Extended abstracts of the OpenWater Symposium in pdf

OpenWater symposium 

19 th April, 2011, 

at 

UNESCO-IHE, Delft, The Netherlands 

Extended abstracts

OpenWater symposium 

19 th April, 2011, 

at 

UNESCO-IHE, Delft, The Netherlands 

Extended abstracts

Contents 

Invited presentations 

Sharing data sources for an integrated water resource assessment of the Black Sea 

catchment ...................................................................................................................... 1 

A. Lehmann, N. Ray, G. Giuliani, A. Barbosa, K. Abbaspour, S. Sozen, D. Gorgan, 

T. Uyttendaele 

The System of Systems approach and multi-disciplinary interoperability: the 

EuroGEOSS experience ............................................................................................... 4 

S. Nativi 

The What, Why and How of the WMO/OGC Hydrology Domain Working Group ... 5 

D. Lemon 

Open Source Software in GIS and Environmental Modeling ..................................... 8 

D. P. Ames 

Open Modelling Interface for integration of models and data ................................. 12 

S. Hummel 

The State of the Art and Practice of Integrated Environmental Modeling .............. 16 

A. Voinov 

Sharing Water Data Through Web Services .............................................................. 18 

D.R. Maidment 

Parallel session 1: Software technologies for integration 

Introducing Pipistrelle and the FluidEarth Software Development 

Kit for OpenMI version 2.0 ......................................................................................... 23 

Q. Harpham, G. Pearce1, G. Glasgow, A. Harper 

FluidEarth – use of the OpenMI tool Pipistrelle to promote integrated 

modelling in the Water/Environment sector .............................................................. 25 

G. Pearce, Q. Harpham, A. Harper 

Applied open standards in integrated water information management .................... 29 

M. Natschke, S. Fuest, R. Funke , J. Gregersen, U. Looser, I. Dornblut 

Comparison of time series ingest performance of various standardized 

file formats using DelftFEWS .................................................................................... 32 

P. Gijsbers, E. de Rooij

On the use of open standards and open-source libraries in Delta Shell ................... 35 

G. Donchyts 

Real time satellite observation of large basins using the ILWIS Open 

GEONETCast Toolbox .............................................................................................. 38 

C. M. Mannaerts, B. H.P. Maathuis, L. Wang, M. Schouwenburg B. V. Retsios 

OGC Hydro-Domain Working Group Surface Water interoperability experiment . 43 

P. Fitch 

WaterML2.0: Enabling water information exchange .............................................. 47 

G. Walker, P. Taylor 

Enabling Near Real-Time Water Resource Management via The Sensor Web ....... 49 

A. Terhorst, P. Taylor, B. Lee, C. Peters, C. Malewski 

A Prototype Provenance Management System for a Continuous Flow 

Forecasting System ..................................................................................................... 52 

C. Kloppers, C. Peters, Q. Liu, Q. Bai, P. Taylor, C. Malewski, H. Mueller, A. 

Terhorst, B. Lee, S. Zednik, P. West, P. Fox 

Managing Trust in Provenance-Aware Water Information Systems ....................... 55 

Q. Bai, Q. Liu, C. Kloppers, X. Su, H. Mueller, A. Terhorst 

HYDROSYS: On-site monitoring and management of water environment ............. 57 

A. Jolma, A. Nurminen, I. Ferencik, V. Lehtinen, M. Kamalov, E. Kruijff, E. 

Mendez, E. Veas, M. Parlange, S. Simoni, S. F. Vidal, T. Papaioannou, M. Bavay, 

N. Dawes, T. Drummond, E. Rosten, O. Kaehler, J. Mykkänen, J. Jemai 

The use of OpenMI in product development and project work ................................. 59 

P. Tacheci, Z. Nagy, S. Vaněček, J. Hurkens, T. Clausen, R. Cora, J. Gross, G. 

Philipsen, J. Hartnack 

Parallel session 2: enviroGRIDS project and related research 

HAWQS: multi-spatial-temporal scale hydrologic and water quality 

decision support system for continental US ............................................................... 65 

D. Deb, J. Veluppillai, R. Srinivasan, J. Arnold 

Automated Analysis of upstream-downstream relationships using 

Bayesian Belief Networks from spatially distributed SWAT models ........................ 68 

Y. Xuan, S. Yalew, X. Zhu, Z. Xu, A.van Griensven 

Information support for regional water resource management in Ukraine ............. 71 

Y. Makarovskiy 

Applying EnviroGRIDS Tools to Rioni Basin ........................................................... 74 

K. Allenbach, M.Gvilava

Integrated data collection and dissemination in North Bulgarian river basins ....... 77 

S.Balabanova, S.Stoyanova 

Software integration for flood management .............................................................. 86 

V. Moya, I. Popescu, A. Jonoski, D. P. Solomatine 

Dealing with uncertainties in remotely linked models .............................................. 89 

N. Kayastha, A. van Griensven, D. P. Solomatine 

Determining the effects of water quality and quantity on public 

health by using GIS technology .................................................................................. 92 

A. Ö. Doğru, S. Alkoy, N. Uluğtekin, F. B. Balçık, Ç. Göksel, S. Sözen 

Ecosystem monitoring using digital change detection method; example 

of Iğneada wetland forest ........................................................................................... 94 

F. B. Balcik, C. Göksel, G. Bozkaya, A. Ö. Doğru, N. Uluğtekin, S. Sozen 

Environmental information for citizens via mobile phones: case studies in the 

Province of Noord Brabant, The Netherlands ........................................................... 96 

A.Almoradie, A.Jonoski, S.J. van Andel, I.Popescu 

Study regarding delineation of flood hazard zones in Dej area ................................ 99 

F. Stoica, M.Sarb, H. Selagea, R. Dulau 

A Cloud-based Virtual Observatory for Environmental Science ............................ 102 

G. S. Blair and Y. El-khatib

Abstracts – Invited presentations

OpenWater symposium 1 

Sharing data sources for an integrated water resource 

assessment of the Black Sea catchment 

Anthony Lehmann 1,2 , Nicolas Ray 1,2 , Gregory Giuliani 1,2 , Ana Barbosa 3 , Karim 

Abbaspour 4 , Seval Sozen 5 , Dorian Gorgan 6 , Trui Uyttendaele 7 

and the enviroGRIDS consortium 8 

1 Institute for Environmental Sciences, University of Geneva, Switzerland 

2 UNEP/GRID-Europe, Switzerland 

3 UMA, University of Malaga, Spain 

4 EAWAG, Aquatic Research Institute, Switzerland 

5 ITU, Technical University of Istanbul, Turkey 

6 UTCN, Technical University of Cluj Napoca, Romania 

7 AnteaGroup, Belgium 

8 www.envirogrids.net 

With 30 partners distributed in 15 countries, the enviroGRIDS project is contributing to 

the Global Earth Observation System of Systems (GEOSS) by promoting the use of webbased 

services, according to the Open GIS Consortium standards (OGC), to share and 

process large amounts of key environmental information in the Black Sea catchment. The 

main aim of the project is to assess water resource in the past, the present and the future, 

according to different development scenarios. The objective is also to develop datasets 

that are compatible with the European INSPIRE Directive on spatial data sharing across 

Europe. The data and metadata gathered and produced on the Black Sea catchment will 

be distributed through the enviroGRIDS geoportal (www.envirogrids.cz). 

The enviroGRIDS project supports the objectives of the Integrated Water Resource 

Management (IWRM) process by bringing needed information on water quantity and 

quality to decision makers. The project is indeed addressing the objectives of the Water 

Framework Directive (WFD) by promoting a water basin approach to water management. 

The aim is to bring the best available scientific information to the concerned end users, 

such as the International Commission for the Protection of the Danube River (ICPDR) 

and the Commission on the Protection of the Black Sea Against Pollution (BSC), which 

are both partners in the enviroGRIDS project. One of the hopes of the project is to extend 

eastward the experience of the ICPDR. 

With the Metronomica – Integrated Decision Support System, the first step is to build 

scenarios of development by integrating the expected climate, demographic and land use 

changes along different story lines that are considering sub regional differences. The aim 

here is to set the past, present and future scenes for this entire region by building spatially 

explicit scenarios at a 1km resolution. The question is to explore how the important 

changes expected in these three dimensions are susceptible of influencing water resources 

in the future. Knowing the distribution of human population and land cover allows one 

also to assess the vulnerability to water shortage for drinking water, agricultural or 

industrial uses. 

UNESCO-IHE Institute for water education and enviroGRIDS project


The second step is to calibrate a hydrological model with the Soil and Water Assessment 

Tool (SWAT) that needs several key environmental information as input data; the needed 

data to run the model are digital elevation model, land cover, soil map, climate station 

data for temperature and rainfall; then hydrological data on water quantity and quality is 

needed to calibrate the parameters of the model. While most of these data are now readily 

available at the global scale and at high resolution, climatic and hydrological data remain 

harder to gather across different countries, as the access to such data is made complicated 

by different national policies that transform the efforts of the World Meteorological 

Organisation (WMO) into a labyrinth of administrative difficulties. In this period of 

international concern to preserve our climate, one could be surprised that the basic 

climatic data is almost kept hidden from the public, the scientists and the decision 

makers. If the data was made available through GEOSS for instance, the enviroGRIDS 

project could be extended globally quite easily. 

The next step of the project is to develop the capacity of geoscientists to use the new 

geoprocessing opportunities brought by distributed and high performance computing. The 

enviroGRIDS system resources and functionality are accessible to the large community 

of users through the portal that provides Web applications for data management, 

hydrologic models calibration and execution, satellite image processing, report generation 

and visualization, and virtual training center. Indeed, with the increasing quantity and 

quality of data sets that GEOSS and INSPIRE will make available at finer resolutions, the 

computing capacities of most organisations will be quickly overwhelmed. By porting 

several applications such as SWAT on the largest grid of computers (EGEE) available in 

the world, the enviroGRIDS project is paving the way to future distributed 

geoprocessing. This should become increasingly important in order to bring timely 

responses to scientists and decision makers who have to manage together our 

environment in a more sustainable way. 

One of the key factors of success of the enviroGRIDS approach is to transform 

heterogeneous sources of information provided by GEOSS web services into sound 

scientific information at the right spatial and temporal scales. This demonstration will be 

made on various so-called GEO Societal Benefit Areas (water, health, biodiversity, 

ecosystems, agriculture and energy) through several pilot studies in different countries. 

Here, the downscaling of raw data at the finest possible resolution followed by its 

upscaling through calculation of adequate sustainability and vulnerability indicators is the 

solution envisaged by enviroGRIDS. It is indeed important to recognize that 

environmental problems often operate at very local scales. Spatially explicit information 

is therefore the only way to bring suitable solutions at the right place and time. The 

advantage of fine scale data is also that it can be upscaled according to various spatial 

frameworks such as administrative boundaries or natural entities (e.g. river catchments, 

biogeographical regions). 

Finally, the interest of building sound scientific information on the state of the 

environment, now and in the future, would be of limited interest if we did not try to take 

advantage of the new communications channels offered by social networks such as 

LinkedIn for professional networking, Facebook for a broader public communication, 

Twitter for the participation of the public and factual communications, Flickr and 



YouTube for sharing images and videos. The numerous environmental issues at stake in 

the Black Sea catchment, as in the rest of the World, seem to largely justify an increased 

communication effort towards primarily concerned people, and not only towards their 

governments. As demonstrated recently, such new communication tools can play a key 

role to extend the democratic spirit in areas that seem to be condemned to survive in 

controlled information fluxes. 



The System of Systems approach and multi-disciplinary 

interoperability: the EuroGEOSS experience 

Stefano Nativi 

National Research Council of Italy 

c/o PIN – University of Florence, Piazza coardi, 25 – 59100 Prato, Italy 

Multi-disciplinary interoperability is essential to underpin the Earth System Science and 

contribute to address the challenges raised by the global changes. In fact, the scope and 

complexity of the Earth system investigations demand for the formation of distributed, 

multidisciplinary collaborative teams. Earth system analysis is a challenge for scientists 

as much as it is for information technology experts. 

A young scientific field called System of Systems (SoS) underpins the development of 

multi-disciplinary digital infrastructures (cyber-infrastructures) building on existing 

systems and teams. 

In spite of a large-scale integrated system, the SoS components can operate independently 

to produce products or services satisfying their customer objectives . The component 

systems may be connected by implementing one or more interoperability arrangements 

that do not require tight coupling or strong integrations. This allows a SoS to maintain its 

inherent operational character even as system components join or disengage from it. 

Since SoS is a construct of both legacy and new systems, an important feature is the 

attention to holistic aspects. For instance, GEOSS (Global Earth Observation System of 

Systems) will be a ‘system of systems’ consisting of existing and future Earth observation 

systems, supplementing but not supplanting their own mandates and governance 

arrangements (Group on Earth Observations). 

An Earth science SoS should build incrementally on existing information systems and 

should incorporate heterogeneous resources (or components) e.g. observation and data 

models, service interfaces, environmental modelling schemes, etc. 

Policymaking requires for multi-disciplinarily; building a SoS approach to support 

environmental policymaking promises to provide the necessary flexibility and scalability 

leaving the existing systems autonomous. 

The researches and developments carried out in the framework of the FP7 EuroGEOSS (a 

European approach to GEOSS) project will be presented and discussed. This project 

addresses a specific EC call for a “European Environment Earth Observation system 

supporting INSPIRE and compatible with GEOSS”. The multi-disciplinary operating 

capacity developed by project is described, with particular attention to its innovative 

Brokering framework. This framework provides the necessary flexibility required by SoS 

as well the de-coupling. Besides, this solution lower the present entry level barrier 

characterizing SOA (Service-Oriented Architecture) based infrastructures. 



The What, Why and How of the WMO/OGC Hydrology 

Domain Working Group 

Dr David Lemon 

Commonwealth Scientific and Industrial Research Organisation - CSIRO, Australia 

In March 2009, the Open Geospatial Consortium (OGC) formally created the Hydrology 

Domain Working Group (HDWG). Its purpose is to ‘provide a venue and mechanism for 

seeking technical and institutional solutions to the challenge of describing and 

exchanging data describing the state and location of water resources, both above and 

below the ground surface.’ 1 The purpose of this presentation is to describe why this group 

was formed, the work it is undertaking and how others can participate. 

The OGC is a consensus based standards organisation with the goal of the development 

of publicly available interface standards for geospatial data. Examples of commonly used 

OGC standards include: Web Map Service (WMS), Web Feature Service (WFS), 

Geography Markup Language (GML) and KML. 

This creation of the HDWG was significant for the OGC for two reasons. Firstly, the 

working group is a joint working group of the OGC and the World Meteorological 

Organisation (WMO) and the first of two such groups to eventually come under the 

governance of the Memorandum of Understanding (MOU) between the OGC and the 

WMO 2 . The primary point of engagement within the WMO for the HDWG is the 

Commission for Hydrology (CHy) which has an international mandate to publish water 

data standards. 

The reason for bringing the two groups together was to provide mutual benefit to both 

CHy and the OGC. In the case of CHy this meant: 

� access to the regular and more frequent OGC meeting schedule, 

� visibility and a level of accountability to a broader group of stakeholders, 

� access to interoperability expertise and a set of technologies and practices 

supporting consensus-based standards development among widely dispersed and 

disparate organizations, and 

� exposure to an ecosystem of tool-builders, from researchers and small companies 

through to some of the largest geospatial and generic software houses. 

1 OGC\WMO Hydrology Domain Working Group Charter, OGC 08-095r5 

[http://external.opengis.org/twiki_public/pub/HydrologyDWG/WebHome/Hydrology_DWG_Cha 

rter.doc] 

2 Memorandum of Understanding between the WMO and the OGC, 2009 

[http://www.wmo.int/pages/prog/www/WIS/documents/MOAWMO_OGC.pdf] 



From the OGC’s perspective, benefits of the relationship include: 

� rigorous testing of OGC technologies within a domain that uses a wide variety of 

"observations" (in-situ monitoring, imagery, gridded simulations); and 

� development of a profiling methodology for OGC standards which can be applied 

within other domains of interest. 

The second significant aspect associated with the creation of the HDWG was that this 

step set a new precedent for the OGC. Namely, they had created a group which would be 

focused on the creation of specifications for use within a particular field of interest, 

namely water, rather than the OGC’s traditional focus on particular technological 

solutions. Since this time, groups with a focus of meteorology, oceanography and 

aviation have also been created. 

Technical work within the HDWG is carried out by one of two paths. The first is one 

developed by members of the HDWG and is aimed at the development of candidate 

specifications for consideration by the OGC’s Technical Committee (TC). 

To date only one such specification has travelled this path. WaterML2.0, a specification 

for the encoding of water observations, is the result of efforts by members of the HDWG 

to harmonise work from the US, Europe and Australia with relevant OGC specifications. 

This work has recently moved into the formal part of the OGC’s standards development 

process with the creation of the WaterML2.0 Standards Working Group in March this 

year. 

Other potential specifications that the HDWG is considering for this process include: 

� Groundwater Markup Language 2.0 (GWML2.0) – an information model 

describing groundwater features and their relationships; 

� An information model describing surface water features and their relationships; 

� Key water related vocabularies; and 

� A reference architecture for a distributed, evolvable water information system 

The second mechanism by which technical work is undertaken within the HDWG is the 

Interoperability Experiment (IE). Interoperability experiments are a mechanism of the 

OGC and have specific requirements and processes associated with them 3 . In particular, 

IEs are tightly scoped to one aspect of work, run over a short timeframe (usually six 

months to one year) and are resourced by in-kind commitments by the participants. 

The HDWG has already undertaken one IE (Groundwater, with USGS and NR Canada 

leading a small number of organisations), which has now successfully concluded. In that 

experiment, WaterML2.0 was used to exchange groundwater information across the US- 

Canadian border, providing important feedback to the WaterML2.0 team and resulting in 

nine change requests to existing OGC standards. 

3 Interoperability Experiment (IE) Policies and Procedures, OGC 05-130r2, 2009 



A second HDWG IE expands on the work of the Groundwater IE focusing on the subdomain 

of surface water observations. This IE, is currently underway and aims to further 

advance the development of WaterML 2.0 and test its use with various OGC service 

standards applied to three use cases. It will also contribute to the development of a 

hydrology domain surface water feature model and vocabularies, which are essential for 

interoperability across the hydrology domain. 

A third experiment, focusing on forecasting, is currently being planned. This work will be 

undertaken in conjunction with the OGC’s Meteorology/Oceans Domain Working Group. 

Membership of the HDWG is open to all those groups with an interest in the sharing of 

water information. To date, the group has been extremely active with healthy, and 

growing, attendance at the quarterly meetings and full and interesting agendas for each of 

these meetings. The group also holds an annual, one week workshop which also well 

attended. Importantly, the group is not dominated by any one sector of the water industry 

with the research, government and industry equally contributing. 



Open Source Software in GIS and Environmental Modeling 

Daniel P. Ames 

Geospatial Software Laboratory, Idaho State University, Idaho Falls, USA 

Recent years have witnessed a growth in the quantity and quality of free and open source 

software (FOSS) tools for use in environmental data analysis and modeling. Such 

developments have significant potential to aid environmental managers, scientists and 

stakeholders in both developed and developing countries by reducing initial acquisition 

costs and providing a level of transparency that is often critical for regulatory, 

administrative, and on-going code development purposes. While commercial off-theshelf 

(COTS) environmental software tools play an important role in the environmental 

community, the concept of FOSS continues to attract interest among software developers, 

sponsors and end-users alike. This presentation will review several interesting and useful 

open source software tools in GIS and environmental modeling that are based on the free 

MapWindow and DotSpatial GIS packages with special focus on water resources systems 

including BASINS 4 and HydroDesktop. 

MapWindow GIS (www.MapWindow.org) is relatively unique among open source GIS 

tools because it is optimized for the Microsoft Windows operating system and is intended 

for use as a customizable data distribution system for developers and end users. The 

MapWindow GIS desktop application plug-in interface supports custom tool 

development by end users and the MapWindow Open Source team. MapWindow GIS 

has been adopted by the United States Environmental Protection Agency, United Nations 

University and others as a development and distribution platform for several water 

resources models including BASINS/HSPF, FRAMES-3MRA and SWAT. Other users 

and developers have modified and applied MapWindow GIS for use in the fields of 

transportation, agriculture, community planning and recreation. End users and 

developers around the world download MapWindow about 7000 times per month. 



The newest free library for embedding GIS capabilities within custom environmental and 

water resources modeling software is DotSpatial (www.DotSpatial.org) – a fully 

Microsoft .NET compatible library which is similar in concept to the ESRI ArcObjects 

product. DotSpatial aims to provide a free, open source, consistent and dependable set of 

libraries for the .NET, Silverlight and Mono platforms, enabling developers to easily 

incorporate spatial data, analysis, and mapping into their applications thereby unleashing 

the potential of embedded GIS in custom software solutions in a non-restrictive way. 

DotSpatial is under rapid development by the open source community – lead by 

researchers at Idaho State University – with a targeted full release date of June 13, 2011 

at the annual MapWindow/DotSpatial Conference in San Diego California (see 

www.MapWindow.org/conference/2011). 

Since 2006, the United States Environmental Protection Agency’s lead watershed 

analysis system, BASINS, has been released as a fully free and open source product built 

on the MapWindow libraries. As shown in the following figure, BASINS includes data 

pre-processing and post-processing modules that are built as MapWindow plugins, 

together with custom graphical user interfaces for specific watershed hydrology and 

water quality models. 



Key GIS functions provided by the MapWindow system and used in BASINS include 

watershed delineation, land use reclassifications, soils and land cover clipping to 

watershed boundaries, extraction of summary statistics from raster data for creating input 

files, creating a buffer around the watershed basin, computing upstream and downstream 

river connectivity, and all map display and printing functions needed to communicate 

model results. The MapWindow executable also provides a powerful plug-in interface 

used by BASINS to enable tight integration of separate models (e.g. AGWA, SWAT, 

HSPF, AQUATOX). This approach has encouraged other agencies and researchers in the 

United States and elsewhere to develop models and analytical tools that are compatible 

with and run as plugins with the BASINS framework. 

HydroDesktop (www.HydroDesktop.org) is a new desktop software application 

developed by the U.S. National Science Foundation-funded Consortium of Universities 

for the Advancement of Hydrological Sciences (CUAHSI) specifically to enable rapid 

search, discovery, download, and analysis of online water resources and climate time 

series data sets. The following figure shows the GIS based search and discovery interface 

and the time series data graphing interface for HydroDesktop. 



In summary, a number of free and open source tools for water resources and 

environmental analysis and modeling have been enabled in recent years through, in part, 

the release of requisite underlying free and open source GIS software platforms. While 

MapWindow and DotSpatial represent only a fraction of the number of interesting and 

useful open source GIS tools presently available (see more at www.osgeo.org), several 

important water resources and watershed analysis tools have been developed on these 

platforms and made available to the water resource community. We hope to see 

continued participation from the water resources community in the further development 

of both the overlying water resources tools as well as the underlying GIS platforms. 

Potential participants in any of the projects described herin are welcome and encouraged 

to contact Dr. Dan Ames at Idaho State University, dan.ames@isu.edu. 



Open Modelling Interface for integration of models and data 

Stef Hummel 

Deltares, P.O.Box 177, 2600 MH Delft, The Netherlands; stef.hummel@deltares.nl 

Introduction 

Integrated analysis often requires integrated modeling. This can be done by developing 

all-inclusive models, but it is preferable to be able to flexibly combine individual models 

or model components, that address specific domains, at run time. This can be realized by 

implementing a common standardized interface for model communication. In the water 

sector, in a series of EU-projects that focused on river basin management, the Open 

Modeling Interface (OpenMI, see [1] and [2]) has been developed in order to link 

together model components from various origins. OpenMI provides a standard model 

interface, and a Software Development Kit (SDK) to support existing models in adhering 

to that standard. The OpenMI standard is published by the OpenMI Association; the SDK 

is provided as an open source project by the OATC, the OpenMI Association Technical 

Committee (see [3] and [4]). 

Since the launch of the first OpenMI version at the end of 2005, the user and 

development community has grown steadily, and various well known models have 

become compliant. Because of limitations of this first version, some of the models did not 

follow the OpenMI standard interfaces exactly, but used slight deviations to achieve their 

goal in a similar style. Improvements were necessary to become a general interface 

standard that would not only cover water related applications, but also other domains. A 

core group of six institutes has worked on OpenMI 2.0, which is suitable for a large range 

of applications, from non-time dependent Geographical Information Systems (GIS) to 

e.g. master-slave controlled modeling frameworks. Version 2.0 of the OpenMI standard 

(see [5] and [6]) has been released in December 2010. 

OpenMI concepts 

OpenMI provides standardized interfaces to define, describe and transfer data between 

software components that run simultaneously, thus supporting systems where feedback 

between the modeled processes is necessary in order to achieve physically sound results. 

OpenMI allows the linking of models with different spatial and temporal representations: 

for example, linking river models and groundwater models, where the river model 

typically uses a one-dimensional grid and a short time step and the groundwater model 

uses a two- or three-dimensional grid and a longer time step. 

The OpenMI standard consists of a set of interface classes, specified in both Java and C#, 

that define the behavior of a model component, and that define which quantities can be 

exchanged by that component, at which locations and in what time frame. 



What, where, when 

The run time data exchange between model components is done by means of a 

GetValues(…) call, where the argument of this call specifies: 

� What is exchanged? 

This is defined by the IQuantity and the IQuality interfaces below. 

� Where is it exchanged? 

The location is specified by the so called IElementSet, a set of ID-based or Georeferenced 

locations (see table below). 

� When, i.e. at what times is the data needed? 

This is expressed by the ITimeSet, a list of time stamps or time spans. 

A quantity is specified by 

• Caption (“Runoff”) 

• Description (optional explanatory 

description) 

• Value Type (double, integer, etc.) 

• Unit: 

• Caption (“CFS” ) 

• Description (“Cubic feet per 

second“) 

• ConversionFactorToSI 

(0.0283168439 ) 

• OffsetToSI ( 0 ) 

• Dimension (e.g. L 3 T -1 ) 

UNESCO-IHE Institute for water education and enviroGRIDS project 

A quality is defined by its: 

• Caption (“Soil Type”) 

• Description (optional 

explanatory description) 

• Categories: 

• Caption (“sand 1”) 

• Description (“coarse 

sand“) 

• IsOrdered 

For the definition of locations, the ElementSet, various types are available: 

ElementType Description 

IDBased ID-based (string comparison). 

Point geo-referenced point in the horizontal (XY)-plane or in the 3dimensional 

(XYZ)-space. 

PolyLine geo-referenced polyline connecting at least two vertices in the 

horizontal (XY)-plane or in the 3-dimensional (XYZ)-space. 

Polygon geo-referenced polygons in the horizontal (XY)-plane or in the 3dimensional 

(XYZ)-space. 

Linking components 

A software component that implements the OpenMI standard is called a Linkable 

Component. It specifies its data exchange capabilities by defining input items and output 

items. Each input item and output item specifies its quantity or quality, its element set and 

its time set. 

The actual data exchange between components is established by defining 

provider/consumer relationships between output items and input items (see Figure 1). The 

GetValues() call mentioned above is performed on the output items.


Figure 1, Linking output and input items 

If the quantity, the ElementSet or the TimeSet of a certain output does not fit the way the 

input item requires it, the output can be adapted by adding an AdaptedOutput to the 

output. As the name indicates, an adapted output in its turn is just an output again, so 

subsequent adapted output items can be added to the initially created adapted output. 

Figure 2 shows an example of some sequences of adapted outputs. It may be clear that the 

adapted output approach offers great flexibility in defining the steps that have to be taken 

to transform that data from, for example, Output-1 to Input-2 and Input-a. 

Component 

1 

Output 1 

Output 2 

Output 3 

Spatial 

adaptation A 

Spatial 

adaptation B 

Time 

interpolation 

Time 

interpolation 

SI-conv. 

SI-conv. 

Figure 2, Flexibility in data transformations by means of Adapted Outputs 

Migrating models 

From the very start the OpenMI has been designed in such a way that it supports the easy 

migration of existing modeling systems. Generally speaking, a model in any 

programming language is made OpenMI compliant by re-organizing the its code in such a 

way that it has a separate initialization, computation and finalization routine, and can 

accept and provide input data and results, after which a wrapper is put around it (see 

Figure 3). The OpenMI SDK offers utilities to facilitate the development of such a 

wrapper. 


Input 1 

Input 2 

Input a 

Input b 

Input c 

Component 

2 

Component 

3


SetValues(…) 

GetValues(…) 

Figure 3, Wrapping native (e.g. Fortran) code in an OpenMI wrapper 

References 

[1] The OpenMI Standard, published by the OpenMI Association. 

Distribution, news and publications: http://www.openmi.org/, 

Developers and users: http://wiki.openmi.org/. 

[2] Gregersen, J.B. , P.J.A. Gijsbers and S.J.P. Westen (2007) OpenMI: Open modelling interface, 

Journal of Hydroinformatics Vol.9 No 3. pp 175–191 

[3] The OpenMI SDK (System Development Kit), provided by the OpenMI Association Technical 

Committee [3]. 

Information on developments: http://wiki.openmi.org/ 

Source development: http://sourceforge.net/projects/openmi/ 

[4] The OpenMI Association Technical Committee (AOTC), 

OpenMI's core developing team, participants (amongst others): 

Deltares (NL), Alterra (NL), DHI (DK), MWHSoft (GB), Bundesanstalt für Wasserbau (D). 

http://wiki.openmi.org/OpenMI+Association+Technical+Committee 

[5] Gijsbers, P., Hummel S., Vaneçek S., Groos J., Harper A., Knapen R., Gregersen J. Schade P., 

Antonelli A., Donchyts, G. (2010) From OpenMI 1.4 to 2.0. International Congress on 

Environmental Modelling and Software, July 5 - 8 2010, Ottawa, Ontario, Canada 

[6] Donchyts, G., Hummel S., Vaneçek S., Groos J., Harper A., Knapen R., Gregersen J. Schade P., 

Antonelli A., Gijsbers P. (2010) OpenMI 2.0 What’s New. International Congress on 

Environmental Modelling and Software, July 5 - 8 2010, Ottawa, Ontario, Canada 


Initialize computational kernel 

PerformTimeStep 

Finalize 

Initialize 

Get(varId) 

Set(varId) 

Get(varId) 

PerformTimeStep 

Finalize


The State of the Art and Practice of Integrated 

Environmental Modeling 

Alexey Voinov 

Faculty of Geo-Information Science and Earth Observation (ITC) 

University of Twente, The Netherlands 

WWW personal: http://www.likbez.com/AV 

As our impacts on the environment become more dramatic more interest is drawn to 

analysis of systems that span over several traditional scientific disciplines. These systems 

are more complex and have to be described by models that may have various components 

characterized by different scales, resolutions, and developed under different assumptions 

and paradigms coming from different scientific traditions and backgrounds. Integrated 

modeling is the method that is developing to bring together diverse types information, 

theories and data originating from scientific areas that are different not just because they 

study different objects and systems, but because they are doing that in very different 

ways, using different languages, assumptions, scales and techniques. 

There may be two ways of doing integrated modeling. One is to build the model as a 

whole. Here the modeling team collects data and information from various scientific 

fields, processes it, and translates it into one formalism. 

The other approach is to put together already built models. Since there are already 

numerous legacy models that have been carefully designed and tested in numerous 

applications by skilled researchers who are specialists in their corresponding fields of 

science, it makes perfect sense to try to reuse their products as building blocks for more 

complex systems. We only need to make sure that they can be linked together in a 

meaningful way matching the variables, scales and resolutions. Currently there are 

several efforts to develop the standards and software tools that would provide for this 

kind of integration. For these purposes, models are merely treated as software 

components that are to be made to work together and talk to each other. 

To distinguish between the two approaches, I will call the first type of models integral 

models, while the second type I call integrated models. The two approached to integrated 

modeling bear their own caveats and complications, which we consider in this talk. 

Building integral models is a rather conventional approach as we can see from their 

history. In most cases such models were built by integrating knowledge from different 

fields of science in the form of data, concepts, functions, approximations, etc. The 

models themselves were built from scratch and little care was given to future reuse of the 

pieces that went into the model formulation. The models were not designed for reuse or 

modification by others outside of the team. At some point it became clear that the model 

components could be reused or modified for future applications, and models started to be 

packaged as modeling systems. The Modular Modeling System (MMS, Leavesley, et al., 

1996) was one of the first attempts to make modules from integral models available for 

reuse. 

MMS and other such frameworks are not really designed to integrate legacy code as is, 

but rather they serve as tools for future model development, offering input, output and 

project support tools and common library standards, into which different modules are 



invited to be contributed. The future of these frameworks will very much depend upon 

how wide their standards will be accepted within the modeling community and whether 

there will appear a critical mass of contributed models. So far, unfortunately, there are 

very few examples of frameworks being used beyond their respective development 

teams. Since each module has to be rewritten in the language of the framework, there is 

less worry about the consistency of models that these frameworks are producing. All the 

scaling and linkage issues are expected to be solved when modules are tailored, adjusted 

and added to the framework libraries. 

The alternative approach is to support model integration at the level of existing models, 

allowing these models to talk to each other directly. There is much excitement currently 

about doing this and a number of model integration tools are available at the moment. 

Among others it is worth mentioning: 

• Common Component Architecture (CCA) from the Department of Energy and 

Lawrence Livermore National Lab; 

• ESMF sponsored by Department of Defense, NASA, the National Science 

Foundation, and NOAA, and widely used for high performance atmospheric and 

climatic modeling; 

• FRAMES from the US Environmental Protection Agency (EPA) used to model 

release, fate, transport, exposure, and risk associated with various contaminants; 

• The Open Modeling Interface and Environment developed by a consortium of 

European universities and private companies, as a standard for model linkage in the 

water domain. 

All these systems are primarily about software issues and the major concern is how to 

link models as software components. Actually, models are more than just software and 

there are issues that may not be easy to deal with at the software level. I use a number of 

metaphors to illustrate what are the concerns associated with module coupling: 

• Ugly constructs - by following only technical principles we may be giving birth to 

'integronsters' - constructs that are perfectly valid as software products but ugly and 

useless as models; 

• Skewed geometry - mismatched boundaries and boundary conditions; 

• Mismatched scales - both in time, space and structure (detail); 

• Confusion of tongues, where different scientific disciplines speak different languages 

and use different standards; 

• Complexity curse - model integration creates ever more complex structures that 

become impossible to analyze; 

• The vicious circle of calibration occurs when recalibrating one module starts a chain 

reaction through the whole integrated model, requiring recalibration of all other 

modules. 

There are a lot of gains that can be made if instead of mechanistically plugging modules 

together we take into account the specific goals and features of the system, and approach 

the problem with some creativity. The considerations presented here by no means 

devalue the importance and need for module linking software tools, component interfaces 

and specifications. I only remind that models are more than just pieces of software and 

they require more to make them work together efficiently. We need to be actively 

seeking for solutions based on modeling theory and not just on improved software 

development. 



Sharing Water Data Through Web Services 

David R. Maidment 

Leader, CUAHSI Hydrologic Information System Project 

and Director, Center for Research in Water Resources, University of Texas at Austin, USA 

The Consortium of Universities for the Advancement of Hydrologic Science, Inc, 

(CUAHSI) (http://www.cuahsi.org) is an organization formed in 2001, now representing 

125 US universities, which is supported by the National Science Foundation to develop 

infrastructure and services to advance hydrologic science. Since 2004, NSF has 

supported a CUAHSI Hydrologic Information System (HIS) project to enhance access to 

hydrologic information. CUAHSI HIS has defined a language called WaterML (Water 

Markup Language) for conveying times series of water observations data through the 

internet, including data measured at point locations concerning streamflow, groundwater 

levels, soil moisture, evaporation, snow, precipitation, climate and water quality. Data 

accessible in WaterML includes information from the USGS National Water Information 

System, EPA Storet, National Climatic Data Center, US Army Corps of Engineers, 

USDA NRCS and ARS, National Weather Service, NASA, and water observations data 

from 15 universities. CUAHSI has also compiled a national water metadata catalog at 

the San Diego Supercomputer Center and a uniform way of searching this catalog using 

standard terms, which provides access to 5.1 billion water observations data contained in 

23 million time series describing 18,000 variables measured at 1.9 million locations in the 

United States. 

The USGS and some other water agencies now publish some of their observations data in 

WaterML. EPA has developed a web services language called WQX (Water Quality 

Exchange) that conveys groups of water quality observations, and CUAHSI has 

developed a translator that converts WQX into WaterML so that time series of physical 

hydrology and water quality data can be acquired in a consistent way. 

The Open Geospatial Consortium (OGC) (http://www.opengeospatial.org/) is an 

international organization, representing about 400 companies and agencies, which has 

developed the most widely used standards for sharing geospatial data through the 

internet. In 2008, CUAHSI proposed to the OGC that there should be established a 

Hydrology Domain Working Group to harmonize WaterML with OGC standards, and 

later the OGC and the World Meteorological Organization expanded this mission to 

included joint development of data standards for hydrology, climatology, oceanography 

and meteorology. International interoperability experiments in data sharing for 

groundwater and surface water are being carried out, and a version 2 of WaterML is 

being proposed that is conformal with OGC standards. 

There are three main components of this water data services architecture – data Providers 

such as water agencies or scientific research groups who publish water data and metadata 

describing their information, data Catalogs such as a national water data portal or 

data.gov, which index and provide search capability for the information, and data 



Consumers who search metadata in the catalogs to identify the information they need, 

and then directly access that data from the water data service Providers. This triangular 

arrangement of Providers, Catalogs and Consumers is illustrated in Figure 2, and it 

follows the classic pattern of internet use where users first search a catalog like Google or 

Bing to find summaries and links, then go to the source to get the information. 

Figure 2. The interaction of water data providers, catalogs and consumers 

Hydrologic Information System 

This web-integrated Hydrologic Information System supports any number of data 

providers, catalogs and consumers. Each data Provider publishes its own data, metadata 

and catalog connection. The catalog connection can be registered in any number of 

portals, thus proving access to the information services to specialized outlets like 

drought.gov as well as generalized ones like data.gov. Researchers associated with the 

CUAHSI Hydrologic Information Systems team have built a prototype of the architecture 

envisaged, with a national water data catalog at the San Diego Supercomputer Center 

called HIS Central that indexes information from the USGS, EPA, NCDC, USDA, 

USACE and regional water sources, including NSF-supported observatories. In Texas, 

the Texas Water Development Board has assumed the responsibility of being the central 

data node for state water agency data, and the web application shown in Figure 3 searches 

across both federal and state catalogs to get dissolved oxygen observations for Texas and 

the surrounding area. 

Although the principal focus of the CUAHSI Hydrologic Information System project has 

been on water observations data, we have desired from the beginning of our project to 

provide access to an integrated array of information including static spatial water data in 

GIS, hydrologic modeling information, weather and climate model outputs, and remote 

sensing products, as illustrated in Figure 4. The information architecture we are 

envisaging is focused on water observations time series but is sufficiently general that it 

can also be applied to acquire the other information types as well, which form 

components of a Hydrologic Information System. 



Figure 3. A web application that searches across federal and state water data catalogs 

Figure 4. Components of a Hydrologic Information System 

Conclusion 

The purpose of this short paper is to describe a services-oriented architecture for web 

services for water data world-wide. The standards proposed are based on existing 

standards of the Open Geospatial Consortium and WMO, and extensions of those found 

to be needed to properly convey hydrologic time series and related water information 

through the internet. 

Reference 

OGC (2008), Open Geospatial Consortium Reference Model, document ref: OGC 08- 

062r4 http://www.opengeospatial.org/standards/orm 



Abstracts – session 1- Software technologies for integration 





Introducing Pipistrelle and the FluidEarth Software 

Development Kit for OpenMI version 2.0 

Quillon Harpham 1 , Geoff Pearce 1 , Gordon Glasgow 1 , and Adrian Harper 2 

1 HR Wallingford, Howbery Park, Wallingford, Oxon., UK 

2 MWH-Soft, Kestrel House, Wallingford, Oxon., UK 

FluidEarth is a platform for Integrated Modelling in the water (or, indeed, wider 

environmental) domain. It consists of a toolkit and a community of practitioners 

supported by an eInfrastructure including a library of models. The aim is to research and 

implement integrated computer modelling approaches to environmental systems. 

In order to understand how pressures such as climate change and developments impact 

the environment we need model not just physical, chemical and biological parameters, 

but how these parameters interact to affect the whole system. Environmental systems 

couple many natural processes and simulating them accurately demands modelling them 

in a similar fashion. Modelling such systems more accurately can be done in two ways: 

either simulate everything in one large model or link smaller models together. FluidEarth 

focuses on the second approach: linking existing computer models together to form 

integrated compositions. 

This is done by utilising the Open Modelling Interface (OpenMI) standard for integrated 

modelling. OpenMI provides a standard interface which allows models to exchange data 

with each other and other modelling tools on a time-step by time-step basis as they run. 

This enables the modelling of process interactions. The models may come from different 

suppliers, represent processes from different domains, be based on different concepts or 

have different spatial and temporal resolutions. Model components that comply with this 

standard can, without any programming, be configured to exchange data during 

computation (at run-time). This means that combined systems can be created, based on 

OpenMI-compliant models from different providers, enabling the modeller to use the 

models that are best suited to a particular project. The standard supports two-way links 

where the involved models mutually depend on calculation results from each other. 

Linked models may run asynchronously with respect to timesteps, and data represented 

on different geometries (grids) can be exchanged seamlessly. 

OpenMI itself is a ‘paper’ standard, independent of technical platform or implementation. 

It was stable and widely used at version 1.4 and, at a specially convened reception in 

Washington DC, was released at version 2.0 in December 2010. In order to facilitate the 

use of OpenMI, the custodian of the standard, the OpenMI Association, pledges to make 

two tools available:- 

� a Software Development Kit (SDK) to aid developers in adapting their models in 

order to comply with the OpenMI standard (a process known as ‘wrapping’); 

� a graphical user interface to facilitate the building and running of compositions of 

such OpenMI components. 



Having been heavily utilised against version 1.4 of OpenMI, the FluidEarth Software 

Development Kit has become the reference implementation SDK for OpenMI version 

2.0. Supporting any engine that can run as a subroutine through a simple C API, it allows 

developers to turn their models into OpenMI components. The idea is to take most of the 

complexity out of this process so that the additional development overhead is minimal. 

Extra skills required are also minimised. Examples exist for C#, VB and FORTRAN. 

The process involves isolating the initialise, time-step (or get-set) and finalise functions 

within the model code and allowing them to be accessed independently. This is to allow 

the model execution to be externally controlled in combination with other OpenMI 

components. 

This process results in the creation of an xml file associated with the model. This file 

includes a definition of: 

� ‘input exchange items’ – parameter values that the model is able to receive from 

other OpenMI components; 

� ‘output exchange items’ – parameter values that the model is able to pass to other 

OpenMI components. 

To accompany the Fluid Earth SDK, Pipistrelle is the reference implementation graphical 

user interface for OpenMI version 2.0. The purpose of this tool is to enable modellers to 

collect OpenMI components together easily and link them to form OpenMI compositions. 

Each component appears as a labelled box with links between output exchange items 

from one component to input exchange items from another denoted as arrows. As links 

are formed, the user is offered a list of the valid input and output exchange items to be 

matched. 

Following completion of the OpenMI composition, it can then be executed within the 

Pipistrelle tool. 

Pipistrelle and the FluidEarth SDK are both open source and available in the FluidEarth 

project on sourceforge: http://sourceforge.net/projects/fluidearth/. 

The FluidEarth eInfrastructure also includes a community portal and model catalogue. 

The portal is at http://fluidearth.net and includes downloadable releases of Pipistrelle and 

the FluidEarth SDK together with training material for new users in the eLearning 

library. There is also a discussion forum, news database and calendar of FluidEarth 

events. The model catalogue is at http://catalogue.fluidearth.net and contains metadata 

listings of model engines (the base code of the model itself) and model instances (the 

application of a model engine to a geographical area). 

Contact FluidEarth administration (details on the portal) if you would like to be issued 

with a FluidEarth id which allows you access to these facilities over and above that 

available to anonymous internet users. 



FluidEarth – use of the OpenMI tool Pipistrelle to promote 

integrated modelling in the Water/Environment sector 

Geoff Pearce 1 , Quillon Harpham 1 , and Adrian Harper 2 

1 HR Wallingford, Howbery Park, Wallingford, Oxon., UK 

2 MWH-Soft, Kestrel House, Wallingford, Oxon., OX10 8BA, UK 

Abstract 

FluidEarth provides facilities for integrated modelling and for discovery of wrapped and 

non-wrapped software model and their applications. As well as a mechanism for 

collaborating institutions to upload and share wrapped software, FluidEarth is a 

community of researchers interacting with model output users from the water industry. 

The FluidEarth portal, through which all FluidEarth facilities are displayed and accessed, 

is at http://fluidearth.net 

1 FluidEarth Programme 

FluidEarth provides access to a range of OpenMI-wrapped software, and software tools, 

to promote integrated modelling. Its objectives include developing the tools needed for 

analysing multi-parameter interactive processes, such as catchment or marine processes. 

A key tool is Pipistrelle, which provides an OpenMI Editor and a run environment in 

which OpenMI-wrapped software codes can be linked together and run in 

synchronisation with each other. This is a step forward on the existing editor presently 

available. The paper outlines how Pipistrelle (v1.0) has been successfully used in a 

number of applications in the UK and enabled multi-model integrated compositions to be 

created and run straightforwardly. The FluidEarth approach to OpenMI integrated 

modelling is intended for use by process scientists to install coupling between 

conventional models without recourse to bespoke re-coding (by software engineers). 

1.1 Community 

FluidEarth is a community initiative established in response to the need for a change in 

modelling capability for the simulation of complex and inter-acting systems in the water 

and environmental domain. It supports the uptake of OpenMI as a mechanism that 

enables complex water/environment systems to be simulated by integrating a number of 

smaller models in a modular manner. By creating a pool of researchers who agree to 

share their respective models with each other and create integrated model compositions, 

the initiative aims to promote the development of modelling facilities’ that are applicable 

to the real world problems that the group of users, who are also part of the community, 

are able to define. This process enables emerging integrated-model solutions to be pulled 

through into application by the water /environment industry. 

1.2 Tools 

The tools that have been developed under FluidEarth principally comprise:- 

Pipistrelle. The Editor/GUI/Run environment enables any software that has been wrapped 

using the OpenMI standard to be linked to another. How data is exchanged between the 

models used is the user’s responsibility. 



FE-SDK. The software development kit enables a user to write an OpenMI wrapper 

around their software code without recourse to advanced software engineering, so that it 

can be coupled to any suitable available code (also wrapped to OpenMI standard). 

1.3 FluidEarth Catalogue 

The FEC is a record of numerical water/environmental models in terms of both model 

engines, and model instances. It enables the discovery of such models and indicates how 

to gain access to them. The listings comprise the meta-data about the model, and enable 

increased use, integration and interoperability. 

1.4 Portal 

The FluidEarth portal is the centre both for exchange of information between users, and 

for exchange of software between modellers. It provides users with access to details 

about model applications for given locations/situations, and gives access to the software 

that can be downloaded. It contains a forum and self-training material. 

1.5 Shared software 

The Development Platform comprises a repository of different OpenMI-wrapped 

software systems available to FluidEarth partners for use in developing new integrated 

model compositions. Under the IPR protection provided by FluidEarth, the software can 

be used by all the partners according to the FluidEarth Agreement. 

The Demonstration Platform is a repository for completed software now available to 

external users – either open source or through commercial licensing. 

1.6 Membership 

Formal members of FluidEarth are institutions who have accepted and signed the legal 

agreement, which gives them full access to all available facilities. Through the public 

parts of the portal, users can register to be able to receive access to any open-source or 

freely available material. 

2 FluidEarth Applications – details of three recent use-cases 

2.1 Building collapse in response to flood wave 

This case study considered the development of depth-velocity-damage relationships. One 

of the more challenging aspects of probabilistic flood risk assessment is obtaining a 

credible estimate of the likely damage to buildings due to flood flows. The prediction of 

flood characteristics was improved by coupling fine- and coarse-grid simulations without 

incurring the computational overhead of high-resolution models. Going beyond the 

traditional depth-damage functions already in use, an OpenMI composition of high 

resolution hydrodynamic and building collapse models was used to construct depthvelocity-damage 

functions to evaluate damage-potential. It could also be used to improve 

the depth-damage functions currently used in system-level probabilistic flood risk 

assessment. The composition was effectively a ‘numerical flume’, where different dambreak 

scenarios could be systematically tested for damage to different types of building). 

It comprised: 

- a high-resolution raster flood-spreading model (‘2DCellModel’). 

- a building collapse model (Bernoulli total-head model; ‘BuildCollapse’). 



- a model to spatially aggregate depth and velocity in areas of interest 

(‘ZoneModel’). 

The composition was used to explore the damage potential close to the source of 

flooding. 

Fig 1 shows the predicted progression of flood waters for a given situation threatening a 

given orientation and design of building. 

2.2 Interaction between groundwater and floodplain under flooding 

This study has applied an integrated modelling approach to studying the flooding 

problems of the city of Oxford by linking a groundwater model (ZOOM3DQ) with a river 

flow model (Infoworks RS). The study supports the Environment Agency's research 

strategy on integrated modelling and has provided new versatile modelling tools for 

developing an improved understanding of groundwater modelling and wetland flow 

processes. 

Fig 2 GIS view of interaction between flood plain nodes in InfoworksRS and the top 

layer of the underlying ZOOM3DQ groundwater grid. 



2.3 Interaction between waves and currents 

The Coastal Evolution Model (CEM) comprises a number of coastal process models that 

can be linked together according to need. Pipistrelle can be extended using easily created 

plug-ins, for example custom interfaces (such as the control panel for the ‘Coastal 

Evolution Model’) can be developed. For this particular application, an alongshore 

transport model was to a cross-shore model. 

Fig 3 The CEM Control Panel facilitates set-up and running of such compositions. 

3 Conclusions 

The FluidEarth tools Pipistrelle and the FE-SDK (Software Development Kit) have been 

demonstrated to work effectively across a range of water/environmental processes, and 

are now freely available. 



Applied open standards in integrated water information 

management 

Michael Natschke 1 ,Stefan Fuest, Ph.D. 2 , Roland Funke 3 ,Jan Gregersen, 

Ph.D. 4 ,Ulrich Looser 5 , Irina Dornblut 6 

1 Product Manager Water Information Systems, KISTERS AG, Charlottenburger Allee 5, D-52068 

Aachen, Germany, Phone +49 241 9671-158, Michael.Natschke@Kisters.de 

2 Product Manager Web/GIS KISTERS AG, Charlottenburger Allee 5, D-52068 Aachen, 

Germany, Phone +49 241 9671-176, Stefan.Fuest@Kisters.de 

3 Product Director, Water, KISTERS AG, Charlottenburger Allee 5, D-52068 Aachen, Germany, 

Phone +49 241 9671-179, Roland.Funke@Kisters.de 

4 Manager HydroInform, Tingstedet 8, 4070 Kirke Hyllinge, Denmark, Phone +45 41 58 80 26, 

Gregersen@hydroinform.com 

5 Head Global Runoff Data Centre, Am Mainzer Tor 1, 56068 Koblenz, Germany, Phone +49 261 

1306-5224 looser@bafg.de 

6 Deputy Global Runoff Data Centre, Am Mainzer Tor 1, 56068 Koblenz, Germany, Phone +49 

261 1306-5265 dornblut@bafg.de 

Keywords: OpenMI, SOS, WaterML2, Interoperability Experiment, HydroDomain 

Working Group, Global Runoff Data Centre 

Since 20 years KISTERS is developing and delivering off the shelf solutions responding 

to requirements of a demanding water industry. The requirements on data acquisition, 

storage, organization, validation, analysis and integration and dissemination from the 

international market have been included in a reliable, scalable and controlled open multitier 

architecture. KiTSM is developed in JAVA and is designed to organize, compute and 

share time series mass data. 

Today the KISTERS user community consist worldwide of more than 3000 active users 

processing water domain meta data, gaugings, rating curves, water quality samples, time 

series and derived time series data products. The size of the hydrometric networks 

managed with KISTERS Solutions range from 10 to 100 000 measurement stations. 

Acting for their customers KISTERS continuously contributes to standardisation 

processes and products. Since 2005 KISTERS contributes and adopts standards 

developed by the CUASHI initiative. Since two years -as a member of the OGC 

HydroDomain working group- KISTERS is working closely together with the Australian 

CSIRO to develop and enhance WaterML 2.0 and its implementation into the Sensor 

Observation Service SOS. 

KISTERS is contributing to the surface water interoperability experiment 1 (cross border 

use case, international exchange of Water Levels on the River Rhine between SANDRE 

(France) and WSV (Germany) ). In addition KISTERS is leading the interoperability 



experiment 3 (publishing fluxes to the ocean from the Global Runoff Data Centre 

(GRDC)). 

Global data centres such as the GRDC are benefiting from the advances in open standards 

for the acquisition and dissemination of hydrological data and information. To date 

hydrological data are provided in a multitude of formats necessitating the development of 

complex import tools and converters. With the advent of open standards, the transfer and 

exchange of hydrological data amongst users such as the GRDC can be streamlined or 

even automated. The management and manipulation of hydrological data is done with 

commercially available software systems, such as WISKI provided KISTERS, used by a 

large community. The provision of the data and data products to the users is utilising 

open standards as tested in the surface water interoperability experiments. The 

combination of domain specific software with open standards is welcomed by the GRDC 

as it ultimately contributes to a more efficient operation of the data centre. 

The KISTERS Web Interoperability Solution “KIWIS“ has been developed as a 

contribution to both interoperability experiments. KIWIS offers combined services such 

as SOS, WOF, WMS and WFS for different data sources in one instance. 

As a standard requirement the KISTERS archives have to serve data to a range of 

modeling applications. In the majority of cases this data transfer is scheduled in time and 

is based upon a variation of model specific ASCII file formats. To ease this process and 



to provide the advantage of linking models to our customers the KISTERS time series 

server became OpenMI compliant by the end of 2010. 

With the KISTERS OpenMI wrapper, local and/or remote modeling users can establish 

the communication to a KISTERS time series server, search and identify the appropriate 

input time series and retrieve data through the internet directly into a chain of integrated 

modeling applications. Time series data from a KISTERS time series server located in 

Germany has been successfully feed into a model application in Denmark. 

The technology and quality of open standards are steadily enhanced. This paper outlines 

the experience made during the harmonisation and implementation process and discusses 

the potential of applied open standards in integrated water information management. 



Comparison of time series ingest performance of various 

standardized file formats using DelftFEWS 

Peter Gijsbers 1 , Erik de Rooij 2 

1 Deltares USA Inc; peter.gijsbers@deltares-usa.us 

2 Deltares, The Netherlands; erik.derooij@deltares.nl 

Over the past few years, the Open Geospatial Consortium has become the main body for 

the development of geospatial related IT-standards. Starting from the geographic data 

model domain, OGC widened its scope to services to enable data exchange or 

interoperability via the web. In 2004, OGC presented its vision on enabling sensor 

information to be exchanged via the web. The SWE initiative was born resulting in a 

number of standards to describe the data (SensorML, Observations & Measurements) as 

well as standards to provide associated services on the data (e.g. Sensor Observation 

Service). 

Within the hydrologic world, a lack of an internationally accepted data exchange 

standards has resulted in a wide variety of national ones. A situation which does not 

necessarily result in basin wide data exchange. To ease the sharing of hydrological data, 

WMO and OGC have joined forces to create a domain working group that will create an 

international data exchange standard with the name WaterML2. The Domain Working 

Group Hydro started officially in 2010 and is lead by CSIRO, Global Runoff Data Center 

(on behalf of WMO) and the San Diego Supercomputer Center. The domain working 

group hosts a design team as well as various Interoperability Experiments (IE) to evaluate 

the design of WaterML2 and other OGC standards and provide feedback to the 

developers. A Groundwater IE is recently completed, a Surface Water IE underway and a 

Forecasting IE is planned to commence later in 2011. The latter will partly be conducted 

in co-operation with the WMO/OGC-MetOcean domain working group. 

WaterML2 will compromise multiple aspects of the hydrological domain, amongst others 

observations and forecast time series, topography, water quality. So far, the focus of the 

domain working group has been on the time series aspect of WaterML2. The associated 

data model is derived from O&M2, the updated Observations and Measurements model 

from the OGC-SWE suite of standards. 

As part of the Surface Water Interoperability Experiment a use case is defined which 

assesses the capability of OGC-Sensor Observation Services (SOS2) to provide 

WaterML2 encoded time series in a timeliness and fashion which meets incremental high 

performance needs of a real time forecasting system. In this context, the forecasting 

system thus will act as a client application to the data delivery service. This activity is a 

step towards the Forecasting IE, where a forecasting system can receive data from OGCweb 

services as well as deliver forecast products via OGC-web services. 

To evaluate the capability of the OGC standards, a comparison will be conducted against 

various other file formats in use with hydrological forecasting. The software used for this 



experiment will be DelftFEWS, an operational forecasting system used by a wide variety 

of agencies around the world. The file formats assessed will be: 

� SHEF, the Standard Hydrologic exchange Format widely used by federal agencies 

in the US. 

� FEWS-PI, the native Published Interface xml-format of DelftFEWS 

� O&M2.0, the OGC standard for Observations & Measurements, version 1.0 

� WaterML2.0, the standard format under development by the OGC-DWG Hydro 

The capability to support real time high performance needs of operational forecasting 

systems will be evaluated against the following criteria: 

� code complexity (i.e. number of code lines needed for the data parser) 

� file size (i.e. number of bytes to transport) 

� ingest performance (i.e. milliseconds to complete the ingest) 

To minimize impact of network performance on the experiment, the main focus is on the 

ingest of a set of scalar time series from files available on local disk. A variety of 

observations has been collected to account for diversity in time series length, time series 

regularity, number of phenomena (in SHEF: data types, in PI parameters) and features of 

interest (FoI; in SHEF: stations, in FEWS-PI: locations). 

Code complexity 

DelftFEWs offers a utility library for time series objects, which can be used by a java 

class that parses the data into the library object. The code complexity is assessed by an 

inventory of the lines of code needed by such java class, the associated size of the java 

class in bytes as well as the additional libraries needed to support the ingest. The result, 

presented in the table below indicates that the WaterML2 parser seems straightforward, 

although one should note that the parser only support the GetObservations call. 

FEWS-PI NWS-SHEF OGC-O&M OGC-WaterML2 

code size (lines) 824 396 303 267 

code size (b) 34,025 15,005 13,871 12,420 

support lib size (Kb) 0 0 8,911 0 

File Size 

To exclude network behaviour from the experiments, but still obtain a feeling of the 

impact of the standard on performance in a networked environment, the file size is 

assessed, both unzipped as well as zipped. Evaluating this aspect also requires attention 

for compression rates. Since compressing and decompressing requires time and resources 

as well, transmission of files with low compression rates might turn out to be faster 

uncompressed than compressed. No figures are available yet at the moment of writing. 



file size unzipped (b) FEWS-PI NWS-SHEF OGC-O&M OGC-WaterML2 

one FoI; one Phenomena 

multiple FoI; one Phenomena 

one FoI; multiple Phenomena 

multiple FoI; multiple Phenomena 

file size zipped (b) FEWS-PI NWS-SHEF OGC-O&M OGC-WaterML2 





avg.compression rate 

Ingest performance 

This criteria purely focuses on the number of milliseconds required to ingest the various 

files. No figures available yet at the time of writing 

ingest time unzipped (millisec) FEWS-PI NWS-SHEF OGC-O&M OGC-WaterML2 





ingest zipped (millisec) FEWS-PI NWS-SHEF OGC-O&M OGC-WaterML2 





Conclusion 

The conclusions of this experiment will be presented on the Poster. The conclusions will 

also be documented in the final report on the Surface Water Interoperability Experiment 

by the OGC-Domain Working Group Hydro. 



On the use of open standards and open-source libraries in 

Delta Shell 

Gennadii Donchyts 

Deltares, The Netherlands; gennadii.donchyts@deltares.nl 

Keywords: integrated environmental modelling, geospatial, model coupling, opensource, 

open standards 

The use of standards in integrated environmental modelling has increased in the last 

decade. The standards that made life of modellers a lot easier are found in the fields of 

model coupling, data exchange, storage and visualization. Delta Shell is a new integrated 

modelling environment being developed at Deltares. The system allows creating, storing 

and visualizing different types of environmental data, like time series, networks, GIS 

features, and other data types defined in the spatio-temporal domain and used by 

environmental models. This presentation will focus on the use of various open-source 

libraries, advantages and pitfalls. In the second part experiences on the use of the various 

open-standards such as OGC and OpenMI will be presented. 

Figure 1 Architecture of the Delta Shell 



Delta Shell is being developed as a model-agnostic modelling environment (without 

knowledge about specific model engine or its data). It consists of a set of common class 

libraries (see Figure 1), mainly used to describe different domains and software standards 

such as OGC-based geometries and features, surface water objects, mathematical and 

physical objects. Delta Shell defines on a very high level how a typical console or 

graphical user interface application is constructed. It also provides an API which allows 

extending the system by means of plugins. This includes both non-gui plugins providing 

new data types or computational models to the system, but also graphical user interface 

components used for data visualization. 

Many Delta Shell components rely on the open-source libraries. The following list 

summarizes most important open-source libraries used in Delta Shell: 

� NHibernate and SQLite – file format of Delta Shell, storage for user project. 

� NetCDF – multi-dimensional file format and library used to store model results 

data, based on UCAR Java implementation converted using IKVM.NET. 

� OGR / GDAL – geospatial libraries providing access to hundreds of the raster and 

vector file formats 

� Log4net – Apache logging library, allows level-based logging 

� PostSharp – aspect-oriented library 

� DLR –extends system with a high-quality scripting engine using IronPython. 

� Netron Graph Library – diagramming library 

With increasing availability of detailed geospatial information, GIS becoming a crucial 

part of any integrated modelling system. In some cases it is used only as a data 

visualization tool, but also it is increasingly used to generate and store schematizations 

used by numerical models, including all topological relations between different features 

of that schematization. The goal of the DeltaShell development in relation to modelling 

of the HydroNetwork was to provide a lightweight class library which is based only the 

open-source class libraries and is not necessary specific to model engine like SOBEK 1D 

FLOW. 

DeltaShell is based on the following open-source geospatial libraries: 

� GeoAPI.NET – definition of OGC simple geometry specifications 

� NetTopologySuite – implementation of OGC simple geometry specifications 

� Proj.NET – coordinate system transformation library 

� SharpMap – mapping library allowing working with vector and raster layers 

� BruTile – library which allows using OpenStreetMaps and Bing within .NET 

applications 



While these libraries were sufficient to implement a lightweight GIS subsystem, they 

were lacking some very important GIS concepts like Feature and Coverage. As a part of 

DeltaShell development these libraries were significantly reworked and extended with the 

following: 

� Geospatial Feature – OGC simple feature specifications 

� Coverage – OGC-like implementation of interpolated continuous coverages 

o Regular grid coverage 

o Discrete curve coverage, used to define data on 1D network 

� Network library, used as a basis of the HydroNetwork 

� Mapping functionality 

o Advanced map control supporting parallel layer rendering 

o Geometry and network editing functionality supporting snapping, 

topologies 

o Vector and raster layer style editors 

During development of Delta Shell many standards were used, mainly from the field of 

the geospatial software. For example OGC standards were great source of inspiration; 

even many of them do not provide implementation. These include OGC Abstract 

Specifications: Topic 1 - Feature Geometry, Topic 5 - Features, Topic 6 - Schema for 

coverage geometry and functions. Coverage model implemented in Delta Shell is 

conceptually based on ISO19123, but differs a lot in implementation details. 

CF conventions are used in the Delta Shell where appropriate to store model results in 

NetCDF files in a way that other software will be able to interpret them correctly. 

Hydrologic Features – harmonized Surface Water Feature Model is not available yet, but 

being developed as a part of Hydrology OGC working group. Current Delta Shell 

implementation is mainly inspired by ArcHydro object model. 

OpenMI – a prototype plugin is available within Delta Sell which can be used to 

integrated external models. 

DeltaShell is currently used at Deltares as the basis for a new generation of integrated 

modelling environment, and it is provided upon request as an open environment to others. 



Real time satellite observation of large basins using the 

ILWIS Open GEONETCast Toolbox 

Chris M. Mannaerts 1 , Ben H.P. Maathuis 1 , Lichun Wang 1 , Martin Schouwenburg 2 

and Bas V. Retsios 2 

1 Department of Water Resources & 2 Department of Geo-information Processing 

Faculty of Geo-information Science and Earth Observation (ITC) 

University of Twente, the Netherlands 

Abstract 

There is a growing availability of earth observation (EO) data and derived products for 

use in hydrology, water resources research and professional practice. GEONETCast 

operated by EUMETSAT and the ESA-DDS or the European Space Agency direct data 

dissemination system are examples of such global real time EO data dissemination 

systems. Currently also more and more software packages can be legally downloaded and 

used with little restrictions, including source code access. The use of ILWIS Open (3.7), 

an open source geospatial analysis system, is shown with emphasis on hydrological 

application. Real time satellite data import and visualization, land surface and digital 

terrain parameterization, as well as spatially distributed computation of water cycle 

components, like rainfall and evapotranspiration rates is also illustrated. 

Satellite data acquisition and access 

Real-time satellite observation capabilities have been greatly enhanced in recent years by 

data infrastructures such as GEONETCast, the global satellite and in situ data 

dissemination system of GEOSS, the emerging Global Earth Observation System of 

Systems, an initiative led by the Group on Earth Observation [1, 2]. 

GEONETCast permits open and free data reception through a global network of 

communication satellites, and provides near real-time earth observation and 

environmental data and derived products. This DVB-S or digital video broadcasting by 

satellites is emerging rapidly as a very effective way for persons around the globe to 

receive satellite and environmental data and 

geo information instantaneously (Fig.1). 

The EUMETCast dissemination system 

started off to provide the European and 

African national meteorological centres with 

satellite information mainly focussing on 

meteorological applications; currently the 

system is rapidly expanding and is now 

disseminating a wide array of environmental 

data from various third party data providers. 

More details on images and products can be consulted at the “Product Navigator”, 

available at the home page of EUMETSAT (http://www.eumetsat.int). 



Data processing and analysis 

The Geonetcast Toolbox is open source software plug-in of ILWIS Open [3]. It enables 

direct import and management of GEONETCast satellite data streams and supports their 

subsequent processing using ILWIS or other geospatial analysis systems. The software 

design principles of the toolbox were an easy operability, open source a/o freeware 

software components and an interface, adaptable by the users to their own selected data 

streams, data analysis, processing needs and information dissemination requirements, like 

e.g. web mapping services. The toolbox setup and key features are shown in Figure 2. 

The key features of the ITC Geonetcast Toolbox can be described as follows: 

� Satellite data reception & archive 

� Data reception via DVB antenna using EUMETCast / GEONETCast (or ftp) 

� Global geographical coverage combining EUMETCast & CMACast services 

� Selective archiving according user preferences e.g. satellite, data type, segment 

selection, time of storage, using an build-in data reception manager 

� Near real-time image processing & 

applications 

� ILWIS Open v.3.7 with full image 

analysis - GIS functionality with 

vector, raster, database, spatial 

modelling, visualization modules 

� Build-in Meteosat MSG data 

browser and retriever 

� Multiple data import and format 

routines (64) i.e. BUFR, GRIB, 

netCDF, GeoTIFF, NAS, HDF and 

other formats using Open GDAL 

geospatial data library libraries routines and freeware tools 

� Toolbox sample library (with processing and example application development 

routines) 

� Visualization and web-based services 

� Web-based client/server model using HTML and XML languages 

� WMS or web mapping services client 

For an overview of the standard ILWIS remote sensing and GIS functionality, we refer to 

[4] and the 52north.org (www.52north.org) website for download and information. This 

geospatial analysis software combines image processing with raster and vector data 

processing, using a high geo referencing accuracy and contains extended projection and 

data format exchange libraries. New versions (v.3.6 and higher) have a modular setup, 

permitting advanced users to create and add own plug-ins like additional processing 

capabilities, model couplings, etc. 



Hydrological application examples 

ILWIS Open also contains hydrological functionality permitting to use satellite 

observations and data for monitoring and management of water resources. Its geographic 

data handling permits to combine spatial and hydrological analysis. This permits to 

address spatial scales ranging from very small regions of interest, to large river basins or 

global area mapping. 

� Weather and rainfall 

An example of direct satellite data import for hydro meteorological purposes are 

Meteosat second generation (MSG) data and derived weather and hydrological products. 

Figures 3a and 3b illustrate the processing of MSG data (2011-03-13 1200 UTC) for the 

Black sea region with Fig.3a, the HRV high resolution visible 1 km channel and Fig.3b, a 

daytime false colour image after resampling of the 2 VIS and NIR MSG 3 km resolution 

bands from the geostationary MSG projection to a UTM projection - 1km resolution 

using the Ilwis Open GEONETCast Toolbox MSG data retriever (automated processing). 

Several atmospheric products as well as land surface and land cover products can be 

obtained from the 12-channel SEVIRI sensor on-board the MSG. 

Fig.3a and 3b: MSG images of Black Sea and eastern Mediterranean on March, 13 2011 1200 UTC 

Hydrological products of satellites are for example satellite rainfall estimates, using 

algorithms which may be combined with ground observations (level 3 products). Figure 4 

shows the MSGMPE or multi sensor 24-hour precipitation estimate from MSG, for 

March, 13 2011 and derived from the geostationary MSG thermal (10.8 μm) channel in 

combination with the TMI Microwave imager on-board the DMSP constellation [5]. 

This satellite precipitation estimate, available at 15-minutes intervals or frequency, 

permits to monitor well the spatial and temporal distribution of precipitation intensities, 

as shown by Figure 4, indicating very high rainfall on the northern slopes of the Alps 

(Swiss, Southern Germany and Austria), a hydro meteorological phenomena with high 

potential hydrological impacts such as high river discharges and local or regional 

flooding in e.g. the Danube river basin. 



Fig.4: MSG Multisensor precipitation estimate (24-hour total) for March, 13 2011, illustrating large 

rainfall amounts on the northern slopes of the Alps during that day. 

� Digital topography and river basin geomorphometry 

The digital elevation or DEM Hydro-processing module for automated hydrological 

network analysis and hydrological basin parameterization is available in ILWIS. The 

drainage network extraction routine has 

advanced and multiple options and permits to 

delineate river networks in very flat 

floodplain areas or depressions, a known 

problem issue of concern in much other 

available software for automated basin 

delineation [6]. DTM generated basins and 

parameters can then be coupled to rainfall – 

runoff models using geomorphologic 

information like e.g. GIUH (geomorphic 

instantaneous unit hydrograph) approaches to 

generate peak discharges and stream flow 

estimates at basin outlets [7]. 

� Evapotranspiration using Surface energy exchanges and balance 

Another hydrological feature included in Ilwis Open modular system is the SEBS plug-in 

giving users the ability to perform semi-automated satellite-based computations of 

evapotranspiration using the surface energy balance method [8]. The SEBS algorithm can 

be implemented, using a variety of satellite source data (e.g. MODIS, ASTER, MSG, 

NOAA and other), pending on availability, user preference and scale of observation. The 

modelling system permits to generate daily (or even sub-daily using geostationary orbits) 

estimates as shown in Figure 5, illustrating the use of MSG and limited ground data to 

estimate daily evapotranspiration fields (3 km – 24hr) over the Zambezi river basin [9]. 

Several more options for retrieval of satellite, in situ data and analysis of hydrological 

phenomena are available under ILWIS Open. New open plug-ins for coupling of 

hydrological models to satellite and in situ data are under continuous development. As 

such the Integrated Land and Water Information System is supporting the water resources 

community and users to generate and analyse satellite data-driven near real-time 



scenarios in river basins and forecast or predict changes or environmental risks more 

soundly. 

Figure 5: Estimate of daily evapotranspiration fluxes over the Zambezi river basin using Meteosat satellite 

data, standard ground meteorological observations and the SEBS modular plug-in of Ilwis Open 

ILWIS and the GEONETCast Toolbox are open source contributions, developed mainly 

at ITC, and available under 52north.org, an initiative for Geospatial Open Source of ITC - 

University of Twente with the University of Munster (Germany) and other partners. 

References 

[1] GEO Group on Earth Observations, “Ten year implementation plan (2005-2015)”, GEO Doc., 

ESA Publication Division, ESTEC, the Netherlands, 2005. 

[2] Mannaerts, C.M., Maathuis, B.H.P., Molenaar. M. and Lemmens, R. 2009. The ITC GEONETCast 

toolbox: a geo capacity building component for education and training in global earth observation and 

geo information provision to society. In: IGARSS 2009: Proceedings of the 2009 IEEE international 

geoscience and remote sensing symposium: 2009, Cape Town, SA ISBN 978-1-4244-3395-7. 

[3] B.P.M. Maathuis, C.M.Mannaerts and B. Retsios. “The ITC GEONETCast - toolbox approach for 

less developed countries”. ISPRS 2008: Proceedings of the XXI congress: Silk road for information 

from imager, 3-11 July, Beijing, China, ISPRS, 2008. pp. 1301-1306. 

[4] Unit Geo Software Development, 2001. ILWIS 3.0 Academic User Guide. International Institute 

for Geo-Information Sciences and Earth Observation (ITC), Enscheda, the Netherlands, 

[5] DMSP. US.Defense Meteorological Satellite Program. 

[6] Hengl, T., Maathuis, B.H.P., Wang. L., 2009. Developments in Soil Science, Volume 33, 2009, 

Elsevier, ISSN 0166-2481. 

[7] Ngyen, H.Q., Maathuis, B.H.P. and Rientjes, T, 2009. Catchment storm runoff using the 

Geomorphic Instantaneous Unit Hydrograph, GeoCarto International, Vol.24 (2009), pp.357-375. 

[8] Z. Su, “The Surface Energy Balance System (SEBS) for estimation of turbulent heat fluxes”, 

Hydrology and Earth System Sciences, 6(1), 85-99, 2002. 



OGC Hydro-Domain Working Group Surface Water 

interoperability experiment 

Peter Fitch 

Commonwealth Scientific and Industrial Research Organisation - CSIRO, Australia 

This paper briefly describes work being undertaken to further develop an international 

standard for water information exchange, the context of the OpenGeospatial Consortium 

and how it and domain interoperability experiments are being used to progress this work. 

We also discuss progress to date and some of the issues already identified. The goal of 

this paper is raise awareness of the work within the hydrological community. 

The OpenGeospatial Consortium (OGC) is a consensus based standards organization with 

the goal of the development of publicly available interface standards for geospatial data. 

OGC undertakes its standards development work within the Standards Program (SP), 

which consists of a number of Standards Working Groups (SWG’s). SWG’s are primarily 

responsible for drafting candidate standards that are endorsed by the Technical 

Committee (TC) and then promulgated. One of the principles adhered to by the OGC is 

that interoperability is achieved through the development and subsequent profiling of 

abstract specifications. Domain Working Groups (DWG’s), also part of the SP, are 

created with the purpose of refining and profiling these abstract standards such that they 

are suitable for domain use. 

The Hydrology Domain Working Group (HDWG), is a joint working group with the 

World Meteorological Organization (WMO) and has been established for this reason, 

focusing on the development of an international standard for water information known as 

WaterML2.0 , which is a profile of Observations and Measurements 2.0 1 . Other goals of 

the working group are the profiling of OGC service standards so they are suitable for the 

exchange of WaterML2.0 as well as the identification of a suitable Feature model and 

controlled Vocabularies. The HDWG was initiated in June 2009 at the Boston OGC 

Technical Committee meeting. 

Technical work within the HDWG is carried out by way of Interoperability Experiments 

(IE’s). Interoperability experiments have specific requirements and processes detailed by 

the OGC 2 . The experiments are tightly scoped to one aspect of work, run over a short 

timeframe (usually six months to one year) and are resourced by in-kind commitments by 

the participants. The HDWG has already run one IE (Groundwater, with USGS and NR 

Canada being participants), which has now successfully concluded. In that IE, 

WaterML2.0 was used to exchange groundwater information across the US-Canadian 

1 OGC Abstract Specification - Observations and Measurements, OGC 07022r1/OGC 07-002r3. 

2 Interoperability Experiment (IE) Policies and Procedures, OGC 05-130r2, 2009. 



border, providing important feedback to the WaterML2.0 team and resulting in nine 

change requests to existing standards. 

The second HDWG IE, expands on the work of the groundwater IE, and focuses on the 

sub-domain of surface water observations. In particular the IE will further advance the 

development of WaterML 2.0 and test its use with various OGC service standards applied 

to three use cases. It will also contribute to the development of a hydrology domain 

feature model and vocabularies, which are essential for interoperability across the 

hydrology domain, although these are not the main focus for the IE. 

Surface water datasets typically contain a large number of observations at a small number 

of locations, which will test WaterML 2.0 in new ways. This contrasts and complements 

the groundwater IE in which there are many locations with a comparative small number 

of observations. We expect that the project will discover issues related transferring large 

timeseries of observations. It is expected that the work required to support IE needs by 

WaterML2.0 will be small, and that the bulk of the requirements will be in the areas of 

profiling O&M and OGC services and the associated performance of these services and 

clients. 

The three surface water IE use cases are: 

1. Cross Border Data Exchange Use Case-led by disy: In this use case the user 

will discover surface water data from cross border regions via web map client and 

then visualize the time-series via web charts. The user can also download the data 

of interest in an appropriate format. 

2. Forecasting Use Case-led by NOAA/NWS: The user will discover and 

download data suitable for a streamflow forecast. The user will be able to feed a 

streamflow forecast model with this data, but the modeling itself is not part of the 

scope of this IE. 

3. Global Runoff Use Case-Led by Kisters: The goal for this use case is to provide 

automated monthly and yearly volume calculations from large rivers discharging 

to the oceans. 

The Surface water IE is being led by CSIRO who was nominated as co-ordinator and 

technical lead, with additional participation by: disy Informationssysteme GmbH (disy, 

DE), US Geological Survey (USGS, US), IOW-Sandre (FR), Commonwealth Scientific 

and Industrial Research Organisation (CSIRO, AU), Kisters (KISTERS, DE), Service 

Center Information Technology of the BMVBS (DLZ-IT BMVBS, DE), Consortium of 

Universities for the Advancement of Hydrologic Science (CUAHSI, US), Bureau of 

Meteorology(BoM, AU), University of Calgary (CA), NOAA/NWS (US), Deltares USA 

(US), 52° North Initiative for Geospatial Open Source Software (52NORTH, DE) 

The work-plan for the IE consists of four phases to coincide with four OGC TC meetings 

running from June 2010-June 2011. The strategy is to use these quarterly TC meetings for 

each use case to demonstrate progress, highlight issues, culminating in a final 

demonstration and report in September 2011. Each phase/demonstration should 



incrementally build upon the previous (an iterative approach) with the goal of having the 

final systems realize the use case goals by June 2011. 

As of late February 2011, a number of Sensor Observation Services (SOS) to deliver 

surface water observation have been set up (USGS, CSIRO, Kisters, IOW-Sandre, disy, 

GRDC), although not all of them are using WaterML2.0. USGS and Kisters are the 

exceptions, having developed a prototype SOS services which are delivering candidate 

WaterML2.0 spec documents. Similarly, a number of clients have been 

developed/modified for use in the demonstrations to date (Kisters, 52 North and 

University of Calgary). 

The IE 3 has found that across different implementations of SOS used for transferal of 

water observations, there are at least three different incompatible architectural approaches 

to the conceptual architecture of an SOS instance. These incompatibilities relate to how 

O&M has been interpreted in SOS, and how to get SOS to work best with our use cases. 

Most of these issues have been settled as part of a surface water IE workshop held in 

September 2010. At this workshop agreement was reached on a number of key issues 

relating to the consistent use of O&M and Sensor Observation Service. The two main 

points of this are that OM_Procedure shall be constrained to a single conceptual process, 

either a sensor or algorithm; Eg. Daily mean and that OM_FeatureOfInterest shall be 

constrained to a sampling feature usually a sampling point in space e.g. sampling station 

or point. The reader will require some familiarity of O&M to make sense of these issues. 

Another issue, which was initially recognized by the groundwater IE, was the unwieldy 

size of the GetCapabilities document, which is the primary discovery mechanism for 

SOS. When a SOS advertises many sites, each with many parameters the size of this 

document grows uncontrollably, which results in difficult and time-expensive client-side 

parsing. To overcome this problem agreement was made that the document would be 

viewed as a quick handshake (to advertise the observation network) and that other queries 

should be used to further discover individual sites within the network 

(GetFeatureOfInterest calls). As well, there was consensus that for the surface water IE 

use cases, Get Capabilities discovery is not particularly useful. To make it so would 

require additional filtering capability on the axes of O&M (eg: show me the capabilities 

in this region). 

The initial feature model to be used by the IE is the INSPIRE Hydrography data model 4 

which after a cursory assessment seemed to be general and useful enough to cater for the 

needs of the IE. Unfortunately recent work by USGS, indicates that it might be too 

general and that some more specific features or attributes are required. One of the areas 

where little progress has been made to date is with agreement on controlled vocabularies. 

This is an item for discussion (as is re-visiting the feature model) at the upcoming 

HDWG workshop in Delft scheduled for April 2011. Another item for discussion at the 

3 

SOS Usage Profile for the Hydrology Domain, September 2010, Stefan Fuest, Kisters, 

http://external.opengis.org/twiki_public/HydrologyDWG/SOSUsageAndServerTypeDiscussion 

4 

D2.8.I.8 INSPIRE Data Specification on Hydrography – Guidelines, 2009, INSPIRE Thematic 

Working Group Hydrography 



workshop is a proposal to establish a cross-domain interoperability test with Oceans Met 

DWG. The goal of this cross-domain work is to ensure that those common aspects which 

are required to achieve interoperability across domains should be identified and 

standardized. Having a consistent representation of a time-series is one. 

The IE has found that OGC provides a useful standardization framework to achieve 

interoperability. The framework provides both abstract standards and processes. 

Unfortunately there is not a standard process for translating abstract specifications such 

as O&M into domain profiles and Interoperability experiments are a good way of 

achieving this. 

The more generic findings in the Surface Water IE will be documented within the OGC 

in order to provide a starting point for the development of domain models dealing with 

time series data. 

The IE brings interested parties (community) together for time bound programs, which 

have specific goals, to achieve the community agreement on some of these abstract issues 

open to interpretation. The surface water IE, is half way though and already has been a 

useful means of identifying those areas where agreement is required and achieving it. 



WaterML2.0: Enabling water information exchange 

Gavin Walker 1 , Peter Taylor 1 

1 CSIRO ICT Centre, Australia 

� 

Our ability to share and understand water information from various sources is�currently 

limited by the use of incompatible information management and publication technologies. 

Water management bodies, hydropower companies, meteorologists and other interested 

stakeholders all monitor water resources and store their data using a wide array of 

technologies. 

� 

Standards are multi-party agreements that provide a common ground on which people 

can share information by minimising ambiguity and sharing definitions for domain 

concepts. They are most effective when surrounded by an active community who, along 

with taking part in defining such concepts, develop supporting tools that allow for easy 

transmission and interpretation of data sets. 

Currently there are no internationally agreed upon standards for encoding water 

observation data, although there are some national standards in use, such as Water 

ML1.1. The Hydrology Domain Working Group (HDWG) was set up by the Open 

Geospatial Consortium (OGC) and the World Meteorological Organisation (WMO) to 

develop an international standard for the exchange of water information. It has produced 

a report 1 outlining the harmonisation of significant existing standards in water and is now 

developing WaterML2.0 as a new international standard. 

WaterML2.0 is based on the ISO and OGC standard Observations and Measurements 

(O&M). O&M relates an observation of a property of a feature of interest, measured by a 

process and producing a result. For example, measuring the temperature of a lake using a 

thermometer as 22 deg. Celsius. O&M can be broadly applied in many scientific domains 

and as a result its default implementation is very general. WaterML2.0 makes a specific 

version of O&M for the hydrology domain. Being more specific allows hydrologists to 

use concepts natural to their field and facilitates consistent mapping of hydrology data 

into a format for exchange. 

Features in O&M are, consistent with the ISO general feature model, meant for real 

world entities such as rivers, lakes and catchments. It is common, however, not to 

measure the properties of features directly, but instead to sample the features and measure 

the property of the sample. For example the thermometer is not measuring the 

temperature of the entire lake but just the sample around the thermometer. In turn this 

result can be used to approximate the value of the property for the whole feature. O&M 

part 2 provides the sampling feature concept which does just that. In WaterML2.0 the 

sampling feature is specialised further to a Water Monitoring Point. Even complex 

1 http://portal.opengeospatial.org/files/?artifact_id=39090 



observations such as thermistor strings can be reduced to observations at a point with 

each observation having a different procedure giving its offset along the string. 

In general O&M the property being observed at the Water Monitoring Point may be 

complex, i.e. composed of multiple properties. For example, observing wind velocity 

could be broken into speed and direction. To simplify the observations in WaterML2.0 

only single, simple properties are allowed. So one observation will measure wind speed 

and a separate observation will measure direction. They can be represented as a pair in 

the client and the server, but keeping them separate in an exchange format creates a 

simpler and more versatile format. The relationship is not lost in WaterML2.0. They are 

related because they related to the same feature. They can also be related explicitly by 

creating collections of observations. 

Encoding the definition of the observation process can be complex. OGC developed 

SensorML for this purpose. To keep WaterML2.0 lightweight this was not used, 

however some process information can be packaged with the data. For example, often 

hydrological data are derived data, such as daily mean flows. WaterML2.0 allows 

representation of such data, while keeping reference to the original source of 

measurement. It also allows for provision of arbitrary parameters. For example, in the 

case of a complex sensor like a thermistor string it can be used to define the offset and 

count of the thermistor in the string. 

This initial version of WaterML2.0 is built around timeseries based readings from in-situ 

sensors. The WaterML2.0 result contains significant metadata for the correct 

interpretation of the data as a series and the ability to add qualifiers or flags on a per point 

basis. Cumulative, isochronic (constant time step) and irregular time series are catered 

for. The notion of data type is also supported. A data type describes the relationship 

between points in a series. With this information the points can be correctly interpolated. 

For example, a data type of ‘preceding mean’ specifies that the value is the mean of the 

preceding interval. Currently only numeric values are allowed. 

WaterML2.0 has been endorsed by the HDWG. A time for public comment will 

commence soon and your input is welcome. Participation is also encouraged in 

Interoperability Experiments (IE). To date WaterML2.0 has been tested in sharing 

ground water and surface water data. An IE in hydrological forecasting using 

WaterML2.0 data will commence soon. The formal process of WaterML2.0 becoming an 

OGC standard is already underway. The HDWG is aiming for this process to complete in 

late 2011. 

Gavin Walker and Peter Taylor are part of the water information research and 

development alliance between CSIRO’s Water for a Healthy Country Flagship and the 

Australian Bureau of Meteorology. 



Enabling Near Real-Time Water Resource Management via 

The Sensor Web 

Andrew Terhorst 1 , Peter Taylor 1 , Brad Lee 1 , Chris Peters 1 , and Christian 

Malewski 2 

1 Tasmanian ICT Centre, Commonwealth Scientific and Industrial Research Organisation, Hobart, 

Australia 

2 Institute for Geoinformatics, University of Muenster, Muenster, Germany 

The Sensor Web is an emerging technology that promises to revolutionise the way water 

information is collected and disseminated. A Sensor Web is a group of interoperable web 

services, which all comply with a specific set of information models and interface 

specifications. The Open Geospatial Consortium (OGC) is developing information 

models and service interface specifications for Sensor Web Enablement (SWE). These 

standards and specifications provide the basic building blocks for Sensor Webs. SWE is 

an approach to break down information silos, improve availability of data from sensors 

and promote the development of new applications that are driven by real-time real world 

information. 

The Commonwealth Scientific and Industrial Research Organisation (CSIRO) is profiling 

the SWE standards to allow correct encoding and transmission of water information. A 

Hydrological Sensor Web has been established in the South Esk River catchment in 

north-eastern Tasmania to test the SWE standards. The Hydrological Sensor Web has 

been configured to provide real-time situation awareness of water flow across the South 

Esk river catchment. 

The South Esk River catchment receives variable rainfall ranging from 500mm in the flat 

lands in the southwest, to 1800mm in the more mountainous northeast. Two weather 

systems bring rain to the catchment: Cold fronts moving in from the west (mostly in 

winter) and sub-tropical depressions that form over the Tasman Sea to the east of 

Tasmania (mostly in summer). River flows can be erratic, especially during dry summer 

months where usually low flow conditions are interspersed with occasional and 

sometimes extreme flood events associated with the aforementioned sub-tropical 

depressions. The Tasmanian Department of Primary Industries, Parks, Water and 

Environment (DPIPWE) want to manage water restrictions in a more proactive and 

selective manner. This requires real-time situation awareness of water flows across the 

South Esk River catchment. The Hydrological Sensor Web must satisfy two DPIPWE use 

cases: Apply Water Restrictions and Announce Flood Take (Temporarily Lift Water 

Restrictions). There are two drivers for the Apply Water Restrictions use case: 

Maintaining environmental flows and sustaining a hydro-power facility at the bottom of 

the catchment. The Announce Flood Take use case aims to increase opportunities to 

harvest unlimited amounts of water during flood events when water is spilling over the 

dam wall at the hydro-power facility. 



The use of SWE standards creates an interoperability layer that allows heterogeneous 

sensor assets owned and operated by the five agencies to be integrated into a seamless 

virtual sensor network, offering significant benefits in terms of redundancy, finer 

granularity, uniform access and a common interface. Data from this virtual sensor 

network are used to generate gridded rainfall surfaces that feed the continuous flow 

forecast model. The system generates automated alerts for when forecasted river flows 

drop below or exceed user-specified thresholds. Users of the system includes water 

resource managers (authorities applying or relaxing water restrictions), agencies 

responsible for flood mitigation/management, and water consumers (e.g. irrigators, 

hydro-power storage operators). The finer granularity achieved through the aggregation 

of sensor assets potentially allows water restrictions to be managed on a sub-catchment 

scale. 

CSIRO is also looking at how to incorporate provenance management into the 

Hydrological Sensor Web. The World Wide Web Consortium (W3C) defines provenance 

in these terms: “a resource is a record that describes entities and processes involved in 

producing and delivering or otherwise influencing that resource. Provenance provides a 

critical foundation for assessing authenticity, enabling trust, and allowing 

reproducibility”. Incorporating provenance into the Hydrological Sensor Web should help 

build confidence in the continuous flow forecast system by making the flow forecast 

process more transparent for the user and enabling retrieval of input observations and 

sensor descriptions. Tracking provenance will also assist fine-tuning of continuous flow 

forecast model parameters. 

CSIRO only uses a subset of the SWE standards and specifications in its Hydrological 

Sensor Web. CSIRO is re-using a software implementation for the Sensor Observation 

Service (SOS) developed by 52North. The SOS publishes sensor descriptions encoded 

using the Sensor Modelling Language (SensorML) and sensor observations encoded in 

XML conforming to the Observations & Measurements (O&M) standard version 1.0. In 

a parallel development, CSIRO has developed a profile for O&M for the water domain 

called WaterML2.0. This profile, developed under the auspices of the joint WMO/OGC 

Hydrology Domain Working Group, is based on O&M version 2.0, which is currently not 

supported by the 52North SOS. This prompted CSIRO to implement a RESTful sensor 

data service that does support WaterML2.0. The intention is to deploy this in the 

Hydrological Sensor Web in the coming months. 

CSIRO is keen to recast the current Hydrological Sensor Web as a Linked Open Data 

Sensor Network based on a RESTful architecture. The OGC is looking at incorporating 

RESTful into its architecture. The Hydrological Sensor Web will provide a test bed for 

various RESTful interfaces with the aim of developing next- generation services. 

In conclusion, a Hydrological Sensor Web facilitates the sharing of sensor information 

without the need for huge infrastructure projects to replace legacy systems. Observations 

from a range of sensor devices, such as tipping-bucket rain gauges, river water-level 

sensors, automatic weather stations and soil moisture sensors, can be shared among 

organisations with remarkably different water information systems. This provides 



opportunities for data re-use, redundancy, and improved reliability of simulation models 

connected to the Hydrological Sensor Web. Incorporating provenance management into 

the Hydrological Sensor Web will increase user confidence/trust in the system. 

This research is supported by CSIRO’s Water for a Healthy Country Flagship and the 

Australian Government through the-- Intelligent Island Program. The Intelligent Island 

Program is administered by the Tasmanian Department of Economic Development, 

Tourism and Arts. 



A Prototype Provenance Management System 

for a Continuous Flow Forecasting System 

Corné Kloppers 1 , Chris Peters 1 , Qing Liu 1 , Quan Bai 1 , Peter 

Taylor 1 , Christian Malewski 1 , Heiko Mueller 1 , Andrew 

Terhorst 1 , Brad Lee 1 , Stephan Zednik 2 , Patrick West 2 , and 

Peter Fox 2 

1 Tasmanian ICT Center, Australian Commonwealth Scientific 

and Research Organization (CSIRO), Hobart, Australia 

2 Rensselaer Polytechnic Institute (RPI) , Troy, NY USA 

1 Introduction 

Limited freshwater resources in many parts of Australia have led to a highly regulated 

system of water allocation. Poor situation awareness can result in over-extraction of water 

from river systems, compromising river ecosystems. To increase situation awareness, the 

Tasmanian ICT Centre is developing a continuous flow forecasting system based on the 

Open Geospatial Consortium (OGC) Sensor Web Enablement (SWE) standards. SWE 

standards provide an interoperability layer over existing observation systems, numerical 

models and processing systems. A prototype Hydrological Sensor Web has been 

established in the South Esk river catchment in northeastern Tasmania. 

The Sensor Web aggregates sensor assets owned and operated by multiple agencies. 

Observations from these assets drive a rainfall-runoff_ model that predicts river flow at 

key monitoring points in the catchment. The intention is to use this information to 

manage water restrictions on a more proactive basis. 

The generation of hydrological information such as predicted river flows, involves 

complex interactions between instruments, simulation models, computational facilities 

and data providers. Correct interpretation of information produced at various stages of the 

information life-cycle requires detailed knowledge of data creation and transformation 

processes. Such provenance information allows hydrologists and decision-makers to 

make sound judgments about the trustworthiness of hydrological information. Managing 

provenance information is an inherent part of the prototype South Esk Hydrological 

SensorWeb. Two major challenges had to be overcome to build the provenance 

management system: modelling provenance and tracking provenance information. 

2 Provenance Information Model 

Provenance information must be expressed using terminology familiar to water resource 

managers and hydrologists. The provenance model for a continuous flow forecast system 

must cover: (1) knowledge of the water domain, (2) information processing and (3) data 

lineage. Our provenance information model integrates four interlinguas to satisfy these 

requirements: A hydrological profile of the ISO 19156 encoding standard for 

observations and measurements known as WaterML 2.0 [1], the W3C Semantic Sensor 

Network Sensor Ontology [2], the Proof Markup Language (PML) [3], and the Process 

Ontology [4]. By integrating domain knowledge with provenance and process 

information, the resulting provenance information model enables water domain 



researchers and water resource managers to analyse and understand how observations and 

derived data products were generated. 

3 Provenance Tracking 

Tracking provenance is a challenge because individual steps in the continuous flow 

forecast system do not handle provenance information explicitly. 

Individual rainfall observations from different sensor sites are retrieved using the OGC 

Sensor Observation Service (SOS). These observations feed into a Kepler scientific 

workflow that generates a rainfall surface consumed by the rainfall runoff model. The 

model outputs are published via another SOS. To track provenance information in such a 

distributed heterogeneous system of software components, we built a Provenance 

Management System (PMS) for warehousing provenance information. 

Figure 1: Overview of the continuous flow forecast workflow and the provenance management system 

The architecture of the PMS is divided into four major components: Harvesting, 

Capturing, Storing and Querying. The overall system architecture is shown in Figure 1. 

Harvesting. The PMS relies on two sources for provenance information: 

(1) workflow execution log files and (2) sensor information exposed via a SOS interface. 

Two types of provenance harvesters were implemented: a log file harvester and a SOS 

harvester. The harvesters parse the data from the respective sources and extract intrinsic 



provenance artefacts. Each harvester generates data exchange documents in JSON 

(JavaScript Object Notation) format. The data exchange documents are then posted to the 

RESTful web service. 

The benefits of the chosen harvesting approach are threefold: the existing flow forecast 

workflow did not have to be modified, the harvesting components have zero runtime 

impact on the actual workflow execution and the harvesting process can be executed 

independently. 

Capturing. Provenance information retrieved by the harvesters is encoded as JSON 

documents and pushed to a RESTful web service, where the provenance information is 

translated into PML (RDF graph) and saved to a persistent database. A harvester specific 

adaptor transforms the JSON document to PML (RDF triples) based on the Provenance 

Information Model. The resulting RDF graph is called a provenance trace. Each 

provenance trace is stored in a RDF database via a Jena interface as a separate named 

graph. Named RDF graphs allow grouping of related RDF triples and are therefore a very 

convenient separation for each provenance trace. 

Storage. Provenance information is stored in AllegroGraph, a triplestore database, in the 

form of an RDF graph. A triplestore is a purpose-built database for the storage and 

retrieval of RDF data. AllegroGraph includes a rich set of useful enterprise features, e.g., 

backup and point-in-time recovery, scales to billions of triples while maintaining superior 

performance and it supports SPARQL queries and RDFS++ reasoning. Querying. 

Querying provenance information stored in an RDF database is done by developing 

predefined Simple Protocol and RDF Query Language (SPARQL) queries. It provides a 

standard query language and data access protocol for use with provenance information 

encoded in the Resource Description Framework (RDF) format. 

4 Conclusion 

In this paper, we investigate how to model and track provenance information for a 

continuous flow forecasting system. The provenance model and its tracking mechanisms 

are presented. Harvesting, Capturing, Storage and Querying can be seen as distributed 

and loosely-coupled components. Our architecture provides a scalable, adaptable, and 

domain agnostic approach for provenance enablement. Whilst our description of PMS is 

tailored towards a specific flow forecast system most of the approaches are generic 

enough to be adopted and applied to other workflows or domains. 

References 

[1] Taylor, P., Walker, G., Valentine, D., Cox, S.: WaterML2.0: Harmonising standards for water 

observation data. Geophysical Research Abstracts 12 (2010) 

[2] W3C Semantic Sensor Network Incubator Group. http://www.w3.org/2005/Incubator/ssn/charter 

[3] McGuinness, D.L., Ding, L., da silva, P.P., Chang, C.: Pml 2: A modular explanation interlingua. In: 

Proceedings of the AAAI 2007 Workshop on Explanation-Aware Computing. (2007) 

[4] Martin, D., Burstein, M., Mcdermott, D., Mcilraith, S., Paolucci, M., Sycara, K., Mcguinness, D.L., 

Sirin, E., Srinivasan, N.: Bringing semantics to web services with OWL-S. World Wide Web 10 

(September 2007) 243-277 



Managing Trust in Provenance-Aware Water 

Information Systems 

Quan Bai, Qing Liu, Corné Kloppers, Xing Su, 

Heiko Mueller and Andrew Terhorst 

Tasmanian ICT Center, 

Australian Commonwealth Scientific and 

Research Organization (CSIRO), Hobart, Australia 

Email: Quan.Bai@csiro.au 

Introduction 

Open water information systems such as the South Esk Hydrological Sensor Web 1 allow 

easy access to, and sharing of, hydrological information. Because important decisions 

may be based on the shared information, it is necessary to provide a trust management 

mechanism that allows users to evaluate the trustworthiness of such information. Trust 

can be defined as a belief a party has that the other party will fulfill its commitments 

[1][2]. 

Trust and Trust Management 

Trust is a well-studied topic in computer science, spanning areas as diverse as security, 

multi agent systems and distributed systems. For a water information system, we define 

trust as an assessment of the likelihood that a piece of information is accurate and 

reliable. There are a number of trust management mechanisms, which can be classified 

into two categories [3]: 

1. Reputation-Based Trust Mechanisms normally use explicit and topic- specific trust 

ratings to represent the reliability or quality of a service (e.g., the rating system in the online 

auction site eBay). Service consumers are required to generate and maintained trust 

ratings based on their experience. Obviously, subjective trust evaluations from different 

users are introduced in Reputation-Based Trust Mechanisms, especially since there is no 

standard and shared evaluation criterion among users. 

2. Context-Based Trust Mechanisms use meta-information to describe the circumstances 

in which information has been claimed (e.g., who published what, when and why), then 

evaluate information trustworthiness based on the roles of information providers. For 

example people may be inclined to trust the flood warnings published by the Australian 

Bureau of Meteorology rather than warnings from other authorities/organizations. 

Context-Based Trust Mechanisms encode trust policies from society into trust evaluation 

rules. This reduces subjective biases from individual users but at the cost of flexibility. 

Provenance-Based Trust Evaluation 

1 1http://wron.net.au/au.csiro.OgcThinClient/OgcThinClient.html 



Provenance describes the origins and processes that relate with the generation of a piece 

of data. Provenance is considered as one of the most promising methods to help users 

assess the trustworthiness of shared data. The Tasmanian ICT Centre is currently 

incorporating a Provenance Management System (PMS) into the South Esk Hydrological 

Sensor Web. Not only does the PMS provide detailed descriptions about generation of 

information products, it also provides domain-specific information about sensors and 

model parameters. Provenance enables hydrologists and decision-makers to make 

judgements about the trustworthiness of information based on a wide range of criteria. 

Provenance information can be an order of magnitude more than the information it 

describes. There is a need to develop a trust model which can automatically extract trust 

information from huge amounts provenance information. 

We have investigated popular trust management mechanisms and propose a provenancebased 

trust evaluation model that allows users to define key trust criteria. The proposed 

model inherits the advantages in both Reputation- and Context-Based Trust Mechanisms. 

The richer context provided by provenance information allows users to make more 

accurate assessments about data trustworthiness. Furthermore, users can include the 

explanation about their trust judgements as provenance information. Trust information 

(such as rating scores) can be associated with meta-annotations which include data 

consumption curriculum, data usage and reasoning behind trust judgements. This helps 

reduce risk of using information unfit for purpose and remove bias in trust assessments. 

From the above two aspects, it can be seen that we can build a more reasonable and 

objective trust management mechanism based on provenance information. For water 

information systems, such a trust management mechanism will help users to estimate the 

trustworthiness of water information without introducing too much overhead. 

References 

[1] S. D. Ramchurn, D. Huynh, and N. R. Jennings, \Trust in multi-agent systems," The Knowledge 

Engineering Review, vol. 19, pp. 1{25, 2004. 

[2] D. Artz and Y. Gil, \A survey of trust in computer science and the semantic web," Journal of Web 

Semantics, vol. 5, pp. 58{71, 2007. 

[3] C. Bizer and R. Oldakowski, \Using context- and content-based trust policies on the semantic web," in 

Proceedings of the 13th World Wide Web Conference, New York, NY,USA, 2004, pp. 228{229. 



HYDROSYS: On-site monitoring and management of water 

environment 

Ari Jolma 4 , Antti Nurminen 4 , Ioan Ferencik 4 , Ville Lehtinen 4 , Murad Kamalov 4 , 

Ernst Kruijff 1 , Erick Mendez 1 , Eduardo Veas 1 , Marc Parlange 2 , Silvia Simoni 2 , 

Susana Fernandez Vidal 2 , Thanasis Papaioannou 2 , Mathias Bavay 3 , Nick 

Dawes 3 , Tom Drummond 5 , Ed Rosten 5 , Olaf Kaehler 5 , Joose Mykkänen 6 , 

Jaouhar Jemai 7 

1 Technische Universität Graz, Austria 

2 Ecole Polytechnique fédérale de Lausanne, Switzerland 

3 Swiss Federal Institute for Forest, Snow and Landscape Research, Switzerland 

4 Aalto University, Finland 

5 University of Cambridge, UK 

6 Luode Consulting Oy, Finland 

7 Ubisense Ltd, UK 

HYDROSYS is a research project that aims to provide a design and a prototype of a 

software system infrastructure including mobile applications to support teams of users in 

on-site environmental monitoring and management. The users are supported analyzing 

processes that may cause environmental degradation, and discussing potential solutions 

for problems found. The project introduces the innovative concept of event-driven 

campaigns with mobile devices, potentially supported by an unmanned aerial vehicle 

(blimp). In such campaigns data from numerous sensors, cameras mounted on the blimp 

and other remote locations, and external sources is gathered, generating dense 

information on a small area. The software system retrieves, checks and stores sensor data, 

and processes simulations based on physical process models. Users can analyze the 

environment using cell phones and handheld computers in addition to their own senses, 

by requesting data stored in the system and by data visualization and analysis. The system 

is developed along multiple actual deployments in Switzerland and Finland, dealing 

among others with issues such as pollution caused by storm water, avalanches, and 

permafrost melting. The project is expected to improve the environmental monitoring and 

management capabilities of environmental scientists, institutions, service providers, 

engineering companies and municipalities through its strong integration of mobile 

devices and sensor networks. It is expected that the tools offered by the project will ease 

inter-disciplinary communication and solution finding for the problems being observed. 

To achieve the goal, a consortium of European universities, institutions and companies 

have been working together with a large group of end-users, binding experience from 

various areas to create a uniform platform. To aid the process, the researchers have 

sought end-user involvement throughout the project. The consortium took considerable 

effort in performing a formal user-centered design and implementation process. 

Numerous discussions were carried out with a group of more than 60 end-users. The 

interviews resulted in an overview of what on-site environmental monitoring and 

management comprises, how it fits in current work processes, and what the specific needs 

of projected end-users are. Of specific interest was the shift from the traditional work 

process, centered in indoors activities, towards the extension of moving some of the tasks 



to the field, without losing the access to data sources needed to analyze a situation and 

make decisions. The identified tasks span from specific monitoring and management 

tasks up to support tasks such as decision-making, communication, and sensor setup and 

maintenance. The system platform design was affected by the observed end-user needs. 

Several sensor deployments were and are carried out in a number of sites in Switzerland 

and Finland to obtain real data and test components of the system. Each deployment 

consisted of installing numerous sensors and carrying out observations. Multivariate 

sensor data was fed to a sensor network software system (GSN) that handles the storage 

and distribution of sensor data. Users can access the near realtime sensor data, and 

associated legacy data via data services that handle filtering and pre-processing of data 

for display on the mobile devices. Particular concern has been the capability to integrate 

simulation models with the rest of the system and the real-time or near real-time sensor 

data. The location of the mobile devices that are taken in the field is “tracked” and users 

can analyze data in the context the data originates. Users can go in the field with their 

mobile devices and select data sources available at the observed site from the software 

system, and interact with visualizations of the sensor data. The data can be visualized as 

numerical representation, graph or simulation overlay on top of a 3D model or video 

stream, as if positioned at the “right position” in the real-environment. Users can also add 

annotations to document their actions in the field. Additionally, the blimp can be 

deployed, capturing high resolution image and thermal data. 

The data being observed is related to the selected sites in Switzerland and Finland and 

always originate from numerous sensors. Together with the end users involved in the 

project, the consortium has selected sites that differ in focus and pose different challenges 

for the technology. In Switzerland three sites were selected. Each site focuses on a 

specific topic: hydrology-related disasters prediction and management at La Fouly (close 

to Mont Blanc), permafrost degradation at Gemsstock near Andermatt, and triggering and 

formation of wet snow avalanches in Dorfberg (Davos). Accordingly, different sets of 

sensors / sensor stations are deployed, providing a wide range of data. In Finland, two 

sites were selected. The Kylmäoja site is a medium-size peri-urban catchment and stream. 

The focus on Kylmäoja is the impact of pollution and changes in hydrological processes 

due to construction and other activities on the catchment. The Nummela site is an actively 

restored suburban stream and wetland system. The sites provide the consortium with 

excellent possibilities to test the system infrastructure with end-users in the field, but also 

offer considerable challenges to the deployed system, from network limitations and 

power issues, up to benign weather conditions. 

The project has progressed beyond the current state of art, by dealing with short term 

events and detailed analysis of small sites using mobile devices. The analysis of such 

events is hardly supported by current methods, but may have a large impact on 

environmental degradation. With the current pressures on water related systems in mind, 

the consortium hopes to provide an asset to improve the situation. 



The use of OpenMI in product development and project work 

Pavel Tacheci 1 , Zsuzy Nagy 1 , Stanislav Vaněček 1 , Jelle Hurkens 2 , Thomas 

Clausen 3 , Riccardo Cora 3 , Jesper Gross 3 , Gunvor Philipsen 3 , Johan Hartnack 3 . 

1 DHI a.s, Prague, Czech Republic 

2 RIKS BV, Maastricht, The Netherlands 

3 DHI, Hørsholm, Denmark 

Numerical hydrodynamic models are today seen as part of an integrated whole 

encompassing water quantity, water quality, sediment transport phenomena, rainfall 

runoff etc. The “radius of integration” has grown further by inclusion of such aspects as 

meteorology, economics and legal issues. Thus the scope of integrated hydrodynamic 

modeling has increased from a typical pure water related domain to also cover aspects 

which typically lie outside the water modeling frame. The distinction between coupling 

typical water modeling aspects and coupling to other phenomena related aspects lay not 

so much in the models coupled but more in the origin of the models. That is the models 

being linked typically have different authors and the developers involved do not have 

insight into the internal workings of each other models. This aspect is acknowledged by 

the DHI Group. Their key experts have taken the approach of making the MIKE models 

open-ended through the OpenMI standard. This standard approach of linking is used for 

external links between 3 rd party models and MIKE models but has also proven useful for 

links between multiple models originating within the DHI Group. This paper will focus 

on the use of OpenMI as a linking mechanism with emphasis on the practical use of the 

OpenMI standard as a tool for linking to both internal and external models in order to 

create a very flexible modeling environment, which could cover the various defined 

processes in the problem analyses. 

OpenMI: The OpenMI standard defines an interface that allows models to exchange data 

at runtime. First version of the openMI standard was developed during the FP5’s 

HarmonIT. Currently this standard is maintained by the OpenMI Association. The 

standard defines set of software interfaces. The key interface in the OpenMI standard is 

ILinkable Component. Any OpenMI compliant component must implement this interface. 

Existing models are typically wrapped to such components. The OpenMI association 

provides a set of the tools (SDK) to simplify the process of wrapping. Through the 

standard data is exchanged dynamically between components during of the model 

execution. Dynamic data is transferred between components by the linking of output 

exchange items and input exchange items. Components can be linked using onedirectional 

or bi-directional links. The standard supports data distributed spatially and 

temporally. Models, running in the different temporal and spatial domains can be linked 

using a mechanism, where the outputs of the models are “adapted” in order to fit the 

requested inputs for the other models. Such adapters can be developed separately as set of 

advanced, reusable tools. The OpenMI standard implementation currently exists in C# 

and Java. Based on experience with the Version 1.* the Version 2 of the standard is not 

focusing only on time-based numerical models and simplify the mixing of temporal and 

non-temporal models (e.g. databases, GIS,…) within a single composition. 



Use case 1: Integrated catchment management and planning – Linking of the MIKE SHE 

and METRONOMICA. 

As an example of integration of two different modeling systems, the authors present the 

work and results of the project focused on a DSS tool which supports the aim of the Tisa 

Catchment Area Development (TICAD) (SEE/A/638/4.2/X) transnational project 

financed from South East Europe Transnational Cooperation Program and is able to 

integrate different development themes on Tisza river basin level and usable also for the 

assessment of one of the main issues on the Tisa basin (environmental-water management 

issues). The developed TICAD Tool for spatial planning is able to integrate the relevant 

drivers of land use change such as human behavior, physical characteristics, zoning 

regulations and accessibility, and simulate local dynamics in a spatial and temporal 

manner. Both allow a realistic representation of land use developments and can show the 

impact of different policy options (zoning and construction of infrastructure) on land use 

developments and in this way support policy makers in developing spatial plans. The 

integrated approach brings to spatial planning the environmental related components 

(which are many cases directly linked to water management) in early phases of the 

planning process as well as the integration between land use and water with mutual 

feedback loops. 

METRONAMICA is a generic forecasting tool for planners to simulate and assess the 

integrated effects of their planning measures in urban and regional development. Linking 

of the MIKE SHE and METRONOMICA provide the possibility to add hydrological 

phenomena simulated by complex hydrological model to the planning, and reflect 

changes in the urban and regional development in hydrological simulation. Integrating 

the hydrological processes modeled by MIKE SHE in a METRONOMICA based landuse 

planning support system enables the planner to address a much wider class of 

problems and thereby improves the support for an integrated approach to spatial planning. 

The MIKE SHE and METRONOMICA are connected using OpenMI. Land-use for a 

given time is an output exchange item from METRONAMICA and an input exchange 

item for MIKE SHE. MIKE SHE output exchange items (such as ground water level) are 

transferred using Adapted Output to the suitability maps for METRONOMICA. The 

given concepts provide possibility to create scenarios as combination of hydrological 

variants (Climate changes) and different urban planning strategies. 

Acknowledgement: 

Reference number: 2009/S 234-3348776 (TED) 

Client: VÁTI Hungarian Nonprofit Ltd. for Regional Development and Town Planning 

H-1016 Budapest, Gellérthegy utca 30-32. 

Supply contract supports the Tisa Catchment Area Development (TICAD) project 

financed by South East Europe Transnational Cooperation Programme 

www.see-ticad.eu 



Use Case 2: Using a combined regional scale hydrological model with a fine resolution 

grid to address local scale issues - Coupling of MIKE SHE and the porous media flow 

model FEFLOW 

To investigate the local scale effects of the water table close to wells a coupled MIKE 

SHE and FEFLOW model has been developed utilizing the OPENMI standard. The 

coupling is designed so that the recharge calculated from MIKE SHE is supplied as a 

boundary condition to FEFLOW. MIKE SHE calculates the recharge over a 2D uniform 

grid and the results are then mapped to the unstructured mesh used in FEFLOW. In return 

FEFLOW supplies the calculated groundwater table. What makes this use case stand out 

is that both FEFLOW and MIKE SHE are DHI models and thus the coupling of these 

could be achieved through other means than OPENMI. Though OpenMI was chosen 

since it ensures that the coupling interfaces for both MIKE SHE and FEFLOW are 

available for other models to be coupled. 

Use Case 3: SMART-OPS Project, Metropolitan Water District of California, Los 

Angeles, USA 

The project SMART-OPS (Simulation and Modeling Assisted Real Time Operations) 

developed for Metropolitan Water District of California (MWD) uses a hydraulic model 

as one of the core components. 

The water distribution system is operated by many valves, gates, weirs and pumps where 

many of these are controlled by Advanced Process Control logic (APC). For the APC 

logic it was decided to use the actual code from the field and implement it in separate 

software (‘APC simulator’) in order to have it run together with the hydraulic model to 

exactly replicate the real system behavior. 

Using the actual code directly ensured that the behavior in the model would mimic the 

real situation closely and any changes in the field operation could easily be transferred to 

the hydraulic model. Further any changes of existing APC’s or development of new 

APC’s could be tested with the model before deployment in the field. 

The APC Simulator is integrated with the model through OpenMI and uses necessary 

operating parameters such as flow or pressure set-points from SCADA. 



Figure: The real APC interacts with SCADA to produce the appropriate control action (left figure). The 

APC functionality is reproduced in the model by the 'APC simulator' and interaction is made with model 

results to produce the appropriate control action (right figure) 

References 

Z. Cello Vitasovic et al. (2009) Integrating hydraulic simulation and on-line process control, Water Practice 

and Technology Volume 4, issue 3 

The authors are certain that in the near future, there will be an increasing need for linked 

models originating from different sectors. Using the linked approach helps planners to 

better understand the integrated nature of the system they are operating and point out the 

side effects of planning options, the trade-offs that have to be made and the win-win 

situations that can be created. OpenMI is definitely a vehicle for these model integrations, 

particularly when the models involved are not developed by the same vendor. 



Abstracts – session 2- enviroGRIDS project and related research 





HAWQS: multi-spatial-temporal scale hydrologic and water 

quality decision support system for continental US 

Debjani Deb 1 , Jeyakanthan Veluppillai 1 , Raghavan Srinivasan 1 , and Jeff Arnold 2 

1 Spatial Sciences Laboratory, Texas A&M University, 1500 Research Parkway, College Station, 

TX 77843, USA 

2 Grassland Soil and Water Research Laboratory, USDA-ARS, 808 East Blackland Road, 

Temple, TX 76502, USA 

The Hydrologic and Water Quality System (HAWQS) is an advanced, state-of-the-art 

total water quantity and water quality modeling system with databases, interfaces and 

models that is being developed for USEPA’s Office of Water to evaluate the impacts of 

management alternatives, pollution control scenarios, and climate change scenarios on 

the quantity and quality of water at a national scale. HAWQS will be capable of 

supporting a wide variety of national- and regional-scale economic and policy analyses 

by simulating baseline and alternative water quality conditions with respect to the 

following seven broad categories of water quality constituents - sediments, pathogens, 

nutrients, biological oxygen demand (BOD), dissolved oxygen, pesticides at three 

different spatial (HUC8, HUC10, NHDplus) as shown in the table below and three 

different temporal scales (day/month/year) 

8-Digit HUC 10 -Digit HUC NHD Plus 

No. of watersheds 2110 15479 2595196 

Hydrologic Response 

Units (HRU) 

1466915 12168566 57074834 

HAWQS is an extension of the Hydrologic Unit Model of the United 

States (HUMUS), which was recently updated to support USDA’s Conservation Effects 

Assessment Project (CEAP). Specifically, HAWQS updates input data, upgrades SWAT 

(Soil and Water Assessment Tool) modeling capabilities, replaces the existing stream 

network with the National Hydrography Dataset (NHD Plus), and creates interfaces and 

data management utilities beyond those included in the SWAT-HUMUS system. 

As an information management tool, HAWQS is a continental-scale system capable of 

handling large data files and intensive computations. It is designed as a distributed 

modeling system that allows multiple users at different locations to simultaneously access 

the system, perform simulations, and store results. It has a multi-tier (or n-tier) 

architecture (as shown in the figure below) that includes the HAWQS server database and 

several middle tier servers that perform various operations, including the transfer of data 

between servers, the creation of SWAT input files, SWAT model runs, and client-side 

output analysis. 



Schematic diagram for the HAWQS architecture 

The HAWQS database is a relational database designed and developed with Oracle 10g 

and contains data from RAD, NHDplus, SWAT 2009 database and other attribute tables 

to support SWAT. Some of the data includes watershed description, landuse, soil, slope, 

weather stations and weather parameters, crops, reservoirs, livestock, fertilizer, 

pesticides, point source, atmospheric deposition, management options etc. It provides the 

input data required to simulate SWAT at a wide range of scales. The HAWQS database 

also stores model outputs in cluster servers along with SWAT Editor and SWAT2009.exe 

and allow users to extract both raw data and processed data from the HAWQS data 

tables. The HAWQS database also utilizes many database services and functions to 

generate stream networks for a given starting and ending stream nodes, perform subwatershed 

accumulation, and conduct watershed routing. This backend database is hosted 

on a Linux platform. The HAWQS application server (Middle tier) contains web 

application server (Apache Tomcat with Axis2 plug-in) for running server web services 

and initiates database service to generate stream network then exchange data between the 

database server and the web server to create input files for SWAT model run. It also 

creates ASCII files and distributes them to a cloud server for SWAT model run and runs 

the SWAT model and generate the outputs. Therefore, analysis can be done at 

anyplace/anywhere for any period. The HAWQS client side allows users to select 

modeling options, specify model parameters, and enter physico-chemical characteristics 

of pollutants for SWAT Model simulation. 

HAWQS will evaluate the impacts of management alternatives (changing landuse, 

fertilization rates etc.), pollution control scenarios, and climate change scenarios on the 

quantity and quality of water in large and complex river systems and also the the 

environmental benefits of conservation practices at the national scale. Apart from 

estimating the benefits of conservation practices, HAWQS will also help policy makers 

and program managers to improve the effectiveness of existing conservation programs 



and design new conservation programs. In addition, HAWQS will also be capable of 

supporting a wide variety of national- and regional-scale economic and policy analyses 

by simulating baseline and alternative water quality conditions using 42 years of weather 

data (from 1960 - 2001) that will be useful for policy makers or economists conducting 

benefit assessments of water programs to enforce existing programs, regulations and 

policies at both regional and national scales and recommend new ones by providing 

defensible outputs that can be utilized to estimate human health risks, drinking water 

treatment costs, and criteria exceedence frequencies etc. Users can run all these scenarios 

in conjunction with SWAT editor very quickly. 

HAWQS is a web based application with minimal maintenance on the client side and will 

be installed in a private cloud computing environment at the Spatial Sciences Laboratory 

(SSL), Texas A&M University. It can be run on server side service from start and end 

watershed as well as can be downloaded to desktop for further customization with 

SwatEditor and VizSwat software to visualize. 



Automated Analysis of upstream-downstream relationships 

using Bayesian Belief Networks from spatially distributed 

SWAT models 

Yunqing Xuan, Seleshi Yalew, Xuan Zhu, Zheng Xu, Ann van Griensven 

UNESCO-IHE Institute for Water Education, P.O. Box 3015, 2601 DA DELFT, The Netherlands 

Email: y.xuan@unesco-ihe.org 

Spatially distributed models are powerful tools in supporting decision-making processes 

in environmental policies, such as for the Water Framework Directive. These models are 

able to predict the effects of changes in a distributed way and to relate the typical 

upstream-downstream variable relationships that are fundamental in order to reach good 

status in all water bodies. For example, they can propagate the effects of remediation 

technologies for pollution abatement through a river network. However, those models 

often require substantial computing resources, which in many cases cannot be afforded by 

a real decision making tool, and therefore, using of models often has to be implemented 

in an offline fashion which inevitably results in detrimental impacts on the DSS itself. 

The need of light-weight tool that can represent the essential dynamics from complex, 

physically based models like SWAT, has become increasingly clear in developing DSS 

tool in environmental management. For example, in EU FP7 AQUAREHAB project, a 

management tool able to assess the remedial measures for degraded water bodies is 

commissioned, which needs not only the dynamics basis formed by SWAT and other 

heavy modeling tools, but also the capability of quick and interactive assessment that can 

only be provided by light-weighted tool as a substitute. 

In the development of this management tool (code named as REACH-ER), several 

methods have been investigated to find an efficient way to represent the upstreamdownstream 

relationship revealed by SWAT simulations. The Bayesian Belief Network 

(BBN) is found to be able to not only characterize and reproduce the relationship; the 

topology of the network itself can naturally resemble the spatial structure of catchment 

(river reaches and sub-catchments). Further, once the BBN is derived from the SWAT 

model simulation, its structure is fixed and does not require heavy computation in dealing 

with interaction. This makes it possible and much easier to represent the BBN in the 

management tool, for instance, by simply storing the structure into the underlying 

database. Since the font end of the REACH-ER tool, the Vis-REACHER is completely 

based upon WEB GIS and database; it can immediately make use of this convenient 

feature to integrate the BBN into the management tool. 

One of the important steps in implementing the BBN in the management tool of REACH- 

ER, is to build the BBN from the model simulations. It can be done manually but an 

automated procedure is more in favor since the BBN might need to be changed in the 

maintenance procedure where a dedicated tool is expected to conduct the job. 



An automated procedure has been developed to enable the construction of a Bayesian 

Belief Network from a spatially distributed model SWAT relating upstream-downstream 

relationships on different stream variables. The Bayesian Belief Network was included in 

a GIS platform (MapWindow) as a ‘plug-in’. Time series data on files from a SWAT 

model result were analyzed on different water quality variables, such as Organic nitrogen 

and Nitrate, at each sub-basin outlet. These time series data were trained and verified 

using Artificial Neural Network (ANN) techniques to build the conditional probability 

table (CPT) for a Bayesian Network model. The time series data were further processed 

towards yearly statistical descriptors (e.g., mean, variance, yearly maxima). The GISplugin 

automatically constructs a Bayesian Belief Network using Netica (NORSYS, 

2011) by converting input and output formats from SWAT data files and re-formats and 

sends back as input to SWAT models the parameter values modified as cases scenarios 

on the NETICA interface. In such a way, it is possible to learn impacts of different levels 

of pollutants from different sub-basin outlets and visualize the effects of pollution 

reduction in a sub-basin throughout the downstream river network. An example of the 

BBN (as structured by Netica) and the part of river network it represents is given in 

Figure 1. 

Figure 1 an example of BBN (right) and the part of river network it represents (reproduced from Xuan et al, 

2011) 

The BBN that has been built from the SWAT model simulation and represented by the 

Netica software is then re-modeled into the database of the REACH-ER tool. The 

database is deigned in such a way that the topological relation of BBN nodes as well as 

parameters are built into separate tables. This procedure is also automated so that the 

maintenance can be done without making changes to the front end. 



Figure 2. The front end of REACH-ER shows a case study for the Odense basin. 

An application to the Odense river basin is presented in order to evaluate the potential of 

remedial technologies such as restoration of riparian wetlands. The visualization function 

is implemented as the front-end of REACH-ER (Figure 2) that uses database to 

communicate to other tools as described above. 


This work has been supported by EU FP7 AQUREHAB project. 

References: 

NORSYS (2011): Netica Tutorial, http://www.norsys.com/tutorials/netica/nt_toc_A.htm 

Xuan, Y. et al (2011): The Design of REACHER, AQUREHAB project technical note, draft available at 

http://goo.gl/FPDAe 

Xu, Z. et al (2011) Modelling Eco-toxic Mass Transport using Probabilistic Network and SWAT, 

AQUREHAB project technical note. Available upon request. 



Information support for regional water resource management 

in Ukraine 

Yevgen Makarovskiy 

Ukrainian Scientific and Research Institute of Ecological Problems (USRIEP), Kharkiv, Ukraine 

Water resources availability is a determining factor of human existence. Despite the huge 

volume of water in the hydrosphere, water resources include only those waters that may 

be used by humans for drinking and household water supply and those ones that support 

the sustainable existence of living ecosystems. 

The issues of drinking water consumption are particularly important. Only fresh surface 

water and groundwater, complied with quality standards, may be used for drinking water 

supply. 

At the same time, precisely surface waters are under influence of anthropogenic load 

primarily due to their depletion and chemical and biological pollution. In some cases, 

these waters pass into the volume of water unusable for its consumption and threaten the 

survival of ecosystems. 

In Ukraine, similarly to the European Community [1], a policy aimed primarily at 

stabilizing the human impact on surface water, and improving the ecological status of 

surface waters in the country was adopted. Procedures for water resources management 

require an appropriate information support to justify the adoption of management 

decisions and subsequent control of their performance. 

The information required for management depends on the task of management. Current 

tasks of management can be divided into two classes - tasks of the current control and of 

tasks long-term planning. 

Water quality data obtained from monitoring network represent information data for the 

tasks of water quality control but only as ecological indicators. 

However monitoring information is insufficient for the long-term planning. Directly 

component-wise monitoring observations do not provide opportunities for comparative 

environmental assessment of sites over time. The latest concept is necessary for the 

ranking of environmental problems, extraction of priorities (hot spots) are required for 

efficient use of limited resources. 

Therefore, the establishment of information support of water resources management is 

divided into two interrelated tasks - methodological work on the establishment of 

adequate environmental assessments and instrument development of information support 

for environmental assessment. 

In Ukraine the work on integrated environmental assessment is not completed yet. 

Recommended ECE system for environmental assessment [2] to EECCA countries is 

mainly oriented on control of implementation of international agreement, specifically 

about reducing human impact on the environment. 



For several years USRIEP team performed work on information support of water 

resources management. Taking into account mainly regional kind of planning and waterprotective 

activities performed investigations oriented on their usage within the region 

(river basin or its part within the region or country). 

Performed activities were based on the prevailing practice in Ukraine of surface waters 

protection from pollution, best practices developed in EU countries, in particular the 

proposed principles in EU Directive 2000/60/EC from 23.10.2000 (WFD) established for 

the EU's activities in the area of water policy, with additional materials on introduction of 

GIS elements (explanatory document # 9). Rules of monitoring and water quality 

assessment of transboundary rivers were also used. These rules were developed by a 

working group established under the auspices of the European Economic Commission, 

and aimed at implementation of the Convention on the Protection and Use of 

Transboundary Watercourses and International Lakes (Helsinki, 1992), materials from 

other organizations that have successfully worked in this area, in particular Ruhrverband 

(FRG). 

The requirements for the presented information are assigned on the basis of national 

procedures for environmental management. 

Methods of integrated assessments of human impact on water resources was developed, 

such as chemical sewage pollution, water use, waste discharge, fresh water use, etc. The 

proposed assessment methods have a physical meaning. Developed integrated assessment 

of human impact was tested in the information and management projects [3]. 

Information-analytical system "Water resources" was developed as an information 

management tool for water resources management [4-5]. Information-analytical system is 

based on GIS technology. The system works interactively with the user. 

The main feature of a system is its possibility to construct queries. Electronic maps of the 

region, hierarchical drop-down menus and dialog boxes are used for queries construction. 

System user has ample opportunities for designing of complex queries in response to 

which the system gives statistical or analytical information. 

This information system provides a variety of quantitative assessment of water resources 

in the territorial division and division of water bodies according to their hierarchy. 

Assessment of the quality characteristics of surface water – their contamination and 

suitability for the main types of water consumption - is conducted in a system of 

ecological monitoring data in relation to the quality control stations, or their selected 

complex. 

Specialized methods of environmental assessment approved at the state level and used for 

assessing the contamination of surface waters and analysis of the suitability of water 

resources for drinking and recreational water consumption. 

The system also implements the possibility of a more refined analysis of surface water 

pollution, identifying the most important ingredients. 

Particular attention in the system was focused on the reporting of analytical results to user 

as one of the most important requests of use. The results are presented in the cartographic 

(thematic map), tabular and symbolic-numeric form. 



This information-analytical system was used for surface water quality management and 

supply of water resources in Kharkiv and Luhansk regions. It was used during 

development of the programs aimed to improve the ecological condition of rivers 

Seversky Donets, Southern Bug, and environmental audits of Tashlykskaya hydroelectric 

power plant, etc. 

Literature 

1. Water Framework Directive. Directive of the European Parliament and the Council of the European 

Union № 2000/60/ES from 23th October 2000. 

2. European Economic Commission. Guidance on the Application of Environmental Indicators in countries 

of Eastern Europe, Caucasus and Central Asia. / Working Paper No WGSO/EXECOM-5/2007/9. 110p. 

3. Makarovsky Y. Methods for complex assessment in information support of environmental safety 

management in the region. // East European Journal of advanced technologies, 2006. № 3 / 3 (21) .- 94-99p. 

4. Kuzin A., Makarovskiy Y. Development of a Monitoring System for Water Quality Control in Ukraine 

//Integrated Approach to Environmental Data Management Systems / Edited by N. Harmancioglu et al. 

NATO ASI. Series 2. Environment. 1997, Vol.31. R.495-506. 

5. Makarovsky Y. On information support of surface water quality management. / / Problems of 

environmental protection and ecological safety. USRIEP Collected scientific publications. Kharkiv, Pub.h.. 

House "Rider" 2005. С. 141–157. 



Applying EnviroGRIDS Tools to Rioni Basin 

Karin Allenbach 1 and Mamuka Gvilava 2 

1 EnviroGRIDS Project WP2 Remote Sensing Task Manager 

UNEP/DEWA/GRID-Europe / International Environment House 

11 Chemin des Anémones / CH-1219 Châtelaine / SWITZERLAND 

Tel: +41-22-917-86-45 

Fax: +41-22-917-80-29 

E-mail: Karin.Allenbach@unepgrid.ch 

Web: www.EnviroGRIDS.net 

www.grid.unep.ch 

2 EnviroGRIDS Project Task Manager 

National ICZM Focal Point for Georgia 

GeoGraphic, GIS and RS Consulting Center 

10, Bulachauri street, 0160, Tbilisi, GEORGIA 

Tel: +995-99-54-66-16 

E-mail: MGvilava@GeoGraphic.ge or MGvilava@ICZM.ge 

Web: www.GeoGraphic.ge 

Objective of this work is to explore the approaches for utilizing remote sensing and 

hydrological modelling tools (such as ESIP [Gorgan et al., 2009] and SWAT [Arnold et 

al., 1998]) promoted through the enviroGRIDS BSC-OS Portal (see [enviroGRIDS D6.1]) 

and testing related methodologies through application to the Rioni River Basin (defined in 

[enviroGRIDS D2.4]). After short characterization of the river basin, case study 

requirements are defined and potential for grid-enabled remote sensing applications are 

outlined, specified with regard to potential user profiles, data sources, applications, 

functionalities, and use scenarios. 

Rioni Basin. Rioni is the largest river in western Georgia with the length of 

approximately 330 km, catchment area of 13,400 Km 2 and human population exceeding 

500 thousand. Upstream tributaries of Rioni flow from mountainous areas of the Great 

and Lesser Caucasus, while downstream reaches of the river flow across the Kolkheti 

lowland before entering the sea, essentially mimicking almost all types of hydrological 

processes taking place in the Black Sea Catchment. The flow regimes are controlled by 

the series of reservoirs, and more hydroenergy developments are planned, which would 

alter the hydrology of the entire catchment, rendering the need in the development of 

integrated river basin management policies and tools ever more required. 

Users Profile. The following groups could be identified as the potential developers as 

well as the users of the systems under consideration: enviroGRIDS project partners, 

professional networks of ICPDR and the Black Sea Commission, external partner 

agencies with various river basin management responsibilities (including funding 

institutions), development proponents and authorities involved in related impact 

assessment studies and permit letting. User categories (defined following [enviroGRIDS 



D2.2]) could be attributed as data administrators, data providers, earth science specialists, 

decision makers, concerned citizens and public. 

Data Sources. Required datasets, keeping into consideration the hydrological modelling 

needs of particular case study sites such as Rioni Basin, and the overall Black Sea 

Catchment, could be identified both at the global level (through remote sensing) as well as 

from the national in-situ data sources (such as hydrology, climate, and pollution). 

Global remote sensing sources could originate from a wide range of NASA, ESA and 

JAXA operated sensors and data processing projects, available publicly through http and 

ftp services: TRMM, SSM/I and GSMaP (rainfall), AVHRR (cloud cover and vegetation 

indices), MODIS and MERIS (land cover), AMSR-E and SMOS (soil moisture and ocean 

salinity), QuikSCAT (near-surface winds), and Aster (Global DEM). 

In-situ datasets of variable range and quality are available for Rioni case study site as 

well, including historic and current hydrology (daily discharge rates for several gauging 

locations), climate (3-hourly and daily meteorology data for four locations), and pollution 

(mostly ad hoc measurements as part of the impact assessment studies, such as 

[Namakhvani, 2010]). 

Applications. Range of issues which are to be handled to successfully extract datasets in a 

form useful for hydrogeological modelling applications include the following (based on 

experiences presented in [Milewski et al., 2009]): large storage requirements; automation 

of processing; wide variety of raw data formats; data conversion and georeferencing 

needs; spatial and temporal subsetting; and quality assurance and control. 

Functionalities. Specific functionalities could be built keeping into consideration generic 

portal functionalities specified in ([enviroGRIDS D2.2] to: gather, visualize and 

administer data resources and administer users; define scenarios; execute applications; and 

run scenarios. 

Use Scenarios. There could be two complementary ('manual' and 'automated') scenarios 

proposed for testing the feasibility of approaches proposed in this communication: 

(i) Testing the usability and sufficiency of above described data resources and their 

utility through the application of 'conventional' hydrological modelling tools to a 

particular case site of Rioni Basin. 'Manual' data extraction and model building could help 

better define specific knowledge on data formats as well their relevance for the required 

hydrological modelling application. Established global data portals such as NASA's 

Giovanni (see http://disc.sci.gsfc.nasa.gov/giovanni) or JAXA's GSMaP (see 

http://sharaku.eorc.jaxa.jp/GSMaP/index.htm) could be used for the case study site of 

interest to generate required precipitation and climate datasets from global sources and to 

correlate them with in-situ data. 



(ii) Utilising enviroGRIDS tools (or some of their elements as found most 

appropriate) described in [enviroGRIDS D2.7] to build and implement data use 

applications within GRID-enabled environment such as ESIP. 

Acknowledgements 

Valuable discussions held with the members of the joint remote sensing task team, 

facilitated and guided by enviroGRIDS WP2 Leader Dr. Gregory Giuliani and WP6 

Leader Dr. Dorian Gorgan, are kindly appreciated. 

References 

enviroGRIDS D2.2, Data Storage Guidelines, 2009. 

enviroGRIDS D2.4, Remote Sensing Data Use and Integration Guideline, 2009. 

enviroGRIDS D2.7 GRID Services Supporting Massive Data Management, 2010. 

enviroGRIDS D6.1 Requirements and specifications for the development of BSC-OS Portal, 2010. 

Arnold, J. G., Srinivasan, R. Muttiah, R. S. and Williams, J. R. (1998), "Large area hydrologic modeling 

and assessment: Part I. Model development", J. American Water Resour. Assoc., 34(1): 73-89. 

(http://www.brc.tamus.edu/swat) 

Gorgan D., Bacu V., Rodila D., Pop Fl., Petcu D., Experiments on ESIP - Environment oriented Satellite 

Data Processing Platform. SEE-GRID-SCI User Forum, Bogazici University, Istanbul, Turkey, 

December 9-10, ISBN: 978-975-403-510-0, pp. 157-166, (2009). 

Milewski A., Sultana M., Jayaprakasha S. M., Balekaia R. and Becker R., RESDEM, a tool for integrating 

temporal remote sensing data for use in hydrogeologic investigations. Computers & Geosciences, 

Volume 35, Issue 10, October 2009, Pages 2001-2010. 

JV Group (Nurol, KEPCO, SK E&C) and ENCON Environmental Consultancy Co., Namakhvani HPP 

Cascade Project ESIA, Undisclosed Draft, October, 2010. 



Integrated data collection and dissemination in North 

Bulgarian river basins 

Dr.eng. Snezhanka Balabanova (email: snezana.balabanova@meteo.bg), 

eng. Silviya Stoyanova (email: silviya.stoyanova@meteo.bg) 

National Institute of Meteorology and Hydrology (NIMH) 

, 66, Tsarigradsko shosee Sofia, Bulgaria 

Abstract 

Integrated river basin management is very important for evaluating the river basin 

hydrological resources. This paper introduces an approach for application of SWAT (Soil 

and Water Assessment Tool) for effective river basin management. It is a flexible model 

that can be used under a wide range of different environmental conditions. Geographic 

Information System (GIS) is successfully integrated with a conceptual hydrologic model. 

This is an effective way to collect, manipulate, visualize, and analyze the input and output 

data for modeling watersheds. SWAT is used to evaluate temporal and spatial variability 

of water cycle parameters. The model is used for analyses and water balance of ungauged 

watersheds, studying climate change impacts. The model is suitable for forecasting 

natural hazards as floods and droughts based on 1-3 days forecast from meteorological 

models for example a forecast from numerical weather prediction model ALADIN. The 

model results and analyses based on these results can meet stakeholder's requirements for 

water resources management, climate change and disaster management. 

Keywords: hydrologic modeling, water resources processes, GIS 

NIMH 

The National Institute of Meteorology and Hydrology (NIMH) (www.meteo.bg) at the 

Bulgarian Academy of Sciences is the official hydrometeorological service in Bulgaria. 

NIMH was established in February 1890. Its primary mission is to provide 

meteorological and hydrological information and products to different organizations and 

users in Bulgaria. Its duties comprise both operational and applied research activities. 

Hydrological and meteorological observations, data acquisition and telecommunication, 

monitoring the air, surface and ground water, meteorological (www.weather.bg) and 

hydrological (hydro.meteo.bg) forecasts, assistance to special sectors through applied 

maritime meteorology and agrometeorology, maintenance of data base, scientific 

researches, numerical and statistical modelling are the general duties of NIMH. 

The structure of the NIMH comprises the Central Office in Sofia and the 4 Regional 

branches, which drive observatories and observation stations. The main observation 

networks are: the meteorological and agrometeorological, the hydrological, 

hydrogeological and the air and water chemical background monitoring. 



The research activities are concentrated mainly in the Central Office in Sofia. Their 

permanent staff consist of 68 highly qualified researchers (38 Ph.D., 5 D.Sc.). NIMH has 

been and actually is a partner in many international research projects and programmes 

like: IHP-UNESCO, various EC PHARE research projects including the transboundary 

co-operation, UNDP research projects, WMO and World Bank funded projects, EC 

INCO and FP4/5/6/7 projects, as well as research and applied research projects in the 

frame of bilateral technical assistance agreements of Bulgaria with France, Spain, UK, 

Germany and others. 

Description of the study area 

The studied area comprises two watersheds - the watershed of the Vit River and the 

watershed of the Osam River. Both watersheds belong to the major Danube river basin 

(Fig 1). 

Fig. 1 Study area 

The region falls in the moderate continental climate of the European sub-continental 

climatic region and is characterized by hot summer and cold winter. Annual distribution 

of precipitation in this area has predominantly continental type with well expressed 

summer maximum and a minimum in the cold halfyear. 

The watershed of the Vit River is an elongated basin having an area of 3228 km2. The 

river begins after the confluence of the Beli Vit and Tcherni Vit Rivers springing from 

the Balkan Mountain. In the most upper part of the river the slope reaches up to 200 ‰. 

The mainstream goes on directly to the northward through large valley with lower slope 

to the Danube river. The average slope of the river is 9.6 ‰. The density of the river 



network is very small – 0.5 km/km 2 . The average altitude of the watershed is about 400m. 

There are five monitoring hydrometric stations and nine monitoring meteorological 

stations in the Vit watershed. 

The area of the Osam River is 2838 km2. The source of the river is in the central part of 

the Balkan Mountain. The mean slope is 5.7‰. The density of the river network is very 

small – 0.4 km/km 2 . The average altitude of the watershed is about 355 m. There are five 

monitoring hydrometric stations and fourteen monitoring meteorological stations in the 

Osam watershed. 

The high flow period of the rivers under the Continental influence is during April – June. 

The main reason for the annual maximums are the rains being often intensive and the 

contribution of the snow melt in the watershed during the spring season. The lowest 

runoff is at the end of the summer -August-September.(Fig. 2) 

Discharge [m 3 /s] 

25 

20 

15 

10 

5 

0 

Danube river basin 


r. Osam 

t. Lovech 

1 3 5 7 9 11 

Months 

Fig. 2. Average annual hydrograph 

The mountain part of the area is covered with mixed forest and broad-leaved forest. 

Running down the river valleys appear Non-irrigated arable land and permanently 

irrigated arable land (Fig.3).


Fig. 3. Land cover (CORINE) in the study area 

The mountain and semi mountain part of the region is covered by mountain meadow, 

brown forest, grey and light grey forest (pseudopodzolic) soils. The hilly and low region 

is covered by grey forest slightly and moderately loamy, typical chernozems and 

calcareus chernozems. Along the rivers there are alluvial and alluvial - meadow, sandy 

and loamy. Soil types corresponding to Dokuchaev classification (Fig.4). 



Fig. 4. Soil types in the study area 

The SWAT model is chosen for effective river basin management. It is a flexible model 

that can be used under a wide range of different environmental conditions. 

Data Preparation 

The first stage in the construction of the hydrological model is creating the various input 

datasets from the multiple sources. 

The rainfall is an important characteristic of the hydrologic process in a watershed. 

Accurate estimation of the spatial distribution of observed precipitation is valuable for 

input to hydrologic modeling. 

In most cases raingauge networks are generally sparse and insufficient to capture the 

spatial variability across all subwatersheds. In this study is presented an application of 

GIS tools in estimation and spatial interpolation of the daily accumulated precipitation. 

ArcGIS Geostatistical Analyst is used for spatial data exploration. ArcGIS Geostatistical 

Analyst effectively bridges the gap between geostatistics and geographic information 

system (GIS) analysis. 

Deterministic - Inverse Distance Weighted (IDW) и Radial Basis Function (RBF), 

Geostatistical - Kriging and CoKriging interpolations are applied. 



Deterministic methods of interpolation create surface using the measured points and the 

condition of similarity. 

Inverse Distance Weighted (IDW) interpolation considers that the points which are close 

to each other have similar values of measurements. IDW forms weights from surrounding 

measured values to predict values at unmeasured location. The closest measured values 

have the most influence. Cross - validation evaluates how well the interpolation model 

predicts the unknown values. For all points cross validation sequentially omits a point, 

predict its value using the rest of the data and then compares the measured and predicted 

value. The result is presented in Fig. 5. 

Fig. 5. Spatial distribution of the daily precipitation on May, 25 in 2005 

The weights in Radial Basis Function - Spline with tension are determined by the 

number of equations and the requirement when a prediction is in a measured point, the 

estimated value must be equal to the measured one. The result is presented in Fig. 6. 


It is well known that the altitude influences the formation of precipitation amount. 

Rainfall is higher over the mountains. Cokriging is an advanced interpolation method that 

improves interpolation by taking into account secondary variable. In this study DEM 

(Digital Elevation Model) is included to improve spatial distribution of the precipitation. 

The result is presented in Fig. 7. 




After cross validation comparison for interpolation of measurements Cokriging is applied 

(Fig. 8). 

Fig. 8. Comparison of interpolation methods 

The precipitation grid with pixel size 500 m is used to obtain average areal precipitation 

in each centroid point of the subwatersheds. A spatial distribution of the measured 

precipitation in a grid for a watershed of the Osam river is presented in Fig. 9. The red 

dots present measured precipitation at meteorological stations and yellow dots present 

gravity points of the subwatersheds. 



Fig. 9. Spatial distribution of the measured precipitation in a grid for a watershed 

of the Osam river 

The methodology of SWAT model is to associate for each subbasin the precipitation 

station that is closest to the subbasin centroid. This assumption may lead to 

underestimation or overestimation of the amount of precipitation in some subwatersheds 

in which there are no direct observations (Fig. 10). 



Fig. 10 

Conclusion 

The presented approach for determining the average rainfall for a subcatchment would 

significantly improve the SWAT modeling results, especially in areas where network 

density is not sufficient. The method described would be very useful in cases of local 

extreme events. 

A good alternative would be the use of remote sensed observed precipitation from 

satellites in areas with poor coverage from rain gauges. 


The authors are thankful to Dr. Ann van Griensven and Getnet Betrie for their support 

during the preparation of the SWAT model for the Bulgarian watersheds. 



Software integration for flood management 

Vladimir Moya 1 , Ioana Popescu 1 , Andreja Jonoski 1 , Dimitri P. Solomatine 1,2 

1 UNESCO-IHE Institute for Water Education, P. O. Box 3015, 2601 DA Delft, The Netherlands 

2 Water Resources Section, Delft University of Technology, The Netherlands 

Abstract 

The following article describes a file based integration used to integrate the hydrological 

(0D) HEC-HMS, the 1D hydraulic HEC-RAS and 2D hydraulic Sobek. Such integration 

allowed to run the models one after the other as a single one, taking the output from one 

model as the input for the next one. Furthermore, such integration not only allowed to 

overcome model limitations, but also opened new opportunities such as an online 

application where the user is able to execute the models inside a website and get the 

results. File based integration was easily coupled with new application as Google Earth in 

order to produce flood maps and animations easily accessed by the common user. 

Introduction 

Water is a vital element for life, and any human activity relates with water and needs 

water. Nevertheless, water might also become a destructive force that threatens human 

life. Therefore, is important to analyze and evaluate different scenarios of this humanwater 

symbiosis. Numerical models are very useful tools in order to get fast and reliable 

results concerning different scenarios. Nevertheless, it would be a very complex task to 

develop a single model able to represent the whole set of processes and interactions 

within a catchment. One way to overcome such limitation is to integrate different models. 

Information intervention 

As stated by Price and Thompson (2002), information intervention is a crucial aspect of 

contemporary life that might be taken to prevent conflict, as part of conflict, or as a post 

conflict reconstruction. For instance, the Google Earth Global awareness entitled "Crisis 

in Dafur" shows the impact of Google Earth not only to represent what is happening in 

the ground, but also to create awareness about specific events (Parks 2009). The 

following article also demonstrates the benefits of Google Earth as an easy and accessible 

tool for presenting results to the people. 

Case study description 

The Timis Bega catchment is a complex system, with a topography changing from a 

mountainous area upstream with a slope about 20 m/Km that decreases to a slope about 

25 cm/Km in the downstream. This area has suffered from many floods throughout time. 

Moreover, on April 2005 due to an extraordinary flood wave on Timis River Romania 

faced a flood that affected circa 30000 ha and several villages (Aldescu, 2006). 



Model integration 

The models were integrated using a file based integration, where each model results are 

written in an output file, which in time is used as input file by the next model. 

The integrated model was composed by HEC - HMS (0D), HEC - RAS (1D) and Sobek 

(1D2D). The measured precipitation was provided to the data storage HEC-DSS, which 

was used as input for the HEC-HMS model which transform such rainfall into runoff. 

The hydrograph obtained were used as input for the hydraulic model HEC-RAS that 

computed the water surface elevations. This computed discharge hydrographs were used 

as input into the SOBEK1D-2D model that simulated the floodplain inundation (Popescu 

et al 2010). Finally, the results were visualized not only in GIS, but also in Google earth. 

Figure 1. Software integration 

The integrated model could easily be copied to other computers, so that different 

scenarios could be tested. For sending the files to one single computer it was decided to 

use the free Google docs application, so that they could be easily accessed from 

anywhere. 

Conclusions 

File based integration proved to be a simple and practical way not only to deal with 

model limitations, but also to opens the possibility for online application which would 

increase the interaction with users. Once the models were created and calibrated 

individually, they can easily be integrated. Furthermore, the integration with new IT tools 

such as Google Earth gives an easy way to communicate results with the common user. 

Therefore, it is also possible to overcome the limitation imposed by GIS of special 

software and technical knowledge needed in order to access the results. 

Figure 2. Flood map visualized on Google Earth 



Figure 3. Flood animation on Google Earth 

Refrences 

Aldescu C. (2006) The necessity of flood risk maps on Timiş River. Paper presented at the XXIVth 

Conference of the Danubian Countries 

Parks L. (2009) Digging into Google Eartth: An analysis of "Crisis in Darfur". Geoforum (40), 535-545 

Price M.E., Thompson M. (2002), Forging Peace: Intervention, Human Rights and the Management of 

Media Space. University of Indiana Press, Bloomington 

Popescu, I., Jonoski, A., van Adel, S.J., Onyari, E., Moya Quiroga, V., (2010). Integrated Modelling for 

Flood Risk Mitigation in Romania: Case Study of the Timis-Bega River Basin. Journal of River Basin 



Dealing with uncertainties in remotely linked models 

Nagendra Kayastha, Ann van Griensven and Dimitri P. Solomatine 

UNESCO-IHE Institute for Water Education, P.O. Box 3015, 2601 DA Delft, The Netherlands 

Computer-based mathematical models have been increasingly used to describe and 

predict hydrological and environmental processes. These processes include a variety of 

complex issues (climate, hydrology, soil and land use, ground water, water quality, etc.) 

which involve knowledge from a range of disciplines and cannot be treated in isolation 

from each other. Couple of decades ago modelers used single models to describe 

relatively simple problems. Due to the awareness of the interactions of many domains 

(e.g. climatology and hydrology; hydrology and ecology), more and more models are 

integrated or linked to each other. As we moved from single models to more integrated 

models in the 90s, we are now moving from a single PC based modeling to a distributed 

(remotely linked) modeling configurations, using, for example, web services. The reasons 

for this relate not only to computational efficiency, but also to the necessity to ensure that 

the individual models are being used, maintained and updated by the relevant expert 

teams. 

When models are used for decision making, it is crucial that the uncertainties of the 

outcomes are properly described and estimated. In a complex, and also remotely linked 

modeling system this is a real challenge, especially due to the multiple sources of 

uncertainty. In a climate-hydrological linked model the sources may be emission 

scenarios, climate models, downscaling, hydrological forcing inputs, model parameters, 

model structure, etc. The problem is that uncertainty arising from different sources and 

models increases the number of simulations to be analyzed, but the limits on the 

computing power may force researchers to limit the number of realizations, the time step 

and length of the assessment period, as well as the spatial scale. Therefore there is a clear 

need to increase efficiency of uncertainty analysis methods. 

Linking of individual models will be done more and more on a cloud (using, for example, 

temporal spatial data infrastructure -TSDI) with help of web services. The main idea of 

this type of configuration is to get away from rigid links between domains and explicitly 

communicate the uncertainties associated with predictions. The final data will be 

designed to be simple enough to be used by non-experts within decision-making. In 

Figure 1 a logical linking and information flow between data and model is described. M1, 

M2, and M3 are the individual models which are run on a desktop PC (maybe at different 

locations). The models not only receive data from the cloud (different data repository), 

but also send model outputs to the cloud. 



Figure 2: Logical linking and information flow between data and models 

This planned research project will deal with the following aspects: 

‐ Evaluation of different sampling strategies that allow for increasing efficiency, 

and using non-sampling methods like UNEEC (Shrestha and Solomatine 2009) 

‐ Use of statistical descriptors to describe the uncertainties of model outputs in a 

concise way and the potential use of different database structures to allow for 

exchanging the uncertainties associated to model results (DUE, WaterML, 

UncertML) 

‐ Testing different techniques that can be used to propagate the uncertainties in 

remotely linked models. 

This study proposes a framework that can be applied for the full cascade but will 

probably focus mainly on the climate modeling component, i.e. different climate models 

and downscaling methods. One of the standard uncertainty analysis methods is based on 

using the model ensembles. Disadvantage of these methods is the large amount of data 

that needs to be transferred and that large number of simulations is needed. In this study, 

we plan to use generic uncertainty descriptors that allow to generate ensembles or to 

reproduce the uncertainty bounds for the model results. Such an approach can be 

enhanced by using standards to transfer the uncertainty related to data and/or model 

results (e.g., Standardized meta-data structures) in order to allow for generating new sets 

of ensembles (realisations) without additional model runs. 

Case study. An illustration of the assessment of the cascade uncertainty across climate 

and hydrological models in the Nzoia catchment in Kenya is given. Here we sought to 

link models by remote configuration across domains (e.g. linking climatology and 

hydrology) and across sources of knowledge (e.g. combining empirical data with the 

model and model to model) to investigate the physics of modern climate data to interpret 

stream flow. 

For this case study, we will describe and quantify sources of uncertainty, and its 

propagation through input data from observations and output ensembles through regional 

climate Model (PRESIS) and hydrological model (SWAT). In addition we will 

implement Temporal Spatial data infrastructure (TSDI) for publishing, retrieving and 



accessing the data with the associated uncertainty using web services under standard, 

protocol of Open Geospatial Consortium (OGC) and WaterML/UncertML standards. 

This research will help hydroinformatics specialists and decision makers to use complex 

integrated models whereby both model results and their uncertainties are propagated from 

one domain model to another. 



Determining the effects of water quality and quantity on 

public health by using GIS technology 

Ahmet Özgür Doğru*, Seval Alkoy**, Necla Uluğtekin*, 

Filiz Bektaş Balçık*, Çiğdem Göksel*, Seval Sözen* 

*Istanbul Technical University, Civil Engineering Faculty, İstanbul, Turkey 

**Abant Izzet Baysal University, Faculty of Medicine, Bolu, Turkey 

Epidemiology is the scientific study of the spread and control of diseases as a function of 

time and location. Epidemiology follows the patterns of a disease on people in their 

healthy and un-healthy periods; and tracks their history of illness especially paying 

attention to when and where they have it. Epidemiologists have traditionally used maps 

while analyzing the relationship between location, surrounding environment, and the 

disease. GIS has been used in surveillance and monitoring of vector-borne as well as 

water-borne diseases, and the ones in relation to the environmental health, disease 

policies and planning, the existing health situation in the area, generation and analysis of 

research hypotheses and etc. GIS has enabled researchers to determine locations of high 

prevalence areas and populations at risk. 

Access to a sufficient supply of safe water is essential in maintaining public health. 

Situations with inadequate water directly and indirectly affect health. Poor hygiene 

caused by the lack of water results in the increased transmission of infectious diseases. 

Where the sources of potable water are of poor quality, or the financing, staffing and 

other infrastructure to maintain the distribution system are lacking, mortality rates 

attributable to infectious diseases, to the availability of sanitation services and to general 

hygiene may increase. Inadequate water supplies increase the likelihood of person-toperson 

disease transmission and can compromise the effectiveness and efficiency of 

water-based sewage collection and treatment processes, posing an additional risk of 

disease. In summary, the changes in water quality and quantity cause public health 

problems directly or indirectly. While drinking water quality is an essential factor for 

maintaining the public health, flooding and drought are also considered as important 

factors causing health problems. 

This study aims at determining the health effects of the change in water quality and 

quantity at Black Sea Cost of Turkey by Integrating GIS technology with health statistics. 

This study is being developed as a subtask of the Project enviroGRIDS @ Black Sea 

Catchment, which is a European Union FP 7 Project aiming at building the capacity of 

scientist to assemble such a system in the Black Sea Catchment, the capacity of decisionmakers 

to use it, and the capacity of the general public to understand the important 

environmental, social and economic issues at stake. In this context, 10 year data 

(covering the period of 2000-2009) of water related diseases such as diarrhea, hepatitis A, 

shigella dysentery, typhoid and paratyphoid, including the number of case and deaths in 

Turkey is going to be examined and high and low risk areas are going to be determined 

by using GIS and incidence graphs of each disease. Detailed studies are going to be 



executed for determining the effects of the water quality and quantity change over health 

by using the water related data provided for the most and the least risky areas. Health 

statistics are going to be used for determining the correlation between the case/deaths and 

water quality/quantity. This paper covers two main parts of the overall study; (1) 

literature search indicating the relation between health and water quality/quantity based 

on the technical reports of World Health Organization (WHO) and Intergovernmental 

Panel on Climate Change (IPCC), (2) the methodology description of the risk area 

determination process in the study area in terms of the water related diseases and 

implemented case study. Additionally, the road map of the future studies of the project 

studies is going to be discussed in this paper together with the proposed GIS related 

methodologies. 



Ecosystem monitoring using digital change detection 

method; 

example of Iğneada wetland forest 

*Filiz Bektas Balcik, *Cigdem Göksel, *Gonca Bozkaya, 

*Ahmet Özgür Doğru, *Necla Uluğtekin, *Seval Sözen 

*Istanbul Technical University, Civil Engineering Faculty, İstanbul, Turkey 

Natural areas are under threat of the human all over the world as a resource of human 

basic needs. Devastating and uncontrolled land use changes have occurred in the natural 

areas for urbanization purposes as a result of population increase. Detection and 

monitoring land use and land cover changes have been widely applied for management 

and decision making over large geographical areas using multitemporal remotely sensed 

data. It is possible to provide economic, accurate, temporal, reliable and updated 

information from satellite images. Chance detection has not only an important role in 

monitoring of the current situation of the study area, but it also provides base data for 

modeling the future land cover/use patterns that will prevail in protected areas as a result 

of different development scenarios. 

The study area, İğneada mangrove forest, is one of the important protected areas of 

Turkey. The area houses different kinds of ecosystems and a wide range of biodiversity; 

the parts of it have previously protected as Nature Protection Park, Natural Site, and 

Wildlife Protection Area. İğneada is located at the Black Sea coast in the northwest of 

Turkey that is 20 km away from Bulgarian border. It lies on an area that is approximately 

3000 ha and between the coordinates of 41ᵒ 52’ 34” N and 27ᵒ 59’ 10” E. Alluvial forests 

with associated aquatic and coastal ecosystems include freshwater and saline lakes, 

coastal dunes, freshwater and low salinity marshes, longos forests, and mixed forests of 

deciduous tall trees such as Quercus robur subsp. robur (English oak), Q. petraea subsp. 

iberica (sessile oak), Fagus orientalis (Oriental beech), Carpinus betulus (European 

hornbeam), C. orientalis (Oriental hornbeam), Fraxinus angustifolia subsp. Oxycarpa 

(narrow-leaved ash), and Alnus glutinosa (black alder). Despite its ecological sensitivity 

and importance, İğneada has been under serious threats of the projects such as supplying 

drinking water project to Istanbul, a harbor project, and a coastal road project. The area 

was declared as the national park in 2007. The announcement of the national park is 

expected to contribute to the sustainable development of the area. In addition, great 

variety of urban pressures have gradually increased such as expansion of summer houses 

on and around wetlands, increasing recreational uses on coastal line (mainly off-road 

racing), and wetland pollution due to sewage of settlements. 

The purpose of this study is to detect the changes in İğneada Longos forest and its 

surroundings using medium resolution remotely sensed data. This study is being 

developed as a subtask of the Project enviroGRIDS @ Black Sea Catchment, which is a 

European Union FP 7 Project aiming at building the capacity of scientist to assemble 

such a system in the Black Sea Catchment, the capacity of decision-makers to use it, and 



the capacity of the general public to understand the important environmental, social and 

economic issues at stake. The primary research materials include multitemporal Landsat 

TM satellite images from 1990 and 2010. Landsat images have 30m spatial resolution 

and seven bands. Five steps were employed to conduct the research and these are: (i) 

geometric and radiometric correction of satellite images, (ii) Layer Stacking (iii) PCA 

based change detection (iv) hybrid classification (v) accuracy assessment of the method 

using ground truth data. The preliminary results show that there are land use and land 

cover changes in the selected region as a result of cultural pressure. These ongoing 

pressures and their consequences should be taken into account for sustainable 

management of İğneada protection area. 



Environmental information for citizens via mobile phones: 

case studies in the Province of Noord Brabant, The 

Netherlands 

Adrian Almoradie, Andreja Jonoski, Schalk-Jan van Andel and Ioana Popescu 

UNESCO-IHE Institute for Water Education P.O.Box 3015, 2601 DA 

Delft, The Netherlands 

Demonstrator applications of an environmental information system for citizens’ access to 

water quality information (in particular – bathing water quality) are presented. These are 

being developed for catchments and fresh water lakes in The Netherlands in collaboration 

with the Dutch water authorities. The presented demonstrators are currently being 

deployed the World Wide Web, but system components are also developed for mobile 

phones. The main goals of the system are: improvement of the water quality information 

via better and faster integration of existing data; integration of data and water quality 

models for providing information in periods without measurements, or for forecasting 

purposes; assimilating user feedback on water quality which may come in semi-structured 

forms (SMS messages, photos, etc) via users’ mobile devices (phones). 

Mobile phone (using Google Android OS) and the Web platform 

Mobile phones are currently one of the best possible means to send and receive real time 

information. The field of hydroinformatics has already embarked on exploring the 

potential for such applications for the water domain (see Abbott and Jonoski, 2001). 

Examples of applications in water- and environment-related fields: use of SMS 

messaging for operational interventions in management of water supply systems (Segura, 

2006); use of Java Midlets technology for flood forecasting (Naz, 2006); use of mobile 

phone web browsers for supporting field sampling of data (Fenrich et al. 2009); 

In the world of Smart phones (operating over the 3G - third generation mobile networks), 

new features such as seamless connection with the internet, the use of GPS (Global 

Positioning System) and other applications has been made possible, The competing OS 

(Operating System) platform in the market are the Symbian, Windows Mobile, 

Blackberry, iPhone and recently the Google Android OS. In this study the Google 

Android OS was used because of its flexibility in developing applications and it is 

expected that in the next couple of years it may emerge as one of the dominant platforms 

for mobile phone applications. The Android OS Phone uses the Java technology. 

For the web platform, standard client-server side technologies were used. For the client 

side technology HTML, Javascript, AJAX and Google API’s (for its visualization and 

interactivity) were used and for the server side is the PHP scripts. 

Case study application: Province of Noord Brabant, The Netherlands 

Demonstrator for bathing water quality monitoring and dissemination has been developed 

for two case studies in the province of Noord Brabant, The Netherlands. 



The province of Noord Brabant is located in the southern part of the Netherlands. The 

province and the Waterboards in the area monitor the water quality of the rivers in 

Dommel and the lakes in Brabantse delta. Lakes are mainly used for re-creational 

activities. The pollution in the river and lakes comes in a form of Phosphorus (P) and 

Nitrogen (N) nutrients that is mainly use for agriculture. Both case studies in this area 

aim to provide information of the water quality of the river or lakes to the citizens, this 

information may either come from the models or through monitoring. 

Dommel case study 

The demonstrator was first developed for the Dommel case study. It was developed to 

access the modeled river water quality results in the server. The Soil and Water 

Assessment Tool (SWAT) was used as the modelling tool (Figure 1). To access the data 

of the lakes via spatial location, the Google maps view was used (Figure 2). In the back 

end, the model results was first pre-processed and then uploaded in the server in a proper 

file format. 

Figure 1. Mobile phone – server communication Figure 2. Screen shots Dommel Application 

Brabantse delta case study 

In this case study a demonstrator has been developed connecting both a web platform and 

the mobile phone. This case study is more complex than the first case study because of 

this inter-connection. It has been developed to access historical and latest water quality 

monitored data both via the web platform and mobile phone (Figure 3). Both platform 

implement a Google maps based API as an interface to access information of each lake. 

Graphical charts were also made available (Figure 4). Provision of user feedback via the 

mobile phone has been implemented as a new feature; it has been designed to send 

monitored data or textual information for any selected lake. Once a monitored data or 

textual information is sent, the server will receive the data, update the database and 

display the new data. The web platform automatically updates its self when new 

information comes in. 



Conclusion 

Figure 3. Phone and web application architecture Figure 4. Screen shots Brabantse delta Application 

The design and functionality of the two demonstrators was developed and tested 

successfully. In application to water quality dissemination, the integration of web 

platform and mobile phone has a promising potential to reach interested citizens. 

Although these demonstrators were developed successfully, they still need to be tested 

extensively with the public/citizens to prove their value, and activity which is currently 

being scheduled. 


This work has been developed within the European FP7 research project LENVIS - 

(Localised ENVironmental and health Information Services). 

References 

Abbott M.B. and Jonoski A., 2001: The democratisation of decision making processes in the water sector 

(II), Journal of Hydroinformatics, 03.1, pp. 35-48. 

Fenrich E., Brodt A. and Nicklas D., 2009: WODCA: A Mobile Web-based Field Sampling Support 

System, Proceedings of the 8th International Conference on Hydroinformatics, Concepción, Chile. 

Naz, N.N., 2006: Urban Flood Warning System with Wireless Technology: Case Study of Dhaka City – 

Bangladesh. MSc Report, UNESCO-IHE. The Netherlands. 

Segura, J.L.A., 2006: Use of Hydroinformatics Technologies for Real Time Water Quality Management 

and Operation of Distribution Networks. Case Study of Villavicencio, Colombia. MSc Report, UNESCO- 

IHE. The Netherlands. 

http://www.lenvis.eu 



Study regarding delineation of flood hazard zones in Dej 

area 

Florin Stoica, Mihai Sarb, Horea Selagea and Radu Dulau 

Romanian Waters National Administration – Somes-Tisa Basinal Administration – Cluj-Napoca 

Floods have the potential to cause fatalities, relocation of people, damage to the 

environment and they can severely compromise economic development and damage the 

economic activities of the society. 

While developing policies on water and land uses, countries should take into 

consideration the potential impacts that such policies might have on flood risks and the 

management of flood risks. There are different types of floods that should be taken into 

consideration while developing the policies, such as: river floods, flash floods, urban 

floods and floods from the sea in coastal areas. The damage caused by flood events may 

also vary across the countries and regions of a river basin. Hence, objectives regarding 

the management of flood risks should be determined by the countries which are sharing 

the basin under consideration and should be based on local and regional circumstances. 

In order to implement the EU directives, Romania is developing at a national level a plan 

for prevention, protection and minimization of the flood effects in the Somes-Tisa 

hydrographic basin. The plan is implemented by the Somes - Tisa Water Board. 

The present paper presents briefly the pilot study basin, which is part of the national 

prevention plan. The study area is in the Dej region, at the confluence of two important 

rivers, Somesul Mare and Somesul Mic, where the flood of 1970 had a catastrophic 

impact. Dej area is a priority in the above mentioned plan. 

The Somes River Basin located upstream of the city of Dej, has a surface of 8856 km 2 ., a 

medium height of 647 mASL and a length of the river Somes of 136 km. The Somes 

River Basin has two major components: Somesul Mare and Somesul Mic and covers 

almost entirely Bistrita – Nasaud county and partially Cluj county. The combined surface 

of the two mentioned river basins totals 42% of the entire Somes-Tisa Basin (fig.1). 

The digital terrain model was created in order to delineate the flood hazard areas, using 

LIDAR technology 



Figure 3. Dej area and altitude differences in the Somes basin 

Somesul Mare (from its spring to the confluence with Somesul Mic) is located in Bistrita 

– Nasaud county. Its hydrographic basin lies in the contact area between the Eastern 

Carpathians with the Transylvania Depression. It is a varied and complex territory, being 

composed of mountains (36%) which open as an amphitheatre towards the Somes Mare 

Valley, ranging from 800 m to 2279 m in height and hills (64%) which belong to the 

Transylvanian Plateau, ranging between 400 m and 800 m in height. 

Somesul Mic (from its spring to the confluence with Somesul Mare) is located in Cluj 

county, occupying 56.6% of the county’s surface, stretching across 6674 km 2 ., at the 

border of three major natural units: the Western Carpathians, the Somes Plateau and the 

Transsylvanian Lowlands. It’s main tributaries are Capus, Nadas, Borsa, Gadalin, Fizes 

and Bandau rivers. The discharge of the Somes Mic river is regulated by the existing 

reservoirs in the upstream part of the basin. The purpose of these reservoirs are to 

produce hydroelectric power. The altitude in this basin varies from 1836 mASL 

(Vladeasa Mountain), to a minimum height of 227 mASL, where Somes rivers are 

flowing out of Cluj county. 

Somes river is formed after the confluence of the two main tributaries – Somesul Mare 

and Somesul Mic, which brings 64%, 36% respectively , of the 74.1 m³/s average 

discharge. The yearly average discharge on several sections along the Somes River are: 

Beclean (48,7 m 3 /s), Salatiu (21,5 m 3 /s), Dej (76,8 m 3 /s), Ulmeni (89,8 m 3 /s), Lapuşel 

(19,2 m 3 /s), Satu Mare (126 m 3 /s). 

An analysis of the 100 years return period flood on Somes Mare (0.3 m 3 /s/km 2 ) and 

Somes Mic (0.2 m 3 /s/km 2 ) river upstream of their confluence, shows that the flood 

occurring on the Somes Mare river are much more important than the Somes Mic basin. 

The obtained flood extent for the 100 years return period 



Figure 4.Flood extent for 100 years return period 

În the past 35-40 years, the most important flood was the one of 1970, corresponding to 

the 100 years flood in Dej area. The most devastating floods were the result of a 

combination of rainfall with snow melt. In the past 15 years, there is a tendency of 

having earlier floods (March-April instead of May-June) and more often winter floods 

(December to February); also there are some areas with recurrent floods and a more 

torrential aspect of the rainfall and drainage. 

After the construction of the upstream reservoirs: Tarnita (1974) and Fantanele (1978) in 

Somesul Mic basin, there were no major floods reported. The modified flow regime 

indicates that statistically the 100 years return period flood discharge, decreased by 15- 

20% than in the natural regime of the river, in the Cluj area and, by 10% , upstream of 

the confluence with Somesul Mare. 



A Cloud-based Virtual Observatory for Environmental 

Science 

Gordon S. Blair and Yehia El-khatib 

School of Computing & Communications, Lancaster University, UK 

Environmental scientists are increasingly being asked to answer more complex scientific 

questions for example related to the implications of environmental management decisions 

on a catchment or on the policy implications of climate change at a national or 

international level, a trend we refer to as ‘big science’ for the purposes of this paper. 

However, there are a number of obstacles that restrict environmental scientists from 

tackling such big science issues. Many of these obstacles are technical and related to 

different forms of fragmentation including spatial and temporal gaps in data, disrupted 

alignment and representation of data, unlinked models, and disjoint disciplines. Other 

obstacles include the difficulty of managing and processing extremely large datasets. 

The EVOp project seeks to solve many of these problems by developing a virtual 

observatory (VO) that enables the integration of a variety of information sources 

(including disparate data sets, sensor data and models) at different granularities and 

scales. The VO will also provide interoperability with associated information services 

and encourage the flow from data to knowledge to policy setting in the quest for 

answering big science questions. 

The VO requires the construction of a cyber-infrastructure that provides simplified access 

to data, integration with current information services, and the ability to handle large 

datasets. Cloud computing offers great opportunities to undertake such tasks through 

harnessing economies of scale offered by commodity hardware data centres. It also offers 

universal access to the VO that can be used by scientists, policy makers and the general 

public alike and, through this, achieving the desired level of integration. 

Cloud computing is a distributed paradigm in which computational and storage 

requirements are provided in an on-demand fashion by large clusters of commodity 

computers. Such a pay-per-use model is generally made possible through virtualisation, 

i.e. using virtual machines to create custom execution environments. To the consumers of 

such service, they obtain customised and isolated computing resources as and when 

required without the need to invest in hardware that might not be fully utilised, which 

will depreciate in value, and will require operation, support and maintenance costs. 

Virtualisation also allows cloud service providers to manage large data centres at a low 

overall maintenance cost, and provide varying and tiered services to different customers. 

Different levels of cloud services are available. Users of Software as a Service (SaaS) are 

able to run software through a thin client, such as a Web browser, without actually 

running the application on their computers. Platform as a Service (PaaS) provides a 

platform (i.e. a virtualised hardware setup along with an operating system) that can be 

used to deploy highly customised software. Finally, Infrastructure as a Service (IaaS) 

provides a pool of virtualised hardware resources for the user to use and manage. 



There are a number of challenges involved in developing an environmental cloud, 

including most prominently what architecture is right to support Environmental Sciences. 

Hardware assets that will be required by an environmental cloud could be leased from 

one or more cloud service provider, such as Amazon and Google. Alternatively, 

dedicated data centres could be assembled to serve the needs of the VO. A hybrid 

approach is also possible where owned data centres would serve as primary resources 

while overflow requests would be redirected to third-party cloud service providers. 

Second is the issue of the specific services that will be provided by the VO. Clearly, some 

SaaS provisioning is required to enable especially non-specialist users to access the VO. 

However, further services are required to enable environmental scientists to exploit the 

capabilities of the cloud to tackle their specific problems by mining into extremely large 

datasets. Adapting a distributed programming framework such as MapReduce to the 

needs of the environmental community is an example. 

The third challenge is data discovery. Observations and models are generated by a large 

number of sources within and outside the scientific community. The VO will provide the 

potential to distinguish associations and overlaps between different datasets and models. 

However, there is little incentive for many producers to share their data and models. 

Through the use of publishing incentives, EVOp partners will work on encouraging such 

producers to register their data and models. Easy-to-use desktop tools will be employed 

to ingest datasets and models into the cloud. 

Data collected by the VO would need to be aligned to standard formats understood by 

different research communities, e.g. INSPIRE and data.gov. This normalisation of 

representation is necessary to facilitate low overhead curation, interoperability, and 

serendipity. The final challenge is to enable access to the data, model and visualisation 

tools. This will be achieved via use of a multi-faceted portal that caters for the needs of 

the different stakeholders, i.e. specialists, policy makers, and local communities. The use 

of a RESTful architecture would allow this access portal to use Web browsers that are 

universal client tools that are accessible to everyone and from a great deal of devices. It 

also allows the portal to harness common visualisation tools, such as Google Maps and 

Google Earth. 

Cloud computing offers promising opportunities to enable the environmental community 

to better tackle challenges that face local communities, nation states and the international 

community. EVOp is a 2-year pilot project aimed to demonstrate how such technology 

can be used in environmental management, to help set international standards for 

exchanges of data and models, and to stimulate discussion within the environmental 

community both nationally in the UK and internationally. We are looking to collaborate 

with the other efforts on the regional and international scales to assist with this journey. 


The authors wish to acknowledge the Natural Environment Research Council for funding 

the EVOp project under grant reference NE/I002200/1. The authors also wish to thank 

the valuable contributions of the EVOp consortium.

Extended abstracts of the OpenWater Symposium in pdf

Create successful ePaper yourself

Delete template?

Save as template?