Palaeoenvironmental and palaeoseismic implications of a 3700-year sedimentary record from proglacial Lake Barrancs (Maladeta Massif, Central Pyrenees, Spain)

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Palaeoenvironmental and palaeoseismic implications of a 3700-year sedimentaryrecord from proglacial Lake Barrancs (Maladeta Massif, Central Pyrenees, Spain)

Juan C. Larrasoaña a,⁎, María Ortuño b, Hilary H. Birks c,d, Blas Valero-Garcés e, Josep M. Parés f,1,Ramon Copons g, Lluís Camarero h, Jaume Bordonau b

a Institut de Ciències de la Terra “Jaume Almera”, CSIC, Solé i Sabarís s/n, 08028 Barcelona, Spainb RISKNAT Group, Departament de Geodinàmica i Geofisica, Universitat de Barcelona, Martí i Franqués s/n, 08028 Barcelona, Spainc Department of Biology, University of Bergen, Allégaten 41, N-5007 Bergen, Norwayd Bierknes Centre for Climate Change, Allégaten 41, N-5007 Bergen, Norwaye Instituto Pirenaico de Ecología, CSIC, Aptdo. 13034, 50080 Zaragoza, Spainf Department of Geological Sciences, University of Michigan, 2534 CC. Little Building, Ann Arbor, MI 48109-1063, USAg GEORISC S.L., Vidal i Quadras 44, 08017 Barcelona, Spainh Centre d'Estudis Avançats de Blanes, CSIC, Camino de Santa Barbara, 17300 Blanes, Spain

a b s t r a c ta r t i c l e i n f o

Article history:Received 18 June 2008Received in revised form 19 March 2009Accepted 1 April 2009Available online 9 April 2009

Keywords:Iberian PeninsulaProglacial lakesEnvironmental magnetismPlant macrofossilsGlacier fluctuationsPalaeoseismicity

A multidisciplinary study including sedimentological, mineral magnetic, and palaeobotanical techniquesapplied to a sediment core recovered from proglacial Lake Barrancs in the seismically active Maladeta Massifhas provided the basis for documenting environmental changes and palaeoseismic activity in the CentralPyrenees for the last ca. 3700 yr. Lake Barrancs is located downstream of the Tempestats and Barrancs cirqueglaciers and sedimentation is dominated by clastic input corresponding to seasonal changes in sedimentsupply. Slow fine particle settling during the winter and sediment-loaded homopycnal flows during thewarm season, triggered by snow-melting and glacier outwash, have resulted in deposition of rhythmitescomposed of clays, silts, and sands. The predominance of finer-grained sediments and the low concentrationof relatively finer magnetite grains suggest that glacier activity was very small, if not absent, before ca. A.D.350. Their replacement by coarser-grained sediments and the overall increased (but highly oscillating)concentrations of relatively coarser magnetite grains in the uppermost 4.3 m of the record suggest the onsetof glacial activity and enhanced snow-melting in the catchment of Lake Barrancs after A.D. 350. We suggestthat this onset of glacial and enhanced snow-melt activity was driven by a complex balance between winterprecipitation and annual mean temperatures, among other climatic variables. Peat layers suggest twodramatic lake-level drops at A.D. 300 and A.D. 450, when Lake Barrancs was drained. The mechanisms forsuch extreme hydrological events are not clear. Changes in the precipitation/evaporation ratio cannotaccount for such desiccation events. Dam failure is unlikely since there are no geomorphological evidence ofbreaching processes. Geomorphological and structural evidence demonstrates active faulting since formationof Lake Barrancs and reactivation during earthquake shaking. Based on this, we propose an alternativeexplanation for the desiccation events that involves the draining of the lake through pre-existing fracturesopened by earthquakes. Further studies in Lake Barrancs and other lakes from the Maladeta massif arenecessary to validate the hypotheses presented here concerning the response of glacial and snow-meltactivity to climate variability and the palaeoseismic record of the Central Pyrenees.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Lake sediments are important archives of geological processes,palaeoenvironmental variations, and past human activity becausethey often have high accumulation rates that enable the study of these

processes at resolutions down to centennial and even decadaltimescales (see De Batist and Chapron, 2008). High-altitude lakerecords are the focus of an increasing number of studies becausemountains are very sensitive to recent environmental changes(Battarbee et al., 2002; Pla and Catalan, 2005). This is especiallyevident for proglacial lakes, which have been shown to provide notonly continuous, but also high-resolution records of glacier dynamicsin locations where its geomorphological expression is very discontin-uous both in time and space (Matthews and Karlén, 1992; Leemannand Niessen, 1994; Leonard and Reasoner, 1999; Dahl et al., 2003; Lieet al., 2004; Nesje et al., 2006; Chapron et al., 2007). The study of

Palaeogeography, Palaeoclimatology, Palaeoecology 294 (2010) 83–93

⁎ Corresponding author. Now at Area de Cambio Global, Instituto Geológico y Minerode España, Oficina de Proyectos de Zaragoza, C/ Manuel Lasala 44, 9°B, 50006 Zaragoza,Spain. Tel.: +34 976 555 153; Fax: +34 976 555 582.

E-mail address: [email protected] (J.C. Larrasoaña).1 Now at: CENIEH, 09004 Burgos, Spain.

0031-0182/$ – see front matter © 2009 Elsevier B.V. All rights reserved.doi:10.1016/j.palaeo.2009.04.003

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology

j ourna l homepage: www.e lsev ie r.com/ locate /pa laeo

Author's personal copy

glacier history from proglacial lake sediments is based on theassumption that glacier size controls sediment particulate productionand hence sediment transport into the lake (Lie et al., 2004; Nesjeet al., 2006; Chapron et al., 2007). However, other processes such asfluvial reworking of glacial-sourced sediment load (e.g. the ‘para-glacial’ processes of Church and Ryder, 1972) might overprint theglacial signal (Lie et al., 2004).

Lake sediments have been also shown to provide long and well-dated records of seismic activity (Monecke et al., 2004; Becker et al.,

2005; Carrillo et al., 2008; Wagner et al., 2008), which is importantbecause other terrestrial palaeoseismic indicators provide onlydiscontinuous records of seismic activity that are difficult to date(Becker et al., 2005; De Batist and Chapron, 2008). In the recent years,a growing number of studies have shown the combined effects ofseismicity, climate, and environmental changes on lacustrine sedi-mentation (Bertrand et al., 2008; Carrillo et al., 2008; Fanetti et al.,2008; Wagner et al., 2008). Thus, especial care has to be taken whenstudying lacustrine sedimentary records to disentangle signals of

Fig. 1. A) Sketch map showing historical large earthquakes (INVIII) in the Pyrenean region (IGN, 2006). The studied area, indicated by a black square, includes two of the greaterearthquakes that occurred in the Pyrenees. B) Ortophoto of the studied area, with indication of the boundary between the Maladeta granitoid and the Palaeozoic country rocks andlocation of two historical earthquakes (small triangle: I=V, 2.12.1919; large triangle: Ribagorza earthquake, I=VIII–IX, 3.3.1373). NMF stands for the North Maladeta Fault locatedb3 km east of the studied area. C) Geology and geomorphology of the Maladeta Massif (after Moya and Vilaplana, 1992; Copons and Bordonau, 1994, 1996; Chueca Cía et al., 2005),with location of Lake Barrancs and Core B5. The dashed line indicates the lake catchment. D) Aerial picture showing the main faults around Lake Barrancs, the bathymetry of the lake(isobaths are 2 m), and location of Core B5.

84 J.C. Larrasoaña et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 294 (2010) 83–93

Author's personal copy

seismo-tectonic activity from those generated by climatic variability(De Batist and Chapron, 2008). This is especially relevant for high-altitude lakes because they are located in young and active mountainbelts that are very sensitive to climatic variations and can also beaffected by intense seismic activity.

Here we present the study of a 6.85 m long sedimentary sequencerecovered from proglacial Lake Barrancs, which is located at analtitude of 2360 m a.s.l. in the Maladeta Massif (Central Pyrenees,Spain). Lake Barrancs is one of the few Pyrenean lakes located justdownstream (b1.5 km) of active cirque glaciers. Moreover, theMaladeta Massif is one of the most seismically active regions withinthe Pyrenees (Souriau and Pauchet, 1998). The sedimentarysequence recovered from Lake Barrancs might therefore constitutea unique record of recent glacier and seismic activity in the Pyrenees.We use a combination of sedimentologic, environmental magnetic,and palaeobotanic techniques to characterize the depositionalevolution of Lake Barrancs. The changes in sedimentary dynamicsprovide the basis for reconstructing palaeoenvironmental variationsand palaeoseismic activity in the Central Pyrenees for the last ca.3700 yr.

2. Geological setting

Lake Barrancs is located in the Axial Zone of the central Pyrenees,on the northern slope of theMaladetaMassif (Fig.1). Thismassif hosts

the highest Pyrenean peak (Aneto Peak, 3404 m a.s.l.) and the largestcirque glaciers still preserved in the Pyrenean mountain belt (Fig. 2).The Maladeta Massif is composed of medium to coarse-grainedgranitoids emplaced within Palaeozoic sediments during the latestages of the Variscan orogeny (Leblanc et al., 1994; Evans et al., 1998),and is affected by three sets of NNW–SSE, N–S, and NE–SW orientedsubvertical fractures. Near Lake Barrancs, these fractures are repre-sented by several major, NNW–SSE oriented fault scarps withrectilinear traces and by minor, NE–SW oriented faults. The NNW–

SSE oriented faults delineate an elongated ridge that separates twodepressions with steep margins, the eastern one occupied by LakeBarrancs and the western one by periglacial talus deposits (Figs. 1, 2).These faults have a maximum length of 1.4 km, show normaldisplacements of b40m that offset polished glacial surfaces originatedduring the Last Pyrenean Pleniglacial, and present glacial striae(Fig. 2C) (Moya and Vilaplana, 1992; Ortuño, 2008). The NE–SW faultshave a maximum length of 200 m, show minor displacements ofb10 m that offset the major NNW–SSE oriented faults, and do not bearglacial striae on their surfaces (Moya and Vilaplana, 1992). LakeBarrancs is located at b3 km NE of the Coronas fault and b7 km SW ofthe North Maladeta fault (Fig. 1B). These normal faults, between 12and 18 km long, are among the few seismogenic faults withgeomorphological expression recognized in the Central Pyrenees,and have been identified as the most likely source of historicalearthquakes in the area, such as the MW=5.3 Vielha (19.11.1923) and

Fig. 2. A) Lake Barrancs and its deltaic plain (DP) viewed from the south, with location of Core B5. B) Lake Barrancs viewed from the north. The Tempestats glacier (TG), the Little IceAge (LIA) and the Holocene (H) moraines are clearly visible. The white rectangle marks the location of the fault shown in Fig. 2C. C) Detail of one of the faults located near the SWshore of Lake Barrancs. The fault displaces a polished glacial surface and shows a colour banding (dark grey, beige and orange from top to bottom) that mimics the shape of thedisplaced blocks. Arrows indicate the slip along the fault surface. D) Detail of the drilling camp and platform set up over the frozen surface of the lake.

85J.C. Larrasoaña et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 294 (2010) 83–93

Author's personal copy

the MW=6.2 Ribagorza (3.3.1373) earthquakes (Fig. 1A) (Ortuño,2008; Ortuño et al., 2008).

During the maximum extent of the glaciers in the central Pyrenees,dated at 60–70 ka in the Cinca valley by OSL (Sancho et al., 2002) and at20–25 in theNoguera Ribagorzanavalley bymeansof 10Be exposure ages(Pallàs et al., 2006), the Maladeta Massif was covered by a 36 km longvalley glacier that flowed down the Esera valley. This glacier over-excavated several basins, among them the one that contains LakeBarrancs. Lake Barrancs probably formed during the final stages of thePyrenean deglaciation, when the Esera glacier was fragmented intoseveral cirque glaciers with small ice tongues. A moraine locatedupstream of Lake Barrancs, at about 2400 m a.s.l. (Figs. 1C, 2B) (Moyaand Vilaplana, 1992; Copons and Bordonau, 1996), formed during atransient stabilization period dated at ca. 10 ka (e.g. Early Holocene) onthe basis of 10Be exposure ages of moraines on the SE slope of theMaladetaMassif (Pallàs et al., 2006). In addition, historical glacial phaseshave been documented in the Maladeta Massif. A prominent (tens ofmetres high) moraine ridge, disconnected from the present-day glacierfronts, attests for glacier advances during the Little Ice Age (LIA, 18th–19th centuries) (Figs.1C, 2B) (Copons and Bordonau,1994,1996; ChuecaCía et al., 2005).

Lake Barrancs is an elongated lake (470 m long, 100 m wide)located downstream of the Barrancs and Tempestats (0.11 and0.14 km2, respectively) cirque glaciers (Figs. 1, 2) (Copons andBordonau, 1994, 1996; Chueca Cía et al., 2005). The lake has verysteep slopes, especially on its western shore, but has a rather flatbottom composed of two small sub-basins with maximum depths ofca. 13.5 and 12.5 m (Fig. 1D). The bathymetry of the lake reflects thetectonic setting in which it is located. Thus, the steeper slope of thewestern shore reflects the fault scarp that bounds the Lake Barrancsbasin to the west, whereas the small bathymetric high that separatesthe two sub-basins attests the occurrence of minor NE–SW orientedfaults (Fig. 1C, D). The lake catchment (b4 km2) is composed entirelyof granitoids and includes some periglacial talus deposits and glacialtills from the Holocene and LIAmoraines. The lake catchment has verystrong topographic gradients (N1000 m in 2 km) and is largely snow-covered from November to May, when large snow avalanches cantransport coarse sediments to the frozen surface of the lake. Snow-melting in the catchment occurs typically betweenMay and June. Meltwaters have eroded the Early Holocene and LIA till sediments. As aresult, a proglacial cone has formed downstream of the Barrancsglacier and a delta has occupied nearly half the depressionwhere LakeBarrancs is located (Figs. 1C, 2A). The eastern shore of Lake Barrancs iscovered by screes generated by rockfalls from the steep slopes.

3. Materials and methods

Drilling in Lake Barrancs was carried out from the frozen surface ofthe lake in March 1998 (Fig. 2D). Five cores, 6 cm in diameter, wererecovered from the small bathymetric high located in the central partof the lake (Figs. 1D, 2A) using a stationary piston corer (Montserrat,1992) designed after the modified Livingstone coring apparatus

(Wright, 1980). The longest core (B5) was drilled at 11 m waterdepth without reaching the substratum, and comprises a nearly con-tinuous composite sequence 6.85 m thick.

In the laboratory, core B5was split in two halves, logged, described,and sampled. Sedimentological descriptions include lithology, colour,grain size, and sediment textures and structures. Total organic carboncontents of some representative samples were determined using aLECO elemental analyzer. Three bulk organic matter samples and twosamples of terrestrial plant macrofossils were AMS 14C dated at theBeta Analytic laboratory (Miami, USA). The AMS dates were calibratedwith CALIB v.5.0.2 (Stuiver and Reimer, 1986; Stuiver et al., 2005).Plant macrofossils were extracted from two samples (8 cm3) from an

Table 1Radiocarbon data from Core B5.

Lab. ref. Depth(cblf)

Material 14C age(yr B.P.)

Corrected 14C age(yr B.P.)

Cal. age(2σ)

Probabilitydistribution

δ13C(‰)

Unit

BETA-122150 170 Bulk sediment 1240±40 630±40 AD 1285–1401 1 −26.3 3BETA-122151 286.5 Bulk sediment 1900±50 1290±50 AD 652–829 0.952 −26.7 3

AD 837–867 0.048BETA-122153 338 Plant material 1540±30 1540±30 AD 430–590 1 −27.2 3BETA-122148 426 Bulk sediment 2300±40 1690±40 AD 249–426 1 −26.5 2BETA-122149 567 Plant material 2560±40 2560±40 BC 809–729 0.501 −26.4 1

BC 692–659 0.157BC 652–543 0.342

Bold numbers indicate radiocarbon ages chosen on the basis of their probability distribution.

Fig. 3. A) Plot of depth versus 14C ages for Core B5. The solid line represents the fit of thetwo samples on plant remains to a line that intercepts the top of the core. The dashedline represents the fit of the three samples on bulk sediment (open symbols) to a line,with a similar slope, that intercepts the core top at ca. 610 yr BP. B) Depth-agemodel forCore B5 constructed after correcting the bulk sediment samples for a reservoir age of610 yr. Horizontal bars represent the error at the 2σ interval.

86 J.C. Larrasoaña et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 294 (2010) 83–93

Author's personal copy

organic-rich layer located in the middle part of Core B5. The sampleswere disaggregated in water and sieved through a 125 μm mesh. Theplant remains were systematically picked out at 12× magnificationunder a stereo-microscope and identified (Birks, 2001).

Sampling for environmental magnetic measurements was done bypushing 2×2×2 cm standard plastic boxes into the working half of coreB5. Sampling was performed continuously through the sedimentarysection. Magnetic properties were measured at the PalaeomagneticLaboratory of the Ludwig Maximilians Universität (Munich, Germany),and include: 1) the low field magnetic susceptibility (χ); 2) ananhysteretic remanent magnetization (ARM), applied in a dc bias fieldof 0.05 mT parallel to an axially-oriented peak alternating field (AF) of100mT; and 3) two isothermal remanentmagnetizations applied at 0.2 T([email protected] T) and 1.5 T (SIRM). χ was measured with a KLY-2 magneticsusceptibility bridge using a field of 0.1 mTat a frequency of 470 Hz. ARMwas producedusing a 2GEnterprises AF demagnetizer, andwasmeasuredwith a 2G Enterprises three-axis cryogenic magnetometer (noise level ofb7×10−6 A/m). Both the [email protected] T and SIRM were produced using ahome-made pulse magnetizer and were measured with the samecryogenic magnetometer. All magnetic properties were normalized bythedryweightof the samples.Wehaveuseddifferentmagneticpropertiesand interparametric ratios to determine downcore relative variations inthe type, concentration, and grain size of magnetic minerals (Thompson

andOldfield,1986; Verosub andRoberts,1995; Peters andDekkers, 2003).χhas beenused as afirst order indicator for the concentration ofmagnetic(s.l.) minerals. SIRM/χ and the S-ratio (operationally defined [email protected] T/SIRM; Bloemendal et al., 1992) have been used to detectchanges in magnetic mineralogy (Verosub and Roberts, 1995; Peters andDekkers, 2003). Then, the ARM and the “hard” IRM (HIRM, operationallydefined as [email protected] T; Thompson and Oldfield, 1986) have beenused as aproxy for the concentrationof relatively low- andhigh-coercivityminerals, respectively (Verosub and Roberts, 1995). Finally, the ratiosbetween SIRM and χ and between SIRM and χARM (i.e. the anhystereticsusceptibility; King et al., 1982), have been used to detect downcorechanges in magnetic grain size, provided that they correspond to a singlemagnetic mineral (Thompson and Oldfield, 1986; Verosub and Roberts,1995; Peters and Dekkers, 2003). All results from core B5 are referred tocentimetres below the lake floor (sediment surface) (cblf).

4. Results

4.1. Chronology

The chronological model for the Lake Barrancs sequence is based onfive AMS radiocarbon dates ranging from 2560±40 14C yr BP at 567 cblfto 1240±40 14C yr BP at 170 cblf (Table 1). The 14C age–depth

Fig. 4. Lithostratigraphy, sedimentary units and facies associations (A), and selected magnetic properties (B) of Core B5.

87J.C. Larrasoaña et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 294 (2010) 83–93

Author's personal copy

relationship shows a reversal between two samples at 286.5 (bulksediment) and 338 (plant remains) cblf, which suggests that the age ofthe bulk sediment samples could be too old due to some reservoir effectaffecting bulk organic matter (Fig. 3A). The most likely explanations forsuch old ages in bulk organic matter samples are the presence of oldcarbon from the catchment as constituent of the fine-grainedsedimentary load and the uptake of old carbon (leached from thebedrock catchment) by phytoplankton (e.g. Lowe and Walker, 1997).The two samples dated onplant remains gave 14C ages that conform to astraight line that intercepts the topof the core at ca. 30 yr BP(Fig. 3A).Onthe contrary, the three samples dated on bulk sediment gave 14C agesthatfit (R=0.98) to a line that intercepts the core top at ca. 610 14C yr BP(Fig. 3A). The slope of this line is very similar to that of the samples onplant remains and the top of the core. This suggests that all the threebulk sediment samples, not only the one at 286.5 cblf, are affected by amore or less constant reservoir age of ca. 610 yr. Since old carbon erodedfrom the catchment would presumably have variable ages, the mostlikely explanation for such constant ageing in lacustrine-derived organicmatter is a hard-water effect. It is comparable to the ca. 940 yrdiscrepancy reported for 14C ages in the neighbouring Lake Redon byCamarero et al. (1998), who consider it to be a hard-water reservoireffect. After correcting thebulk sediment radiocarbon ages for the 610 yrreservoir age, the Barrancs age model was constructed using calibratedages chosen on the basis of their maximum probability distributionwithin the 2σ interval (Table 1). The corrected age model is based onlinear interpolation between calibrated radiocarbondates, and indicatesthat the sediment sequence of Lake Barrancs recovered in core B5 spansthe last 3700 yr (Fig. 3B). The average accumulation rate is about1.85 mm/yr, lower but comparable to a previous report from LakeBarrancs in which no reservoir effect was considered (i.e. 2.6 mm/yr;Coponset al.,1997). In anycase, sedimentation rates in LakeBarrancs areseveral times larger than in other high-altitude Pyrenean lakes such asLake Tramacastilla (ca. 0.4 mm/yr; García-Ruiz et al., 2003) and LakeRedon (ca. 0.055 mm/yr; Pla and Catalan, 2005), suggesting highsediment delivery to the lake during the last millennia.

4.2. Lithostratigraphy

Core B5 includes a nearly continuous sequence of proglacialsediments composed of fine sands, silts, and clays organized in three

facies associations (Fig. 4). Facies association F1 is composed of lightbrown to grey clays, silts, and sandy silts with occasional, discrete,mm-thick intervals of organic debris. These sediments have a poorlydeveloped horizontal lamination that is marked by faint colour andtextural variations. Facies association F2 is composed of light brown togrey clays and silts that contain dispersed organic debris and includeabundant layers of sandy silts and fine sands. The sandy silts and finesands occasionally contain distinctive whitish layers, which can be upto 1 cm thick and often show a turbidite-like fining-upward grain-sizegradation. Facies association F2 sediments have a well-developedmillimetre to centimetre horizontal lamination that is marked bytextural and colour variations. Facies association F2 is scarce in thelowermost 3 m of the record but dominates between 430 and 70 cblf.Facies associations F1 and F2 have low TOC contents (b0.4% inweight). It is worth mentioning that, although no dropstone has beenobserved throughout the studied sequence of core B5, the presence ofdropstones in Lake Barrancs sediments is evidenced by the difficultiesfaced when drilling other sites, which had to be abandoned whenencountering large blocks that prevented further drilling.

Distinctive organic-rich dark layers (facies association F3) occurbetween 344 and 363 cblf, and 424–427 cblf. The upper one isintercalated within a thick (70 cm) facies association F2 interval(Fig. 4). This 19-cm thick layer is composed of silts and sands, whichinclude large quartz, feldspar and biotite grains of up to 3 mm at thebase of the layer. The distinctive dark colour is caused bydisseminated organic matter and mm-scale vegetal remains, whichgive a mean TOC content of nearly 3% by weight. At the top of thelayer, vertical root structures penetrate 10 cm into the layer. Plantmacrofossils associations are composed of Calluna vulgaris, Rhodo-dendruon ferrugineum, Selaginella selaginoides, Juniperus communis,Salix sp., Betula (peduncula/pubescens), Ranunculus sp., and Carexsp. (Table 2). An associated fauna of oribatids, trichopterans,chironomids, and fragments of other insects also occurs. Thisassemblage is typical of Pyrenean heathlands located aroundmountain streams up to 2200 m a.s.l. (Villar et al., 1997). Thelower dark layer (424–427 cblf) appears at the base of a thin (20 cm)facies association F2 interval (Fig. 4). This layer is also composed ofsilts and sands that contain organic debris (including root frag-ments) and large quartz, feldspar, and biotite grains (b2 mm) in itslower part.

Table 2Macrofossil plants and invertebrate remains in two selected samples from the organic-rich layer between 344 and 363 cblf.

Sample Depth(cm)

Macrofossil plants Other constituents

Plant type Plant part Abundance Invertebrate remains Abundance

4.18 345 Calluna vulgaris Seeds 1 Oribatid mites fCalluna vulgaris Flowers 3 Chironomids vrCalluna vulgaris Leaves, stems, twigs f Insect fragments pJuniperus communis Leaves 4Salix sp. Bark fragments rBetula (peduncula/pubescens) Bud scales 3Rhododendron ferrugineum Leaf glands ocRhododendron ferrugineum Seeds 2Rhododendron cf. ferrugineum Leaf fragments ocRanunculus sp. Achene (small) 1Carex sp. (tristigmata) Nutlets 1Roots ocLeaf and twig fragments 1

4.26 362 Calluna vulgaris Twigs oc Chironomids rSelaginella selaginoides Megaspores 2 Trichoptera vrTwigs oc Insect fragments pLeaf fragments oc

f = frequent.oc = occasional.r = rare.vr = very rare.p = present.

88 J.C. Larrasoaña et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 294 (2010) 83–93

Author's personal copy

Facies associations F1 to F3 define two stratigraphic units (Fig. 4).The uppermost unit (Unit 2) is dominated by facies association F2,which is progressively replaced by facies association F1 towards thetop of the unit. The presence of the two organic-rich layers in thelower part of Unit 2 enables identification of two deepeningsequences that range from subaerial conditions (indicated by faciesassociation F3) to relatively deeper lacustrine sedimentation(indicated by the increased abundance of facies association F1towards the top). Unit 1 is dominated by facies association F1, andappears also to form a deepening sequence (Fig. 4).

4.3. Environmental magnetism

Magnetic susceptibility values oscillate around 8×10−8 m3/kgthroughout the record except in the two organic-rich layers at around350 and 430 cblf and at a facies association F1 interval located around650 cblf, where magnetic susceptibility values drop sharply wellbelow 7×10−8 m3/kg (Fig. 4). These values are similar to those ofsamples from the Maladeta granitoid, which typically range between4 and 12×10−8 m3/kg (i.e. 10–30×10−5 SI; Leblanc et al., 1994). Thisis consistent with the fact that the granitoid rocks that surround LakeBarrancs are dominantly paramagnetic (Leblanc et al., 1994) and arethe main source for sediments accumulated in Lake Barrancs. Lowmagnetic susceptibility values such as those in the two organic-richlayers might indicate a different concentration of detrital minerals or,alternatively, a post-depositional change in the primary magneticsignal.

Sediments from facies associations F1 and F2 are characterizedby S-ratios ranging between 0.7 and 0.9, which contrast with thedistinctively low (down to 0.6) S-ratios in the two organic-richlayers (facies F3) (Fig. 4). SIRM/χ values oscillate between 0.5 and3.5 kA/m except in the uppermost organic-rich layer, which showsdistinctively higher SIRM/χ values of up to 9 kA/m. Similarly,SIRM/χARM ratios range between 1 and 3 kA/m except in theuppermost organic-rich layer, which shows significantly higherratios of up to 4 kA/m. High S-ratios combined with low SIRM/χvalues indicate that the main magnetic mineral characterizing faciesassociations F1 and F2 is magnetite (Snowball, 1993; Verosub andRoberts, 1995; Peters and Dekkers, 2003). This magnetite is mainlyderived from glacial abrasion of the granitoid catchment rock and itssubsequent transport to the lake mainly by snow-melt waters andglacier outwash. Downcore variations in ARM intensity reveal a ratherlow (b2×10−6 Am2/kg) and constant concentration of magnetite inUnit 1, which contrasts with oscillating values between 2×10−6 and10×10−6 Am2/kg in Unit 2. Concerning parameters indicative of grain-size variations, SIRM/χ values show rather constant values (around1 kA/m) in Unit 1 and highly oscillating values (1–4 kA/m, excluding F3intervals) in Unit 2. In contrast, SIRM/χARM ratios show rather constantvalues of around 1.8 kA/m throughout the core, with only subtlevariations in Unit 2 (facies F3 excluded). The combination of such SIRM/χARM and SIRM/χ ratios indicates that magnetite grains in faciesassociations F1 and F2 range between 0.01 and 0.03 µm in size (Petersand Dekkers, 2003). The high variability of SIRM/χARM ratioswithin thisgrain-size range (Peters and Dekkers, 2003), coupled with their ratherconstantvalues throughoutunits 1 and2, prevents theuse of SIRM/χARM

ratios as a proxy for variations in magnetite grain size. On the contrary,SIRM/χ ratios, which tend to increase for coarser magnetite within the0.01–0.03 µm grain-size range (Peters and Dekkers, 2003), show adistinctively different behaviour between units 1 and 2. Thus, SIRM/χratios of up to 4 kA/m in facies association F2 sediments suggest thatthey have coarser magnetite grains compared with facies association F1intervals (around 1 kA/m), which is consistent with the sedi-mentological description of both facies types. In any case, it should bestressed that these grain-size variations are very subtle, probably inresponse to the original grain-size distribution of magnetite in theparent Maladeta granitoid.

The distinctively low S-ratios that characterize the two organic-rich layers indicate that the main magnetic mineral present in faciesassociation F3 is not magnetite, but rather a relatively-highercoercivity mineral phase (Verosub and Roberts, 1995). Diageneticreactions in sediments are mainly driven by the metabolic activity ofmicrobes, which consume oxygen (under oxic conditions), nitrate,manganese and iron oxides (under suboxic conditions), and sulphate(under anoxic conditions) to degrade buried organic matter (Froelichet al., 1979). Microbially-mediated reduction of sulphate duringearliest diagenesis releases sulphide, which reacts with iron-bearingminerals (including magnetic grains) and dissolved iron to form ironsulphides. Significant accumulations of organic matter in lake sedi-ments easily drives diagenetic reactions to the point where detritalmagnetic grains are dissolved and the magnetic iron sulphide greigiteis formed (Snowball, 1991). This process is favoured when sulphide isproduced in low amounts (e.g. Roberts and Weaver, 2005; Larrasoañaet al., 2007), which typically happens in small lakes due to the lowavailability of dissolved sulphate. Given the high organic content ofthe two organic-rich layers, and keeping in mind that hematite andgoethite are unstable under reducing conditions (Canfield et al., 1992),we interpret that the most likely magnetic mineral in F3 sediments isgreigite because it better explains the combined lower S-ratios and thedistinctively higher SIRM/χ and SIRM/χARM values (Snowball, 1991,1993; Roberts,1995) of the two organic-rich layers. Authigenic growthof greigite and reductive dissolution of magnetite accounts for the lowχ values of the organic-rich layers, because magnetite has a largerspecific magnetic susceptibility compared to greigite (Roberts, 1995).HIRM values give an indication of the concentration of the relatively-higher coercivity minerals, and indicate that large amounts of greigitewere formed in the upper organic-rich layer, but not in the lower one.

5. Discussion

5.1. Sedimentary processes in Lake Barrancs

Facies associations F1 and F2 are typically deposited in proglaciallakes in which rhythmic sedimentation is controlled by the seasonalalternation of intense snow and glacier melting during the late springand summer with freezing conditions over thewinter and early spring(Ashley, 1995; Chapron et al., 2007). Between May and June, floodevents triggered by snow-melting cause cold and sediment-loadedrunoff waters to flow into the lake. They form an underflow currentthat is gradually mixed with the lake water. Over the course of thesummer, glacier outwash takes over as the main source of underflowcurrents. As a result of these floods, also called homopycnal flows(Bates, 1954), the sedimentary load moves by advection throughoutthe water column until it eventually settles down draping the lakebottom (Ashley, 1995, Chapron et al., 2007). Homopycnal flows are aneffective mechanism for separating coarse- and fine-grained sedi-ments due to their differential settling times (Ashley, 1995), whichaccounts for the alternation of clays, silts, and sands that characterizesfacies associations F1 and F2. During winter and early spring, whenthe surface of the lake is frozen, sedimentation is restricted to finerparticles settling out of suspension, which accounts for the presence ofthin clay layers within facies associations F1 and F2 sediments. Thehigher abundance of sandy silts and fine sands in facies association F2is interpreted as an increased intensity and/or frequency of homo-pycnal flows (Chapron et al., 2007), which might also explain thebetter-developed laminations in these sediments.

Snow and glacier melting throughout the spring and summer,coupled with the availability of easily erodable moraine material in thecatchment, might explain the high accumulation rates (1.85 mm/yr)that characterize the sediment core from Lake Barrancs. In this regard, itis worth mentioning that a substantial part of the sedimentary loadproduced by the Barrancs and Tempestats glaciers is deposited in theproglacial cone formed downstream of the Barrancs glacier and in the

89J.C. Larrasoaña et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 294 (2010) 83–93

Author's personal copy

delta that occupies the proximal part of Lake Barrancs (Figs. 1, 2A).Despite this, intense glacial grinding, together with meltwater runoffand erosion of glacial deposits, provides a very effective mechanism fortransferring sedimentary load from the glaciated areas of the catchmentinto the lake.

Facies association F3 sediments were formed in subaerial condi-tions, as shown by the plant remains, the high TOC, and the in situroots. The presence of aquatic invertebrates and Selaginella mightindicate sporadic (e.g. seasonal) water-logged conditions whichwould allow the preservation of organic remains as a peaty deposit.At least for the organic-rich layer at around 350 cblf, the plantmacrofossil assemblage indicates that subaerial conditions lasted longenough to enable the development of a dwarf-shrub community andthe accumulation of 19 cm of peat. The tree-birch remains probablyoriginated from trees growing on the valley sides. It might be arguedthat these two organic-rich layers correspond to reworked peat andpalaeosoil fragments that were transported by snowavalanches or slidfrom the deltaic plain located upstream.We discard these possibilitiesand propose that these organic-rich layers are palaeosoils formed insitu during low-lake-level conditions because: 1) the presence of rootsin a vertical position and the absence for internal deformation is notconsistent with a slump (i.e. folded) geometry; and 2) the basal partsof the organic-rich layers contain the coarsest sand grains reported inthe section (Fig. 4), which is compatible with sudden base level fallsand the concomitant arrival of coarser detrital material.

5.2. Palaeoenvironmental implications

Unit 2 is characterized by the predominance of coarser-grainedsediments (facies association F2) and by overall increased (but highlyoscillating) concentrations of relatively coarser magnetite grains. Thisindicates that periods of enhanced detrital supply into Lake Barrancshave been common between ca. A.D. 350 and the present. Conversely,Unit 1 is dominated by finer-grained sediments (facies association F1)and by constantly low concentrations of relatively finer magnetitegrains, which indicates lower detrital supply into the lake before A.D.350 and back to, at least, 1700 B.C., which is the lower boundary of ourrecord. Present-day detrital supply into Lake Barrancs is governed byhomopycnal flows triggered by snow-melt and glacier outwash.Enhanced glacier and snow-melt activity are typically manifested bycoarser grain sizes (Lie et al., 2004; Nesje et al., 2006; Paasche et al.,2007; Chapron et al., 2007). Based on the sedimentary and magneticrecord of Lake Barrancs, we conclude that glacier activity was verysmall (if not absent) and snow-meltingwas largely reduced beforeA.D.350, but that they have been active in the catchment of the lake sincethen.

No obvious link is seen between sedimentary facies and magneticparameters from the Lake Barrancs sequence with Late Holocene globalclimatic variations recognized at different locations throughout theIberian Peninsula (Gutiérrez-Elorza and Peña-Monné, 1998; Despratet al., 2003; Riera et al., 2004; Gil-García et al., 2007) (Fig. 5). Thus,

Fig. 5. Age variations of selected magnetic properties from Core B5 compared with a regional climatic record of winter mean temperatures (Lake Redon, Pla and Catalan, 2005) andthe sequence of climate periods recorded in the Iberian Peninsula (Gutiérrez-Elorza and Peña-Monné, 1998; Desprat et al., 2003; Riera et al., 2004; Gil-García et al., 2007 ). The leftcolumn indicates the timing of historical earthquakes recorded in the area and the timing of the two possible earthquakes inferred from the occurrence of facies F3 sediments. ARMand SIRM/χ data for facies association F3 are not shown because they do not indicate variations in magnetite concentration and grain size.

90 J.C. Larrasoaña et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 294 (2010) 83–93

Author's personal copy

before A.D. 350, facies association F1 sedimentswith lowconcentrationsof finer-grained magnetite dominate during both cold (i.e. theSubatlantic Cold Period, 975 B.C.–250 B.C.) andwarm (i.e. the SubborealClimate Optimum, b975 B.C.; the Roman Warm Period, 250 B.C.–A.D.450) periods. Similarly, after A.D. 300, both cold (i.e. the Dark Ages, A.D.450–A.D. 950; the Little Ice Age, A.D.1400–A.D.1850) andwarm (i.e. theMedievalWarm Period, A.D. 950–A.D.1400; the recentwarming period,NA.D. 1850) events are characterized by the alternation of faciesassociation F1 and F2 sediments with variable concentrations ofgenerally coarser magnetite grains. This lack of correlation might berelated to the different response of sedimentary facies dynamics in high-altitude lakes to temperature changes during the Holocene, andunderlines the difficulties in linking globally-established climatic eventswith environmental changes at a specific mountain location with itsown physiographic and environmental responses to climate change.

We have also compared the Lake Barrancs record with a regionalrecord of climate variability based on chrysophyte cysts from LakeRedon (Fig. 5) (Pla and Catalan, 2005), which is located just 8 km eastof Lake Barrancs at a similar altitude (2240 m a.s.l.). Since distributionof chrysophyte cysts is related mainly to altitude, downcore variationsin chrysophyte cysts have been used to estimate a local altitudeanomaly that reflects changes in winter mean temperatures throughtime (negative and positive altitude anomalies indicate warmer andcolder temperatures, respectively) (Pla and Catalan, 2005). The LakeRedon record shows rather small winter temperature anomalies(WTA) between 1700 B.C. and A.D. 300, when mean wintertemperatures usually oscillated between 0.2 °C warmer and −0.5 °Ccolder than present. Around A.D. 300, mean winter temperaturesshow a significant shift toward warmer conditions that lasted till A.D.950, when they decreased ca. 1.4 °C in 100 yr. After A.D. 1050, whenmean winter temperatures were 1.1 °C colder than today, they show aprogressive warming trend that is characterized by large-amplitudeoscillations of up to 1 °C/100 yr. The coincidence between the onset ofglacial and enhanced snow-melt activity in the catchment of LakeBarrancs at A.D. 350 with a significant increase of winter meantemperatures seems to discard winter temperature conditions as themainmechanism leading to glacier development and enhanced snow-melt activity in the Maladeta Massif. Recent (i.e. NA.D. 1800) glacieractivity in the massif has been shown to depend largely on winterprecipitation, which supplies most of the snow necessary for glacierdevelopment, and annual temperatures, which influence glacierablation not only during the summer months (Chueca Cía et al.,2005). Concerning the intensity of snow-melting activity, it is likely todepend on the balance between winter precipitation (which influ-ences snow accumulation) and summer temperatures. We interpretthat the possible response of glacier and snow-melt activity todifferent climatic variables might have conditioned the complexity ofthe Lake Barrancs magnetic record. Following the suggestion ofMatthews and Karlén (1992) and Lie et al. (2004), further studies ofproglacial and “control lakes” (i.e. devoid of recent glacier activity inthe catchment) in the Maladeta Massif are necessary to isolate glacieractivity from snow-melt dynamics and to establish the link betweenboth processes and climate variability.

5.3. Palaeoseismic implications

Keeping in mind that core B5 was drilled near the deepest partof the lake, the proposed formation of the palaeosoils at 350 and430 cblf would imply the nearly complete desiccation of LakeBarrancs at around A.D. 450 and A.D. 300, respectively. Estimatedlake-level drops are of about 13.5 and 14.25 m (i.e. 11 m of waterdepth plus the thickness of the overlying sediments). It might beargued that, instead of such dramatic lake-level drops, formationof the palaeosoils would respond to relatively smaller level drops(i.e. 5–6 m) imposed on a long-term deepening trend, whichmight explain the lake-level fall required to explain the onset of

Unit 1. In any case, these dramatic lake-level drops could beexplained by desiccation during periods of dam failure or periodsof negative hydrological balance. The outlet of the lake corre-sponds to a bedrock barrier originating from glacial over-deepen-ing during the Last Pyrenean Pleniglacial (Figs. 1D, 2A), whichtherefore eliminates the possibility of recent (b3 kyr) events ofdam failure and building required to explain successive episodes oflake-level drop and its subsequent refilling. Concerning the secondpossibility, a comparison with the palaeoclimate proxy producedfrom Lake Redon reveals that the two palaeosoils formed at thebeginning of the warm period starting at ca. A.D. 300 (Fig. 5).Although this warm period might have favoured an increase inevaporation, it is extremely unlikely that it modified the hydro-logical balance of the alpine catchment, including the completemelting of the glaciers, strongly enough to enable desiccation ofthe lake.

In order to find an alternative explanation for the formation of thetwo palaeosoils, it is useful to consider the tectonic and geomorphologicsetting of Lake Barrancs. The NNW–SSE faults that delineate the lakebasin offset glacial polished surfaces attributed to the Last PyreneanPleniglacial, and their surfaces bear glacial striae generated during theYounger Dryas (Ortuño, 2008). In at least one of the fault surfaces, threebands of different bedrock colour mimic the shape of the displacedblocks (Fig. 2C), which indicates different weathering stages linked toexhumation of the fault in three distinctive events.Moreover, theNNW–

SSEoriented faults are displacedbyminorNE–SWoriented faults thatdonot display glacial striae at their surfaces. In addition, the surface of thefault located near the Aneto glacier (Fig. 1C) does not have glacial striaedespite of being located upwards from the LIAmoraine and very close tothe present-day glacier front. All these observations, together with theimportant historical and instrumental seismicity in the area, indicatethat faulting around Lake Barrancs has been active since the LastPyrenean Pleniglacial (20–25 ka. Pallàs et al., 2006) and up to thepresent (Moya and Vilaplana, 1992). Regardless of their timing, thelength of all these faults is always b1.4 km, well below the lowerboundary for rupture length of seismogenic faults (~3.8 km for normalfaults, Wells and Coppersmith,1994). The location of the faults near thebottom of the valley makes a gravitational origin also unlikely (Moyaand Vilaplana, 1992). The most plausible explanation for these faults istheir activation as secondary faults during seismic shaking along theseismogenic Coronas fault, which is located just 4 km SW of LakeBarrancs and has historical and instrumental seismicity (Ortuño, 2008).In addition to seismic shaking, displacement along these faults mighthave been aggravated by uplift of the valley bottom in response tovertical unloading following over-deepening and glacier melting, aprocess that has been studied in similar alpine settings (Ustaszewskiet al., 2008). Considering the widespread evidence for neotectonicactivity in the studied area, we propose an alternative explanation forthe desiccation of Lake Barrancs that involves the drainage of the lakethroughpre-existing fractures openedbyearthquakes. In this regard, thefaults that delineate the western shore of the lake and its centralbathymetric high (Fig. 1D) clearly stand out as the putative fracturescausing drainage of the lake.

The rapid transition from the palaeosoils back to facies F2 sedi-ments indicates that lacustrine conditions were rapidly re-estab-lished, which is consistent with a scenario of rapid sealing of thefractures acting as a subaquatic outlet by sediments dragged by theoutflowing water. If our interpretation is correct, then the LakeBarrancs record shows evidence for the occurrence of two seismicevents along the Coronas Fault in the short time interval between A.D.300 and A.D. 450. To explain the origin of the lowermost deepeningsequence, other earthquakes that occurred before B.C. 1700 might alsoneed to be considered. Since these seismic events are detected bylake-drainage and the formation of palaeosoils, we consider that noestimate of the intensity of the putative earthquakes can be made. Itshould be kept in mind that the surface expression of an earthquake

91J.C. Larrasoaña et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 294 (2010) 83–93

Author's personal copy

depends not only on its magnitude, but also on the location of therupture within the fault and the hypocentral depth. It is thus possible,for instance, that a moderate (MW=5–6) seismic event sufficed toopen a pre-existing fracture and cause drainage of the lake, but that alarger earthquake did not result in significant fracture opening. Thisseems to be the case for the historical Ribagorza event (MW=6.2, A.D.1373), for which no surface rupture has been described in historicalarchives probably as a result of its large hypocentral depth (16 km, seeOrtuño, 2008). Future detailed studies should be carried out in orderto identify sedimentological evidence for seismic shaking in the formof widespread seismites (such as homogenites, slump deposits and/orchaotic deposits; Rodríguez-Pascua et al., 2000; Bertrand et al., 2008).Combined with very-high-resolution seismic surveys, these studiescan be very helpful for testing the hypothesis presented here con-cerning the drainage of Lake Barrancs, and also for providing newinsights into the palaeoseismic history of the Maladeta Massif.

6. Conclusions

The sequence recovered from proglacial Lake Barrancs is composedof three sedimentary facies associations. Facies associations F1 and F2are composed by clays, silts, and sands that have low (b0.4%) TOCcontents, whereas facies association F3 is composed of organic-rich(TOC~3%) silts and sands and occurs in two distinctively dark layers at344–363 and 424–427 centimetres below the lake floor (cblf). Faciesassociations F1 and F2 respond to seasonal changes in sedimentsupply, which is characterized by slow particle settling during thewinter and by the arrival of sediment-loaded homopycnal flows,triggered by snow-melting and glacier outwash, during the warmseason.

Combined low IRM/χ values and high S-ratios indicate thatmagnetite is the main magnetic mineral in facies associations F1and F2. Higher contents of relatively coarser magnetite grains,indicated by high ARM and IRM/χ values, are preferentially associatedwith facies association F2, which is richer in coarser-grainedsediments and displays better-developed horizontal laminationscompared to facies association F1. The predominance of coarser-grained graded sediments (facies association F2) and of overallincreased (but highly oscillating) concentrations of relatively coarsermagnetite grains upwards from 430 cblf suggest the onset of glacialactivity and enhanced snow-melting at ca. A.D. 350. Low concentra-tions of relatively finer magnetite grains and the predominance offacies association F1 sediments in the lower part of the record suggestthat glacier activity was insignificant (if not absent) and snow-meltingwas largely reduced before that age.

Plant macrofossil assemblages and the presence of roots in avertical position indicate that facies association F3 represents in situformation of two palaeosoils at A.D. 300 and A.D. 450. Combined highIRM/χ and low S-ratios of facies association F3 indicate that greigitelikely formed authigenically during degradation of organic matter inthe two palaeosoils. These in situ soils imply sudden and substantiallake-level drops lasting at least for decades. Geomorphological andpalaeoclimatic evidence indicates that neither dam failure nor climatevariability can account for these lake-level falls. An alternativehypothesis proposed here for these lake-level drops is that LakeBarrancs was drained through a pre-existing fracture networkreactivated by earthquakes in the short time interval between A.D.300 and A.D. 450. This scenario is consistent with the widespreadevidence of historical seismic activity in the area and with structuralevidence from the Maladeta Massif, which indicates that formationand subsequent evolution of the Barrancs basin is related to faultreactivation during seismic shaking along the neighbouring seismo-genic Coronas fault. Our results strengthen the view that proglaciallakes constitute excellent archives of past glacier and snow-meltdynamics in alpine settings, and suggest that they might also providea reliable record of seismic activity in young and active mountain belts.

Acknowledgements

We are very grateful to Nikolai Petersen and the staff of thePalaeomagnetic Laboratory of the Ludwig Maximilians Universität(Munich, Germany), where the rock-magnetic analyses were carriedout, for their hospitality, technical assistance, and discussions at theearly stages of this study. This research was supported by projectsAMB93-0814-C02-01 and PB96-0815, and by a MEC Ramón y Cajalcontract (JCL). We are very grateful to Jordi Catalan, who kindlyprovided results from Lake Redon and helped to obtain thebathymetry of Lake Barrancs. We also thank Alberto Sáez and ananonymous reviewer for their useful comments, and Santiago Giraltfor his helpful and efficient editorial handling.

References

Ashley, G.M., 1995. Glaciolacustrine environments. In: Menzies, J. (Ed.), Modern GlacialEnvironments: Processes, Dynamics and Sediments. Butterworth-Heinemann Ltd.,pp. 417–444.

Bates, C.C., 1954. Rational theory of delta formation. AAPG Bull. 37, 2219-2162.Battarbee, R.W., Grytnes, J.-A., Thompson, R., Appleby, P.G., Catalan, J., Korhola, A., Birks,

J.B., Lami, A., 2002. Climate variability and ecosystem dynamics at remote alpineand arctic lakes: the last 200 years. J. Paleolimnol. 28, 161–179.

Becker, A., Ferry, M., Monecke, K., Schnellmann, M., Giardini, D., 2005. Multiarchivepaleoseismic record of late Pleistocene and Holocene strong earthquakes inSwitzerland. Tectonophysics 400, 153–177.

Bertrand, S., Charlet, F., Chapron, E., Fagel, N., De Batist, M., 2008. Reconstruction of theHolocene seismotectonic activity of the Southern Andes from seismites recorded inLago Icalma, Chile, 39°S. Palaeogeogr. Palaeoclimatol. Palaeoecol. 259, 301–322.

Birks, H.H., 2001. Plant macrofossils. In: Smol, J.P., Birks, H.J.B., Last, W.M. (Eds.), TrackingEnvironmental ChangeUsingLakeSediments, Volume3: Terrestrial, Algal, and SiliceousIndicators. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 49–74.

Bloemendal, J., King, J.W., Hall, F.R., Doh, S.J., 1992. Rockmagnetism of Late Neogene andPleistocene deep-sea sediments: relationship to sediment source, diageneticprocesses and sediment lithology. J. Geophys. Res. 97, 4361–4375.

Camarero, L., Masqué, P., Devos, W., Ani-Ragolta, I., Catalan, J., Moor, H.C., Pla, S.,Sanchez-Cabeza, J.A., 1998. Historical variations in lead fluxes in the Pyrenees (NESpain) from a dated lake sediment core. Water Air Soil Pollut. 105, 439–449.

Canfield, D.E., Raiswell, R., Bottrell, S.H., 1992. The reactivity of sedimentary ironminerals towards sulfide. Am. J. Sci. 292, 659–683.

Carrillo, E., Beck, C., Audemard, F.A., Moreno, E., Ollarves, R., 2008. Disentangling LateQuaternary climatic and seismo-tectonic controls on Lake Mucubají sedimentation(Mérida, Andes, Venezuela). Palaeogeogr. Palaeoclimatol. Palaeoecol. 259, 284–300.

Chapron, E., Faïn, X., Magand, O., Charlet, L., Debret, M., Mélieres, M.A., 2007.Reconstructing recent environmental changes from proglacial lake sediments inthe western Alps (Lake Blanc Huez, 2543 m a.s.l., Grandes Rousses Massif, France).Palaeogeogr. Palaeoclimatol. Palaeoecol. 252, 586–600.

Chueca Cía, J., Julián Andrés, A., Saz Sánchez, M.A., Creus Novau, J., López Moreno, J.L.,2005. Responses to climatic changes since the Little Ice Age on Maladeta Glacier(Central Pyrenees). Geomorphology 68, 167–182.

Church, M., Ryder, J.M., 1972. Paraglacial sedimentation: a consideration of fluvialprocesses conditioned by glaciation. Geol. Soc. Am. Bull. 83, 3059–3072.

Copons, R., Bordonau, J., 1994. La Pequeña Edad del Hielo en el macizo de la Maladeta(Alta cuenca del Ésera, Pirineos centrals). In: Martí-Bono, C., García-Ruiz, J.M.(Eds.), El Glaciarismo Surpirenaico: Nuevas Aportaciones. Geoforma Ediciones,Logroño, Spain, pp. 111–124.

Copons, R., Bordonau, J., 1996. El ultimo ciclo glaciar (Pleistoceno superior-Holoceno) enel macizo de la Maladeta (Pirineos Centrales). Rev. Soc. Geol. Esp. 10, 55–66.

Copons, R., Parés, J.M., Dinarès-Turell, J., Bordonau, J., 1997. Sampling induced AMS insoft sediments: a case study in Holocene glaciolacustrine rhythmites from lakeBarrancs (Central Pyrenees, Spain). Phys. Chem. Earth 22, 137–141.

Dahl, S.O., Bakke, J., Lie, O., Nesje, A., 2003. Reconstruction of former glacier equilibrium-line altitudes based on proglacial sites: an evaluation of approaches and selection ofsites. Quat. Sci. Rev. 22, 275–287.

De Batist, M., Chapron, E., 2008. Lake systems: sedimentary archives of climate changeand tectonics. Palaeogeogr. Palaeoclimatol. Palaeoecol. 259, 93–95.

Desprat, S., Sánches-Goñi, M.F., Loutre, M.F., 2003. Revealing climatic variability in thelast three millennia in northwestern Iberia using pollen influx data. Earth Planet.Sci. Lett. 213, 63–78.

Evans, N.G., Gleizes, G., Leblanc, D., Bouchez, J.L., 1998. Syntectonic emplacement of theMaladeta granite (Pyrenees) deduced from relationships between Hercyniandeformation and contact metamorphism. J. Geol. Soc. 155, 209–216.

Fanetti, D., Anselmetti, F.S., Chapron, E., Sturm, M., Vezzoli, L., 2008. Megaturbiditedeposits in the Holocene basin fill of Lake Como (Southern Alps, Italy). Palaeogeogr.Palaeoclimatol. Palaeoecol. 259, 323–340.

Froelich, P.N., Klinkhammer, G.P., Bender, M.L., Luedtke, N.A., Heath, G.R., Cullen, D.,Dauphin, P., Hammond, D., Hartman, B., Maynard, V., 1979. Early oxidation oforganic matter in pelagic sediments of the eastern equatorial Atlantic: suboxicdiagenesis. Geochim. Cosmochim. Acta 43, 1075–1090.

92 J.C. Larrasoaña et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 294 (2010) 83–93

Author's personal copy

García-Ruiz, J.M., Valero-Garcés, B.L., Martí-Bono, C., González-Sampériz, P., 2003.Asynchroneity of maximum glacier advances in the central Spanish Pyrenees.J. Quat. Sci. 18, 61–72.

Gil-García, M.J., Ruiz-Zapata, M.B., Santisteba, J.I., Mediavilla, R., López-Pamo, E., Dabrio,C.J., 2007. Late Holocene environments in Las Tablas de Daimiel (south centralIberian peninsula, Spain). Veg. Hist. Archaeobot. 16, 241–250.

Gutiérrez-Elorza, M., Peña-Monné, J.L., 1998. Geomorphology and Late Holoceneclimatic change in Northeastern Spain. Geomorphology 23, 205–217.

IGN, 2006. Servicio de información sísmica del Instituto Geográfico Nacional de España.(http://www.ign.es/ign/es/IGN/SisIndice.jsp).

King, J., Banerjee, S.K., Marvin, J., Özdemir, Ö., 1982. A comparison of different magneticmethods for determining the relative grain-size of magnetite in natural materials -some results from lake sediments. Earth Planet. Sci. Lett. 59, 404–419.

Larrasoaña, J.C., Roberts, A.P., Musgrave, R.J., Gràcia, E., Piñero, E., Vega, M., Martínez-Ruiz, F., 2007. Diagenetic formation of greigite and pyrrhotite in gas hydrate marinesedimentary systems: results from ODP Leg 204 (southern Hydrate Ridge). EarthPlanet. Sci. Lett. 261, 350–366.

Leblanc, D., Gleizes, G., Lespinasse, P., Olivier, P., Bouchez, J.L., 1994. The maladeta granitpolydiapir, Spanish Pyrenees: a detailed magneto-structural study. J. Struct. Geol. 16,223–235.

Leemann, S., Niessen, F., 1994. Holocene glacial activity and climatic variations in theSwiss Alps: reconstructing a continuous record from proglacial lake sediments.Holocene 4, 259–268.

Leonard, E.M., Reasoner, M.A., 1999. A continuous Holocene glacial record inferred fromproglacial lake sediments in Banff National park, Alberta, Canada. Quat. Res. 51,1–13.

Lie, O., Dahl, S.O., Nesje, A., Matthews, J.A., Sandvold, S., 2004. Holocene fluctuations of apolythermal glacier in high-alpine eastern Jotunheim, central-southern Norway.Quat. Sci. Rev. 23, 1925–1945.

Lowe, J.J., Walker, M.J.C., 1997. Reconstructing Quaternary Environments. 2nd ed.Longman, Essex. 446pp.

Matthews, J.A., Karlén, W., 1992. Asynchronous Neoglaciation and Holocene climatechange reconstructed from Norwegian glacio-lacustrine sedimentary sequences.Geology 20, 991–994.

Monecke, K., Anselmetti, F., Becker, A., Sturm, M., Giardini, D., 2004. Signature of historicearthquakes in lake sediments in central Switzerland. Tectonophysics 394, 21–40.

Montserrat, J., 1992. Evolución glaciar y postglaciar del clima y la vegetación en lavertiente sur del Pirineo: estudio palinológico. Monografías del Instituto Pirenaicode Ecología-CSIC. Zaragoza. 147 pp.

Moya, J., Vilaplana, J.M.,1992. Tectónica reciente enelmacizode laMaladeta, sectordel AltoEsera (Pirineo Central). In: Cearreta, A., Ugarte, F.M. (Eds.), The Late Quaternary in theWestern Pyrenean Region. Servicio Editorial Universidad del País Vasco, Bilbao, Spain,pp. 385–403.

Nesje, A., Bjune, A.E., Bakke, J., Dahl, S.O., Lie, O., Birks, H.J.B., 2006. Holocenepalaeoclimate reconstructions at Vanndalsvatnet, western Norway, with particularreference to the 8200 cal. yr BP event. Holocene 16, 717–729.

Ortuño, M., 2008. Deformación activa en el Pirineo Central: la falla Norte de la Maladetay otras fallas activas. PhD. Thesis, Universitat de Barcelona.

Ortuño, M., Queralt, P., Martí, A., Ledo, J., Masana, E., Perea, H., Santanach, P., 2008. TheNorth Maladeta Fault (Spanish Central Pyrenees) as the Vielha 1923 earthquakeseismic source: recent activity revealed by geomorphological and geophysicalresearch. Tectonophysics 45, 246–262.

Paasche, Ø., Dahl, S.O., Løvlie, R., Bakke, J., Nesje, A., 2007. Rockglacier activity during theLast Glacial–Interglacial transition and Holocene spring snowmelting. Quat. Sci.Rev. 26, 793–807.

Pallàs, R., Rodés, A., Braucher, R., Carcaillet, J., Ortuño, M., Bordonau, J., Bourlès, D.,Vilaplana, J.M., Masana, E., Santanach, P., 2006. Late Pleistocene and Holoceneglaciation in the Pyrenees: a critical review and new evidence from 10Be exposureages, south-central Pyrenees. Quat. Sci. Rev. 25, 2937–2963.

Peters, C., Dekkers, M.J., 2003. Selected room temperature magnetic parameters as afunction of mineralogy, concentration and grain size. Phys. Chem. Earth 28,659–667.

Pla, S., Catalan, J., 2005. Chrysophyte cysts from lake sediments reveal the submillenialwinter/spring climate variability in the northwestern Mediterranean regionthroughout the Holocene. Clim. Dyn. 24, 263–278.

Riera, S., Wansard, G., Julià, R., 2004. 2000-year environmental history of a karstic lakein the Mediterranean pre-Pyrenees: the Estanya lakes (Spain). Catena 55, 293–324.

Roberts, A.P., 1995. Magnetic properties of sedimentary greigite (Fe3S4). Earth Planet.Sci. Lett. 144, 7–17.

Roberts, A.P., Weaver, R., 2005. Multiple mechanisms of remagnetization involvingsedimentary greigite (Fe3S4). Earth Planet. Sci. Lett. 231, 263–277.

Rodríguez-Pascua, M.A., Calvo, J.P., De Vicente, G., Gómez-Gras, D., 2000. Soft-sedimentstructures interpreted as seismites in lacustrine sediments of the Prebetic Zone, SESpain, and their potential use as indicators of earthquake magnitudes during theLate Miocene. Sediment. Geol. 135, 117–135.

Sancho, C., Peña-Monné, J.L., Lewis, C.,McDonald, E., Rhodes, E., 2002. Preliminarydatingofglacial and fluvial deposits in the Cinca River Valley (NE Spain): chronologicalevidences for the Glacial Maximum in the Pyrenees? In: Ruiz-Zapata, B., Dorado-Valiño, M., Valdeolmillos, A., Gil-García, M.J., Bardají, T., Bustamante, I., Mendizábal, I.(Eds.), Quaternary Climatic Changes and Environmental Crises in the MediterraneanRegion. Universidad de Alcalá de Henares, Madrid, Spain, pp. 453–456.

Snowball, I.F., 1991. Magnetic hysteresis properties of greigite (Fe3S4) and a newoccurrence in Holocene sediments from Swedish Lappland. Phys. Earth Planet.Inter. 68, 32–40.

Snowball, I.F., 1993. Mineral magnetic properties of Holocene lake sediments and soilsfrom the Kårsa valley, Lapland, Sweden, and their relevance to palaeoenviron-mental reconstruction. Terra Nova 5, 258–270.

Souriau, A., Pauchet, H., 1998. A new synthesis of Pyrenean seismicity and its tectonicimplications. Tectonophysics 290, 221–244.

Stuiver, M., Reimer, P.J., 1986. A computer program for radiocarbon age calibration.Radiocarbon 28, 1022–1030.

Stuiver, M., Reimer, P.J., Reimer, R.W., 2005. CALIB Radiocarbon Calibration version 5.0.University of Washington Quaternary Isotope Lab. (www.calib.org/).

Thompson, R., Oldfield, F., 1986. Environmental Magnetism. Allen and Unwin, London.Ustaszewski, M., Hampel, A., Pfiffner, O.A., 2008. Composite faults in the Swiss Alps

formed by the interplay of tectonics, gravitation and post-glacial rebound: anintegrated field and modelling study. Swiss J. Geosci. 100/1, 223–235.

Verosub, K.L., Roberts, A.P., 1995. Environmental magnetism: past, present and future.J. Geophys. Res. 100, 2175–2192.

Villar, L., Sesé, J.A., Ferrández, J.V., 1997. Atlas de la flora del Pirineo aragonés, I: consejosde protección de la naturaleza de Aragón. Inst. Est. Altoaragoneses, Huesca.

Wagner, B., Reicherter, K., Daut, G., Wessels, M., Matzinger, A., Schwalb, A., Spirkovski, S.,Sanxhaku, M., 2008. The potential of Lake Ohrid for long-term palaeoenviron-mental reconstructions. Palaeogeogr. Palaeoclimatol. Palaeoecol. 259, 341–356.

Wells, D.L., Coppersmith, K.J., 1994. New empirical relationships among magnitude,rupture length, rupture area and surface displacement. Bull. Seismol. Soc. Am. 84,974–1002.

Wright, H.E., 1980. Cores of soft lake sediments. Boreas 9, 107–114.

93J.C. Larrasoaña et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 294 (2010) 83–93