Lake Sediments as Archives of Recurrence Rates and ... - Eawag
Lake Sediments as Archives of Recurrence Rates and ... - Eawag
Lake Sediments as Archives of Recurrence Rates and ... - Eawag
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
<strong>Lake</strong> <strong>Sediments</strong> <strong>as</strong> <strong>Archives</strong> <strong>of</strong> <strong>Recurrence</strong> <strong>Rates</strong><br />
<strong>and</strong> Intensities <strong>of</strong> P<strong>as</strong>t Flood Events<br />
Adrian Gilli, Flavio S. Anselmetti, Luk<strong>as</strong> Glur, <strong>and</strong> Stefanie B. Wirth<br />
1 Introduction<br />
Palae<strong>of</strong>lood hydrology is an exp<strong>and</strong>ing field <strong>as</strong> the damage potential <strong>of</strong> floods <strong>and</strong><br />
flood-related processes is incre<strong>as</strong>ing with the population density <strong>and</strong> the value <strong>of</strong> the<br />
infr<strong>as</strong>tructure. Assessing the risk <strong>of</strong> these hazards in mountainous terrain requires<br />
knowledge about the frequency <strong>and</strong> severness <strong>of</strong> such events in the p<strong>as</strong>t. A wide<br />
range <strong>of</strong> methods is employed using diverse biologic, geomorphic or geologic<br />
evidences to track p<strong>as</strong>t flood events. Impact <strong>of</strong> floods are studied <strong>and</strong> dated on<br />
alluvial fans <strong>and</strong> cones using for example the growth disturbance <strong>of</strong> trees (St<strong>of</strong>fel<br />
<strong>and</strong> Bollschweiler 2008; Schneuwly-Bollschweiler <strong>and</strong> St<strong>of</strong>fel 2012: this volume)<br />
or stratigraphic layers deposited by debris flows, allowing to reconstruct p<strong>as</strong>t flood<br />
frequencies (Bardou et al. 2003). Further downstream, the cl<strong>as</strong>sical approach <strong>of</strong><br />
palae<strong>of</strong>lood hydrology (Kochel <strong>and</strong> Baker 1982) utilizes geomorphic indicators<br />
such <strong>as</strong> overbank sediments, silt lines <strong>and</strong> erosion features <strong>of</strong> floods along a<br />
river (e.g. Benito <strong>and</strong> Thorndycraft 2005). Fine-grained sediment settles out <strong>of</strong><br />
the river suspension in eddies or backwater are<strong>as</strong>, where the flow velocity <strong>of</strong> the<br />
river is reduced. Records <strong>of</strong> these deposits at different elevations across a river’s<br />
A. Gilli ()•S.B.Wirth<br />
Geological Institute, ETH Zurich, CH-8092 Zurich, Switzerl<strong>and</strong><br />
e-mail: adrian.gilli@erdw.ethz.ch<br />
F.S. Anselmetti<br />
<strong>Eawag</strong>, Department <strong>of</strong> Surface Waters, CH-8600 Dübendorf, Switzerl<strong>and</strong><br />
Institute <strong>of</strong> Geological Sciences, University <strong>of</strong> Bern, Baltzerstr<strong>as</strong>se 1C3, CH-3012 Bern,<br />
Switzerl<strong>and</strong><br />
e-mail: flavio.anselmetti@geo.unibe.ch<br />
L. Glur<br />
Department <strong>of</strong> Surface Waters, <strong>Eawag</strong>, CH-8600 Dübendorf, Switzerl<strong>and</strong><br />
e-mail: luk<strong>as</strong>.glur@eawag.ch<br />
M. Schneuwly-Bollschweiler et al. (eds.), Dating Torrential Processes on Fans <strong>and</strong><br />
Cones, Advances in Global Change Research 47, DOI 10.1007/978-94-007-4336-6 15,<br />
© Springer ScienceCBusiness Media Dordrecht 2013<br />
225
226 A. Gilli et al.<br />
pr<strong>of</strong>ile can be used to <strong>as</strong>sess the discharge <strong>of</strong> the p<strong>as</strong>t floods. This approach <strong>of</strong><br />
palae<strong>of</strong>lood hydrology studies w<strong>as</strong> successfully applied in several river catchments<br />
(e.g. Ely et al. 1993; Macklin <strong>and</strong> Lewin 2003; O’Connor et al. 1994; Sheffer et al.<br />
2003; Thorndycraft et al. 2005; Thorndycraft <strong>and</strong> Benito 2006). All these different<br />
reconstruction methods have their own advantages <strong>and</strong> disadvantages, but <strong>of</strong>ten<br />
these studies have a limited time coverage <strong>and</strong> the records are potentially incomplete<br />
due to lateral limits <strong>of</strong> depositional are<strong>as</strong> <strong>and</strong> due to the erosional power <strong>of</strong> fluvial<br />
processes that remove previously deposited flood witnesses. Here, we present a<br />
method that follows the sediment particle transported by a flood event to its final<br />
sink: the lacustrine b<strong>as</strong>in.<br />
<strong>Lake</strong>s <strong>of</strong>fer a great possibility to record flood events <strong>as</strong> their suspension load is<br />
deposited on the lake floor <strong>as</strong> a distinct detrital layer. This detrital layer contr<strong>as</strong>ts in<br />
the ideal c<strong>as</strong>e strongly with the ‘normal’ background sedimentation, which mainly<br />
consists <strong>of</strong> authigenically-produced minerals, organic remains <strong>and</strong> cl<strong>as</strong>tic particles.<br />
Detrital flood layers are detected visually or by a suite <strong>of</strong> different sedimentological<br />
<strong>and</strong> geochemical methods. Flood events that have been documented either by<br />
historical sources or by instrumentation were detected in lacustrine sediment cores<br />
(Bussmann <strong>and</strong> Anselmetti 2010; Chapron et al. 2002; Gilli et al. 2003; Lamoureux<br />
2000; Siegenthaler <strong>and</strong> Sturm 1991) confirming the concept <strong>of</strong> using lake sediments<br />
<strong>as</strong> flood archive. Major advantage <strong>of</strong> the lacustrine archive is the continuous mode <strong>of</strong><br />
the background sedimentation in the deep lake are<strong>as</strong>, which <strong>of</strong>fers the potential for<br />
the reconstruction <strong>of</strong> a temporally complete flood record. The resolution <strong>of</strong> a lakeb<strong>as</strong>ed<br />
flood reconstruction is highly dependent on the overall sedimentation rate, but<br />
in the c<strong>as</strong>e <strong>of</strong> annually laminated sediments the flood record could be resolved down<br />
to the se<strong>as</strong>on. P<strong>as</strong>t flood events can be scaled by thickness <strong>and</strong> grain-size distribution<br />
<strong>of</strong> the detrital layer <strong>as</strong> they reflect flood dynamics.<br />
Although modern limnogeology started to underst<strong>and</strong> the involved processes<br />
between high river-run<strong>of</strong>f <strong>and</strong> the deposition <strong>of</strong> detrital layers in lakes through<br />
underflows in the 1970s (e.g. Lambert <strong>and</strong> Hsü 1979; Sturm <strong>and</strong> Matter 1978),<br />
the limnogeologic community did not immediately realize the value <strong>of</strong> the lake<br />
sediments for the reconstruction <strong>of</strong> flood history. For Switzerl<strong>and</strong>, the severe<br />
damage <strong>of</strong> several flood events in the Swiss Alps <strong>and</strong> its forel<strong>and</strong> in summer<br />
1987 spurred different disciplines to reconstruct flood frequencies <strong>and</strong> intensities.<br />
Siegenthaler <strong>and</strong> Sturm (1991) systematically sampled the sedimentary archive <strong>of</strong><br />
<strong>Lake</strong> Uri, which drained an area <strong>of</strong> severe flood damage. They documented in<br />
detail the spatial characteristics <strong>of</strong> the 1987 flood layer in <strong>Lake</strong> Uri <strong>and</strong> were able<br />
to identify further strong flood events that occurred in the l<strong>as</strong>t 700 years. This<br />
intensified flood-risk evaluation after the 1987-flood event w<strong>as</strong> amplified by the<br />
fact that Switzerl<strong>and</strong> experienced only few incidences <strong>of</strong> natural dis<strong>as</strong>ters between<br />
1882 <strong>and</strong> 1976 (Pfister 2009) <strong>and</strong>, therefore, w<strong>as</strong> not aware <strong>of</strong> the natural variability<br />
<strong>of</strong> such dev<strong>as</strong>tating events. Especially in Europe <strong>and</strong> North America, the number<br />
<strong>of</strong> flood reconstructions using lake sediments (e.g. Arnaud et al. 2005; Bussmann<br />
<strong>and</strong> Anselmetti 2010; Chapron et al. 2002; Dapples et al. 2002; Eden <strong>and</strong> Page<br />
1998; Gilli et al. 2003; Mazzucchi et al. 2003; Nesje et al. 2001; Noren et al. 2002;<br />
Osleger et al. 2009; Parris et al. 2010) incre<strong>as</strong>ed exponentially in the l<strong>as</strong>t 15 years<br />
underlining the great potential <strong>of</strong> this method.
<strong>Lake</strong> <strong>Sediments</strong> <strong>as</strong> <strong>Archives</strong> <strong>of</strong> <strong>Recurrence</strong> <strong>Rates</strong> <strong>and</strong> Intensities <strong>of</strong> P<strong>as</strong>t Flood Events 227<br />
2 Depositional Processes in <strong>Lake</strong>s<br />
Extreme flood events mobilize <strong>and</strong> entrain large amounts <strong>of</strong> sediments that are<br />
consequently fed into the river drainage. The amount <strong>and</strong> the grain size <strong>of</strong> the<br />
sediment particles entrained by a flow are strongly dependent on the hydraulic<br />
energy. The coarser fraction <strong>of</strong> these mobilized particles may remain in the<br />
catchment, forming alluvial fans <strong>and</strong> cones. The finer particles, which are kept in<br />
suspension, are transported from the catchment area to the next downstream lake,<br />
which is the final sink for nearly the entire riverine sediment load (Fig. 1).<br />
When a sediment-laden river enters a lake, the sediment plume is redistributed<br />
depending on the density stratification <strong>of</strong> the lake <strong>and</strong> the density <strong>of</strong> the entering<br />
water, which is a function <strong>of</strong> temperature, suspension load <strong>and</strong> ion content (salinity).<br />
The density contr<strong>as</strong>t between sediment-laden river water <strong>and</strong> the ambient lake water<br />
determines if the sediment plume forms an over-, inter- or underflow in the lake<br />
Fig. 1 Schematic drawing <strong>of</strong> a lake <strong>and</strong> its catchment area illustrating in particular the different<br />
distribution paths <strong>of</strong> a river-borne sediment plume in a lake <strong>as</strong> over-, inter-, or underflow (Giovanoli<br />
1990; Sturm <strong>and</strong> Matter 1978). Coarser sediment particles mobilized through a flood are mainly<br />
deposited on alluvial fans or in the delta, where<strong>as</strong> the finer particles are entering the lake<br />
<strong>as</strong> sediment plume. High-density flows, generated by heavy rainfall events, form underflows<br />
depositing their particles <strong>as</strong> specific turbidite layers, which focus in the deepest lake area.<br />
<strong>Lake</strong>s dominated by underflows have a flat <strong>and</strong> near horizontal lake bottom. Seismic investigations<br />
image the subsurface providing a seismic stratigraphy that allows prediction <strong>of</strong> sediment type <strong>and</strong><br />
selection <strong>of</strong> ideal coring sites
228 A. Gilli et al.<br />
Fig. 2 Grain-size me<strong>as</strong>urements across two flood layers in <strong>Lake</strong> Thun, Switzerl<strong>and</strong> (Wirth 2008;<br />
Wirth et al. 2011). Vertical white double-arrow indicates extent <strong>of</strong> flood layer. (a) This turbidite<br />
shows an upcore incre<strong>as</strong>e in grain size between 3 <strong>and</strong> 4 cm indicating, in addition to the coarse<br />
b<strong>as</strong>e, a second peak in river run<strong>of</strong>f. (b) Turbidite layer with clearly inversed grading at the b<strong>as</strong>e <strong>as</strong><br />
a result <strong>of</strong> the waxing river run<strong>of</strong>f<br />
(Giovanoli 1990). Regular run<strong>of</strong>f waters are usually less dense than the lake water<br />
<strong>and</strong> thus rather form over- <strong>and</strong> interflows, from where the suspended fine particles<br />
slowly settle out <strong>and</strong> eventually become deposited on the lake floor. Extreme<br />
events, on the other h<strong>and</strong>, are so suspension-laden that the high density <strong>of</strong> the flow<br />
allows the water to break through the lake stratification to proceed <strong>as</strong> turbiditic<br />
underflow (hyperpycnal flow) to the deepest area <strong>of</strong> the lake (Fig. 1). Once the<br />
deepest lake area is reached, the underflow spreads out <strong>and</strong> flow velocities decre<strong>as</strong>e.<br />
Consequently, the sediment particles are deposited forming characteristic layers<br />
termed ‘turbidites’. The depositional characteristics <strong>of</strong> a turbidite are dependent<br />
on temporal run<strong>of</strong>f evolution (i.e. the shape <strong>of</strong> the hydrograph) <strong>of</strong> a flood event<br />
<strong>as</strong> the grain size is a direct function <strong>of</strong> the current velocity (Mulder et al. 2001,<br />
2003). Turbidites comprise a fining upward sequence representing the waning <strong>of</strong><br />
the discharge (Fig. 2). Sometimes, turbidites also record the incre<strong>as</strong>ing underflow<br />
velocity during the initiation <strong>of</strong> a flood by a coarsening upward b<strong>as</strong>al unit (Fig. 2b).<br />
This concept <strong>of</strong> directly linking the rising <strong>and</strong> falling river discharge with the reverse<br />
<strong>and</strong> subsequently normal graded turbidite deposition w<strong>as</strong> recently questioned for the<br />
marine realm (Lamb <strong>and</strong> Mohrig 2009), although for the smaller lacustrine setting<br />
this concept probably still is valid.
<strong>Lake</strong> <strong>Sediments</strong> <strong>as</strong> <strong>Archives</strong> <strong>of</strong> <strong>Recurrence</strong> <strong>Rates</strong> <strong>and</strong> Intensities <strong>of</strong> P<strong>as</strong>t Flood Events 229<br />
2.1 Selecting <strong>Lake</strong>s with Promising Flood Records<br />
The selection <strong>of</strong> appropriate lakes is crucial for this approach <strong>of</strong> flood<br />
reconstruction. The underst<strong>and</strong>ing <strong>of</strong> the transport <strong>and</strong> depositional processes <strong>of</strong><br />
flood-transported particles allows us to set criteria to select lakes holding promising<br />
flood records. The following two criteria are highly relevant:<br />
1. Well-defined morphobathymetric depocentre<br />
Underflows transport the suspension load <strong>of</strong> a flood along the delta slope towards<br />
the deepest part <strong>of</strong> the lake, where flow velocity decre<strong>as</strong>es <strong>and</strong> larger particles<br />
start to settle down on the lake floor. <strong>Lake</strong>s with underflow-dominated sedimentation<br />
are characterized morphologically by a flat <strong>and</strong> horizontal b<strong>as</strong>in floor, where<br />
the flood turbidites accumulate <strong>and</strong> consequently level out inherited topography<br />
(Fig. 1). This flat floor is the prime target area to recover cores for flood<br />
reconstruction studies. A detailed knowledge about the lake’s bathymetry gained<br />
from map studies, detailed depth soundings or reflection seismic surveying is<br />
therefore a prerequisite for a successful flood reconstruction. Although in most<br />
<strong>of</strong> the c<strong>as</strong>es, the deepest lake part is the depocentre for the underflows, possible<br />
sills between sub-b<strong>as</strong>ins may prevent the underflows reaching the lake’s centre.<br />
This needs to be taken into account with complex lake b<strong>as</strong>in morphologies.<br />
Furthermore, reflection seismic information from the subsurface is needed to<br />
show the stratigraphic architecture <strong>of</strong> the sedimentary b<strong>as</strong>in fill (see Fig. 1).<br />
Potential unconformities related to large m<strong>as</strong>s movements, bottom currents, or<br />
g<strong>as</strong> escape features may seriously hamper the continuous nature <strong>of</strong> the sediment<br />
succession. Only a detailed seismic stratigraphy allows to select ideal coring<br />
locations, define sediment thickness <strong>and</strong> predict sediment types.<br />
2. Geomorphic indications around the lake<br />
As a consequence <strong>of</strong> the high detrital influx to a lake during flood events, the<br />
following geomorphic indications around the lake should be present.<br />
(a) A certain relief in the catchment area is necessary in order to erode material<br />
that can be transported into the lake.<br />
(b) The lake should have evident inflows, however, they should ideally only be<br />
activated in the c<strong>as</strong>e <strong>of</strong> an extreme event. That provides a certain threshold<br />
with an ‘on/<strong>of</strong>f signal’ so that only the larger ‘extreme’ events are recorded<br />
in the lacustrine archive.<br />
(c) A delta structure should be present at the major inflows <strong>of</strong> a lake <strong>as</strong>suring a<br />
persistent detrital influx in the p<strong>as</strong>t.<br />
The following criterion deals with the recognition <strong>of</strong> the flood layers in the sedimentary<br />
record. This can <strong>of</strong>ten only be evaluated after preliminary investigations<br />
using short sediment cores.<br />
3. Contr<strong>as</strong>t between flood deposits <strong>and</strong> regular background sedimentation<br />
The lithology <strong>of</strong> the flood deposits should contr<strong>as</strong>t to the regular background<br />
sediments, so that these events can be lithologically <strong>and</strong> geochemically recognized.<br />
This implies that the regular background sedimentation should not
230 A. Gilli et al.<br />
Fig. 3 Flood layers in a core section (core LL081-A3a) <strong>of</strong> <strong>Lake</strong> Ledro, Italy, with relative<br />
concentration <strong>of</strong> silicium (Si) <strong>and</strong> iron (Fe) me<strong>as</strong>ured continuously by an XRF core scanner<br />
(me<strong>as</strong>uring interval: 0.2 mm). Dark-coloured flood layers st<strong>and</strong> out prominently in the finelylaminated<br />
calcareous background sediment (note the lightness difference reported in reflectance<br />
L * ). Flood layers (grey bars) are characterized by incre<strong>as</strong>ed Si <strong>and</strong> Fe content, especially<br />
pronounced in the fine-grained top part <strong>of</strong> the flood layer (‘clay cap’) marked in light grey bars<br />
consist solely <strong>of</strong> detrital material. If organic materials dominate the background<br />
sediments, the detrital layers are in general lighter than the otherwise very dark<br />
background sediments (e.g. <strong>Lake</strong> Schwendi, Gilli et al. 2003). But if the debris<br />
background sediments manly consist <strong>of</strong> authigenic carbonates, the detrital layers<br />
are considerably darker than the normal sediment (e.g. <strong>Lake</strong> Ledro, see Fig. 3).<br />
In either setting, detrital layers st<strong>and</strong> out prominently when compared to the<br />
background sediments <strong>and</strong> can e<strong>as</strong>ily be identified.<br />
Beside the described criteria for selecting a lake system to record recognizable<br />
flood layers, more technical <strong>as</strong>pects to recover the expected flood record are <strong>as</strong> well<br />
<strong>of</strong> importance for a successful study.
<strong>Lake</strong> <strong>Sediments</strong> <strong>as</strong> <strong>Archives</strong> <strong>of</strong> <strong>Recurrence</strong> <strong>Rates</strong> <strong>and</strong> Intensities <strong>of</strong> P<strong>as</strong>t Flood Events 231<br />
4. Coring ability to recover desired time interval<br />
The sedimentation rate <strong>of</strong> a lacustrine archive can vary strongly between the<br />
different lake systems. In particular, the contribution <strong>of</strong> the general detrital influx<br />
to the total sedimentation can make a large difference to total sediment thickness.<br />
St<strong>and</strong>ard coring techniques with various piston-coring methods are capable <strong>of</strong><br />
reaching 20 m sub lake-floor depth, so that sediment thickness for the desired<br />
time interval should not exceed this range. Deeper core recovery can otherwise<br />
only be reached by complex <strong>and</strong> expensive commercial drilling operations. If a<br />
complete Holocene section is the target, rather small lakes with little inflow are<br />
ideal (Gilli et al. 2003; Theiler 2003). Larger perialpine lakes (e.g. <strong>Lake</strong> Uri, <strong>Lake</strong><br />
Brienz, both Switzerl<strong>and</strong>) have substantially larger sedimentation rates reaching<br />
values <strong>of</strong> up to 10 m per 1,000 years (Girardclos et al. 2007; Siegenthaler <strong>and</strong><br />
Sturm 1991).<br />
3 The Signature <strong>of</strong> Flood Events in the Lacustrine Archive<br />
A key step in palae<strong>of</strong>lood studies using lake sediments is the secure identification <strong>of</strong><br />
detrital layers <strong>as</strong> the result <strong>of</strong> a flood event. As the sedimentation <strong>of</strong> each lake system<br />
is influenced by a large variety <strong>of</strong> different factors, we can only discuss general rules<br />
here, which clearly need to be adapted to each individual lacustrine system (Mulder<br />
<strong>and</strong> Chapron 2011). Ideally, the detrital layers should sharply contr<strong>as</strong>t in terms <strong>of</strong><br />
sedimentology <strong>and</strong> geochemical composition with the regular background sediment<br />
(Sletten et al. 2003). In these c<strong>as</strong>es, the detrital layer can be <strong>of</strong>ten identified by<br />
visual observation even at the millimetre scale, a process, which can be supported<br />
by image analysis (e.g. Rodbell et al. 1999)(Fig.3).<br />
The sedimentology <strong>of</strong> the flood layers is characterized by a near pure terrestrial<br />
mineralogical composition <strong>and</strong> consequently a low organic content. The underflowrelated<br />
depositional processes (see part “Depositional processes in lakes”) cause<br />
a fining-upward grain-size pattern with sometimes an inverse graded b<strong>as</strong>e from<br />
temporal waxing <strong>of</strong> the flood-related run<strong>of</strong>f. The top part <strong>of</strong> the detrital layer<br />
consists <strong>of</strong> the finest particles <strong>and</strong> h<strong>as</strong> <strong>of</strong>ten a distinctive light colour (Fig. 3),<br />
commonly referred <strong>as</strong> ‘clay cap’. Sometimes, several stacked graded intervals<br />
within the same flood layer point towards multiple water discharge peaks (Fig. 2). In<br />
rather rare c<strong>as</strong>es, the b<strong>as</strong>e <strong>of</strong> flood layers consists <strong>of</strong> aquatic macrophytes (Gilli et al.<br />
2003), which were entrained by the underflow. On the b<strong>as</strong>is <strong>of</strong> the characteristics <strong>of</strong><br />
these detrital layers, previous studies applied, next to visual observations, one or<br />
a combination <strong>of</strong> several proxies to identify such layers. The three most common<br />
applied proxies are (1) grain size, (2) total organic carbon (<strong>of</strong>ten expressed <strong>as</strong> loss<br />
<strong>of</strong> ignition, i.e. LOI), <strong>and</strong> (3) magnetic susceptibility. These proxies are interpreted<br />
<strong>as</strong> follows:<br />
1. Grain size: Grain-size me<strong>as</strong>urements are almost routinely used to confirm a<br />
fining-upward gradation overlying a coarser b<strong>as</strong>e (Fig. 2). Recently, Parris
232 A. Gilli et al.<br />
et al. (2010) proposed end-member modelling <strong>of</strong> the particle-size distribution<br />
<strong>and</strong> demonstrated that the coarsest end-member is a sensitive indicator for the<br />
b<strong>as</strong>e <strong>of</strong> detrital flood layers.<br />
2. Total organic carbon (TOC): Caused by the high minerogenic content <strong>of</strong> the<br />
flood-related material, the organic-carbon content is usually low in the flood<br />
layers. This proxy is successfully applicable only in rather eutrophic lakes, where<br />
the background sedimentation is high in organic material <strong>of</strong>fering sufficient<br />
contr<strong>as</strong>t to the detrital signature.<br />
3. Magnetic susceptibility: Depending on the catchment’s geology, the detrital<br />
layers are <strong>of</strong>ten enriched in para- <strong>and</strong> ferromagnetic minerals reflecting high<br />
magnetic susceptibility values (Chapron et al. 2005; Wolfe et al. 2006). More<br />
elaborated proxies like isothermal remanent magnetization (IRM) have also been<br />
applied to identify surface-soil-derived material in lacustrine sediments (Boe<br />
et al. 2006; Thorndycraft et al. 1998).<br />
Beside these widely applied proxies, various additional methods are proposed to<br />
better identify detrital layers in the lacustrine archive.<br />
• Elemental composition: Elements like Si, Al, Fe, K, <strong>and</strong> Ti are typical for detrital<br />
sediments <strong>and</strong> thus their concentration is elevated in the detrital layers (Revel-<br />
Roll<strong>and</strong> et al. 2005). Elemental concentration is determined either on discrete<br />
samples or by continuous non-destructive XRF core scanning techniques (Fig. 3)<br />
with resolutions down to the sub-millimetre scale.<br />
• Bulk density: Especially the coarser b<strong>as</strong>e <strong>of</strong> detrital layers is characterized by<br />
elevated bulk densities, which can be determined by continuous gamma-ray<br />
attenuation me<strong>as</strong>urements performed on a st<strong>and</strong>ard multi-sensor core logging<br />
device (Gilli et al. 2003).<br />
• Image analysis: Detrital layers can be detected <strong>and</strong> mapped with image analysis<br />
(Fig. 3) if there is a clear colour contr<strong>as</strong>t between detrital layer <strong>and</strong> background<br />
sediments.<br />
• Carbon/nitrogen ratios: Organic matter <strong>of</strong> lacustrine origin reveals a distinct<br />
lower C/N ratio (i.e. 4–10) compared to v<strong>as</strong>cular l<strong>and</strong> plants (>20) (Meyers <strong>and</strong><br />
Teranes 2001). Several studies showed an incre<strong>as</strong>e <strong>of</strong> the C/N ratio in the detrital<br />
flood layers (Brown et al. 2000; Osleger et al. 2009; Wolfe et al. 2006)<br />
• Stable carbon isotopes on organic material: As with the C/N ratio, carbon<br />
isotopes can be used to differentiate between aquatic or terrestrial origin <strong>of</strong> the<br />
organic matter (Bierman et al. 1997). In general, rather higher carbon isotopic<br />
ratios (• 13 C) are typical for terrestrial plants <strong>and</strong> found in the detrital layers<br />
(Osleger et al. 2009; Wolfe et al. 2006)<br />
• Pollen <strong>as</strong>semblages: Pollen spectra <strong>of</strong> detrital layers show a clear shift towards<br />
more robust pollen types, like Liguliflorae <strong>and</strong> Filicales pollens, <strong>as</strong> they remain<br />
in the catchment <strong>and</strong> are subsequently w<strong>as</strong>hed-in, when more vulnerable pollen<br />
types are already corroded <strong>and</strong> decayed (Thorndycraft et al. 1998).<br />
• Sr <strong>and</strong> Nd isotopes: If the lake’s catchment consists <strong>of</strong> different geologic<br />
terrains, strontium <strong>and</strong> neodymium isotopes <strong>of</strong> the detrital particles st<strong>and</strong> out<br />
prominently when compared to the values from background sediment (Revel-<br />
Roll<strong>and</strong> et al. 2005).
<strong>Lake</strong> <strong>Sediments</strong> <strong>as</strong> <strong>Archives</strong> <strong>of</strong> <strong>Recurrence</strong> <strong>Rates</strong> <strong>and</strong> Intensities <strong>of</strong> P<strong>as</strong>t Flood Events 233<br />
The clear identification <strong>and</strong> attribution <strong>of</strong> a detrital layer <strong>as</strong> a p<strong>as</strong>t flood event<br />
can sometimes be challenging <strong>as</strong> similar layers could also be the result <strong>of</strong> m<strong>as</strong>s<br />
movements (i.e. slope failures) triggered by earthquakes (Chapron et al. 1999;<br />
Schnellmann et al. 2002; Str<strong>as</strong>ser et al. 2006), delta collapse (Girardclos et al.<br />
2007) or lake level fluctuations (Anselmetti et al. 2009). But these m<strong>as</strong>s-movement<br />
related turbidite layers have different structural <strong>and</strong> compositional characteristics<br />
<strong>and</strong> they are – in general – rather thicker than flood layers (Siegenthaler <strong>and</strong> Sturm<br />
1991; Sturm et al. 1995). M<strong>as</strong>s-movement related turbidite layers show only a thin<br />
fining-upward sequence followed by a thick homogenous interval with no inverse<br />
grading at the b<strong>as</strong>e. With the exception <strong>of</strong> turbidites triggered by delta collapses,<br />
the m<strong>as</strong>s-movement layers contain reworked lacustrine background sediment with a<br />
geochemical composition different from the catchment-derived detrital signal.<br />
Detailed sedimentological analysis <strong>of</strong> detrital layers is further <strong>of</strong> importance to<br />
scale intensities <strong>of</strong> p<strong>as</strong>t flood events. Assessing p<strong>as</strong>t flood events intensity is always<br />
b<strong>as</strong>ed on a constant availability <strong>of</strong> sediment during the observed period. The amount<br />
<strong>of</strong> sediment, however, that is mobilized by intense precipitation <strong>and</strong> run<strong>of</strong>f may<br />
be also dependent on anthropogenic activity, <strong>as</strong> various forms <strong>of</strong> l<strong>and</strong>-use expose<br />
the soils <strong>and</strong> incre<strong>as</strong>e their sensitivity to become eroded. Nevertheless, a positive<br />
relationship between suspended particle concentration <strong>and</strong> river discharge (Mulder<br />
et al. 2003) enables a scaling <strong>of</strong> the flood event by the detrital layer thickness. This<br />
relationship is specific for each river b<strong>as</strong>in, but follows a power-law function making<br />
the sedimentary record very sensitive to variations in the river discharge <strong>of</strong> extreme<br />
floods (Mulder et al. 2003). So far, only a few studies used this relationship to<br />
<strong>as</strong>sess the flood magnitude <strong>of</strong> p<strong>as</strong>t events (e.g. Brown et al. 2000; Bussmann <strong>and</strong><br />
Anselmetti 2010; Irmler et al. 2006). These studies always use the imprint <strong>of</strong> a<br />
documented historic flood event in the sedimentary archive to relatively scale prehistoric<br />
flood events. A proxy even less applied in evaluating p<strong>as</strong>t flood events in the<br />
sedimentary archive is grain-size distribution. Grain-size me<strong>as</strong>urements are <strong>of</strong>ten<br />
carried out to identify detrital layers, but <strong>as</strong> the grain-size distribution also reflects<br />
hydrologic conditions, in particular transport capacity, during a flood event, it could<br />
be applied to <strong>as</strong>sess the intensity <strong>of</strong> p<strong>as</strong>t floods.<br />
4 Dating Methods for <strong>Lake</strong> <strong>Sediments</strong><br />
Once the flood layers have been identified, age dating <strong>of</strong> the sedimentary succession<br />
is required to establish a flood chronology. This can be achieved applying various<br />
dating techniques, which eventually result in a ‘best-fit’ age model for the recovered<br />
stratigraphic section. For this purpose, one h<strong>as</strong> to consider that a sedimentary<br />
sequence consists <strong>of</strong> background sediments, characterized by low sedimentation<br />
rate, which is punctuated or ‘interrupted’ by individual event-layers that were<br />
deposited in short time (i.e. hours to days) thus resulting in a high sedimentation<br />
rate. Consequently, in order to interpolate between age markers, those rapidly<br />
deposited layers need to be removed from the sedimentary succession before the
234 A. Gilli et al.<br />
age model is defined. The resulting interpolation <strong>as</strong>sumes a stable background<br />
sedimentation rate between age markers, which is <strong>of</strong>ten the best-possible approach.<br />
The uppermost part <strong>of</strong> the sediments that were deposited in the l<strong>as</strong>t century can<br />
be dated me<strong>as</strong>uring the gamma-activity <strong>of</strong> radionuclides, in particular 210 Pb <strong>and</strong><br />
137 Cs (Appleby 2001, <strong>and</strong> references therein). 137 Cs emissions into the environment<br />
started with the onset <strong>of</strong> atmospheric nuclear bomb testing in the mid-1950s<br />
culminating in 1963. A second maximum w<strong>as</strong> produced upon the accident <strong>of</strong><br />
Chernobyl in 1986, <strong>of</strong>fering thus two marker horizons that can be detected by<br />
gamma spectroscopy. 210 Pb-activities, in contr<strong>as</strong>t, are a product <strong>of</strong> in situ produced<br />
‘supported’ <strong>as</strong> well <strong>of</strong> ‘unsupported’ sources from a decay chain, the latter <strong>of</strong> which<br />
allows calculation <strong>of</strong> sedimentation rates for the l<strong>as</strong>t 150 years <strong>as</strong>suming various<br />
sedimentation models (Appleby 2001; von Gunten et al. 2009). Furthermore, the<br />
occurrence <strong>of</strong> combustion products, such <strong>as</strong> spherical carbonaceous particles (SPE)<br />
or various other combustion-derived aerosols reflect local to regional use <strong>of</strong> burning<br />
agents <strong>and</strong> can be compared to anthropogenic <strong>and</strong> industrial history (Thevenon <strong>and</strong><br />
Anselmetti 2007; von Gunten et al. 2009), providing additional dating possibilities<br />
for the most recent section.<br />
Radiocarbon dating <strong>of</strong> terrestrial organic matter is probably the most common<br />
<strong>and</strong> powerful dating tool that is used with lacustrine sediments. The success <strong>of</strong><br />
this method relies on identification <strong>of</strong> organic macro-remains <strong>of</strong> terrestrial origin,<br />
<strong>as</strong> aquatically produced biom<strong>as</strong>s <strong>of</strong>ten incorporates ‘old’ carbon, which is not in<br />
equilibrium with the atmosphere (Deevey et al. 1954). This radiocarbon reservoir<br />
effect causes an <strong>of</strong>fset towards too old radiocarbon ages, which for this re<strong>as</strong>on<br />
prevents building a robust age model.<br />
Further possibilities for providing age markers are the occurrences <strong>of</strong> volcanic<br />
tephra layers <strong>of</strong> known age, which can be tracked over large distances. In Swiss<br />
lakes, the Laacher See Tephra (originating from the Eifel area in Germany) <strong>and</strong><br />
the V<strong>as</strong>set-Killian Tephra from an eruption in the M<strong>as</strong>sif Central in France (Hajd<strong>as</strong><br />
et al. 1993) can <strong>of</strong>ten be identified <strong>and</strong> may refine a radiocarbon-b<strong>as</strong>ed age model<br />
(Schnellmann et al. 2006).<br />
Annually laminated background sediments (i.e. varves, see Zolitschka 2006 for<br />
an overview) provide probably the most accurate dating by layer counting, <strong>as</strong> annual<br />
resolution cannot be obtained by the other methods. However, <strong>as</strong> varves rarely occur<br />
throughout the entire section, varve chronology may be ‘floating’ <strong>and</strong> therefore be<br />
used together with other independently determined age markers providing thus a<br />
relative rather than an absolute chronology <strong>of</strong> annual resolution.<br />
Eventually, the age model may be adjusted by making reference to prominent<br />
historic event markers <strong>as</strong> earthquakes, rock falls or large floods leave their unique<br />
fingerprint behind. Several studies (e.g. Bussmann <strong>and</strong> Anselmetti 2010; Gilli<br />
et al. 2003; Siegenthaler <strong>and</strong> Sturm 1991) linked detrital layers within the 137 Cs<br />
<strong>and</strong> 210 Pb-dated interval to documented extreme flood events producing higher<br />
accuracies <strong>of</strong> the independently dated succession. Overall, defining an age model is<br />
usually a combined application <strong>of</strong> various methods described above, which together<br />
maximize range, resolution, <strong>and</strong> accuracy <strong>of</strong> the ages that are then <strong>as</strong>signed to each<br />
individual flood layer.
<strong>Lake</strong> <strong>Sediments</strong> <strong>as</strong> <strong>Archives</strong> <strong>of</strong> <strong>Recurrence</strong> <strong>Rates</strong> <strong>and</strong> Intensities <strong>of</strong> P<strong>as</strong>t Flood Events 235<br />
5 Determining the Se<strong>as</strong>onality <strong>of</strong> Flood Events<br />
Flood reconstructions mainly focus on the frequency <strong>and</strong> intensity <strong>of</strong> p<strong>as</strong>t<br />
heavy-precipitation events. This allows the identification <strong>of</strong> p<strong>as</strong>t time periods<br />
<strong>of</strong> incre<strong>as</strong>ed flood activity with implications for general atmospheric circulation<br />
patterns. However, knowledge <strong>of</strong> p<strong>as</strong>t atmospheric processes leading to severe<br />
floods would substantially be improved if the individual flood layers could be<br />
se<strong>as</strong>onally resolved. With such high-resolution records, conclusions about – for<br />
example – a change in frequency <strong>of</strong> floods caused by summer thunderstorms could<br />
be drawn.<br />
Varved sediments allow a precise chronology <strong>of</strong> the lacustrine record down to<br />
the se<strong>as</strong>on b<strong>as</strong>ed on the alternating deposition <strong>of</strong> distinct laminae during the annual<br />
cycle (Zolitschka 2006). The position <strong>of</strong> detrital flood layers within the se<strong>as</strong>onal<br />
alternating laminae <strong>of</strong> a varve may be used to date the se<strong>as</strong>on they were deposited.<br />
Depending on the thickness <strong>of</strong> the varves, the position <strong>of</strong> the flood layer within a<br />
varve is visually determined either on thin sections, on high-resolution core images,<br />
or by the naked eye. To apply the concept <strong>of</strong> se<strong>as</strong>onal flood layer dating, a possible<br />
erosion <strong>of</strong> the underlying laminae needs to be considered. Thin section analysis is<br />
used to identify an erosional b<strong>as</strong>e <strong>of</strong> the detrital layer, but a safe se<strong>as</strong>onal <strong>as</strong>signment<br />
<strong>of</strong> the flood layer can be achieved by determining the se<strong>as</strong>on <strong>of</strong> the next overlying<br />
varve lamina.<br />
Unfortunately, continuously varved lacustrine sections over longer time intervals<br />
are quite rare <strong>and</strong> if present, the lake setting is <strong>of</strong>ten not sensitive to flood events.<br />
For this re<strong>as</strong>on only very few se<strong>as</strong>onal resolved flood records are documented in the<br />
literature (Lamoureux 2000; Mangili et al. 2005). With the above outlined large gain<br />
<strong>of</strong> knowledge about flood-triggering atmospheric processes, the search for lacustrine<br />
flood records in varved lacustrine sections <strong>and</strong> its detailed, but time-consuming<br />
analysis, is highly desirable.<br />
6 Occurrence <strong>of</strong> Flood Events in the P<strong>as</strong>t – C<strong>as</strong>e Studies<br />
The following two studies illustrate the application <strong>of</strong> lake sediments <strong>as</strong> archive for<br />
flood events.<br />
6.1 Flood History <strong>of</strong> <strong>Lake</strong> Lauerz, Central Switzerl<strong>and</strong><br />
<strong>Lake</strong> Lauerz is a small perialpine lake (surface area: 3 km 2 , catchment area:<br />
69 km 2 ) located in Central Switzerl<strong>and</strong> that records heavy precipitation events<br />
with typical flood layers. Piston cores were retrieved in the deepest part <strong>of</strong> the<br />
main lake b<strong>as</strong>in allowing a composite record <strong>of</strong> nearly 10 m length (Bussmann
236 A. Gilli et al.<br />
Fig. 4 Chronology <strong>of</strong> flood layers determined from a sediment core from <strong>Lake</strong> Lauerz (Switzerl<strong>and</strong>)<br />
covering the l<strong>as</strong>t 2,000 years (Bussmann <strong>and</strong> Anselmetti 2010). Intervals <strong>of</strong> frequent flood<br />
events are shaded <strong>and</strong> labelled with even roman numbers. The layer thickness <strong>of</strong> the flood<br />
event from 1934 AD allows a scaling <strong>of</strong> the storm magnitude <strong>of</strong> p<strong>as</strong>t events. Six flood events<br />
(marked with stars) were <strong>as</strong> severe <strong>as</strong> the historically documented flood event from 1876, which<br />
caused large damages in Central Switzerl<strong>and</strong> (reprinted by permission from the Swiss Geological<br />
Society, © 2010)<br />
<strong>and</strong> Anselmetti 2010). The sediments were dated with four radiocarbon ages to<br />
establish a core chronology comprising the l<strong>as</strong>t 2,000 years (Fig. 4). A high bulk<br />
sedimentation-rate <strong>of</strong> up to 1 cm year 1 provides a unique high temporal resolution<br />
<strong>of</strong> palae<strong>of</strong>loods in <strong>Lake</strong> Lauerz. 54 single flood events with a layer thickness<br />
between 0.5 <strong>and</strong> 9 cm were identified. Interestingly, the flood events are not evenly<br />
distributed in the p<strong>as</strong>t 2,000 years, but rather cluster in three periods (Fig. 4). An<br />
incre<strong>as</strong>ed occurrence <strong>of</strong> heavy precipitation events is clearly documented between<br />
1940–1630 AD, 1420–990 AD <strong>and</strong> 850–580 AD in the palae<strong>of</strong>lood record <strong>of</strong> <strong>Lake</strong><br />
Lauerz (Bussmann <strong>and</strong> Anselmetti 2010).<br />
Under the <strong>as</strong>sumption that layer thickness is a proxy for storm magnitude (Brown<br />
et al. 2000), the flood layers detected during the instrumental period in the twentieth<br />
century were used to scale p<strong>as</strong>t flood events. The largest flood during the period<br />
<strong>of</strong> precipitation me<strong>as</strong>urements occurred in 1934 AD. This event caused severe<br />
dev<strong>as</strong>tation around <strong>Lake</strong> Lauerz <strong>and</strong> w<strong>as</strong> recorded with a 2.5 cm thick detrital layer<br />
in the lake b<strong>as</strong>in. On the b<strong>as</strong>is <strong>of</strong> the established 2,000-year palae<strong>of</strong>lood record, 17<br />
<strong>of</strong> the 54 detected events were recorded with an even larger layer thickness <strong>and</strong> are<br />
thus interpreted to reflect a larger flood intensity than the 1934 AD event. One <strong>of</strong><br />
the most severe flood events in Central Switzerl<strong>and</strong> during historic times, which<br />
is recorded with layer thickness <strong>of</strong> 4.5 cm in <strong>Lake</strong> Lauerz, occurred in 1876 AD.<br />
Considering the event layer thicknesses, five earlier flood events had a similar or<br />
even larger storm magnitude pointing towards the occurrence <strong>of</strong> severe run<strong>of</strong>f events<br />
in the p<strong>as</strong>t 2,000 years (Bussmann <strong>and</strong> Anselmetti 2010).
<strong>Lake</strong> <strong>Sediments</strong> <strong>as</strong> <strong>Archives</strong> <strong>of</strong> <strong>Recurrence</strong> <strong>Rates</strong> <strong>and</strong> Intensities <strong>of</strong> P<strong>as</strong>t Flood Events 237<br />
Fig. 5 Holocene palae<strong>of</strong>lood records from northe<strong>as</strong>tern United States (Noren et al. 2002). (a)<br />
Palae<strong>of</strong>lood records from 13 individual lakes. (b) Weighted compilation <strong>of</strong> all flood records<br />
arranged in 100-year bins. Linear incre<strong>as</strong>ing trend <strong>of</strong> the storminess toward modern is explained<br />
by the authors by delta progradation leading to an incre<strong>as</strong>ed sensitivity <strong>of</strong> the coring site<br />
to record flood events. (c) GISP2 non-sea-salt (n.s.s.) K concentration with values above the<br />
superimposed linear regression shaded (reprinted by permission from Macmillan Publishers Ltd:<br />
Noren et al. 2002)<br />
6.2 Flood History from <strong>Lake</strong>s in Northe<strong>as</strong>tern United States<br />
Noren et al. (2002) conducted a comprehensive study about millennial-scale<br />
periodicity <strong>of</strong> exceptional rainfall events in the northe<strong>as</strong>tern United States covering<br />
the entire Holocene. Unique in the design <strong>of</strong> this study is the stacking <strong>of</strong> the<br />
chronological occurrence <strong>of</strong> detrital layers from 13 individual lakes (Fig. 5a, b). This
238 A. Gilli et al.<br />
multi-archive approach is less susceptible to the influence <strong>of</strong> only locally occurring<br />
events. The results have therefore more paleoclimatic significance when compared<br />
to studies dealing solely with a single lake b<strong>as</strong>in.<br />
<strong>Lake</strong>s investigated within this study fulfil clearly the criteria for promising<br />
lacustrine flood reconstruction (cf. section “Selecting lakes with promising flood<br />
records”) with well-defined morphobathymetric depocentres, a steep surrounding<br />
relief, evident inflows <strong>and</strong> delta structures, <strong>as</strong> well <strong>as</strong> predominantly organic background<br />
sedimentation contr<strong>as</strong>ting with the detrital flood layers. The identification<br />
<strong>of</strong> flood layers w<strong>as</strong> achieved by visual logging <strong>and</strong> high-resolution X-radiography<br />
(providing information on internal sediment structures), <strong>as</strong> well <strong>as</strong> magnetic susceptibility,<br />
loss-on-ignition <strong>and</strong> grain-size me<strong>as</strong>urements.<br />
Spectral analysis <strong>and</strong> correlations with proxies from the Greenl<strong>and</strong> Ice Sheet<br />
Project 2 (GISP2) ice-core record could connect the established flood record with<br />
large-scale atmospheric circulation patterns (Fig. 5b, c). Periods with high flood<br />
activity show a recurrence interval <strong>of</strong> about 3,000 years <strong>and</strong> correlate in the<br />
GISP2 record with periods characterized by higher concentrations <strong>of</strong> non-sea-salt<br />
pot<strong>as</strong>sium (n.s.s. K). Incre<strong>as</strong>ed concentrations <strong>of</strong> these aerosols are attributed to<br />
enhanced storminess in the high-latitude North-Atlantic region. Hence, the results<br />
indicate that precipitation intensity in the northe<strong>as</strong>tern United States is controlled<br />
by large-scale atmospheric circulation modes such <strong>as</strong> the Arctic Oscillation (AO)<br />
(Noren et al. 2002).<br />
7 Conclusions<br />
<strong>Lake</strong>s have a great potential to record flood events by the deposition <strong>of</strong> detrital layers<br />
through turbiditic underflows. The major advantages <strong>of</strong> the lacustrine approach in<br />
palae<strong>of</strong>lood reconstructions compared to analysis <strong>of</strong> alluvial fans <strong>and</strong> cones are<br />
the continuous mode <strong>of</strong> sedimentation, the high temporal resolution <strong>and</strong> the long<br />
time period covered by lake sediments. These advantages led to an incre<strong>as</strong>ing<br />
number <strong>of</strong> studies exploring the lacustrine archives for palae<strong>of</strong>lood reconstructions.<br />
The results <strong>of</strong> these studies are very promising <strong>and</strong> clearly indicate periods <strong>of</strong><br />
different flood frequencies for the Holocene. However, these studies cover almost<br />
exclusively a single catchment <strong>and</strong> therefore may include local events. Compilations<br />
<strong>of</strong> different lacustrine palae<strong>of</strong>lood records, <strong>as</strong> for example applied by Noren et al.<br />
(2002), overcome this limitation <strong>and</strong> are able to build regionally robust flood reconstructions.<br />
With the now available limnogeologic data sets <strong>of</strong> p<strong>as</strong>t floods, linkages<br />
with other scientific communities like geomorphologists <strong>and</strong> dendrochronologists,<br />
need to be established to compile palae<strong>of</strong>lood records from different archives. This<br />
combined approach (<strong>as</strong> outlined by Irmler et al. 2006) will be <strong>of</strong> great importance to<br />
underst<strong>and</strong> the sensitivity <strong>of</strong> each flood archive <strong>and</strong> will ultimately lead to the most<br />
complete flood reconstructions in terms <strong>of</strong> frequency <strong>and</strong> intensity.
<strong>Lake</strong> <strong>Sediments</strong> <strong>as</strong> <strong>Archives</strong> <strong>of</strong> <strong>Recurrence</strong> <strong>Rates</strong> <strong>and</strong> Intensities <strong>of</strong> P<strong>as</strong>t Flood Events 239<br />
Acknowledgements The authors thank the Swiss National Science Foundation (SNF) for the<br />
financial support within the project ‘FloodAlp’ (Project No. 200021–121909) to initiate research<br />
on the Holocene flood history in Switzerl<strong>and</strong> <strong>and</strong> northern Italy using lake sediments. Figures 1<br />
<strong>and</strong> 3 contains scientific results form a joint project on <strong>Lake</strong> Ledro, Italy whose co-funding by<br />
the French ANR (project LAMA, directed by M. Magny <strong>and</strong> N. Combourieu-Nebout) is kindly<br />
acknowledged. The cores from <strong>Lake</strong> Thun shown in Fig. 2 were collected <strong>as</strong> part <strong>of</strong> a project in<br />
cooperation with Stephanie Girardclos, University <strong>of</strong> Geneva.<br />
References<br />
Anselmetti FS, Ariztegui D, De Batist M, Gebhardt AC, Haberzettl T, Niessen F, Ohlendorf C,<br />
Zolitschka B (2009) Environmental history <strong>of</strong> southern Patagonia unraveled by the seismic<br />
stratigraphy <strong>of</strong> Laguna Potrok Aike. Sedimentology 56:873–892<br />
Appleby PG (2001) Chronostratigraphic techniques in recent sediments. In: L<strong>as</strong>t WM, Smol JP<br />
(eds) Tracking environmental change using lake sediments, vol 1, B<strong>as</strong>in analysis, coring <strong>and</strong><br />
chronological techniques. Kluwer Academic, Dordrecht, pp 171–201<br />
Arnaud F, Revel M, Chapron E, Desmet M, Tribovillard N (2005) 7200 years <strong>of</strong> Rhone river<br />
flooding activity in <strong>Lake</strong> Le Bourget, France: a high-resolution sediment record <strong>of</strong> NW Alps<br />
hydrology. The Holocene 15:420–428<br />
Bardou E, Fournier F, Sartori M (2003) Pale<strong>of</strong>lood reconstruction at Illgraben torrent (Switzerl<strong>and</strong>):<br />
a current need for event frequency estimation. In: Thorndycraft VR, Benito G,<br />
Barriendos M, Ll<strong>as</strong>at C (eds) Palae<strong>of</strong>loods, historical floods <strong>and</strong> climate variability: applications<br />
in flood risk <strong>as</strong>sessment. Proceedings <strong>of</strong> the PHEFRA international workshop, Barcelona,<br />
pp 53–59, 16–19 October 2002<br />
Benito G, Thorndycraft VR (2005) Palae<strong>of</strong>lood hydrology <strong>and</strong> its role in applied hydrological<br />
sciences. J Hydrol 313:3–15<br />
Bierman P, Lini A, Zehfuss P, Church A, Davis PT, Southon J, Baldwin L (1997) Postglacial ponds<br />
<strong>and</strong> alluvial fans: recorders <strong>of</strong> Holocene l<strong>and</strong>scape history. GSA Today 7:1–8<br />
Boe AG, Dahl SO, Lie O, Nesje A (2006) Holocene river floods in the upper Glomma<br />
catchment, southern Norway: a high-resolution multiproxy record from lacustrine sediments.<br />
The Holocene 16:445–455<br />
Brown SL, Bierman PR, Lini A, Southon J (2000) 10,000 yr record <strong>of</strong> extreme hydrologic events.<br />
Geology 28:335–338<br />
Bussmann F, Anselmetti FS (2010) Rossberg l<strong>and</strong>slide history <strong>and</strong> flood chronology <strong>as</strong> recorded<br />
in <strong>Lake</strong> Lauerz sediments (Central Switzerl<strong>and</strong>). Swiss J Geosci 103:43–59<br />
Chapron E, Beck C, Pourchet M, Deconinck J-F (1999) 1822 earthquake-triggered homogenite in<br />
<strong>Lake</strong> Le Bourget (NW Alps). Terra Nova 11:86–92<br />
Chapron E, Desmet M, De Putter T, Loutre MF, Beck C, Deconinck JF (2002) Climatic variability<br />
in the northwestern Alps, France, <strong>as</strong> evidenced by 600 years <strong>of</strong> terrigenous sedimentation in<br />
<strong>Lake</strong> Le Bourget. The Holocene 12:177–185<br />
Chapron E, Arnaud F, Noel H, Revel M, Desmet M, Perdereau L (2005) Rhone River flood deposits<br />
in <strong>Lake</strong> Le Bourget: a proxy for Holocene environmental changes in the NW Alps, France.<br />
Bore<strong>as</strong> 34:404–416<br />
Dapples F, Lotter AF, van Leeuwen JFN, van der Knaap WO, Dimitriadis S, Oswald D (2002)<br />
Paleolimnological evidence for incre<strong>as</strong>ed l<strong>and</strong>slide activity due to forest clearing <strong>and</strong> l<strong>and</strong>-use<br />
since 3600 cal BP in the western Swiss Alps. J Paleolimnol 27:239–248<br />
Deevey ES, Gross MS, Hutchinson GE, Kraybill HL (1954) The natural C 14 content <strong>of</strong> materials<br />
from hard-water lakes. Proc Natl Acad Sci USA 40:285–288<br />
Eden DN, Page MJ (1998) Palaeoclimatic implications <strong>of</strong> a storm erosion record from late<br />
Holocene lake sediments, North Isl<strong>and</strong>, New Zeal<strong>and</strong>. Palaeogeogr Palaeoclimatol Palaeoecol<br />
139:37–58
240 A. Gilli et al.<br />
Ely LL, Enzel Y, Baker VR, Cayan DR (1993) A 5000-year record <strong>of</strong> extreme flood <strong>and</strong> climatechange<br />
in the southwestern United States. Science 262:410–412<br />
Gilli A, Anselmetti FS, Ariztegui D, McKenzie JA (2003) A 600-year sedimentary record <strong>of</strong> flood<br />
events from two sub-alpine lakes (Schwendiseen, Northe<strong>as</strong>tern Switzerl<strong>and</strong>). Eclogae Geol<br />
Helv 96(Supplement 1):S49–S58<br />
Giovanoli F (1990) Horizontal transport <strong>and</strong> sedimentation by interflows <strong>and</strong> turbidity currents in<br />
<strong>Lake</strong> Geneva. In: Tilzer MM, Serruya C (eds) Large lakes: ecological structure <strong>and</strong> function.<br />
Springer, Berlin/Heidelberg, pp 175–195<br />
Girardclos S, Schmidt OT, Sturm M, Ariztegui D, Pugin A, Anselmetti FS (2007) The 1996 AD<br />
delta collapse <strong>and</strong> large turbidite in <strong>Lake</strong> Brienz. Mar Geol 241:137–154<br />
Hajd<strong>as</strong> I, Ivy SD, Beer J, Bonani G, Imboden D, Lotter AF, Sturm M, Suter M (1993) AMS<br />
radiocarbon dating <strong>and</strong> varve chronology <strong>of</strong> <strong>Lake</strong> Soppensee – 6000 to 12000 14 C years BP.<br />
Clim Dyn 9:107–116<br />
IrmlerR,DautG,Mäusbacher R (2006) A debris flow calendar derived from sediments <strong>of</strong> lake<br />
Lago di Braies (N. Italy). Geomorphology 77:69–78<br />
Kochel RC, Baker VR (1982) Pale<strong>of</strong>lood hydrology. Science 215:353–361<br />
Lamb MP, Mohrig D (2009) Do hyperpycnal-flow deposits record river-flood dynamics? Geology<br />
37:1067–1070<br />
Lambert A, Hsü KJ (1979) Non-annual cycles <strong>of</strong> varve-like sedimentation in Walensee, Switzerl<strong>and</strong>.<br />
Sedimentology 26:453–461<br />
Lamoureux S (2000) Five centuries <strong>of</strong> interannual sediment yield <strong>and</strong> rainfall-induced erosion in<br />
the Canadian High Arctic recorded in lacustrine varves. Water Resour Res 36:309–318<br />
Macklin MG, Lewin J (2003) River sediments, great floods <strong>and</strong> centennial-scale Holocene climate<br />
change. J Quaternary Sci 18:101–105<br />
Mangili C, Brauer A, Moscariello A, Naumann R (2005) Micr<strong>of</strong>acies <strong>of</strong> detrital event layers<br />
deposited in Quaternary varved lake sediments <strong>of</strong> the Piànico-Sèllere B<strong>as</strong>in (northern Italy).<br />
Sedimentology 52:927–943<br />
Mazzucchi D, Spooner IS, Gilbert R, Osborn G (2003) Reconstruction <strong>of</strong> Holocene climate change<br />
using multiproxy analysis <strong>of</strong> sediments from Pyramid <strong>Lake</strong>, British Columbia, Canada. Arct<br />
Antarct Alp Res 35:520–529<br />
Meyers PA, Teranes JL (2001) Sediment organic matter. In: L<strong>as</strong>t WM, Smol JP (eds) Tracking<br />
environmental change using lake sediments, vol 2, Physical <strong>and</strong> geochemical methods. Kluwer<br />
Academic, Dordrecht, pp 239–269<br />
Mulder T, Chapron E (2011) Flood deposits in continental <strong>and</strong> marine environments: character<br />
<strong>and</strong> significance. In: Slatt RM, Zavala C (eds) Sediment transfer from shelf to deep water -<br />
Revisiting the delivery system: AAPG Studies in Geology 61:1–30<br />
Mulder T, Migeon S, Savoye B, Faugères J-C (2001) Inversely graded turbidite sequences in the<br />
deep Mediterranean: a record <strong>of</strong> deposits from flood-generated turbidity currents? Geo-Mar<br />
Lett 21:86–93<br />
Mulder T, Syvitski JPM, Migeon S, Faugères J-C, Savoye B (2003) Marine hyperpycnal flows:<br />
initiation, behavior <strong>and</strong> related deposits. A review. Mar Pet Geol 20:861–882<br />
Nesje A, Dahl SO, Matthews JA, Berrisford MS (2001) A 4500-yr record <strong>of</strong> river floods obtained<br />
from a sediment core in <strong>Lake</strong> Atnsjoen, e<strong>as</strong>tern Norway. J Paleolimnol 25:329–342<br />
Noren AJ, Bierman PR, Steig EJ, Lini A, Southon J (2002) Millennial-scale storminess variability<br />
in the northe<strong>as</strong>tern United States during the Holocene epoch. Nature 419:821–824<br />
O’Connor JE, Ely LL, Wohl EE, Stevens LE, Melis TS, Kale VS, Baker VR (1994) A 4500-year<br />
record <strong>of</strong> large floods on the Colorado River in the Gr<strong>and</strong>-Canyon, Arizona. J Geol 102:1–9<br />
Osleger DA, Heyvaert AC, Stoner JS, Verosub KL (2009) Lacustrine turbidites <strong>as</strong> indicators<br />
<strong>of</strong> Holocene storminess <strong>and</strong> climate: <strong>Lake</strong> Tahoe, California <strong>and</strong> Nevada. J Paleolimnol<br />
42:103–122<br />
Parris AS, Bierman PR, Noren AJ, Prins MA, Lini A (2010) Holocene paleostorms identified by<br />
particle size signatures in lake sediments from the northe<strong>as</strong>tern United States. J Paleolimnol<br />
43:29–49
<strong>Lake</strong> <strong>Sediments</strong> <strong>as</strong> <strong>Archives</strong> <strong>of</strong> <strong>Recurrence</strong> <strong>Rates</strong> <strong>and</strong> Intensities <strong>of</strong> P<strong>as</strong>t Flood Events 241<br />
Pfister C (2009) The “Dis<strong>as</strong>ter Gap” <strong>of</strong> the 20th century <strong>and</strong> the loss <strong>of</strong> traditional dis<strong>as</strong>ter memory.<br />
Gaia 18:239–246<br />
Revel-Roll<strong>and</strong> M, Arnaud F, Chapron E, Desmet M, Givelet N, Alibert C, McCulloch M (2005) Sr<br />
<strong>and</strong> Nd isotopes <strong>as</strong> tracers <strong>of</strong> cl<strong>as</strong>tic sources in <strong>Lake</strong> Le Bourget sediment (NW Alps, France)<br />
during the Little Ice Age: palaeohydrology implications. Chem Geol 224:183–200<br />
Rodbell DT, Seltzer GO, Anderson DM, Abbott MB, Enfield DB, Newman JH (1999) An 15,000-<br />
year record <strong>of</strong> El Nino-driven alluviation in southwestern Ecuador. Science 283:516–520<br />
Schnellmann M, Anselmetti FS, Giardini D, McKenzie JA, Ward SN (2002) Prehistoric earthquake<br />
history revealed by lacustrine slump deposits. Geology 30:1131–1134<br />
Schnellmann M, Anselmetti FS, Giardini D, McKenzie JA (2006) 15,000 years <strong>of</strong> m<strong>as</strong>s-movement<br />
history in <strong>Lake</strong> Lucerne: implications for seismic <strong>and</strong> tsunami hazards. Eclogae Geol Helv<br />
99:409–428<br />
Schneuwly-Bollschweiler M, St<strong>of</strong>fel M (2012) Dendrogeomorphology – tracking p<strong>as</strong>t events with<br />
tree rings. In: Schneuwly-Bollschweiler M, St<strong>of</strong>fel M, Rudolf-Miklau M (eds) Dating torrential<br />
processes on fans <strong>and</strong> cones – methods <strong>and</strong> their application for hazard <strong>and</strong> risk <strong>as</strong>sessment,<br />
Advances in global change research. Springer, Dordrecht/Heidelberg/London/New York<br />
Sheffer NA, Enzel Y, Benito G, Grodek T, Poart N, Lang M, Naulet R, Coeur D (2003) Pale<strong>of</strong>loods<br />
<strong>and</strong> historical floods <strong>of</strong> the Ardeche River, France. Water Resour Res 39(12):1376<br />
Siegenthaler C, Sturm M (1991) Die Häufigkeit von Ablagerungen extremer Reuss-Hochw<strong>as</strong>ser.<br />
Die Sedimentationsgeschichte im Urnersee seit dem Mittelalter. In: Ursachenanalyse der<br />
Hochw<strong>as</strong>ser 1987 – Ergebnisse der Untersuchungen. Mitteilungen des Bundesamtes für<br />
W<strong>as</strong>serwirtschaft, vol 4, Bern, pp 127–139<br />
Sletten K, Blikra LH, Ballantyne CK, Nesje A, Dahl SO (2003) Holocene debris flows recognized<br />
in a lacustrine sedimentary succession: sedimentology, chronostratigraphy <strong>and</strong> cause <strong>of</strong><br />
triggering. The Holocene 13:907–920<br />
St<strong>of</strong>fel M, Bollschweiler M (2008) Tree-ring analysis in natural hazards research – an overview.<br />
Nat Hazard Earth Syst Sci 8:187–202<br />
Str<strong>as</strong>ser M, Anselmetti FS, Fäh D, Giardini D, Schnellmann M (2006) Magnitudes <strong>and</strong> source are<strong>as</strong><br />
<strong>of</strong> large prehistoric northern Alpine earthquakes revealed by slope failures in lakes. Geology<br />
34:1005–1008<br />
Sturm M, Matter A (1978) Turbidites <strong>and</strong> varves in <strong>Lake</strong> Brienz (Switzerl<strong>and</strong>): deposition <strong>of</strong> cl<strong>as</strong>tic<br />
detritus by density currents. Spec Publ Int Assoc Sedimentol 2:147–168<br />
Sturm M, Siegenthaler C, Pickrill RA (1995) Turbidites <strong>and</strong> ‘homogenites’ – a conceptual model<br />
<strong>of</strong> flood <strong>and</strong> slide deposits. In: Publication <strong>of</strong> IAS-16th regional meeting sedimentology, vol<br />
22, Paris, p 140<br />
Theiler A (2003) Die Sedimente des Seeli (Seelisberg, Uri) – Starkniederschläge im Holozän.<br />
Unpublished diploma, Geological Institute, ETH Zurich<br />
Thevenon F, Anselmetti FS (2007) Charcoal <strong>and</strong> fly-<strong>as</strong>h particles from <strong>Lake</strong> Lucerne sediments<br />
(Central Switzerl<strong>and</strong>) characterized by image analysis: anthropologic, stratigraphic <strong>and</strong> environmental<br />
implications. Quaternary Sci Rev 26:2631–2643<br />
Thorndycraft VR, Benito G (2006) Late Holocene fluvial chronology <strong>of</strong> Spain: the role <strong>of</strong> climatic<br />
variability <strong>and</strong> human impact. Catena 66:34–41<br />
Thorndycraft V, Hu Y, Oldfield F, Crooks PRJ, Appleby PG (1998) Individual flood events detected<br />
in the recent sediments <strong>of</strong> the Petit Lac d’Annecy, e<strong>as</strong>tern France. The Holocene 8:741–746<br />
Thorndycraft VR, Benito G, Rico M, Sopena A, Sánchez-Moya Y, C<strong>as</strong><strong>as</strong> A (2005) A long-term<br />
flood discharge record derived from slackwater flood deposits <strong>of</strong> the Llobregat River, NE Spain.<br />
J Hydrol 313:16–31<br />
von Gunten L, Grosjean M, Beer J, Grob P, Morales A, Urrutia R (2009) Age modeling <strong>of</strong> young<br />
non-varved lake sediments: methods <strong>and</strong> limits. Examples from two lakes in Central Chile.<br />
J Paleolimnol 42:401–412<br />
Wirth S (2008) <strong>Lake</strong> Thun sediment record: 300 years <strong>of</strong> human impact, flood events <strong>and</strong><br />
subaquatic slides. Unpublished MSc thesis, Department <strong>of</strong> Earth Sciences, ETH Zurich,<br />
doi:10.3929/ethz-a-005676930
242 A. Gilli et al.<br />
Wirth SB, Girardclos S, Rellstab C, Anselmetti FS (2011) The sedimentary response to a pioneer<br />
geo-engineering project: tracking the K<strong>and</strong>er River deviation in the sediments <strong>of</strong> <strong>Lake</strong> Thun<br />
(Switzerl<strong>and</strong>). Sedimentology 58:1737–1761<br />
Wolfe BB, Hall RI, L<strong>as</strong>t WM, Edwards TWD, English MC, Karst-Riddoch TL, Paterson A,<br />
Palmini R (2006) Reconstruction <strong>of</strong> multi-century flood histories from oxbow lake sediments,<br />
Peace-Athab<strong>as</strong>ca Delta, Canada. Hydrol Process 20:4131–4153<br />
Zolitschka B (2006) Varved lake sediments. In: Eli<strong>as</strong> SA (ed) Encyclopedia <strong>of</strong> Quaternary science.<br />
Elsevier, Amsterdam, pp 3105–3114