Lacustrine sedimentation in active volcanic settings: the ...diapiro.ictja.csic.es/~sgiralt/articles/Saez_et_al... · and its location in an active volcanic setting, provides an opportunity

Lacustrine sedimentation in active volcanic settings: the LateQuaternary depositional evolution of Lake Chungara (northernChile)

A. SAEZ*, B. L. VALERO-GARCES� , A. MORENO� , R. BAO� , J . J . PUEYO*, P. GONZALEZ-SAMPERIZ� , S. GIRALT§, C. TABERNER§, C. HERRERA– and R. O. GIBERT**Facultat de Geologia, Universitat de Barcelona, c/Martı Franques s/n, 08028 Barcelona, Spain(E-mail: [email protected])�Instituto Pirenaico de Ecologıa, Consejo Superior de Investigaciones Cientıficas, Apdo 202, 50080Zaragoza, Spain�Facultade de Ciencias, Universidade da Coruna, Campus da Zapateira s/n, 15071 A Coruna, Spain§Instituto de Ciencias de la Tierra ‘Jaume Almera’ – CSIC, c/Lluıs Sole Sabaris s/n, 08028 Barcelona,Spain–Departamento de Ciencias Geologicas, Universidad Catolica del Norte, Casilla 1280, Antofagasta, Chile

ABSTRACT

Lake Chungara (18�15¢S, 69�09¢W, 4520 m above sea-level) is the largest

(22Æ5 km2) and deepest (40 m) lacustrine ecosystem in the Chilean Altiplano

and its location in an active volcanic setting, provides an opportunity to

evaluate environmental (volcanic vs. climatic) controls on lacustrine

sedimentation. The Late Quaternary depositional history of the lake is

reconstructed by means of a multiproxy study of 15 Kullenberg cores and

seismic data. The chronological framework is supported by 10 14C AMS dates

and one 230Th/234U dates. Lake Chungara was formed prior to 12Æ8 cal kyr bp as

a result of the partial collapse of the Parinacota volcano that impounded the

Lauca river. The sedimentary architecture of the lacustrine succession has been

controlled by (i) the strong inherited palaeo-relief and (ii) changes in the

accommodation space, caused by lake-level fluctuations and tectonic

subsidence. The first factor determined the location of the depocentre in the

NW of the central plain. The second factor caused the area of deposition to

extend towards the eastern and southern basin margins with accumulation of

high-stand sediments on the elevated marginal platforms. Synsedimentary

normal faulting also increased accommodation and increased the rate of

sedimentation in the northern part of the basin. Six sedimentary units were

identified and correlated in the basin mainly using tephra keybeds. Unit 1 (Late

Pleistocene–Early Holocene) is made up of laminated diatomite with some

carbonate-rich (calcite and aragonite) laminae. Unit 2 (Mid-Holocene–Recent)

is composed of massive to bedded diatomite with abundant tephra (lapilli and

ash) layers. Some carbonate-rich layers (calcite and aragonite) occur. Unit 3

consists of macrophyte-rich diatomite deposited in nearshore environments.

Unit 4 is composed of littoral sediments dominated by alternating charophyte-

rich and other aquatic macrophyte-rich facies. Littoral carbonate productivity

peaked when suitable shallow platforms were available for charophyte

colonization. Clastic deposits in the lake are restricted to lake margins (Units

5 and 6). Diatom productivity peaked during a lowstand period (Unit 1 and

subunit 2a), and was probably favoured by photic conditions affecting larger

areas of the lake bottom. Offshore carbonate precipitation reached its maximum

during the Early to Mid-Holocene (ca 7Æ8 and 6Æ4 cal kyr bp). This may have been

favoured by increases in lake solute concentrations resulting from evaporation

Sedimentology (2007) 54, 1191–1222 doi: 10.1111/j.1365-3091.2007.00878.x

� 2007 The Authors. Journal compilation � 2007 International Association of Sedimentologists 1191

and calcium input because of the compositional changes in pyroclastic supply.

Diatom and pollen data from offshore cores suggest a number of lake-level

fluctuations: a Late Pleistocene deepening episode (ca 12Æ6 cal kyr BP), four

shallowing episodes during the Early to Mid-Holocene (ca 10Æ5, 9Æ8, 7Æ8 and

6Æ7 cal kyr BP) and higher lake levels since the Mid-Holocene (ca 5Æ7 cal kyr BP)

until the present. Explosive activity at Parinacota volcano was very limited

between c. >12Æ8 and 7Æ8 cal kyr bp. Mafic-rich explosive eruptions from the

Ajata satellite cones increased after ca 5Æ7 cal kyr bp until the present.

Keywords Andean Altiplano, carbonate, diatomite, Holocene, lacustrineecosystem, tephra.

INTRODUCTION

Sedimentary successions of lakes in active vol-canic areas in the Andean Altiplano have provi-ded detailed records of global changes(environmental, climatic and cultural) during theLate Quaternary (Grosjean, 1994; Grosjean et al.,1997, 2001; Valero-Garces et al., 1999b). Many ofthe palaeoenvironmental and palaeohydrologicalfluctuations have been attributed to climatic vari-ability. However, given the active volcanism andrelated tectonics in the region, the role of volcanicprocesses in lacustrine sedimentation requiresevaluation. Volcanic activity may strongly influ-ence lake deposition by several processes:(i) changes in the vegetation of the lake catchmentscaused by increased wildfires and variations insoil conditions favourable to pioneer plants (Hab-erle et al., 2000); (ii) changes in bathymetry andmorphology of the lake basin caused by faultingand the construction and erosion of volcanicstructures (Colman et al., 2002); (iii) variabilityin the sediment supply to the lake and in thechemistry of the waters caused by the supply ofnew volcanic materials in the watershed and bythe direct input of pyroclastic material into thelake (Telford et al., 2004); (iv) addition of hydro-thermal fluids to the lake system changing thechemistry of water and sediments (Valero-Garceset al., 1999a); and (v) ecological impacts on theaquatic ecosystems (Baker et al., 2003).

Lake Chungara (18�15¢S, 69�09¢W, 4520 m a.s.l.),at the base of the active Parinacota volcano (6342 ma.s.l.), is the deepest and highest lacustrine eco-system in the Chilean Altiplano. The sedimentaryrecord, including diatomite-dominated sedimentsand tephra layers, provides a unique opportunityto analyse the interplay of climate change (Gros-jean, 1994; Grosjean et al., 1997, 2001; Valero-Garces, 1999b) and the activity of the Parinacotavolcano during the Holocene (Worner et al., 1988,2000; Clavero et al., 2002, 2004).

A seismic survey and some littoral coresobtained in 1993 facilitated a preliminary recon-struction of the Late Quaternary evolution of thelake (Valero-Garces et al., 1999b, 2000, 2003). InNovember 2002, a coring expedition with theLimnological Research Center (University of Min-nesota, USA) retrieved 15 Kullenberg cores up to8 m long along several transects in the lake basin.Stratigraphical and sedimentological analyses ofthe new cores and integration with the seismicprofiles allowed the three-dimensional (3D)reconstruction of the lake sediment architecture.The chronological framework was based on 14CAMS and 230Th/234U methods. The interpreteddepositional environments may serve as modernanalogues of lacustrine sedimentation of Quater-nary and pre-Quaternary volcanic settings else-were. Diatomaceous sediments occur in a varietyof Quaternary and pre-Quaternary lake succes-sions (Bellanca et al., 1989; Owen & Crossley,1992; Owen & Utha-aroon, 1999; Saez et al., 1999;Zheng & Lei, 1999), often in volcanic-influencedbasins where volcanic silica and hot watersprovide the dissolved silica necessary for diatomgrowth. The variety of diatomite facies from LakeChungara illustrates the large compositionalrange of diatomite in a single basin. Diatomitecommonly occurs with other offshore and littoralfacies such as alluvial, carbonate-rich (Gasseet al., 1987; Bao et al., 1999) and aquatic macro-phyte-rich facies. A multiproxy approach enabledthe roles of tectonics, volcanism and climatechange in lake evolution to be characterized.

GEOLOGICAL SETTING

The Lauca Basin and the origin of LakeChungara

Lake Chungara is located on the NE edge of theLauca Basin (Fig. 1), an intra-arc basin bounded

1192 A. Saez et al.

� 2007 The Authors. Journal compilation � 2007 International Association of Sedimentologists, Sedimentology, 54, 1191–1222

by faults and volcanoes. The peaks of the WesternAndean Cordillera (up to 5000 m a.s.l.) form thewestern margin, and a north–south ridge, madeup of Parinacota, Quisiquisini, Guallatire andPuquintica volcanoes, forms the eastern margin(Fig. 1A). The Lauca Basin is filled with an UpperMiocene to Pliocene, volcaniclastic alluvial andlacustrine sedimentary succession, >120 m thick,that rests unconformably on an Upper Creta-ceous–Lower Miocene volcanic substrate (Kott

et al., 1995; Gaupp et al., 1999). During the LatePleistocene, the Palaeo-Lauca River flowed north-wards from Guallatire to Cotacotani, between theAjoya and Parinacota volcanoes, turned west-wards at Parinacota, then flowed southwardsfor about 100 km, and finally eastwards toBolivia. Lacustrine depositional environmentsalso occurred during the Late Pleistocene in thenorthern areas close to the village of Parinacota(Fig. 1A).

A

B

Fig. 1. (A) Map of the Lauca Basin and the two sub-basins that were created by the collapse of Parinacota volcano:the Chungara and the Cotacotani sub-basins. Lake Chungara occupies the highest sub-basin. It is topographicallyclosed and surrounded by >5500 m a.s.l. volcanoes (modified from Kott et al., 1995). (B) Geological map of the LakeChungara area.

Depositional evaluation in Chungara Lake 1193


Parinacota volcano (6342 m a.s.l.) is a largecomposite stratocone of Late Quaternary age. It isbuilt on an earlier stratocone in the Lauca Basinthat underwent a single catastrophic sector col-lapse event and produced a ca 6 km3 debris-avalanche deposit that covered more than140 km2 (Francis & Wells, 1988; Worner et al.,1988, 2000; Clavero et al., 2002, 2004). The debrisavalanche buried fluvial and lacustrine forma-tions, dammed the Palaeo-Lauca River and isola-ted the Chungara sub-basin (273 km2) to thesouth, which became topographically closedwithout any surface outlets (Fig. 1A). The pro-nounced hummocky topography of the avalanchedeposit also created new lakes within the Parin-acota debris (the present-day Cotacotani lakes;Fig. 1). At the lowest topographic level of bothsub-basins, lakes formed almost immediately: theCotacotani lakes and Lake Chungara (Fig. 1B). Aminimum age for the collapse, between 11 155and 13 500 14C yr BP, was obtained by 14C datingof the post-avalanche Cotacotani lake sediments(Francis & Wells, 1988; Baied & Wheeler, 1993;Ammann et al., 2001) and 18 000 cal yr BP usingHe-exposure techniques (Worner et al., 2000).Clavero et al. (2002, 2004) dated palaeosoil hori-zons and suggested a maximum age of 8000 14C yrBP for the collapse. The differences reflect thedifferent interpretations of the stratigraphic loca-tion of the analysed samples; Clavero et al. (2002,2004) considered that the dated lacustrine sedi-ments were buried by the Parinacota debrisavalanche deposit, the palaeosoils being incor-porated into the avalanche, whereas Francis &Wells (1988) regarded them as post-avalanchelacustrine deposits. Baied & Wheeler (1993) andAmmann et al. (2001) derived the dates from thebase of the Laguna Seca lacustrine sequence (inthe Cotacotani Lake District area; Fig. 1B), whichclearly post-dates the debris avalanche, providinga more reliable minimum age for the collapse.

The presence of numerous moraines and glacio-fluvial deposits indicates that glaciers extendeddown to 4450 m a.s.l. during the Pleistoceneglaciation (N-II moraines mapped by Ammannet al., 2001). Moraines from the eastern slopes ofthe Ajoya volcano reached the Lauca Basin at4450 m a.s.l.; they are overlain by the Parinacotadebris-avalanche deposit and so are older thanthe collapse (Ammann et al., 2001). Pollen stra-tigraphy in Laguna Seca (Baied, 1991; Baied &Wheeler, 1993) indicates a gradual transitiontowards drier and warmer climates starting inthe Late Pleistocene and culminating in the Mid-Holocene dry period. This dry period was

followed by a short, wet episode during the LateHolocene. One of the most significant changes inthe sequence is a transition from carbonate-rich,laminated lacustrine sediments to peat sedimentsthat occurred at ca 7030 ± 245 14C yr BP (Baied &Wheeler, 1993).

The Holocene activity of the Parinacotavolcano

Parinacota has been the only active volcano inthe Lake Chungara watershed during the Holo-cene (Worner et al., 1988; de Silva & Francis,1991). The last eruption was at 290 ± 300 yearsBP (Sieber & Simkin, 2002). According toWorner et al. (1988, 2000), the first phase ofHolocene activity of the Parinacota volcano (upto 6000 yr BP) emplaced andesite Aa type lavaflows that reached the northern edge of the lakewhere lobate lava morphologies are clearlyidentified. The second phase of activity (after6000 yr BP) consisted of a number of eruptionsfrom two small satellite cones (Ajata cones) onthe southern flanks. Lavas from these cones aremore mafic in composition, and three eruptionshave been identified: the Lower Ajata lava(6000 yr BP) with abundant clinopyroxene andhornblende, and the Upper and High Ajata lavas(dated at 3000 and 1400 years ago, respectively)composed of black basaltic andesites with oliv-ine phenocrysts (Worner et al., 2000). Holocenetephra fall-out deposits are scarce: a few layersof fine ash to medium lapilli around LakeChungara are probably associated with the mainvent explosive episodes of Parinacota (Claveroet al., 2004). One major tephra deposit reached15 km to the east of the volcano and other thintephra layers are located close to the west flankof the volcano (Clavero et al., 2004). The LakeChungara succession record, however, containsmany tephra layers.

Lake Chungara

Lake Chungara is located at 4520 m a.s.l. in thenorthern part of the Chungara sub-basin (Fig. 1A),a watershed bounded by high snow-capped vol-canoes (Parinacota, Quisiquisini, Guallatire andAjoya). It has an irregular shape with a maximumlength of 8Æ75 km, maximum depth of 40 m, asurface area of 21Æ5 km2 and a volume of400 · 106 m3 (Herrera et al., 2006). The westernand northern lake margins are steep, formed bythe eastern slopes of the Ajoya and Parinacotavolcanoes. The eastern and southern margins are

1194 A. Saez et al.


gentle, formed by the distal fringe of recentalluvial fans and the River Chungara valley.

The morphology of the lake floor has beendetermined by bathymetric data (Villwock, inDorador et al., 2003) and seismic profiles (Valero-Garces et al., 2000). Six morphological compo-nents can be differentiated along a west-to-eastprofile (Fig. 2A): (i) a narrow ‘western littoralplatform’, ca 175–300 m wide, 0–7 m deep (slope<1�); (ii) a ‘western slope’, 115 m wide and7–20 m deep, dipping 10�; (iii) a 2–3� ‘rise’at the base of the slope, 115–235 m wide and25–40 m deep; (iv) a ‘central plain’, 4 km wide,25–40 m deep; (v) an ‘eastern slope’ of 3�, 200 mwide and between 7 and 25 m deep; and (vi) asubhorizontal (<1�) ‘eastern platform’, 450–850 mwide, between 0 and 7 m deep. All cores are fromsites in the central plain except cores 7, 13 and14, which are from the rise near the western lakemargin, and core 1993, which is from the easternplatform (Fig. 2B).

The region is semi-arid with annual rainfallbetween 345 and 394 mm, and an average tem-perature of 4Æ2 �C. Evaporation is estimated to be1200 mm year)1 (Mladinic et al., 1987). The maininlets to the lake are the Chungara River (300–460 l sec)1) draining the Guallatire Volcano, andthe Ajata and Sopocalane creeks draining the7 million year old Ajoya stratovolcano (Fig. 1B)(Herrera et al., 2006). The lake has no surfaceoutlet. Underground water discharge from LakeChungara to the Cotacotani lake system is thoughtto be ca 25 l sec)1 (Herrera et al., 2006). Waterinputs to the lake (sampled in November 2002)display, on average, the following chemistry:42 p.p.m. HCO�3 , 3 p.p.m. Cl), 17 p.p.m. SO2�

4 ,7 p.p.m. Na+, 4 p.p.m. Mg2+, 8 p.p.m. Ca2+,3 p.p.m. K+ and 22 p.p.m. Si. The Mg:Ca ratio ofwater inputs ranges from 0Æ22 to 0Æ71, dependingon the lithology of the catchment. The lake ispolimictic (Muhlhauser et al., 1995) with strongsurface currents. Temperature profiles measuredin November 2002 showed a gradient from thelake surface (9Æ1–12Æ1 �C) to the lake bottom (6Æ2–6Æ4 �C at 35 m of water depth), with a weakthermocline (0Æ5–0Æ6 �C) at 19 m depth. Oxygenranged from 11Æ9–12Æ5 p.p.m. (surface) to 7Æ6 p.p.m.(bottom). The pH ranged between 8Æ99 (surface)and 9Æ30 (bottom) and the water lake chemistrywas: 448 p.p.m. HCO�3 , 68 p.p.m. Cl), 354 p.p.m.SO2�

4 , 140 p.p.m. Na+, 99 p.p.m. Mg2+, 50 p.p.m.Ca2+, 32 p.p.m. K+ and 2 p.p.m. Si2+ 2. Mag-nesium and sodium, the most conservativeelectrolytes, were concentrated in the lake byevaporation, 30 and 20 times respectively.

Calcium was concentrated by only six times,and dissolved silica was extremely depleted. Thephytoplankton community is made up of a fewspecies; diatoms and Chlorophyceae are domin-ant during the cold and warm seasons, respect-ively. Macrophyte communities in the littoralzone form dense patches that also contribute toprimary productivity. The fauna includesendemic cyprinodontid fish (e.g. 19 species ofOrestias; Villwock et al., 1985). Seasonal meas-urements of conductivity, nitrate, phosphate andchlorophyll reveal changes in productivity and inthe composition of the algal communities relatedmainly to changes in water temperature andsalinity (Dorador et al., 2003). The absence ofhigh-level shorelines at the lake margins suggeststhat the current level of the lake is the highestsince the lake formed.

MATERIALS AND METHODS

In November 2002, a coring expedition retrieved15 cores from Lake Chungara using a raft andKullenberg coring equipment. The physical prop-erties (GRAPE-density, p-wave velocity and mag-netic susceptibility) were measured at 1 cmintervals using a Multi-Sensor Core Logger [atthe Limnological Research Center of the Univer-sity of Minnesota, USA]. Cores were split intotwo, scanned using a colour scanner, and thetextures, colours and sedimentary structures weredescribed. About 400 smear slides were preparedfor the description of the sediment compositionand for semi-quantitative estimates of biogenic,clastic and authigenic mineral contents using apolarizing microscope. Subsamples were takenevery 5 cm for mineralogical, chemical and bio-logical analyses. Total inorganic carbon (TIC) andtotal organic carbon (TOC) contents were sampledevery 5 cm and measured by coulometry atMinnesota and at the USGS (Denver).

X-ray fluorescence (XRF) measurements wereconducted with a 2 mm resolution on cores 10and 11 using the XRF core scanner at theUniversity of Bremen (Germany), using a 60 seccount time, 10 kV X-ray voltage and an X-raycurrent of 1 mA to obtain statistically significantdata. Samples for X-ray diffraction were dried at60 �C for 24 h and manually ground in an agatemill. XRD analysis was performed on an automa-tic X-ray diffractometer: Cu–Ka, 40 kV, 30 mAand graphite monochromator. Initial inspectionof the X-ray diffraction patterns revealed a dom-inant amorphous fraction, characterized by the



A

B

Fig. 2. (A) Bathymetric map of Lake Chungara showing the main morphological units of the lake floor. (B) Map ofLake Chungara depicting littoral alluvial deposits, the main syn-sedimentary faults, the location of the 1993 and 2002cores and the 1993 seismic profiles.

1196 A. Saez et al.


presence of a broad peak centred between 20� and25� 2h. The identification and quantification ofthe different mineralogical species present in thecrystalline fraction followed a standard proce-dure (Chung, 1974). The amorphous fraction wasquantified by measuring total counts using theXRD software. The sample that showed thehighest amorphous peak area (a pure diatomite)was mixed with increasing quantities of purecalcite (5, 10, 20, 40 and 60 wt%) and thepercentage of amorphous content was plottedagainst total counts. A logarithmic function wasadjusted in order to calculate the percentage ofthe amorphous fraction in all samples. Morpho-logical description and mineralogical identifica-tions were undertaken on representative samplesof each facies using SEM-EDS equipment. Sam-ples were dried in two phases: firstly, the mainamount of water was eliminated by capillarityfiltration, and secondly, samples were freeze-dried and vacuum-stored prior to (and after)carbon coating. Secondary and backscatteredelectron images and X-ray emission spectra wereused systematically to characterize the sedimentsamples.

Diatom assemblages were quantified every10 cm in core 11 and in a core taken in 1993.Samples were treated following the procedures ofRenberg (1990). A minimum of 300 diatom valveswere counted at ·1000 with a Nomarski differ-ential interference contrast microscope. Qualitat-ive lake-level reconstructions were based onchanges in the percentages of diatom life forms(Wolin & Duthie, 1999). As the lake level falls, anincrease in shallow-water habitats induces aproliferation of benthic (bottom dwelling forms)and tychoplanktonic diatoms (those usually hav-ing a benthic life form but which can occasion-ally be facultatively planktonic). By contrast,during high lake levels, the percentage ofeuplanktonic diatoms (those having a strictplanktonic character) increases. Pollen samplepreparation followed a classic chemical method(Moore et al., 1991; Dupre, 1992), using high-density liquids and adding Lycopodium (Stock-marr, 1971) to quantify the pollen concentration.Because of the low pollen concentration in thesediments, samples of up to 15 g weight weretaken in core 11 to ensure a sufficient amount ofpollen grains. The extreme abundance of Botryo-coccus cysts hindered statistically significantresults in many samples.

The chronological model for the Chungarasedimentary sequence is based on 17 AMS 14Cdates (Table 1A and B) and four U-series disequi-

librium dates (Table 1C). The AMS 14C dates wereanalysed at the Poznan Radiocarbon Laboratory(Poland, Poz samples) and at the Arizona Radio-carbon Facility (Arizona, AA samples). The pre-sent-day radiocarbon reservoir effect wasdetermined by radiocarbon dating of the dis-solved inorganic carbon (DIC) at Beta AnalyticInc. (Miami, FL, USA). Two litres of lake waterwas filtered using a polycarbon filter to remove allsuspended particles, and a small amount ofNaOH was added, following the standard proto-col of the radiocarbon laboratory. 234U/230Thdating was performed at the Minnesota IsotopeLab (MIL) of the University of Minnesota (Ta-ble 1C). Four carbonate samples were analysed byhigh-resolution ICP-MS using a technique devel-oped at MIL (Edwards et al., 1987; Cheng et al.,2000; Shen et al., 2002). Up to 17 AMS 14C dateswere obtained from: (i) bulk organic matter sam-ples from offshore; and (ii) aquatic organicmacrorests picked from the most marginal cores(Table 1B). The radiocarbon dates were calibratedusing CALIB 5Æ02 software (Reimer et al., 2004).The mid-point of the 95Æ4% (2r probabilityinterval) was selected for constructing the agemodel (Table 1B; calibrated columns age 1 and 2).Seven of the AMS 14C dates were not used for thechronological model because they showed revers-als or were non-coherent (Table 1B).

RESULTS

Dating lacustrine sequences in the Altiplanousing 14C AMS has been hindered by the scarcityof terrestrial macrorests and the large and time-variable reservoir effect (Grosjean et al., 1995,2001; Geyh et al., 1999; Valero-Garces et al.,2000).

Chronology

In Lake Chungara, the present-day reservoir effectcalculated from the DIC age in lake waters is2320 ± 40 14C yr BP (Table 1A), which is close tothe values obtained by Geyh et al. (1999) from thewaters of Lake Chungara (1754 ± 160 14C yr BP)and from living aquatic vegetation (2560 ± 24514C yr BP). However, the reservoir effect in theAltiplano lakes has proved to be highly variablein time because of the interplay of several factors,the lake water volume being the most significant(Geyh & Grosjean, 2000). Accordingly, the correc-tion of the reservoir effect in the Chungarasequence was not performed in the same way in



Table

1.

(A)

AM

S14C

rad

iocarb

on

age

of

the

pre

sen

t-d

ay

DIC

inL

ake

Ch

un

gara

surf

ace

wate

r(s

am

ple

din

2003).

(B)

14C

AM

Sra

dio

carb

on

age

measu

red

inbu

lkorg

an

icm

att

er

an

daqu

ati

corg

an

icm

acro

rem

ain

sof

Lake

Ch

un

gara

core

sam

ple

s.(C

)230T

h/2

34U

ages

measu

red

incarb

on

ate

cry

stals

of

Lake

Ch

un

gara

core

sam

ple

s.

Un

itL

ab.

IDS

am

ple

Typ

eof

sam

ple

d13C

(&)

Un

cali

bra

ted

14C

age

(yr)

Cali

bra

ted

age

(1)

(cal

yr

bp)

min

.C

ali

bra

ted

age

(2)

(cal

yr

bp)

max.

(A)

Beta

188745

DIC

Su

rface

wate

r2320

±40

––

–(B

)2b

Poz8726

14A

-1,

5B

ulk

org

an

icm

att

er

)13Æ6

±0Æ2

4620

±40

2330

±30

2330

±30

2b

Poz8720

11A

-2,

39

Bu

lkorg

an

icm

att

er

)12Æ9

±0Æ4

4850

±40

2620

±130

2620

±130

2b

AA

56904

15A

-2,

48

Aqu

ati

corg

an

icm

acro

rem

ain

s)

25Æ4

66635

±40

4900

±70

4900

±70

2b

Poz8721

11A

-2,

84

Bu

lkorg

an

icm

att

er

)14Æ8

±0Æ2

7290

±80

5700

±80

5700

±80

2a

Poz8723*

11A

-3,

2B

ulk

org

an

icm

att

er

)16Æ1

±0Æ1

8920

±50

––

2a

AA

56903*

15A

-4,

27

Aqu

ati

corg

an

icm

acro

rem

ain

sN

od

ata

9999

±50

––

2a

Poz8724*

11A

-3,

86

Bu

lkorg

an

icm

att

er

)16Æ9

±0Æ1

10

860

±60

1b

Poz7170

11A

-3,1

23

Bu

lkorg

an

icm

att

er

)16Æ8

±0Æ1

8570

±50

7200

±65

9550

±90

1b

Poz8647

11A

-4,

10

Bu

lkorg

an

icm

att

er

)14Æ1

±0Æ3

9860

±60

8310

±115

11

290

±115

1b

Poz7171*

11A

-4,

63

Bu

lkorg

an

icm

att

er

)13Æ6

±0Æ2

11

070

±70

––

1a

AA

56905*

15A

-5,

76

Aqu

ati

corg

an

icm

acro

rem

ain

sN

od

ata

4385

±100

––

1a

Poz8725*

13A

-4,

66

Bu

lkorg

an

icm

att

er

)22Æ9

8810

±50

––

1a

Poz11891*

11A

-4,1

45

Bu

lkorg

an

icm

att

er

)16Æ2

±0Æ4

11

460

±60

––

1a

Poz13032

11A

-5,

41

Bu

lkorg

an

icm

att

er

)22Æ7

±2Æ3

10

950

±80

9650

±170

12

940

±115

1a

Poz11982

11A

-5,

84

Bu

lkorg

an

icm

att

er

)28Æ7

±3Æ7

11

180

±70

9940

±240

13

070

±140

1a

Poz-1

3033

11A

-6,

41

Bu

lkorg

an

icm

att

er

)19Æ6

±1Æ7

12

120

±80

11

240

±170

13

970

±180

1a

Poz-7

169

11A

-6,

79

Bu

lkorg

an

icm

att

er

)23Æ1

±0Æ2

13

100

±80

12

790

±115

15

510

±360

Un

itS

am

ple

IDC

arb

on

ate

mate

rial

238U

p.p

.b.

232T

hp

.p.m

.d2

34U

230T

h/2

38U

230T

h/2

32T

hp

.p.m

.E

rror

Cale

nd

ar

age

(yr

bp)

(C)

2a

14A

-3,

6*

Cry

stal

Ou

tof

scale

–323Æ3

–0

––

2a

13A

-2,

45*

Cry

stal

576Æ4

53Æ9

335Æ1

0Æ0

774

14

1244

4450

2a

13A

-2,

105

Cry

stal

467Æ4

35Æ2

413Æ5

0Æ1

036

23

974

6730

315A

-4,

77*

Sh

ell

717Æ4

203Æ1

320Æ2

0Æ1

607

93445

7720

Sh

ad

ed

an

d(*

)sa

mp

les

are

not

inclu

ded

inth

efi

nal

ch

ron

olo

gic

al

mod

el

(revers

al

date

sor

hig

hvalu

es

of

232T

h).

See

text

for

cali

bra

tion

pro

ced

ure

s.

1198 A. Saez et al.


the lower lacustrine (Unit 1) as in the upperlacustrine deposits (Unit 2; see definition oflithostratigraphic units below).

Correction of dates for the variable radicarbonreservoir effect was based on two assumptions: (i)the present-day lake level is the highest in thehistory of the lake, and (ii) the reservoir effectduring periods of relatively high lake level (Unit2) is similar to that of the present. A constantreservoir effect of 2320 years (the present-dayvalue) is considered for the upper Unit 2 becausethe average lake characteristics (depth, watervolume) probably did not vary much during thedeposition of this unit.

At present, it is only possible to speculateabout the variation in time of the reservoir effectin Unit 1. Following Geyh & Grosjean (2000), thereservoir effect was probably lower than in Unit2 because the lake was, on average, shallowerthan it was during the deposition of the upperunit. This speculation is supported by: (i) theabsence of shorelines presently emergent formerlake; (ii) the diatom composition; and (iii)seismic data (see section below). In this studyit is proposed to correct the Unit 1 dates for twoextreme reservoir age values: a minimum valueof 0 years and a maximum of 2320 years(Table 1B; columns calibrated age 1 and calibra-ted age 2, respectively). A range of age variationfor every radiocarbon date of Unit 1 is obtainedfor these two extreme reservoir effect values. Thefinal chronological model was constructedcalculating the mid-point between these twocalculated extremes.

Four 230Th/234U measurements were carriedout on calcite crystals that appeared in some thinlayers from the Chungara cores (Table 1C). Onlyone 230Th/234U date was finally suitable forestablishing the chronological model becausethree samples were rejected due to the highcontent of 232Th, indicative of high terrigenousinputs.

The uppermost sediments of Lake Chungarawere dated by the 210Pb method (Valero-Garceset al., 2003; Barra et al., 2004). Very low 210Pbactivities at the top of the core (5Æ41 pCi g)1)posed a significant problem in establishing areliable chronology. The values of supported210Pb activity and the calculated fluxes of unsup-ported 210Pb were very low (0Æ18 pCi g)1 and0Æ20 pCi cm)2 year)1, respectively), probably aconsequence of low atmospheric concentrationsof 210Pb in the southern hemisphere and lowrainfall on the Altiplano, resulting in a very lowatmospheric 210Pb flux. With a stable background

activity of 0Æ18 pCi g)1, the average sedimenta-tion rate in lake Chungara is 0Æ033 g cm)2 year)1

or 2Æ9 mm year)1 (Valero-Garces et al., 2003), anda similar value (2Æ4 mm year)1) was obtained byBarra et al. (2004).

Long-term sedimentation rates for the wholesedimentary succession based on AMS 14C and230Th/234U dates are considerably lower(0Æ5 mm year)1 core 11 and 0Æ8 mm year)1 core10; Fig. 3A) than the 210Pb sedimentation rateobtained by Valero-Garces et al. (2003) and Barraet al. (2004). The sedimentation rate depends,among other parameters, upon the depositionalenvironments (central plain, rise, littoral zone)and the sedimentary processes involved and,consequently, are different for each sedimentaryunit (Fig. 3A). Nevertheless, the highest short-term sedimentation rate (2Æ7 mm year)1) is sim-ilar to the modern values, whereas the lowestshort-term sedimentation rate is only0Æ02 mm year)1 (Fig. 3A). Sedimentation rateestimates are lower if volcaniclastic deposits arenot considered, particularly in the upper part ofthe sequence that contains more volcaniclasticlayers (Fig. 3B). However, the main sedimentaryrate trend remains similar (Fig. 3B).

Seismic stratigraphy

The cores retrieved during the 2002 expeditionallowed a more accurate reinterpretation of theseismic profiles obtained in 1993 with the high-resolution survey and a graphic recorder (Valero-Garces et al., 2000). Seismic penetration was notdeep, and the presence of gas blanketing, prob-ably due to methane accumulation, obscured theseismic stratigraphy in some profiles. Despitethese problems, three seismic units were identi-fied.

Seismic unit A (profiles SP5 and SP3)This unit is characterized by the alternation oftwo seismic facies: (i) transparent; and (ii) irre-gular and discontinuous reflectors with cross-cutting relations, and some onlap terminations,suggesting channel morphologies. It has a mini-mum thickness of 5 m in both the eastern riseand the western part of the central plain (Fig. 4Aand B).

Seismic features, stratigraphic location and thesedimentological context of the Chungara Basinsuggest that this seismic unit corresponds toalluvial–fluvial deposits accumulated prior tothe collapse of Parinacota volcano. The depositsare inferred to have derived from the Palaeo-



Lauca River and they probably extended over thewhole central plain.

Seismic unit B (profiles SP5 and SP3)This wedge-shaped unit with a minimum thick-ness of 5 m occurs in the rise of the western

margin. It is characterized by the presence ofshort and mainly convex reflectors with nooverall internal structure. The unit grades later-ally towards the centre of the basin into the lowerpart of seismic unit C (SP3) and, towards thewestern margin of the basin, it is overlapped by

A

B

Fig. 3. Inferred sedimentation ratecurves for the Chungara successionshowing short-term values. Curvesfor the northern (core 10) andsouthern (core 11) sectors of thecentral plain and the rise (core 7) arerepresented. The model was con-structed using 8 AMS 14C and 1230Th/234U radiometric dates, andcorrelation between cores. The AMS14C dates of subunit 2b were cor-rected using the present-day reser-voir effect value (2320 years),whereas dates for subunits 1a and1b were corrected using mean val-ues between 2320 and 0 years. Thedate point for subunit 2a corres-ponds to the 230Th/234U. (A) Sedi-mentation rate curves includingvolcaniclastic deposits; (B) sedi-mentation rate curves without thevolcaniclastic deposits. Arrow (a)indicates the main trend of increas-ing sedimentation rate (see text forfurther explanation).

Fig. 4. (A) Seismic profile SP5, a cross-section perpendicular to the western margin of the lake; (B) seismic profileSP3, a cross-section parallel to the western margin of the lake; (C) seismic profile SP1, a cross-section perpendicularto the eastern lake margin. Locations of the seismic profiles are indicated in Fig. 2B. TWTT, two-way trend timescale.

1200 A. Saez et al.


A

B

C



the upper part of seismic unit C. Core 13 reachedsediments of unit B (basal 15 cm) which consistof massive dark-coloured, coarse gravels andsands, composed of volcanic clasts (up to pebblesize). The location, shape and morphology of theinternal reflectors are similar to those describedby Chapron et al. (2004) and Schnellmann et al.(2005) and are interpreted as wedges of sublacus-trine mass flow deposits that have been deformedby gravity-spreading induced by loading of theslope-adjacent lake floor during mass flow depos-ition. The frequent and high-intensity earth-quakes in this area of the Central Andes wouldhave favoured these mass flow episodes (Fig. 4Aand B).

Seismic unit C (profiles SP5, SP3 and SP1)This unit occurs in the rise (cores 3 and 13), thecentral plain (cores 10 to 12) and in the easternplatform (core 1993) (Fig. 4A–C). It is made up oftransparent seismic facies and parallel horizontalreflectors all over the basin. Its maximum thick-ness is about 10 m in the north of the centralplain and it thins towards both the eastern andwestern margins. Profiles SP5 and SP3 (Fig. 4Aand B) show a discontinuous surface betweenunits B and C, and the upper 2 m of this unit,overlapping unit B. This geometry could beinterpreted as an indication of a former lowerlake-level stand followed by a rise in lake level.Several normal faults in profile SP5 affect onlythe lower part of unit C, and not the uppersediments (Fig. 4A). Unit C becomes thicker closeto some of the faults, suggesting that the activityof those faults increased the accommodation.

Seismic unit C corresponds to the lacustrinedeposits. Correlation of the seismic profiles withcores shows that the main seismic reflectorscorrespond to volcanic layers. The thicknessand number of these reflectors increase towardsthe north (eastern margin in profile SP5 and SP3;Fig. 4A and B), suggesting that volcaniclasticdeposits are derived from the Parinacota volcano.None of the cores reached the base of the lacus-trine sequence in the central plain. The totalthickness of the lacustrine sediments in thecentral plain is estimated to be a minimum of10 m, the maximum recovery being 8 m in cores10 and 11.

Sedimentary units, facies and faciesassociations

Fourteen facies were defined from the coresretrieved in 1993 and 2002 on the basis of

detailed sedimentological descriptions, smearslide observations and compositional analyses(Table 2). Most facies are diatomites differenti-ated by colour and lamination type (facies A, B, C,D and E) and by the relative abundance of someaquatic macrophyte remains (facies I and J).Carbonate-rich facies occur in discrete intervals(facies F, G and H). Alluvial facies are restrictedto the lake basin margins (facies K and L).Volcaniclastic facies (facies M and N) are partic-ularly abundant in the upper part of the cores.Core log correlation is based on matching themagnetic susceptibility peaks corresponding totephra layers (M1 to M14, Fig. 5). The 14 sedi-mentary facies were grouped into seven litho-stratigraphic units corresponding to sevenlacustrine and alluvial facies associations on thebasis of the stratigraphical correlation (Fig. 5Aand B) and seismic stratigraphy (Fig. 4A–C).

Unit 1: shallow to deep offshore depositsUnit 1 deposits correspond to the lower half ofseismic unit C (Fig. 4A–C). Seismic profiles showthat it is thickest in the NW sector of the centralplain, and thins towards the south (4Æ07 m in core11) and west (1 m in core 13), probably onlappingthe Miocene substrate. Two subunits were iden-tified in accordance with facies analyses(Figs 5A,B and 6 sections 3 to 6).

Subunit 1a: finely laminated green diatomite.Subunit 1a reaches at least 2Æ56 m thick in thesouth of the central plain (core 11), and thins tothe east (0Æ65 m, core 15) and west (0Æ58 m, core13) (Fig. 5A). The deposits are ca 14Æ1 to 10Æ2 calkyr BP in age. The average sedimentation rate was0Æ43 mm year)1 in core 11 (Fig. 3A).

Facies A dominates subunit 1a and is com-posed of diatomite with alternating green andwhite fine laminae. One 2 cm thick, glass shard-dominated tephra layer occurs (M1 in cores 11and 14; Fig. 5B). The diatomite laminations are0Æ8 to 10 mm (average 2Æ9 mm) thick (sections 5and 6 in Fig. 6). Smear slide and SEM observa-tions show that white laminae are mostly com-posed of large diatoms. Lamination is defined bychanges in the percentage and size of diatomfrustules (Fig. 7A), and by variations in thecontent of amorphous algal organic matter.Diatom frustules show good preservation and nopreferred orientation (Fig. 7B), suggesting accu-mulation from diatomite bloom episodes. Thegreen laminae are composed of smaller diatomsand larger amounts of amorphous algal organicmatter than white laminae. Alternating green and

1202 A. Saez et al.


Table

2.

Facie

san

dfa

cie

sass

ocia

tion

sin

Lake

Ch

un

gara

sed

imen

ts.

Facie

sU

nit

Com

posi

tion

an

dse

di-

men

tolo

gic

al

featu

res

%D

iato

ms

Oth

er

flora

an

dfa

un

aD

ep

osi

tion

al

en

vir

on

men

t

Dia

tom

ites

(A)

Gre

en

-wh

ite

lam

inate

dd

iato

mit

e

1a

(core

s10

to15)

Lam

inae

thic

kn

ess

ran

ges

from

0Æ8

to10

mm

(avera

ge

2Æ9

mm

);w

hit

ela

min

ae

are

most

lycom

-p

ose

dof

dia

tom

s;so

me

layers

show

inte

rnal

gra

-d

ing;

gre

en

lam

inae

are

com

pose

dof

dia

tom

san

dam

orp

hou

salg

al

or-

gan

icm

att

er.

Low

TO

Cvalu

es

(4–8%

).O

pal

isd

om

inan

tan

dcarb

on

ate

isabse

nt

Ben

thic

:6–43

(av.

14)

Pla

nkto

nic

:30–95

(av.

73)

Poll

en

:ri

ch

inM

yri

op

hyll

um

.P

oor

inB

otr

iococcu

sbra

un

ii.

Som

ebiv

alv

es

an

dgast

rop

od

sre

work

ed

from

litt

ora

lzon

es

Off

shore

,p

luri

-an

nu

al

lam

init

es,

rela

tively

shall

ow

(B)

Bro

wn

-wh

ite

lam

inate

dan

dcarb

on

-ate

-beari

ng

dia

tom

ite

1b

(core

s10

to15)

Lam

ina

thic

kn

ess

ran

ges

from

1Æ1

to6Æ0

mm

(avera

ge

2Æ6

mm

).T

OC

valu

es

are

sim

ilar

tofa

-cie

sA

.O

pal

isd

om

in-

an

t.S

om

ecarb

on

ate

lam

inae.

Low

TO

Cval-

ues

(4–8%

).S

om

ew

hit

ela

min

acon

tain

sabou

t4%

of

carb

on

ate

(calc

ite

an

dara

gon

ite)

Ben

thic

:1–6

(peak

ran

gin

g:

15–31)

Pla

nkto

nic

:92–99

–O

ffsh

ore

,p

luri

-an

nu

al

lam

init

es

(C)

Gre

en

,m

ass

ive,

org

an

ic-r

ich

dia

tom

ite

1b

(core

s10

to15)

Fro

m2

to30

cm

thic

kla

yers

as

am

ass

ive

toban

ded

dia

tom

ite,

rich

inam

orp

hou

sgre

en

ish

org

an

icm

att

er.

Hig

hT

OC

.O

pal

isd

om

inan

tan

dcarb

on

ate

isabse

nt.

Som

ein

terb

ed

ded

car-

bon

ate

layers

(calc

ite

an

dara

gon

ite,

facie

sF

)to

the

top

of

subu

nit

1b

Ben

thic

:1–4

(peak

ran

gin

g:

4–6)

Pla

nkto

nic

:96–99

Very

poor

inB

otr

iococcu

sbra

un

ii.

Rela

tively

abu

nd

an

t,d

isp

ers

ed

biv

alv

es

an

dgast

rop

od

s

Off

shore

,sh

all

ow

er

than

facie

sB

(D)

Bro

wn

red

dis

h,

mass

ive

an

dban

ded

,org

an

ic-r

ich

dia

tom

ite

2a

(core

s1

to8

an

d10

to14)

Cen

tim

etr

eto

decim

etr

eth

ick,

mass

ive

tosl

igh

tly

ban

ded

layers

.H

igh

TO

Cvalu

es

(8–11%

)

Ben

thic

:1–7%

Pla

nkto

nic

:91–99

Ric

hin

B.

bra

un

ii.

Abu

nd

an

tost

racod

sO

ffsh

ore

,fl

uctu

ati

ng

bath

ym

etr

y



Table

2.

(Con

tin

ued

)

Facie

sU

nit

Com

posi

tion

an

dse

di-

men

tolo

gic

al

featu

res

%D

iato

ms

Oth

er

flora

an

dfa

un

aD

ep

osi

tion

al

en

vir

on

men

t

(E)

Dark

gre

yto

bla

ck,

mass

ive

dia

tom

ite

2b

(core

s1

to8

an

d10

to14)

Cen

tim

etr

eto

decim

etr

eth

ick,

mass

ive

tosl

igh

tly

ban

ded

layers

.H

igh

TO

Cvalu

es

(8–10%

)

Ben

thic

:10–21

Pla

nkto

nic

:73–89

Ric

hin

B.

bra

un

ii.

Abu

nd

an

tost

racod

sin

severa

lle

vels

.

Off

shore

.R

ela

tively

deep

Carb

on

ate

-ric

h(F

)W

hit

e,

wh

ite-p

inkis

hcarb

on

ate

-ric

hla

min

ae

2a,

3(a

llcore

s,excep

tcore

1993)

Th

in(1

–5

cm

thic

k)

lam

inae

com

pose

dof

carb

on

ate

min

era

ls(c

al-

cit

e,

magn

esi

um

calc

ite,

ara

gon

ite

an

dd

olo

mit

etr

aces,

up

to50%

)an

dd

iato

ms

Ben

thic

:2–6

Pla

nkto

nic

:94–98

Very

rich

inB

.bra

un

ii.

Abu

nd

an

tost

racod

s.O

ffsh

ore

(G)

Carb

on

ate

bre

ccia

2a

(core

14)

A40

cm

thic

kin

terv

al

com

pose

dof

cem

en

ted

carb

on

ate

cla

sts,

cm

-lon

gw

ith

ad

iato

mit

ebro

wn

matr

ix

No

data

–O

ffsh

ore

dep

osi

tre

moved

from

more

marg

inal

zon

es

(H)

Dark

gre

yto

ligh

tgre

ycarb

on

ate

,si

lt-s

an

dgra

insi

zed

4 (core

1993)

Cen

tim

etr

e-t

hic

kla

yers

com

pose

dof

macro

ph

yte

rem

ain

s,d

iato

ms

an

dabu

nd

an

tcalc

ified

ch

ar-

op

hyte

rem

ain

s(u

pto

25%

carb

on

ate

calc

ite

an

dM

g-c

alc

ite)

Ben

thic

:18–75

Pla

nkto

nic

:25–82

Ric

hin

Ch

ara

sp.,

ost

racod

s,gast

rop

od

san

dbiv

alv

es

Lit

tora

l,C

hara

-dom

inate

dsh

elf

Macro

ph

yte

-ric

h(I

)D

ark

gre

en

mass

ive

dia

tom

ite

3 (core

s9,

15)

Decim

etr

e-t

hic

k,

mass

ive

layer

com

pose

dof

vari

-able

am

ou

nts

of

mm

-to

cm

-lon

g,

macro

ph

yte

re-

main

s(1

0–25%

)in

ad

iato

mit

em

atr

ix.

Som

ein

terb

ed

ded

carb

on

ate

layers

(calc

ite

an

dara

-gon

ite)

No

data

Poll

en

:abu

nd

an

tM

yri

op

hyll

um

.A

bu

n-

dan

taqu

ati

cm

acro

ph

yte

rem

ain

s(n

on

-ch

aro

-p

hyte

),biv

alv

es

an

dgast

rop

od

s

Nears

hore

(J)

Dark

gre

yto

bla

ck,

peaty

3,

4(c

ore

s9,

15,

1993)

Decim

etr

e-t

hic

k,

mass

ive

layer

com

pose

dof

vari

-able

am

ou

nts

of

cen

ti-

metr

e-l

on

g,

macro

ph

yte

rem

ain

s(2

5–50%

)in

ad

iato

mit

em

atr

ix.

Hig

h-

est

TO

Cvalu

es

(10–17%

)

Ben

thic

:72–98

Pla

nkto

nic

:1–25

Ric

hin

aqu

ati

cm

acro

ph

yte

rem

ain

s(n

on

-ch

aro

ph

yte

),ost

ra-

cod

s,biv

alv

es

an

dgas-

trop

od

s

Lit

tora

l,m

acro

ph

yte

-d

om

inate

dsh

elf

1204 A. Saez et al.


Table

2.

(Con

tin

ued

)

Facie

sU

nit

Com

posi

tion

an

dse

di-

men

tolo

gic

al

featu

res

%D

iato

ms

Oth

er

flora

an

dfa

un

aD

ep

osi

tion

al

en

vir

on

men

t

Cla

stic

(K)

Fin

esa

nd

an

dsi

lt5 (s

urf

ace,

core

1993)

Dark

,w

ell

-sort

ed

fin

esa

nd

an

dsi

lt–

–D

elt

ap

lain

of

ou

tcro

pp

ing

delt

as

inw

est

an

dso

uth

ern

lake

marg

ins;

dis

tal

all

uvia

l-fa

nfr

inge

ineast

ern

lake

marg

in(L

)M

ass

ive

gra

vel,

san

dan

dsi

lt6 (c

ore

13)

Mass

ive

gra

vel

com

pose

dby

volc

an

iccla

sts

(up

top

ebble

gra

in-s

ize)

an

dd

ark

,bad

-sort

ed

san

dy

matr

ix

Talu

s-sl

op

em

ass

flow

dep

osi

tson

slop

ean

dri

seof

west

ern

lake

marg

in

Volc

an

icla

stic

(M)

Lap

illi

2,

3,

4(a

llcore

s,excep

tcore

1993)

Lig

ht

gre

yto

gre

en

ish

gre

y,

mass

ive

togra

ded

,0Æ5

to12

cm

thic

kte

ph

rala

yers

.T

hey

are

com

-p

ose

dof

0Æ5

–2

cm

lon

g,

an

gu

lar,

pu

mic

efr

ag-

men

tsw

ith

ou

tan

ym

at-

rix.

Gla

ssis

dom

inan

t;m

inor

am

ou

nts

of

qu

art

z,

an

dm

afi

cm

inera

ls

––

Pyro

cla

stic

fall

-ou

ton

the

lake

from

Pari

nacota

volc

an

oexp

losi

ve

eru

pti

on

s.

(N)

Ash

1a,

2,

3,

4(a

llcore

s)D

ark

gre

yto

bla

ck,

mass

ive

or

poorl

yla

m-

inate

d,

1to

18

cm

thic

ksi

lty

tosa

nd

ygra

in-s

ized

tep

hra

layers

.M

ost

lycom

pose

dof

feld

spar

min

era

ls(a

nd

esi

ne)

an

dse

con

dari

lyam

ph

ibole

,gla

ss,

qu

art

zan

dm

usc

o-

vit

e.

Magn

eti

csu

scep

ti-

bil

ity

peaks

corr

esp

on

dto

these

layers

––

Pyro

cla

stic

fall

-ou

ton

the

lake

from

Pari

nacota

volc

an

oexp

losi

ve

eru

pti

on

s



A

Fig

.5.

(A)

West

–east

stra

tigra

ph

iccro

ss-s

ecti

on

;(B

)n

ort

h–so

uth

stra

tigra

ph

iccro

ss-s

ecti

on

.S

trati

gra

ph

iccorr

ela

tion

sare

base

don

lith

ost

rati

gra

ph

ican

dse

dim

en

tolo

gic

al

cri

teri

a(l

imit

sbetw

een

un

its

an

dso

me

key

levels

an

dfa

cie

s)an

dm

agn

eti

csu

scep

tibil

ity

pro

file

s.In

cro

ss-s

ecti

on

(A)

note

that

there

isn

ola

tera

lcon

tin

uit

yw

ith

Un

it4

(core

1993)

inth

eeast

ern

pla

tform

.T

oim

pro

ve

cla

rity

,th

eh

ori

zon

tal

scale

sare

not

the

sam

ein

the

cen

tral

trou

gh

as

inth

ep

latf

orm

s.

1206 A. Saez et al.


B

Fig

.5.

(Con

tin

ued

)



Fig. 6. Image of core 11 taken with a DMT scanner (LRC, Minnesota). Facies, lithological units and 14C AMSradiocarbon dates and main magnetic susceptibility peaks (M1 to M11 and WAF) are indicated. Black arrows indicateradiocarbon dates in core 11 that are expressed in cal yr BP (Table 1).

1208 A. Saez et al.


white laminae group in 2 to 4 cm thick bundles inwhich the white laminae are variously thicker orthinner. Carbonate is absent (Fig. 8) with theexception of dispersed fragments of gastropodand bivalve shells. No ostracods are present inthese facies. TOC percentages are the lowest(4–8%), increasing towards the top of this unit(Fig. 9).

Euplanktonic diatoms dominate both laminae(average 73%). However, benthic diatoms reachthe highest abundance in this facies of all theoffshore deposits (average 14%). Two intervalsrich in benthic diatoms occur towards the base(16–43%) and top of the unit (11–24%) andcorrespond with the two low percentage peaksof planktonic diatoms in Fig. 9. Pollen concen-

A B

C D

E F

Fig. 7. SEM microphotographs of the diatomite, volcaniclastic and carbonate facies. (A) Facies A: white-greendiatomite laminites. Lamination is marked by diatom content and size; in this case note the sharp change from smallto large diatoms (subunit 1a, core 11, section 6, cm 22). (B) Facies A: detail of the upper part of previous photograph.The good preservation of diatoms together with the absence of a preferred orientation of frustules suggests accu-mulation from a diatomite bloom episode. (C) Facies B: white-brown diatomite laminites (subunit 1b, core 11, section3, cm 14), the high fragmentation and orientation of diatoms suggest some clastic reworking of this deposit. (D) FaciesN: detail of fine-grained tephra (subunit 2a, core 11, section 3, cm 40); note in the centre, a vesicular glass grain. (E)Facies F: detail of calcite crystals in a carbonate layer. Crystals are composed of fibre-bundles (subunit 2a, core 11,section 3, cm 5). (F) Facies H: littoral carbonate-rich sediments with abundant Charophyte (Chara sp.) and ostracodremains. Observe the mineralized intercell areas and calcite-covered charophyte stems (Unit 4, core 1993).



tration is very low in this unit. The presence ofthe aquatic macrophyte Myriophyllum sp. (Fig. 9)suggests relatively shallow water depths, between0Æ4 and 4 m, according to data from Lake Titicaca(Ybert, 1992) and Laguna Miscanti (Grosjeanet al., 2001). The very low content of Botryococ-cus braunii (Fig. 9) is another indication forrelatively low lake levels, given that the develop-ment of this algae peaks at water depths greaterthan 10 m (Carrion, 2002).

Subunit 1a is interpreted as offshore biogenicdeposits accumulated in relatively shallow water.Pollen indicators suggest an increase in waterdepth towards the top of the subunit: Myriophyl-lum is restricted to the base of the subunitwhereas the percentages of B. braunii increasetowards the top. Diatom assemblages indicate adeepening–shallowing water cycle towards thebase of the unit (D1 interval, ca 12Æ6 kyr cal BP;Fig. 9). The interval rich in benthic diatoms at thetop of the unit marks a shallowing–deepeningcycle (S1 interval, ca 10Æ5 kyr cal BP; Fig. 9). Theradiocarbon dates and the number of laminaesuggest a pluri-annual frequency, and the occur-

rence of laminae bundles could be related tomulti-decadal cyclicity processes, such as chan-ges in productivity, water temperature and/orlake volume.

Subunit 1b: laminated and massive browndiatomite and carbonate-rich intervals. Subunit1b has an age between ca 10Æ2 and 7Æ8 cal yr BP,and is composed mainly of two diatomite facies(facies B and C) and some carbonate-rich intervalsat the top. No volcaniclastic layers occur (seeK-area curve in Fig. 8, as an indicator of volcaniccontent due to the presence of potassium involcanic minerals such as amphiboles and feld-spars). The thickness of subunit 1b varies in thecentral plain from 1Æ87 m in the north (core 10) to1Æ15 m in the south (core 11; Fig. 5B). In the riseand the western platform, the thickness thins to0Æ73 m in core 15 and 0Æ62 m in core 13 (Fig. 5A).Sedimentation rate ranges from 0Æ26 (core 11) to0Æ45 mm year)1 (core 10, Fig. 3A).

Facies B is a laminated to thin-bedded diatom-ite (1Æ1 to 6Æ0 mm, average 2Æ6 mm) with brown,white, and some minor reddish laminae. Some

Fig. 8. Magnetic susceptibility, potassium curve and calcium content in tephras (K-area) from a composite log ofcores 10 and 11. CaCO3 % curve and mineralogy curves are from core 11 (Fig. 2). The magnetic susceptibility andpotassium curves mark the presence of tephra layers in the lacustrine succession. Calcium content in the volcan-iclastics and mineralogy curves mainly show changes in tephra composition. Calcium content in tephras showcalcium-rich volcanic inputs in subunit 2a. CaCO3 % curve shows a maximum presence of offshore carbonates in thesame interval.

1210 A. Saez et al.


intervals show laminae bundles <5 cm thick,defined by gradual changes in the thickness ofthe white laminae (either fining or thickeningupwards; Fig. 6, core sections 4 and 5). Brownlaminae are commonly fragmented having ahigher percentage of oriented diatoms (Fig. 7C)and contain more amorphous organic matter. Thewhite laminae are almost exclusively composedof diatoms, although some of them also containcarbonate (<4%), mostly calcite, with only fewintercalated laminae of aragonite. In these carbo-nate-rich laminae, calcite and aragonite crystalsare dispersed in the diatomitic matrix. Calcitecrystals are euhedral and about 50 lm long.Aragonite crystals are acicular, about 10 lm longand 2 lm wide. Planktonic diatoms account for88% of total diatoms and benthic diatoms repre-sent about 9% on average (Fig. 9). However, aninterval from 3Æ15 to 3Æ36 m of sediment depthshows higher percentages of benthic diatoms(between 15% and 31%) corresponding to thehighest carbonate content (4%; Fig. 9). TOC val-

ues are relatively low and resembled those offacies A (4–8%) (Fig. 9).

Facies C occurs in the upper part of subunit 1b,in layers 2 to 30 cm thick, as a massive to bandeddiatomite, rich in green amorphous organic mat-ter, with fragmented and entire bivalve andgastropod shells. Ostracods are absent. Facies Clayers are extensive throughout the basin andwere correlated between all cores at the centralplain (layers A to C in Fig. 5B). A 30 cm thickinterval located at the top of Unit 1 in core 11contains two carbonate layers (up to 32% carbon-ate) of interbedded facies C deposits (Fig. 6,section 3). These layers are rich in authigeniccalcite and aragonite, similar to those describedbelow as facies F, with significant percentages ofbenthic diatoms (5Æ6%; Fig. 8).

Although the pollen content is low, B. brauniipercentages increase from subunit 1b to theoverlying Unit 2 (from 40% up to 90%; Fig. 9),indicating a rise in lake level. Subunit 1b isinterpreted as offshore deposition, relatively

Fig. 9. Biological indicators (TOC %; Botryoccocus and planktonic diatom percent abundances of the total paly-nomorph and diatom assemblages, respectively) and CaCO3 % profiles from core 11 (Fig. 2). The lake level evolutioncurve, based on these proxies, is shown on the right side. Deepening–shallowing episode (D1) and shallowing–deepening episodes (S1, S2, S3 and S4) are indicated. The lake deepened overall. Botryococcus content is expressedas a percentage of the total particles observed in the pollen slides, including pollen, microcharcoal particles, cystsand unknown plant remains. The percentage of planktonic diatoms versus the addition of tycoplanktonic andbenthonic diatoms is shown.



deeper than subunit 1a. However, benthic dia-toms and a carbonate-rich interval suggest aperiod of shallower waters (S2 interval, at ca9Æ8 cal kyr BP; Fig. 9). The depositional bathy-metry of other carbonate-rich intervals is moredifficult to interpret (e.g. the top of subunit 1b, S3interval, ca 7Æ8 cal. kyr BP; Fig. 9) because theydid not correlate with clear changes in diatomitecomposition.

Unit 2: deep (with a shallow event) offshoredepositsThis unit corresponds to the upper part of seismicunit C (Fig. 4A and B) and occurs in the centralplain and the rise. It is composed of massivediatomite with some volcaniclastic layers andcarbonates, and grades laterally to the westand south into alluvial and deltaic deposits, andtowards the eastern platform into macrophyte,organic-rich facies (Unit 4, Fig. 5A). Two sub-units are identified.

Subunit 2a: brown massive diatomite withcarbonate-rich intervals and volcaniclastics.Subunit 2a has an approximate age between ca7Æ8 and 5Æ7 cal yr BP and its thickness variesgreatly from a maximum of 3Æ44 m in the centralplain (core 10), to intermediate values in thecentral areas (2Æ18 m in the rise, core 13, 2Æ29 m inthe eastern sector, core 15), and to the lowestvalues of 1Æ56 m in the southern areas (core 11).Sedimentation rate of subunit 2a ranged from 0Æ94(core 11) to 2Æ18 mm year)1 (core 10; Fig. 3A).Decrease in thickness of subunit 2a (comparecores 10, 12, 05 and 11 in Fig. 5B) and thedisappearance of some of the lower levels ofsubunit 2a (M2; Fig. 5) to the south indicates theexistence of onlap close its base. The onlapsurface would have developed at ca 7Æ5 cal kyrBP. The decrease in thickness of subunit 2A to thewest (compare cores 10, 12, 13 and 14; Fig. 5Aand B) is the result of normal faulting that affectsthe lower part of seismic unit C (Fig. 4A). Subunit2a is composed of massive diatomite (facies D)with intercalated carbonate-rich (facies E and H)and tephra layers (facies M and N).

Facies D is brown-red massive and bandeddiatomite (Fig. 6, sections 3 and 2). Ostracods areparticularly abundant in some layers and bivalvesand gastropods occur at some levels. Planktonicdiatoms commonly represent more than 95% ofthe total diatoms (90% in average) (Fig. 9).B. braunii is much more abundant than in Unit 1(Fig. 9) and TOC values are relatively high(8–11%) (Fig. 9).

Facies F consists of 1–5 cm thick, carbonate-rich layers (25–55% carbonate) that occur in a25 cm thick, carbonate-rich interval (uppermost25 cm of section 3, Fig. 6). The carbonate-richlayers form up to 5% of the thickness of subunit2a and are white to pink and made up of calcite,magnesium calcite, aragonite and some traces ofdolomite (Fig. 8). Stratigraphic correlation showsthat the dominant carbonate mineral in one of thelayers changes along a littoral-offshore transectin the basin. Calcite-rich levels are composedof fibre-bundle crystals (Fig. 7E), fusiformaggregates, rice-shaped crystals and dumbbells(10–200 lm long and 6–80 lm wide), euhedralcrystals (50–100 lm in size) and irregular aragon-ite spheroids (70–140 lm in diameter). Ostracodsare abundant in some levels. Aragonite-richlayers show needle-shaped crystals 10 lm longand 1–3 lm wide. TOC content varies from 4% to8% (Fig. 9). Benthic diatom percentages rangebetween 10% and 21% and they are more abun-dant than in facies D (Fig. 9).

Facies G is a carbonate-breccia in a single layer incore 14, between levels M3 and M4 (Fig. 5B). Thebreccia is about 40 cm thick, has an erosive contactwith the underlying sediments (Fig. 5B), and iscomposed of angular, centimetre-thick carbonateclasts within a brown diatomite matrix. Inorganiccomponents of carbonate clasts are similar to thoseforming carbonate levels of facies F. The texturalfeatures of this breccia suggest a reworked depositderived from a former, thicker and cementedcarbonated level located in more littoral areas.

Volcaniclastic layers in subunit 2a are Ca-richandesitic tephras (Ca content 61–247 p.p.m.;Fig. 8), composed of glass, silicate crystals (pla-gioclases, muscovite, quartz) and pyrite (Fig. 8).Overall, volcaniclastic layers (M2–M7; Fig. 5)account for 20–30% of the total thickness ofsubunit 2a. Two facies can be differentiated: (i)Facies M consists of 0Æ5 to 12 cm thick, massive tograded layers of light grey to greenish grey lapilli;the thickest ones have erosive bases. The lapilliare 0Æ5–2 cm long, angular, pumice fragments,mostly of glass with minor amounts of quartz andmafic minerals, and have no matrix (layer M5 insection 2, Fig. 6). Most layers grade upwards intofine ash (facies N). The occurrence of somescouring at the base of the thickest tephra layerssuggests current transport. Nevertheless, the uni-formity and wide lateral extent of the layers, thecommon presence of graded bedding, the pres-ence of some aquatic fossils and the scarcity ofcurrent-derived features are interpreted as indi-cative of fall-out deposits (e.g. Fisher &

1212 A. Saez et al.


Schmincke, 1984). Two centimetre-thick lapillilayers can be traced throughout the basin (M4 andM5; Fig. 5). The absence of other non-volcaniclacustrine sediments mixed in volcaniclasticlayers suggests that there were no significantbottom currents during their deposition.

(ii) Facies N is made up of dark grey to black,massive to poorly laminated, 1 to 18 cm thicksilty to sandy grain-sized tephra layers, primarilycomposed of andesine with subordinate amphi-boles, glass, quartz and muscovite. Most peaks ofmagnetic susceptibility and potassium content(measured by the XRF scanner) correspond tothese layers (see MS1 and K-area curves in Fig. 8).Glass shards are a common component of thesedeposits (Fig. 7D). Most fine tephra layers arecentimetre-thick, display normal grading (seelayer M5 in section 2; Fig. 6) and contain diatomfossils. There is no evidence of bottom currentreworking. The sedimentological features suggestrapid deposition largely by fall-out. Accumula-tions of bivalve shells occur at the base of severalfine tephra layers. These accumulations are inter-preted as ‘obrution’ deposits caused by the deathof bivalves because of sudden burial under ash(Brett, 1990).

Greater abundance of planktonic diatoms andBotriococcus sp. and the absence of Myriophyl-lum sp. of subunit 2a with respect to Unit 1suggest relatively deeper conditions in the off-shore areas of the lake. However, the peaks inbenthic diatoms, corresponding to the carbonate-rich intervals, point to an episode of shallowerlake levels (S4 interval, ca 6Æ7 cal kyr BP; Fig. 9).Carbonate formation may have been favoured insubunit 2a by factors such as changes in salinity,stronger diatom blooms, and increases in theinput of calcium caused by weathering of cal-cium-rich volcanic material.

Subunit 2b: dark grey-black, massive diatomite.Subunit 2b ranges from ca 5Æ7 cal yr BP to thepresent. Its thickness ranges from 0Æ86–1Æ00 m inthe central plain to 1Æ86 in the platform (core 15)and to 3 m in the rise (core 7; Fig. 5A and B). Thesedimentation rate in the central plain is thelowest recorded in the succession (from 0Æ09 to0Æ34 mm year)1 in core 11; Fig. 3A). The subunitcomprises dark grey to black diatomite (facies E)with abundant volcaniclastic layers (facies M andN). The volcaniclastic deposits represent between47% and 56% of the total thickness of the unit(levels M8–M14; Fig. 5A and B) and they areandesitic and rhyolitic, with amphibole and a lowcalcium content (11–52 p.p.m.). A centimetre-

thick rhyolitic white volcaniclastic sand-gradedeposit (2Æ6 cal kyr BP; WAF in Figs 5A and B,and 6) occurs in all cores and can be correlatedwith a white volcaniclastic deposit reported inthe ‘upper Ajata lava flows’ (Worner et al., 1988,2000).

Facies E has intermediate TOC values (8–10%)and is progressively darker towards the topbecause of the increase in mafic minerals dis-persed in the diatomitic sediment. Planktonicdiatoms dominate the assemblages (95%; Fig. 9)and benthic diatoms show the lowest percent-ages of the entire succession. By contrast,B. braunii percentages are the highest in thesuccession (Fig. 9). No carbonate-rich intervalsor benthic diatom peaks were identified,although ostracods are relatively abundant andthere are some levels with bivalves and gastro-pods. This subunit represents deep offshoredeposition, about 30–40 m, as the present-daylake water conditions.

Unit 3: shallow platform depositsUnit 3 (ca 7Æ8 cal kyr BP to Recent) overlies Unit 1in the central plain. It grades laterally intooffshore facies of Unit 2 towards the west, andinto the littoral facies of Unit 4 towards the east(Fig. 5A). A sandy interval close to the base ofcore 15 (Fig. 5A) might correspond to alluvialdeposits at the eastern lake margin (Fig. 5A). Unit3 sediments are the thickest in the eastern areas ofthe central plain (up to 4Æ5 in cores 15 and 9;Fig. 5A), where they are mainly composed ofdiatomite rich in aquatic macrophyte remains(facies I).

Facies I deposits are made up of massivediatomite with aquatic macrophyte remains, gas-tropod and bivalve shells. Aquatic plant remainsand shells appear dispersed in the diatomitesediment and in discrete, millimetre-thick tocentimetre-thick layers. These discrete intervalsare more abundant towards the base and the topof the unit, and are similar to facies J (describedbelow), which is dominant in Unit 4. Intercalatedtephra layers can be clearly correlated with thoseidentified in the cores from the central plain.Several calcite-rich and aragonite-rich intervalscorrelate with carbonate levels described above asfacies F in subunit 2a.

Two lines of evidence suggest that the deposi-tional environment for Unit 3 was shallower thanthat inferred for the laterally equivalent depositsof the central plain: (i) the palaeo-relief of theChungara basin indicates that the eastern area hasalways been 10–15 m shallower than the western



central area; and (ii) the abundance of aquaticmacrophyte remains. Currently in Lake Chungara,aquatic macrophytes dominate the photic, easternlittoral zone (up to 10–15 m depth) and manymacrophyte remains are transported towards thewest, suggesting that the most probable source ofremains for Unit 3 deposits was the easternlittoral platform (Unit 4).

Unit 4: shallow to very shallow littoral depositsUnit 4 (>5 cal yr BP to Recent) has a minimumthickness of 3Æ69 m (core 1993) occurring onlyin the littoral eastern platform, and disappear-ing towards the slope (Fig. 4C). This unit ismade up of non-charophyte-dominated macro-phytic peaty deposits (facies J), carbonate-rich,charophyte-dominated deposits (facies H) andsome intercalated fine-grained tephra layers(facies N).

Facies J occurs as a decimetre-thick, massivelayer composed of variable amounts of centi-metre-long aquatic macrophyte remains (10–25%)in a diatomite matrix. It contains abundant gas-tropod and bivalve shells. Aquatic plant remainsand shells are scattered within the diatomite andconcentrated in discrete, millimetre-thick to cen-timetre-thick layers. TOC values range between10% and 17% (Valero-Garces et al., 2003).

Facies H occurs in centimetre-thick layerscomposed of organic matter, diatoms and abun-dant calcified charophyte remains. Carbonateaccounts for 25% and comprises calcite andminor amounts of Mg-calcite. The most commoncarbonate components are mineralized intercellareas and calcite-covered charophyte stems (Cha-ra sp.). Gastropods, ostracods and bivalves arealso abundant (Fig. 7F). Diatoms in Unit 4 fromcore 1993 contain more benthic diatoms (18–98%, average 56%) than those from the centralplain core (Units 1 and 2). Several discrete levelscontain higher planktonic diatom percentages(67% to 82%).

Facies H and J accumulated in the shallow(<7 m) littoral areas of Lake Chungara, wherecharophyte and other macrophyte meadows de-veloped, with abundant gastropod and bivalvefauna. The alternation between facies H and Jindicates a succession of prevalent charophytemeadows, and other macrophyte areas, that mayrelate to changes in water depth and salinity. Theeastern littoral platform was only flooded duringthe final stages of the lake’s development. Diatomassemblages also show periods of increasedplanktonic taxa that could correlate with thoseidentified in the central offshore zones (Unit 2).

Carbonate from this unit was clearly bio-mediatedby charophyte activity related to the developmentof Chara meadows in littoral zones.

Unit 5: deltaic and distal alluvial-fan depositsAlluvial deltaic fine clastic deposits of Unit 5(facies K) include: (i) deltaic deposits thataccumulated on the shallow west platform atthe mouths of major creeks draining Ajoyavolcano (Fig. 2B); (ii) deposits from ChungaraRiver at the southern margin of the lake; and(iii) sublacustrine alluvial–deltaic accumulationsin the rise in front of the former deltas (inter-preted from lake floor morphology; Fig. 2B).Facies K is composed of dark, well-sorted finesand and silt.

Shallow deltas developed during the current,high-stand stage of Lake Chungara. Deeper allu-vio-deltaic deposits on the slope and rise formedduring periods of lake level stability in theprevious low-stand stage of Lake Chungara. Lat-eral relationships between these alluvial–deltaicdeposits with mass flow wedge deposits and withlower offshore lake deposits cannot be estab-lished.

Distal clastic deposits of a gentle-dip alluvialfan occur in the eastern margin of the lake, andare very similar to the fine clastic sediments offacies K. The fine sands at the bottom of core 15(Fig. 5A) could be related to these alluvial fans.

Unit 6: Talus-slope mass flow depositsWedge-shaped deposits from sublacustrine massflows on the slope and rise are identified in thewestern lake by seismic profiles (unit B; Fig. 4Aand B). These deposits (facies L) overlie theMiocene substrate and underlie the laminateddeposits of subunit 1a.

Facies L deposits constitute the lowermost40 cm of core 13. They are massive, clast-suppor-ted gravels (up to pebble-size volcaniclasts) in apoorly sorted, dark sandy matrix. This faciescould correspond to pre-lacustrine alluvial mater-ial deposited in the basin and reworked by massflow processes.

The sedimentary architecture

The topography of Parinacota volcano prior toits collapse was mainly responsible for thelocation of the depocentre in the NW of thelake, the lowest point along the Palaeo-LaucaRiver. The central plain and the rise in thenorthern sector had the highest accommodationand, consequently, the thickest lacustrine

1214 A. Saez et al.


succession (10 m). Eastwards and southwards,the thickness of the lower part of the lacustrinesequence thins out to almost 50% because oflower rates of sedimentation (Fig. 5). The inher-ited palaeo-slope of the Palaeo-Lauca Rivervalley suggests that lacustrine deposits probablyonlap the Miocene volcanic substrate south andwestwards.

A general thickening towards the NW of thecentral plain affects the lower part of the sedi-mentary succession (Fig. 5A and B). Synsedimen-tary normal faulting in the north was responsiblefor the occurrence of an onlap surface within thelacustrine sequence towards the south and theeast (Fig. 5A and B). The onlap surface corres-ponds to a sedimentary discontinuity identifiedby comparing detailed sedimentological logs fromcores 10, 11, 12 and 15 (Fig. 5). An interval of40 cm, which includes key level M2, is absent incores 10 and 12 because of this onlap. Subunit 1b,the lower part of subunit 2a and, probably,subunit 1a become progressively thinner towardsthe south and the west (Fig. 5). The uppermostdeposits of subunit 2a and the whole subunit 2bretain their lateral thicknesses (Fig. 5A and B),and indicate cessation of normal faulting betweenthe deposition of key levels M2 and M3 (subunit2a).

Lacustrine deposits thin out towards the upperpart of the eastern and western tali (Fig. 4A). Thethickness of the sediments on the shelves isunknown because seismic penetration was lowand there are no long cores from these areas. Theeastern platform has at least 3Æ69 m of sediments(core 1993). In the western platform, streamsdraining the Ajoya volcano have deposited 20–25 m wide, deltaic sands. Between the deltalobes, the platform is colonized by macrophytesubaqueous vegetation (Myriophyllum sp. andothers). These modern deltas and sublacustrinedeposits suggest that lake level stabilized period-ically during its overall increase. Absence of aseismic record of the oldest alluvial–deltaicdeposits prevents precise correlation with theoffshore sediments. Towards the south, the cen-tral plain changes abruptly to the slope in front ofthe delta generated by the Chungara River and theSopacalane and Chachapay creeks (Fig. 2B).There are no deltaic deposits on the easternplatform because there are no significant streamsentering the lake. A large alluvial fan covering theeastern margin of the lake provides only fine-grained material near the lake shoreline. Thelittoral sediments are fine-grained and the wholeplatform is covered with macrophytes.

DISCUSSION

The lacustrine record of Parinacota eruptions

The absence of coarse volcaniclastic deposits anddeformation structures in Unit 1 lacustrine depos-its suggests that Parinacota collapsed prior to thedeposition of the oldest sediments recovered fromLake Chungara (between ca 12Æ8 and 15Æ5 cal kyrBP). These data fit the chronological framework ofWorner et al. (2000) who assigned an 18 kyr BPage to the collapse. The abundance of volcani-clastic layers in Unit 2 (upper half of seismicunit C; Figs 4, 5A,B and 8) could indicate thatthe highest frequency of explosive eruptionsoccurred between Early–Mid-Holocene (ca 7Æ8 calkyr BP) and recent times. Because of the betterpreservation in the lake (e.g. Saez et al., 1999) thenumber of ash fall deposits in the lacustrinerecord (about 14) is much higher than thatrecognized in the subaerial watershed (describedby Clavero et al., 2004). The Lake Chungararecord demonstrates that volcanic activity – atleast explosive activity – was minimal betweenthe Late Pleistocene and Early Holocene (from ca14Æ1 to 7Æ8 cal kyr BP). Seismic profiles andstratigraphic correlation show that the thickestvolcaniclastic deposits blanketed the entire lakebasin (Figs 4 and 5). Changes in reflector ampli-tude show a general increase towards the nor-thern margin, suggesting that the source of thevolcaniclastic material was Parinacota volcano orits satellite cones. Mafic mineral enrichment involcaniclastic layers from subunit 2a to subunit2b suggests an increasing dominance of mafic-rich eruptions from Ajata satellite cones.

Diatomite deposition in volcanic influencedlakes

The dominance of diatomite in the Lake Chungaraoffshore deposits is a direct consequence of silicaavailability. Supply of dissolved silica reflects thehydrolysis of the dominant volcanic minerals inthe catchment area (mostly plagioclase, biotite,amphibole and pyroxene minerals). Silica in lakewater is extremely depleted compared withinflowing water (see chemical data in a previoussection) as a result of the uptake by diatoms.Nevertheless, changes in diatomite depositioncould also reflect the role of biological processesduring the depositional history of the lake, and asimple link between siliceous productivity andvolcanic silica availability is not straightforward(Telford et al., 2004). Moreover, the high altitude



of Lake Chungara adds some unique features tothese biological processes. Such high-altitudelakes are mainly characterized by high insolationvalues and low relative atmospheric moisture.Given that the photic zone reaches lower parts ofwater column because of the high insolation,benthic algal communities survive at greaterdepths (>20 m) (Vila & Mulhauser, 1987). Onthe other hand, water stratification in high-altitude polimictic lakes occurs for shorter dura-tions. The biota in Lake Chungara must adapt tothis high solar radiation and to abrupt tempera-ture changes. This situation gives rise to contrast-ing periods of strong mixing of the water columnand of water stratification. The phytoplanktoncommunities reflect these changes, with Chlor-ophyceae as the dominant group during warmperiods and diatoms dominating the cold seasons(Vila & Mulhauser, 1987; Dorador et al., 2003).The relative abundance of diatom frustules vs.Chlorophyceace (white vs. green laminae) in Unit1 could reflect temperature controls similar tothose at present: warmer periods would be moreconducive to Chlorophyceae and colder periodswould be more favourable to diatoms. Warmerlake waters during relatively lower lake-levelstages would also be more conducive to Chloro-phyceae producers (massive, green facies C) plusgastropod and bivalves. The increased abundanceof macrophyte-derived remains within the upper-most sediments of Units 2 and 4 marks anincrease in littoral productivity. Extensive flatareas of the Chungara Basin were flooded duringthe final stages of the rise in the lake level. This isparticularly true for the western and easternplatforms, which have slopes <1�. These shelfareas were suitable for macrophyte colonizationand gave rise to quasi-palustrine conditionsbefore they were completely flooded.

Carbonate formation in volcanicallyinfluenced lakes

A variety of mechanisms may trigger carbonateprecipitation in lakes (Kelts & Hsu, 1978; Last,1982). These include a photosynthetic uptake ofCO2 and a consequent increase in carbonate ionsby rising pH, an increase of calcium concentrationin water, temperature effects on carbonate equilib-rium, and the mixing of brines of different compo-sitions. Several processes may work together tocause carbonate precipitation: evaporation of quietwaters may lead to oversaturation of the thinsurface layer during the summer months; photo-synthetic uptake of CO2 may increase the pH in the

chemocline where the organic productivity is highor in the littoral zone where Chara and otheraquatic macrophytes proliferate.

Recent waters in Lake Chungara show a non-conservative behaviour of calcium because of itsprecipitation as carbonate. The chemistry of thewater inputs versus the lake water (see LakeChungara section) also reflects a significant res-ervoir effect for calcium (and other solutes) in thelake water. Lake Chungara is hydrologically open,with variable water supply (by run-off, rainfall,rivers and groundwater) and outflow (as ground-water and evaporation) and, consequently, thewater chemistry reflects the net solute balance.

The different carbonate occurrences in LakeChungara sediments suggest a variety of deposi-tional environments, of which the offshore andlittoral environments are the most important.Moreover, calcium availability in the lake waterscould have fluctuated over time. The lack ofauthigenic carbonates in offshore deposits (lowestsubunit 1a) indicates that the calcium contentwas very low in the earlier stages of lake evolu-tion. Carbonate content increases slowly upwardsalong the offshore deposits of subunit 1b, reach-ing the highest concentration in subunit 2a. Thisevolution is consistent with a trend of increasingsalinity in Lake Chungara.

Biological factors were crucial for carbonateprecipitation, both in the littoral and in theoffshore areas. When calcium was not a limitingfactor, carbonate production in littoral zones (e.g.core 1993) was related mainly to the respiration–photosynthesis balance of Chara and other aqua-tic macrophytes. Carbonate formation in theeastern margin (reworked carbonate facies of core14) may have been related to littoral carbonatecementation as there is no clear evidence ofsubaqueous macrophyte calcification. Carbonateprecipitation in the offshore areas was triggeredby calcium availability and biological activity(mostly algae). Irregular spacing among carbonatelaminae in subunit 1b suggests that precipitationevents were not the result of regular or seasonaltemperature fluctuations or regular biota blooms.

Some of the carbonate-rich layers in the off-shore sediments of subunit 2a also occur withcalcium-rich (plagioclase-bearing) volcaniclasticlayers (see calcium evolution in tephra deposits;Fig. 8); this suggests a link between volcaniceruptions and carbonate production in the lake.Weathering of the calcium-rich volcanic layers inthe lake and in the catchment could have causedan increase in the calcium content in the lakewater, favouring carbonate precipitation during

1216 A. Saez et al.


deposition of subunit 2a. The increase in watervolume during deposition of subunit 2b (seeplanktonic vs. benthonic diatom evolution;Fig. 9) and the decrease in calcium input to thelake could have reduced the offshore precipita-tion of calcite (see calcium evolution in tephradeposits Fig. 8).

Although calcite (with some magnesium calcite)is the main carbonate in Lake Chungara, aragoniteis also present as thin layers offshore. Aragonitelaminae occur in many lakes, and aragoniteprecipitation has been related to fluctuations ofthe Mg:Ca ratio within lake water (Utrilla et al.,1998; Yu et al., 2002). Precipitation of aragonitelaminae in a meromictic saline lake in Canada hasbeen linked to periodic mixing with waters withhigher Ca2+ concentration, modifying the Mg:Caratio (Last & Schweyen, 1985). In Lake Chungarathe occurrence of isolated aragonite laminae(subunit 1b) and aragonite layers (subunit 2a andUnit 3) that grade laterally into calcite-rich layersmay reflect local and small fluctuations of theMg:Ca ratio in the lake water as a result ofdifferences in Mg:Ca values in the water inputs(see Lake Chungara section).

Controls on the sedimentation rate changes

The sediments of Lake Chungara record signifi-cant vertical and lateral changes in the sedimen-tation rate (Fig. 3). Overall, the rate of thesedimentation diminished, with a particularlymarked decrease in the upper part of subunit 2a(see arrow a in Fig. 3B). The highest short-termsedimentation rate occurred in subunit 1a(2Æ25 mm year)1 in core 11; Fig. 3A) and thelowest rate occurred in subunit 2b (0Æ09 in core11; Fig. 3A). The main factor that controlled thistrend in the central plain of the lake was normalfaulting, which affected the lower part of thesuccession (up to 2 m of vertical displacement).In addition, changes in pyroclastic supply and indiatom productivity would have influenced thesedimentation rate. The sedimentation rate in thelower part of the succession was enhanced bythe accommodation generated by the synsedi-mentary faulting. Moreover, because photic con-ditions reach deeper areas at the bottom of thelake, diatom productivity could have been higherduring the lowstand period (Unit 1 and subunit2a) than during the highstand period (subunit 2b).The maximum sedimentation rates attained insubunit 2a correlate with: (i) the main activity offaults, as determined by faulting and lateralchanges in thickness in lower part of subunit 2a;

(ii) the deposition of a significant volume ofvolcaniclastic deposits; and (iii) the occurrence ofshallow lake-level episodes (S3 and S4).

Lateral changes in sedimentation rates can beinferred by comparing the sedimentation ratecurve of cores 10, 11 and 7 (Fig. 3). The averagesedimentation rate in the rise (core 7) is greaterthan in the central plain (cores 10 and 11), andsedimentation rate values in the northern centralplain (core 10) exceed those in the southerncentral plain (11). The sedimentation rate in thenorthern central plain was thought to have beenhigher because of the accommodation generatedby the activity of normal faults. The high sedi-mentation rate in the rise was probably due to theclastic inputs at the foot of the adjacent westernslope and due to the high diatom productivity inshallow rise areas.

DEPOSITIONAL HISTORY: TECTONICS,VOLCANIC AND HYDROLOGICALFORCING

Fluctuations in the level of Lake Chungara duringthe Late Glacial and the Holocene have beenreconstructed using: (i) diatom ratios (euplank-tonic vs. tychoplanktonic plus benthic species);(ii) the presence and percentages of littoral macro-phytes as Myriophyllum sp.; (iii) percentages ofthe alga Botriococcus sp.; (iv) variations in sedi-mentary facies (e.g. facies C); and (v) the occur-rence of carbonate. A progressive increase in lakelevel in core 11 (central plain) was defined by theincrease in euplanktonic diatoms and Botryococ-cus and by the decrease in macrophyte remains(Fig. 9). Core 11 did not reach the substrate, so thebathymetry of the early stages of Lake Chungara isunknown. Many lake sequences record shallowbasins that became deeper with time (Anadonet al., 1991; Saez & Cabrera, 2002), but thecatastrophic origin of Lake Chungara, because ofthe impoundment of the river, caused rapid earlydevelopment of deep depositional environments.This general increase in water volume during thelast 15 000 cal yr BP could be attributed to higherwater inputs caused by enlargement of the water-shed. However, there is no geomorphologicalevidence of substantial changes in the watershedafter the collapse of Parinacota volcano, so theincrease in water volume is probably related tochanges in the hydrological balance caused byclimate change.

The overall increase in lake level is punctuatedby a Late Pleistocene deepening episode (D1;



Fig. 9) and by at least three shallowing episodesduring the Lower–Mid-Holocene (S1, S2 and S4;Fig. 9), identified by the changes in plank-tonic:benthic diatom ratios. Another shallowingepisode (S3?; Fig. 9), in the Mid-Holocene, is notso well-defined, but it shows an increase inbenthic diatoms and is associated with carbon-ates. The following stages have been interpretedin the sequence.

Stage 1: Late Pleistocene to onset of Holocene

The Lake Chungara Basin was drained by thePalaeo-Lauca River from north to south prior tothe Parinacota debris avalanche (Fig. 10A). Thedebris avalanche created an endorheic basin withsteep northern and western margins, and the NWpart of the basin was almost immediately occu-pied by a lake. Shortly afterwards, the lakeexpanded and occupied the entire central plain,

and lacustrine facies (subunit 1a) were depositedon the alluvial sediments of the Lauca River.Maximum water depth was 20 m, as indicated bythe absence of sediments of subunit 1a in theeastern lake margin sector. The waters were diluteand calcium-poor. The photic zone extended upto the lake bed and benthic diatoms and greenalgae proliferated. Diatoms and green algae dom-inated sedimentation, and the accumulation ratewas high. Green laminae represent periods ofstability conducive to diatom and green algaedeposition. These periods were interrupted bypluri-annual strong water mixing conditions,which enhanced nutrient availability, triggeringextensive diatom blooms (e.g. Harris, 1986) thatdeposited to form the white laminae, exclusivelymade up of diatom skeletons. After the deepeningevent D1, the dominance of planktonic diatomssuggests that generally deeper conditions pre-vailed in the lake, punctuated by a shallowing

Fig. 10. Evolution of Lake Chungara in the last 14 000 years. From Stage 1 to Stage 4, the lake expanded anddeepened. Active faulting affected lake sediments during Stages 2 and 3. The main volcanic activity occurred duringStages 3 and 4. Calcium-rich volcaniclastic deposits are recognized only in Stage 3. Main aquatic macrophyteproduction was mostly located in the eastern margin of the lake.

1218 A. Saez et al.


event at ca 11Æ7 cal kyr BP. Volcanic activity afterthe collapse was minor (only one tephra layer)and there was little tectonic activity.

Stage 2: Early Holocene

During the Early Holocene, subunit 1b sedimentswere deposited in the central plain (Fig. 10B).The overall lake-level rise continued reaching adepth of 30 m, although minor lake-level fluctu-ations occurred. The eastern lake margin wasprogressively flooded, and two fluctuations ofshallowing–deepening lake level occurred at ca9Æ8 cal kyr BP and at the end of this stage (ca7Æ8 cal kyr BP). Sedimentation was mainly bio-genic dominated by diatoms and other algalproducers. Green and brown laminae representmixed green algae and diatom sediments; thebrown colours could indicate oxidation of thedeposits during periods of more oxygenatedwaters that could correspond to lower lake levels.White laminae represent diatom blooms withpluri-annual periodicity. Conditions for carbon-ate precipitation occurred during several shortintervals. Some of these periods were associatedwith shallowing lake episodes. No volcanicevents were recorded during this period.

Stage 3: Mid-Holocene

During Stage 3, subunit 2a was deposited in thecentral plain (Fig. 10C). Lake level greatly fluctu-ated and the western platform was periodicallyflooded. A shallower lake level episode occurredat ca 6Æ7 cal kyr BP. Carbonate production peakedin offshore areas during the initial and middleparts of this stage. The flooding of the platformsincreased the area of littoral subenvironments inthe lake, which were the most carbonate-produ-cing habitats. Increased explosive volcanic activ-ity could also have been conducive to carbonateproduction due to calcium-rich volcaniclasticinputs, and other changes in water composition.The sedimentation rate was the highest duringthis period and could be the result of fault activity(increased accommodation), biogenic activity(high TOC) and carbonate productivity.

Stage 4: Mid-Holocene to Late Holocene toRecent

During Stage 4, subunit 2b sediments weredeposited on the central plain (Fig. 10D). Thelake level increased and reached its highest level,although it probably fluctuated around this high

stand (30–40 m). The eastern platform was com-pletely flooded and littoral deposition was initi-ated. Biological productivity in the bottom watersand water–sediment interphase diminishedbecause the photic zone did not reach the lakebottom. Offshore zone carbonate productionceased, possibly due to increased dilution withthe greater lake volume and smaller calciuminputs. Sedimentation rates decreased as a resultof lower biological activity. Carbonate productiv-ity continued in the littoral zone (core 1993) andthere is evidence of carbonate deposition in thenorthern margin of the central plain (20 m waterdepth) during the last 500 years (Valero-Garceset al., 2003).

CONCLUSIONS

1 Lake Chungara was formed by the collapse ofParinacota volcano prior to 12Æ8 cal kyr BP. Seis-mic profiles and core stratigraphy show alluvial–fluvial deposits beneath the 10 m thick, lacustrinesuccession of Lake Chungara. The age of theoldest dated lake sediments (between ca 12Æ8 and15Æ5 cal kyr BP) is consistent with the chronolo-gical framework assigning an 18 kyr BP age to theParinacota collapse (Worner et al., 2000).

2 The sedimentary architecture of the Holocenelacustrine succession was controlled by: (i) theinherited palaeo-relief; and (ii) changes in theaccommodation caused by lake-level fluctuationsand tectonic subsidence. The first factor deter-mined the location of the depocentre in the NWsector of the central plain. The second factorcaused the expansion of lacustrine depositiontowards the eastern and southern margins, andthe accumulation of sediments on the elevatedmarginal platforms.

3 A period of normal fault activity occurredduring the deposition of Unit 1 and the lower partof Unit 2. The deposits were affected by normalfaults with displacements of a few metres in thenorthern sector of the basin. The faultingincreased the accommodation and the sedimen-tation rate in the northern depocentre, and gen-erated an onlap surface within the lacustrinesuccession.

4 Diatoms were the main producers in LakeChungara; this reflects the availability of silica inthe volcanic setting. Increases in benthic diatomproductivity during lowstand periods led to par-allel increases in the sediment rate. The relativelyhigh TOC values (8–11%) suggest that other algaesuch as Botriococcus sp. could also have played a



significant role in the depositional history of thelake. The biogenic productivity in littoral zoneswas dominated by charophytes (Chara sp.) andother macrophytes (Myriophyllum sp.). Green andbrown laminae reflect periods of depositioncharacterized by the dominance of diatom andother algae producers whereas white laminaeindicate diatom blooms.

5 Well-preserved volcaniclastic sedimentaryrecords in Lake Chungara have allowed a recon-struction of the local eruptive history after thecollapse of Parinacota volcano. Some lava flowsoccurred in the Early Holocene, but explosiveeruptions were rare during the Late Pleistoceneand Early Holocene (ca >12Æ8 to 7Æ8 cal kyr BP).From Mid-Holocene (ca 7Æ8 cal kyr BP) to thepresent, explosive activity from Ajata satellitecones increased. The first phase of this eruptiveperiod was mafic-poor and calcium-rich (subunit2a). After ca 5Æ7 cal kyr BP, the composition of thevolcaniclastic materials became mafic-rich andcalcium-poor (subunit 2b).

6 Carbonate in volcanic-influenced lakes hasdifferent origins to that in lakes in non-volcanicsettings. Littoral productivity associated withCharophytes peaked when suitable shallowshelves were available for colonization. Offshorecarbonate precipitation in Lake Chungara peakedduring Mid-Holocene, approximately at ca 7Æ8and 6Æ4 cal kyr BP. Two main factors favoured thecarbonate formation: (i) the rise in lake watersalinity due to evaporation; and (ii) increasedinput of calcium due to emplacement of vol-caniclastic material from Parinacota volcano intothe lake and into the broader catchment.

7 Changes in the aquatic flora (diatoms andaquatic macrophytes) and the presence of car-bonate layers reflect the lake level fluctuations: (i)a deepening episode at Late Pleistocene (ca12Æ6 cal kyr BP); (ii) four shallowing episodes atEarly to Mid-Holocene (ca 10Æ5, 9Æ8, 7Æ8 and6Æ7 cal kyr BP); and (iii) higher lake levels at Lateto Mid-Holocene (since ca 5 cal kyr till present).Laminated diatomite from Unit 1 shows a pluri-annual to decadal frequency of biogenic depos-ition.

ACKNOWLEDGEMENTS

The Spanish Ministry of Science and Educationfunded the research at Lake Chungara through theprojects ANDESTER (BTE2001-3225), BTE2001-5257-E and LAVOLTER (CGL2004-00683/BTE).The Limnological Research Center (U of MN,

USA) provided the technology and expertise toretrieve the cores, and the facilities for the InitialCore Descriptions. We are very grateful to CONAF(National Forestry Authority of Chile), the Direc-cion General de Aguas de Chile (Water Authority)and to the staff at the Lauca National Park. ThePalaeostudies Programme (European ScienceFoundation) provided the funding necessary tocarry out the analysis at the University of Bremenby A. Moreno in December 2003. We wish tothank Walter Dean (USGS, Denver) for help withelemental carbon analyses. We are particularlyindebted to D. Schnurrenberger, A. Myrbo and M.Shapley (LRC, USA) for ensuring the success ofthe coring expedition to the lake. Thanks toG. Worner for suggestions on volcanics. P. Anadon,C. Beck, and M. Branney are acknowledged forsuggestions to improve and clarify the earlierversion of this paper.

REFERENCES

Ammann, C., Jenny, B., Kammer, K. and Messerli, B. (2001)

Late Quaternary Glacier response to humidity changes in

the arid Andes of Chile (18–29 �S). Palaeogeogr. Palaeo-climatol. Palaeoecol., 172, 313–326.

Anadon, P., Cabrera, L., Julia, R. and Marzo, M. (1991)

Sequential arrangement and asymmetrical fill in the

Miocene Rubielos de Mora Basin (Northeast Spain). In:

Lacustrine Facies Analysis (Eds P. Anadon, L. Cabrera-L

and K. Kelts), Spec. Publ. Int. Assoc. Sedimentol., 13, 257–

275.

Baied, C.A. (1991) Late-Quaternary Environment, Climate

and Human Occupation of the South-Central Andes.

Ph.D. Dissertation, University of Colorado, Boulder, CO,

131 pp.

Baied, C.A. and Wheeler, J.C. (1993) Evolution of High An-

dean Puna ecosystem environment, climate and culture

change over the last 12 000 years in the Central Andes. Mt.Res. Dev., 13, 145–156.

Baker, P., Williamson, D., Gasse, F. and Gibert, E. (2003)

Climatic and volcanic forcing reveleated in a 50 000 year

diatom record from Lake Massoko, Tanzania. Quatern. Res.,60, 368–376.

Bao, R., Saez, A., Servant-Vildary, S. and Cabrera, L. (1999)

Lake-level and salinity reconstruction from diatom analyses

in Quillagua Formation (Late Neogene, Central Andean

Forearc, Northern Chile). Palaeogeogr. Palaeoclimatol. Pal-

aeoecol., 153, 309–335.

Barra, R., Pozo, K., Munoz, P., Salamanca, M.A., Araneda, A.,Urrutia, R. and Forcadi, S. (2004) PCBs in a dated sediment

core of a high altitude lake in the Chilean Altiplano: The

Chungara Lake. Fresenius Environ. Bull. 13, 83–88.

Bellanca, A., Calvo, J.P., Censi, P., Elizaga, E. and Neri, R.(1989) Evolution of lacustrine diatomite carbonate cycles of

miocene age, southeastern Spain - Petrology and isotope

geochemistry. J. Sed. Petrol., 59, 45–52.

Brett, C.E. (1990) Obrution deposits. In: Palaeobiology. ASynthesis (Eds D.E.G. Briggs and P.R. Crowther), pp. 239–

243. Blackwell Scientific Publications, Oxford.

1220 A. Saez et al.


Carrion, J.S. (2002) Patterns and processes of Late Quaternary

environmental change in a montane region of southwestern

Europe. Quatern. Sci. Rev., 21, 2047–2066.

Chapron, E., Van Rensbergen, P., De Batist, M., Beck, C. and

Henriet, J.P. (2004) Fluid-escape features as a precursor of a

large sublacustrine sediments slide in Lake Le Bourget, NW

Alps, France. Terra Nova, 16, 305–311.

Cheng, H., Edwards, R.L., Hoff, J., Gallup, C.D., Richards,D.A. and Asmeron, Y. (2000) The half-lives of uranium-234

and thorium-230. Chem. Geol., 169, 17–33.

Chung, F.H. (1974) Quantitative interpretation of X-Ray dif-

fraction patterns of mixtures. J. Appl. Cryst. 7, 519–531.

Clavero, J.E., Sparks, S.J. and Huppert, H.E. (2002) Geological

constraints on the emplacement mechanism of the Parina-

cota debris avalanche, northern Chile. Bull. Volcanol., 64,40–54.

Clavero, J.E., Sparks, S.J., Polanco, E. and Pringle, M. (2004)

Evolution of Parinacota Volcano, Central Andes, Northern

Chile. Rev. Geol. Chile, 31, 317–347.

Colman, S.M., Kelts, K.R. and Dinter, D.A. (2002) Depositional

history and neotectonics in Great Salt Lake, Utah, from

high-resolution seismic stratigraphy. Sed. Geol., 12, 61–78.

Dorador, C., Pardo, R. and Vila, I. (2003) Temporal variations

of physical, chemical and biological parameters of a high

altitude lake: the case of Chungara Lake. Rev. Chilena Hist.

Nat., 76, 15–22.

Dupre, M. (1992) Palinologıa. Soc. Espanola de Geom-

orfologıa, Logro�no, 30 pp.

Edwards, R.L., Chen, J.H. and Wasserburg, G.J. (1987)238U-234U-230Th-232Th systematics and the precise meas-

urement of time over the past 500 000 years. Earth Planet.

Sci. Lett., 81, 175–192.

Fisher, R.V. and Schmincke, H.U. (1984) Pyroclastic Rocks.

Springer-Verlag, Berlin, 472 pp.

Francis, P.W. and Wells, G. (1988) Landsat thematic mapper

observations of debris avalanche deposits in the Central

Andes. Bull. Volcanol., 50, 258–278.

Gasse, F., Fontes, J.C., Plaziat, J.C., Carbonel, P., Kaczmarska,I., Deckker, P.D., Soulie-Marsche, I., Callot, Y. and

Dupeuble, P.A. (1987) Biological remains, geochemistry

and stable isotopes for the reconstruction of environmental

and hydrological changes in the holocene lakes from

North Sahara. Palaeogeogr. Palaeoclimatol. Palaeoecol., 60,1–46.

Gaupp, R., Kott, A. and Worner, G. (1999) Palaeoclimatic

implications of Mio-Pliocene sedimentation in the high-

altitude intra-arc Lauca Basin of northern Chile. In: Ancient

and Recent Lacustrine Systems in Convergent Margins (Eds

L. Cabrera and A. Saez), Palaeogeogr. Palaeoclimatol.

Palaeoecol., 151, 79–100.

Geyh, M.A. and Grosjean, M. (2000) Establishing a reliable

chronology of lake level changes in the Chilean Altiplano: a

result of close collaboration between geochronologists and

geomorphologists. Zbl. Geol. Palaont. Teil I, 7/8, 985–995.

Geyh, M.A., Grosjean, M. Nunez, L. and Schotterer, U. (1999)

Radiocarbon reservoir effect and the timing of the late-

glacial/early Holocene humid phase in the Atacama Desert

(Northern Chile). Quatern. Res., 52, 143–153.

Grosjean, M. (1994) Paleohydrology of the Laguna Lejıa (north

Chilean Altiplano) and climatic implications for late-glacial

times. Paleogeogr. Paleoclimatol. Paleoecol., 109, 89–100.

Grosjean, M., Geyh, M.A., Messerli, B. and Schotterr, U.(1995) Late-glacial and early Holocene lake sediments,

groundwater formation and climate in the Atacama Altip-

lano. J. Paleolimnol., 14, 214–252.

Grosjean, M., Valero-Garces, B.L., Geyh, M.A., Messerli, B.,Schreier, H. and Kelts, K. (1997) Mid and Late Holocene

Limnogeology of Laguna del Negro Francisco, northern

Chile, and its paleoclimatic implications. Holocene, 7, 151–

159.

Grosjean, M., van Leeuwen, J.F., van der Knaap, W., Geyh,M., Ammann, B., Taner, W., Messerli, B., Nunez, L., Vale-ro-Garces, B. and Veit, H. (2001) A 22 000 14C yr B.P.

sediment and pollen record of climate change from Laguna

Miscanti (23 � S), northern Chile. Global Planet. Change, 28,35–51.

Haberle, S.G., Szeicz, J.M. and Bennett, K.D. (2000) Late

Holocene vegetation dynamics and lake geochemistry at

Laguna Miranda, XI Region, Chile. Rev. Chilena Hist. Nat.,

73, 655–669.

Harris, G.P. (1986). Phytoplankton Ecology. Chapman and

Hall, London. 384 pp.

Herrera, C., Pueyo, J.J., Saez, A. and Valero-Garces, B.L.(2006) Relacion de aguas superficiales y subterraneas en el

area del Lago Chungara y lagunas de Cotacotani, norte de

Chile: un estudio isotopico. Rev. Geol. Chile, 33, 299–325.

Kelts, K. and Hsu, K. (1978) Freshwater carbonate sedimen-

tation. In: Lakes-Chemistry, Geology, Physics (Ed. A. Ler-

man), pp. 295–323. Springer, New York.

Kott, A., Gaupp, R. and Worner, G. (1995) Miocene to Recent

history of the western Altiplano in northern Chile revealed

by lacustrine sediments of the Lauca Basin (18 degrees

15¢-18 degrees 40¢ S/69 degrees 30¢-69 degrees 05¢ W).

Geol. Rundsch., 84, 770–780.

Last, W.M. (1982) Holocene carbonate sedimentation in Lake

Manitoba, Canada. Sedimentology, 29, 691–704.

Last, W.M. and Schweyen, T.H. (1985) Late Holocene history

of Waldsea Lake, Saskatchewan, Canada. Quatern. Res., 24,219–234.

Mladinic, P., Hrepic, N. and Quintana, E. (1987) Water

physical and chemical characterization of the lakes Chun-

gara and Cotacotani. Archivos de biologia y medicina ex-perimentales, 20, 89–94.

Moore, P., Webb, J.A. and Collinson, A. (1991) An Illustrated

Guide to Pollen Analysis. Hodder and Stroughton, London,

216 pp.

Muhlhauser, H., Hrepic, N., Mladinic, P., Montecino, V. and

Cabrera, S. (1995) Water-quality and limnological features

of a high-altitude andean lake, Chungara in northern chile.

Rev. Chilena Hist. Nat., 68, 341–349.

Owen, R.B. and Crossley, R. (1992) Spatial and temporal

distribution of diatoms in sediments of Lake Malawi, Cen-

tral Africa, and ecological implications. J. Paleolimnol., 7,55–71.

Owen, R.B. and Utha-aroon, C. (1999) Diatomaceous sedi-

mentation in the Tertiary Lampang Basin, Northern Thai-

land. J. Paleolimnol., 22, 81–95.

Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W.,Bertrand, C.J.H., Blackwell, P.G., Buck, C.E., Burr, G.S.,Cutler, K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G.,Friedrich, M., Guilderson, T.P., Hogg, A.G., Hughen, K.A.,Kromer, B., McCormac, G., Manning, S., Ramsey, C.B.,Reimer, R.W., Remmele, S., Southon, J.R., Stuiver, M.,Talamo, S., Taylor, F.W., van der Plicht, J. and Wey-henmeyer, C.E. (2004) IntCal04 terrestrial radiocarbon age

calibration, 0–26 cal kyr BP. Radiocarbon, 46, 1029–1058.

Renberg, I. (1990) A procedure for preparing large sets of

diatom slides from sediment cores. J. Paleolimnol., 4, 87–90.

Saez, A. and Cabrera, L. (2002) Sedimentological and pal-

aeohydrological responses to tectonics and climate in a



small, closed, lacustrine system: Oligocene As Pontes Basin

(Spain). Sedimentology, 49, 1073–1094.

Saez, A., Cabrera, L., Jensen, A. and Chong, G. (1999) Late

Neogene lacustrine record and paleogeography in the

Quillagua-Llamara basin, Central Andean fore-arc (Northern

Chile). In: Ancient and Recent Lacustrine Systems in Con-

vergent Margins (Eds L. Cabrera and A. Saez). Palaeogeol.

Palaeoclim. Palaeoecol., 151, 5–37.

Schnellmann, M., Anselmetti, F.S., Giardini, G. and Mcken-zie, J.A. (2005) Mass movement-induced fold-and-thrust

belt structures in unconsolidated sediments in Lake Lu-

cerne (Switzerland). Sedimentology, 52, 271–289.

Shen, C.C., Edwards, R.L., Cheng, H., Dorale, J.A., Thomas,R.B., Moran, S.B. and Edmonds, H.N. (2002) Uranium and

thorium isotopic and concentration measurements by mag-

netic sector inductively coupled plasma mass spectrometry.

Chem. Geol., 185, 165–178.

Sieber, L. and Simkin, T. (2002) Volcanoes of the World: An

Illustrated Catalog of Holocene Volcanoes and their Erup-tions. Smithsonian Institution. Global Volcanism Program

Digital Information Series, GVP-3. Available at: http://

www.volcano.si.edu/world/.

de Silva, S. and Francis, P. (1991) Volcanoes of the Central

Andes. Springer-Verlag, Berlin, 216 p.

Stockmarr, J. (1971) Tablets with spores used in absolute

pollen analysis. Pollen Spores, 13, 614–621.

Telford, R.J., Barker, P., Metcalfe, S. and Newton, A. (2004)

Lacustrine responses to tephra deposition: examples from

Mexico. Quatern. Sci. Rev., 23, 2337–2353.

Utrilla, R., Vazquez, A. and Anadon, P. (1998) Paleohydrology

of the Upper Miocene Bicorb Lake (easter Spain) as inferred

from stable istopic data from inorganic carbonates. Sed.

Geol., 121, 191–206.

Valero-Garces, B.L., Delgado-Huertas, A., Navas, A. and

Ratto, N. (1999a) Large 13C enrichment in primary carbon-

ates from Andean Altiplano lakes, Northwest Argentina.

Earth Planet. Sci. Lett., 17, 253–266.

Valero-Garces, B.L., Grosjean, M., Schreier, H., Kelts, K. and

Messerli, B. (1999b) Holocene lacustrine deposition in the

Atacama Altiplano: facies models, climate and tectonic

forcing. Palaeogeogr. Palaeoclimatol. Palaeoecol., 151, 101–

125.

Valero-Garces, B.L., Grosjean, M., Messerli, B., Schwalb, A.and Kelts, K. (2000) Late Quaternary Lacustrine Deposition

in the Chilean Altiplano (18�–28�S). In: Lake Basins Trough

Space and Time (Eds E.H. Gierlowski-Kordesch and K.

Kelts), AAPG Stud. Geol., 46, 625–639.

Valero-Garces, B.L., Delgado-Huertas, A., Navas, A.,Edwards, L., Schwalb, A. and Ratto, N. (2003) Patterns of

regional hydrological variability in central-southern Altip-

lano (18 �–26 �S) lakes during the last 500 years. Palaeoge-

ogr. Palaeoclimatol. Palaeoecol., 194, 319–338.

Vila, I. and Muhlhauser, H. (1987) Dinamica de lagos de

altura, perspectivas de investigacion. Archivos de Biologıa

y Medicina Experimentales (Chile), 20, 95–103.

Villwock, W., Kies, L., Thiedig, F. and Thomann, R. (1985)

Geologish-okologishe Untersuchungen am Lago Chungara/

Nord Chile: Zielsetzungen und erste Ergebnisse. IDESIA

(Chile), 9, 21–34.

Wolin, J.A. and Duthie, H.C. (1999) Diatoms as indicators of

water level change in freshwater lakes. In: The Diatoms:

Applications for the Environmental and Earth Sciences (Eds

E.F. Stoermer and J.P. Smol), pp. 183–202. Cambridge

University Press, Cambridge.

Worner, G., Harmon, R.S., Davidson, J., Moorbath, S., Turner,D.L., McMillan, N., Nye, C., Lopez-Escobar, L. and Moreno,H. (1988) The Nevados de Payachata volcanic region (18 �S/

69 �W, N. Chile). I. Geological, geochemical, and isotopic

observations. Bull. Volcanol., 50, 287–303.

Worner, G., Hammerschmidt, K., Henjes-Huns, F., Lezaun, J.and Wilke, H. (2000) Geochronology (40Ar/39Ar, K-Ar and

He-exposure ages) of Cenozoic magmatic rocks fron Northern

Chile (18 �–22 �S): implications for magmatism and tectonic

evolution of the central Andes. Rev. Geol. Chile, 27, 205–240.

Ybert, J.P. (1992) Ancient lake environments as deduced from

pollen analysis. In: Lake Titicaca A Synthesis of Limnolog-

ical Knowledge (Eds C. Dejoux and A. Iltis), pp. 49–60.

Kluwer Academic Publishing, Dordrecht.

Yu, Z.C., Ito, E. and Engstrom, D.R. (2002) Water isotopic and

hydrochemical evolution of a lake chain in the northern

Great Plains and its paleoclimate implications. J. Paleolim-

nol., 28, 207–217.

Zheng, Z. and Lei, Z.Q. (1999) A 400 000 year record of veg-

etational and climatic changes from a volcanic basin, Leiz-

hou Peninsula, southern China. Palaeogeogr.

Palaeoclimatol. Palaeoecol., 145, 339–362.

Manuscript received 6 July 2005; revision accepted 19March 2007

1222 A. Saez et al.


Documents

Lacustrine sedimentation in active volcanic settings: the ...diapiro.ictja.csic.es/~sgiralt/articles/Saez_et_al... · and its location in an active volcanic setting, provides an opportunity