lable at ScienceDirect

Quaternary Science Reviews 80 (2013) 78e90

Contents lists avai

Quaternary Science Reviews

journal homepage: www.elsevier .com/locate/quascirev

Mid- to late Holocene flood frequency changes in the northeasternAlps as recorded in varved sediments of Lake Mondsee (Upper Austria)

Tina Swierczynski a, *, Stefan Lauterbach a, Peter Dulski a, José Delgado c, Bruno Merz b,Achim Brauer a

a GFZ German Research Centre for Geosciences, Section 5.2 e Climate Dynamics and Landscape Evolution, Telegrafenberg, 14473 Potsdam, Germanyb GFZ German Research Centre for Geosciences, Section 5.4 e Hydrology, Telegrafenberg, 14473 Potsdam, Germanyc University of Potsdam, Institute of Earth and Environmental Sciences, 14476 Potsdam, Germany

a r t i c l e i n f o

Article history:Received 29 November 2012Received in revised form1 August 2013Accepted 27 August 2013Available online

Keywords:Varved lake sedimentsDetrital layersmXRFMicrofaciesPalaeofloodsFlood frequency

* Corresponding author. Tel.: þ49 331 288 1338.E-mail address: swier@gfz-potsdam.de (T. Swiercz

0277-3791/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.quascirev.2013.08.018

a b s t r a c t

Annually laminated (varved) lake sediments with intercalated detrital layers resulting from sedimentaryinput by runoff events are ideal archives to establish precisely dated records of past extreme runoffevents. In this study, the mid- to late Holocene varved sediments of Lake Mondsee (Upper Austria) wereanalysed by combining sedimentological, geophysical and geochemical methods. This approach allows todistinguish two types of detrital layers related to different types of extreme runoff events (floods anddebris flows) and to detect changes in flood activity during the last 7100 years. In total, 271 flood and 47debris flow layers, deposited during spring and summer, were identified, which cluster in 18 main floodepisodes (FE 1e18) with durations of 30e50 years each. These main flood periods occurred during theNeolithic (7100e7050 vyr BP and 6470e4450 vyr BP), the late Bronze Age and the early Iron Age (3300e3250 and 2800e2750 vyr BP), the late Iron Age (2050e2000 vyr BP), throughout the Dark Ages ColdPeriod (1500e1200 vyr BP), and at the end of the Medieval Warm Period and the Little Ice Age (810e430 vyr BP).

Summer flood episodes in Lake Mondsee are generally more abundant during the last 1500 years,often coinciding with major advances of Alpine glaciers. Prior to 1500 vyr BP, spring/summer floods anddebris flows are generally less frequent, indicating a lower number of intense rainfall events that trig-gered erosion. In comparison with the increase of late Holocene flood activity in western and north-western (NW) Europe, commencing already as early as 2800 yr BP, the hydro-meteorological shift in theLake Mondsee region occurred much later. These time lags in the onset of increased hydrological activitymight be either due to regional differences in atmospheric circulation pattern or to the sensitivity of theindividual flood archives. The Lake Mondsee sediments represent the first precisely dated and severalmillennia long summer flood record for the northeastern (NE) Alps, a key region at the climatic boundaryof Atlantic, Mediterranean and East European air masses, aiding a better understanding of regional andseasonal peculiarities of flood occurrence under changing climate conditions.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Modelling studies suggest that the flood risk in Europe is likelygoing to increase under global warming conditions due to theassociated intensification of the water cycle and hence morefrequent extreme precipitation events (Arnell, 1999; Milly et al.,2002; Christensen and Christensen, 2003; Kundzewicz et al.,2005, 2006). However, the spatial and temporal distribution of

ynski).

All rights reserved.

flood events might be rather heterogeneous for Europe. Forinstance, instrumental records of floods for the last 50 years inGermany exhibit a spatial and seasonal pattern with increasedsummer flood occurrence in southern Germany but more winterfloods in western Germany (Petrow and Merz, 2009). This un-derlines that the response of flood patterns to climatic changes inCentral Europe is variable on temporal and spatial scales (Merz andBlöschl, 2003; Beurton and Thieken, 2009). To decipher themechanisms triggering flood recurrence and the resulting regionalflood patterns, mainly instrumental data have been analysed so far(Pfister et al., 2004; Mudelsee et al., 2006; Pinter et al., 2006;Delgado et al., 2010). In addition, for some regions historical flood

Fig. 1. Simplified geographical and geological map of Lake Mondsee and climate diagram for the 1971e2000 observation period (ZAMG, Vienna). The northern sub-basin iscomposed of Flysch sediments, which are drained by the three main tributaries Griesler Ache, Wangauer Ache and Zeller Ache. The southern sub-basin is fed by several smallercreeks, originating from the Northern Calcareous Alps. The investigated piston core has been retrieved from the southern sub-basin (white circle).

T. Swierczynski et al. / Quaternary Science Reviews 80 (2013) 78e90 79

event chronologies have been utilized to extend instrumental timeseries (Mudelsee et al., 2003; Brázdil et al., 2006; Glaser et al., 2010;Schmocker-Fackel and Naef, 2010). However, the natural variabilityof flood recurrence on centennial or millennial time scales cannotbe estimated from instrumental and historical data because thesedo not extend beyond the last 150e1000 years, respectively.

To improve the knowledge about the natural frequency andseasonality of severe floods, long and reliable records of extremeevents such as those preserved in geoarchives are required. Firstattempts to establish long flood time series have been made bypalaeoflood hydrology, using riverine deposits in the United States(Baker, 1987, 2008; Hirschboeck, 1988; Ely et al., 1993; Enzel andWells, 1997; Knox, 2000) and Europe (Dearing et al., 2001;Starkel, 2002; Sheffer et al., 2003; Thorndycraft et al., 2005;Chiverrell et al., 2008; Macklin et al., 2010). Despite the identifi-cation of significant flood occurrence patterns (Rumsby andMacklin, 1996; Knox, 2000), this approach is limited by poorlyresolved chronologies and the reconstruction of those flood events,which overtop the preceding ones due to river erosion. In contrastto riverine deposits, which predominantly record highest magni-tude floods, lake sediment records provide continuous records ofdetrital material deposition during floods of a certain magnitude(Thorndycraft et al., 1998; Brown et al., 2000; Nesje et al., 2001;Chapron et al., 2002; Arnaud et al., 2005; Chapron et al., 2005;Bøe et al., 2006; Moreno et al., 2008; Corella et al., 2011). Howev-er, low-resolution chronologies based on 14C dating do not provideinformation about decadal to sub-decadal time scales. Additionally,the seasonality of floods remains unknownwhen using age modelsestablished by 14C dating. In contrast, annually laminated (varved)lake sediments with intercalated detrital flood event layers alloweven establishing continuous and precisely dated event chronolo-gies with sub-annual resolution (Mangili et al., 2005; Czymzik et al.,2010, 2013). The European Alps is a climatically sensitive region(Auer et al., 2007), which is influenced by Mediterranean andAtlantic air masses, both triggering extreme flood events. Whereasmost palaeohydrological and flood reconstructions are reported

from the NW Alps (Gilli et al., 2003; Magny, 2004; Arnaud et al.,2005, 2012; Debret et al., 2010; Wilhelm et al., 2012), data fromthe NE Alps are still lacking, leaving a gap of knowledge about pastflood activity in this region. The sediment record of pre-Alpine LakeMondsee (Upper Austria) in the NE Alps has recently been intro-duced as a valuable regional archive of climatic and environmentalchanges during the last ca 15 000 years (Lauterbach et al., 2011). Inaddition, detailedmicrofacies analysis of detrital layers intercalatedin the varved Lake Mondsee sediments of the last 1600 yearsrevealed the high potential of this sediment record as a seasonallyresolved palaeoflood record (Swierczynski et al., 2012). Here, wepresent an extended flood reconstruction for the last 7100 yearsfrom the varved sediments of Lake Mondsee. The recurrencepattern of intercalated detrital layers within the Lake Mondseesediments is discussed with respect to regional climatic boundaryconditions as reflected by Alpine glacier advances (Holzhauseret al., 2005; Joerin et al., 2006; Ivy-Ochs et al., 2009) lake-levelfluctuations (Magny, 2004) and wet phases reconstructed frompeat bogs (Haas et al., 1998). Continent-wide comparison withother European flood records aims at deciphering regional varia-tions and their either climatic or archive-specific causes.

2. Study area

Lake Mondsee (surface area 14.2 km2, lake volume 5.1 km3) islocated in the NE Alps (47�490N, 13�240E) in Upper Austria at analtitude of 481 m above sea level (Fig. 1). The lake basin has beensubstantially altered by the Pleistocene glaciations (Kohl, 1998; vanHusen, 1989, 2004) and comprises a shallower northern sub-basin(up to 40 m deep) and a deeper southern sub-basin with amaximumwater depth of 68 m. The catchment area (241 km2) canbe sub-divided into two clearly distinguishable geological units,separated by a Tertiary thrust fault along the southern shoreline ofthe lake (van Husen, 1989). The gently sloped hills in the northernpart of the catchment (ca 75%) are characterized by siliciclasticsediments of the Rhenodanubic Flysch, locally overlain by

Fig. 2. Lithostratigraphy of the investigated sediment sequence from Lake Mondsee (Lithozone V in Lauterbach et al. (2011)), the abundance of detrital layers, sediment mircrofaciesand mXRF measurements for Ti, Mg and Ca (grey line: raw data in counts s�1 (cps), black line: 300 pt-lowpass filtered data).

T. Swierczynski et al. / Quaternary Science Reviews 80 (2013) 78e9080

Quaternary deposits. In contrast, the southern catchment (ca 25%)reveals a steeper relief and is dominated by the Triassic MainDolomite and Mesozoic limestones of the Northern CalcareousAlps. Three main streams (Griesler Ache, Zeller Ache, WangauerAche) drain the northern part of the catchment, whereas onlysmaller creeks (e.g. Kienbach, Klausbach) discharge into thesouthern sub-basin of the lake.

The regional climate is temperate and characterized by a meanannual air temperature of 8.7 �C and an average annual precipita-tion of w1500 mm (ZAMG, 1971e2000). The climatic effects of theNorth Atlantic Oscillation (NAO) are predominantly apparent dur-ing winter (Wanner et al., 2001; Beniston and Jungo, 2002; Castyet al., 2005). The precipitation regime of the NE Alps is mainlycontrolled by Atlantic and Mediterranean cyclonic activity(Sodemann and Zubler, 2010). Because of orographic effects,enhanced convective rainfall during July and August is the maincause for extreme floods at Lake Mondsee as confirmed by theexceptional summer rainfall events of 1954, 1959, 1997, 2002and 2013, which all produced severe floods and distinct flooddeposits in the lake sediments (Swierczynski et al., 2012). Suchextreme summer rains are predominantly caused by the influenceof Mediterranean air masses, namely the SWeNE route of low-pressure cells moving over the Golf of Genoa (Parajka et al.,2010), a circulation pattern classified as “Vb-type” weatherregime (Petrow and Merz, 2009). Extreme runoff events at LakeMondsee rarely occur after rain-on-snow events in winter and dueto snowmelt in early spring (Merz and Blöschl, 2003; Swierczynskiet al., 2012).

3. Methods

3.1. Coring and micro-facies analysis

A continuous sediment profile of about 15m length, covering theHolocene and Lateglacial, has been recovered from the southern

sub-basin of Lake Mondsee in June 2005 by using a 90-mm-diam-eter UWITEC piston corer (Lauterbach et al., 2011). This study fo-cuses on the uppermost 849 cmof the sediment record, covering thelast 7100 yr BP. The Holocene chronology of the Lake Mondseesediment record is based onmicroscopic varve counting, confirmedby additional AMS 14C dating (Lauterbach et al., 2011). In addition tovarve counting, detailed sedimentmicro-facies analysis was carriedout on a continuous set of overlapping large-scale petrographic thinsections (100 � 20 � 10 mm), prepared according to the methoddescribed by Brauer et al. (1999). Microfacies analysis, varvecounting and detrital layer thickness measurements wereundertaken under a ZEISS Axiophot polarization microscope at25e400� magnification.

3.2. Geophysical and geochemical properties

Semi-quantitative major element scanning was carried out at200 mm resolution on the impregnated sediment blocks from thinsection preparation by using a vacuum-operating EAGLE III XLmicro X-ray fluorescence (mXRF) spectrometer with a low power RhX-ray tube at 40 kV and 300 mA (250 mm spot size, 60 s countingtime, single scan line). Element intensities for Mg, Ti and Ca areexpressed as counts s�1 (cps), representing relative changes inelement composition. The approach of combining parallel microf-acies analysis and mXRF element scanning on the same sedimentblocks (Brauer et al., 2009) enables the characterization and reliabledetection of even microscopic detrital layers. Selected sedimentsamples were additionally investigated by scanning electron mi-croscopy (SEM).

3.3. Statistical analysis of detrital layer time series

Extreme floods are rare events by nature and their recurrencerate has to be modelled accordingly. Due to the length of theinvestigated dataset, inhomogeneities in flood occurrence are likely

T. Swierczynski et al. / Quaternary Science Reviews 80 (2013) 78e90 81

to be present at different scales and with unknown patterns. Tobetter estimate the occurrence of floods and to improve the com-parison with other proxy records, we analysed the detrital layerrecord with the software Matlab by applying a Gaussian kernelregression according to Nadaraya (1964) and Watson (1964) with

lðtÞ ¼ SiY*Kððt � TðiÞÞ=hÞ=SiKððt � TðiÞÞ=hÞ

(l ¼ regression estimator, K ¼ Gaussian kernel function,T(i) ¼ observed events, t ¼ time, Y ¼ flood cluster, h ¼ band width).

The applied Gaussian kernel regression is a non-parametricmethod, representing a meaningful predictor for the occurrenceof rare events. It is based on the assumption that flood occurrencefollows an inhomogeneous Poisson process, i.e. flood events areunrelated to each other. A similar modified kernel regression(Diggle, 1985) was previously used to quantify trends in floodoccurrence in German rivers during the last 1000 years (Mudelseeet al., 2003; Mudelsee, 2010). We used the original NadarayaeWatson estimator and selected a bandwidth of 30 years, which isregarded as a period of climatic significance. To estimate the sig-nificance of the flood pattern and to reduce probable errors in thereconstructed time series, confidence intervals were calculated bybootstrapping. From this envelope of simulated flood series, indi-vidual Kernel regressions were calculated. The 5% confidence in-tervals around the flood series were confined from the multiplekernel regressions of the simulated flood series by applying the t-percentile technique, which is especially used for non-stationaryPoisson processes (Cowling et al., 1996).

4. Results

4.1. Sediment microfacies and geochemistry

The investigated uppermost 849 cm of the Holocene sedimentrecord from Lake Mondsee (Lithozone V in Lauterbach et al. (2011))are composed of endogenic calcite varves, visible as lightedarkcouplets, with frequently intercalated sub-mm to cm thickbrownish detrital layers (Fig. 2). The detrital layer record starts at acore depth of 849 cm because below this depth detrital materialappears only diffuse and finely dispersed in the sediment andalmost no discrete layers have been formed.

The light layers of the varves are composed of biochemicallyprecipitated calcite, reflected by elevated Ca and low Ti and MgmXRF counts, and represent spring/summer sedimentation (Fig. 3).These calcite layers reveal sharp basal boundaries and consist oftwo or three sub-layers with well defined rhombohedral calcitecrystals (5e10 mm) at the base, followed by a layer of smallercrystals (2e5 mm) and intermediate crystals (ca 5 mm) on top. Thissub-lamination is related to several pulses of calcite precipitation(Kelts and Hsü, 1978; Koschel et al., 1983; Lotter, 1989; Ohlendorfand Sturm, 2001). Abundant pelagic diatoms (Stephanodiscus sp.,Aulacoseira sp.), organic detritus (e.g. leaves), ostracods and clasticdetrital material occur within the basal calcite sub-layer. Both, thesharp basal boundary of the calcite layers and the good preserva-tion of endogenic calcite and pelagic diatoms are indicative foranoxic conditions at the wateresediment interface, related to astratifiedwater column in spring/summer (Koschel et al., 1983). Theoccurrence of reworked littoral material in the spring layer suggestsincreased fluvial activity or wave activity that causes reworking oflittoral sediments. The dark layers of the varves, representingautumn/winter sedimentation, are predominantly composed ofclay-sized minerogenic detritus, reflected by elevated mXRF Mg andTi counts.

Two types of detrital layers, intercalated within the summer/spring calcite layers, can be distinguished. Thin detrital layers

(0.05e2.2 mm) are composed of silt- to sand-sized (<50e100 mm)angular shaped quartz, feldspars, dolomite and calcite grains,which can be easily distinguished from the endogenic calcite(<10 mm). Increased mXRF Ti and Mg counts suggest a provenancefrom both, the siliciclastic (Flysch Zone) and the dolomitic(Northern Calcareous Alps) part of the catchment, respectively.Thin detrital layers within Lake Mondsee sediments reflect springand summer flood deposits (Swierczynski et al., 2012) originatingfrom the main tributary Griesler Ache (syn.: Fuschler Ache), whichflows through both catchment geologies, the mainly siliciclasticFlysch deposits and the carbonatic Northern Calcareous Alps. Thedistance between the Griesler Ache inflow into the northern sub-basin and the coring location is w3 km (Fig. 1). The good sorting ofthe detrital material indicates inter- and overflows as the maintransport mechanism (Sturm and Matter, 1978; Mulder andAlexander, 2001), spreading in a well-stratified water body alongthe thermocline before the particles settle down to the lake bot-tom. Several short gravity cores from different locations in LakeMondsee show a basin-wide occurrence of these detrital layers,confirming the catchment-scale relevance of flood events(Swierczynski et al., 2009).

In contrast, thick detrital layers (0.65e32.0 mm) are composedof organic fragments and up to sand-sized (100e200 mm) angularshaped mineral grains (Swierczynski et al., 2012). Microscopic in-spection reveals dolomite and calcite as the main clastic compo-nents. Thick detrital layers are graded and show a sharp basalcontact to the underlying endogenic calcite layer. mXRF Mg and Cacounts are high, whereas mXRF Ti counts (representing siliciclasticcomponents) remain low (Fig. 3). Thick detrital layers are inter-preted as deposits of highly concentrated turbidity currents/underflows (Mulder and Alexander, 2001), representing the sub-aqueous propagation of debris flows from the catchment (Gomiet al., 2004). The sharp erosional basal contacts to the underlyingcalcite layers and the continuation of calcite precipitation above theevent layers indicate instantaneous deposition in summer. Mgenrichment as well as the spatial distribution of these detritallayers within the southern sub-basin reveal sediment supply fromthe Northern Calcareous Alps, induced by local debris flow eventsthrough the Kienbach creek close to the coring site (Swierczynskiet al., 2012).

4.2. Chronology

The chronology for the investigated part of the Lake Mondseesediment record is based on microscopic varve counting (Fig. 4A).The annual character of the calcite laminae is confirmed bysediment microfacies analysis and 137Cs dating (Lauterbach et al.,2011) as well as a detrital event layer related to a documenteddebris flow in summer 1986 (Swierczynski et al., 2009) interca-lated within the 1986 calcite layer. The varve-based age-depthmodel is further supported by 20 AMS 14C dates in the intervalunder investigation (Lauterbach et al., 2011). An additional errorestimate has been derived from two independent varve countscarried out for the last 4000 years by different examiners. Despitefew intervals of increased deviation, the total difference of lessthan ten varves between the two counts (Fig. 4B) confirms thepreviously estimated maximum error of 1.25% for the upper 1600years of the record (Swierczynski et al., 2012), thus providing arobust chronology for the event layer record. Differences in thetwo varve counts are on the one hand due to seasonal calcite sub-layers, which could be mistaken as individual years, thus leadingto an overestimation of the number of varves. On the other hand,occasional massive calcite layers or calcite mud with only weaklayering might lead to an underestimation of the time included.Varve thickness during the late Holocene varies between 1.0 and

Fig. 3. Thin section images and mXRF scanner data of selected sediment intervals from different coring depths including intercalated detrital layer deposition (DL) (AeE). Summersub-layers (Su) are composed of autochthonous calcite (cc) as reflected by high Ca counts, whereas winter sub-layers (Wi) are composed of organic and minerogenic detritus asreflected by higher Ti counts. Turbidites (detrital layer type 2) show dolomitic (do) and calcitic (ca) composition with increased Mg counts. (F) High Ti and low Mg countscharacterize detrital layers type 1, whereas high Mg and low Ti counts characterize detrital layers type 2.

T. Swierczynski et al. / Quaternary Science Reviews 80 (2013) 78e9082

Fig. 4. (A) Age depth model for the investigated sequence of the Lake Mondsee sediment record based on two varve counts and confirmed by AMS 14C dates (for details onradiocarbon dating see Lauterbach et al. (2011)). Sedimentation rates are given as 10 cm averages. (B) Differences between the two independent varve counts. Highest differencesbetween the two varve counts occur within faintly laminated intervals.

T. Swierczynski et al. / Quaternary Science Reviews 80 (2013) 78e90 83

2.5 mm (Fig. 4A) with a mean sedimentation rate of about1.5 mm/a.

4.3. Detrital layer record and reconstruction of flood episodes

The sediment record covering 7096 vyr BP comprises 271 floodand 47 debris flow layers (Fig. 5). The distinct annual layering inthe uppermost 2000 years of the record allows allocating >95% ofthe detrital layers to the spring/summer season based on theirmicro-stratigraphic position within the varve succession. Based on

these data we suggest a similar seasonal pattern also for the lowerpart of the record although less clear varve preservation pre-vented to prove this assumption. Both types of event depositsreveal a pronounced hydrological variability on decadal, centen-nial and millennial time scales. Flood occurance is increasedduring the last 1500 vyr BP (2e15 events per 100 years) but floodsare less abundant prior to ca 1500 vyr BP (1e9 events per 100years) and even less prior to 3300 vyr BP (1e5 events per 100years). On shorter time scales, enhanced flood layer depositionoccurred in 18 characteristic periods of mainly 30e50 years

Fig. 5. Reconstruction of flood and debris flow events from the Lake Mondsee sediments: (A) Estimation of the flood occurrence rate by Gaussian kernel regression (30-yr bandwidth, grey shading: 5% confidence intervals based on student’s t distribution of 2000 bootstrappings results.) (B) 31-year running mean of flood events indicating flood recurrencetimes. (C) Blue columns show flood activity for 100-year intervals. Black columns show flood events for 30-year intervals. Light blue and numbered (1e18) bars denote floodepisodes (FE) as defined in the text. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

T. Swierczynski et al. / Quaternary Science Reviews 80 (2013) 78e9084

duration (Fig. 5), hereafter labelled as flood episodes (FE). Theseflood episodes are inferred from the 31-yr running mean of floodevents and are dated at 7100e7050 (FE 18), 6470e6270 (FE 17),6170e6120 (FE 16), 5850e5800 (FE 15), 5410e5360 (FE 14), 5110e5040 (FE 13), 4900e4850 (FE 12), 4670e4620 (FE 11), 4500e4450(FE 10), 3300e3250 (FE 9), 2800e2750 (FE 8), 2050e2000 (FE 7),1500e1470 (FE 6), 1360e1310 (FE 5), 1250e1200 (FE 4), 810e780(FE 3), 650e620 (FE 2), 470e430 (FE 1) vyr BP. A slightly higherflood recurrence of 3e5 years within these flood episodes isobserved for FE 1e6 (Swierczynski et al., 2012), whereas floodepisodes FE 7e9 exhibit 5e7 years of flood recurrence and FE 10e

18 exhibit 10e14 years of flood recurrence (cf. Swierczynski et al.,2013). Based on the Kernel regression slightly lower flood re-currences are calculated for FE 1e6 (6e10 years), FE 7e9 (10e12years) and FE 10e18 (17e25 years). Relatively narrow bandsfor the t-percentiles of the bootstrapping results confirm thatthese occurrence rates are robust and representative (Fig. 5).Compared to the flood events, debris flows originating from theKienbach creek on the southern lake shore after local thunder-storms are rare events. Such thunderstorms are not necessarilycausing major regional floods as observed, for example, for the AD1986 event.

Fig. 6. Comparison of proxy-based European flood reconstructions. (A) Kernel regression of the Lake Mondsee flood record with flood episodes (FE 1 to FE 18 indicated by light bluecolumns). (B) Flood layer thickness from the Lake Ammersee sediment record (Czymzik et al., 2013). (c) Flood reconstruction from Lake Le Bourget sediments (northwestern Alps)with 13 clastic phases (Arnaud et al., 2012). (D) Flood phases reconstructed from deposits of Upper Vistula River, Poland (Starkel et al., 2006). (E) Flood reconstruction from lakesediments of Lake Meringsdalsvatnet, southern Norway (Støren et al., 2012). (For interpretation of the references to colour in this figure legend, the reader is referred to the webversion of this article.)

T. Swierczynski et al. / Quaternary Science Reviews 80 (2013) 78e90 85

5. Discussion

It has been demonstrated that the deposition of detrital eventlayers in Lake Mondsee is triggered by extreme rainfall events andrelated surface runoff in spring/summer mainly from the Medi-terranean Sea (Swierczynski et al., 2012). Flood layer deposition iscorroborated by instrumental flood series as well as historical data(Swierczynski et al., 2012, repository data), which reveal a maininfluence of Mediterranean moisture inflow for floods transportingmaterial from both catchments, the northern and the southern lakecatchment.

5.1. Testing human impacts on detrital layer formation in LakeMondsee

In addition to climatic factors, catchment erosion caused byhuman activity can increase detrital sediment transport processes

and supply to the lake (Zolitschka et al., 2003; Brown et al., 2009;Houben et al., 2009; Hoffmann et al., 2010; Lauterbach et al.,2012) and thus might bias flood reconstructions. Therefore, eachsediment record must be tested for eventual human impacts ondetrital matter flux before climatic interpretation. Since detritalmatter in the Lake Mondsee sediments does not show a clear in-crease during the last 7100 years (Fig. 7) we assume that our recordof flood frequencies is not significantly influenced by human ac-tivities. Similarly low impact of human activities on flooding hasbeen reported also from other Alpine lake settings as, for example,Lake Ammersee (Czymzik et al., 2013) and Lake Bourget (Debretet al., 2010) (Fig. 6).

More detailed information about human impact in the catch-ment of Lake Mondsee (forest clearance and agricultural activity)has been derived from two pollen records, one from Lake Egelsee, asmall lake located only 100 m south of the southern lake shore(Schmidt, 1981) and another one from the Moosalm peatbog

Fig. 7. Comparison of the Lake Mondsee flood record (A) with other palaeoclimate records (AeC). (A) Flood occurrence rate reconstruction from the Lake Mondsee sediments (seeFig. 5). Diamonds indicate debris flows. (B) Mg counts as a proxy for dolomitic detrital input from the Northern Calcareous Alps. (C) Ti counts as a proxy for siliciclastic detrital inputfrom the Flysch Zone, (D) Lake-level high stands in the southwestern Alps (Magny, 2004), (E) Advances of Grosser Aletsch glacier (Holzhauser et al., 2005) and main Alpine glacieradvances during the Little Ice Age (LIA), Göschenen I and II (G1 and G2), Löbben (L), Rotmoos I and II (R1 and R2) and Frosnitz (F) cold oscillations (Ivy-Ochs et al., 2009),(F) Reconstruction of tree line in the Kauner valley (Central-Eastern Alps) gaps reflect uncertainties possibly due to human impact (Nicolussi et al., 2005). Please note the reversedelevation scale.

T. Swierczynski et al. / Quaternary Science Reviews 80 (2013) 78e9086

ca 5 km to the south (Draxler, 1977). First, but still only weak signsof human impact on vegetation are already visible for Neolithictimes (ca 5800e4700 cal. yr BP) when lake dwellings were builtaround Lake Mondsee (Schmidt, 1986). Periods of slightly increasedhuman impact are observed for the Pre-Roman Iron Age and thetimes when the monastery was founded in AD 748 (1202 cal yr BP),but intense forest clearance and related strong human impacts oncatchment vegetation only started around AD 1600 (350 cal. yr BP)(Draxler, 1977; Kunze, 1986). Most of these periods of anthropo-genic changes in the regional vegetation cover do not coincide withany of the flood episodes except the period around 2800 vyr BPwhich falls together with FE 8. We cannot exclude that humanimpact amplified flooding around that time, however, since none ofthe other intervals of human impact resulted in an increase in floodfrequencies this might also be pure coincidence. Furthermore, it

appears that forest clearance during times of monastery foundationoccurred directly after a series of three flood episodes (FE 4e6)commencing at around 1500 vyr BP (AD 400) and that the mainperiod of forest clearing began without any doubt at the end ofthree flood episodes with highest flood recurrence (FE 1e3). Notsurprisingly, also the early flood episodes FE 16e18 occurred beforethe onset of Neolithic settlement and thus cannot have a causalrelationship (Swierczynski et al., 2013).

Comparing our flood frequency record with historical informa-tion reveals also FE 3 as a candidate for potential overprinting byhuman activity. FE 3 (810e780 vyr BP, 1140e1170 AD) (Fig. 6) co-incides with a time of mill constructions along the Griesler Ache(Kunze, 1986) which acts as the main tributary transporting sus-pended sediment during regional-scale floods into Lake Mondsee.It cannot be excluded that the initial phase of mill construction

T. Swierczynski et al. / Quaternary Science Reviews 80 (2013) 78e90 87

resulted in increased erosion near the riverside resulting in highersediment transport into the lake. In summary, we exclude ananthropogenic bias of the record of flood frequencies with theconfinement that FE 3 and FE 8 might be slightly influenced byhuman activity in the catchment.

5.2. Long-term changes in flooding

The presented flood time series from Lake Mondsee does notaccidentally start at 7100 vyr BP. Before that time almost no discretedetrital layers were deposited. Instead, detrital matter is finelydispersed in the sediments, preventing from reliable flood recon-struction. Therefore, we do not have information about floodingduring the early Holocene. From7100 vyr BP until the late Holocenerecurrence times of summer and spring floods in the Mondsee re-gion show a clear increase. This increase, however, is not expressedas a gradual trend but as stepwise increase in the occurrence ofdecadal-scale intervals labelled as flood episodes of significantlyhigher recurrence times of flooding (Figs. 5 and 6). In addition tothe number of these flood episodes, also the recurrence times offloods increased (Fig. 5). The main shift towards intensified floodepisodes in the Lake Mondsee record occurred at 1500 vyr BP withFE 6 (Fig. 5), but the first flood episode with flood recurrence timeshigher than every 10 years (FE 9) already occurred at 3300 vyr BP(Fig. 5). However, from 3300 until 1500 vyr BP flood episodes withsuch recurrence rates remained rare and only twomore (FE 8 and 7)occurred. A late Holocene flood increase has been recorded inmanyother European flood records. However, this shift did not occursynchronously in different regions but shows clear time lags. Onthewesternmargin of the continent, enhanced flooding is observedsince about 2800 cal. yr BP in England, Spain and the Northern andNW Alps (Thorndycraft and Benito, 2006; Debret et al., 2010;Macklin et al., 2010; Arnaud et al., 2012; Czymzik et al., 2013) andslightly later since about 2500 cal. yr BP in Norway (Støren et al.,2010). In more continental regions, the onset of increased flood-ing activity apparently occurred later, e.g. around 2000 cal. yr BP inPoland (Starkel et al., 2006) and around 1500 vyr BP at LakeMondsee. Despite these differences in timing, the late Holoceneincrease in hydro-climatic extreme events is most likely related tothe northern hemisphere climate cooling of the so-called Neo-glacial (Wanner et al., 2008), which is explained by reduced solarinsolation during northern hemisphere summer caused by changesin the Earth’s orbit around the sun. This trend towards a coolerlate Holocene is well expressed in the Alps by expanding glaciers(Ivy-Ochs et al., 2009) and lowering of tree lines (Nicolussi et al.,2005). The observed differences in the timing of the onset ofenhanced flood activity across Europe thus suggests a complex andregionally varying reorganization of atmospheric circulationpattern in response to the long-term decrease in summer insola-tion. However, in addition to climatic causes for the time lags, anumber of other effects related to the sensitivity of various sedi-ment archives and proxies to record floods, should be considered.

Alluvial deposits record floods in a rather different way thanlake systems because mainly high-magnitude floods are reflectedby sediment deposition in the flood plain and overtop older floodevents. Since strong floods may cause erosion of previous flooddeposits (Gilli et al., 2013), the established flood recordsmight evenbe discontinuous. In contrast, lakes form sediment traps thatpotentially allow continuous recording of flood deposits. However,the threshold in river discharge during floods that results in thedeposition of discrete detrital layers is depending on localmorphological and hydrological characteristics including catch-ment size and relief, lake shape and depth, position of the riverinflow and position of the sediment core related to river inflow.This is demonstrated by comparing the high-resolution flood

records from lakes Mondsee and Ammersee (Czymzik et al., 2010,2013) located 170 km northwest (Fig. 6). Due to different catch-ment morphologies and daily river discharge the number of flood-triggered detrital layers in both lakes recorded for the last 450 yearsdistinctly differs: whereas 28 flood layers are formed in LakeMondsee (ca 5% of these layers are thicker than 1 mm), 97 floodlayers have been deposited in Lake Ammersee (Czymzik et al., 2010)of which 35% are thicker than 1 mm. The higher number of floodlayers formed in Lake Ammersee is related to the larger catchment(Ammersee: 990 km2; Mondsee: 241 km2) the higher relief energy(Ammersee 1652 m; Mondsee ca 1300 m) and the stronger riverrunoff (mean daily River Ammer discharge: 16 m3 s�1, max. dailydischarge: 650 m3 s�1; mean daily discharge of Griesler Ache:4 m3 s�1; max. daily discharge: 82 m3 s�1). In addition, the corefrom Lake Ammersee is located in the direct prolongation of theinflowing River Ammerwhereas in LakeMondsee the core has beentaken in a more distal position to the main river inflow due to thespecific basin morphometry with two sub-basins (Fig. 1).

Other factors potentially influencing palaeoflood re-constructions are the seasonality and the predominant type offloods. The well-known regional variations in main flood seasons ofrecent flooding across Europe (Merz and Blöschl, 2003) are alsoobserved in palaoflood records from lake sediments. While lakes inthe NW Alps record either meso-scale precipitation events afteradvective moisture entries from the North Atlantic in autumn andwinter (Debret et al., 2010) or summer flash floods (Wilhelm et al.,2012), the pre-Alpine lakes in the Northern Alps record long-rainfloods (Czymzik et al., 2010) or convective short-rain floods inspring and summer (Swierczynski et al., 2012).

Even if disentangling climatic and archive effects is still difficult,it is likely that the observed regional differences of the onset of lateHolocene intensification of flood activity in Europe are at leastpartly related to regional differences in the reorganization of theatmospheric circulation pattern in response to the long-terminsolation change. Even today the precipitation regime indifferent regions of Europe is controlled by different atmosphericmodes: while the influence of moisture transport from the NorthAtlantic is dominant in northern (Bøe et al., 2006; Støren et al.,2012) and western Europe (Debret et al., 2010; Wilhelm et al.,2012), central and eastern Europe are additionally influenced bycontinental air masses from Siberia (Starkel et al., 2006) and/or theMediterranean moisture regime (Swierczynski et al., 2012).

5.3. Multi-decadal scale variability

As discussed before, the observed long-term increase in floodactivity at Lake Mondsee is superimposed by a multi-decadalvariability. This is expressed in 18 flood episodes, each lasting be-tween 30 and 50 years and showing increasing recurrence timesfrom 10e14 years (FE 10e18) and 5e7 years (FE 7e9) to 3e5 years(FE 1e6). Multi-decadal oscillations of flood activity have also beenreported from the varved sediments of Lake Ammersee (Czymziket al., 2010) and Lake Ledro (Wirth et al., 2013) covering the last450 and 2000 years, respectively. In contrast, comparing decadal-scale variability with long non-varved flood records (e.g. Wessels,1998; Starkel et al., 2006; Debret et al., 2010; Macklin et al., 2010;Arnaud et al., 2012) is not straightforward due to the lowertemporal-resolution of these deposits. Non-varved lake records,like for example Lake Bourget, often exhibit centennial-scale floodphases (Fig. 6; Arnaud et al., 2012), which are not observed in theLake Mondsee record. This might be not solely due to the lowerdating resolution but also caused by the different type of proxyused for flood reconstructions. In the Lake Mondsee sediment re-cord discrete detrital layers are detected, whereas the Lake Bourgetflood record is based on continuous m-XRF scanning and, in

T. Swierczynski et al. / Quaternary Science Reviews 80 (2013) 78e9088

particular, concentrations in titanium as a proxy for terrigeneousmineral matter (Fig. 6). While discrete detrital layers solely reflectindividual hydro-meteorological events, total terrigeneous influx isnot necessarily transported only by floods into the lake but mightalso reflect generally wetter periods with higher surface runoff.Therefore, a perfect match of these different flood records cannotbe expected. Nevertheless, for some time intervals like the last1000 years the multi-centennial flood phases in the NW Alps(Debret et al., 2010; Arnaud et al., 2012; Wilhelm et al., 2013)appear quite similar to the multi-decadal flood episodes found atLake Mondsee (Fig. 6). In contrast, flood episodes prior to3300 vyr BP (FE 10e18) are not corresponding to the Lake Bourgetflood record which, however, might also be due to the fact thatthese flood episodes are not very prominent in the Lake Mondseerecord. Due to the different proxies used for flood reconstructionand the different dating resolution in the Lake Mondsee record andthe lakes in NW Alps, it is difficult to say to what extend theobserved short-term variability coincide within the region or not.However, the comparisonwith the Lake Ammersee record, which isbased on the same proxies and dating like Lake Mondsee (Fig. 6;Czymzik et al., 2013), suggests distinct regional differences even indecadal-scale variations. This is likely related to the fact thatflooding in each region is controlled by different large-scaleweather regimes and atmospheric circulation pattern.

5.4. The flood record in relation to climatic development in the Alps

Most of the late Holocene flood episodes in the late Holocenesince ca 3000 vyr BP at Lake Mondsee occurred during intervals ofgenerally colder and wetter climate conditions in the Alpine regionas reflected by glacier advances (Holzhauser et al., 2005; Ivy-Ochset al., 2009), lowered tree lines (Nicolussi et al., 2005) and lake-level high stands (Magny, 2004) (Fig. 7). Flood episode 8 (FE 8)developed at the onset of the Göschenen I oscillation (Zoller et al.,1966), while FE 4e6 and FE 2e3 occurred during the Göschenen II(Zoller, 1960; Zoller et al., 1966) and the beginning of the Little IceAge (Holzhauser et al., 2005; Ivy-Ochs et al., 2009), respectively.Two flood episodes, FE 2 (430e470 vyr BP) and FE 7 (2000e2050vyr BP) do not coincide with main glacier advances in the CentralAlps, but with lower tree lines in the Austrian Alps suggestingcooler summer temperatures (Fig. 7; Nicolussi et al., 2005). Duringthese periods winters might have been too dry to lead to majorglacier advances.

Prior to 3300 yr BP it is hardly possible to discuss coincidencesbetween flood episodes and glacier advances, partly due to thelower amplitudes and less clear structure of the flood episodes butpossibly also due to dating issues of the glacier advances. For thisperiod we observe weak coincidence between some flood episodes(FE 16, 13, 12, 10) with periods of lower tree lines (Nicolussi et al.,2005), however, others (FE 17, 15, 14 11) also occur during timesof higher tree lines and thus warmer summer temperatures (Fig. 7).Apparently, no clear relation between both proxy records exists. FE9 is an important flood episode since it is the first, which showshigher recurrence rates than all previous flood episodes. Interest-ingly, it occurred about a century later than the Löbben glacieradvance (Patzelt and Bortenschlager, 1973; Patzelt, 1977; Fig. 7). Wemay speculate about a delayed response of flooding to climatecooling as expressed by glacier advances whereas later flood epi-sodes tend to appear during the early phase of glacier advances.

Although comparing decadal-scale flood episodes withcentennial-scale glacier advances and tree line oscillations is notstraightforward at high time resolution, we still can state that thelong-term increasing trend in the occurrence and amplitudes ofperiods of intensified spring and summer flooding generally fallstogether with more frequent and stronger glacier advances as well

as lowering of tree lines in the Alps. This suggests that changes inatmospheric circulation, leading to an increase in flood-proneweather regimes in the late Holocene, including might have beencaused by the same mechanism that caused the trend towardscooler andwetter climate during this time (Wanner et al., 2008), i.e.the reduction in northern hemisphere summer insolation. How-ever, the lack of coincidence of some flood episodes with temper-ature proxies also suggests that there is no simple relationshipbetween temperature and flood occurrence. Instead, we have toconsider a number of other factors controlling flooding, includingregional and local climate conditions. Last but not least, despitesome major advances in the last years, palaeoflood reconstructionsfrom sediment archives still comprise some uncertainties mainlyrelated to the understanding of proxy responses. In particular, abetter knowledge of the entire chain of processes from the type andseason of floods, their impact on erosion and resulting transportand deposition of suspended sediment remains elusive.

6. Conclusions

The varved sediments from Lake Mondsee provide a high-resolution record of extreme summer/spring runoff events (floodsand debris flows) during the last 7100 vyr BP in the NE Alps. In total,eighteen flood episodes, i.e. 30e50 year long periods of increasedflood recurrence, occurred during this time interval. At a long-termperspective, the number of flood episodes and their amplitude interms of flood recurrence intervals increased from the mid- to thelate Holocene, likely in response to decreasing northern hemi-sphere summer insolation. The increase in flood frequencies,however, was not gradual but occurred stepwise. The main step ofincreasing spring/summer flood activity is observed at about1500 vyr BP, thus apparently occurred later in the Lake Mondseeregion than in western and NW Europe. The asynchronous onset ofincreased flood activity in different regions might partly reflecthydro-meteorological contrasts due to spatially variable influencesof Mediterranean, Atlantic and Siberian air masses across Europe.However, threshold effects due to the different sensitivity in floodrecording of different sediment archives and the seasonality offloods might have also played a role.

Despite the coincidence of mid-to late Holocene cooling and theincrease in flood activity, the relation between temperature andflood episodes at decadal scales is complex and still remainselusive. Hence, long and highly resolved palaeoflood records fromother regions need to be studied. There is still a lack of knowledgehow flood types and their seasonality but also the sensitivity ofindividual sediment archives and proxies influence the recon-struction of long flood records. The establishment of a densenetwork of palaeoflood records will shed further light on thespatio-temporal variations of flood activity and their possiblecauses and thereby improve the anticipation of future hydro-meteorological extremes under a changing climate.

Acknowledgements

This is a contribution to BMBF (FederalMinistry of Education andResearch), Project PROGRESS (A3, Extreme events in Geoarchives;03IS2191G). Tina Swierczynski obtained a Ph.D. fellowship fundedby the German Research Centre for Geosciences (GFZ). Lake coringwas carried out within the ESF EuroCLIMATE Project DecLakes (04-ECLIM-FP29), funded through DFG grants BR2208/2-2 and AN 554/1-2. We thank Richard Niederreiter (UWITEC, Mondsee), JohannKnoll (Institute for Limnology, Mondsee) and the DecLakes teammembers for assistance during the coring campaigns. Dieter Berger,Gabriele Arnold and Michael Köhler (GFZ, Potsdam) are acknowl-edged for preparing thin sections and Juliane Herwig for SEM

T. Swierczynski et al. / Quaternary Science Reviews 80 (2013) 78e90 89

analysis. We are also grateful to the Institute for Water Ecology,Fisheries and Lake Research in Scharfling, particularly Albert Jagsch,and the Institute for Limnology in Mondsee for providing logisticalsupport during the coring campaigns. We thank two anonymousreviewers for their detailed comments on the previous version ofour manuscript. The data presented in the figures are available atdoi: 10.5880/GFZ.5.2.2012.003 and via the PANGAEA database:http://doi.pangaea.de/10.1594/PANGAEA.818922.

References

Arnaud, F., Revel, M., Chapron, E., Desmet, M., Tribovillard, N., 2005. 7200 years ofRhone river flooding activity in Lake Le Bourget, France: a high-resolutionsediment record of NW Alps hydrology. The Holocene 15, 420e428.

Arnaud, F., Révillon, S., Debret, M., Revel, M., Charon, E., Giguet-Covex, C.,Poulenard, J., Magny, M., 2012. Lake Bourget regional erosion patterns recon-struction reveals Holocene NW European Alps soil evolution and paleohy-drology. Quat. Sci. Rev. 51, 81e92.

Arnell, N.W., 1999. The effect of climate change on hydrological regimes in Europe:a continental perspective. Global Environ. Change 9, 5e23.

Auer, I., Böhm, R., Jurkovic, A., Lipa, W., Orlik, A., Potzmann, R., Schöner, W.,Ungersböck, M., Matulla, C., Briffa, K., Jones, P., Efthymiadis, D., Brunetti, M.,Nanni, T., Maugeri, M., Mestre, L.M.O., Moisselin, J.-M., Begert, M., Müller-Westermeier, G., Kveton, V., Bochnicek, O., Stastny, P., Lapin, M., Szalai, S.,Szentimrey, T., Cegnar, T., Dolinar, M., Gajic-Capka, M., Zaninovic, K.,Majstorovic, Z., Nieplova, E., 2007. HISTALP e historical instrumental climato-logical surface time series of the Greater Alpine Region. Int. J. Climatol. 27,17e46.

Baker, V.R., 1987. Paleoflood hydrology and extraordinary flood events. J. Hydrol. 96,79e99.

Baker, V.R., 2008. Paleoflood hydrology: origin, progress, prospects. Geomorphology101, 1e13.

Beniston, M., Jungo, P., 2002. Shifts in the distributions of pressure, temperature andmoisture and changes in the typical weather patterns in the Alpine region inresponse to the behavior of the North Atlantic Oscillation. Theor. Appl. Clim. 71,29e42.

Beurton, S., Thieken, A.H., 2009. Seasonality of floods in Germany. Hydrol. Sci. J. 54,62e76.

Bøe, A.-G., Dahl, S.O., Lie, O., Nesje, A., 2006. Holocene river floods in the upperGlomma catchment, southern Norway: a high-resolution multiproxy recordfrom lacustrine sediments. The Holocene 16, 445e455.

Brauer, A., Dulski, P., Mangili, C., Mingram, J., Liu, J., 2009. The potential of varves inhigh-resolution paleolimnological studies. PAGES 17, 96e98.

Brauer, A., Endres, C., Negendank, J.F.W., 1999. Lateglacial calendar year chronologybased on annually laminated sediments from Lake Meerfelder Maar, Germany.Quat. Int. 61, 17e25.

Brázdil, R., Kundzewicz, Z.W., Benito, G., 2006. Historical hydrology for studyingflood risk in Europe. Hydrol. Sci. J. 51, 739e764.

Brown, A.G., Carey, C., Erkens, G., Fuchs, M., Hoffmann, T., Macaire, J.-J.,Moldenhauer, K.-M., Walling, D.E., 2009. From sedimentary records to sedimentbudgets: multiple approaches to catchment sediment flux. Geomorphology108, 35e47.

Brown, S.L., Bierman, P.R., Lini, A., Southon, J., 2000. 10 000 yr record of extremehydrologic events. Geology 28, 335e338.

Casty, C., Wanner, H., Luterbacher, J., Esper, J., Böhm, R., 2005. Temperature andprecipitation variability in the European Alps since 1500. Int. J. Climatol. 25,1855e1880.

Chapron, E., Arnaud, F., Noël, H., Revel, M., Desmet, M., Perdereau, L., 2005. RhoneRiver flood deposits in Lake Le Bourget: a proxy for Holocene environmentalchanges in the NW Alps, France. Boreas 34, 404e416.

Chapron, E., Desmet, M., De Putter, T., Loutre, M.F., Beck, C., Deconinck, J.F., 2002.Climatic variability in the northwestern Alps, France, as evidenced by 600 yearsof terrigenous sedimentation in Lake Le Bourget. The Holocene 12, 177e185.

Chiverrell, R.C., Harvey, A.M., Hunter, S.Y., Millington, J., Richardson, N.J., 2008. LateHolocene environmental change in the Howgill Fells, Northwest England.Geomorphology 100, 41e69.

Christensen, J.H., Christensen, O.B., 2003. Climate modelling: severe summertimeflooding in Europe. Nature 421, 805e806.

Corella, J., Moreno, A., Morellón, M., Rull, V., Giralt, S., Rico, M., Pérez-Sanz, A.,Valero-Garcés, B., 2011. Climate and human impact on a meromictic lake duringthe last 6,000 years (Montcortès Lake, Central Pyrenees, Spain). J. Paleolimnol.46, 351e367.

Cowling, A., Hall, P., Phillips, M.j, 1996. Bootstrap confidence regions for the in-tensity of a poisson point process. J. Am. Stat. Assoc. 91, 1516e1524.

Czymzik, M., Dulski, P., Plessen, B., von Grafenstein, U., Naumann, R., Brauer, A.,2010. A 450 year record of springesummer flood layers in annually laminatedsediments from Lake Ammersee (southern Germany). Water Resour. Res. 46.http://dx.doi.org/10.1029/2009WR008360.

Czymzik, M., Brauer, A., Dulski, P., Plessen, B., Naumann, R., von Grafenstein, U.,Scheffler, R., 2013. Orbital and solar forcing of shifts in Mid- to Late Holoceneflood intensity from varved sediments of pre-alpine Lake Ammersee (southernGermany). Quat. Sci. Rev. 61, 96e110.

Dearing, J., Hu, Y., Doody, P., James, P.A., Brauer, A., 2001. Preliminary reconstructionof sediment-source linkages for the past 6000 yr at the Petit Lac d’Annecy,France, based on mineral magnetic data. J. Paleolimnol. 25, 245e258.

Debret, M., Chapron, E., Desmet, M., Rolland-Revel, M., Magand, O., Trentesaux, A.,Bout-Roumazeille, V., Nomade, J., Arnaud, F., 2010. North western Alps Holo-cene paleohydrology recorded by flooding activity in Lake Le Bourget, France.Quat. Sci. Rev. 29, 2185e2200.

Delgado, J.M., Apel, H., Merz, B., 2010. Flood trends and variability in the MekongRiver. Hydrol. Earth Syst. Sci. 14, 407e418.

Diggle, P., 1985. A Kernel Method for smoothing point process data. Appl. Stat. 34,138e147.

Draxler, I., 1977. Pollenanalytische Untersuchungen von Mooren zur spät- undpostglazialen Vegetationsgeschichte im Einzugsgebiet der Traun. Jb. Geol. B-A120, 131e163.

Ely, L.L., Enzel, Y., Baker, V.R., Cayan, D.R., 1993. A 5000-year record of extremefloods and climate change in the southwestern United States. Science 262, 410e412.

Enzel, Y., Wells, S.G., 1997. Extracting Holocene paleohydrology and paleoclima-tology information from modern extreme flood events: an example fromsouthern California. Geomorphology 19, 203e226.

Gilli, A., Anselmetti, F., Ariztegui, D., McKenzie, J.A., 2003. A 600-year sedimentaryrecord of flood events from two sub-alpine lakes (Schwendiseen, NortheasternSwitzerland). Ecologae Geol. Helv. 96, 49e58.

Gilli, A., Anselmetti, F.S., Glur, L., Wirth, S.B., 2013. Lake sediments as archives ofrecurrence rates and intensities of past flood events. In: Schneuwly-Bollschweiler, M., Stoffel, M., Rudolf-Miklau, F. (Eds.), Dating Torential Processeson Fans and Cones. Springer, Netherlands, pp. 225e242.

Glaser, R., Riemann, D., Schönbein, J., 2010. The variability of European floods sinceAD 1500. Clim. Change 101, 235e256.

Gomi, T., Sidle, R.C., Swanston, D.N., 2004. Hydrogeomorphic linkages of sedimenttransport in headwater streams, Maybeso Experimental Forest, southeastAlaska. Hydrobiol. Process. 18, 667e683.

Haas, J.N., Richoz, I., Turner, W., Wick, L., 1998. Synchronous Holocene climaticoscillations recorded on the Swiss Plateau and at timberline in the Alps. TheHolocene 8, 301e309.

Hirschboeck, K., 1988. Flood Hydroclimatology. John Wiley & Sons, New York,pp. 27e49.

Hoffmann, T., Thorndycraft, V.R., Brown, A.G., Coulthard, T.J., Damnati, B., Kale, V.S.,Middelkoop, H., Notebaert, B., Walling, D.E., 2010. Human impact on fluvialregimes and sediment flux during the Holocene: review and future researchagenda. Global Planet. Change 72, 87e98.

Holzhauser, H., Magny, M., Zumbühl, H.J., 2005. Glacier and lake-level varia-tions in west-central Europe over the last 3500 years. The Holocene 15,789e801.

Houben, P., Wunderlich, J.R., Schrott, L., 2009. Climate and long-term human impacton sediment fluxes in watershed systems. Geomorphology 108, 1e7.

Ivy-Ochs, S., Kerschner, H., Maisch, M., Christl, M., Kubik, P.W., Schlüchter, C., 2009.Latest Pleistocene and Holocene glacier variations in the European Alps. Quat.Sci. Rev. 21e22, 2137e2149.

Joerin, U.E., Stocker, T.F., Schlüchter, C., 2006. Multicentury glacier fluctuations inthe Swiss Alps during the Holocene. The Holocene 16, 697e704.

Kelts, K., Hsü, K., 1978. Freshwater carbonate sedimentation. In: Lerman, A. (Ed.),Lakes: Physics, Chemistry, Geology. Springer, New York, pp. 295e323.

Knox, J.C., 2000. Sensitivity of modern and Holocene floods to climate change. Quat.Sci. Rev. 19, 439e457.

Kohl, H., 1998. Das Eiszeitalter in Oberösterreich Teil II: Die eiszeitliche Verglet-scherung in Oberösterreich. Jahrbuch des Oberösterreichischen Museal- ver-eins, Linz, pp. 175e390.

Koschel, R., Brenndorf, J., Proft, G., Recknagel, R., 1983. Calcite precipitation as anatural mechanism of eutrophication. Arch. Hydrobiol. 98, 380e408.

Kundzewicz, Z., Radziejewski, M., Pinskwar, I., 2006. Precipitation extremes in thechanging climate. Clim. Res., 51e56.

Kundzewicz, Z.W., Ulbrich, U., Brücher, T., Graczyk, D., Krüger, A., Leckebusch, G.C.,Menzel, L., Pi�nskwar, I., Radziejewski, M., Szwed, M., 2005. Summer floods incentral Europe e climate change track? Nat. Hazard. 36, 165e189.

Kunze, W., 1986. Mondsee e 5000 Jahre Geschichte und Kultur. Selbstverlag derGemeinde, Mondsee.

Lauterbach, S., Brauer, A., Andersen, N., Danielopol, D.L., Dulski, P., Hüls, M.,Milecka, K., Namiotko, T., Obremska, M., von Grafenstein, U., Declakes partici-pants, 2011. Environmental responses to Lateglacial climatic fluctuationsrecorded in the sediments of pre-Alpine Lake Mondsee (northeastern Alps).J. Quat. Sci. 26, 253e267.

Lauterbach, S., Chapron, E., Brauer, A., Hüls, M., Gilli, A., Arnaud, F., Piccin, A.,Nomade, J., Desmet, M., von Grafenstein, U., DecLakes participants, 2012.A sedimentary record of Holocene surface runoff events and earthquake activityfrom Lake Iseo (Southern Alps, Italy). The Holocene 22, 749e760.

Lotter, A.F., 1989. Evidence of annual layering in Holocene sediments of Soppensee,Switzerland. Aquat. Sci. 51, 19e30.

Macklin, M.G., Jones, A.F., Lewin, J., 2010. River response to rapid Holocene envi-ronmental change: evidence and explanation in British catchments. Quat. Sci.Rev. 29, 1555e1576.

Magny, M., 2004. Holocene climate variability as reflected by mid-European lake-level fluctuations and its probable impact on prehistoric human settlements.Quat. Int. 113, 65e79.

T. Swierczynski et al. / Quaternary Science Reviews 80 (2013) 78e9090

Mangili, C., Brauer, A., Moscariello, A., Naumann, R., 2005. Microfacies of detritalevent layers deposited in Quaternary varved lake sediments of the Piànico-Sèllere Basin (northern Italy). Sedimentology 52, 927e943.

Merz, B., Blöschl, G., 2003. A process typology of regional floods. Water Recour. Res.39, 1340.

Milly, P.C.D., Wetherald, R.T., Dunne, K.A., Delworth, T.L., 2002. Increasing risk ofgreat floods in a changing climate. Nature 415, 514e517.

Moreno, A., Valero-Garcés, B., González-Sampériz, P., Rico, M., 2008. Flood responseto rainfall variability during the last 2000 years inferred from the Taravilla Lakerecord (Central Iberian Range, Spain). J. Paleolimnol. 40, 943e961.

Mudelsee, M., 2010. Climate Series Analysis: Classical Statistical and BootstrapMethods. In: Atmospheric and Oceanographic Sciences Library 42. Springer,p. 474.

Mudelsee, M., Börngen, M., Tetzlaff, G., Grunewald, U., 2003. No upward trends inthe occurrence of extreme floods in central Europe. Nature 425, 166e169.

Mudelsee, M., Deutsch, M., Börngen, M., Tetzlaff, G., 2006. Trends in flood risk of theRiver Werra (Germany) over the past 500. Hydrol. Sci. J. 51, 818e833.

Mulder, T., Alexander, J., 2001. The physical character of subaqueous sedimentarydensity flows and their deposits. Sedimentology 48, 269e299.

Nadaraya, E.A., 1964. On estimating regression. Theor. Probab. Appl. 9, 141e142.Nesje, A., Dahl, S.O., Matthews, J.A., Berrisford, M.S., 2001. A w4500 yr record of

river floods obtained from a sediment core in Lake Atnsjøen, eastern Norway.J. Paleolimnol. 25, 329e342.

Nicolussi, K., Kaufmann, M., Patzelt, G., Plicht van der, J., Thurner, A., 2005. Holocenetree-line variability in the Kauner Valley, Central Eastern Alps, indicated bydendrochronological analysis of living trees and subfossil logs. Veg. Hist.Archaeobot. 14, 221e234.

Ohlendorf, C., Sturm, K., 2001. Precipitation and dissolution of calcite in a Swiss highalpine lake. Arct. Antarct. Alp. Res. 33, 410e417.

Parajka, J., Kohnová, S., Bálint, G., Barbuc, M., Borga, M., Claps, P., Cheval, S.,Dumitrescu, A., Gaume, E., Hlavcová, K., Merz, R., Pfaundler, M., Stancalie, G.,Szolgay, J., Blöschl, G., 2010. Seasonal characteristics of flood regimes across theAlpine-Carpathian range. J. Hydrol. 394, 78e89.

Patzelt, G., 1977. Der zeitliche Ablauf und das Ausmass postglazialer Kli-maschwankungen in den Alpen. In: Frenzel, B. (Ed.), Dendrochronologie undpostglaziale Klimaschwankungen in Europa, Erdwissenschaftliche Forschung,vol. 13, pp. 249e259.

Patzelt, G., Bortenschlager, S., 1973. Die postglazialen Gletscher- und Kli-maschwankungen in der Venedigergruppe (Hohe Tauern, Ostalpen). Zeitschriftfür Geomorphologie N.F., Supplementband 16, 25e72.

Petrow, T., Merz, B., 2009. Trends in flood magnitude, frequency and seasonality inGermany in the period 1951e2002. J. Hydrol. 371, 129e141.

Pfister, L., Drogue, G., El Idrissi, A., Iffly, J.-F.ß., Poirier, C., Hoffmann, L., 2004. Spatialvariability of trends in the rainfall-runoff relationship: a mesoscale study in theMosel Basin. Clim Change 66, 67e87.

Pinter, N., Ickes, B.S., Wlosinski, J.H., van der Ploeg, R.R., 2006. Trends in floodstages: contrasting results from the Mississippi and Rhine River systems.J. Hydrol. 331, 554e566.

Rumsby, B.T., Macklin, M.G., 1996. River response to the last neoglacial (the ‘LittleIce Age’) in northern, western and central Europe. In: Geological Society Lon-don, Special Publications 115, pp. 217e233.

Schmidt, R., 1981. Grundzüge der spät-und postglazialen Klimageschichte des Sal-zkammergutes (Oberösterreich) aufgrund palynologischen Untersuchunger vonSee- und Moorprofilen. In: Mitteilungen der Kommission für Quartärforschungder Österreichischen Akademie der Wissenschaften 3, pp. 1e96.

Schmidt, R., 1986. Palynologie, Stratigraphie und Großreste von Profilen der neo-lithischen Station See am Mondsee, Oberösterreich. ArchAustr 70, 227e235.

Schmocker-Fackel, P., Naef, F., 2010. Changes in flood frequencies in Switzerlandsince 1500. Hydrol. Earth Syst. Sci. 14, 1581e1594.

Sheffer, Enzel, Y., Benito, G., 2003. Palaeofloods and historical floods of ArchècheRiver, France. Water Recour. Res. 39, 1376.

Sodemann, H., Zubler, E., 2010. Seasonal and inter-annual variability of the moisturesources for Alpine precipitation during 1995e2002. Int. J. Climatol. 30, 947e961.

Starkel, L., 2002. Change in the frequency of extreme events as the indicator ofclimatic change in the Holocene (in fluvial systems). Quat. Int. 91, 25e32.

Starkel, L., Soja, R., Michczynska, D.J., 2006. Past hydrological events reflected inHolocene history of Polish rivers. CATENA 66, 24e33.

Støren, E.N., Dahl, S.O., Nesje, A., Paasche, Ø., 2010. Identifying the sedimentaryimprint of high-frequency Holocene river floods in lake sediments: develop-ment and application of a new method. Quat. Sci. Rev. 29, 3021e3033.

Støren, E.N., Kolstad, E.W., Paasche, Ø., 2012. Linking past flood frequencies inNorway to regional atmospheric circulation anomalies. J. Quat. Sci. 27, 71e80.

Sturm, M., Matter, A., 1978. Turbidites and varves in Lake Brienz (Switzerland):deposition of clastic detritus by density currents. In: Matter, A.,Tucker, M.E. (Eds.), Modern and Ancient Lake Sediments, Special Publica-tion of the International Association of Sedimentologists 2. Blackwell, Ox-ford, pp. 147e168.

Swierczynski, T., Brauer, A., Lauterbach, S., Martin-Puertas, C., Dulski, P., vonGrafenstein, U., Rohr, C., 2012. A 1600-year seasonally resolved record ofdecadal scale flood variability from the Austrian pre-Alps. Geology 40, 1047e1050.

Swierczynski, T., Lauterbach, S., Dulski, P., Brauer, A., 2009. Die Sedimentablager-ungen des Mondsees (Oberösterreich) als ein Archiv extremer Abflussereignisseder letzten 100 Jahre. Alp. Space 6, 115e126.

Swierczynski, T., Lauterbach, S., Dulski, P., Brauer, A., 2013. Late Neolithic MondseeCulture in Austria: living on lakes and living with flood risk? Clim. Past 3, 1601e1612.

Thorndycraft, V., Hu, Y., Oldfield, F., Crooks, P.R.J., Appleby, P.G., 1998. Individualflood events detected in the recent sediments of the Petit Lac d’Annecy, easternFrance. The Holocene 8, 741e746.

Thorndycraft, V.R., Benito, G., Rico, M., Sopena, A., Sánchez-Moya, Y., Casas, A., 2005.A long-term flood discharge record derived from slackwater flood deposits ofthe Llobregat River, NE Spain. J. Hydrol. 313, 16e31.

van Husen, D., 1989. Blatt 65-Mondsee, Geologische Karte der Republik Österreich1:50 000. Geologische Bundesanstalt, Wien.

van Husen, D., 2004. Quaternary glaciations in Austria. In: Ehlers, J., Gibbard, P.L.(Eds.), Quaternary Glaciations: Extent and Chronology, Part 1: Europe, first ed.Elsevier, Amsterdam, pp. 1e13.

Wanner, H., Beer, J., Bütikofer, J., Crowley, T.J., Cubasch, U., Flückiger, J., Goosse, H.,Grosjean, M., Joos, F., Kaplan, J.O., Küttel, M., Müller, S.A., Prentice, I.C.,Solomina, O., Stocker, T.F., Tarasov, P., Wagner, M., Widmann, M., 2008. Mid- toLate Holocene climate change: an overview. Quat. Science Rev. 27, 1791e1828.

Wanner, H., Brönnimann, S., Casty, C., Gyalistras, D., Luterbacher, J., Schmutz, C.,Stephenson, D.B., Xoplaki, E., 2001. North Atlantic oscillation e concepts andstudies. Surv. Geophys. 22, 321e381.

Watson, G.S., 1964. Smooth regression analysis. Indian J. Stat. 26, 359e378.Wessels, M., 1998. Natural environmental changes indicated by Late Glacial and

Holocene sediments from Lake Constance, Germany. Palaeogeogr. Palae-oclimatol. Palaeoecol. 140, 421e432.

Wilhelm, B., Arnaud, F., Enters, D., Allignol, F., Legaz, A., Magand, O., Revillon, S.,Giguet-Covex, C., Malet, E., 2012. Does global warming favour the occurrence ofextreme floods in European Alps? First evidences from a NW Alps proglaciallake sediment record. Clim. Change 113, 563e581.

Wilhelm, B., Arnaud, F., Sabatier, P., Magand, O., Chapron, E., Courp, T., Tachikawa, K.,Fanget, B., Malet, E., Pignol, C., Bard, E., Delannoy, J.-J., 2013. Palaeoflood activityand climate change over the last 1400 years recorded by lake sediments in thenorth-west European Alps. J. Quat. Sci. 28, 189e199.

Wirth, S., Gilli, A., Simonneau, A., Ariztegui, D., Vannière, B., Glur, L., Chapron, E.,Magny, M., Anselmetti, F., 2013. A 2000-year long seasonal record of floods in thesouthern European Alps. A 2000-year long seasonal record of floods in thesouthern European Alps. Geophys. Res. Lett.. http://dx.doi.org/10.1002/grl.50741.

Zolitschka, B., Behre, K.-E., Schneider, J., 2003. Human and climatic impact on theenvironment as derived from colluvial, fluvial and lacustrine archives e ex-amples from the Bronze Age to the migration period, Germany. Quat. Sci. Rev.22, 81e100.

Zoller, H., 1960. Pollenanalytische Untersuchungen zur Vegetationsgeschichte derinsubrischen Schweiz. In: Denkschriften der Schweizerischen Natur-forschenden Gesellschaft 83, pp. 45e156.

Zoller, H., Schindler, C., Röthlisberger, 1966. Postglaziale Gletschstände und Kli-maschwankungen im Gotthardmassiv und Vorderrheingebiet. In: Verhand-lungen der Naturforschungs Gesellschaft Basel 77, pp. 97e164.