Lake Sediments as Archives of Recurrence Rates and ... - Eawag

Lake Sediments as Archives of Recurrence Rates
and Intensities of Past Flood Events
Adrian Gilli, Flavio S. Anselmetti, Lukas Glur, and Stefanie B. Wirth
1 Introduction
Palaeoflood hydrology is an expanding field as the damage potential of floods and
flood-related processes is increasing with the population density and the value of the
infrastructure. Assessing the risk of these hazards in mountainous terrain requires
knowledge about the frequency and severness of such events in the past. A wide
range of methods is employed using diverse biologic, geomorphic or geologic
evidences to track past flood events. Impact of floods are studied and dated on
alluvial fans and cones using for example the growth disturbance of trees (Stoffel
and Bollschweiler 2008; Schneuwly-Bollschweiler and Stoffel 2012: this volume)
or stratigraphic layers deposited by debris flows, allowing to reconstruct past flood
frequencies (Bardou et al. 2003). Further downstream, the classical approach of
palaeoflood hydrology (Kochel and Baker 1982) utilizes geomorphic indicators
such as overbank sediments, silt lines and erosion features of floods along a
river (e.g. Benito and Thorndycraft 2005). Fine-grained sediment settles out of
the river suspension in eddies or backwater areas, where the flow velocity of the
river is reduced. Records of these deposits at different elevations across a river’s
A. Gilli ()•S.B.Wirth
Geological Institute, ETH Zurich, CH-8092 Zurich, Switzerland
e-mail: adrian.gilli@erdw.ethz.ch
F.S. Anselmetti
Eawag, Department of Surface Waters, CH-8600 Dübendorf, Switzerland
Institute of Geological Sciences, University of Bern, Baltzerstrasse 1C3, CH-3012 Bern,
Switzerland
e-mail: flavio.anselmetti@geo.unibe.ch
L. Glur
Department of Surface Waters, Eawag, CH-8600 Dübendorf, Switzerland
e-mail: lukas.glur@eawag.ch
M. Schneuwly-Bollschweiler et al. (eds.), Dating Torrential Processes on Fans and
Cones, Advances in Global Change Research 47, DOI 10.1007/978-94-007-4336-6 15,
© Springer ScienceCBusiness Media Dordrecht 2013
225

226 A. Gilli et al.
profile can be used to assess the discharge of the past floods. This approach of
palaeoflood hydrology studies was successfully applied in several river catchments
(e.g. Ely et al. 1993; Macklin and Lewin 2003; O’Connor et al. 1994; Sheffer et al.
2003; Thorndycraft et al. 2005; Thorndycraft and Benito 2006). All these different
reconstruction methods have their own advantages and disadvantages, but often
these studies have a limited time coverage and the records are potentially incomplete
due to lateral limits of depositional areas and due to the erosional power of fluvial
processes that remove previously deposited flood witnesses. Here, we present a
method that follows the sediment particle transported by a flood event to its final
sink: the lacustrine basin.
Lakes offer a great possibility to record flood events as their suspension load is
deposited on the lake floor as a distinct detrital layer. This detrital layer contrasts in
the ideal case strongly with the ‘normal’ background sedimentation, which mainly
consists of authigenically-produced minerals, organic remains and clastic particles.
Detrital flood layers are detected visually or by a suite of different sedimentological
and geochemical methods. Flood events that have been documented either by
historical sources or by instrumentation were detected in lacustrine sediment cores
(Bussmann and Anselmetti 2010; Chapron et al. 2002; Gilli et al. 2003; Lamoureux
2000; Siegenthaler and Sturm 1991) confirming the concept of using lake sediments
as flood archive. Major advantage of the lacustrine archive is the continuous mode of
the background sedimentation in the deep lake areas, which offers the potential for
the reconstruction of a temporally complete flood record. The resolution of a lakebased
flood reconstruction is highly dependent on the overall sedimentation rate, but
in the case of annually laminated sediments the flood record could be resolved down
to the season. Past flood events can be scaled by thickness and grain-size distribution
of the detrital layer as they reflect flood dynamics.
Although modern limnogeology started to understand the involved processes
between high river-runoff and the deposition of detrital layers in lakes through
underflows in the 1970s (e.g. Lambert and Hsü 1979; Sturm and Matter 1978),
the limnogeologic community did not immediately realize the value of the lake
sediments for the reconstruction of flood history. For Switzerland, the severe
damage of several flood events in the Swiss Alps and its foreland in summer
1987 spurred different disciplines to reconstruct flood frequencies and intensities.
Siegenthaler and Sturm (1991) systematically sampled the sedimentary archive of
Lake Uri, which drained an area of severe flood damage. They documented in
detail the spatial characteristics of the 1987 flood layer in Lake Uri and were able
to identify further strong flood events that occurred in the last 700 years. This
intensified flood-risk evaluation after the 1987-flood event was amplified by the
fact that Switzerland experienced only few incidences of natural disasters between
1882 and 1976 (Pfister 2009) and, therefore, was not aware of the natural variability
of such devastating events. Especially in Europe and North America, the number
of flood reconstructions using lake sediments (e.g. Arnaud et al. 2005; Bussmann
and Anselmetti 2010; Chapron et al. 2002; Dapples et al. 2002; Eden and Page
1998; Gilli et al. 2003; Mazzucchi et al. 2003; Nesje et al. 2001; Noren et al. 2002;
Osleger et al. 2009; Parris et al. 2010) increased exponentially in the last 15 years
underlining the great potential of this method.

Lake Sediments as Archives of Recurrence Rates and Intensities of Past Flood Events 227
2 Depositional Processes in Lakes
Extreme flood events mobilize and entrain large amounts of sediments that are
consequently fed into the river drainage. The amount and the grain size of the
sediment particles entrained by a flow are strongly dependent on the hydraulic
energy. The coarser fraction of these mobilized particles may remain in the
catchment, forming alluvial fans and cones. The finer particles, which are kept in
suspension, are transported from the catchment area to the next downstream lake,
which is the final sink for nearly the entire riverine sediment load (Fig. 1).
When a sediment-laden river enters a lake, the sediment plume is redistributed
depending on the density stratification of the lake and the density of the entering
water, which is a function of temperature, suspension load and ion content (salinity).
The density contrast between sediment-laden river water and the ambient lake water
determines if the sediment plume forms an over-, inter- or underflow in the lake
Fig. 1 Schematic drawing of a lake and its catchment area illustrating in particular the different
distribution paths of a river-borne sediment plume in a lake as over-, inter-, or underflow (Giovanoli
1990; Sturm and Matter 1978). Coarser sediment particles mobilized through a flood are mainly
deposited on alluvial fans or in the delta, whereas the finer particles are entering the lake
as sediment plume. High-density flows, generated by heavy rainfall events, form underflows
depositing their particles as specific turbidite layers, which focus in the deepest lake area.
Lakes dominated by underflows have a flat and near horizontal lake bottom. Seismic investigations
image the subsurface providing a seismic stratigraphy that allows prediction of sediment type and
selection of ideal coring sites

228 A. Gilli et al.
Fig. 2 Grain-size measurements across two flood layers in Lake Thun, Switzerland (Wirth 2008;
Wirth et al. 2011). Vertical white double-arrow indicates extent of flood layer. (a) This turbidite
shows an upcore increase in grain size between 3 and 4 cm indicating, in addition to the coarse
base, a second peak in river runoff. (b) Turbidite layer with clearly inversed grading at the base as
a result of the waxing river runoff
(Giovanoli 1990). Regular runoff waters are usually less dense than the lake water
and thus rather form over- and interflows, from where the suspended fine particles
slowly settle out and eventually become deposited on the lake floor. Extreme
events, on the other hand, are so suspension-laden that the high density of the flow
allows the water to break through the lake stratification to proceed as turbiditic
underflow (hyperpycnal flow) to the deepest area of the lake (Fig. 1). Once the
deepest lake area is reached, the underflow spreads out and flow velocities decrease.
Consequently, the sediment particles are deposited forming characteristic layers
termed ‘turbidites’. The depositional characteristics of a turbidite are dependent
on temporal runoff evolution (i.e. the shape of the hydrograph) of a flood event
as the grain size is a direct function of the current velocity (Mulder et al. 2001,
2003). Turbidites comprise a fining upward sequence representing the waning of
the discharge (Fig. 2). Sometimes, turbidites also record the increasing underflow
velocity during the initiation of a flood by a coarsening upward basal unit (Fig. 2b).
This concept of directly linking the rising and falling river discharge with the reverse
and subsequently normal graded turbidite deposition was recently questioned for the
marine realm (Lamb and Mohrig 2009), although for the smaller lacustrine setting
this concept probably still is valid.

Lake Sediments as Archives of Recurrence Rates and Intensities of Past Flood Events 229
2.1 Selecting Lakes with Promising Flood Records
The selection of appropriate lakes is crucial for this approach of flood
reconstruction. The understanding of the transport and depositional processes of
flood-transported particles allows us to set criteria to select lakes holding promising
flood records. The following two criteria are highly relevant:
1. Well-defined morphobathymetric depocentre
Underflows transport the suspension load of a flood along the delta slope towards
the deepest part of the lake, where flow velocity decreases and larger particles
start to settle down on the lake floor. Lakes with underflow-dominated sedimentation
are characterized morphologically by a flat and horizontal basin floor, where
the flood turbidites accumulate and consequently level out inherited topography
(Fig. 1). This flat floor is the prime target area to recover cores for flood
reconstruction studies. A detailed knowledge about the lake’s bathymetry gained
from map studies, detailed depth soundings or reflection seismic surveying is
therefore a prerequisite for a successful flood reconstruction. Although in most
of the cases, the deepest lake part is the depocentre for the underflows, possible
sills between sub-basins may prevent the underflows reaching the lake’s centre.
This needs to be taken into account with complex lake basin morphologies.
Furthermore, reflection seismic information from the subsurface is needed to
show the stratigraphic architecture of the sedimentary basin fill (see Fig. 1).
Potential unconformities related to large mass movements, bottom currents, or
gas escape features may seriously hamper the continuous nature of the sediment
succession. Only a detailed seismic stratigraphy allows to select ideal coring
locations, define sediment thickness and predict sediment types.
2. Geomorphic indications around the lake
As a consequence of the high detrital influx to a lake during flood events, the
following geomorphic indications around the lake should be present.
(a) A certain relief in the catchment area is necessary in order to erode material
that can be transported into the lake.
(b) The lake should have evident inflows, however, they should ideally only be
activated in the case of an extreme event. That provides a certain threshold
with an ‘on/off signal’ so that only the larger ‘extreme’ events are recorded
in the lacustrine archive.
(c) A delta structure should be present at the major inflows of a lake assuring a
persistent detrital influx in the past.
The following criterion deals with the recognition of the flood layers in the sedimentary
record. This can often only be evaluated after preliminary investigations
using short sediment cores.
3. Contrast between flood deposits and regular background sedimentation
The lithology of the flood deposits should contrast to the regular background
sediments, so that these events can be lithologically and geochemically recognized.
This implies that the regular background sedimentation should not

230 A. Gilli et al.
Fig. 3 Flood layers in a core section (core LL081-A3a) of Lake Ledro, Italy, with relative
concentration of silicium (Si) and iron (Fe) measured continuously by an XRF core scanner
(measuring interval: 0.2 mm). Dark-coloured flood layers stand out prominently in the finelylaminated
calcareous background sediment (note the lightness difference reported in reflectance
L * ). Flood layers (grey bars) are characterized by increased Si and Fe content, especially
pronounced in the fine-grained top part of the flood layer (‘clay cap’) marked in light grey bars
consist solely of detrital material. If organic materials dominate the background
sediments, the detrital layers are in general lighter than the otherwise very dark
background sediments (e.g. Lake Schwendi, Gilli et al. 2003). But if the debris
background sediments manly consist of authigenic carbonates, the detrital layers
are considerably darker than the normal sediment (e.g. Lake Ledro, see Fig. 3).
In either setting, detrital layers stand out prominently when compared to the
background sediments and can easily be identified.
Beside the described criteria for selecting a lake system to record recognizable
flood layers, more technical aspects to recover the expected flood record are as well
of importance for a successful study.

Lake Sediments as Archives of Recurrence Rates and Intensities of Past Flood Events 231
4. Coring ability to recover desired time interval
The sedimentation rate of a lacustrine archive can vary strongly between the
different lake systems. In particular, the contribution of the general detrital influx
to the total sedimentation can make a large difference to total sediment thickness.
Standard coring techniques with various piston-coring methods are capable of
reaching 20 m sub lake-floor depth, so that sediment thickness for the desired
time interval should not exceed this range. Deeper core recovery can otherwise
only be reached by complex and expensive commercial drilling operations. If a
complete Holocene section is the target, rather small lakes with little inflow are
ideal (Gilli et al. 2003; Theiler 2003). Larger perialpine lakes (e.g. Lake Uri, Lake
Brienz, both Switzerland) have substantially larger sedimentation rates reaching
values of up to 10 m per 1,000 years (Girardclos et al. 2007; Siegenthaler and
Sturm 1991).
3 The Signature of Flood Events in the Lacustrine Archive
A key step in palaeoflood studies using lake sediments is the secure identification of
detrital layers as the result of a flood event. As the sedimentation of each lake system
is influenced by a large variety of different factors, we can only discuss general rules
here, which clearly need to be adapted to each individual lacustrine system (Mulder
and Chapron 2011). Ideally, the detrital layers should sharply contrast in terms of
sedimentology and geochemical composition with the regular background sediment
(Sletten et al. 2003). In these cases, the detrital layer can be often identified by
visual observation even at the millimetre scale, a process, which can be supported
by image analysis (e.g. Rodbell et al. 1999)(Fig.3).
The sedimentology of the flood layers is characterized by a near pure terrestrial
mineralogical composition and consequently a low organic content. The underflowrelated
depositional processes (see part “Depositional processes in lakes”) cause
a fining-upward grain-size pattern with sometimes an inverse graded base from
temporal waxing of the flood-related runoff. The top part of the detrital layer
consists of the finest particles and has often a distinctive light colour (Fig. 3),
commonly referred as ‘clay cap’. Sometimes, several stacked graded intervals
within the same flood layer point towards multiple water discharge peaks (Fig. 2). In
rather rare cases, the base of flood layers consists of aquatic macrophytes (Gilli et al.
2003), which were entrained by the underflow. On the basis of the characteristics of
these detrital layers, previous studies applied, next to visual observations, one or
a combination of several proxies to identify such layers. The three most common
applied proxies are (1) grain size, (2) total organic carbon (often expressed as loss
of ignition, i.e. LOI), and (3) magnetic susceptibility. These proxies are interpreted
as follows:
1. Grain size: Grain-size measurements are almost routinely used to confirm a
fining-upward gradation overlying a coarser base (Fig. 2). Recently, Parris

232 A. Gilli et al.
et al. (2010) proposed end-member modelling of the particle-size distribution
and demonstrated that the coarsest end-member is a sensitive indicator for the
base of detrital flood layers.
2. Total organic carbon (TOC): Caused by the high minerogenic content of the
flood-related material, the organic-carbon content is usually low in the flood
layers. This proxy is successfully applicable only in rather eutrophic lakes, where
the background sedimentation is high in organic material offering sufficient
contrast to the detrital signature.
3. Magnetic susceptibility: Depending on the catchment’s geology, the detrital
layers are often enriched in para- and ferromagnetic minerals reflecting high
magnetic susceptibility values (Chapron et al. 2005; Wolfe et al. 2006). More
elaborated proxies like isothermal remanent magnetization (IRM) have also been
applied to identify surface-soil-derived material in lacustrine sediments (Boe
et al. 2006; Thorndycraft et al. 1998).
Beside these widely applied proxies, various additional methods are proposed to
better identify detrital layers in the lacustrine archive.
• Elemental composition: Elements like Si, Al, Fe, K, and Ti are typical for detrital
sediments and thus their concentration is elevated in the detrital layers (Revel-
Rolland et al. 2005). Elemental concentration is determined either on discrete
samples or by continuous non-destructive XRF core scanning techniques (Fig. 3)
with resolutions down to the sub-millimetre scale.
• Bulk density: Especially the coarser base of detrital layers is characterized by
elevated bulk densities, which can be determined by continuous gamma-ray
attenuation measurements performed on a standard multi-sensor core logging
device (Gilli et al. 2003).
• Image analysis: Detrital layers can be detected and mapped with image analysis
(Fig. 3) if there is a clear colour contrast between detrital layer and background
sediments.
• Carbon/nitrogen ratios: Organic matter of lacustrine origin reveals a distinct
lower C/N ratio (i.e. 4–10) compared to vascular land plants (>20) (Meyers and
Teranes 2001). Several studies showed an increase of the C/N ratio in the detrital
flood layers (Brown et al. 2000; Osleger et al. 2009; Wolfe et al. 2006)
• Stable carbon isotopes on organic material: As with the C/N ratio, carbon
isotopes can be used to differentiate between aquatic or terrestrial origin of the
organic matter (Bierman et al. 1997). In general, rather higher carbon isotopic
ratios (• 13 C) are typical for terrestrial plants and found in the detrital layers
(Osleger et al. 2009; Wolfe et al. 2006)
• Pollen assemblages: Pollen spectra of detrital layers show a clear shift towards
more robust pollen types, like Liguliflorae and Filicales pollens, as they remain
in the catchment and are subsequently washed-in, when more vulnerable pollen
types are already corroded and decayed (Thorndycraft et al. 1998).
• Sr and Nd isotopes: If the lake’s catchment consists of different geologic
terrains, strontium and neodymium isotopes of the detrital particles stand out
prominently when compared to the values from background sediment (Revel-
Rolland et al. 2005).

Lake Sediments as Archives of Recurrence Rates and Intensities of Past Flood Events 233
The clear identification and attribution of a detrital layer as a past flood event
can sometimes be challenging as similar layers could also be the result of mass
movements (i.e. slope failures) triggered by earthquakes (Chapron et al. 1999;
Schnellmann et al. 2002; Strasser et al. 2006), delta collapse (Girardclos et al.
2007) or lake level fluctuations (Anselmetti et al. 2009). But these mass-movement
related turbidite layers have different structural and compositional characteristics
and they are – in general – rather thicker than flood layers (Siegenthaler and Sturm
1991; Sturm et al. 1995). Mass-movement related turbidite layers show only a thin
fining-upward sequence followed by a thick homogenous interval with no inverse
grading at the base. With the exception of turbidites triggered by delta collapses,
the mass-movement layers contain reworked lacustrine background sediment with a
geochemical composition different from the catchment-derived detrital signal.
Detailed sedimentological analysis of detrital layers is further of importance to
scale intensities of past flood events. Assessing past flood events intensity is always
based on a constant availability of sediment during the observed period. The amount
of sediment, however, that is mobilized by intense precipitation and runoff may
be also dependent on anthropogenic activity, as various forms of land-use expose
the soils and increase their sensitivity to become eroded. Nevertheless, a positive
relationship between suspended particle concentration and river discharge (Mulder
et al. 2003) enables a scaling of the flood event by the detrital layer thickness. This
relationship is specific for each river basin, but follows a power-law function making
the sedimentary record very sensitive to variations in the river discharge of extreme
floods (Mulder et al. 2003). So far, only a few studies used this relationship to
assess the flood magnitude of past events (e.g. Brown et al. 2000; Bussmann and
Anselmetti 2010; Irmler et al. 2006). These studies always use the imprint of a
documented historic flood event in the sedimentary archive to relatively scale prehistoric
flood events. A proxy even less applied in evaluating past flood events in the
sedimentary archive is grain-size distribution. Grain-size measurements are often
carried out to identify detrital layers, but as the grain-size distribution also reflects
hydrologic conditions, in particular transport capacity, during a flood event, it could
be applied to assess the intensity of past floods.
4 Dating Methods for Lake Sediments
Once the flood layers have been identified, age dating of the sedimentary succession
is required to establish a flood chronology. This can be achieved applying various
dating techniques, which eventually result in a ‘best-fit’ age model for the recovered
stratigraphic section. For this purpose, one has to consider that a sedimentary
sequence consists of background sediments, characterized by low sedimentation
rate, which is punctuated or ‘interrupted’ by individual event-layers that were
deposited in short time (i.e. hours to days) thus resulting in a high sedimentation
rate. Consequently, in order to interpolate between age markers, those rapidly
deposited layers need to be removed from the sedimentary succession before the

234 A. Gilli et al.
age model is defined. The resulting interpolation assumes a stable background
sedimentation rate between age markers, which is often the best-possible approach.
The uppermost part of the sediments that were deposited in the last century can
be dated measuring the gamma-activity of radionuclides, in particular 210 Pb and
137 Cs (Appleby 2001, and references therein). 137 Cs emissions into the environment
started with the onset of atmospheric nuclear bomb testing in the mid-1950s
culminating in 1963. A second maximum was produced upon the accident of
Chernobyl in 1986, offering thus two marker horizons that can be detected by
gamma spectroscopy. 210 Pb-activities, in contrast, are a product of in situ produced
‘supported’ as well of ‘unsupported’ sources from a decay chain, the latter of which
allows calculation of sedimentation rates for the last 150 years assuming various
sedimentation models (Appleby 2001; von Gunten et al. 2009). Furthermore, the
occurrence of combustion products, such as spherical carbonaceous particles (SPE)
or various other combustion-derived aerosols reflect local to regional use of burning
agents and can be compared to anthropogenic and industrial history (Thevenon and
Anselmetti 2007; von Gunten et al. 2009), providing additional dating possibilities
for the most recent section.
Radiocarbon dating of terrestrial organic matter is probably the most common
and powerful dating tool that is used with lacustrine sediments. The success of
this method relies on identification of organic macro-remains of terrestrial origin,
as aquatically produced biomass often incorporates ‘old’ carbon, which is not in
equilibrium with the atmosphere (Deevey et al. 1954). This radiocarbon reservoir
effect causes an offset towards too old radiocarbon ages, which for this reason
prevents building a robust age model.
Further possibilities for providing age markers are the occurrences of volcanic
tephra layers of known age, which can be tracked over large distances. In Swiss
lakes, the Laacher See Tephra (originating from the Eifel area in Germany) and
the Vasset-Killian Tephra from an eruption in the Massif Central in France (Hajdas
et al. 1993) can often be identified and may refine a radiocarbon-based age model
(Schnellmann et al. 2006).
Annually laminated background sediments (i.e. varves, see Zolitschka 2006 for
an overview) provide probably the most accurate dating by layer counting, as annual
resolution cannot be obtained by the other methods. However, as varves rarely occur
throughout the entire section, varve chronology may be ‘floating’ and therefore be
used together with other independently determined age markers providing thus a
relative rather than an absolute chronology of annual resolution.
Eventually, the age model may be adjusted by making reference to prominent
historic event markers as earthquakes, rock falls or large floods leave their unique
fingerprint behind. Several studies (e.g. Bussmann and Anselmetti 2010; Gilli
et al. 2003; Siegenthaler and Sturm 1991) linked detrital layers within the 137 Cs
and 210 Pb-dated interval to documented extreme flood events producing higher
accuracies of the independently dated succession. Overall, defining an age model is
usually a combined application of various methods described above, which together
maximize range, resolution, and accuracy of the ages that are then assigned to each
individual flood layer.

Lake Sediments as Archives of Recurrence Rates and Intensities of Past Flood Events 235
5 Determining the Seasonality of Flood Events
Flood reconstructions mainly focus on the frequency and intensity of past
heavy-precipitation events. This allows the identification of past time periods
of increased flood activity with implications for general atmospheric circulation
patterns. However, knowledge of past atmospheric processes leading to severe
floods would substantially be improved if the individual flood layers could be
seasonally resolved. With such high-resolution records, conclusions about – for
example – a change in frequency of floods caused by summer thunderstorms could
be drawn.
Varved sediments allow a precise chronology of the lacustrine record down to
the season based on the alternating deposition of distinct laminae during the annual
cycle (Zolitschka 2006). The position of detrital flood layers within the seasonal
alternating laminae of a varve may be used to date the season they were deposited.
Depending on the thickness of the varves, the position of the flood layer within a
varve is visually determined either on thin sections, on high-resolution core images,
or by the naked eye. To apply the concept of seasonal flood layer dating, a possible
erosion of the underlying laminae needs to be considered. Thin section analysis is
used to identify an erosional base of the detrital layer, but a safe seasonal assignment
of the flood layer can be achieved by determining the season of the next overlying
varve lamina.
Unfortunately, continuously varved lacustrine sections over longer time intervals
are quite rare and if present, the lake setting is often not sensitive to flood events.
For this reason only very few seasonal resolved flood records are documented in the
literature (Lamoureux 2000; Mangili et al. 2005). With the above outlined large gain
of knowledge about flood-triggering atmospheric processes, the search for lacustrine
flood records in varved lacustrine sections and its detailed, but time-consuming
analysis, is highly desirable.
6 Occurrence of Flood Events in the Past – Case Studies
The following two studies illustrate the application of lake sediments as archive for
flood events.
6.1 Flood History of Lake Lauerz, Central Switzerland
Lake Lauerz is a small perialpine lake (surface area: 3 km 2 , catchment area:
69 km 2 ) located in Central Switzerland that records heavy precipitation events
with typical flood layers. Piston cores were retrieved in the deepest part of the
main lake basin allowing a composite record of nearly 10 m length (Bussmann

236 A. Gilli et al.
Fig. 4 Chronology of flood layers determined from a sediment core from Lake Lauerz (Switzerland)
covering the last 2,000 years (Bussmann and Anselmetti 2010). Intervals of frequent flood
events are shaded and labelled with even roman numbers. The layer thickness of the flood
event from 1934 AD allows a scaling of the storm magnitude of past events. Six flood events
(marked with stars) were as severe as the historically documented flood event from 1876, which
caused large damages in Central Switzerland (reprinted by permission from the Swiss Geological
Society, © 2010)
and Anselmetti 2010). The sediments were dated with four radiocarbon ages to
establish a core chronology comprising the last 2,000 years (Fig. 4). A high bulk
sedimentation-rate of up to 1 cm year 1 provides a unique high temporal resolution
of palaeofloods in Lake Lauerz. 54 single flood events with a layer thickness
between 0.5 and 9 cm were identified. Interestingly, the flood events are not evenly
distributed in the past 2,000 years, but rather cluster in three periods (Fig. 4). An
increased occurrence of heavy precipitation events is clearly documented between
1940–1630 AD, 1420–990 AD and 850–580 AD in the palaeoflood record of Lake
Lauerz (Bussmann and Anselmetti 2010).
Under the assumption that layer thickness is a proxy for storm magnitude (Brown
et al. 2000), the flood layers detected during the instrumental period in the twentieth
century were used to scale past flood events. The largest flood during the period
of precipitation measurements occurred in 1934 AD. This event caused severe
devastation around Lake Lauerz and was recorded with a 2.5 cm thick detrital layer
in the lake basin. On the basis of the established 2,000-year palaeoflood record, 17
of the 54 detected events were recorded with an even larger layer thickness and are
thus interpreted to reflect a larger flood intensity than the 1934 AD event. One of
the most severe flood events in Central Switzerland during historic times, which
is recorded with layer thickness of 4.5 cm in Lake Lauerz, occurred in 1876 AD.
Considering the event layer thicknesses, five earlier flood events had a similar or
even larger storm magnitude pointing towards the occurrence of severe runoff events
in the past 2,000 years (Bussmann and Anselmetti 2010).

Lake Sediments as Archives of Recurrence Rates and Intensities of Past Flood Events 237
Fig. 5 Holocene palaeoflood records from northeastern United States (Noren et al. 2002). (a)
Palaeoflood records from 13 individual lakes. (b) Weighted compilation of all flood records
arranged in 100-year bins. Linear increasing trend of the storminess toward modern is explained
by the authors by delta progradation leading to an increased sensitivity of the coring site
to record flood events. (c) GISP2 non-sea-salt (n.s.s.) K concentration with values above the
superimposed linear regression shaded (reprinted by permission from Macmillan Publishers Ltd:
Noren et al. 2002)
6.2 Flood History from Lakes in Northeastern United States
Noren et al. (2002) conducted a comprehensive study about millennial-scale
periodicity of exceptional rainfall events in the northeastern United States covering
the entire Holocene. Unique in the design of this study is the stacking of the
chronological occurrence of detrital layers from 13 individual lakes (Fig. 5a, b). This

238 A. Gilli et al.
multi-archive approach is less susceptible to the influence of only locally occurring
events. The results have therefore more paleoclimatic significance when compared
to studies dealing solely with a single lake basin.
Lakes investigated within this study fulfil clearly the criteria for promising
lacustrine flood reconstruction (cf. section “Selecting lakes with promising flood
records”) with well-defined morphobathymetric depocentres, a steep surrounding
relief, evident inflows and delta structures, as well as predominantly organic background
sedimentation contrasting with the detrital flood layers. The identification
of flood layers was achieved by visual logging and high-resolution X-radiography
(providing information on internal sediment structures), as well as magnetic susceptibility,
loss-on-ignition and grain-size measurements.
Spectral analysis and correlations with proxies from the Greenland Ice Sheet
Project 2 (GISP2) ice-core record could connect the established flood record with
large-scale atmospheric circulation patterns (Fig. 5b, c). Periods with high flood
activity show a recurrence interval of about 3,000 years and correlate in the
GISP2 record with periods characterized by higher concentrations of non-sea-salt
potassium (n.s.s. K). Increased concentrations of these aerosols are attributed to
enhanced storminess in the high-latitude North-Atlantic region. Hence, the results
indicate that precipitation intensity in the northeastern United States is controlled
by large-scale atmospheric circulation modes such as the Arctic Oscillation (AO)
(Noren et al. 2002).
7 Conclusions
Lakes have a great potential to record flood events by the deposition of detrital layers
through turbiditic underflows. The major advantages of the lacustrine approach in
palaeoflood reconstructions compared to analysis of alluvial fans and cones are
the continuous mode of sedimentation, the high temporal resolution and the long
time period covered by lake sediments. These advantages led to an increasing
number of studies exploring the lacustrine archives for palaeoflood reconstructions.
The results of these studies are very promising and clearly indicate periods of
different flood frequencies for the Holocene. However, these studies cover almost
exclusively a single catchment and therefore may include local events. Compilations
of different lacustrine palaeoflood records, as for example applied by Noren et al.
(2002), overcome this limitation and are able to build regionally robust flood reconstructions.
With the now available limnogeologic data sets of past floods, linkages
with other scientific communities like geomorphologists and dendrochronologists,
need to be established to compile palaeoflood records from different archives. This
combined approach (as outlined by Irmler et al. 2006) will be of great importance to
understand the sensitivity of each flood archive and will ultimately lead to the most
complete flood reconstructions in terms of frequency and intensity.

Lake Sediments as Archives of Recurrence Rates and Intensities of Past Flood Events 239
Acknowledgements The authors thank the Swiss National Science Foundation (SNF) for the
financial support within the project ‘FloodAlp’ (Project No. 200021–121909) to initiate research
on the Holocene flood history in Switzerland and northern Italy using lake sediments. Figures 1
and 3 contains scientific results form a joint project on Lake Ledro, Italy whose co-funding by
the French ANR (project LAMA, directed by M. Magny and N. Combourieu-Nebout) is kindly
acknowledged. The cores from Lake Thun shown in Fig. 2 were collected as part of a project in
cooperation with Stephanie Girardclos, University of Geneva.
References
Anselmetti FS, Ariztegui D, De Batist M, Gebhardt AC, Haberzettl T, Niessen F, Ohlendorf C,
Zolitschka B (2009) Environmental history of southern Patagonia unraveled by the seismic
stratigraphy of Laguna Potrok Aike. Sedimentology 56:873–892
Appleby PG (2001) Chronostratigraphic techniques in recent sediments. In: Last WM, Smol JP
(eds) Tracking environmental change using lake sediments, vol 1, Basin analysis, coring and
chronological techniques. Kluwer Academic, Dordrecht, pp 171–201
Arnaud F, Revel M, Chapron E, Desmet M, Tribovillard N (2005) 7200 years of Rhone river
flooding activity in Lake Le Bourget, France: a high-resolution sediment record of NW Alps
hydrology. The Holocene 15:420–428
Bardou E, Fournier F, Sartori M (2003) Paleoflood reconstruction at Illgraben torrent (Switzerland):
a current need for event frequency estimation. In: Thorndycraft VR, Benito G,
Barriendos M, Llasat C (eds) Palaeofloods, historical floods and climate variability: applications
in flood risk assessment. Proceedings of the PHEFRA international workshop, Barcelona,
pp 53–59, 16–19 October 2002
Benito G, Thorndycraft VR (2005) Palaeoflood hydrology and its role in applied hydrological
sciences. J Hydrol 313:3–15
Bierman P, Lini A, Zehfuss P, Church A, Davis PT, Southon J, Baldwin L (1997) Postglacial ponds
and alluvial fans: recorders of Holocene landscape history. GSA Today 7:1–8
Boe AG, Dahl SO, Lie O, Nesje A (2006) Holocene river floods in the upper Glomma
catchment, southern Norway: a high-resolution multiproxy record from lacustrine sediments.
The Holocene 16:445–455
Brown SL, Bierman PR, Lini A, Southon J (2000) 10,000 yr record of extreme hydrologic events.
Geology 28:335–338
Bussmann F, Anselmetti FS (2010) Rossberg landslide history and flood chronology as recorded
in Lake Lauerz sediments (Central Switzerland). Swiss J Geosci 103:43–59
Chapron E, Beck C, Pourchet M, Deconinck J-F (1999) 1822 earthquake-triggered homogenite in
Lake Le Bourget (NW Alps). Terra Nova 11:86–92
Chapron E, Desmet M, De Putter T, Loutre MF, Beck C, Deconinck JF (2002) Climatic variability
in the northwestern Alps, France, as evidenced by 600 years of terrigenous sedimentation in
Lake Le Bourget. The Holocene 12:177–185
Chapron E, Arnaud F, Noel H, Revel M, Desmet M, Perdereau L (2005) Rhone River flood deposits
in Lake Le Bourget: a proxy for Holocene environmental changes in the NW Alps, France.
Boreas 34:404–416
Dapples F, Lotter AF, van Leeuwen JFN, van der Knaap WO, Dimitriadis S, Oswald D (2002)
Paleolimnological evidence for increased landslide activity due to forest clearing and land-use
since 3600 cal BP in the western Swiss Alps. J Paleolimnol 27:239–248
Deevey ES, Gross MS, Hutchinson GE, Kraybill HL (1954) The natural C 14 content of materials
from hard-water lakes. Proc Natl Acad Sci USA 40:285–288
Eden DN, Page MJ (1998) Palaeoclimatic implications of a storm erosion record from late
Holocene lake sediments, North Island, New Zealand. Palaeogeogr Palaeoclimatol Palaeoecol
139:37–58

240 A. Gilli et al.
Ely LL, Enzel Y, Baker VR, Cayan DR (1993) A 5000-year record of extreme flood and climatechange
in the southwestern United States. Science 262:410–412
Gilli A, Anselmetti FS, Ariztegui D, McKenzie JA (2003) A 600-year sedimentary record of flood
events from two sub-alpine lakes (Schwendiseen, Northeastern Switzerland). Eclogae Geol
Helv 96(Supplement 1):S49–S58
Giovanoli F (1990) Horizontal transport and sedimentation by interflows and turbidity currents in
Lake Geneva. In: Tilzer MM, Serruya C (eds) Large lakes: ecological structure and function.
Springer, Berlin/Heidelberg, pp 175–195
Girardclos S, Schmidt OT, Sturm M, Ariztegui D, Pugin A, Anselmetti FS (2007) The 1996 AD
delta collapse and large turbidite in Lake Brienz. Mar Geol 241:137–154
Hajdas I, Ivy SD, Beer J, Bonani G, Imboden D, Lotter AF, Sturm M, Suter M (1993) AMS
radiocarbon dating and varve chronology of Lake Soppensee – 6000 to 12000 14 C years BP.
Clim Dyn 9:107–116
IrmlerR,DautG,Mäusbacher R (2006) A debris flow calendar derived from sediments of lake
Lago di Braies (N. Italy). Geomorphology 77:69–78
Kochel RC, Baker VR (1982) Paleoflood hydrology. Science 215:353–361
Lamb MP, Mohrig D (2009) Do hyperpycnal-flow deposits record river-flood dynamics? Geology
37:1067–1070
Lambert A, Hsü KJ (1979) Non-annual cycles of varve-like sedimentation in Walensee, Switzerland.
Sedimentology 26:453–461
Lamoureux S (2000) Five centuries of interannual sediment yield and rainfall-induced erosion in
the Canadian High Arctic recorded in lacustrine varves. Water Resour Res 36:309–318
Macklin MG, Lewin J (2003) River sediments, great floods and centennial-scale Holocene climate
change. J Quaternary Sci 18:101–105
Mangili C, Brauer A, Moscariello A, Naumann R (2005) Microfacies of detrital event layers
deposited in Quaternary varved lake sediments of the Piànico-Sèllere Basin (northern Italy).
Sedimentology 52:927–943
Mazzucchi D, Spooner IS, Gilbert R, Osborn G (2003) Reconstruction of Holocene climate change
using multiproxy analysis of sediments from Pyramid Lake, British Columbia, Canada. Arct
Antarct Alp Res 35:520–529
Meyers PA, Teranes JL (2001) Sediment organic matter. In: Last WM, Smol JP (eds) Tracking
environmental change using lake sediments, vol 2, Physical and geochemical methods. Kluwer
Academic, Dordrecht, pp 239–269
Mulder T, Chapron E (2011) Flood deposits in continental and marine environments: character
and significance. In: Slatt RM, Zavala C (eds) Sediment transfer from shelf to deep water -
Revisiting the delivery system: AAPG Studies in Geology 61:1–30
Mulder T, Migeon S, Savoye B, Faugères J-C (2001) Inversely graded turbidite sequences in the
deep Mediterranean: a record of deposits from flood-generated turbidity currents? Geo-Mar
Lett 21:86–93
Mulder T, Syvitski JPM, Migeon S, Faugères J-C, Savoye B (2003) Marine hyperpycnal flows:
initiation, behavior and related deposits. A review. Mar Pet Geol 20:861–882
Nesje A, Dahl SO, Matthews JA, Berrisford MS (2001) A 4500-yr record of river floods obtained
from a sediment core in Lake Atnsjoen, eastern Norway. J Paleolimnol 25:329–342
Noren AJ, Bierman PR, Steig EJ, Lini A, Southon J (2002) Millennial-scale storminess variability
in the northeastern United States during the Holocene epoch. Nature 419:821–824
O’Connor JE, Ely LL, Wohl EE, Stevens LE, Melis TS, Kale VS, Baker VR (1994) A 4500-year
record of large floods on the Colorado River in the Grand-Canyon, Arizona. J Geol 102:1–9
Osleger DA, Heyvaert AC, Stoner JS, Verosub KL (2009) Lacustrine turbidites as indicators
of Holocene storminess and climate: Lake Tahoe, California and Nevada. J Paleolimnol
42:103–122
Parris AS, Bierman PR, Noren AJ, Prins MA, Lini A (2010) Holocene paleostorms identified by
particle size signatures in lake sediments from the northeastern United States. J Paleolimnol
43:29–49

Lake Sediments as Archives of Recurrence Rates and Intensities of Past Flood Events 241
Pfister C (2009) The “Disaster Gap” of the 20th century and the loss of traditional disaster memory.
Gaia 18:239–246
Revel-Rolland M, Arnaud F, Chapron E, Desmet M, Givelet N, Alibert C, McCulloch M (2005) Sr
and Nd isotopes as tracers of clastic sources in Lake Le Bourget sediment (NW Alps, France)
during the Little Ice Age: palaeohydrology implications. Chem Geol 224:183–200
Rodbell DT, Seltzer GO, Anderson DM, Abbott MB, Enfield DB, Newman JH (1999) An 15,000-
year record of El Nino-driven alluviation in southwestern Ecuador. Science 283:516–520
Schnellmann M, Anselmetti FS, Giardini D, McKenzie JA, Ward SN (2002) Prehistoric earthquake
history revealed by lacustrine slump deposits. Geology 30:1131–1134
Schnellmann M, Anselmetti FS, Giardini D, McKenzie JA (2006) 15,000 years of mass-movement
history in Lake Lucerne: implications for seismic and tsunami hazards. Eclogae Geol Helv
99:409–428
Schneuwly-Bollschweiler M, Stoffel M (2012) Dendrogeomorphology – tracking past events with
tree rings. In: Schneuwly-Bollschweiler M, Stoffel M, Rudolf-Miklau M (eds) Dating torrential
processes on fans and cones – methods and their application for hazard and risk assessment,
Advances in global change research. Springer, Dordrecht/Heidelberg/London/New York
Sheffer NA, Enzel Y, Benito G, Grodek T, Poart N, Lang M, Naulet R, Coeur D (2003) Paleofloods
and historical floods of the Ardeche River, France. Water Resour Res 39(12):1376
Siegenthaler C, Sturm M (1991) Die Häufigkeit von Ablagerungen extremer Reuss-Hochwasser.
Die Sedimentationsgeschichte im Urnersee seit dem Mittelalter. In: Ursachenanalyse der
Hochwasser 1987 – Ergebnisse der Untersuchungen. Mitteilungen des Bundesamtes für
Wasserwirtschaft, vol 4, Bern, pp 127–139
Sletten K, Blikra LH, Ballantyne CK, Nesje A, Dahl SO (2003) Holocene debris flows recognized
in a lacustrine sedimentary succession: sedimentology, chronostratigraphy and cause of
triggering. The Holocene 13:907–920
Stoffel M, Bollschweiler M (2008) Tree-ring analysis in natural hazards research – an overview.
Nat Hazard Earth Syst Sci 8:187–202
Strasser M, Anselmetti FS, Fäh D, Giardini D, Schnellmann M (2006) Magnitudes and source areas
of large prehistoric northern Alpine earthquakes revealed by slope failures in lakes. Geology
34:1005–1008
Sturm M, Matter A (1978) Turbidites and varves in Lake Brienz (Switzerland): deposition of clastic
detritus by density currents. Spec Publ Int Assoc Sedimentol 2:147–168
Sturm M, Siegenthaler C, Pickrill RA (1995) Turbidites and ‘homogenites’ – a conceptual model
of flood and slide deposits. In: Publication of IAS-16th regional meeting sedimentology, vol
22, Paris, p 140
Theiler A (2003) Die Sedimente des Seeli (Seelisberg, Uri) – Starkniederschläge im Holozän.
Unpublished diploma, Geological Institute, ETH Zurich
Thevenon F, Anselmetti FS (2007) Charcoal and fly-ash particles from Lake Lucerne sediments
(Central Switzerland) characterized by image analysis: anthropologic, stratigraphic and environmental
implications. Quaternary Sci Rev 26:2631–2643
Thorndycraft VR, Benito G (2006) Late Holocene fluvial chronology of Spain: the role of climatic
variability and human impact. Catena 66:34–41
Thorndycraft V, Hu Y, Oldfield F, Crooks PRJ, Appleby PG (1998) Individual flood events detected
in the recent sediments of the Petit Lac d’Annecy, eastern France. The Holocene 8:741–746
Thorndycraft VR, Benito G, Rico M, Sopena A, Sánchez-Moya Y, Casas A (2005) A long-term
flood discharge record derived from slackwater flood deposits of the Llobregat River, NE Spain.
J Hydrol 313:16–31
von Gunten L, Grosjean M, Beer J, Grob P, Morales A, Urrutia R (2009) Age modeling of young
non-varved lake sediments: methods and limits. Examples from two lakes in Central Chile.
J Paleolimnol 42:401–412
Wirth S (2008) Lake Thun sediment record: 300 years of human impact, flood events and
subaquatic slides. Unpublished MSc thesis, Department of Earth Sciences, ETH Zurich,
doi:10.3929/ethz-a-005676930

242 A. Gilli et al.
Wirth SB, Girardclos S, Rellstab C, Anselmetti FS (2011) The sedimentary response to a pioneer
geo-engineering project: tracking the Kander River deviation in the sediments of Lake Thun
(Switzerland). Sedimentology 58:1737–1761
Wolfe BB, Hall RI, Last WM, Edwards TWD, English MC, Karst-Riddoch TL, Paterson A,
Palmini R (2006) Reconstruction of multi-century flood histories from oxbow lake sediments,
Peace-Athabasca Delta, Canada. Hydrol Process 20:4131–4153
Zolitschka B (2006) Varved lake sediments. In: Elias SA (ed) Encyclopedia of Quaternary science.
Elsevier, Amsterdam, pp 3105–3114