>the</strong> Habitat Requirements of Salmonid Fish. F<strong>in</strong>al R&D<br />
Technical report to <strong>the</strong> Environment Agency, Environment<br />
Agency, Bristol, UK.<br />
Dyson M, Bergkamp G, Scanlon H (Eds). 2003 Flow: The Essentials<br />
of Environmental Flows. IUCN: Gland, Switzerland and<br />
Cambridge, UK.<br />
Elliott CRN, Johnson IW, Sekul<strong>in</strong> AE et al. 1996. Guide to <strong>the</strong><br />
use of <strong>the</strong> Physical Habitat Simulation System, EA Release<br />
version. Environment Agency R&D Technical Report W20.<br />
Environment Agency: Bristol, UK.
136 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Elliott CRN, Dunbar MJ, Gow<strong>in</strong>g I, Acreman MC. 1999. A<br />
habitat assessment approach to <strong>the</strong> management of groundwater<br />
dom<strong>in</strong>ated rivers, Hydrological Processes 13: 459–475.<br />
Emery JC. 2003. Characteristics and controls of gravel-bed riffl epool<br />
sequences for habitat assessment and river rehabilitation<br />
design. Unpublished PhD <strong>the</strong>sis, University of Birm<strong>in</strong>gham,<br />
UK.<br />
Emery JC, Gurnell AM, Clifford NJ et al. 2003. Classify<strong>in</strong>g <strong>the</strong><br />
performance of riffl e-pool bedforms for habitat assessment and<br />
river rehabilitation design. <strong>River</strong> Research and Applications<br />
19: 533–549.<br />
Enders EC, Boisclair D, Roy AG. 2003. The effect of turbulence<br />
on <strong>the</strong> cost of swimm<strong>in</strong>g for juvenile Atlantic salmon (Salmos<br />
salar). Canadian Journal of Fisheries and Aquatic Sciences 60:<br />
1149–1160.<br />
Enders EC, Buffi ngton-Belanger T, Boisclair D, Roy AG. 2005.<br />
The feed<strong>in</strong>g behaviour of juvenile Atlantic salmon <strong>in</strong> relation<br />
to turbulent fl ow. Journal of Fish Biology 66: 242–253.<br />
Everard M. 1996. The importance of periodic droughts for ma<strong>in</strong>ta<strong>in</strong><strong>in</strong>g<br />
diversity <strong>in</strong> <strong>the</strong> freshwater environment. Freshwater<br />
Forum 7: 33–50.<br />
Extence C, Balbi DM, Chadd RP. 1999. <strong>River</strong> fl ow <strong>in</strong>dex<strong>in</strong>g us<strong>in</strong>g<br />
British benthic macro-<strong>in</strong>vertebrates: a framework for sett<strong>in</strong>g<br />
hydro-ecological objectives. Regulated <strong>River</strong>s Research and<br />
Management 15: 543–574.<br />
Frissell CA, Liss WJ, Warren CE, Hurley MD. 1986. A hierarchical<br />
framework for stream classifi cation: view<strong>in</strong>g streams <strong>in</strong> a<br />
watershed context. Environmental Management 10: 199–214.<br />
French JR, Clifford NJ. 2000. Hydrodynamic modell<strong>in</strong>g as a basis<br />
for expla<strong>in</strong><strong>in</strong>g estuar<strong>in</strong>e environmental dynamics: some conceptual<br />
and methodological issues. Hydrological Processes 14:<br />
2089–2108.<br />
Fjeldstad H-P. 2001. Numerical modell<strong>in</strong>g tools for predict<strong>in</strong>g<br />
physical habitat adjustments. In: Taugbol T, L’Abee-Lund J-H<br />
(Eds), Proceed<strong>in</strong>gs from <strong>the</strong> CONNECT workshop ‘Physical<br />
habitat restoration <strong>in</strong> canalised watercourses – possibilities<br />
and constra<strong>in</strong>ts, Norweigan Institute of Nature Research, Oslo,<br />
Norway, 91–99.<br />
Gevrey M, Dimopoulosb I, Leka S. 2003. Modell<strong>in</strong>g <strong>the</strong> structure<br />
of aquatic communities: concepts, methods and problems. Ecological<br />
Modell<strong>in</strong>g 160: 249–164.<br />
Ghanem A, Steffl er P, Hicks F, Katopodis C. 1996. Two-dimensional<br />
hydraulic simulation of physical habitat conditions <strong>in</strong><br />
fl ow<strong>in</strong>g streams. Regulated <strong>River</strong>s-Research & Management<br />
12: 185–200.<br />
Gilvear DJ, Heal KV, Stephen A. 2002. Hydrology and <strong>the</strong> ecological<br />
quality of Scottish river ecosystems. The Science of <strong>the</strong><br />
Total Environment 294: 131–159.<br />
Gregory S, Boyer K, Gurnell AM (Eds). 2003. The Ecology and<br />
Management of Wood <strong>in</strong> World <strong>River</strong>s. American Fisheries<br />
Society Symposium 37, American Fisheries Society: Be<strong>the</strong>sda,<br />
Maryland.<br />
Gries G, Juanes F. 1998. Microhabitat use by juvenile Atlantic<br />
salmon (Salmo salar) shelter<strong>in</strong>g dur<strong>in</strong>g <strong>the</strong> day <strong>in</strong> summer.<br />
Canadian Journal of Zoology 76: 1441–1449.<br />
Gore JA, Crawford DJ, Addison DS. 1998. An analysis of artifi cial<br />
riffl es and enhancement of benthic community diversity by<br />
physical habitat simulation (PHABSIM) and direct observation.<br />
Regulated <strong>River</strong>s: Research and Management 14: 69–77.<br />
Graf WL. 2001. Damage control: restor<strong>in</strong>g <strong>the</strong> physical <strong>in</strong>tegrity<br />
of America’s rivers. Annals of <strong>the</strong> Association of American<br />
Geographers 91: 1–27.<br />
Hardy RJ, Lane SN, Ferguson RI, Parsons DR. 2003. Assess<strong>in</strong>g<br />
<strong>the</strong> credibility of a series of computational fl uid dynamic simulationsof<br />
open channel fl ow. Hydrological Processes 17:<br />
1539–60.<br />
Hardy TB. 1998. The future of habitat model<strong>in</strong>g and <strong>in</strong>stream<br />
fl ow assessment techniques Regulated <strong>River</strong>s: Research and<br />
Management 14: 405–420.<br />
Hayes JW, Stark JD, Shearer KA. 2000. Development and test of<br />
a whole-lifetime forag<strong>in</strong>g and bioenergetics growth model for<br />
drift feed<strong>in</strong>g brown trout. Transactions of <strong>the</strong> American Fisheries<br />
Society 125: 315–332.<br />
Hearne J, Johnson IW, Armitage PD. 1994. Determ<strong>in</strong>ation of<br />
ecologically acceptable fl ows <strong>in</strong> rivers with seasonal changes<br />
<strong>in</strong> <strong>the</strong> density of macrophyte Regulated <strong>River</strong>s: Research and<br />
Management 9: 117–184.<br />
Hendry K, Cragg-H<strong>in</strong>e D, O’Grady MO et al. 2003. Management<br />
of habitat for rehabilitation and enhancement of salmonid<br />
stocks. Fisheries Research 12: 171–192.<br />
Hill MT, Platt S, Beschta RL. 1991. Ecological and geomorphological<br />
concepts for <strong>in</strong>stream and out of channel fl ow requirements.<br />
<strong>River</strong>s 2: 198–210.<br />
Janauer GA. 2000. Ecohydrology: fus<strong>in</strong>g concepts and scales.<br />
Ecological Eng<strong>in</strong>eer<strong>in</strong>g 16: 9–16.<br />
Johnson IW, Elliott CRN, Gustard A et al. 1993. Ecologically<br />
Acceptable Flows National <strong>River</strong>s Authority R&D Project<br />
Record 282/1/Wx: Bristol, UK.<br />
Johnson IW, Elliott CRN, Gustard A. 1995. Modell<strong>in</strong>g <strong>the</strong> effect<br />
of groundwater abstraction on salmonid habitat availability <strong>in</strong><br />
<strong>the</strong> <strong>River</strong> Allen, Dorset, England. Regulated <strong>River</strong>s: Research<br />
& Management 10: 229–238.<br />
Junk WJ, Bayley PB, Sparks RE. 1989. The fl ood pulse concept<br />
<strong>in</strong> river-fl oodpla<strong>in</strong> systems. In: Dodge DP (Ed), Proceed<strong>in</strong>gs<br />
of <strong>the</strong> International Large <strong>River</strong> Symposium (LARS), Canadian<br />
Special Publication of Fisheries and Aquatic Sciences 106,<br />
National Research Council of Canada, Ottawa; 110–127.<br />
Kemp JL, Harper DM, Crosa GA. 1999. Use of ‘functional habitats’<br />
to l<strong>in</strong>k ecology with morphology and hydrology <strong>in</strong> river<br />
rehabilitation. Aquatic Conservation: Mar<strong>in</strong>e and Freshwater<br />
Ecosystems 9: 159–178.<br />
Kemp JL, Harper DM, Crosa GA. 2000. The habitat-scale ecohydraulics<br />
of rivers. Ecological Eng<strong>in</strong>eer<strong>in</strong>g 16: 17–29.<br />
Kondolf GM, Larsen EW, Williams JG. 2000. Measur<strong>in</strong>g and<br />
modell<strong>in</strong>g <strong>the</strong> hydraulic environment for assess<strong>in</strong>g <strong>in</strong>stream<br />
fl ows. North American Journal of Fisheries Management 20:<br />
1016–1028.<br />
Kirkby MJ. 1993. Network hydrology and geomorphology. In:<br />
Beven K, Kirkby MJ (Eds), Channel Network Hydrology. John<br />
Wiley & Sons Ltd: Chichester, UK; 1–11.<br />
K<strong>in</strong>g JM, Tharme RE, de Villiers MS (Eds). 2000. Environmental<br />
fl ow assessments for rivers: manual for <strong>the</strong> Build<strong>in</strong>g Block<br />
Methodology. Water Research Commission Report TT 131/00:<br />
Pretoria, South Africa.
Hydrological and Hydraulic Aspects of <strong>River</strong> <strong>Restoration</strong> <strong>Uncerta<strong>in</strong>ty</strong> for Ecological Purposes 137<br />
Lamouroux N, Capra H, Pouilly M. 1998. Predict<strong>in</strong>g Habitat<br />
Suitability for lotic fi sh: l<strong>in</strong>k<strong>in</strong>g statistical hydraulic models<br />
with multivariate habitat use models. Regulated <strong>River</strong>s 14:<br />
1–11.<br />
Lamouroux N, Doutriaux E, Terrier C, Zylberblat M. 1999.<br />
Modélisation des impacts de la gestion des débits réservés du<br />
Rhône sur les peuplements piscicoles. Bullet<strong>in</strong> Français de la<br />
Pêche et de la Pisciculture 352: 45–61.<br />
Lane SN. 1998. Hydraulic modell<strong>in</strong>g <strong>in</strong> hydrology and geomorphology:<br />
a review of high resolution approaches. Hydrological<br />
Processes 12: 1131–1150.<br />
Lane SN, Richards KS. 2001. The ‘validation of hydrodynamic<br />
models: some critical perspectives. In: Anderson MG, Poates<br />
PD (Eds), Model validation: perspectives <strong>in</strong> hydrological<br />
science. John Wiley & Sons Ltd: Chichester, UK; 414–438.<br />
LeClerc 2002. Ecohydraulics, last frontier for fl uvial hydraulics:<br />
research challenges and multidiscipl<strong>in</strong>ary perspectives. In:<br />
Bousmar D, Zech Y (Eds), <strong>River</strong> Flow 2002. Swets & Zeitl<strong>in</strong>ger/<br />
Balkema: Liss; 13–26.<br />
Logan P, Furse M. 2002. Prepar<strong>in</strong>g for <strong>the</strong> European Water<br />
Framework Directive – mak<strong>in</strong>g <strong>the</strong> l<strong>in</strong>ks between habitat and<br />
aquatic biota. Aquatic Conservation: Mar<strong>in</strong>e and Freshwater<br />
Ecosystems 12: 425–437.<br />
Lytle DA, Poff NL. 2004. Adaptation to natural fl ow regimes.<br />
Trends <strong>in</strong> Ecology and Evolution 19: 94–100.<br />
Maddock I. 1999. The importance of physical habitat assessment<br />
for evaluat<strong>in</strong>g river health. Freshwater Biology 41: 373–391.<br />
Maddock IP, Bird D. 1996. The application of habitat mapp<strong>in</strong>g to<br />
identify representative PHABSIM sites on <strong>the</strong> <strong>River</strong> Tavy,<br />
Devon, UK. In: Leclerc et al. (Eds), Proceed<strong>in</strong>gs of <strong>the</strong> 2nd<br />
International Symposium on habitats and hydraulics, Quebec,<br />
Canada, June 1996.<br />
Marsh TJ. 2002. Capitalis<strong>in</strong>g on river fl ow data to meet chang<strong>in</strong>g<br />
national needs – a UK perspective. Flow Measurement and<br />
Instrumentation 13: 291–298.<br />
Mat<strong>the</strong>ws RC, Bao Y. 1991. The Texas method of prelim<strong>in</strong>ary<br />
<strong>in</strong>stream fl ow determ<strong>in</strong>ation. <strong>River</strong>s 2: 295–310.<br />
McBa<strong>in</strong> & Trush, Inc. (Eds). 2002. San Joaqu<strong>in</strong> <strong>River</strong> <strong>Restoration</strong><br />
Study Background Report. Prepared for Friant Water Users<br />
Authority, L<strong>in</strong>dsay, CA, and Natural Resources Defense<br />
Council: San Francisco, California.<br />
McPherson JI. 1997. Appeal by Thames Water Utilities under<br />
Section 43 of <strong>the</strong> Water Resources Act 1991, Inspector’s<br />
Report, WAT/95/22. Department of <strong>the</strong> Environment: Bristol,<br />
UK.<br />
Mesick CF. 1988. Effects of food and cover on numbers of apache<br />
and brown trout establish<strong>in</strong>g residency <strong>in</strong> artifi cial stream channels.<br />
Transactions of <strong>the</strong> American Fisheries Society 117:<br />
421–431.<br />
Milhous RT. 1999. History, <strong>the</strong>ory, use, and limitations of <strong>the</strong><br />
Physical Habitat Simulation System. Proceed<strong>in</strong>gs of <strong>the</strong> 3rd<br />
International Symposium on Ecohydraulics, Salt Lake City,<br />
Utah, USA. Available on CD-ROM only.<br />
Milhous RT, Updike MA, Scheider DM. 1989. Physical Habitat<br />
Simulation (PHABSIM) Reference Manual, Version II.<br />
Instream Flow Information Paper No. 26. US Fish and Wildlife<br />
Service Biology Report 89 (16).<br />
Morhardt JE, Hanson DF, Coulston CJ. 1983. Instream fl ow<br />
analysis: Increased accuracy us<strong>in</strong>g habitat mapp<strong>in</strong>g. Waterpower<br />
83- An International Conference on Hydropower.<br />
Conference Proceed<strong>in</strong>gs Vol 3, Tennessee Valley Authority,<br />
Knoxville; 1294–1304.<br />
Mousouridis A. 2001. Identify<strong>in</strong>g Upland Channel Network<br />
Dynamics Over Decadal and Century Timescales. Unpublished<br />
PhD <strong>the</strong>sis, University of London, UK.<br />
Naiman RJ, Lonzarich DG, Beechie TJ, Ralph SC. 1992. General<br />
pr<strong>in</strong>ciples of classifi cation and <strong>the</strong> assessment of conservation<br />
potential <strong>in</strong> rivers. In: Boon PJ, Calow P, Petts GE (Eds), <strong>River</strong><br />
Conservation and Management. John Wiley & Sons Ltd:<br />
Chichester, UK; 93–123.<br />
Nestler J, Sutton VK. 2000. Describ<strong>in</strong>g scales of features <strong>in</strong> river<br />
channels us<strong>in</strong>g fractal geometry concepts. Regulated <strong>River</strong>s:<br />
Research & Management 16: 1–22.<br />
Newbury R, Gaboury M. 1993. Exploration and rehabilitation of<br />
hydraulic habitats <strong>in</strong> streams us<strong>in</strong>g pr<strong>in</strong>ciples of fl uvial behaviour.<br />
Freshwater Biology 29: 195–210.<br />
Newson MD. 2002. Geomorphological concepts and tools for<br />
susta<strong>in</strong>able river ecosystem management. Aquatic Conservation:<br />
Mar<strong>in</strong>e and Freshwater Ecosystems 12: 365–379.<br />
Newson MD, Newson CL. 2000. Geomorphology, ecology<br />
and river channel habitat: mesoscale approaches to bas<strong>in</strong>scale<br />
challenges. Progress <strong>in</strong> Physical Geography 24: 195–<br />
217.<br />
Newson MD, Harper DM, Padmore CL et al. 1998. A costeffective<br />
approach for l<strong>in</strong>k<strong>in</strong>g habitats, fl ow types and species<br />
requirements. Aquatic Conservation: Mar<strong>in</strong>e and Freshwater<br />
Ecosystems 8: 431–446.<br />
Nicholas AP. 2001. Computational fl uid dynamics modell<strong>in</strong>g of<br />
boundary roughness <strong>in</strong> gravel-bed rivers: an <strong>in</strong>vestigation of<br />
<strong>the</strong> effects of random variability <strong>in</strong> bed elevation. Earth Surface<br />
Processes and Landforms 26: 345–362.<br />
Oreskes N, Shrader-Frechette K, Belitz K. 1994. Verifi cation,<br />
validation, and confi rmation of numerical models <strong>in</strong> <strong>the</strong> Earth<br />
Sciences. Science 263: 641–646.<br />
Padmore CL. 1997. Biotopes and <strong>the</strong>ir hydraulics: a method for<br />
defi n<strong>in</strong>g <strong>the</strong> physical component of freshwater quality. In:<br />
Boon PJ, Howell DL (Eds), Freshwater Quality: Defi n<strong>in</strong>g<br />
<strong>the</strong> Indefi nable? The Stationery Offi ce: Ed<strong>in</strong>burgh, UK;<br />
251–257.<br />
Parasiewicz P. 2001. MesoHABSIM: A concept for application<br />
of <strong>in</strong>stream fl ow models <strong>in</strong> river restoration plann<strong>in</strong>g. Fisheries<br />
26: 6–13.<br />
Parasiewicz P, Dunbar MJ. 2001. Physical habitat modell<strong>in</strong>g<br />
for fi sh: A develop<strong>in</strong>g approach. Large <strong>River</strong>s, 12 2–4, Arch.<br />
Hydrobiol. Suppl. 135/2–4: 239–268.<br />
Peters MR, Abt SR, Watson CC et al. 1995. Assessment of<br />
restored river<strong>in</strong>e habitat us<strong>in</strong>g RCHARC. Water Resources<br />
Bullet<strong>in</strong> 31: 745–752<br />
Petts GE. 1984. Impounded <strong>River</strong>s: Perspectives for Ecological<br />
Management. John Wiley & Sons Ltd: Chichester, UK.<br />
Petts GE, Amoros C. 1996. Fluvial Hydrosystems. Chapman and<br />
Hall: London, UK.<br />
Poff NL, Allan JD, Ba<strong>in</strong> MB et al. 1997. The natural fl ow regime.<br />
BioScience 47: 769–784.
138 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Prudhomme C, Jakob D, Svensson C. 2003. <strong>Uncerta<strong>in</strong>ty</strong> and<br />
climate change impact on <strong>the</strong> fl ood regime of small UK catchments.<br />
Journal of Hydrology 277: 1–23.<br />
Pusey BJ. 1998. Methods address<strong>in</strong>g <strong>the</strong> fl ow requirements of<br />
fi sh. In: Arth<strong>in</strong>gton AH, Zalucki JM (Eds), Comparative evaluation<br />
of environmental fl ow assessment techniques review of<br />
methods. Occassional Paper 27/98, Land and Water Resources<br />
and Development Corporation: Canberra, Australia; 64–103.<br />
Railsback SF, Lamberson RH, Harvey BC, Duffy WE, 1999.<br />
Movement rules for <strong>in</strong>dividual-based models of stream fi sh.<br />
Ecological Modell<strong>in</strong>g 123: 73–89.<br />
Reice SR, Wissmar RC, Naiman RJ. 1990. Disturbance regimes,<br />
resilience and recovery of animal communities and habitats <strong>in</strong><br />
lotic ecosystems. Environmental Management 14: 647–660.<br />
Reyjol Y, Lim P, Belaud A, Lek S. 2001. Modell<strong>in</strong>g microhabitat<br />
used by fi sh <strong>in</strong> natural and regulated fl ows <strong>in</strong> <strong>the</strong> river Garonne<br />
(France). Ecological Modell<strong>in</strong>g 146: 131–142.<br />
Reynolds CS. 2002. Ecological pattern and ecosystem <strong>the</strong>ory.<br />
Ecological Modell<strong>in</strong>g 158: 181–200.<br />
Richter BD, Baumgartner JV, Powell J, Braun DP. 1996. A method<br />
for assess<strong>in</strong>g hydrological alteration with<strong>in</strong> ecosystems. Conservation<br />
Biology 10: 1163–1174.<br />
Richter BD, Baumgartner JV, Wig<strong>in</strong>gton R, Braun DP. 1997. How<br />
much water does a river need? Freshwater Biology 37:<br />
231–249.<br />
Rodriguez-Iturbe, Valdes JB. 1979. The geomorphic structure<br />
of hydrologic response. Water Resources Research 15:<br />
1409–1420.<br />
Roussel JM, Bardonnet A, Claude A. 1999. Microhabitats of<br />
brown trout when feed<strong>in</strong>g on drift and when rest<strong>in</strong>g <strong>in</strong> a<br />
lowland salmonid brook: effects on Weighted Usable Area.<br />
Archiv Fur Hydrobiologie 146: 413–429.<br />
Rowntree K, Wadeson R. 1998. A geomorphological framework<br />
for <strong>the</strong> assessment of <strong>in</strong>stream fl ow requirements. Aquatic Ecosystem<br />
Health and Management 1: 125–141.<br />
Sabaton C, Souchon Y, Lascaux JM et al. 2003. The ‘Guaranteed<br />
Flow Work<strong>in</strong>g Group’: A French feedback of IFIM based on<br />
habitat and brown trout population time series observations. In:<br />
Lamb BL, D Garcia de Jalon C, Sabaton C et al. (Eds), Proceed<strong>in</strong>gs<br />
of International IFIM users conference: Colorado<br />
State University, Fort Coll<strong>in</strong>s, Colorado.<br />
Schmidt JC, Webb RH, Valdez RA et al. 1998. Science and values<br />
<strong>in</strong> river restoration <strong>in</strong> <strong>the</strong> Grand Canyon. BioScience 48:<br />
735–747.<br />
Smakht<strong>in</strong> VU. 2001. Low fl ow hydrology: a review. Journal of<br />
Hydrology 240: 147–186.<br />
Souchon Y, Capra H. 2003. Aquatic habitat modell<strong>in</strong>g: biological<br />
validations of IFIM/PHABSIM methodology and new perspectives<br />
In: Lamb BL, D Garcia de Jalon C, Sabaton C et al. (Eds),<br />
Proceed<strong>in</strong>gs of International IFIM users conference: Colorado<br />
State University, Fort Coll<strong>in</strong>s, Colorado.<br />
Souchon Y, Keith P. 2001. Freshwater fi sh habitat: science, management<br />
and conservation <strong>in</strong> France. Aquatic Ecosystem Health<br />
and Management 4: 401–412.<br />
Spence R, Hickley P. 2000. The use of PHABSIM <strong>in</strong> <strong>the</strong> management<br />
of water resources and fi sheries <strong>in</strong> England and Wales.<br />
Ecological Eng<strong>in</strong>eer<strong>in</strong>g 16: 153–158.<br />
Stanford JA, Ward JV, Liss WJ et al. 1996. A general protocol for<br />
restoration of regulated rivers. Regulated <strong>River</strong>s: Research and<br />
Management 12: 391–413.<br />
Stillwater Sciences 2001. Tuolumne <strong>River</strong> restoration program<br />
summary report, summary of studies, conceptual models, restoration<br />
projects, and ongo<strong>in</strong>g monitor<strong>in</strong>g. Prepared for <strong>the</strong><br />
CALFED/AFRP Adaptive Management Forum, with assistance<br />
from <strong>the</strong> Tuolumne <strong>River</strong> Technical Advisory Committee.<br />
Strevens AP. 1999. Impacts of groundwater abstraction on <strong>the</strong><br />
trout fi shery of <strong>the</strong> <strong>River</strong> Piddle; and an approach to <strong>the</strong>ir alleviation.<br />
Hydrological Processes 13: 487–496.<br />
Tennant DL. 1976. Instream fl ow regimens for fi sh, wildlife,<br />
recreation and related environmental resources. Fisheries 1:<br />
6–10.<br />
Thoms MC, Sheldon F. 2002. An ecosystem approach for determ<strong>in</strong><strong>in</strong>g<br />
environmental water allocations <strong>in</strong> Australian dryland<br />
river systems: <strong>the</strong> role of geomorphology. Geomorphology 47:<br />
153–168.<br />
Turnpenny AWH, Blay SR, Carron JJ, Clough SC. 2001. Literature<br />
Review of Swimm<strong>in</strong>g Speeds of Freshwater Fish. Environment<br />
Agency R&D Technical Report W2–026/TR2,<br />
Environment Agency: Bristol, UK.<br />
US Army Corps of Eng<strong>in</strong>eers. 2002. HEC-RAS <strong>River</strong> Analysis<br />
System. CDP-68. US Army Corps of Eng<strong>in</strong>eers: Davis,<br />
California.<br />
Valdimarsson SK, Metcalfe NB. 2001. Is <strong>the</strong> level of aggression<br />
and dispersion <strong>in</strong> territorial fi sh dependent on light <strong>in</strong>tensity?<br />
Animal Behaviour 61: 1143–1149.<br />
Van W<strong>in</strong>kle W, Jager HI, Railsback SF et al. 1998. Individualbased<br />
model of sympatric populations of brown and ra<strong>in</strong>bow<br />
trout for <strong>in</strong>stream fl ow assessment: model description and calibration.<br />
Ecological Modell<strong>in</strong>g 110: 175–207.<br />
Vismara R, Azzell<strong>in</strong>o A, Bosi R et al. 2001. Habitat suitability<br />
curves for brown trout (Salmo trutta fario L.) <strong>in</strong> <strong>the</strong> <strong>River</strong> Adda,<br />
Nor<strong>the</strong>rn Italy: Compar<strong>in</strong>g univariate and multivariate approaches.<br />
Regulated <strong>River</strong>s, Research & Management 17: 37–50.<br />
Waters BF. 1976. A methodology for evaluat<strong>in</strong>g <strong>the</strong> effects of<br />
different stream fl ows on salmonid habitat. In: Orsborn JF,<br />
Allman CH (Eds), Instream Flow Needs. American Fisheries<br />
Society: Be<strong>the</strong>sda, Maryland. 254–266.<br />
Werner H, Oback M. 2001. New neural network types estimat<strong>in</strong>g<br />
<strong>the</strong> accuracy response for ecological modell<strong>in</strong>g. Ecological<br />
Modell<strong>in</strong>g 146: 289–298.<br />
Wood PJ, Hannah DM, Agnew MD, Petts GE. 2001. Scales of<br />
hydroecological variability with<strong>in</strong> a groundwater-dom<strong>in</strong>ated<br />
stream. Regulated <strong>River</strong>s: Research and Management 17:<br />
347–367.<br />
Wu FC, Wang CF. 2002. Effect of fl ow-related substrate alteration<br />
on physical habitat: A case study of <strong>the</strong> endemic river loach<br />
S<strong>in</strong>ogastromyzon puliensis (Cypr<strong>in</strong>iformes, Homalopteridae)<br />
downstream of Chi-Chi diversion weir, Chou-Shui Creek,<br />
Taiwan. <strong>River</strong> Research and Applications 18: 155–169.<br />
Yrjänä T, Lahti M, Kamula R. 1999. Effects of a habitat enhancement<br />
experiment <strong>in</strong> <strong>the</strong> Laukka area of <strong>the</strong> river Oulujoki,<br />
F<strong>in</strong>land. University of Oulu. Publications of Water Resources<br />
and Environmental Eng<strong>in</strong>eer<strong>in</strong>g Laboratory. A5. (In F<strong>in</strong>nish,<br />
with English Summary).
<strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat Edited by Stephen Darby and David Sear<br />
© 2008 John Wiley & Sons, Ltd<br />
8<br />
<strong>Uncerta<strong>in</strong>ty</strong> Surround<strong>in</strong>g <strong>the</strong> Ecological<br />
Targets and Response of <strong>River</strong> and<br />
Stream <strong>Restoration</strong><br />
8.1 INTRODUCTION<br />
Mart<strong>in</strong> R. Perrow 1 , Eleanor R. Skeate 1 , David Leem<strong>in</strong>g 2 , Judy England 3 and<br />
Mark L. Toml<strong>in</strong>son 1<br />
In his sem<strong>in</strong>al paper, Bradshaw (1987) described restoration<br />
ecology as <strong>the</strong> ‘acid test’ of ecology <strong>in</strong> that if <strong>the</strong>re<br />
is suffi cient understand<strong>in</strong>g of form, structure, process and<br />
function <strong>the</strong>n it should be possible to restore any particular<br />
habitat to a close approximation of that which is desired.<br />
The recent Handbook of Ecological <strong>Restoration</strong> <strong>in</strong> two<br />
volumes detail<strong>in</strong>g pr<strong>in</strong>ciples (Perrow and Davy, 2002a)<br />
and practice (Perrow and Davy, 2002b) suggests that restoration<br />
ecology has come of age as a science, although<br />
<strong>the</strong>re is great disparity <strong>in</strong> <strong>the</strong> level of understand<strong>in</strong>g and<br />
thus success <strong>in</strong> different biomes. The book also reveals<br />
<strong>the</strong> fundamentally different approach to <strong>the</strong> restoration of<br />
lotic and lentic environments.<br />
In lakes, especially shallow ones, <strong>the</strong>re has been a <strong>the</strong>oretical<br />
shift from ‘bottom-up’, where <strong>the</strong> form and function<br />
was controlled by physical and chemical properties,<br />
to ‘top-down’, where components at <strong>the</strong> top of <strong>the</strong> food<br />
web, especially fi sh, have great bear<strong>in</strong>g on lower trophic<br />
levels and even physical properties (Perrow et al., 2002).<br />
The focus of lake restoration is <strong>the</strong>refore often on fi sh<br />
(Jeppesen and Sammalkarpi, 2002) embrac<strong>in</strong>g and us<strong>in</strong>g<br />
fundamentally ecological <strong>in</strong>teractions.<br />
In rivers and streams <strong>the</strong>re has also been a shift <strong>in</strong> focus<br />
away from water quality – which is now not seen as <strong>the</strong><br />
primary limit<strong>in</strong>g factor for a healthy, natural system as it<br />
was immediately after <strong>the</strong> <strong>in</strong>dustrial revolution and prior<br />
1 ECON, Ecological Consultancy, UK.<br />
2 Consultant Ecologist, UK<br />
3 Environment Agency, UK<br />
to <strong>the</strong> development of improved water quality standards –<br />
to <strong>the</strong> limitations of physical habitat structure. For example<br />
<strong>in</strong> <strong>the</strong> United K<strong>in</strong>gdom, <strong>in</strong>itial restoration efforts on <strong>the</strong><br />
<strong>River</strong> Thames <strong>in</strong> 1858 were driven by water problems<br />
caused by pollution from human and animal wastes<br />
(Gameson and Wheeler, 1977), whereas now <strong>the</strong> focus is<br />
on restor<strong>in</strong>g habitat quality, especially for fi sh, <strong>in</strong> <strong>the</strong> upper<br />
catchment where tributaries have suffered from river eng<strong>in</strong>eer<strong>in</strong>g<br />
schemes (Rob<strong>in</strong>son and Whitton, 2004).<br />
What has become <strong>the</strong> billion dollar <strong>in</strong>dustry of river<br />
restoration (Malakoff, 2004; Palmer et al., 2005) generally<br />
uses a geomorphological (<strong>in</strong> comb<strong>in</strong>ation with eng<strong>in</strong>eer<strong>in</strong>g)<br />
approach (Downs et al., 2002; Malakoff 2004), <strong>in</strong> <strong>the</strong><br />
hierarchical gradient of process–form–habitat–biota promoted<br />
by <strong>the</strong> highly <strong>in</strong>fl uential National Research Council<br />
(of <strong>the</strong> United States) publication <strong>Restoration</strong> of Aquatic<br />
Ecosystems (1992). This is restore natural water and sediment<br />
regime; restore natural channel geometry; restore<br />
natural riparian plant communities; and restore native<br />
aquatic plants and animals.<br />
The widely perceived ecological focus of river restoration<br />
(Newson et al., 2002; Downs and Sk<strong>in</strong>ner 2002;<br />
Ormerod, 2004), is <strong>the</strong>refore <strong>the</strong> fi nal step <strong>in</strong> <strong>the</strong> cha<strong>in</strong> of<br />
works. As with ripples on a pond or <strong>the</strong> cascade response<br />
through trophic levels (Carpenter and Kitchell, 1993), <strong>the</strong><br />
strength of <strong>the</strong> <strong>in</strong>teraction and <strong>the</strong> response is likely to<br />
weaken <strong>the</strong> fur<strong>the</strong>r from <strong>the</strong> source action. To take a more<br />
specifi c example, <strong>the</strong> huge amount of river restoration
140 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
focused on anadromous fi shes <strong>in</strong>clud<strong>in</strong>g sturgeon<br />
(Waldman and Wirg<strong>in</strong>, 1998) and particularly salmonids,<br />
upon which millions of dollars are spent annually <strong>in</strong> <strong>the</strong><br />
US Pacifi c Northwest alone (Roni et al., 2002), effectively<br />
h<strong>in</strong>ges on <strong>the</strong> relationship between <strong>the</strong> fi sh and limit<strong>in</strong>g<br />
habitat variables. However, such relationships are not<br />
always clear-cut and may expla<strong>in</strong> a low proportion of <strong>the</strong><br />
variation <strong>in</strong> abundance and biomass (Milner et al., 1985).<br />
In very simple terms, a ‘restored’ river may be perfectly<br />
capable of support<strong>in</strong>g a high abundance and biomass of<br />
fi sh but may actually conta<strong>in</strong> very few. In recognition of<br />
<strong>the</strong> fact that biological and ecological limitations (i.e.<br />
stock limited recruitment, disease, predation, competition<br />
etc) may be more important than <strong>the</strong> generic problem of<br />
not hav<strong>in</strong>g enough habitat, release of large numbers of<br />
artifi cially raised juveniles may have to be undertaken, e.g.<br />
on <strong>the</strong> <strong>River</strong> Mattole <strong>in</strong> California (Mattole <strong>Restoration</strong><br />
council http://www.mattole.org/) and <strong>in</strong> rivers <strong>in</strong> North-<br />
East England (Russell, 1994). In a nutshell, <strong>the</strong> paradigm<br />
of gett<strong>in</strong>g <strong>the</strong> physical structure right and <strong>the</strong>n <strong>the</strong> ‘ecology’<br />
will <strong>the</strong>n surely follow is likely to be a false premise<br />
(Ormerod, 2004).<br />
Geomorphological and ecological processes are <strong>in</strong>extricably<br />
<strong>in</strong>tertw<strong>in</strong>ed and best considered <strong>in</strong> an eco-hydromorphic<br />
approach to restoration (Clarke et al., 2003). To<br />
illustrate, consider a low order boulder-strewn upland<br />
stream. Here, geological structure, gradient and geomorphological<br />
processes expressed <strong>in</strong> fl ow and sediment<br />
regime may <strong>in</strong>itially be viewed as dom<strong>in</strong>ant over ecological<br />
processes. But consider <strong>the</strong> impact of <strong>the</strong> presence or<br />
absence of trees and <strong>the</strong>ir coarse woody debris (CWD)<br />
feed<strong>in</strong>g back <strong>in</strong>to geomorphological processes by constra<strong>in</strong><strong>in</strong>g<br />
channel width and <strong>in</strong>creas<strong>in</strong>g fl ows and promot<strong>in</strong>g<br />
erosion <strong>in</strong> one place and slow<strong>in</strong>g fl ows and caus<strong>in</strong>g<br />
deposition <strong>in</strong> ano<strong>the</strong>r. Allochthonous <strong>in</strong>put <strong>in</strong> <strong>the</strong> form of<br />
leaf fall provides coarse particulate matter (CPOM) to <strong>the</strong><br />
channel provid<strong>in</strong>g <strong>the</strong> basis of nutrient cycl<strong>in</strong>g and spiral<strong>in</strong>g<br />
(see Newbold, 1992 for an overview) and energy fl ow<br />
(Calow, 1992 for an overview) and thus biological production,<br />
ultimately determ<strong>in</strong>es <strong>the</strong> biomass of <strong>the</strong> biota.<br />
Coarse woody debris from <strong>the</strong> trees promotes retentiveness<br />
and decomposition of litter (Lepori et al., 2005), key ecological<br />
functions <strong>in</strong>fl uenc<strong>in</strong>g <strong>the</strong> abundance and diversity<br />
of shredd<strong>in</strong>g <strong>in</strong>vertebrates responsible for produc<strong>in</strong>g fi ne<br />
particulate matter (FPOM) used by o<strong>the</strong>r feed<strong>in</strong>g guilds,<br />
consumed <strong>in</strong> turn by fi sh, birds and mammals (Richardson<br />
and Jackson, 2002). In this hypo<strong>the</strong>tical example, understand<strong>in</strong>g<br />
and subsequently manipulat<strong>in</strong>g ecological processes,<br />
such as succession, herbivory (predation) and<br />
competition, to <strong>in</strong>fl uence <strong>the</strong> structure and composition of<br />
<strong>the</strong> riparian tree fl ora may be as (if not more) effective as<br />
manipulat<strong>in</strong>g physical habitat structure.<br />
Moreover, whilst ecological processes may be embodied<br />
with<strong>in</strong> a particular component of <strong>the</strong> fauna and fl ora, this<br />
is <strong>in</strong>tuitively unlikely to operate over trophic scales with<strong>in</strong><br />
biotic components as well as <strong>in</strong>teract with abiotic components.<br />
Individuals, species populations and <strong>the</strong>ir communities<br />
cannot be seen as mere products of <strong>the</strong> physical (and<br />
chemical) framework but <strong>in</strong>extricably l<strong>in</strong>ked with it via<br />
particular processes. Thus, an eng<strong>in</strong>eer<strong>in</strong>g analogue of<br />
fi tt<strong>in</strong>g habitat components back toge<strong>the</strong>r (i.e. a build<strong>in</strong>g<br />
block functional habitat approach; Harper et al., 1992) to<br />
make a naturally function<strong>in</strong>g ecosystem is likely to be too<br />
simplistic. The nub of an eng<strong>in</strong>eer<strong>in</strong>g approach, that a<br />
habitat component or unit may be expressed as a certa<strong>in</strong><br />
number of <strong>in</strong>dividuals or composition of a community,<br />
thus takes no account of process, but ra<strong>the</strong>r assumes it.<br />
Our experiences of river restoration <strong>in</strong> <strong>the</strong> United<br />
K<strong>in</strong>gdom lead us to believe <strong>the</strong>re is a widespread lack of<br />
understand<strong>in</strong>g of what ‘ecology’ is and does <strong>in</strong> river restoration.<br />
Pert<strong>in</strong>ent to this volume, we aim to explore<br />
whe<strong>the</strong>r this <strong>in</strong>tuitively leads to uncerta<strong>in</strong>ty (<strong>in</strong> this case<br />
reducible ignorance, see Chapter 3) and a lack of confi -<br />
dence <strong>in</strong> ecologists amongst o<strong>the</strong>r types of restoration<br />
practitioner. For example, Downs and Sk<strong>in</strong>ner (2002) suggested<br />
‘river restoration ecology suffers from <strong>the</strong> <strong>in</strong>ability<br />
of ecologists to defi ne <strong>the</strong>ir system requirements as closely<br />
as, for <strong>in</strong>stance, eng<strong>in</strong>eers can specify and model (although<br />
not necessarily achieve) <strong>the</strong>ir fl ood defence requirements’.<br />
This neatly encapsulates <strong>the</strong> perception that not only is<br />
ecological restoration uncerta<strong>in</strong> but that <strong>the</strong> scientifi c basis<br />
to predict and subsequently evaluate ecological response<br />
is severely lack<strong>in</strong>g, perhaps even absent.<br />
One of <strong>the</strong> aims of this chapter is to determ<strong>in</strong>e whe<strong>the</strong>r<br />
such bold statements are justifi ed <strong>in</strong> <strong>the</strong> context of <strong>the</strong><br />
basic premise of this volume that much restoration of<br />
rivers has and is be<strong>in</strong>g conducted with considerable uncerta<strong>in</strong>ty<br />
<strong>in</strong> one form or ano<strong>the</strong>r. To do this, <strong>the</strong> sources of<br />
uncerta<strong>in</strong>ty surround<strong>in</strong>g ecology (and ecologists) affect<strong>in</strong>g<br />
river restoration schemes are explored <strong>in</strong> stepwise manner,<br />
broadly ak<strong>in</strong> to <strong>the</strong> plann<strong>in</strong>g, design and target sett<strong>in</strong>g,<br />
implementation and outcome and appraisal phases <strong>in</strong> any<br />
project (Figure 8.1). An attempt is made to determ<strong>in</strong>e if<br />
<strong>the</strong>re is widespread uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong> ecological response<br />
and outcome of restoration schemes over and above that<br />
expected as a result of natural variation, and, if so, whe<strong>the</strong>r<br />
uncerta<strong>in</strong>ty and unreliability can be reduced.<br />
8.2 SOURCES OF ECOLOGICAL UNCERTAINTY<br />
8.2.1 Inherent Variability<br />
Natural systems are <strong>in</strong>herently variable <strong>in</strong> time and space,<br />
and probably none more so than rivers. Variation <strong>in</strong> space
Figure 8.1 Is Ecological <strong>River</strong> <strong>Restoration</strong> ‘Uncerta<strong>in</strong>’? Infl uences of Institutional and Ecological Drivers <strong>in</strong> Project Plann<strong>in</strong>g and Implementation<br />
<strong>Uncerta<strong>in</strong>ty</strong> Surround<strong>in</strong>g <strong>the</strong> Ecological Targets and Response of <strong>River</strong> and Stream <strong>Restoration</strong> 141
142 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
is particularly well developed both longitud<strong>in</strong>ally, laterally<br />
and vertically <strong>in</strong> rivers, as recognised by attempts to<br />
understand <strong>the</strong>m with<strong>in</strong> <strong>the</strong> river cont<strong>in</strong>uum concept<br />
(RCC) (Vannote et al., 1980). In larger rivers with extensive<br />
fl oodpla<strong>in</strong> systems <strong>the</strong> fl ood pulse concept (Junk et<br />
al., 1989) has been used to frame <strong>in</strong>teractions between<br />
fl oodpla<strong>in</strong> and channel, which typically vary over time.<br />
Temporal variation may also conta<strong>in</strong> o<strong>the</strong>r important <strong>in</strong>terannual<br />
(random, periodic or semi-periodic), annual, seasonal<br />
and diel components, not simply amongst fauna and<br />
fl ora but also <strong>in</strong> <strong>the</strong> rates of ecosystem processes such as<br />
productivity and nutrient and energy fl ow.<br />
Buijse et al. (2002) outl<strong>in</strong>e that whilst considerable<br />
effort and emphasis have been directed at attempt<strong>in</strong>g to<br />
quantify and understand biological, hydrological and geomorphological<br />
<strong>in</strong>teractions that toge<strong>the</strong>r determ<strong>in</strong>e <strong>the</strong><br />
function<strong>in</strong>g of a river fl oodpla<strong>in</strong> system, <strong>the</strong>se are <strong>in</strong>evitably<br />
variable and complex. Complexity occurs at <strong>the</strong><br />
physical level (variation <strong>in</strong> types of river<strong>in</strong>e system), <strong>the</strong><br />
restoration level (<strong>the</strong> comb<strong>in</strong>ation of features requir<strong>in</strong>g<br />
restoration are site-specifi c) and <strong>the</strong> ecological level (community<br />
<strong>in</strong>teractions). Hughes et al. (Chapter 6) argue that<br />
‘complexity’ accounts for a degree of <strong>the</strong> perceived ecological<br />
‘uncerta<strong>in</strong>ty’ <strong>in</strong> river restoration and it must be<br />
accepted that much may rema<strong>in</strong> unknowable (Chapter 3).<br />
In a wider discussion of management of nature <strong>in</strong> <strong>the</strong><br />
United K<strong>in</strong>gdom, Adams (1997) suggested a recent shift<br />
<strong>in</strong> ecological th<strong>in</strong>k<strong>in</strong>g from look<strong>in</strong>g to restore a ‘stable<br />
equilibrium’, often taken to represent historical conditions,<br />
to a new view that a river<strong>in</strong>e system is by its nature<br />
dynamic and chang<strong>in</strong>g, i.e. one that embraces natural variability.<br />
This is <strong>the</strong> basis of ‘recovery enhancement’ where<br />
<strong>the</strong> scope for natural recovery is exploited, after <strong>the</strong> shackl<strong>in</strong>g<br />
constra<strong>in</strong>ts such as structures (e.g. weirs) and modifi<br />
ed eng<strong>in</strong>eered banks, management actions such as<br />
<strong>in</strong>tensive vegetation management regimes or even simply<br />
graz<strong>in</strong>g by livestock are removed. The river may also be<br />
given <strong>the</strong> space to move (i.e. dur<strong>in</strong>g <strong>the</strong> re-connection<br />
of <strong>the</strong> fl oodpla<strong>in</strong> as on <strong>the</strong> <strong>River</strong>s Cole and Brede, see<br />
Kronvang et al., 1998; Holmes and Nielsen, 1998). Such<br />
action to favour a potentially less controllable and predictable<br />
system (Hughes et al., Chapter 6) may appear to <strong>the</strong><br />
eng<strong>in</strong>eer, whose target is to achieve control over <strong>the</strong> system<br />
to meet specifi ed fl ood defence and shipp<strong>in</strong>g access<br />
targets, as simple acceptance of uncerta<strong>in</strong>ty.<br />
8.2.2 Limited Understand<strong>in</strong>g<br />
In spite of <strong>the</strong> wealth of work carried out on river<strong>in</strong>e<br />
systems, several authors (de Waal et al., 1995; Buijse<br />
et al., 2002) have argued that <strong>the</strong>re is actually a lack<br />
of scientifi c literature about river restoration, and that a<br />
lack of robust datasets has rendered much of <strong>the</strong> work<br />
that has been carried out unpublishable <strong>in</strong> peer reviewed<br />
journals. Although l<strong>in</strong>ked, <strong>the</strong>re are actually two issues<br />
here, <strong>in</strong>formation exchange and scientifi c substance and<br />
credibility.<br />
Information exchange has undoubtedly <strong>in</strong>creased enormously<br />
as <strong>the</strong> number of river restoration projects across<br />
<strong>the</strong> globe has <strong>in</strong>creased. Malakoff (2004) reports 30 000<br />
projects <strong>in</strong> <strong>the</strong> United States alone with tens of thousands<br />
more to come <strong>in</strong> <strong>the</strong> next few years. The literature search<br />
for this chapter uncovered a huge variety of <strong>in</strong>formation<br />
on various river restoration projects, especially via websites<br />
such as <strong>the</strong> European Centre for <strong>River</strong> <strong>Restoration</strong><br />
(http://www.rws.nl/rws/riza/home/ecrr/) and its UK<br />
(http://www.<strong>the</strong>rrc.co.uk/) and Danish counterparts (http://<br />
www2.dmu.dk/) and <strong>the</strong> National <strong>River</strong> <strong>Restoration</strong><br />
Science Syn<strong>the</strong>sis, which operates <strong>in</strong> seven states of <strong>the</strong><br />
United States, with a satellite branch <strong>in</strong> Victoria, Australia<br />
(http://www.nrrss.umd.edu/). However, much <strong>in</strong>formation<br />
is produced <strong>in</strong> an abbreviated form and it diffi cult to<br />
determ<strong>in</strong>e whe<strong>the</strong>r many projects outl<strong>in</strong>ed <strong>in</strong> such a<br />
manner are of real scientifi c substance.<br />
At <strong>the</strong> opposite end of <strong>the</strong> spectrum lies <strong>the</strong> academic<br />
peer-reviewed process as a vehicle for <strong>in</strong>formation<br />
exchange. The fact that Ormerod (2004) found only 300+<br />
papers <strong>in</strong>clud<strong>in</strong>g <strong>the</strong> terms river or stream restoration or<br />
rehabilitation <strong>in</strong> <strong>the</strong> title, abstract or key words <strong>in</strong> <strong>the</strong> ISI®<br />
database, <strong>in</strong>dicates that only a fraction of projects reach<br />
<strong>the</strong> academic literature. This could illustrate that much<br />
work is not submitted for publication and/or many submissions<br />
are not accepted as a result of <strong>in</strong>adequate scientifi c<br />
rigour. It may also suggest that few academic river ecologists<br />
are <strong>in</strong>volved with restoration. In <strong>the</strong> United K<strong>in</strong>gdom<br />
at least, ecological <strong>in</strong>put <strong>in</strong> many schemes is undertaken<br />
<strong>in</strong>ternally by statutory bodies such as <strong>the</strong> Environment<br />
Agency or by consult<strong>in</strong>g fi rms (like <strong>the</strong> authors of this<br />
chapter!).<br />
Overall, whilst <strong>the</strong>re is a proliferation of <strong>in</strong>formation,<br />
facilitat<strong>in</strong>g <strong>the</strong> rapid development of project plann<strong>in</strong>g,<br />
target sett<strong>in</strong>g, implementation and post-project appraisal<br />
on a global scale, <strong>the</strong> question-mark over data quality and<br />
<strong>the</strong> lack of <strong>in</strong>volvement of academic ecologists may mean<br />
<strong>the</strong> uncerta<strong>in</strong>ty due to limited knowledge is not reduc<strong>in</strong>g<br />
as rapidly as it might.<br />
8.2.3 Project Selection<br />
In <strong>the</strong> United K<strong>in</strong>gdom, until very recently, virtually all<br />
restoration was opportunistic despite <strong>the</strong> need for a strategic<br />
approach be<strong>in</strong>g recognised over a decade ago (ECON,<br />
1993). One of <strong>the</strong> reasons for this was simply that restoration<br />
was most effectively directed at rivers (reaches) of
<strong>Uncerta<strong>in</strong>ty</strong> Surround<strong>in</strong>g <strong>the</strong> Ecological Targets and Response of <strong>River</strong> and Stream <strong>Restoration</strong> 143<br />
highest conservation value, i.e. those closest to <strong>the</strong> predisturbance<br />
state. Investment <strong>in</strong> <strong>the</strong> most degraded<br />
reaches/rivers was misplaced and wasted resources, as<br />
<strong>the</strong>y could not achieve high status as a result of overwhelm<strong>in</strong>g<br />
and complex limit<strong>in</strong>g factors (Kern, 1992).<br />
Moreover, as shown for lakes (Benndorf, 1992), <strong>the</strong><br />
pathway from highly degraded to high quality is long and<br />
littered with <strong>the</strong> opportunity for unpredictable and <strong>in</strong>direct<br />
effects, which may constra<strong>in</strong> or even reduce ecological<br />
value (i.e. colonisation by alien species). In simple terms,<br />
long restoration trajectories (see Section 8.2.6) are more<br />
diffi cult to plan and pull off.<br />
8.2.4 Project Structure<br />
It is generally acknowledged that river restoration projects<br />
require an <strong>in</strong>tegrated and <strong>in</strong>terdiscipl<strong>in</strong>ary approach,<br />
which requires a number of experts from different fi elds<br />
work<strong>in</strong>g alongside each o<strong>the</strong>r (NRC, 1992). The scale of<br />
<strong>the</strong> project appears to have great <strong>in</strong>fl uence as whilst this<br />
may have been achieved <strong>in</strong> <strong>the</strong> case of large, high-profi le<br />
schemes (e.g. <strong>the</strong> Kissimme <strong>River</strong> <strong>Restoration</strong> Project <strong>in</strong><br />
<strong>the</strong> USA – http://www.sfwmd.gov/org/erd/krr/ – and <strong>the</strong><br />
schemes tackled by <strong>the</strong> <strong>River</strong> <strong>Restoration</strong> Centre <strong>in</strong> <strong>the</strong><br />
UK – http://www.<strong>the</strong>rrc.co.uk/) it does not appear to be<br />
rout<strong>in</strong>e. Tak<strong>in</strong>g <strong>the</strong> schemes evaluated <strong>in</strong> <strong>the</strong> UK <strong>River</strong>s<br />
and Wildlife Handbook as a sample, out of <strong>the</strong> 25 schemes<br />
where <strong>the</strong> project team was described, 80% did not employ<br />
an <strong>in</strong>terdiscipl<strong>in</strong>ary approach (RSPB et al., 1995).<br />
Resources are also clearly an issue, with a common<br />
desire to ensure that <strong>the</strong> bulk of resources available are<br />
channeled <strong>in</strong>to ‘do<strong>in</strong>g’ restoration works ra<strong>the</strong>r than be<br />
absorbed <strong>in</strong>to project <strong>in</strong>frastructure and <strong>in</strong>stitutions<br />
(Bruce-Burgess and Sk<strong>in</strong>ner, 2002). The <strong>in</strong>volvement of<br />
too many organisations and stakeholders with confl ict<strong>in</strong>g<br />
<strong>in</strong>terests may also ultimately constra<strong>in</strong> <strong>the</strong> options for<br />
restoration and <strong>the</strong> ability to undertake it (McDonald<br />
et al., 2004).<br />
From an ecological perspective it seems obvious to<br />
suggest that <strong>the</strong> more central <strong>the</strong> position of ecologists <strong>in</strong><br />
<strong>the</strong> project, <strong>the</strong> greater <strong>the</strong> chance of uncerta<strong>in</strong>ties surround<strong>in</strong>g<br />
ecological issues be<strong>in</strong>g identifi ed early <strong>in</strong> <strong>the</strong> life<br />
of <strong>the</strong> project and subsequently tackled. The perception<br />
that much river restoration has been undertaken with <strong>the</strong><br />
idea of improv<strong>in</strong>g at least some elements of <strong>the</strong> biota,<br />
especially fi sh (Bruce-Burgess and Sk<strong>in</strong>ner, 2002;<br />
Ormerod, 2004) gives <strong>the</strong> impression that such schemes<br />
are ‘front-loaded’ by ecologists. Our experience is more<br />
that ecologists enter <strong>the</strong> restoration fray to ‘count bugs’ <strong>in</strong><br />
<strong>the</strong> process–form–habitat–biota sequence (see previously),<br />
particularly where an opportunity for restoration such as<br />
on <strong>the</strong> back of a ma<strong>in</strong>tenance scheme has been taken.<br />
Clearly, it is too late to set mean<strong>in</strong>gful ecological targets<br />
if <strong>the</strong> ecologists are simply used <strong>in</strong> project monitor<strong>in</strong>g.<br />
8.2.5 Goals and Target Sett<strong>in</strong>g<br />
The nature of <strong>the</strong> goal underp<strong>in</strong>s <strong>the</strong> ecological restoration<br />
process. <strong>Restoration</strong> <strong>in</strong> its strictest sense <strong>in</strong>volves <strong>the</strong><br />
return of <strong>the</strong> structure and function of an ecosystem to a<br />
condition that existed prior to disturbance (NRC, 1992)<br />
and this forms a ready-made goal for <strong>the</strong> project (White<br />
and Walker, 1997). However, reliable historic <strong>in</strong>formation<br />
is often absent and <strong>the</strong>re is also <strong>the</strong> issue that climatic<br />
conditions, amongst o<strong>the</strong>r planetary processes, may have<br />
changed suffi ciently to mean that <strong>in</strong> ecological terms <strong>the</strong><br />
river may no longer function <strong>in</strong> <strong>the</strong> same manner, even if<br />
it had rema<strong>in</strong>ed undisturbed. The use of ecologically<br />
similar but undisturbed contemporary reference sites or<br />
even professional expert op<strong>in</strong>ion based on empirical and/<br />
or computational models, represent an alternative means<br />
of establish<strong>in</strong>g a suitable goal (White and Walker, 1997;<br />
Anderson and Dugger, 1998), although any method still<br />
requires understand<strong>in</strong>g of <strong>the</strong> current nature of <strong>the</strong> system,<br />
range of natural variation, mechanisms whereby impacts<br />
have occurred and <strong>the</strong> specifi c effects of such impacts.<br />
O<strong>the</strong>rwise, restoration may be doomed to failure.<br />
The identifi cation of causes, <strong>in</strong> particular, can be diffi -<br />
cult if <strong>the</strong>y are subtle and far removed <strong>in</strong> space and time<br />
from ecological damage (NRC, 1992). Some of <strong>the</strong> most<br />
commonly overlooked factors are: <strong>the</strong> number of stresses<br />
on biotic components; <strong>the</strong> multi-causal nature of degradation;<br />
and <strong>the</strong> diversity of result<strong>in</strong>g problems. These issues<br />
may only be resolved by adequate basel<strong>in</strong>e monitor<strong>in</strong>g.<br />
Poor basel<strong>in</strong>e data potentially leads to poor target sett<strong>in</strong>g.<br />
Sett<strong>in</strong>g and achiev<strong>in</strong>g targets ultimately provides <strong>the</strong><br />
bench-measure for <strong>the</strong> success of <strong>the</strong> project. Without a<br />
clear and appropriate goal and precise and accurate targets,<br />
<strong>the</strong>re may be a lack of benefi cial impacts, and even detrimental<br />
ones. Kondolf (1998) describes <strong>the</strong> project at Rush<br />
Creek, California, USA, as an example of this type of<br />
mistake. The aim to protect <strong>the</strong> banks from erosion was<br />
<strong>in</strong>appropriate and <strong>in</strong>consistent with geomorphological and<br />
ecological pressures at <strong>the</strong> site, and <strong>the</strong> aim to create a wet<br />
meadow to stop bank erosion was probably unrealistic due<br />
to hydrologic changes. By stopp<strong>in</strong>g bank erosion, <strong>the</strong> end<br />
result of <strong>the</strong> project was that <strong>the</strong> stream’s natural processes<br />
of recovery and re-establishment of woody riparian vegetation<br />
was arrested.<br />
These sort of experiences may have lead Downs and<br />
Sk<strong>in</strong>ner (2002) to suggest that one of <strong>the</strong> two key aspects<br />
limit<strong>in</strong>g restoration was <strong>the</strong> <strong>in</strong>ability of ecologists to set<br />
targets (<strong>the</strong> o<strong>the</strong>r was monitor<strong>in</strong>g and evaluation, see<br />
below). Certa<strong>in</strong>ly, <strong>the</strong>re does seem to be a general lack of
144 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
ecological target sett<strong>in</strong>g <strong>in</strong> projects. In <strong>the</strong> United<br />
K<strong>in</strong>gdom, an evaluation of <strong>the</strong> 40 different case studies by<br />
Holmes (1998) and a fur<strong>the</strong>r 40 <strong>in</strong> <strong>the</strong> New <strong>River</strong>s &<br />
Wildlife Handbook (RSPB et al., 1995) showed that<br />
although projects were driven by ecological goals, only<br />
around 65% actually set any sort or target, and only on<br />
one scheme were quantitative targets (based around <strong>the</strong><br />
restoration of spawn<strong>in</strong>g habitat for salmonid fi shes) set<br />
and publicised (Table 8.1). Even <strong>in</strong> <strong>the</strong> recent Cole and<br />
Skerne projects, perhaps represent<strong>in</strong>g <strong>the</strong> p<strong>in</strong>nacle of restoration<br />
<strong>in</strong> <strong>the</strong> United K<strong>in</strong>gdom to date, <strong>the</strong>re is little sign<br />
of ecological target sett<strong>in</strong>g, nei<strong>the</strong>r qualitative nor qualtitative,<br />
<strong>the</strong> benefi ts brought about by <strong>the</strong> scheme be<strong>in</strong>g<br />
assessed purely on <strong>the</strong> results of monitor<strong>in</strong>g (Holmes and<br />
Nielsen, 1998; Kronvang et al., 1998; Hoffmann et al.,<br />
1998; Biggs et al., 1998; Vivash et al., 1998).<br />
This is <strong>in</strong> sharp contrast with large scale schemes such<br />
as <strong>the</strong> Kissimmee <strong>River</strong> <strong>Restoration</strong> Project RRP (KRRP)<br />
<strong>in</strong> <strong>the</strong> USA (Trexler, 1995; Toth, 1996; Toth and Anderson,<br />
1998; Toth et al., 1998) and <strong>the</strong> Rh<strong>in</strong>e <strong>in</strong> Cont<strong>in</strong>ental<br />
Europe (Buijse et al., 2002). In <strong>the</strong> latter, ecological conditions<br />
<strong>in</strong> a similarly function<strong>in</strong>g section of <strong>the</strong> Danube were<br />
used to establish target conditions. In <strong>the</strong> former, around<br />
60 target aims were selected (performance measures), all<br />
of which had explicit goals associated with restor<strong>in</strong>g <strong>the</strong><br />
full range of structural and functional processes, and<br />
which were developed with peer review by an <strong>in</strong>dependent<br />
scientifi c review panel. Targets covered <strong>the</strong> range of ecosystem<br />
components and trophic levels and <strong>in</strong>cluded habitat<br />
characteristics (12 – hydrology, geomorphology, water<br />
quality), wetland vegetation (10), food base (13 – phytoplankton,<br />
periphyton, <strong>in</strong>vertebrates, herpetofauna) and<br />
fi sh and wildlife (25) (Whalen et al., 2002). The feasibility<br />
Table 8.1 Proportion (%) of <strong>the</strong> 80 projects reviewed by<br />
Holmes (1998) and presented <strong>in</strong> <strong>the</strong> New <strong>River</strong>s and Wildlife<br />
Handbook (1995) <strong>in</strong> England, Wales and Nor<strong>the</strong>rn Ireland<br />
satisfy<strong>in</strong>g selected criteria<br />
Criteria Holmes (1998) NRWH (1995)<br />
Wholly/partially<br />
ecologically driven<br />
65 55<br />
Sites strategically selected<br />
us<strong>in</strong>g ecological<br />
restoration criteria<br />
43 30<br />
Use of an <strong>in</strong>terdiscipl<strong>in</strong>ary<br />
team<br />
50 13<br />
With specifi c ecological<br />
targets<br />
65 68<br />
With some sort of appraisal 70 100<br />
Achiev<strong>in</strong>g some measure of<br />
ecological improvement<br />
100 92<br />
of <strong>the</strong> targets was considered <strong>in</strong> great detail. To evaluate<br />
<strong>the</strong> restoration of <strong>the</strong> fi sh community a conceptual model<br />
outl<strong>in</strong><strong>in</strong>g aspects of ecosystem function was developed<br />
(Trexler, 1995). This <strong>in</strong>tegrated <strong>in</strong>formation from several<br />
levels of biotic organisation (<strong>in</strong>dividuals, populations,<br />
communities and systems) and focused on <strong>the</strong> dynamics<br />
of fl oodpla<strong>in</strong>–channel nutrients and <strong>the</strong> movement of<br />
larvae, juvenile and adult fi sh and <strong>the</strong>ir macro<strong>in</strong>vertebrate<br />
prey.<br />
The experiences from large schemes, especially <strong>the</strong><br />
KRRP, suggest <strong>the</strong>re are o<strong>the</strong>r reasons ra<strong>the</strong>r than a lack of<br />
ability as to why targets are not rout<strong>in</strong>ely set. In simple<br />
terms, it may be diffi cult, time consum<strong>in</strong>g and expensive.<br />
Particularly if <strong>the</strong> ‘ultimate’ approach to target sett<strong>in</strong>g,<br />
namely sett<strong>in</strong>g quantitative targets and formulat<strong>in</strong>g expectations<br />
as hypo<strong>the</strong>ses that can be evaluated statistically<br />
to provide objective evaluations, is pursued (Toth and<br />
Anderson, 1998). Such an approach requires an extensive<br />
<strong>in</strong>formation base on <strong>the</strong> river <strong>in</strong> question. Where this is<br />
absent, ga<strong>the</strong>r<strong>in</strong>g such data may not be cost effective, especially<br />
<strong>in</strong> <strong>the</strong> case of reach-scale projects, s<strong>in</strong>ce understand<strong>in</strong>g<br />
a complex system could easily take several years and<br />
exceed <strong>the</strong> lifetime of <strong>the</strong> project (Kondolf, 1998). As a<br />
result, a more pragmatic approach may be to set more<br />
qualitative targets, although it should be recognised that<br />
even qualitative targets may be evaluated <strong>in</strong> a rigorous<br />
manner. Toth and Anderson (1998) suggest that if qualitative<br />
expectations are to be used as success criteria <strong>the</strong>y must<br />
be expressed <strong>in</strong> an objective scientifi c manner relative to<br />
<strong>the</strong> restoration goal. Assessments of <strong>the</strong> presence or absence<br />
of species, for example, are highly sensitive to sampl<strong>in</strong>g<br />
effort (see Section 8.2.6) and an <strong>in</strong>creased frequency of<br />
occurrence may be a response that cannot be used to differentiate<br />
restoration from habitat enhancement.<br />
Fur<strong>the</strong>r pragmatism to target sett<strong>in</strong>g may be to simply<br />
narrow <strong>the</strong> targets (Toth and Anderson, 1998), such as<br />
focus<strong>in</strong>g on a s<strong>in</strong>gle species, as adopted <strong>in</strong> <strong>the</strong> recovery of<br />
endangered species populations. Many such species, from<br />
mammals (e.g. Eurasian otter Lutra lutra; Ott<strong>in</strong>o and<br />
Giller, 2004) to fi sh (e.g. Colorado pikem<strong>in</strong>now Ptychocheilus<br />
lucius <strong>in</strong> <strong>the</strong> <strong>River</strong> Colorado – van Steeter<br />
and Pitlick, 1998; Roanoke logperch Perc<strong>in</strong>a rex <strong>in</strong> <strong>the</strong><br />
Nottoway & Roanoke rivers <strong>in</strong> Virg<strong>in</strong>ia – Rosenberger and<br />
Angermeier, 2003) to <strong>in</strong>vertebrates (e.g. freshwater pearl<br />
mussel Margaritifera margaritifera – Cosgrove and Hastie,<br />
2001; White-clawed crayfi sh Austropotamobius pallipes<br />
– Smith et al., 1996) have been subject to <strong>in</strong>-depth assessment<br />
of habitat requirements <strong>in</strong> river<strong>in</strong>e environments.<br />
After detailed research, specifi c habitat attributes may be<br />
readily defi ned for particular life history stages.<br />
Defi n<strong>in</strong>g habitat relationships between any organism<br />
and any particular physical parameter, especially fl ow, has
<strong>Uncerta<strong>in</strong>ty</strong> Surround<strong>in</strong>g <strong>the</strong> Ecological Targets and Response of <strong>River</strong> and Stream <strong>Restoration</strong> 145<br />
been exploited by <strong>the</strong> Instream Flow Incremental Methodology<br />
(IFIM) (Gore and Judy, 1981; Bovee, 1982) and its<br />
underly<strong>in</strong>g model, <strong>the</strong> Physical Habitat Simulation Model<br />
(PHABSIM) (see Appendix 8.1). PHABSIM provides a<br />
potentially powerful tool that may enable specifi c targets<br />
to be set for particular species where enough <strong>in</strong>formation<br />
of <strong>the</strong> habitat relationships of <strong>the</strong> species concerned exists.<br />
Specifi c sets of habitat relationships, which may feed <strong>in</strong>to<br />
restoration efforts, are particularly well advanced for charismatic<br />
and/or commercially valuable species such as salmonid<br />
fi shes, particularly <strong>in</strong> <strong>the</strong> United States, where<br />
detailed and structured attempts at relatively large scale<br />
habitat restoration for salmonids has been go<strong>in</strong>g on for<br />
40 years or more (White & Brynildson, 1967; Wesche,<br />
1985) with a wealth of detailed guides and manuals on<br />
<strong>the</strong> web (e.g. http://www.wildfi sh.montana.edu/resources/<br />
mammals.asp).<br />
In some cases <strong>the</strong> use of <strong>in</strong>dicators such as fi sh (or birds<br />
or mammals) at <strong>the</strong> top of <strong>the</strong> food web may at least <strong>in</strong>dicate<br />
<strong>the</strong> direction and strength of <strong>the</strong> response of lower<br />
trophic levels as a result of habitat restoration. This has<br />
<strong>the</strong> basic assumption that targeted improvements <strong>in</strong> higher<br />
trophic levels as a result of improvements <strong>in</strong> habitat diversity<br />
are also likely to have led to a response <strong>in</strong> <strong>the</strong> diversity<br />
of lower trophic levels (i.e. species diversity is related to<br />
habitat diversity; Gorman and Karr, 1978). Monitor<strong>in</strong>g<br />
keystone species may not always be appropriate, however,<br />
as it may not provide enough <strong>in</strong>formation about o<strong>the</strong>r<br />
important processes such as organic matter spiral<strong>in</strong>g and<br />
ground–surface water <strong>in</strong>teractions (Tockner and Schiemer,<br />
1997).<br />
Never<strong>the</strong>less, where multiple measures are taken, <strong>the</strong><br />
multiple expectations generated may perversely lead to an<br />
ambiguous evaluation of restoration success if all expectations<br />
are not met (Toth and Anderson, 1998). In such a<br />
case, ei<strong>the</strong>r all expectations need to be judged collectively,<br />
for example by track<strong>in</strong>g a cumulative count of expectations<br />
that are achieved over time, or perhaps expectations<br />
could also be weighted accord<strong>in</strong>g to set priorities. Targets<br />
that can be expressed as constants or thresholds are<br />
perhaps easiest to evaluate but probably will be limited to<br />
a few attributes, for example <strong>in</strong> many systems enough<br />
<strong>in</strong>formation is available to formulate constant or threshold<br />
expectations for species richness. This approach is readily<br />
employed for <strong>in</strong>vertebrates, with <strong>the</strong> recent development<br />
of numerous assessment criteria particularly <strong>in</strong> relation to<br />
fl ow at <strong>the</strong> assemblage level (Appendix 8.2).<br />
Due to <strong>the</strong> natural spatial and temporal variability that<br />
is associated with <strong>the</strong> scale of ecosystem restoration, Toth<br />
and Anderson (1998) suggest many targets need to be<br />
expressed as ranges, although <strong>the</strong> breadth of <strong>the</strong> range that<br />
is used for restoration expectations cannot be so great as<br />
to mask or preclude evaluation of restoration success. For<br />
example, <strong>the</strong> expectation of small fi sh densities of 0.7<br />
to 1.3 fi sh per square foot of restored marshes on <strong>the</strong><br />
Kissimmee fl oodpla<strong>in</strong> refl ected sampl<strong>in</strong>g variability whilst<br />
still provid<strong>in</strong>g a useful range for evaluat<strong>in</strong>g restoration<br />
success. With h<strong>in</strong>dsight, our own experiences on small<br />
rivers <strong>in</strong> <strong>the</strong> United K<strong>in</strong>gdom show that such targets could<br />
have been readily set. For example, at several sites on <strong>the</strong><br />
Misbourne, <strong>the</strong> fi sh community responded <strong>in</strong> a ra<strong>the</strong>r predictable<br />
way to <strong>the</strong> <strong>in</strong>crease <strong>in</strong> fl ow, with one assemblage<br />
replaced by ano<strong>the</strong>r as habitat conditions changed (Appendix<br />
8.3).<br />
Target sett<strong>in</strong>g <strong>in</strong> large and complex schemes, especially<br />
<strong>the</strong> KRRP, but not <strong>in</strong> smaller schemes is <strong>the</strong> <strong>in</strong>verse of that<br />
expected. Perhaps it is simply that sett<strong>in</strong>g mean<strong>in</strong>gful<br />
targets and goals may be diffi cult, time consum<strong>in</strong>g and<br />
expensive. In large schemes <strong>in</strong> which huge <strong>in</strong>vestment of<br />
time and resources is made, it is prudent that every effort<br />
is made to reduce <strong>the</strong> uncerta<strong>in</strong>ty of <strong>the</strong> outcome, particularly<br />
where this relates to reducible ignorance <strong>in</strong> <strong>the</strong> form<br />
of <strong>in</strong>creas<strong>in</strong>g <strong>the</strong> knowledge base. Conversely, smaller<br />
schemes, where <strong>the</strong> risks of failure may not be as great and<br />
where resources are fewer and tend to be directed at ‘do<strong>in</strong>g<br />
restoration’, may embrace all aspects of uncerta<strong>in</strong>ty.<br />
In general, contrary to <strong>the</strong> conclusions of Downs and<br />
Sk<strong>in</strong>ner (2002), we suggest that ecologists may set targets<br />
as readily and as well as eng<strong>in</strong>eers given <strong>the</strong> opportunity,<br />
especially s<strong>in</strong>ce Stewardson and Ru<strong>the</strong>rfurd (Chapter 5)<br />
argue that <strong>the</strong>re may be unreasonable confi dence <strong>in</strong> <strong>the</strong><br />
target sett<strong>in</strong>g ability of geomorphologists, who tend to rely<br />
on uncalibrated (with real fi eld data) <strong>the</strong>oretical models<br />
dogged by poor estimation of variance.<br />
8.2.6 Time Scales and Trajectories<br />
In any restoration, consideration should be given to <strong>the</strong><br />
time scale for recovery or restoration, which is <strong>in</strong>fl uenced<br />
by how far removed <strong>the</strong> system is from <strong>the</strong> goal and targets<br />
set, and <strong>the</strong> trajectory and pathway(s) that may be undertaken<br />
(Gregory and Downs, Chapter 13). Anand and<br />
Desrochers (2004) observe that <strong>the</strong> lack of long term<br />
observational studies <strong>in</strong> virtually any ecological system<br />
often compels researchers to rely on <strong>the</strong>oretical models to<br />
speculate upon <strong>the</strong> trajectory that will lead <strong>the</strong> damaged<br />
system to <strong>the</strong> desired state. Accord<strong>in</strong>g to complex systems<br />
<strong>the</strong>ory, which <strong>in</strong>corporates concepts such as chaos, <strong>the</strong><br />
nature and direction of a trajectory is governed by <strong>the</strong><br />
system’s attractor, which may be thought of as its dest<strong>in</strong>ation.<br />
It is hoped that <strong>the</strong> dest<strong>in</strong>ation and desired state are<br />
one and <strong>the</strong> same th<strong>in</strong>g or <strong>the</strong> restoration attempt may be<br />
doomed at <strong>the</strong> outset. Anand and Desrochers (2004)<br />
eloquently illustrate <strong>the</strong> different types of attractor, with
146 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
<strong>the</strong> simplest be<strong>in</strong>g <strong>the</strong> progression of a system to <strong>the</strong> same<br />
stable state regardless of <strong>in</strong>itial conditions, analogous to a<br />
climax state of vegetational succession. Where a system<br />
has more than one attractor, it may cycle between two<br />
alternative states. However, systems may have any number<br />
of attract<strong>in</strong>g states and <strong>the</strong> system may progress to a particular<br />
‘bas<strong>in</strong> of attraction’ depend<strong>in</strong>g on <strong>the</strong> start<strong>in</strong>g po<strong>in</strong>t<br />
or <strong>the</strong> <strong>in</strong>itial conditions at which restoration started. This<br />
offers <strong>the</strong> possibility of multiple alternative stable states,<br />
which proved to be highly <strong>in</strong>fl uential <strong>in</strong> <strong>the</strong> restoration of<br />
shallow lakes (Figure 8.2, Scheffer et al., 1993). Ecological<br />
stability itself is also a diffi cult concept to grasp, as<br />
large spatial and temporal variation <strong>in</strong> populations of even<br />
key species tuned <strong>in</strong> to <strong>the</strong>ir life cycles or disturbance<br />
events (Figure 8.3) may mask underly<strong>in</strong>g trends. Stability<br />
is thus best judged over decades ra<strong>the</strong>r than from one year<br />
to <strong>the</strong> next.<br />
To <strong>the</strong> best of our knowledge such concepts have yet to<br />
be explored <strong>in</strong> rivers, although river ‘types’ are <strong>the</strong> basis<br />
of <strong>the</strong> Rosgen classifi cation used extensively and not<br />
without controversy <strong>in</strong> river restoration (Malakoff, 2004).<br />
Notable gaps (i.e. unquantifi ed uncerta<strong>in</strong>ties) <strong>in</strong> our conceptual<br />
understand<strong>in</strong>g of ecological dynamics with<strong>in</strong><br />
rivers appear to constra<strong>in</strong> any ability to predict how long<br />
nature will take to return <strong>the</strong> system to a specifi ed historic<br />
condition or even what is to be expected.<br />
Y<br />
X<br />
8.3 DEVISING A MONITORING PROGRAMME<br />
8.3.1 Basic Design<br />
To m<strong>in</strong>imise ecological uncerta<strong>in</strong>ty, monitor<strong>in</strong>g needs to<br />
be undertaken prior to <strong>the</strong> work <strong>in</strong> <strong>the</strong> form of collection<br />
of a robust set of basel<strong>in</strong>e data (see previously). If <strong>the</strong><br />
pre-restoration system is not fully understood, <strong>the</strong> results<br />
of <strong>the</strong> scheme cannot be correctly assessed or evaluated.<br />
Use of reference reaches (see previously) may help set<br />
targets and, to monitor progress aga<strong>in</strong>st those targets, a<br />
suitable monitor<strong>in</strong>g programme needs to be devised. Monitor<strong>in</strong>g<br />
of controls (e.g. unaltered reaches) undoubtedly<br />
helps reduce uncerta<strong>in</strong>ty, as this should help dist<strong>in</strong>guish<br />
what degree of change is due to natural <strong>in</strong>ter-annual variation<br />
and what is due to <strong>the</strong> restoration works (Stewardson<br />
and Ru<strong>the</strong>rfurd, Chapter 5).<br />
Choos<strong>in</strong>g which measurements of biotic and abiotic<br />
patterns and processes to monitor is <strong>the</strong>n critical, as <strong>the</strong>se<br />
need to be relevant to <strong>the</strong> goals of <strong>the</strong> project and <strong>the</strong><br />
targets set, and must be able to be l<strong>in</strong>ked to progress<br />
towards those targets. Where <strong>the</strong> target of restoration is a<br />
particular species, this may lead to very specifi c targets<br />
and <strong>the</strong> desire to monitor a few aspects. However, several<br />
workers have highlighted <strong>the</strong> need for wider surveillance<br />
ra<strong>the</strong>r than aim<strong>in</strong>g monitor<strong>in</strong>g at one aspect of <strong>the</strong> ecosystem.<br />
For example, Boon (1998) suggested that fi sheries-<br />
(Left) A po<strong>in</strong>t attractor. The same f<strong>in</strong>al state is reached <strong>in</strong> both <strong>in</strong>stances X and Y, although <strong>the</strong><br />
time taken to achieve this (shown by differ<strong>in</strong>g arrow lengths) varies depend<strong>in</strong>g on <strong>the</strong> location<br />
of <strong>the</strong> start<strong>in</strong>g position on <strong>the</strong> restoration trajectory.<br />
(Right) A dual attractor. This functions like a pendulum sw<strong>in</strong>g<strong>in</strong>g between two magnets, or<br />
alternative stable states (A and B). In this <strong>in</strong>stance <strong>the</strong> restoration time scales are <strong>the</strong> same,<br />
although <strong>the</strong> f<strong>in</strong>al state reached may depend on o<strong>the</strong>r ecological factors, for example composition<br />
of <strong>the</strong> community at <strong>the</strong> po<strong>in</strong>t when restoration is commenced.<br />
Figure 8.2 Types of attractors (Reproduced from M. Anand and R. Desrochers (2004) ‘Quantifi cation of <strong>Restoration</strong> Success Us<strong>in</strong>g<br />
Complex Systems Concepts and Models’ <strong>Restoration</strong> Ecology 12 (1), 117–123, with k<strong>in</strong>d permisison from Blackwell Publish<strong>in</strong>g.)<br />
Y<br />
X
<strong>Uncerta<strong>in</strong>ty</strong> Surround<strong>in</strong>g <strong>the</strong> Ecological Targets and Response of <strong>River</strong> and Stream <strong>Restoration</strong> 147<br />
(A) The potential effects of disturbance events (d) on ecosystem structure over time.<br />
Disturbance events could constitute anyth<strong>in</strong>g from pollution or alien species <strong>in</strong>troduction to <strong>the</strong><br />
impacts of restoration e.g. dredg<strong>in</strong>g, and recovery times differ accord<strong>in</strong>gly.<br />
(B) Variation <strong>in</strong> flow at two temporal scales.<br />
Figure 8.3 The effects of extensive spatial and temporal scale variation on ecological stability (Reproduced from P. S. White and<br />
J. L. Walker (1997) ‘Approximat<strong>in</strong>g nature’s variation: select<strong>in</strong>g and us<strong>in</strong>g reference <strong>in</strong>formation <strong>in</strong> restoration ecology’ <strong>Restoration</strong><br />
Ecology 5, 338–349, with k<strong>in</strong>d permission from Blackwell Publish<strong>in</strong>g.)<br />
led restoration projects should take a comprehensive<br />
ecosystem approach. Tockner et al. (1998) reached a<br />
similar conclusion ‘A key challenge <strong>in</strong> <strong>the</strong> evaluation of<br />
<strong>the</strong> effects of restoration is <strong>the</strong> development and test<strong>in</strong>g of<br />
an appropriate monitor<strong>in</strong>g scheme, which has to <strong>in</strong>clude<br />
a wide range of physical, chemical, geomorphic, and ecological<br />
parameters.’ Monitor<strong>in</strong>g a range of variables provides<br />
more <strong>in</strong>formation about <strong>the</strong> system response and <strong>the</strong><br />
potential mechanisms beh<strong>in</strong>d <strong>the</strong> response of <strong>the</strong> target<br />
variable or species.<br />
Consequently, a number of groups across <strong>the</strong> variety of<br />
trophic levels are typically used as <strong>in</strong>dicators of ecological<br />
status and response. For example, benthic <strong>in</strong>vertebrates,<br />
fi sh and aquatic macrophytes were selected as <strong>in</strong>dicators<br />
of <strong>the</strong> ecological status of rivers <strong>in</strong> <strong>the</strong> European<br />
Union Water Framework Directive (EU, 2000), and benthic
148 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
<strong>in</strong>vertebrates, fi sh, plankton, birds, amphibians and terrestrial<br />
and aquatic plants have been used on <strong>the</strong> project<br />
on <strong>the</strong> Austrian Danube (Buijse et al., 2002). For evaluat<strong>in</strong>g<br />
biodiversity of river–fl oodpla<strong>in</strong> complexes, a number<br />
of species-specifi c groups that differ <strong>in</strong> <strong>the</strong>ir responses to<br />
hydrological connectivity, water quality and habitat heterogeneity<br />
should be considered (Buijse et al., 2002;<br />
Tockner et al., 1999). The absence of certa<strong>in</strong> species may<br />
also be a good <strong>in</strong>dication of ecological deterioration (EU,<br />
2000). For example, <strong>the</strong> loss of long-distance migratory<br />
fi sh (salmonids, coregonids, shads and sturgeons) <strong>in</strong>dicates<br />
disruption of longitud<strong>in</strong>al connectivity or a deterioration<br />
of spawn<strong>in</strong>g/nursery areas.<br />
8.3.2 Sampl<strong>in</strong>g Requirements<br />
Brooks et al. (2002) po<strong>in</strong>ted out that many commonly<br />
used river restoration techniques may never have been<br />
scientifi cally tested and, <strong>in</strong> a demonstration of <strong>the</strong> need<br />
for rigorous scientifi c test<strong>in</strong>g, carried out a fi eld experiment<br />
designed to mimic restoration to <strong>in</strong>vestigate <strong>the</strong><br />
importance of habitat heterogeneity for macro<strong>in</strong>vertebrates.<br />
The results showed that although a diversity of<br />
habitats was required, macro<strong>in</strong>vertebrate populations were<br />
actually more sensitive to <strong>in</strong>dividual site conditions at<br />
each riffl e than to <strong>the</strong> heterogeneity treatments. This led<br />
to <strong>the</strong> conclusion that <strong>the</strong> extremely high variability<br />
between replicate riffl es meant that a monitor<strong>in</strong>g programme<br />
for localised restoration projects would be<br />
unlikely to detect gradual shifts <strong>in</strong> community structure<br />
until <strong>the</strong> differences between <strong>the</strong> reference and treatment<br />
sites were extreme. Brooks et al. <strong>the</strong>n suggested that <strong>in</strong>novative<br />
measurement of o<strong>the</strong>r parameters, such as ecosystem<br />
function variables (e.g. production, respiration,<br />
decomposition), may be more appropriate <strong>in</strong>dicators of<br />
change at local scales.<br />
This illustrates <strong>the</strong> age-old problem faced by ecologists<br />
of huge spatial and temporal variation <strong>in</strong> <strong>the</strong>ir subject<br />
matter (Figure 8.3). However, ra<strong>the</strong>r than look<strong>in</strong>g to o<strong>the</strong>r<br />
variables that may be technically diffi cult to measure,<br />
parameterise and ultimately <strong>in</strong>terpret, <strong>the</strong> only option to<br />
overcome variability is to devise appropriate sampl<strong>in</strong>g<br />
regimes of suffi cient <strong>in</strong>tensity and frequency and to use<br />
robust methods. In <strong>the</strong> case of <strong>in</strong>vertebrates, methods that<br />
allow changes to be evaluated on a community ra<strong>the</strong>r than<br />
species-specifi c level are likely to be useful (Appendix<br />
8.2). For fi sh, Bohl<strong>in</strong> et al. (1990) suggested three precision<br />
classes for fi sheries studies depend<strong>in</strong>g on <strong>the</strong> nature<br />
of <strong>the</strong> change that needs to be detected, i.e. a factor as<br />
small as 1.2, 1.5 or as large as 2.0, with about 80% probability<br />
when us<strong>in</strong>g a 5% signifi cance level. Unfortunately,<br />
when <strong>the</strong> number of fi sh is relatively low with even moder-<br />
ate variation around a mean value, even <strong>the</strong> lowest precision<br />
class tends to require a high sampl<strong>in</strong>g effort. On <strong>the</strong><br />
<strong>River</strong> Lambourn, UK, dur<strong>in</strong>g an attempt to monitor <strong>the</strong><br />
status of bullhead Cottus gobio, one of <strong>the</strong> species for<br />
which <strong>the</strong> river was designated as a site of <strong>in</strong>ternational<br />
conservation <strong>in</strong>terest (candidate Special Area of Conservation<br />
[cSAC]), it was calculated that twenty-seven 100 m<br />
sites would need to be surveyed over <strong>the</strong> 21 km river<br />
to detect change on a population level (Perrow and<br />
Toml<strong>in</strong>son, 2002). Us<strong>in</strong>g <strong>the</strong> same calculations to estimate<br />
<strong>the</strong> number of po<strong>in</strong>ts required <strong>in</strong> PASE (see Appendix 8.3)<br />
at each site, as few as 39 po<strong>in</strong>ts were required where <strong>the</strong><br />
population of bullheads was dense, but up to 150 po<strong>in</strong>ts<br />
were required where bullheads were at low density. The<br />
latter probably exceeds <strong>the</strong> number of po<strong>in</strong>ts that can be<br />
<strong>in</strong>dependently sampled, <strong>in</strong> that <strong>the</strong> po<strong>in</strong>ts are far enough<br />
apart to not <strong>in</strong>fl uence each o<strong>the</strong>r, at any particular site.<br />
Where such factors cannot be calculated <strong>in</strong> advance, rules<br />
of thumb are often generated and <strong>in</strong> <strong>the</strong> case of PASE this<br />
is often regarded as 50 po<strong>in</strong>ts (Garner, 1997). Below this,<br />
<strong>the</strong> confi dence <strong>in</strong> <strong>the</strong> estimates may be reduced, mak<strong>in</strong>g<br />
<strong>the</strong> response of <strong>the</strong> fi sh community to restoration diffi cult,<br />
if not impossible, to evaluate and <strong>in</strong>terpret. Too few<br />
samples may have been ano<strong>the</strong>r factor contribut<strong>in</strong>g to <strong>the</strong><br />
lack of a signifi cant impact of <strong>the</strong> <strong>in</strong>stallation of channel<br />
enhancements (i.e. artifi cial riffl es and fl ow defl ectors)<br />
upon fi sh abundance and species richness <strong>in</strong> low energy<br />
lowland alluvial channels (Pretty et al., 2003).<br />
The use of standardised methods and techniques <strong>in</strong> a<br />
‘one size fi ts all approach’ may also be a particular<br />
problem. Standardised approaches are particularly liked<br />
by statutory organisations and whilst <strong>the</strong> desire for comparability<br />
is understandable, many of <strong>the</strong> standard<br />
approaches have been designed for objectives o<strong>the</strong>r than<br />
monitor<strong>in</strong>g <strong>the</strong> response of a restoration scheme. Ultimately,<br />
such approaches may simply provide qualitative<br />
targets with no basis for statistical comparison. In relation<br />
to <strong>the</strong> recovery enhancement of <strong>the</strong> <strong>River</strong> Misbourne, UK<br />
(Appendix 8.3), standard <strong>River</strong> Corridor Survey (RCS),<br />
<strong>River</strong> Habitat Survey (RHS) and <strong>River</strong> Macrophyte Survey<br />
(RMS) were used at each site at not <strong>in</strong>considerable effort.<br />
However, all <strong>the</strong>se methods simply described and mapped<br />
<strong>the</strong> nature of <strong>the</strong> vegetation present and did not rout<strong>in</strong>ely<br />
provide quantitative measures that could be evaluated statistically.<br />
With <strong>the</strong> likely difference <strong>in</strong> hydrological regime<br />
between <strong>the</strong> different monitor<strong>in</strong>g sites along <strong>the</strong> course of<br />
<strong>the</strong> river, <strong>the</strong> response of each site had to be evaluated<br />
<strong>in</strong>dependently. Thus, <strong>the</strong> use of <strong>the</strong> standard fi sheries<br />
survey through depletion fi sh<strong>in</strong>g result<strong>in</strong>g <strong>in</strong> n = 1 on each<br />
sampl<strong>in</strong>g occasion could provide no more than a subjective<br />
comparison over time as fl ow was recovered. For this<br />
reason, a more specifi c sampl<strong>in</strong>g regime us<strong>in</strong>g PASE and
<strong>Uncerta<strong>in</strong>ty</strong> Surround<strong>in</strong>g <strong>the</strong> Ecological Targets and Response of <strong>River</strong> and Stream <strong>Restoration</strong> 149<br />
number of replicate samples with<strong>in</strong> each site was also used<br />
(Appendix 8.3).<br />
Even where <strong>the</strong> sampl<strong>in</strong>g regime has been carefully and<br />
perhaps specifi cally developed, particularly amongst fi sheries<br />
studies, it is essential to be clear that <strong>the</strong> effects of<br />
<strong>the</strong> scheme are at a population level with an <strong>in</strong>crease (or<br />
decrease) <strong>in</strong> <strong>the</strong> population, and not simply a change <strong>in</strong><br />
distribution of <strong>the</strong> fi sh present. Whilst <strong>the</strong> latter may still<br />
be viewed as a benefi cial response, as it illustrates that <strong>the</strong><br />
restoration works have produced <strong>the</strong> required sort of<br />
habitat, this may not have actually achieved <strong>the</strong> target<br />
set.<br />
8.3.3 Evaluation and Deviations from<br />
Expected Outcomes<br />
Put simply, <strong>the</strong> measure of success of any scheme is<br />
whe<strong>the</strong>r <strong>the</strong> targets have been met after a specifi ed time.<br />
Even where <strong>the</strong> plann<strong>in</strong>g and science were sound, poor<br />
implementation may have jeopardized <strong>the</strong> success of a<br />
restoration project. For example, <strong>in</strong> 1991 an attempt was<br />
made to restore vegetation on mud fl ats border<strong>in</strong>g <strong>the</strong><br />
Milwaukee <strong>River</strong>, USA, which had been previously submerged<br />
due to a dam (Drezner, 2004). Although <strong>the</strong> area<br />
was <strong>in</strong>itially seeded with native plants, and non-native<br />
plants were regularly pulled up, an oversight result<strong>in</strong>g <strong>in</strong><br />
<strong>the</strong> failure to remove Reed canary grass Phalaris arund<strong>in</strong>acea<br />
resulted <strong>in</strong> its dom<strong>in</strong>ation <strong>in</strong> certa<strong>in</strong> areas. When<br />
<strong>the</strong> area was surveyed 11 years later, on average 89% of<br />
plant stems per quadrat were non-native. The reasons<br />
beh<strong>in</strong>d such problems are not necessarily simple carelessness.<br />
Frequently, budgetary decisions relat<strong>in</strong>g to a whole<br />
project have to be made at <strong>the</strong> plann<strong>in</strong>g stage, which<br />
means that <strong>the</strong> fi nancial fl exibility required to deal with<br />
additional and unpredictable implementation issues may<br />
simply not be available.<br />
Streams and rivers have an <strong>in</strong>herent ability for natural<br />
recovery, which has been widely exploited <strong>in</strong> restoration<br />
(Brooks and Shields, 1996; Downs and Sk<strong>in</strong>ner, 2002).<br />
Consequently, <strong>the</strong>re seems to be little concern of whe<strong>the</strong>r<br />
a river will be restored once <strong>the</strong> limit<strong>in</strong>g factors are<br />
removed but ra<strong>the</strong>r when. Thus, once habitat restoration<br />
has taken place, ecological communities are left to recolonise.<br />
The rate of recovery follow<strong>in</strong>g a disturbance is<br />
dependent upon <strong>the</strong> resilience of <strong>the</strong> community to <strong>the</strong><br />
perturbation <strong>in</strong> <strong>the</strong> fi rst place as well as <strong>the</strong> rate of colonisation.<br />
Many <strong>in</strong>vertebrate species are highly resilient to<br />
disturbance through physical adaptations and behavioural<br />
change. Magoulik and Kobza (2003) concluded that many<br />
seek refuge from disturbance and/or have adaptations that<br />
provide refuge, whilst Townsend (1989) highlighted <strong>the</strong><br />
critical role played by refugia as sources of recolonisation<br />
after spates, and <strong>the</strong>refore as buffers aga<strong>in</strong>st disturbance.<br />
Important refugia <strong>in</strong>clude low fl ow areas (W<strong>in</strong>terbottom<br />
et al., 1997).<br />
The extent and <strong>in</strong>tensity of channel modifi cation (i.e.<br />
disturbance) is thought to have an impact on <strong>the</strong> time scale<br />
of recovery. Tikkanen et al. (1994) found that recolonisation<br />
of <strong>in</strong>vertebrates occurred rapidly, with<strong>in</strong> ten days,<br />
after a small-scale rehabilitation scheme. In contrast, on<br />
<strong>the</strong> <strong>River</strong> Rib where major habitat reconstruction was<br />
undertaken, although <strong>the</strong>re was rapid colonisation by<br />
some macro<strong>in</strong>vertebrates, <strong>the</strong> community took almost two<br />
years before show<strong>in</strong>g <strong>in</strong>dications of stabilisation. Fur<strong>the</strong>rmore,<br />
Niemi et al. (1990) found that many systems took<br />
more than fi ve years to recover to <strong>the</strong> desired ‘endpo<strong>in</strong>ts’,<br />
and <strong>the</strong>y concluded that <strong>the</strong> longest recovery times are<br />
associated with disturbance that leads to long term alterations<br />
<strong>in</strong> physical habitat. For example, <strong>the</strong> effects of an<br />
episodic pollution event are relatively short lived whilst<br />
recovery follow<strong>in</strong>g dredg<strong>in</strong>g takes much longer s<strong>in</strong>ce geomorphic<br />
channel readjustment is often slow. Indeed work<br />
carried out by Laasonon et al. (1998) on <strong>the</strong> recovery of<br />
macro<strong>in</strong>vertebrates follow<strong>in</strong>g stream habitat restoration<br />
works – which took <strong>the</strong> form of boulder dams, fl ow defl ectors,<br />
excavations and channel enlargements – <strong>in</strong>dicates<br />
that even after 16 years abundances of <strong>in</strong>vertebrates were<br />
still lower than <strong>in</strong> natural streams. Abundances of shredders<br />
were particularly low <strong>in</strong> recently restored streams <strong>in</strong><br />
comparison to streams restored a number of years previously,<br />
although even some streams that had been restored<br />
8 or 16 years ago still conta<strong>in</strong>ed relatively sparse shredder<br />
populations. Fur<strong>the</strong>r work (Muotka et al., 2002) highlighted<br />
<strong>the</strong> importance of aquatic mosses, which are frequently<br />
uprooted dur<strong>in</strong>g restoration works, and which<br />
perform important ecosystem functions, e.g. <strong>the</strong> retention<br />
of fi ne particles and <strong>the</strong> provision of fl ow refugia for<br />
<strong>in</strong>vertebrates, which ultimately <strong>in</strong>fl uence <strong>the</strong> rate of recovery<br />
of <strong>the</strong> macro<strong>in</strong>vertebrate community. At present <strong>the</strong><br />
recovery rates of aquatic mosses are unknown, mak<strong>in</strong>g it<br />
diffi cult to predict <strong>the</strong> responses of macro<strong>in</strong>vertebrate<br />
communities <strong>in</strong> <strong>the</strong>se types of systems.<br />
The time scales implicated <strong>in</strong> <strong>the</strong>se studies <strong>in</strong>dicate that<br />
all too often schemes are probably assessed too quickly,<br />
at a po<strong>in</strong>t dur<strong>in</strong>g <strong>the</strong> colonisation process ra<strong>the</strong>r that at <strong>the</strong><br />
end po<strong>in</strong>t target, even if this was defi ned accurately <strong>in</strong> <strong>the</strong><br />
fi rst place. Short term fl uctuations <strong>in</strong> species populations<br />
are to be expected before a longer-term stable equilibrium<br />
is reached.<br />
There is a wide range of factors <strong>in</strong>fl uenc<strong>in</strong>g <strong>the</strong> rate of<br />
recolonsiation that may simply over-ride <strong>the</strong> level of disturbance<br />
undertaken, <strong>in</strong>clud<strong>in</strong>g <strong>the</strong> source of colonists<br />
with<strong>in</strong> <strong>the</strong> river itself, <strong>the</strong> proximity of watercourses<br />
nearby for aerial colonsiation by w<strong>in</strong>ged adults and <strong>the</strong>
150 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
life cycle of <strong>the</strong> <strong>in</strong>vertebrates concerned and <strong>the</strong>ir habitat<br />
preferences. On <strong>the</strong> Rib, <strong>the</strong>re were marked differences <strong>in</strong><br />
<strong>the</strong> rates of recovery of different species and four groups<br />
were readily recognised: early colonists, medium term<br />
colonists, seasonal colonists and long term colonists<br />
(Figure 8.4). The Olive dun mayfl y Baetis rhodani was a<br />
typical early colonist be<strong>in</strong>g an active swimmer as well as<br />
a typical component of <strong>the</strong> drift. Whilst <strong>the</strong> Riffl e beetle<br />
Elmis aenea also used <strong>the</strong> drift, as a resident of fast fl ow<strong>in</strong>g<br />
areas it is adapted to be<strong>in</strong>g washed away and thus tends to<br />
colonise more slowly. Blackfl y larvae Simulium spp. are<br />
typically abundant <strong>in</strong> <strong>the</strong> spr<strong>in</strong>g and only tend to colonise<br />
newly available areas at this time. A typical long term<br />
colonist <strong>in</strong> this system was <strong>the</strong> predatory caseless caddis<br />
Rhyacophila dorsalis, which relies on <strong>the</strong> adult w<strong>in</strong>ged<br />
stage to colonise, as <strong>the</strong> larvae are adapted to fast fl ows<br />
Figure 8.4 The response of <strong>the</strong> mayfl y Baetis rhodani and <strong>the</strong> riffl e beetle Elmis aenea at two re-wetted sites on <strong>the</strong> <strong>River</strong> Rib, UK.
<strong>Uncerta<strong>in</strong>ty</strong> Surround<strong>in</strong>g <strong>the</strong> Ecological Targets and Response of <strong>River</strong> and Stream <strong>Restoration</strong> 151<br />
and are unable to tolerate slower fl ows between suitable<br />
habitat patches.<br />
Like <strong>in</strong>vertebrates, fi sh are often assumed to be capable<br />
of rapid colonisation and, aga<strong>in</strong> like <strong>in</strong>vertebrates, <strong>the</strong><br />
speed and recovery of colonisation may be primarily<br />
dependent on <strong>the</strong> geographical isolation of <strong>the</strong> communities<br />
and <strong>the</strong> proximity of potential colonisers. Both downstream<br />
and upstream movement may be readily undertaken<br />
provided <strong>the</strong>re are no barriers to <strong>the</strong> latter. What constitutes<br />
a barrier varies enormously. Bullhead, a small<br />
benthic fi sh is effectively limited by barriers of just 10 cm<br />
(Utz<strong>in</strong>ger et al., 1998), whereas Barbel Barbus barbus a<br />
large (to 1 m) powerful cypr<strong>in</strong>id <strong>in</strong> European rivers has<br />
been recorded mak<strong>in</strong>g movements of up to nearly 20 km,<br />
cross<strong>in</strong>g large weirs (to 2 m) <strong>in</strong> <strong>the</strong> process (Lucas and<br />
Bately, 1996). The ability of many anadromous salmonids<br />
(salmon and trout) to leap large obstacles is well known.<br />
Even though <strong>the</strong> restored area may have been recolonised,<br />
recruitment success can still fl uctuate ow<strong>in</strong>g to a wide<br />
range of factors not related to <strong>the</strong> scheme (e.g. temperature,<br />
predation of eggs and larvae etc). Such factors can<br />
affect both <strong>the</strong> abundance and distribution of <strong>in</strong>dividuals,<br />
and may signifi cantly slow <strong>the</strong> colonisation process.<br />
Moreover, <strong>the</strong> major structur<strong>in</strong>g forces of predation and<br />
competition may rage for some time before a stable confi<br />
guration is reached. In <strong>the</strong> Misbourne for example, sticklebacks<br />
may have out-competed o<strong>the</strong>r species for some<br />
time, even as physical conditions changed, simply as <strong>the</strong>y<br />
arrived fi rst and achieved high density as a result of <strong>the</strong><br />
abundance of potential nest sites and fl ow refugia <strong>in</strong> <strong>the</strong><br />
form of emergent and <strong>the</strong>n submerged vegetation. Only<br />
when <strong>the</strong>se were removed, did <strong>the</strong> community shift rapidly,<br />
with bullheads dom<strong>in</strong>at<strong>in</strong>g through rapid recruitment<br />
(Appendix 8.3).<br />
In general, vagaries <strong>in</strong> <strong>the</strong> abilities of different organisms<br />
to colonise <strong>in</strong>tuitively mean that this stage of <strong>the</strong><br />
project is highly vulnerable to <strong>the</strong> <strong>in</strong>fl uence of ecological<br />
uncerta<strong>in</strong>ty. Predict<strong>in</strong>g <strong>the</strong> likely pattern of colonisation<br />
may be problematic, due to both a limited understand<strong>in</strong>g<br />
of o<strong>the</strong>r populations <strong>in</strong> <strong>the</strong> local area, and a general lack<br />
of scientifi c knowledge. For example, <strong>the</strong> factors <strong>in</strong>fl uenc<strong>in</strong>g<br />
<strong>the</strong> colonisation of submerged macrophytes, which are<br />
of enormous ecological importance, are poorly understood.<br />
Clearly, better knowledge of <strong>the</strong> colonisation mechanisms<br />
and movement patterns is vital to understand <strong>the</strong><br />
processes beh<strong>in</strong>d restoration of river<strong>in</strong>e systems (Fenoglio<br />
et al., 2002) and allow a better understand<strong>in</strong>g of <strong>the</strong> time<br />
scale required for recovery to a stable end po<strong>in</strong>t. A lack of<br />
scientifi c knowledge is not necessarily easy to overcome,<br />
s<strong>in</strong>ce <strong>the</strong> situation at one site may differ substantially from<br />
ano<strong>the</strong>r and at present <strong>the</strong>re is a lack of longer term studies<br />
that may feed <strong>in</strong>to <strong>the</strong> knowledge base (Figure 8.1).<br />
8.3.4 Appraisal and Feedback<br />
The importance of appraisal has been highlighted by<br />
Anderson and Dugger (1998) who concluded that ‘Failure<br />
to evaluate projects not only precludes learn<strong>in</strong>g anyth<strong>in</strong>g<br />
about a particular restoration, but it also limits <strong>the</strong> opportunity<br />
for improv<strong>in</strong>g plans for future projects’. In our experience<br />
<strong>in</strong> <strong>the</strong> United K<strong>in</strong>gdom <strong>the</strong> number of projects that<br />
<strong>in</strong>corporate post-project appraisal (PPA) is <strong>in</strong>creas<strong>in</strong>g follow<strong>in</strong>g<br />
early efforts to raise <strong>the</strong> importance of this aspect<br />
(ECON, 1993). Review<strong>in</strong>g projects from 1991–1996,<br />
Holmes (1998) recorded that 53% of projects had undergone<br />
PPA, 23% of projects did not <strong>in</strong>corporate any form<br />
of PPA and for 25% of projects this <strong>in</strong>formation was not<br />
known. Today, partly as a result of <strong>the</strong> efforts of <strong>the</strong> RRP<br />
demonstration projects and subsequently <strong>the</strong> <strong>River</strong> <strong>Restoration</strong><br />
Centre (RRC) with<strong>in</strong> <strong>the</strong> European Centre for <strong>River</strong><br />
<strong>Restoration</strong> (see previously), we would suggest that perhaps<br />
only <strong>the</strong> smaller scale schemes do not rout<strong>in</strong>ely undertake<br />
PPA. As well as <strong>the</strong> importance of shared <strong>in</strong>formation,<br />
drivers have <strong>in</strong>cluded <strong>the</strong> need for organisations such as<br />
<strong>the</strong> Environment Agency to demonstrate cost effi ciency<br />
for <strong>the</strong>ir ‘bus<strong>in</strong>ess’ needs, and that <strong>the</strong> publicity mach<strong>in</strong>es<br />
of <strong>the</strong> organisations <strong>in</strong>volved like/need to have someth<strong>in</strong>g<br />
to show to <strong>the</strong> public and o<strong>the</strong>r <strong>in</strong>terested parties.<br />
However, <strong>the</strong>re is potentially little control over <strong>the</strong><br />
quality of <strong>the</strong> <strong>in</strong>formation produced, with accumulat<strong>in</strong>g<br />
ecological uncerta<strong>in</strong>ties <strong>in</strong> <strong>the</strong> earlier stages of <strong>the</strong> project<br />
potentially ultimately result<strong>in</strong>g <strong>in</strong> mislead<strong>in</strong>g <strong>in</strong>formation<br />
that could be used to <strong>in</strong>form future projects. This is a<br />
variant of <strong>the</strong> ‘Almost as bad as no evaluation are poorly<br />
planned efforts that waste limited resources while provid<strong>in</strong>g<br />
mean<strong>in</strong>gless or even mislead<strong>in</strong>g <strong>in</strong>formation’ warned<br />
by Anderson and Dugger (1998). It is diffi cult if not<br />
impossible to assess how much this type of situation<br />
occurs. Although ecological uncerta<strong>in</strong>ty may frequently<br />
occur, <strong>the</strong> concern is that it goes unrecognised, which<br />
could lead to <strong>in</strong>accurate results and mis<strong>in</strong>formation. If<br />
ecological uncerta<strong>in</strong>ty is recognised, and if <strong>the</strong> scheme<br />
results <strong>in</strong> a clear improvement, even if it is not exactly as<br />
predicted, this is perhaps less of a concern. However, <strong>the</strong>re<br />
is <strong>the</strong> possibility that <strong>the</strong> project may be held <strong>in</strong> less<br />
esteem or considered not as successful if targets are not<br />
met, and project fund<strong>in</strong>g and <strong>the</strong> prospect of cont<strong>in</strong>ued<br />
restoration may be jeopardised, even though <strong>the</strong> work may<br />
have ultimately had a positive impact on <strong>the</strong> ecology of<br />
<strong>the</strong> site. Schemes are beg<strong>in</strong>n<strong>in</strong>g to allow for this eventuality<br />
and are constructed to allow adaptive management if<br />
<strong>the</strong> scheme needs to be adjusted to obta<strong>in</strong> more desirable<br />
results.<br />
Even if appraisal is conducted, it is often not easy to<br />
evaluate success and for this reason <strong>the</strong>re is a major move
152 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
from river ecologists to establish a set of criteria for measur<strong>in</strong>g<br />
<strong>the</strong> ecological success of schemes (Palmer et al.,<br />
2005). It is hoped that this will fi nally be endorsed by <strong>the</strong><br />
United Nations Environmental Programme. The criteria<br />
specifi ed are that:<br />
<strong>the</strong> design should be based on a specifi ed guid<strong>in</strong>g image<br />
that a more dynamic, healthy river could exist at <strong>the</strong><br />
site;<br />
<strong>the</strong> river’s ecological condition must be measurably<br />
improved;<br />
<strong>the</strong> river system must be more self-susta<strong>in</strong><strong>in</strong>g and resilient<br />
to external perturbations so that only m<strong>in</strong>imal<br />
follow-up ma<strong>in</strong>tenance is needed.<br />
dur<strong>in</strong>g <strong>the</strong> construction phase, no last<strong>in</strong>g harm should<br />
be <strong>in</strong>fl icted on <strong>the</strong> ecosystem.<br />
both pre- and post-assessment must be completed and<br />
data made publicly available.<br />
The emphasis beh<strong>in</strong>d <strong>the</strong>se criteria is <strong>the</strong> desire to have<br />
projects that are funded and implemented <strong>in</strong> <strong>the</strong> name of<br />
ecological restoration evaluated accord<strong>in</strong>g to ecological<br />
<strong>in</strong>dicators of success ra<strong>the</strong>r than accord<strong>in</strong>g to o<strong>the</strong>r social,<br />
economic and <strong>in</strong>stitutional factors, such as cost effectiveness,<br />
stakeholder satisfaction and aes<strong>the</strong>tic/recreational<br />
value (Palmer et al., 2005). These criteria are fl exible<br />
enough to apply to both large scale major projects and<br />
smaller projects where complex and expensive design is<br />
not appropriate. They also encourage <strong>the</strong> view of restoration<br />
success as an adaptive process ra<strong>the</strong>r than a s<strong>in</strong>gle<br />
defi ned end po<strong>in</strong>t.<br />
8.4 CONCLUSIONS AND RECOMMENDATIONS<br />
Throughout this chapter we have sought to outl<strong>in</strong>e sources<br />
of unwanted uncerta<strong>in</strong>ty that ultimately reduce confi dence<br />
<strong>in</strong> restoration and h<strong>in</strong>der <strong>the</strong> development of a cost effective,<br />
repeatable approach. The latter can only be achieved<br />
by contribut<strong>in</strong>g to a shared knowledge base, <strong>the</strong> potential<br />
for which is now enormous with a number of dedicated<br />
restoration journals and even more specifi c websites. The<br />
approach adopted accepts that much uncerta<strong>in</strong>ty result<strong>in</strong>g<br />
from natural variation and sheer system complexity, especially<br />
<strong>in</strong> large systems, will rema<strong>in</strong> unknowable. Fortunately,<br />
this has not dampened <strong>the</strong> desire for restoration of<br />
large systems, which make greatest contribution to biodiversity,<br />
as recent <strong>in</strong>itiatives with<strong>in</strong> <strong>the</strong> European Union to<br />
restore rivers such as <strong>the</strong> Danube (Tockner and Schiemer,<br />
1997), <strong>the</strong> Rh<strong>in</strong>e (Buijse et al., 2002) and <strong>the</strong> Drava and<br />
Lech rivers <strong>in</strong> Austria (Mohl, 2004) testify.<br />
Indeed, <strong>in</strong>vestment <strong>in</strong> large systems such as <strong>the</strong> KRRP<br />
has illustrated that many uncerta<strong>in</strong>ties may be quantifi ed<br />
and overcome, with <strong>the</strong> accumulation of an extensive<br />
knowledge and basel<strong>in</strong>e data, ga<strong>the</strong>red over a number of<br />
years, which <strong>in</strong> turn required considerable resources. With<br />
such effort, predictions and targets can be detailed, quantitative<br />
and set to with<strong>in</strong> a realistic time scale. This quells<br />
<strong>the</strong> notion of <strong>the</strong> <strong>in</strong>ability of ecologists to set targets which<br />
has been raised as a particular source of uncerta<strong>in</strong>ty. Our<br />
experiences, especially <strong>in</strong> <strong>the</strong> United K<strong>in</strong>gdom, suggest<br />
that <strong>the</strong> lack of target sett<strong>in</strong>g cannot be attributed to a<br />
particular <strong>in</strong>ability to do so but ra<strong>the</strong>r appears to be more<br />
due to a lack of resources <strong>in</strong> small projects especially.<br />
With an <strong>in</strong>crease <strong>in</strong> post-project appraisal of whe<strong>the</strong>r <strong>the</strong><br />
project succeeded or failed, it is more diffi cult to justify<br />
not us<strong>in</strong>g targets at all. Project ‘culture’ thus appears to<br />
be chang<strong>in</strong>g for <strong>the</strong> better.<br />
There is probably a better knowledge base of <strong>the</strong> fauna–<br />
habitat <strong>in</strong>formation than is immediately apparent and this<br />
may be developed specifi cally where this is lack<strong>in</strong>g,<br />
although this may be expensive of time and resources.<br />
Habitat relationships have, after all, been a fundamental<br />
component of many species-centred restoration projects,<br />
<strong>in</strong>clud<strong>in</strong>g <strong>in</strong> rivers. Powerful tools such as PHABSIM and<br />
a wide range of classifi cation systems are available to<br />
predict and quantify change even <strong>in</strong> species-rich groups<br />
such as <strong>in</strong>vertebrates, although of course <strong>the</strong> limitations<br />
of such tools do need to be recognised.<br />
There seems to be historical sense of ecologists be<strong>in</strong>g<br />
advocates of a ‘black art’, perhaps particularly to hardened<br />
river eng<strong>in</strong>eers deal<strong>in</strong>g <strong>in</strong> ma<strong>the</strong>matical formulae. This<br />
may have orig<strong>in</strong>ated from ecologists be<strong>in</strong>g employed to<br />
count ‘bugs’ or fi sh or record plant abundance, orig<strong>in</strong>at<strong>in</strong>g<br />
from eng<strong>in</strong>eer<strong>in</strong>g works. Ecology is essentially a numerate<br />
science that can be quantifi ed by relationships and patterns<br />
and is no more ‘uncerta<strong>in</strong>’ than any o<strong>the</strong>r science, although<br />
it does deal with extremely complex systems and <strong>the</strong> relationships<br />
and <strong>in</strong>teractions with<strong>in</strong> <strong>the</strong>m. Ecological understand<strong>in</strong>g<br />
also often takes time and resources. Br<strong>in</strong>g<strong>in</strong>g<br />
ecologists <strong>in</strong>to a more central role, with <strong>in</strong>volvement from<br />
<strong>the</strong> early stages of <strong>the</strong> project, may help reduce uncerta<strong>in</strong>ty.<br />
Engag<strong>in</strong>g a greater range of academic river ecologists<br />
with wide understand<strong>in</strong>g of catchment processes,<br />
<strong>the</strong>oreticians and ecological modellers, as well as <strong>the</strong><br />
‘samplers and sorters’, who are frequently consultants, is<br />
also thought likely to be particularly benefi cial.<br />
Even a lack of historical reference data may be partly<br />
alleviated by monitor<strong>in</strong>g reference (control) reaches before<br />
and after <strong>the</strong> impact of <strong>the</strong> restoration scheme <strong>in</strong> a before–<br />
after–control–impact (BACI) design (e.g. as on <strong>the</strong> <strong>River</strong>s<br />
Cole and Brede; biggs et al., 1998). Particularly where this<br />
<strong>in</strong>corporates any historical <strong>in</strong>formation, it lends itself well<br />
to target sett<strong>in</strong>g with <strong>the</strong> subsequent means of assess<strong>in</strong>g<br />
whe<strong>the</strong>r <strong>the</strong> target <strong>in</strong> <strong>the</strong> controls is reached. As more
<strong>Uncerta<strong>in</strong>ty</strong> Surround<strong>in</strong>g <strong>the</strong> Ecological Targets and Response of <strong>River</strong> and Stream <strong>Restoration</strong> 153<br />
<strong>in</strong>formation is ga<strong>the</strong>red, a fl exible adaptive management<br />
approach allows targets to be adjusted (or perhaps even<br />
hardened from qualitative to quantitative, if necessary),<br />
enables fur<strong>the</strong>r studies to be planned when needed and can<br />
recognise and deal with emerg<strong>in</strong>g problems.<br />
The rigorous selection or even design of appropriate<br />
monitor<strong>in</strong>g techniques is essential if <strong>the</strong> major source<br />
of uncerta<strong>in</strong>ty <strong>in</strong> <strong>the</strong> monitor<strong>in</strong>g phase is to be reduced.<br />
The use of standard techniques simply because <strong>the</strong>y are<br />
rout<strong>in</strong>ely used by <strong>the</strong> organisation <strong>in</strong>volved, is to be<br />
guarded aga<strong>in</strong>st, as such methods have often been developed<br />
for a different purpose. It is far better to treat each<br />
scheme as a unique <strong>in</strong>dividual experiment for which <strong>the</strong><br />
appropriate method, <strong>in</strong>tensity and frequency of sampl<strong>in</strong>g<br />
is calculated. In o<strong>the</strong>r words, monitor<strong>in</strong>g should be seen<br />
as detailed research ra<strong>the</strong>r than a mere estimation of<br />
general changes.<br />
Although rivers have enormous resilience and an<br />
<strong>in</strong>herent ability for natural recovery <strong>the</strong> time scale and<br />
trajectory of recovery/restoration represents a major<br />
source of uncerta<strong>in</strong>ty, particularly where a stable endpo<strong>in</strong>t<br />
has been poorly defi ned or <strong>the</strong>re is little understand<strong>in</strong>g of<br />
what that may be. The ‘how long is a piece of str<strong>in</strong>g’<br />
analogy does not sit well with project budgets, which are<br />
typically fi nite and allocated to a particular time frame.<br />
We suggest that <strong>the</strong> trajectory of restoration may be more<br />
direct and shortened by considered project selection <strong>in</strong> <strong>the</strong><br />
fi rst place, where restoration is part of a generic <strong>the</strong>me<br />
adopted by a statutory body or <strong>in</strong>terested organisation.<br />
Tackl<strong>in</strong>g less damaged sites with good prospects for<br />
recovery, particularly where this can be undertaken through<br />
recovery enhancement and not major habitat reconstruction,<br />
is also likely to lead to a reduced time scale for<br />
recovery.<br />
There is clearly a need for more work on <strong>the</strong> factors<br />
surround<strong>in</strong>g colonisation to enable better predictions to be<br />
ultimately made. This may need to be conducted specifi -<br />
cally for <strong>the</strong> river concerned with<strong>in</strong> <strong>the</strong> project <strong>in</strong> an adaptive<br />
manner and may also require action to speed up <strong>the</strong><br />
colonisation process. For example, with problem species<br />
such as <strong>in</strong>vertebrates with particularly poor dispersal abilities,<br />
it may be prudent to <strong>in</strong>troduce <strong>the</strong>se, perhaps through<br />
‘seed<strong>in</strong>g’ of substrate, to reach a stable confi guration as<br />
soon as possible. For fi sh, it is important to defi ne any<br />
barriers to natural movement that aga<strong>in</strong>, may be readily<br />
overcome by translocation.<br />
Obviously, <strong>the</strong> strength of post-project appraisal ultimately<br />
depends on <strong>the</strong> uncerta<strong>in</strong>ties accumulated earlier<br />
<strong>in</strong> <strong>the</strong> project. Where <strong>the</strong>se are great this could deliver a<br />
fl awed evaluation of little or no value to future projects.<br />
Follow<strong>in</strong>g <strong>the</strong> approach outl<strong>in</strong>ed it should be possible to<br />
avoid this scenario <strong>in</strong> <strong>the</strong> future.<br />
ACKNOWLEDGEMENTS<br />
We are grateful to <strong>the</strong> Environment Agency for fund<strong>in</strong>g<br />
several of <strong>the</strong> studies documented and to its staff for assistance<br />
with sampl<strong>in</strong>g. ECON staff undertook much of <strong>the</strong><br />
work on <strong>the</strong> <strong>River</strong> Misbourne.<br />
REFERENCES<br />
Adams W. 1997. Rationalization and conservation: Ecology and<br />
<strong>the</strong> management of nature <strong>in</strong> <strong>the</strong> UK. Transactions of <strong>the</strong><br />
Institute of British Geographers 22: 227–291.<br />
Anand M, Desrochers R. 2004. Quantifi cation of restoration<br />
success us<strong>in</strong>g complex systems concepts and models. <strong>Restoration</strong><br />
Ecology 12: 117–123.<br />
Anderson DH, Dugger BD. 1998. A conceptual basis for evaluat<strong>in</strong>g<br />
restoration success. Transactions of <strong>the</strong> North American<br />
Wildlife and Natural Resources Conference 63.<br />
Benndorf J. 1992. The control of <strong>in</strong>direct effects of biomanipulation.<br />
In: Sutcliffe DW, Jones JG (Eds), Eutrophication:<br />
Research and Application to Water Supply. Freshwater Biological<br />
Association: Cumbria, UK; 82–93.<br />
Biggs J, Cornfi eld A, Grøn R et al. 1998. <strong>Restoration</strong> of <strong>the</strong> rivers<br />
Brede, Cole and Skerne: a jo<strong>in</strong>t Danish and British EU–LIFE<br />
demonstration project, V – Short–term impacts on <strong>the</strong> conservation<br />
value of aquatic macro<strong>in</strong>vertebrate and macrophyte<br />
assemblages. Aquatic Conservation: Mar<strong>in</strong>e and Freshwater<br />
Ecosystems 8: 241–255.<br />
Bohl<strong>in</strong> T. 1990. Estimation of population parameters us<strong>in</strong>g electric<br />
fi sh<strong>in</strong>g: aspects of <strong>the</strong> sampl<strong>in</strong>g design with emphasis on<br />
salmonids <strong>in</strong> streams. In: Cowx IG, Lamarque P (Eds), Fish<strong>in</strong>g<br />
with Electricity Fish<strong>in</strong>g News Books: Oxford, UK; 156–173.<br />
Boon PJ. 1998. <strong>River</strong> restoration <strong>in</strong> fi ve dimensions. Aquatic<br />
Conservation: Mar<strong>in</strong>e and Freshwater Ecosystems 8:<br />
257–264.<br />
Bovee KD. 1982. A Guide to Stream Habitat Analysis Us<strong>in</strong>g <strong>the</strong><br />
Instream Flow Incremental Methodology. Instream Flow Information<br />
Paper No.12. US Fish and Wildlife Service, FES/OBS–<br />
82/86: Wash<strong>in</strong>gton, DC.<br />
Bradshaw AD. 1987. <strong>Restoration</strong>: an acid test for ecology. In:<br />
Jordan WR, Gilp<strong>in</strong> ME, Aber JD (Eds), <strong>Restoration</strong> Ecology<br />
Cambridge University Press: Cambridge, UK; 23–29.<br />
Brooks A, Shields FD Jr. 1996. Towards an approach to susta<strong>in</strong>able<br />
river restoration. In: Brookes A, Shields FD Jr., <strong>River</strong><br />
Channel <strong>Restoration</strong>: Guid<strong>in</strong>g Pr<strong>in</strong>ciples for Susta<strong>in</strong>able Projects.<br />
John Wiley & Sons Ltd: Chichester, UK; 385–402.<br />
Brooks SS, Palmer MA, Card<strong>in</strong>ale BJ et al. 2002. Assess<strong>in</strong>g<br />
stream ecosystem rehabilitation: Limitations of community<br />
structure. <strong>Restoration</strong> Ecology 10: 156–168.<br />
Bruce-Burgess L, Sk<strong>in</strong>ner L. 2002. Appraisal: <strong>River</strong> restoration’s<br />
miss<strong>in</strong>g l<strong>in</strong>k. Proceed<strong>in</strong>gs of Conference held 27th November<br />
2002 at <strong>the</strong> University of Nott<strong>in</strong>gham, Downloadable from<br />
http://www.<strong>the</strong>rrc.co.uk.<br />
Buijse AD, Coups H, Staras M et al. 2002. <strong>Restoration</strong> strategies<br />
for river fl oodpla<strong>in</strong>s along large lowland rivers <strong>in</strong> Europe.<br />
Freshwater Biology 47: 889–907.
154 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Calow P. 1992. Energy budgets. In: Calow P, Petts G (Eds), The<br />
<strong>River</strong>s Handbook, Volume 1. Blackwell Scientifi c Publications:<br />
Oxford, UK; 370–378.<br />
Carpenter SR, Kitchell JF. 1993. The Trophic Cascade <strong>in</strong> Lakes.<br />
Cambridge University Press: Cambridge, UK.<br />
Carle FL, Strub MR. 1978. A new method for estimat<strong>in</strong>g population<br />
size from removal data. Biometrics 34: 621–630.<br />
Chandler JR. 1970. A biological approach to water quality management.<br />
Water Pollution Control, 69: 415–422.<br />
Clarke SJ, Bruce-Burgess L, Wharton G. 2003. L<strong>in</strong>k<strong>in</strong>g form and<br />
function: towards an ecohycromorphic approach to susta<strong>in</strong>able<br />
river restoration. Aquatic Conservation: Mar<strong>in</strong>e and Freshwater<br />
Ecosystems. 13: 439–450.<br />
Copp GH, Peñáz 1988. Ecology of fi sh spawn<strong>in</strong>g and nursery<br />
zones <strong>in</strong> <strong>the</strong> fl ood pla<strong>in</strong>, us<strong>in</strong>g a new sampl<strong>in</strong>g approach.<br />
Hydrobiologia 169: 209–224.<br />
Cosgrove PJ, Hastie LC. 2001. Conservation of threatened freshwater<br />
pearl mussel population: river management, mussel<br />
translocation and confl ict resolution. Biological Conservation<br />
99: 183–190.<br />
de Waal LC, Large ARG, Gippel CJ, Wade PM. 1995. <strong>River</strong> and<br />
fl oodpla<strong>in</strong> restoration <strong>in</strong> western Europe: opportunities and<br />
constra<strong>in</strong>ts. Archiv für Hydrobiologie 101: 679–693.<br />
Downs P, Sk<strong>in</strong>ner KS. 2002. <strong>River</strong>s and streams. In: Perrow MR,<br />
Davy AJ (Eds), Handbook of Ecological <strong>Restoration</strong>. Volume<br />
2: <strong>Restoration</strong> <strong>in</strong> Practice. Cambridge University Press: Cambridge,<br />
UK; 267–296.<br />
Drezner TD. 2004. Few native species colonize on mud fl ats ten<br />
years after dam removal. <strong>Restoration</strong> Ecology 22: 50–51.<br />
ECON 1993. The <strong>River</strong> <strong>Restoration</strong> Project, Phase 1: <strong>the</strong> Feasibility<br />
Study. Report to <strong>River</strong> <strong>Restoration</strong> Project (RRP), UK.<br />
Extence CA, Balbi DM, Chadd RP. 1999. <strong>River</strong> fl ow <strong>in</strong>dex<strong>in</strong>g<br />
us<strong>in</strong>g British benthic macro-<strong>in</strong>vertebrates: a framework for<br />
sett<strong>in</strong>g hydroecological objectives. Regulated <strong>River</strong>s Research<br />
and Management 15: 543–574.<br />
EU. 2000. Directive 2000/60/EC of <strong>the</strong> European Parliament<br />
and of <strong>the</strong> council establish<strong>in</strong>g framework for <strong>the</strong> community<br />
action <strong>in</strong> <strong>the</strong> fi eld of water policy. (Website: http://europa.eu.<br />
<strong>in</strong>t/comm/environment/water/water-framework/<strong>in</strong>dex_en.<br />
html).<br />
Fenolglio S, Agosta P, Bo T, Cucco M. 2002. Field experiments<br />
on colonization and movements of stream <strong>in</strong>vertebrates <strong>in</strong> an<br />
Apenn<strong>in</strong>e river (Visone, NW Italy). Hydrobiologia 474:<br />
125–130.<br />
Gameson ALH, Wheeler A. 1977. <strong>Restoration</strong> and <strong>the</strong> recovery of<br />
<strong>the</strong> Thames Estuary. In: Cairns J Jr., Dickson KL, Herricks EE<br />
(Eds), Recovery and <strong>Restoration</strong> <strong>in</strong> Damaged Ecosystems<br />
University Press of Virg<strong>in</strong>ia: Charlottesville, Virg<strong>in</strong>ia; 72–101.<br />
Garner P. 1997. Sample sizes for length and density estimation<br />
of 0+ fi sh when us<strong>in</strong>g po<strong>in</strong>t sampl<strong>in</strong>g by electrofi sh<strong>in</strong>g. Journal<br />
of Fisheries Biology 50: 95–106.<br />
Gore SA, Judy RD Jr. 1981. Predictive models of benthic macro<strong>in</strong>vertebrate<br />
density for use <strong>in</strong> <strong>in</strong>stream fl ow studies and<br />
regulated fl ow management. Canadian Journal of Fisheries<br />
and Aquatic Sciences 38: 1363–1370.<br />
Gorman OT, Karr JR. 1978. Habitat structure and stream fi sh<br />
communities. Ecology 59 (3): 507–515.<br />
Harper DM, Smith CD, Barham PJ. 1992. Habitats as <strong>the</strong> build<strong>in</strong>g<br />
blocks for river conservation assessment. In: Boon PJ, Calow<br />
P, Petts GE (Eds), <strong>River</strong> Conservation and Management. John<br />
Wiley & Sons Ltd: Chichester, UK; 311–319.<br />
Hearne J, Johnson IW, Armitage PD. 1994. Determ<strong>in</strong>ation of<br />
ecologically acceptable fl ows <strong>in</strong> rivers with seasonal changes<br />
<strong>in</strong> <strong>the</strong> density of macrophytes. Regulated <strong>River</strong>s: Research and<br />
Management 9: 117–184.<br />
Hoffmann CC, Pedersen ML, Kronvang B, Øvig L. 1998. <strong>Restoration</strong><br />
of <strong>the</strong> <strong>River</strong>s Brede, Cole and Skerne: a jo<strong>in</strong>t Danish and<br />
British EU–LIFE demonstration project, IV – implications for<br />
nitrate and iron transformation. Aquatic Conservation: Mar<strong>in</strong>e<br />
and Freshwater Ecosystems 8: 223–240.<br />
Holmes NTH. 1998. A Review of <strong>River</strong> Rehabilitation <strong>in</strong> <strong>the</strong> UK,<br />
1990–1996. Technical Report to <strong>the</strong> Environment Agency<br />
(W175), Environment Agency: Bristol, UK.<br />
Holmes NTH, Nielsen MB. 1998. <strong>Restoration</strong> of <strong>the</strong> rivers Brede,<br />
Cole and Skerne: a jo<strong>in</strong>t Danish and British EU–LIFE demonstration<br />
project, I – Sett<strong>in</strong>g up and delivery of <strong>the</strong> project.<br />
Aquatic Conservation: Mar<strong>in</strong>e and Freshwater Ecosystems 8:<br />
185–196.<br />
Jeppesen E, Sammalkarpi I. 2002. Lakes In: Perrow MR, Davy<br />
AJ (Eds), Handbook of Ecological <strong>Restoration</strong>. Volume 2: <strong>Restoration</strong><br />
<strong>in</strong> Practice. Cambridge University Press: Cambridge,<br />
UK; 297–324.<br />
Junk WJ, Bayley PB, Sparks RE. 1989. The fl ood pulse concept<br />
<strong>in</strong> river-fl oodpla<strong>in</strong> systems. Special Publication of <strong>the</strong> Canadian<br />
Journal of Fisheries and Aquatic Sciences 106:<br />
110–127.<br />
Kern K. 1992. Rehabilitation of streams <strong>in</strong> South-west Germany.<br />
In: Boon PJ, Calow P, Petts GE (Eds), <strong>River</strong> Conservation and<br />
Management. John Wiley & Sons Ltd: Chichester, UK;<br />
321–335.<br />
Kondolf GM. 1998. Lessons learned from river restoration projects<br />
<strong>in</strong> California. Aquatic Conservation: Mar<strong>in</strong>e and Freshwater<br />
Ecosystems 8: 39–52.<br />
Kronvang B, Svendsen LM, Brookes A et al. 1998. <strong>Restoration</strong><br />
of <strong>the</strong> rivers Brede, Cole and Skerne: a jo<strong>in</strong>t Danish and British<br />
EU–LIFE demonstration project, III – channel morphology,<br />
hydrodynamics and transport of sediment and nutrients.<br />
Aquatic Conservation: Mar<strong>in</strong>e and Freshwater Ecosystems 8:<br />
209–222.<br />
Laasonen P, Muotka T, Kilvijärvi I. 1998. Recovery of macro<strong>in</strong>vertebrate<br />
communities from stream habitat restoration.<br />
Aquatic Conservation: Mar<strong>in</strong>e and Freshwater Ecosystems 8:<br />
101–113.<br />
Lepori F, Palm D, Malmqvist B. 2005. Effects of stream restoration<br />
on ecosystem function<strong>in</strong>g: detritus retentiveness and<br />
decomposition. Journal of Applied Ecology 42: 175–189.<br />
Lucas MC, Batley E. 1996. Seasonal movements and behaviour<br />
of adult barbel Barbus barbus, a river<strong>in</strong>e cypr<strong>in</strong>id fi sh: implications<br />
for river management. Journal of Applied Ecology 33:<br />
1345–1358<br />
Magoulick DD, Kobza R. 2003. The role of refugia for fi shes<br />
dur<strong>in</strong>g drought: a review and syn<strong>the</strong>sis. Freshwater Biology 48:<br />
1186–1198.<br />
Malakoff D. 2004. The <strong>River</strong> Doctor. Science 305: 937–939.
<strong>Uncerta<strong>in</strong>ty</strong> Surround<strong>in</strong>g <strong>the</strong> Ecological Targets and Response of <strong>River</strong> and Stream <strong>Restoration</strong> 155<br />
McDonald A, Lane SN, Haycock NE, Chalk EA. 2004. <strong>River</strong>s of<br />
dreams: on <strong>the</strong> gulf between <strong>the</strong>oretical and practical aspects<br />
of an upland river restoration. Transactions of <strong>the</strong> Institute of<br />
British Geographers 29: 257–281.<br />
Milner NJ, Hemsworth RJ, Jones BE. 1985. Habitat evaluation<br />
as a fi sheries management tool. Journal of Fish Biology 27<br />
(Supplement A): 85–108.<br />
Mohl A. 2004. LIFE <strong>River</strong> <strong>Restoration</strong> projects <strong>in</strong> Austria. In:<br />
Geres D (Ed) <strong>River</strong> <strong>Restoration</strong> 2004 – Pr<strong>in</strong>ciples, Processes,<br />
Practices. Proceed<strong>in</strong>gs of <strong>the</strong> 3rd ECRR International Conference<br />
on <strong>River</strong> <strong>Restoration</strong> <strong>in</strong> Europe, Zagreb: Croatia (Papers<br />
downloadable from http://www.ecrr.org/).<br />
Muotka T, Paavol R, Haapala A et al. 2002. Long-term recovery<br />
of stream-habitat structure and benthic <strong>in</strong>vertebrate communities<br />
from <strong>in</strong>-stream restoration. Biological Conservation 105:<br />
243–253.<br />
Nestler JM, Schneider LT, Latka D, Johnson P. 1996. Impact analysis<br />
and restoration plann<strong>in</strong>g us<strong>in</strong>g <strong>the</strong> river<strong>in</strong>e community<br />
habitat assessment and restoration concept (RCHARC). In:<br />
Leclerc M, Capra H, Valent<strong>in</strong> S et al. (Eds), Ecohydraulics 2000.<br />
Proceed<strong>in</strong>gs of <strong>the</strong> 2nd International Symposium on Habitat<br />
Hydraulics, INRS – Eau: Québec, Canada; A871-A876.<br />
Newbold JD. 1992. Cycles and spirals of nutrients. In: Calow P,<br />
Petts GE (Eds), The <strong>River</strong>s Handbook Volume 1. Blackwell<br />
Scientifi c Publications: Oxford, UK; 79–408.<br />
Newson MD, Pitlick J, Sear DA. 2002. Runn<strong>in</strong>g water: fl uvial<br />
geomorphology and river restoration. In: Perrow MR, Davy AJ<br />
(Eds), Handbook of Ecological <strong>Restoration</strong>. Volume 1: Pr<strong>in</strong>ciples<br />
of <strong>Restoration</strong>. Cambridge University Press: Cambridge,<br />
UK: 324–354.<br />
Niemi GJ, DeVore P, Detenheck N et al. 1990. Overview of case<br />
studies on recovery of aquatic systems from disturbance.<br />
Environmental Management 14: 571–587.<br />
NRA (National <strong>River</strong>s Authority)/Kirkpatrick Scott Wilson.<br />
1992. Method for Assessment of Low Flow Conditions Caused<br />
by Abstraction. National <strong>River</strong>s Authority, R & D Note 45.<br />
Environment Agency: Bristol, UK.<br />
NRA (National <strong>River</strong>s Authority). 1995. Surface Water Abstraction<br />
Licens<strong>in</strong>g Policy Development. Project Report 505,<br />
National <strong>River</strong>s Authority: Bristol, UK.<br />
NRC (National Research Council US). 1992. Chapter 5: rivers<br />
and streams. In: <strong>Restoration</strong> of Aquatic Ecosytems National<br />
Academy Press: Wash<strong>in</strong>gton, DC.<br />
Ormerod SJ. 2004. A golden age of river restoration science?<br />
Aquatic Conservation: Mar<strong>in</strong>e & Freshwater Ecosystems. 14<br />
(6): 543–549.<br />
Ott<strong>in</strong>o P, Giller P. 2004. Distribution, density, diet and habitat use<br />
of <strong>the</strong> otter <strong>in</strong> relation to land use <strong>in</strong> <strong>the</strong> Aragl<strong>in</strong> Valley, Sou<strong>the</strong>rn<br />
Ireland. Biology and Environment: Proceed<strong>in</strong>gs of <strong>the</strong><br />
Royal Irish Academy 104B (1): 1–17.<br />
Palmer MA, Bernhardt ES, Allan JD et al. 2005. Standards for<br />
ecologically successful river restoration. Journal of Applied<br />
Ecology, 42 (2): 208–217.<br />
Perrow MR, Jowitt AJD, Zambrano González L. 1996. Sampl<strong>in</strong>g<br />
fi sh communities <strong>in</strong> shallow lowland lakes: po<strong>in</strong>t-sample electric<br />
fi sh<strong>in</strong>g vs electric fi sh<strong>in</strong>g with<strong>in</strong> stop-nets. Fisheries<br />
Management & Ecology 3: 303–313.<br />
Perrow MR, Toml<strong>in</strong>son ML, Zambrano L. 2002. Fish. In: Perrow<br />
MR, Davy AJ (Eds), Handbook of Ecological <strong>Restoration</strong>.<br />
Volume 1: Pr<strong>in</strong>ciples of <strong>Restoration</strong>. Cambridge University<br />
Press: Cambridge, UK; 324–354.<br />
Perrow MR, Davy AJ (Eds). 2002a. Handbook of Ecological<br />
<strong>Restoration</strong>. Volume 1: Pr<strong>in</strong>ciples of <strong>Restoration</strong>. Cambridge<br />
University Press: Cambridge, UK.<br />
Perrow MR, Davy AJ (Eds). 2002b. Handbook of Ecological<br />
<strong>Restoration</strong>. Volume 2: <strong>Restoration</strong> <strong>in</strong> Practice. Cambridge<br />
University Press: Cambridge, UK.<br />
Perrow MR, Toml<strong>in</strong>son ML. 2002. <strong>River</strong> Lambourn cSAC basel<strong>in</strong>e<br />
survey: Bullhead and Lamprey populations. Report to<br />
English Nature.<br />
Pretty JL, Harrison SSC, Shepherd DJ et al. 2003. <strong>River</strong> rehabilitation<br />
and fi sh populations: assess<strong>in</strong>g <strong>the</strong> benefi t of <strong>in</strong>stream<br />
structures. Journal of Applied Ecology 40: 251–265.<br />
Punchard NT, Perrow MR, Jowitt AJD. 2000. Fish habitat associations,<br />
community structure, density and biomass <strong>in</strong> natural and<br />
channelized lowland streams <strong>in</strong> <strong>the</strong> catchment of <strong>the</strong> <strong>River</strong><br />
Wensum, UK. In: Cowx IG (Ed.), Management and Ecology<br />
of <strong>River</strong> Fisheries. Blackwell Science: Oxford, UK; 143–157.<br />
Richardson JS, Jackson MJ. 2002. Aquatic <strong>in</strong>vertebrates. In:<br />
Perrow MR, Davy AJ (Eds), Handbook of Ecological <strong>Restoration</strong>.<br />
Volume 1: Pr<strong>in</strong>ciples of <strong>Restoration</strong>. Cambridge University<br />
Press: Cambridge, UK; 300–323.<br />
Rob<strong>in</strong>son C, Whitton S. 2004. Fisheries Action Plans: a new<br />
approach to public consultation and <strong>the</strong> impetus for habitat<br />
enhancement works <strong>in</strong> <strong>the</strong> upper Thames catchment (England).<br />
In: Geres D (Ed.), <strong>River</strong> <strong>Restoration</strong> 2004 – Pr<strong>in</strong>ciples, Processes,<br />
Practices. Proceed<strong>in</strong>gs of <strong>the</strong> 3rd ECRR International<br />
Conference on <strong>River</strong> <strong>Restoration</strong> <strong>in</strong> Europe: Zagreb, Croatia.<br />
Papers downloadable from http://www.ecrr.org.<br />
Roni P, Beechie TJ, Bilby RE et al. 2002. A review of stream<br />
restoration techniques and a hierarchical strategy for prioritiz<strong>in</strong>g<br />
restoration <strong>in</strong> <strong>the</strong> Pacifi c Northwest watersheds. North<br />
American Journal of Fisheries Management 22: 1–20.<br />
Rosenberger A, Angermeier PL. 2003. Ontogenetic shifts <strong>in</strong><br />
habitat use by <strong>the</strong> endangered Roanoke logperch (Perc<strong>in</strong>a rex).<br />
Freshwater Biology 48: 1563–1577.<br />
RSPB (Royal Society for <strong>the</strong> Protection of Birds), NRA (National<br />
<strong>River</strong>s Authority) and <strong>the</strong> RSNC (Royal Society for Nature<br />
Conservation). 1995. The New <strong>River</strong>s and Wildlife Handbook<br />
The Royal Society for <strong>the</strong> Protection of Birds: Sandy, Bedfordshire,<br />
UK.<br />
Russell IC. 1994. Salmon stock<strong>in</strong>g <strong>in</strong> north-east England – some<br />
factors affect<strong>in</strong>g return rates. In: Cowx IG (Ed.), Rehabilitation<br />
of Freshwater Fisheries. Fish<strong>in</strong>g News Books, Blackwell Scientifi<br />
c Publications Ltd: Oxford, UK; 255–267.<br />
Scheffer M, Hosper SH, Meijer M-L et al. 1993. Alternative<br />
equilibria <strong>in</strong> shallow lakes. Trends <strong>in</strong> Ecology and Evolution 8:<br />
275–279.<br />
Smith GRT, Learner MA, Slater FM, Foster J. 1996. Habitat features<br />
important for <strong>the</strong> conservation of <strong>the</strong> native crayfi sh <strong>in</strong><br />
Brita<strong>in</strong>. Biological Conservation 75: 239–246.<br />
Tikkanen P, Laasonen O, Muotka T et al. 1994. Short term recovery<br />
of benthos follow<strong>in</strong>g disturbance from stream habitat rehabilitation.<br />
Hydrobiologia 273: 121–130.
156 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Tockner K, Schiemer F. 1997. Ecological aspects of <strong>the</strong> restoration<br />
strategy for a river-fl oodpla<strong>in</strong> system on <strong>the</strong> Danube <strong>River</strong><br />
<strong>in</strong> Austria. Global Ecology and Biogeography Letters 6:<br />
321–329.<br />
Tockner K, Schiemer F, Ward JV. 1998. Conservation by restoration:<br />
<strong>the</strong> management concept for a river-fl oodpla<strong>in</strong> system on<br />
<strong>the</strong> Danube <strong>River</strong> <strong>in</strong> Austria. Aquatic Conservation 8: 71–86.<br />
Tockner K, Schiemer F, Baumgartner C et al. 1999. The Danube<br />
<strong>Restoration</strong> Project: Species diversity patterns across connectivity<br />
gradient6s <strong>in</strong> <strong>the</strong> fl oodpla<strong>in</strong> system. Regulated <strong>River</strong>s:<br />
Research and Management. 15: 245–258.<br />
Toth LA. 1996. Restor<strong>in</strong>g <strong>the</strong> hydrogeomorphology of <strong>the</strong> channelised<br />
Kissimmee <strong>River</strong>. In: Brookes A, Shields FD Jr (Eds),<br />
<strong>River</strong> Channel <strong>Restoration</strong>: Guid<strong>in</strong>g Pr<strong>in</strong>ciples for Susta<strong>in</strong>able<br />
Projects John Wiley & Sons Ltd: Chichester, UK; 369–383.<br />
Toth LA, Anderson DH. 1998. Develop<strong>in</strong>g expectations for ecosystem<br />
restoration. In: Transactions of <strong>the</strong> 63rd North American<br />
Wildlife Natural Resources Conference: Wash<strong>in</strong>gton, DC;<br />
122–134.<br />
Toth LA, Melv<strong>in</strong> SL, Arr<strong>in</strong>gton DA, Chamberla<strong>in</strong> J. 1998. Hydrological<br />
manipulations of <strong>the</strong> channelised Kissimmee <strong>River</strong>:<br />
Implications for restoration. Bioscience 48: 758–764.<br />
Townsend CR. 1989. The patch dynamics concept of stream community<br />
ecology. Journal of <strong>the</strong> North American Benthological<br />
Society 8: 36–50.<br />
Trexler JC. 1995. <strong>Restoration</strong> of <strong>the</strong> Kissimmee <strong>River</strong>: a conceptual<br />
model of past and present fi sh communities and its consequences<br />
for evaluat<strong>in</strong>g restoration success. <strong>Restoration</strong> Ecology<br />
3: 195–210.<br />
Utz<strong>in</strong>ger J, Roth C, Peter A. 1998. Effects of environmental<br />
parameters on <strong>the</strong> distribution of bullhead Cottus gobio with<br />
particular consideration of <strong>the</strong> effects of obstructions. Journal<br />
of Applied Ecology 35: 882–892.<br />
Van Steeter MM, Pitlick J. 1998. Geomorphology and endangered<br />
fi sh habitats of <strong>the</strong> Upper Colorado <strong>River</strong>. 1: Historic changes<br />
<strong>in</strong> streamfl ow, sediment load and channel morphology. Water<br />
Resources Research 34: 287–302.<br />
Vannote RL, M<strong>in</strong>shall GW, Cumm<strong>in</strong>s KW et al. 1980. The river<br />
cont<strong>in</strong>uum concept. Canadian Journal of Fisheries and Aquatic<br />
Sciences 37: 130–137.<br />
Vivash R, Ottosen O, Janes M, Sørensen HV. 1998. <strong>Restoration</strong><br />
of <strong>the</strong> rivers Brede, Cole and Skerne: a jo<strong>in</strong>t Danish and British<br />
EU–LIFE demonstration project, II – <strong>the</strong> river restoration<br />
works and o<strong>the</strong>r related practical aspects. Aquatic Conservation:<br />
Mar<strong>in</strong>e and Freshwater Ecosystems 8: 197–208.<br />
Waldman JR, Wirg<strong>in</strong> II. 1998. Status and restoration options for<br />
Atlantic sturgeon <strong>in</strong> North America. Conservation Biology 12:<br />
631–638.<br />
Wesche TA. 1985. Stream channel modifi cations and reclamation<br />
structures to enhance fi sh habitat. In: Gore JA (Ed.), The<br />
<strong>Restoration</strong> of <strong>River</strong>s and Streams: Theories and Experience.<br />
Butterworth Publishers: Massachusetts, USA; 103–163.<br />
Whalen PJ, Toth LA, Koebel JW, Strayer PK. 2002. Kissimmee<br />
<strong>River</strong> <strong>Restoration</strong>: A Case Study. Water Science and Technology<br />
45: 55–62.<br />
White RJ, Brynildson OM. 1967. Guidel<strong>in</strong>es for Management of<br />
Trout Stream Habitat <strong>in</strong> Wiscons<strong>in</strong>. Technical Bullet<strong>in</strong> No. 39,<br />
Division of Conservation, Department of Natural Resources:<br />
Madison, Wiscons<strong>in</strong>.<br />
White PS, Walker JL. 1997. Approximat<strong>in</strong>g nature’s variation:<br />
select<strong>in</strong>g and us<strong>in</strong>g reference <strong>in</strong>formation <strong>in</strong> restoration<br />
ecology. <strong>Restoration</strong> Ecology 5: 338–349.<br />
W<strong>in</strong>terbottom JH, Orton SE, Hildrew AG. 1997. Field Experiments<br />
on fl ow refugia <strong>in</strong> streams. Freshwater Biology 37:<br />
569–588.<br />
APPENDIX 8.1 THE USE AND POTENTIAL<br />
LIMITATIONS OF IFIM AND PHABSIM<br />
The ‘Instream fl ow <strong>in</strong>cremental methodology’ (IFIM) and<br />
its underly<strong>in</strong>g model, <strong>the</strong> ‘Physical habitat simulation<br />
model’ (PHABSIM) represent <strong>the</strong> most popular approach<br />
for simulat<strong>in</strong>g habitat preferences (Downs and Sk<strong>in</strong>ner,<br />
2002). The relationship between streamfl ow and available<br />
physical habitat (defi ned by depth, velocity, substrate and<br />
cover) is used to compute <strong>the</strong> ‘weighted usable area’<br />
(WUA) <strong>in</strong> a reach as a function of <strong>the</strong> river discharge, for<br />
different life stages and species of fi sh. For each life<br />
stage of <strong>the</strong> target species, <strong>the</strong> model requires expressions<br />
of <strong>the</strong> relative suitability for that species of <strong>the</strong> full<br />
range of values taken by <strong>the</strong>se variables. These univariate<br />
curves or habitat suitability <strong>in</strong>dices may be derived from<br />
exist<strong>in</strong>g literature, expert op<strong>in</strong>ion or by sampl<strong>in</strong>g techniques<br />
such as electro-fi sh<strong>in</strong>g or snorkell<strong>in</strong>g. PHABSIM<br />
also conta<strong>in</strong>s a number of hydraulic models that predict<br />
values of depth and velocity at different simulation discharges.<br />
These models require calibration us<strong>in</strong>g fi eld data<br />
collected at two or more calibration discharges. Observations<br />
of substrate and cover are recorded us<strong>in</strong>g a cod<strong>in</strong>g<br />
system and are assumed to be <strong>in</strong>dependent of discharge.<br />
Once calibrated <strong>the</strong> model can simulate values of microhabitat<br />
variables over <strong>the</strong> full range of discharge with<strong>in</strong> a<br />
river reach. Comb<strong>in</strong><strong>in</strong>g <strong>the</strong> results with habitat suitability<br />
data produces <strong>the</strong> WUA versus Discharge relationship<br />
(CEH website: http://www.nwl.ac.uk/ih). PHABSIM,<br />
by relat<strong>in</strong>g habitat to discharge, provides a quantitative<br />
entity, allow<strong>in</strong>g river ecologists to negotiate prescribed<br />
fl ows <strong>in</strong> equivalent terms to o<strong>the</strong>r water resource demands,<br />
and offers a practical means of <strong>in</strong>tegrat<strong>in</strong>g ecological<br />
requirements of aquatic species with o<strong>the</strong>r water resource<br />
demands.<br />
However, numerous procedural, biological and physical<br />
limit<strong>in</strong>g assumptions prevented IFIM and PHABSIM<br />
from be<strong>in</strong>g reliable and, <strong>in</strong> response to <strong>the</strong>se criticisms, a<br />
wide range of more complex approaches based on<br />
advanced understand<strong>in</strong>g of <strong>the</strong> bioenergetics of fi sh<br />
biology and on more realistic representation and modell<strong>in</strong>g<br />
of fl ow patterns is now at <strong>the</strong> research stage (3rd<br />
International Symposium on Ecohydraulics, 1999).
<strong>Uncerta<strong>in</strong>ty</strong> Surround<strong>in</strong>g <strong>the</strong> Ecological Targets and Response of <strong>River</strong> and Stream <strong>Restoration</strong> 157<br />
At <strong>the</strong> Instream Flow Incremental Methodology Workshop<br />
held <strong>in</strong> New Zealand (February 2004), jo<strong>in</strong>tly hosted<br />
by Fish and Game New Zealand and <strong>the</strong> Department of<br />
Conservation, Cawthron Institute NIWA (Conference<br />
website: http://www.doc.govt.nz/Explore/Hunt<strong>in</strong>g-and-<br />
Fish<strong>in</strong>g/Taupo-Fishery/), several speakers (e.g. Maclean<br />
and Death) expressed concern about some of <strong>the</strong> underly<strong>in</strong>g<br />
assumptions beh<strong>in</strong>d PHABSIM, <strong>in</strong>clud<strong>in</strong>g that: fi sh are<br />
limited by habitat; that habitat preferences are adequately<br />
described by depth, velocity and substrate type; that habitat<br />
preferences do not change <strong>in</strong> accordance with diurnal/seasonal<br />
patterns; and that a large area of suboptimal habitat<br />
is preferable to a small area of optimal habitat.<br />
It was also argued that PHABSIM was not appropriate<br />
for use <strong>in</strong> connection with New Zealand (NZ) <strong>in</strong>vertebrates<br />
on <strong>the</strong> basis that depth, velocity and substrate type<br />
were of m<strong>in</strong>or importance <strong>in</strong> comparison with o<strong>the</strong>r factors<br />
such as riparian catchment vegetation, disturbance and<br />
food supply. Habitat requirements also appeared to be<br />
fl exible for many NZ species. The use of habitat suitability<br />
curves was also brought <strong>in</strong>to question and it was noted<br />
that it needs to be made explicitly clear how <strong>the</strong>y were<br />
developed, for example which sites were used, <strong>the</strong> number<br />
of samples taken and whe<strong>the</strong>r diurnal patterns were<br />
accounted for <strong>in</strong> <strong>the</strong> sampl<strong>in</strong>g process. The lack of variance<br />
estimation (i.e. error bars) was also raised as a limitation.<br />
Reduced (or <strong>in</strong>creased) fl ows might have effects on<br />
o<strong>the</strong>r variables that are not taken <strong>in</strong>to account, such as<br />
nutrient content, sediment build up, and temperature.<br />
PHABSIM and variants such as <strong>the</strong> <strong>River</strong><strong>in</strong>e Community<br />
Habitat Assessment and <strong>Restoration</strong> Concept<br />
(RCHARC) (Nestler et al., 1996) <strong>in</strong> <strong>the</strong> United K<strong>in</strong>gdom<br />
on lowland chalk streams suffer<strong>in</strong>g from low fl ow have<br />
also proved to be of limited use for <strong>in</strong>vertebrates. There is<br />
<strong>the</strong> risk that habitat area/habitat response predictions will<br />
be mean<strong>in</strong>gless if <strong>the</strong> sampl<strong>in</strong>g transects are not widely<br />
representative. These problems may be more marked for<br />
smaller rivers and streams <strong>in</strong> <strong>the</strong> UK context. It is known<br />
that for chalk streams and o<strong>the</strong>r nutrient-rich lowland<br />
rivers <strong>the</strong> seasonal changes <strong>in</strong> <strong>the</strong> growth and biomass of<br />
macrophytes require, at <strong>the</strong> very least, a careful calibration<br />
of PHABSIM to prevent distortion of results (Hearne et<br />
al., 1994).<br />
Overall, whilst PHABSIM is a potentially powerful tool<br />
it may be <strong>the</strong> most effective when evaluat<strong>in</strong>g and compar<strong>in</strong>g<br />
different management options, <strong>in</strong> <strong>in</strong>stances where<br />
physical habitat limits populations. The United K<strong>in</strong>gdom<br />
method only assesses physical habitat, so factors such as<br />
water quality and temperature and sediment transport<br />
require complementary studies. Indeed, where such factors<br />
are <strong>the</strong> prime constra<strong>in</strong>t on populations <strong>the</strong> use of<br />
PHABSIM is <strong>in</strong>appropriate.<br />
Much criticism centres on this latter po<strong>in</strong>t and<br />
PHABSIM like any model is only as good as <strong>the</strong> data that<br />
is entered and makes <strong>the</strong> assumption that habitat variables<br />
are <strong>the</strong> best descriptor of <strong>the</strong> abundance of a particular<br />
group, which of course need not be <strong>the</strong> case. Also, <strong>the</strong><br />
quantity and quality of water cannot limit <strong>the</strong> distribution<br />
or abundance of <strong>the</strong> specifi ed organism <strong>in</strong> <strong>the</strong> fi rst place.<br />
Whilst <strong>the</strong>re is an <strong>in</strong>herent desire for scientists to adopt a<br />
quantitative approach <strong>in</strong> order to reduce uncerta<strong>in</strong>ty, this<br />
is only successful if a high level of confi dence can be<br />
placed <strong>in</strong> <strong>the</strong> predictions. In <strong>the</strong> case of PHABSIM, many<br />
users have found that predictions for even ‘straightforward’<br />
schemes do not come close to <strong>the</strong> quantitative<br />
targets set. However, PHABSIM can also be used as a<br />
qualitative tool, and may be useful when it comes to<br />
select<strong>in</strong>g <strong>the</strong> best likely restoration option from a number<br />
of possible schemes.<br />
APPENDIX 8.2 CLASSIFICATION TECHNIQUES<br />
FOR MONITORING AND POTENTIALLY<br />
PREDICTING THE RESPONSE OF<br />
INVERTEBRATES TO FLOW RESTORATION<br />
In <strong>the</strong> United Kongdom low fl ows are seen to be a problem,<br />
especially <strong>in</strong> chalk streams, which are characterised by<br />
high <strong>in</strong>vertebrate biomass and abundance and a relatively<br />
high diversity of species. The ma<strong>in</strong> physical effects of low<br />
fl ows <strong>in</strong> chalk streams are: long term dry<strong>in</strong>g of ephemeral<br />
or w<strong>in</strong>terbourne reaches; long term dry<strong>in</strong>g of spr<strong>in</strong>g seepages<br />
or fl ushes; downstream migration of <strong>the</strong> perennial<br />
head of a chalk stream or river; and reduced water levels<br />
and/or velocity <strong>in</strong> downstream reaches. Much restoration<br />
<strong>in</strong> <strong>the</strong> United K<strong>in</strong>gdom has been concerned with ‘Alleviation<br />
of Low Flow’ (ALF) schemes, which were driven<br />
by <strong>the</strong> concern of <strong>the</strong> environmental impacts of overabstraction.<br />
The advantages and disadvantages of methodologies<br />
developed to monitor <strong>in</strong>vertebrates <strong>in</strong> low fl ow<br />
systems are outl<strong>in</strong>ed below.<br />
Scott Wilson Kirkpatrick (SWK) Methodology<br />
The fi rst national methodology for assess<strong>in</strong>g low fl ow<br />
conditions caused by abstraction was not written until<br />
1992 (NRA R&D Note 45, 1992 – a procedural manual<br />
by <strong>the</strong> consultants Scott Wilson Kirkpatrick). This set out<br />
a po<strong>in</strong>t-scor<strong>in</strong>g prioritis<strong>in</strong>g procedure based on <strong>the</strong> assessment<br />
of four <strong>in</strong>dicators: Hydrological; Ecological; Landscape<br />
and Amenity; and Public Perception.<br />
For each <strong>in</strong>dicator, <strong>in</strong>dividual assessments of a variety<br />
of parameters were conducted. The scores for each <strong>in</strong>dicator<br />
were comb<strong>in</strong>ed, with separate weight<strong>in</strong>gs. The ecological<br />
<strong>in</strong>dicator parameters and <strong>the</strong>ir <strong>in</strong>dividual weight<strong>in</strong>gs
158 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
were: <strong>in</strong>vertebrate community parameter (0.4), fi shery<br />
parameter (0.2), fi sh stocks parameter (0.3), plant parameter<br />
(0.1) and an optional conservation parameter (0.3).<br />
The major disadvantage was that <strong>the</strong> methodology was<br />
based solely upon Average Scores Per Taxon (ASPT)<br />
scores, a measure of <strong>the</strong> balance between pollutionsensitive<br />
and pollution-tolerant <strong>in</strong>vertebrates. This provides<br />
what is essentially an <strong>in</strong>dex of organic pollution<br />
without any accommodation of <strong>the</strong> overall faunal richness<br />
of samples, <strong>the</strong> fl ow velocity associations of species<br />
present, or <strong>the</strong>ir preference for ei<strong>the</strong>r perennial or ephemeral<br />
fl ow. The subjective manner of deduc<strong>in</strong>g severity<br />
<strong>in</strong>dex scores, based on targets that were produced on <strong>the</strong><br />
basis of adjustments to a start<strong>in</strong>g fi gure suited to sites<br />
considered to offer highest potential for ASPT, was also<br />
of major concern, s<strong>in</strong>ce it effectively skewed most lowland<br />
rivers towards low impact scores, despite <strong>the</strong> known severity<br />
of impacts with<strong>in</strong> chalk catchments at <strong>the</strong> time.<br />
Instream Flow Incremental Methodology<br />
(IFIM) and PHABSIM<br />
Instream Flow Incremental Methodology (IFIM) and<br />
PHABSIM have also been used to establish fl ow objectives<br />
from direct measurement (or modell<strong>in</strong>g) of resident<br />
biotic communities and <strong>the</strong>ir specifi c habitat requirements,<br />
although <strong>the</strong>se models have a tendency to be <strong>in</strong>accurate<br />
and <strong>in</strong>adequate, and at best require careful calibration to<br />
prevent distortion of results.<br />
The Surface Water Abstraction Licens<strong>in</strong>g Policy<br />
(SWALP) Methodology<br />
The SWALP methodology developed for <strong>the</strong> NRA by <strong>the</strong><br />
consultant eng<strong>in</strong>eers Sir William Halcrow & Partners Ltd<br />
<strong>in</strong> 1995 <strong>in</strong>corporates an environmental weight<strong>in</strong>g (EW)<br />
system, for which scores for sensitivity to abstraction are<br />
assigned to physical character, ecology and fi sheries characteristics<br />
(NRA, 1995). The catchment is fi rst assigned<br />
an EW score and thresholds are produced for a set of<br />
defi ned critical assessment po<strong>in</strong>ts (APs).<br />
The pr<strong>in</strong>ciple beh<strong>in</strong>d <strong>the</strong> SWALP ecological scor<strong>in</strong>g<br />
system was that <strong>the</strong> highest scores were allocated to<br />
species perceived to require coarse bed materials and<br />
rapid/fast current velocities, and lower scores were given<br />
to species which are common <strong>in</strong> still water habitats or are<br />
capable of withstand<strong>in</strong>g dessication. In contrast to <strong>the</strong><br />
SWK method, SWALP targeted environmental weight<strong>in</strong>g<br />
towards physical and biotic (ecology and fi sheries) attributes,<br />
exclud<strong>in</strong>g use-related amenity and recreational criteria,<br />
which were somewhat subjective. It also represented<br />
<strong>the</strong> fi rst attempt to br<strong>in</strong>g toge<strong>the</strong>r <strong>the</strong> perceived fl ow pref-<br />
erences of <strong>in</strong>dividual macrophyte and <strong>in</strong>vertebrate species<br />
<strong>in</strong> a biotic <strong>in</strong>dex concerned solely with fl ow dependency.<br />
However, <strong>the</strong> environmental weight<strong>in</strong>gs produced by <strong>the</strong><br />
SWALP methodology are ra<strong>the</strong>r crude and offer only a<br />
subjective basis for decision mak<strong>in</strong>g and a limited capacity<br />
to monitor improvement or determ<strong>in</strong>e <strong>the</strong> relative fl ow<br />
requirements with<strong>in</strong> or between river bas<strong>in</strong>s.<br />
The Lotic–Invertebrate Index for Flow Evaluation<br />
Index (LIFE)<br />
Follow<strong>in</strong>g <strong>the</strong> <strong>in</strong>vertebrate sensitivities to fl ow suggested<br />
by <strong>the</strong> SWALP methodology, Extence et al. (1999) developed<br />
LIFE. The method sought to l<strong>in</strong>k changes <strong>in</strong> river<strong>in</strong>e<br />
benthic macro<strong>in</strong>vertebrate assemblages to prevail<strong>in</strong>g fl ow<br />
regimes. The <strong>in</strong>dex is based upon <strong>the</strong> ‘primary fl ow associations’<br />
of different maco<strong>in</strong>vertebrate taxa at ei<strong>the</strong>r<br />
BMWP1 family or mixed-species level – deduced, for <strong>the</strong><br />
most part, from taxonomic keys. Calculation of <strong>the</strong> LIFE<br />
<strong>in</strong>dex score for a sample <strong>in</strong>volves <strong>in</strong>dividual fl ow scores<br />
(fs) for each scor<strong>in</strong>g taxon present <strong>in</strong> a sample obta<strong>in</strong>ed<br />
from a matrix. This matrix is based on <strong>the</strong> <strong>in</strong>frastructure<br />
of <strong>the</strong> biotic score system proposed by Chandler (1970)<br />
for assess<strong>in</strong>g water quality (Extence et al., 1999). Increas<strong>in</strong>g<br />
abundance of stand<strong>in</strong>g water or drought resistant<br />
species produces a lower fs score, whilst <strong>in</strong>creased numbers<br />
of <strong>in</strong>dividuals from runn<strong>in</strong>g water fl ow groups produces<br />
higher scores.<br />
Whilst changes <strong>in</strong> LIFE score may faithfully refl ect <strong>the</strong><br />
observed shifts <strong>in</strong> composition at sites undergo<strong>in</strong>g gross<br />
changes <strong>in</strong> habitat provision, <strong>the</strong> depletion of fauna caused<br />
by an earlier channel-dry<strong>in</strong>g event can produce spurious<br />
LIFE scores. This may occur, for example, by <strong>the</strong> temporal<br />
exclusion of molluscs normally associated with slack or<br />
slow-fl ow<strong>in</strong>g water or by rapid colonisation of blackfl ies<br />
(Simuliidae) and mayfl ies (e.g. Baetis spp) at a re-wetted<br />
site (Figure 8.4). These aspects of recolonisation usually<br />
produce a faunal composition with a higher proportion of<br />
velocity-dependent species and an artifi cially elevated<br />
LIFE score, even though taxon richness and diversity may<br />
rema<strong>in</strong> severely depleted. In <strong>the</strong> case of <strong>the</strong> <strong>River</strong> Misbourne<br />
(Appendix 8.3), LIFE scores were also found to<br />
vary signifi cantly between sites <strong>in</strong> close proximity (100 m<br />
apart) l<strong>in</strong>ked to spatial differences <strong>in</strong> habitat representa-<br />
1 The BMWP (Biological Monitor<strong>in</strong>g Work<strong>in</strong>g Party) scores<br />
refere to a system for assess<strong>in</strong>g water quality based on <strong>the</strong> macro<strong>in</strong>vertebrate<br />
assemblage. Each taxa is allocated a value from 1<br />
to 10 depend<strong>in</strong>g on its known tolerance to organic pollution. High<br />
scor<strong>in</strong>g taxa have lower pollution tolerance, and are thus an <strong>in</strong>dicator<br />
of good water quality.
<strong>Uncerta<strong>in</strong>ty</strong> Surround<strong>in</strong>g <strong>the</strong> Ecological Targets and Response of <strong>River</strong> and Stream <strong>Restoration</strong> 159<br />
tion <strong>in</strong> a relatively unmodifi ed chalk stream. Accord<strong>in</strong>gly,<br />
<strong>the</strong> LIFE score response largely depended upon <strong>the</strong> magnitude<br />
of temporal habitat change observed at an <strong>in</strong>dividual<br />
site. This highlights <strong>the</strong> need to select a representative<br />
variety of sites to assess <strong>the</strong> l<strong>in</strong>ear extent of environmental<br />
damage attributable to low fl ow.<br />
The MISINDEX<br />
An alternative fl ow dependency <strong>in</strong>dex, MISINDEX, was<br />
developed for use on <strong>the</strong> <strong>River</strong> Misbourne (Appendix 8.3)<br />
to accommodate cases where assemblages are transitional<br />
and depleted by previous dry<strong>in</strong>g. The MISINDEX works<br />
<strong>in</strong> a similar manner to <strong>the</strong> LIFE <strong>in</strong>dex, with <strong>the</strong> added<br />
advantage that it featured extended taxonomic coverage of<br />
aquatic species and <strong>the</strong> <strong>in</strong>clusion of water dependent (but<br />
not fully aquatic) species of Coleoptera. As with <strong>the</strong> LIFE<br />
<strong>in</strong>dex, <strong>the</strong> MISINDEX used fi ve fl ow-dependency groups<br />
and used abundance data for taxa <strong>in</strong> a similar manner to<br />
Extence et al. and Chandler’s matrices, but <strong>in</strong>troduced a<br />
category of water-related or marsh species that were not<br />
necessarily aquatic but depend upon wet places with<strong>in</strong> a<br />
river corridor, lake bas<strong>in</strong> or marsh. Moreover, ra<strong>the</strong>r than<br />
attempt<strong>in</strong>g to allocate <strong>in</strong>dividual species to a primary fl ow<br />
type association that was evidently subjective, <strong>the</strong> <strong>in</strong>tention<br />
was to defi ne <strong>the</strong>m by <strong>the</strong> limitations of <strong>the</strong>ir tolerance<br />
to fl ow cessation and to tell it like it is if a taxon was of<br />
low or unknown fi delity.<br />
Nei<strong>the</strong>r <strong>the</strong> LIFE or MISINDEX <strong>in</strong>dices explicitly take<br />
<strong>in</strong>to account <strong>the</strong> dispersal ability of species, or <strong>the</strong>ir ability<br />
to recolonise fl ow derogated sites after fl ow restoration, an<br />
important factor that requires fur<strong>the</strong>r <strong>in</strong>vestigation and<br />
could improve prediction of <strong>the</strong> likely time scale of<br />
impacts result<strong>in</strong>g from fl ow derogation or periodic channel<br />
dry<strong>in</strong>g.<br />
Detrended Correspondence Analysis (DECORANA)<br />
As a multivariate statistical technique DECANORA can<br />
be used to quantify similarities between samples or sites<br />
ei<strong>the</strong>r spatially or temporally. From such an analysis, it<br />
may be possible to produce mean<strong>in</strong>gful fl ow objectives<br />
determ<strong>in</strong>ed by observed empirical evidence of <strong>the</strong> l<strong>in</strong>ks<br />
between faunal composition or population sizes of <strong>in</strong>dividual<br />
species and river fl ows or o<strong>the</strong>r recorded environmental<br />
characteristics l<strong>in</strong>ked to hydrological variables <strong>in</strong><br />
a particular river reach. For this, longer term time-series<br />
of <strong>in</strong>formation are required than usually exist at present,<br />
as experience suggests that <strong>the</strong>re is considerable fl ux <strong>in</strong><br />
<strong>the</strong> composition of assemblages, <strong>the</strong> abundance of constituent<br />
species and <strong>in</strong>stream habitat characteristics l<strong>in</strong>ked<br />
to <strong>the</strong> fl ow history of sites and many o<strong>the</strong>r factors.<br />
APPENDIX 8.3 MONITORING THE<br />
ECOLOGICAL RESPONSE OF THE RECOVERY<br />
ENHANCEMENT OF THE RIVER MISBOURNE<br />
Introduction<br />
The <strong>River</strong> Misbourne, a 28 km long, ecologically valuable<br />
chalk stream situated <strong>in</strong> <strong>the</strong> <strong>River</strong> Thames catchment <strong>in</strong><br />
Berkshire, was believed to be amongst <strong>the</strong> rivers most<br />
affected by abstraction <strong>in</strong> <strong>the</strong> United K<strong>in</strong>gdom. Compounded<br />
by <strong>the</strong> effects of drought by 1996/1997, migration<br />
of <strong>the</strong> perennial head saw <strong>the</strong> river dry from its source<br />
to <strong>the</strong> middle reaches. A scheme to restore fl ow was implemented<br />
by <strong>the</strong> Environment Agency and Three Valleys<br />
Water, which comprised abstraction reduction and relocation<br />
of abstraction po<strong>in</strong>ts. A series of boreholes commissioned<br />
by <strong>the</strong> Environment Agency (EA) monitored <strong>the</strong><br />
hydrological response of <strong>the</strong>se measures, with water<br />
quality rout<strong>in</strong>ely monitored at four sites.<br />
The ecological response to <strong>the</strong> attempted fl ow recovery<br />
was monitored at six sites along <strong>the</strong> course of <strong>the</strong> river,<br />
<strong>in</strong>clud<strong>in</strong>g sites that had dried down and those that had<br />
reta<strong>in</strong>ed fl ow. Sites with differ<strong>in</strong>g characteristics were<br />
chosen to assess response more thoroughly, although <strong>the</strong><br />
variation between sites meant that differences between<br />
sites needed to be compared. For example, Sites 1 and 2<br />
were dry <strong>in</strong> 1998; Site 3 rema<strong>in</strong>ed dry until 2001, reta<strong>in</strong>ed<br />
fl ow until spr<strong>in</strong>g 2003, but had dried down aga<strong>in</strong> by<br />
autumn 2003; Site 4 was dry only <strong>in</strong> autumn 1997; and<br />
Sites 5 and 6 reta<strong>in</strong>ed fl ow throughout.<br />
An attempt was made to monitor ecological response<br />
us<strong>in</strong>g a series of standard survey methodologies typically<br />
applied twice annually (often spr<strong>in</strong>g and autumn) over a<br />
three-year period (1996 to 1999). Changes <strong>in</strong> vegetation<br />
with<strong>in</strong> <strong>the</strong> corridor was monitored with Phase 1 and 2<br />
Habitat Survey coupled with <strong>River</strong> Corridor Survey (RCS)<br />
and <strong>in</strong> <strong>the</strong> channel by Macrophyte Survey (MS) accord<strong>in</strong>g<br />
to changes <strong>in</strong> habitat conditions shown by <strong>River</strong> Habitat<br />
Survey (RHS). Birds were monitored us<strong>in</strong>g Common Bird<br />
Census (CBC) and W<strong>in</strong>ter Atlas (WA) surveys. For<br />
mammals, bat transect surveys were conducted and <strong>the</strong><br />
presence of otter Lutra lutra and Water voles Arvicola<br />
terrestris was to be recorded dur<strong>in</strong>g RCS.<br />
Macro<strong>in</strong>vertebrates and fi sh were monitored with more<br />
specifi cally designed sampl<strong>in</strong>g regimes. In <strong>the</strong> case of fi sh<br />
this provided an opportunity to compare two monitor<strong>in</strong>g<br />
techniques: standard depletion electric fi sh<strong>in</strong>g with<strong>in</strong> stopnets,<br />
and po<strong>in</strong>t-abundance sampl<strong>in</strong>g by electric fi sh<strong>in</strong>g<br />
(PASE) (Copp and Peñáz, 1988; Perrow et al., 1996). Follow<strong>in</strong>g<br />
this, <strong>the</strong> longer term response over a fur<strong>the</strong>r fi ve<br />
years was undertaken us<strong>in</strong>g PASE.<br />
The project had a general aim to restore a native fi sh<br />
community of Brown trout Salmo trutta and bullhead
160 <strong>River</strong> <strong>Restoration</strong>: <strong>Manag<strong>in</strong>g</strong> <strong>the</strong> <strong>Uncerta<strong>in</strong>ty</strong> <strong>in</strong> Restor<strong>in</strong>g Physical Habitat<br />
Cottus gobio, but no specifi c targets were set; not only<br />
was <strong>in</strong>formation on <strong>the</strong> fi sh population scarce, <strong>the</strong>re was<br />
also no move to ga<strong>the</strong>r more. However, <strong>the</strong> connection of<br />
<strong>the</strong> Misbourne to <strong>the</strong> <strong>River</strong> Colne was thought likely to<br />
provide a source of colonists at least to <strong>the</strong> lower reaches.<br />
Shardloes Lake, with a direct connection to <strong>the</strong> river, was<br />
also thought likely to operate as a refuge for fi sh <strong>in</strong> <strong>the</strong><br />
upper reaches, although this was also thought to have dried<br />
out <strong>in</strong> 1997/98.<br />
Comparison Between Fish Survey Methods<br />
Each site was sampled twice annually (spr<strong>in</strong>g and autumn)<br />
between July 1996 and October 1999 us<strong>in</strong>g both depletion<br />
fi sh<strong>in</strong>g and PASE. The same electrofi sh<strong>in</strong>g equipment was<br />
used for both methods. Depletion electric fi sh<strong>in</strong>g was<br />
undertaken with<strong>in</strong> a 100 m section created by <strong>the</strong> use of<br />
stop-nets at ei<strong>the</strong>r end to prevent <strong>the</strong> escape of fi sh. Two<br />
operators, each with a s<strong>in</strong>gle anode attached to a separate<br />
100 m cable and a hand net fi shed <strong>the</strong> section, explor<strong>in</strong>g<br />
<strong>the</strong> entire area with sweep<strong>in</strong>g movements. Two runs were<br />
conducted at all sites, apart from at Site 6 <strong>in</strong> spr<strong>in</strong>g and<br />
autumn <strong>in</strong> 1997, where emergent vegetation seriously<br />
hampered fi sh<strong>in</strong>g and only one run was attempted. All fi sh<br />
captured were reta<strong>in</strong>ed <strong>in</strong> one or two b<strong>in</strong>s carried by additional<br />
personnel and processed (identifi ed and measured,<br />
with some fi sh also be<strong>in</strong>g weighed to calculate length–<br />
weight regressions) at <strong>the</strong> end of each run, before release<br />
downstream of <strong>the</strong> stop-netted area. Us<strong>in</strong>g <strong>the</strong> numbers of<br />
each species captured <strong>in</strong> each run, an estimate of <strong>the</strong> total<br />
population size of each species was generated us<strong>in</strong>g <strong>the</strong><br />
maximum weighted likelihood model of Carle and Strub<br />
<strong>in</strong>d. m -2<br />
3<br />
2.5<br />
2<br />
1.5<br />
1<br />
0.5<br />
0<br />
(1978). Us<strong>in</strong>g <strong>the</strong> area of <strong>the</strong> section sampled (100 m ×<br />
width of <strong>the</strong> channel), quantitative estimates of numerical<br />
(<strong>in</strong>d. m −2 ) and biomass (g m −2 ) density were calculated.<br />
Prior to PASE, a volt meter was used to calibrate <strong>the</strong><br />
gear to provide a voltage gradient of 0.12 V at 45 cm away<br />
from <strong>the</strong> 40 cm anode. Such a voltage corresponds to <strong>the</strong><br />
m<strong>in</strong>imum effective voltage at which <strong>in</strong>hibited swimm<strong>in</strong>g<br />
occurs (Copp and Peñáz, 1988). The effective sampl<strong>in</strong>g<br />
area was thus calculated to be 1.3 m 2 from which quantitative<br />
estimates of abundance and biomass were calculated.<br />
A total of 50 po<strong>in</strong>ts were sampled with<strong>in</strong> a 200 m section<br />
of river directly upstream of <strong>the</strong> stop-netted section at each<br />
site. As <strong>in</strong> <strong>the</strong> depletion fi sh<strong>in</strong>g, <strong>the</strong> anode was attached to<br />
a 100 m extension cable from <strong>the</strong> control box, which was<br />
sited mid-way along <strong>the</strong> section. Po<strong>in</strong>ts were taken <strong>in</strong> a<br />
stratifi ed random manner at 4 m <strong>in</strong>tervals, with <strong>the</strong> electric<br />
fi sh<strong>in</strong>g operator mov<strong>in</strong>g diagonally upstream from bank<br />
to bank, and sampl<strong>in</strong>g a pre-determ<strong>in</strong>ed number of po<strong>in</strong>ts<br />
<strong>in</strong> <strong>the</strong> littoral marg<strong>in</strong>, which was determ<strong>in</strong>ed from <strong>the</strong> relative<br />
width of <strong>the</strong> marg<strong>in</strong> to that of <strong>the</strong> channel.<br />
Overall, both techniques tended to sample a similar<br />
number of species (Wilcoxon signed ranks test, n = 17,<br />
Z = −2.84, p = ns) and produce similar estimates of total<br />
biomass (n = 17, Z = −1.87, p = ns), but signifi cantly different<br />
estimates of total abundance (n = 17, Z = −3.10,<br />
p < 0.01) (Figure 8.5). PASE produced <strong>the</strong> higher estimates<br />
of total abundance on account of its tendency to<br />
produce signifi cantly higher estimates for bullhead (n =<br />
13, Z = −2.69, p < 0.01) and Ten-sp<strong>in</strong>ed stickleback Pungitius<br />
pungitius (n = 9, Z = −2.67, p < 0.01). There was<br />
little consequence of <strong>the</strong> production of signifi cantly higher<br />
biomass estimates for both <strong>the</strong>se species (p = 0.01 and<br />
gm -2<br />
Spr<strong>in</strong>g Autumn Spr<strong>in</strong>g Autumn Spr<strong>in</strong>g Autumn<br />
Spr<strong>in</strong>g Autumn Spr<strong>in</strong>g Autumn Spr<strong>in</strong>g Autumn<br />
1997 1998 1999<br />
1997 1998 1999<br />
PASE<br />
8<br />
6<br />
4<br />
2<br />
0<br />
DEPLETION<br />
Figure 8.5 Comparison of numerical (<strong>in</strong>d. m −2 ) and biomass (g m −2 ) density estimates obta<strong>in</strong>ed by PASE (mean ± 1SE shown) and<br />
depletion fi sh<strong>in</strong>g (density derived from total population estimate shown) over time at Site 5, dom<strong>in</strong>ated by small species such as<br />
sticklebacks, stone loach and bullheads