at SciVerse ScienceDirect

Quaternary Science Reviews 30 (2011) 3973e3989

Contents lists available

Quaternary Science Reviews

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

Lake highstands on the Altiplano (Tropical Andes) contemporaneous withHeinrich 1 and the Younger Dryas: new insights from 14C, UeTh dating andd18O of carbonates

P.-H. Blard a,b,*, F. Sylvestre c, A.K. Tripati b,d,e, C. Claude c, C. Causse f, A. Coudrain g, T. Condomh,J.-L. Seidel i, F. Vimeux f,i, C. Moreau j, J.-P. Dumoulin j, J. Lavé a

aCentre de Recherches Pétrographiques et Géochimiques (CRPG), UPR 2300, CNRS, Université de Lorraine, Vandoeuvre-lès-Nancy, FrancebDivision of Geological and Planetary Sciences, California Institute of Technology (Caltech), Pasadena, USAcCentre Européen de Recherche et d’Enseignement des Géosciences de l’Environnement (CEREGE), UMR 6635, CNRS, IRD, Aix-Marseille Université, Aix-en-Provence, FrancedDepartment of Earth and Space Sciences, Institute of the Environment, University of California (UCLA), Los Angeles, USAeDepartment of Atmospheric and Oceanic Sciences, Institute of the Environment, University of California (UCLA), Los Angeles, USAf Institut Pierre Simon Laplace (IPSL), Laboratoire des Sciences de Climat et de l’Environnement (LSCE), UMR 8212 (CEA-CNRS-UVSQ), Gif-sur-Yvette, Franceg Institut de Rercherche pour le Développement (IRD), Marseille, Franceh Laboratoire d’étude des Transferts en Hydrologie et Environnement (LTHE), UMR 5564, Université Joseph Fourier e Grenoble 1, IRD, CNRS, G-INP, Grenoble, Francei Institut de Recherche pour le Développement (IRD), Laboratoire HydroSciences Montpellier (HSM), UMR 5569 (CNRS-IRD-UM1-UM2), Université de Montpellier, Montpellier, Francej Laboratoire de Mesure du Carbone 14 (LMC14), UMS 2572 (CEA, CNRS, IRD, IRSN et Ministère de la Culture et de la Communication), Saclay, Gif-sur-Yvette, France

a r t i c l e i n f o

Article history:Received 22 July 2011Received in revised form29 October 2011Accepted 2 November 2011

Keywords:TropicsAltiplanoLakeTaucaCoipasaHeinrich 1Younger DryasUeThU-series dating14Cd13Cd18O

* Corresponding author. Centre de Recherches Pétro(CRPG), UPR 2300, CNRS, Université de Lorraine, Van

E-mail address: blard@crpg.cnrs-nancy.fr (P.-H. Bla

0277-3791/$ e see front matter � 2011 Elsevier Ltd.doi:10.1016/j.quascirev.2011.11.001

a b s t r a c t

This study provides new geochronological and stable isotope constraints on Late Pleistocene fluctuationsin lake level that occurred in the closed-watershed of the Central Altiplano between w25 and w12 ka.U-series isochrons and 14C ages from carbonates are used to confirm and refine the previous chronologypublished (Placzek et al., 2006b). Our new data support three successive lake highstands during the LatePleistocene: (i) the Lake Sajsi cycle, from w25 to 19 ka, that culminated at 3670 m at about 22 ka, almostsynchronously with the global last glacial maximum, (ii) the Lake Tauca cycle, that lasted from 18 to14.5 ka and was characterized by the highest water level, reached at least 3770 m from 16.5 to 15 ka, (iii)the Lake Coipasa cycle, from 12.5 to 11.9 ka, that reached an elevation of w3700 m, 42 m above theelevation of the Salar de Uyuni (3658 m). These high amplitude lake level fluctuations are in phase withthe coldewarm oscillations that occurred in the North Atlantic and Greenland during the Late Pleisto-cene (Heinrich 1, BøllingeAllerød, Younger Dryas). Such temporal coincidence supports the hypothesisthat wet events recorded in the Central Altiplano are controlled by the northesouth displacement of theInter Tropical Convergence Zone resulting from changes in the meridional temperature gradient. Finally,the oxygen isotope ratios measured in these lacustrine carbonates allows for calculation of the d18O valueof paleolake waters. Estimates of water d18O (V-SMOW) are �2.8 � 0.7& for Lake Tauca and �1.6 � 0.9&for Lake Coipasa. These data are used to constrain changes in lake hydrology and can be interpreted toindicate that the proportion of precipitation arising from local water recycling was less than 50%.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Many recent observations have led to a reconsideration of therole of the tropics in paleoclimate. The tropics are now recognizedto be the driver of high frequency climatic variability, including ElNiño-Southern Oscillation (ENSO) (Chiang, 2009). It has also been

graphiques et Géochimiquesdoeuvre-lès-Nancy, France.rd).

All rights reserved.

proposed that the tropics may be an amplifier of the abruptmillennial changes recorded in the Greenland ice and in the sedi-ments of the Northern Atlantic (e.g. Leduc et al., 2007). In particular,several paleoprecipitation records (Cruz et al., 2005) have led to thesuggestion that the northesouth oscillation of the tropical rainfallbelt might be a key mechanism for a tropical “butterfly effect” (e.g.Peterson et al., 2000; Leduc et al., 2007). According to thishypothesis, millennial scale fluctuations in moisture transportacross the Isthmus of Panama may modulate the North Atlanticfreshwater budget and therefore serve as a positive feedback intoabrupt climate changes. Accurate andwell-dated records of tropical

P.-H. Blard et al. / Quaternary Science Reviews 30 (2011) 3973e39893974

precipitation are thus of critical importance to improve ourunderstanding of the role the tropical hydrologic cycle may have asa driver or amplifier of climate change. Records of past lake levelsreflect temporal and spatial changes in precipitation patterns.Paleoshorelines provide direct benchmarks of water depth and canthus be used as quantitative records of regional hydrologicalbudget.

In the closed TiticacaeAltiplano watershed (Central Andes),numerous very well-preserved paleoshorelines and carbonatedeposits represent direct and spectacular archives of significantchanges in net moisture over the southern tropical Andes. Severalstudies conducted over the two last decades have allowed chro-nological constraints to be placed on the timing and causes of theselarge fluctuations in lake level and demonstrated their abruptness(e.g. Sylvestre et al., 1999; Baker et al., 2001; Placzek et al., 2006a,2006b). In particular, comprehensive studies based on a largenumber of UeTh and radiocarbon dates from paleolake carbonatespermitted identification of two large oscillations in lake levelsynchronous with the abrupt millennial cooling events recorded inthe North Atlantic, namely Lake Tauca (coincident with the Hein-rich 1 event, 17e15 ka) and Lake Coipasa (coincident with theYounger Dryas, 13e12 ka) (Sylvestre et al., 1999; Placzek et al.,2006a, 2006b). Sr and U isotopes have also recently been used tocharacterize the spatial pattern of these pluvial events (Placzeket al., 2011). However, several questions remain unanswered:

i) What is the timing of other oscillations in lake level? Themajority of existing UeTh and 14C data belongs to the LakeTauca cycle, but the chronologies of the other oscillations inlake level are not as well established. This is notably the casefor the lowstand episode between the so-called Sajsi episode(25e18 ka) and the Lake Tauca cycle (17.5e15 ka), as both theamplitude and duration of the SajsieTauca lowstand is stillunclear (Placzek et al., 2006b). Similarly, the chronology andamplitude of the Coipasa cycle needs to be bettercharacterized.

ii) What triggered the lake highstands in this region, and how dooscillations in lake level reflect changes in atmosphericcirculation (Sylvestre et al., 1999; Placzek et al., 2006b)?Various isotopic systems have been used to propose severalmodels for precipitation patterns over the Altiplano(Coudrain et al., 2002; Placzek et al., 2011), but many ques-tions remain open about these highstands, particularly overthe amount of local water that was recycled, thus impactingthe hydrological budget of these lakes.

In this contribution we present a new set of 14C and UeTh agesof the Altiplano lake cycles that occurred during the late Pleistocene(Sajsi, Tauca and Coipasa cycles). We combine these new ages withpublished datasets (Sylvestre et al., 1999; Placzek et al., 2006b).These new data are crucial to refine the chronology of theseepisodes and to establish coherent regional scenarios for netprecipitation changes. We also provide new stable isotope data(d18O and d13C) from well-dated lake carbonates. These data allowcalculation of the d18O of past lake waters and are used to constrainthe contribution of local recycling to the hydrologic budget of thelake.

2. Regional settings

2.1. The Altiplano endorheic watershed

The TiticacaeAltiplano watershed (15e23�S, 66e70�W) is thelargest endorheic basin of the Tropical Andes, with a total area ofw196,000 km2 (Fig. 1). This large plateau, located at an average

elevation of w3800 m, is flanked by two northesouth mountainranges rising up to 6500 m: the Oriental and the OccidentalCordilleras. The Altiplano hydrological system may be divided intwo main sub-catchments: the Titicaca basin (58,000 km2) in thenorth and the UyunieCoipasa basin (138,000 km2) in the south,with a lowest point situated at 3656 m. Since the Lake Titicaca(3810 m) is overfilled with a spillway in its south-western end,these two sub-basins are hydrologically connected through theTiticaca outlet, the Rio Desaguadero (Fig. 1). Current annual rainfallis controlled by the southward deflection of the Inter TropicalConvergence Zone (ITCZ) during the summer season (Garreaudet al., 2009). This atmospheric pattern implies a strong season-ality of precipitation: 80% of rainfall indeed occurs during theAustral summer, between October and April. Moreover, as themajority of precipitation originates from the northeast, theorographic effect is responsible for a strong northeast to southwestgradient: precipitation is 800 mm yr�1 in the northern Altiplano,over the lake Titicaca, and decrease below 100 mm yr�1 in thesouthern part, the Lipez region. Over the Altiplano, the meanannual potential evaporation is larger than 1000mmyr�1 (Delclauxet al., 2007). Given that the present rainfall in the southern basin isless than 330 mm yr�1 (Condom et al., 2004), such intense evap-oration implies that the hydrological balance of the southernUyunieCoipasa basin is in deficit, resulting in the presence of largedry salty basins: Salar de Uyuni and Salar de Coipasa.

In the southern sub-watershed (UyunieCoipasa basin), well-preserved paleoshorelines are present in many locations (Fig. 1)and meter-scale bioherms are often observed just below thesepaleoshorelines. Those features have been extensively described inmany previous studies and were interpreted as relicts of giantpaleolakes (Servant and Fontes, 1978; Rouchy et al., 1996; Sylvestreet al., 1999; Placzek et al., 2006a, 2006b) (Fig. 1). Detailed strati-graphic analyses, 14C and UeTh dating of these deposits establishedthat the highstand of the largest Pleistocene lacustrine episode, theLake Tauca, occupied a total surface of 52,000 km2 (water level at3775 m) during about 1000 years between w16 ka and 15 ka BP(Fig. 1) (Sylvestre et al., 1999; Placzek et al., 2006b). Severalmodeling studies have established that the average precipitationincrease during this Lake Tauca episode could have reached 1.5e1.8times the present average value (Hastenrath and Kutzbach, 1985;Blodgett et al., 1997; Coudrain et al., 2002; Blard et al., 2009).

2.2. Air and lake temperature data

A compilation of the rare available air temperature records forthe Altiplano is shown in Table 1. Fig. 1 shows the location of theseweather stations. Once corrected for elevation by using an annuallapse rate of 6.5 �C km�1 (Klein et al., 1999), the annual temperaturemeans of the 7 weather stations are characterized by a regionalaverage of 9 �C after normalization at 3770 m. Temperatures werescaled at this elevation because it corresponds to the altitude ofthe Lake Tauca highstand (Sylvestre et al., 1999; Placzek et al.,2006b). The low variability between these different weatherstations (standard deviation of the annual means is only 1.2 �C)and the absence of a significant correlation between the spatialposition and the temperature suggests that local climatic particu-larities are not significant. However, the seasonal amplitude(amplitude ¼ TDJF � TJJA) is characterized by a significant variabilitybetween each location (Table 1). Although this seasonal amplitudeseems to be anti-correlated with elevation (R2 ¼ 0.7), it is not clearwhether this variability is not controlled by another parameter,such as the temporal and spatial fluctuations of cloudiness.

Since one of the goals of the present study is to provide recon-struction of paleolake d18O, it is important to rely on accuratereconstructions of paleolake temperature. For this, it is crucial to

Fig. 1. A) Major features of the present-day atmospheric circulation over South America. B) Hydrographic map of the Altiplano is overlain on the Landsat picture. Limit of theendorheic Tauca watershed is the orange dashed line. Contour lines for present-day yearly precipitation are in white (New et al., 2002). Location of paleolake and modern samples inred (1e7). 1 - Urmiri: T7, T8; 2 - Tauca: Hua-44, Bo-93-26, Bo-93-23H, 23M; 3 - Jahuila: Jah-12a, 12b, 14a, 14b, PK; 4 - Nueva Esperanza: Piv-Ga, Piv-Ostra; 5 - Julaca: Jula-23; 6 - LakePoopo: Poopo-Gas; 7 - Lake Uru-Uru: Uru-Uru-Gas. Location of weather stations in yellow (AeG). A: Puno, B: El Alto - La Paz, C: Patacamaya, D: Sajama, E: Oruro, F: Potosi, G: Uyuni.(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

P.-H. Blard et al. / Quaternary Science Reviews 30 (2011) 3973e3989 3975

constrain the relationship between the temperature of surficial lakewater and the local nearby air temperature. Available data fortropical lakes are scarce compared to temperate regions (Wilhelmet al., 2006; Huntington et al., 2010). The modern Lake Titicaca isthe closest regional analog of the paleolakes Sajsi, Tauca and Coi-pasa. Several temperatures surveys of the Lake Titicaca surficialwaters have established a mean annual surface temperature of13.6 �C (Table 1) (Richerson et al., 1977; Carmouze, 1991; Blodgettet al., 1997). There is a w5 �C difference between the Lake Titi-caca temperature and the 8.5 �C mean annual air temperaturerecorded at Puno, at the same elevation as the Titicaca surface(3814 m). Although we cannot rule out the possibility that lakealbedo and local cloudiness have a major effect, another hypothesisto explain this temperature difference is a bottom-up heating of thelake resulting from the geothermal flux (Delclaux et al., 2007).Geothermal heating may indeed exist over the whole Altiplano,from the Lake Titicaca to the southern Uyuni basin, where few hotsprings are present. In the following, wewill thus consider this 5 �C

gap to calculate the paleolake temperature from the reconstructionof past glacial equilibrium line altitudes synchronous with theTauca and the Coipasa events (Blard et al., 2009).

3. Samples

Because the two main goals of this study are (i) to refine thewater level chronology and (ii) to provide new constraints on thed18O of the paleolake water, the new samples studied here wereselected according to specific criteria. A detailed list of the pro-cessed samples along with the type of analyses (UeTh, 14C, d13C,d18O) is provided in Table 2.

Carbonate samples deposited in shallow waters (<10 m waterdepth) are not only the best indicators of the lake surface paleo-elevation, and hence of the total water depth, but they also repre-sent the water zones (epilimnion) that were tightly linked with theair paleotemperature (Richerson et al., 1977; Huntington et al.,2010). Consequently, the majority of the samples targeted in this

Table

1Presen

tmea

n-m

onthly

temperaturesreco

rded

ontheAltiplano.

Location

Period

ofob

servation

Latitude

(�S)

Longitude

(�W

)Altitude

(m)

January

February

March

April

May

June

July

Augu

stSe

ptembe

rOctob

erNov

embe

rDecem

ber

Mea

nan

nual

temperature

1sTb

at37

70m

Tcat

3700

mSe

ason

alam

plitude

(�C)

Lake

Titicaca

epilimniona

1973

1669

3814

1516

1515

1412

1212

1213

14.3

1413

.61.3

Puno

1964

e19

9216

7038

149.9

9.8

9.6

8.9

7.4

6.1

5.8

6.9

8.2

9.4

10.0

10.1

8.5

1.6

8.8

9.3

3.7

ElAl

to-La

Paz

1943

e19

9017

6840

388.3

8.5

8.5

8.3

7.3

6.1

5.9

7.0

7.6

9.0

9.6

9.0

7.9

1.2

9.7

10.1

2.3

Patacamay

a19

58e19

9017

6837

8912

.011

.811

.49.9

7.0

5.0

4.8

6.5

8.8

10.8

12.1

12.3

9.4

2.9

9.5

9.9

6.6

Uyu

ni

1944

e19

9721

6736

6913

.212

.712

.18.9

4.5

1.7

1.9

4.4

6.3

9.0

11.0

12.8

8.2

4.3

7.5

8.0

10.3

Oruro

1951

e19

8818

6737

1012

.812

.412

.611

.38.4

6.1

6.0

8.1

10.6

12.8

13.8

13.4

10.7

2.8

10.3

10.8

6.1

Potosi

1950

e19

7919

6539

0011

1011

108

66

78

1011

119.0

2.0

9.8

10.3

4.4

Sajama

1975

e19

8118

6942

205.3

5.3

4.9

3.7

2.5

1.9

2.7

3.0

4.6

5.4

6.8

6.0

4.4

1.5

7.3

7.7

3.0

Mea

nT

(�C)

9.0

1 s1.2

aTe

mperature

oftheupper

0e10

mlake

surface(Richersonet

al.,19

77).

bAirtemperature

scaled

to37

70m

(Lak

eTa

uca)usingalapse

rate

of6.5

� Ckm

�1.

cAirtemperature

scaled

to37

00m

(Lak

eCoipasa)

usingalapse

rate

of6.5

� Ckm

�1.

P.-H. Blard et al. / Quaternary Science Reviews 30 (2011) 3973e39893976

study were either bioherms or other biogenetic carbonates thatwere formed in associationwith constructional flat benches or withshoreline notches. Shoreline algal bioherms (tufa microbialites) arecharacterized by a tabular morphology indicating that they grew ina shallow water environments (Rouchy et al., 1996), close to thepaleolake shoreline. While these bioherms were initially describedas stromatolites (Rondeau, 1990), detailed petrographical andmineralogical study of the microstructures revealed that these lakedeposits should rather be defined as “algal bioherms” or “tufa” inthe case of the larger constructions, and “algal crusts” in the case ofthinner and plannar encrustations (Rouchy et al., 1996). Several ofthe samples studied here are tufas having these characteristicsindicative of shallow water formation: Jula-23, T7, T8, Hua-44, Bo-93-26 (Table 2, Fig. 1). We also analyzed a horizontal transect of 1 mwithin a large bioherm located in the Quebrada Negrojahuila(67.542�O, 19.570�S) (Rouchy et al., 1996; Sylvestre et al., 1999), at3660 m (Sample Jah-12a-b and Jah-14a-b) (Table 2, Fig. 2). Thisbioherm may have potentially recorded several lake oscillationsduring the Late Pleistocene, including the transgression and thesubsequent regression of Lake Tauca. Although the formation depthof large rounded tufa can be uncertain (Laval et al., 2000), severalbiological textures observed within this bioherm have been inter-preted as indicators of a carbonate growth within the photic zone(Rouchy et al., 1996), i.e. that they probably grew in a water depth<20 m, which is the limit depth of the light penetration in tropicallakes (Tapia et al., 2003). We also collected and processed shorelinesediments at an elevation (3765 m) very close to the altitude of thehighest Tauca shoreline w3770 m (67.0644�O, 20.0495�S) (Piv-Ga,Piv-Ostra, Table 2). Gastropods and ostracods were abundant in the0.1e1 mm fraction of these carbonate-rich lake sediments. Asufficient amount of material was retrieved for 14C dating and d18Omeasurement (w5 mg).

As the quality of the chronology is key, we targeted samples thatwere already described and successfully dated by UeTh or 14C(Sylvestre et al., 1999) (Table 2). However, it is important to notethat all the samples analyzed for d18O were also re-dated by 14C.This was done as a check for closed-system conditions and toevaluate if there was any recent carbonate precipitation. Addi-tionally, we also provide in the present article new UeTh ages (Jah-12a-b) along with a reevaluation of some the UeTh ages previouslypublished for several samples (Bo-93-23H, 23M, Hua-44, PK)(Sylvestre, 1997; Sylvestre et al., 1999). Finally, the UeTh isotopicanalyses of sub-samples allowed construction of three new UeThisochrons for three different tufas samples (T7, T8, Jula-23).

All samples are quite largely distributed and thus cover an arearepresentative of the entire Lake Tauca (Fig. 1).

4. Methods

Tufa sub-samples of few mg were collected by micro-drilling.Microscope observations indicate that some of these algalcarbonates are finely laminated (Fig. 2C), a structure that canpossibly be interpreted to result from seasonality. However, theperformed drilling was large enough (holes between 1 and 2 mmdiameter) to incorporate several layers and to minimize a seasonalsampling bias. Some tufas were sampled at several spots separatedby several centimeters as a test for spatial variability (Fig. 2;Table 2). For each of these sub-samples, d18O, d13C and 14C wereanalyzed on the same aliquots. The same procedure was used toestablish the UeTh isochrons for samples T7, T8 and Jula-23.

Lake sediment samples (Piv-Ga, Piv-Ostra, Uru-Uru-Gas, Poopo-Gas) were sieved and the 0.1e1 mm fraction was processed toisolate individual species. Pure separates of ostracods and gastro-pods were obtained by a careful handpicking under a binocularmicroscope.

Table 2List and location of processed samples.

Sample name Location Longitude Latitude Altitude (m) Material type Type of analysis Reference

Tauca samplesTufa from Huasa JulacaJula-23 Huasa Julaca 67.50991�O 20.88831�S 3665 Tufa UeTh

Isochron - 3 replicatesThis study

Bioherm from Quebrada Negrojahuila - 3670 mJah-12a Quebrada Negrojahuila 67.542�O 19.570�S 3670 Internal part of dense

bioherm - Branching tufaUeTha This study

Jah-12b Quebrada Negrojahuila 67.542�O 19.570�S 3670 Internal part of densebioherm - Columnar microbialite

UeTha This study

Jah-12b Quebrada Negrojahuila 67.542�O 19.570�S 3670 Internal part of densebioherm - Columnar microbialite

14C - 3 replicates This study

Jah-12b Quebrada Negrojahuila 67.542�O 19.570�S 3670 Internal part of densebioherm - Columnar microbialite

d18O d13C - 2 replicates This study

Jah-14a Quebrada Negrojahuila 67.542�O 19.570�S 3670 External cortex - Laminated bioherm 14C and UeTha This studyJah-14a Quebrada Negrojahuila 67.542�O 19.570�S 3670 External cortex - Laminated bioherm d18O d13C - 4 replicates This studyPK Quebrada Negrojahuila 67.542�O 19.570�S 3657 Aragonite crust 14C and UeTha Sylvestre

et al., 1999Tufa from Urmiri village - 3775 mT7 Urmiri 66.8710�O 18.5734�S 3775 Tufa UeTh

Isochron - 5 replicatesThis study

T8 Urmiri 66.8710�O 18.5734�S 3775 Tufa UeThIsochron - 3 replicates

Shoreline lacustrine deposits - 3765 mPiv-Ga Nueva Esperanza 67.0644 O 20.0495 S 3765 Gastropods 14C and d18O d13C This studyPiv-Ostra Nueva Esperanza 67.0644 O 20.0495 S 3765 Ostracods d18O d13C This studyRadial tufa from Tauca-Huacuyo - 3770 mHua-44 Tauca-Huacuyo 68.006�O 19.558�S 3770 Radial fibrous porous tufa 14C and UeTha Sylvestre

et al., 1999Hua-44 Tauca-Huacuyo 68.006�O 19.558�S 3770 Radial fibrous porous tufa 14C This studyHua-44 Tauca-Huacuyo 68.006�O 19.558�S 3770 Radial fibrous porous tufa d18O d13C - 2 replicates This studyCoipasa samplesBioherm from Quebrada Negrojahuila - 3670 mJah-14b Quebrada Negrojahuila 67.542�O 19.570�S 3670 External cortex of dense

bioherm - Tufa layers

14C This study

Jah-14b Quebrada Negrojahuila 67.542�O 19.570�S 3670 External cortex of densebioherm - Tufa layers

d18O d13C - 3 replicates This study

Stromatolites and crust from Tauca village - 3680 mBo-93-26 Tauca village 68.0345�O 19.5519�S 3680 Bioherm 14C - 3 replicates This studyBo-93-26 Tauca village 68.0345�O 19.5519�S 3680 Bioherm d18O d13C - 4 replicates This studyBo-93-23H Tauca village 68.0345�O 19.5519�S 3680 Calcareous crust 14C and UeTha Sylvestre

et al., 1999Bo-93-23M Tauca village 68.0345�O 19.5519�S 3680 Calcareous crust 14C and UeTha Sylvestre

et al., 1999Modern samplesUru-Uru-Gas Lake Poopo 66.85222�O 18.99988�S 3708 Gastropods d18O d13C This studyPoopo-Gas Lake Uru-Uru 67.07650�O 18.15596�S 3720 Gastropods d18O d13C This study

a Bulk sample analysis.

P.-H. Blard et al. / Quaternary Science Reviews 30 (2011) 3973e3989 3977

4.1. 14C dating

Carbonate samples have been cleaned in 2 mL of HNO3 0.01 Mduring 15 min and 10 s in an ultrasonic bath. The typical size oftreated samples was sufficient to yield about 1 mg of carbon. Thecleaned samples were then completely dissolved by acid digestionin dehydrated H3PO4 at 60 �C. The obtained CO2 was then collectedand reduced in pure C by reacting with H2 in the presence of ironpowder at 600 �C. The pure C target was then loaded in a cathodeand the three carbon isotopes 12C, 13C and 14C were analyzed byAccelerator Mass Spectrometry on the ARTEMIS facility (Saclay,France). Radiocarbon ages are calculated after correction for theinstrumental and natural mass fractionation by measuring the d13Cof each sample (Mook and van der Plicht, 1999).

4.2. UeTh dating

The UeTh single ages reported here for samples Jah-12a, Jah-12b and Jah-14a were determined by thermal-ionization massspectrometry (TIMS) at Geotop (Université du Québec, Montréal,

Canada) following the same method than the one used in Sylvestreet al. (1999). Those data had only been published in a PhD thesis(Sylvestre, 1997). In the present study we reexamine their signifi-cance by taking into account new constraints on the initial230Th/232Th activity ratio.

The new UeTh isochron ages reported for samples T7, T8 andJula-23 have been obtained at CEREGE (Aix-en-Provence, France)using the total dissolution isochron technique (TDS). This techniqueis the most appropriate U-series method for dating the authigeniccarbonate fractions because it allows us to correct for the impurecarbonates from the detrital contamination (Bischoff andFitzpatrick, 1991; Ludwig and Titterington, 1994). Samples areassumed to contain a coeval authigenic carbonate fraction (Bischoffand Fitzpatrick, 1991). Tufa samples were manually crushed intoa mortar and carefully handpicked under a binocular. Differentfractions were selected to have contrasted proportion of detritalmaterial respectively to the authigenic carbonate fraction. U and Thseparation and purification were obtained by conventional anionexchange chromatography following procedures developed atCEREGE (Delanghe et al., 2002). Each fraction (0.5e3 g) was spiked

Fig. 3. A and B) Lake Tauca sandbeach deposits near Nueva Esperanza (20.0495�S, 67.0644�O, 3765 m). C) Ostracods (Piv-Ostra). D) Gastropods (Piv-Ga).

Fig. 2. UeTh and 14C results from a transect in a giant bioherm located at 3660 m (Quebrada NegroJahuila, 19.570�S, 67.542�O). Samples Jah-12a, 12b, 14a, 14b are part of thisbioherm transect.

P.-H. Blard et al. / Quaternary Science Reviews 30 (2011) 3973e39893978

P.-H. Blard et al. / Quaternary Science Reviews 30 (2011) 3973e3989 3979

with a mixed 233Ue229Th tracer before dissolution. Total dissolu-tion was performed using a two-step procedure with HNO3eHFacids and HNO3eHClO4 acids. After redissolution of each fractioninto HNO3 7N, U and Th were co-precipitated using a Fe-carrier. Uand Th measurements were performed on a Thermal IonizationMass Spectrometer (TIMS) VG-54-30 fitted with an ion-countingDaly detector at CEREGE using a triple Ta-Re-filament procedureaccording to the technique of Delanghe et al. (2002). U concen-trations and 234U/238U isotope ratios were analyzed fromw200e500 ng U. 234U/238U activity ratios and U concentrationswere calculated assuming a constant 238U/235U ratio of 137.88. Theexternal reproducibility of the isotopic measurements wascontrolled by analyzing the SRM960-U standard: this materialprovided an atomic 234U/238U ratio of 52.74 � 0.2 � 10�6. Thismeasured ratio is fully compatible with the NBS certified value of52.9� 0.2�10�6 andwith the highly precise and accurate values of52.86 � 0.1 � 10�6 determined by Cheng et al. (2000). Errors aregiven at the 95% confidence level.

Thorium concentrations and isotopic ratios were analyzed oncirca 1 mg Th. 230Th and 229Th were measured on the ion counterand 232Th on Faraday cups using the static collection mode.This configuration with a magnetic sector followed by an electro-static filter enabled an abundance sensitivity of about 0.16 ppmunder optimum conditions (measured at mass 237 for naturaluranium). Accuracy of Th isotopic ratio measurements was checkedby analyzing the in-house standard solution Th105(232Th/230Th¼ 215,092� 3000) (Innocent et al., 2004) and the CRMIRMM 035 standard (232Th/230Th ¼ 87,853 � 1300). Our measure-ments are fully consistentwith the certified values. The total processblanks for UeTh isotopic and elemental analysis are 60 pg for U and4 ppt for 232Th. There are thus negligible compared to the amount ofU and Th analyzed in the studied samples. Reported uncertaintiesinclude both the internal and external measurement errors.

Table 3Radiocarbon data.

Reference Laboratorynumber

Samplename

Sample type Elevation(m)

Tauca lakeSylvestre

et al., 1999Beta 73088 PK Aragonite crust 3657

This study SacA 11985 Jah-12b Columnarmicrobialite

3670

This study SacA 11986 Jah-12b Columnarmicrobialite

3670

This study SacA 11987 Jah-12b Columnarmicrobialite

3670

This study SacA 14866 Jah-14a Laminatedbioherm

3670

This study SacA 11983 Piv-Ga Gastropods 3765Sylvestre

et al., 1999e Hua-44 Radial tufa 3770

This study SacA 11988 Hua-44 Radial tufa 3770

Coipasa lakeSylvestre

et al., 1999e Bo-93-23H Calcareous crust 3680

Sylvestreet al., 1999

e Bo-93-23M Calcareous crust 3680

This study SacA 14867 Jah-14b Tufa layers 3670This study SacA 14863 Bo-93-26 Bioherm 3680This study SacA 14864 Bo-93-26 Bioherm 3680This study SacA 14865 Bo-93-26 Bioherm 3680

a Calibrated using IntCal09 (Reimer et al., 2009).b Calibrated using IntCal04 (Reimer et al., 2004).

4.3. d18O and d13C data

d18O and d13C dataweremeasured by dual-inlet gas sourcemassspectrometry at the California Institute of Technology. Powderedcarbonate samples of less than 10 mg were dissolved by anhydrousphosphoric acid digestion at 25 �C for 12 h. Resulting CO2 wasisolated and purified using conventional cryogenic traps and a gaschromatograph according to the procedures described elsewhere(Ghosh et al., 2006; Huntington et al., 2009, 2010; Tripati et al.,2010). These oxygen and carbon isotope data were acquired asa part of an ongoing project dedicated to obtain clumped-isotopedata (Blard et al., 2008), and the isotopic analysis of CO2 was per-formed by measuring all isotopologues from masses 44 to 49, ona Finnigan MAT 253 mass spectrometer. Precision for d18O and d13Cwas better than 0.01e0.02&.

5. Lake chronology

5.1. 14C results

For clarity, all new 14C data (Table 3) and previously published14C ages (Sylvestre et al., 1999; Placzek et al., 2006b) were recali-brated using the most recent calibration curve IntCal09 (Reimeret al., 2009). The IntCal04 calibration (Reimer et al., 2004) is alsoshown for comparison (Table 3). However, all the 14C ages given inthe text are reported as calibrated ages (cal BP) using IntCal09. Wedid not apply any reservoir age correction. Given the good agree-ment between the 14C and the UeTh ages from the same samples(Placzek et al., 2006b), there is no reason to consider a significantamount of old carbon in the lake paleowater having been present(see Section 5.3 below for further discussion).

A preliminary important observation is the clustering of the new14C ages replicates (Table 3): the three replicates of samples Jah-12b

Carbonmass (mg)

14C age (yr) � 1s Calibrated age(ka cal BP)a

� 2s IntCal09

Calibrated age(ka cal BP)b

� 2s IntCal04

e 15,430 � 80 18.7 � 0.2 18.8 � 0.1

0.62 13,320 � 45 16.4 � 0.5 15.8 � 0.4

0.71 12,970 � 50 15.7 � 0.6 15.3 � 0.3

0.69 13,300 � 50 16.4 � 0.5 15.8 � 0.4

Average of Jah-12b 16.2 � 0.5 15.6 � 0.30.49 12,440 � 60 14.6 � 0.4 14.5 � 0.4

1.40 13,080 � 50 15.9 � 0.6 15.5 � 0.3e 12,930 � 50 15.4 � 0.6 15.3 � 0.3

0.21 12,530 � 40 14.8 � 0.4 14.7 � 0.4Average of Hua-44 15.1 � 0.5 15.0 � 0.3

e 11,370 � 50 13.2 � 0.1 13.2 � 0.1

e 11,340 � 50 13.2 � 0.1 13.2 � 0.1

0.40 10,185 � 50 11.9 � 0.2 11.9 � 0.20.60 10,815 � 45 12.7 � 0.1 12.8 � 0.050.85 11,030 � 50 12.9 � 0.2 13.0 � 0.10.90 11,180 � 50 13.1 � 0.2 13.1 � 0.1

Average of Bo-93-26 12.9 � 0.2 13.0 � 0.1

P.-H. Blard et al. / Quaternary Science Reviews 30 (2011) 3973e39893980

and Bo-93-26 agree within uncertainties. This result suggests thatsamples are homogeneous and that the sampling and cleaningprocedures prevented any contamination by recent carbonateprecipitation.

The new 14C ages may be grouped as follows:

5.1.1. Lake Tauca cycleJah-12b (16.2 � 0.5 ka), Piv-Ga (15.9 � 0.6 ka), Hua-44

(15.1 � 0.5 ka) and Jah-14a (14.6 � 0.4 ka) have 14C ages rangingbetween 16.1 and 14.6 ka, in good agreement with the already pub-lished calibrated 14C ages attributed to the Lake Tauca cycle bySylvestreet al. (1999)andPlaczeketal. (2006b). It is interesting tonotethe close agreement of 14C ages of the samples located above 3760m.Indeed, differentmaterials belonging to close stratigraphic units yieldsimilar 14C ages. This is the case of shoreline gastropods (Piv-Ga:15.9� 0.6 ka), and tufas (Hua-44: 15.1�0.5 ka) located above 3760m,close to the maximum elevation reported for the Tauca paleoshore-lines (Sylvestre et al., 1999; Placzek et al., 2006b) (Tables 2 and 3).

5.1.2. Lake Coipasa cycleThe second group of 14C ages includes samples Bo-93-23H and

M (13.2 � 0.1 ka), Jah-14b (11.9 � 0.2 ka), and Bo-93-26(12.9 � 0.2 ka). These new ages are identical to the 14C dates re-ported for the Lake Coipasa by Sylvestre et al. (1999) and Placzeket al. (2006b). All these younger 14C ages are located at lowerelevation (<3700 m) than those of the lake Tauca cycle, confirmingthat this subsequent lacustrine event was not as deep as the Taucaevent (Sylvestre et al., 1999; Placzek et al., 2006b).

5.2. UeTh results

All UeTh isotopic data are available in Table 4.

5.2.1. Isochrons from T7 sample (Tufa from Urmiri village - 3775 m)Concentrations of 238U, 232Th and 230Th isotopes have been

determined in five replicates of T7 sample, while 234U has beenanalyzed in four replicates only (Table 4). T7 sample is character-ized by a remarkable homogeneous and compact texture. Thesedata have been used to plot both “Rosholt-type” (232Th-normali-zation) (Rosholt, 1976) and “Osmond-type” (238U-normalization)(Osmond et al., 1970) isochrons using Isoplot 3 (Ludwig, 2003)(Fig. 4). Both plot types yielded 4 best-fit isochrons with particu-larly lowMSWD factors (<0.35) and high probability of fit (p> 0.7),suggesting that the obtained isochrons are statistically robust,indicating that the measured replicates satisfy the two funda-mental assumptions for accurate UeTh isochrons: i) a simple twoendmembers mixing between a homogeneous contaminant andthe lacustrine carbonate and ii) the permanence of closed-systemconditions since the carbonate formation. The two isochrons yield234U/238U activity ratios undistinguishable within uncertainties, foran average value of 1.59 � 0.02 (2s). Such 234U/238U activity ratiohigher than 1 are not rare in continental groundwater, a reservoir inwhich enrichment in 234U may result from radioactive recoil fromvery fine - high surface area silico-clastic material (Kronfled, 1974;Henderson et al., 2001).

Importantly, the isochrons built for sample T7 allow a quite firmdetermination of the initial activity ratio 230Th/232Th. Indeed, both“Osmond” and “Rosholt-type” yield similar values within uncer-tainties, i.e. a mean initial activity ratio of 0.73 � 0.06 (2s), for T7tufa. This determination is characterized by a very low uncertainty,and the obtained value falls within the range of the initial activityratio 230Th/232Th determined by Placzek et al. (2006a) (Table 4).Moreover, 0.73 � 0.06 is not very different from the 0.84 valuereported for the mean continental crust (Taylor and McLennan,1995). Finally, the so-obtained isochron age for this 3775 m high

tufa T7 is 15.5 � 0.2 ka (2s), in excellent agreement with the age ofthe highest biogenic carbonates (>3770 m) reported in Placzeket al. (2006b).

5.2.2. Isochrons from T8 sample (Tufa from Urmiri village - 3775 m)Isochrons from T8 samples has been built from three sub-

samples (Fig. 5). Although the MSWD factor is close to meaning-less when n ¼ 3, the regressions statistics (MSWD>9) indicate thatthe best-fit is an “error-chron”. This suggests that these sub-samples could not fulfill the criteria of a well-defined isochron.This sample is characterized by a very porous texture underlined byan alternation of yellowish and whitish laminates. These lamina-tions probably incorporate variable amounts of detrital componentthat probably have heterogeneous 230Th/232Th ratios. The porositycould in turn account for U mobility yielding a fractionation of the234U/238U ratio. Such characteristics might be responsible of theobserved scatter and the large uncertainty on the calculatedisochron age: 15.4 � 6.3 ka. Nevertheless, this age matches verywell with the isochron age of T7, a tufa that was sampled on thesame outcrop as T8, at a similar elevation. However, given the poorquality of T8 and the large uncertainty attached to its age, thisisochron age will not be considered further. The robust age of T7(15.5 � 0.2 ka) will be considered instead.

5.2.3. Isochrons from Jula-23 sample (Tufa from Huasa Julaca -3665 m)

Isochrons of Jula-23 are built from three sub-samples (Fig. 6).Although the MSWD factor is close to meaningless when n ¼ 3,these sub-samples define best-fit isochrons, both in the “Rosholt”and in the “Osmond” space: MSWD factors range between 0.11 and0.30. These Jula-23 isochrons yield an age of 19.0� 0.2 ka, an initial230Th/232Th activity ratio of 0.098 � 0.11 and a 234U/238U of1.571 � 0.021. However, it must be noted that T8 and Jula-23isochrons should be considered with caution, given that threepoints are theoretically not sufficient to build statistically robustisochrons.

5.2.4. Determination of the initial 230Th/232ThAlthough a correct determination of the initial 230Th/232Th

activity ratio is crucial, it is challenging to get both accurate andprecise determinations of this ratio (Placzek et al., 2006a). Thisdetermination is indeed complicated by the fact that this initial Thmay have two different sources with contrasted 230Th/232Th ratio:(i) a siliciclastic fraction, with an activity 230Th/232Th ratio that mayrange between 0.25 and 1.5 (Szabo and Rosholt, 1982), and (ii) an“hydrogenous” Th with higher 230Th/232Th ratios.

Imprecise determinations of this initial ratio may thus lead tosignificant inaccuracies in the computation of UeTh ages. Thissource of uncertainty may be particularly significant for sampleshaving low 238U/232Th ratios, such as samples Bo-93-23H and Bo-93-23M. Moreover previous work (Placzek et al., 2006a) suggeststhat the initial Th isotopic ratio may vary between carbonatesamples of the same lacustrine cycle. By measuring the 238U/232Thactivity ratio from several silico-clastic residues (Table 5), Placzeket al. (2006a) estimated the initial 230Th/232Th activity ratioassuming radioactive equilibrium. This estimates range between0.42 and 0.75, with an average value of 0.6 � 0.3 (2s) (Table 4 inPlaczek et al., 2006a). However, several pieces of evidence (Szaboand Rosholt, 1982; Haase-Schramm et al., 2004) as well as ourresults on sample T7, show that the detritus component was notnecessarily at radioactive equilibrium at the moment of thecarbonate precipitation. By using the average 232Th/238U value re-ported in Placzek et al. (2006a) and a composite (232Th/238U) vs(230Th/238U) isochron defined by plotting together all the data fromT7, T8 and the Tauca samples from Placzek et al. (2006b), we

Table 4U-Th data. Uncertainties are given at the 2s level.

Reference Sample Sample type Elevation (m) 238U (ppb) 232Th (ppb) 234U/238U 238U/232Th 234U/232Th 230Th/232Th 230Th/238U U-Th age(ka)a� 2s

Calibrated 14C age(ka cal BP)b� 2s

Tauca lakeSylvestre et al. (1999) PK Aragonite crust 3657 2645.3� 1.3 352.7� 2.2 1.542� 0.006 22.9� 0.1 35.3� 0.1 6.53� 0.11 0.285� 0.004 19.9� 0.9 18.7� 0.2Sylvestre et al. (1999) Hua-44 Radial tufa 3770 313.4� 0.9 21.6� 0.3 1.571� 0.007 44.3� 0.6 69.7� 1 9.96� 0.26 0.225� 0.006 15.6� 0.7 15.1� 0.5

This study Jah-12a Branching tufa 3670 436� 1.5 142.1� 1.8 1.531� 0.007 9.4� 0.1 14.4� 0.2 3.4� 0.2 0.363� 0.017 23.7� 2.6 -This study Jah-12b Columnar

stromatolithes3670 7216� 35 63.3� 0.4 1.555� 0.011 348.4� 2.6 542� 4.7 77.5� 0.7 0.222� 0.002 16.6� 0.2 16.1� 0.5

This study Jah-14a Laminatedstromatolithes

3670 362.8� 1.2 23.8� 0.2 1.608� 0.008 46.6� 0.4 75� 0.7 11.4� 0.2 0.244� 0.004 16.8� 0.5 14.6� 0.4

This study T7 Tufa 3775 433.1� 3.3 19.8� 0.04 - 67.0� 0.5 - 14.86� 0.17 0.222� 0.003This study T7 Tufa 3775 343.8� 0.8 34.3� 0.33 1.573� 0.006 30.6� 0.3 48.2� 0.5 7.17� 0.08 0.234� 0.004This study T7 Tufa 3775 355.5� 0.6 31.6� 0.28 1.577� 0.007 34.3� 0.3 54.2� 0.5 8.0� 0.1 0.232� 0.003This study T7 Tufa 3775 369.3� 0.4 74.8� 0.5 1.571� 0.005 15.1� 0.1 23.7� 0.2 3.92� 0.04 0.26� 0.003 Isochron ageThis study T7 Tufa 3775 384.8� 0.7 15.6� 0.06 1.589� 0.015 75.5� 0.3 120� 1.2 16.73� 0.14 0.222� 0.002 15.5� 0.2 -This study T8 Tufa 3775 513.1� 0.8 116.59� 0.44 1.512� 0.014 13.45� 0.06 20.3� 0.2 3.01� 0.08 0.224� 0.006This study T8 Tufa 3775 514.8� 0.2 39.85� 0.04 1.569� 0.004 39.48� 0.04 62.0� 0.1 8.46� 0.08 0.214� 0.002 Isochron ageThis study T8 Tufa 3775 538.6� 1.6 47.36� 0.06 1.531� 0.014 34.75� 0.11 53.2� 0.5 7.63� 0.08 0.22� 0.002 15.4� 6.3 -This study Jula-23 Tufa 3665 257.5� 0.1 72.0� 0.3 1.547� 0.01 10.93� 0.05 16.90� 0.08 2.87� 0.04 0.263� 0.004This study Jula-23 Tufa 3665 331.9� 0.2 51.2� 0.17 1.558� 0.008 19.8� 0.07 30.86� 0.12 5.12� 0.05 0.259� 0.003 Isochron ageThis study Jula-23 Tufa 3665 206.7� 0.3 39.4� 0.19 1.551� 0.012 16.04� 0.08 24.88� 0.16 4.18� 0.07 0.261� 0.005 19.0� 0.6 -Coipasa lakeSylvestre et al. (1999) BO-23H Calcareous crust 3680 307� 0.9 132.4� 0.7 1.657� 0.01 7.09� 0.04 11.7� 0.1 2.08� 0.09 0.293� 0.013 14.4� 3 13.2� 0.1Sylvestre et al. (1999) BO-23 M Calcareous crust 3680 235.6� 0.9 245.1� 3.4 1.565� 0.012 2.94� 0.04 4.6� 0.1 1.35� 0.15 0.458� 0.052 19.5� 8 13.2� 0.1

a Single U-Th ages are calculated using an initial 230Th/232Th ratio of 0.7� 0.3 (2s).b Calibrated using IntCal09 (Reimer et al., 2009).

P.-H.Blard

etal./

Quaternary

ScienceReview

s30

(2011)3973

e3989

3981

Fig. 5. Isochron plots for T8 sample. Plotted uncertainties are 2s.

Fig. 4. Isochron plots for T7 sample. Plotted uncertainties are 2s.

P.-H. Blard et al. / Quaternary Science Reviews 30 (2011) 3973e39893982

Fig. 6. Isochron plots for Julae23 sample. Plotted uncertainties are 2s.

P.-H. Blard et al. / Quaternary Science Reviews 30 (2011) 3973e3989 3983

calculated an average value of 0.7 � 0.3 (2s) for the 230Th/232Thsilico-clastic component of the Tauca cycle (Table 5). The mainsource of this uncertainty in the initial 230Th/232Th ratio resultsfrom the standard deviation of the 232Th/238U ratios used in thecalculations. This calculation suggests that the 230Th/232Th activityratio of the detritus component may be affected by significantheterogeneities from a sample to another. Although the T7isochrons yield a very precise determination of the initial230Th/232Th (0.73 � 0.06 at 2s), this value must be used withcaution to correct the other samples for the initial thorium. Indeed,given its low uncertainty, the T7 value probably does not addressthe large inter-sample variability. Consequently, in the absence ofisochrons for other lake cycles, we decided to follow a conservative

Table 5Compilation of the 230Th/232Th activity ratios (�2s) of the detrital fraction deter-mined for the Uyuni watershed.

Method - sample (230Th/232Th)initial Age (ka) Reference

Upper continental crust 0.84 e Taylor andMcLennan, 1995

Silico-clastic residues 0.7 � 0.3 w12 to w20 Placzeket al., 2006a

Isochron intercept - B9 0.5 � 0.4 16.3 � 1.7 Placzeket al., 2006a

Isochron intercept - B8 0.6 � 0.2 15.3 � 4.6 Placzeket al., 2006a

Isochron intercept -Tauca highstand

0.4 � 0.9 16.0 � 2.0 Placzeket al., 2006a

Isochron intercept - T7 0.73 � 0.06 15.5 � 0.2 This studyIsochron intercept - T8 0.2 � 2.4 15.4 � 6.3 This studyIsochron intercept -

Julae230.10 � 0.11 19.0 � 0.6 This study

approach and calculate all the new single UeTh ages by usinga value of 0.7 � 0.3 (2s) (Table 5). This approach also permittedrecalculation of some of the single UeTh ages previously published(Sylvestre et al., 1999). The propagation of this large uncertaintyprovides a decisive sensitivity test that permits an evaluation of theaccuracy of published Coipasa ages (Sylvestre et al., 1999), partic-ularly for those having a low 230Th/232Th ratio, such as samples Bo-93-23H and Bo-93-23M (Table 4).

UeTh ages were computed using the 230Th and 234U half livesdetermined by Cheng et al. (2000). Uncertainties of these UeThages are reported at the 2s level. They include the analyticaluncertainties (internal and external reproducibility) and the w25%error on the initial 230Th/232Th activity ratio (Table 4). The latter isthe major source of uncertainty since some of these samples arecharacterized by a large amount of detrital thorium. In other words,radiogenic 230Th produced since carbonate formation can be quiteclose to the initial amount of 230Th for samples having high232Th/238U ratios.

5.3. Discussion: comparison of UeTh and 14C ages and implicationfor the chronology of lakes Tauca and Coipasa

The previous study of Sylvestre et al. (1999) had only poorconstraints on the initial 230Th/232Th ratio in their samples, andtherefore calculated single UeTh ages using an initial activity ratioof 1. Under this assumption, co-determined 14C ages were about2 ka older than the reported UeTh ages for the lake Coipasa cycle.However, as discussed in the previous section, the 230Th/232Thinitial ratio is certainly lower than 1, and is probably closer to 0.7(Table 5). The recalculation of the UeTh ages of Bo-93-H and Bo-93-M with a 230Th/232Th activity ratio of 0.7 � 0.3 (2s) yielded UeTh

Fig. 7. Plot of 14C vs UeTh ages in the same samples. 14C age is calibrated using InCal09(Reimer et al., 2009). UeTh ages are calculated using an initial 230Th/232Th activity ratioof 0.7 � 0.3 (2s) for the detrital Th. The large error bar of the Bo-23H and Bo-23Msamples results from the uncertainty attached to this initial 230Th/232Th ratio.

P.-H. Blard et al. / Quaternary Science Reviews 30 (2011) 3973e39893984

ages with very large uncertainties (Table 4, Fig. 7). Consequently, asshown in a UeTh vs calibrated 14C plot (Fig. 7), UeTh ages ofsamples Bo-93-H and Bo-93-M do not constitute a firm constraintto diagnostic a possible 2 ka reservoir age for the carbon that wasdissolved in Lake Coipasa waters.

On the other hand, in the case of the samples belonging to theTauca (18e14.5 ka) and Sajsi (25e19 ka) cycle, there is an acceptableagreement between the 14C and UeTh ages (Fig. 7). One exceptionis for sample Jah-14a: with a 14C age of 14.6 � 0.4 ka, it is signifi-cantly younger than the UeTh age of 16.8 � 0.5 ka. This mismatchcould result from an inaccurate 230Th/232Th correction for thissample. Another possibility would be that the two analyzedaliquots of Jah-14a were not sampled at the same exact position inthe Jah bioherm, and, that they fortuitously record two differentformation episodes.

However, all other samples are close to the 1:1 line. A carefullook at the results of both dating methods even suggests that themajority of UeTh ages are slightly older than the calibrated 14C ages.These two observations suggest that the presence of dead carbon isunlikely for these paleolake waters, and, thus, that a reservoircorrection should not be applied to the 14C ages of the lakes Coipasaand Tauca. This interpretation is also supported by a publisheddataset (Placzek et al., 2006b) that shows a good agreementbetween the UeTh and 14C ages for samples younger than 20 ka.

The large bioherm from Quebrada Negrojahuila (Jah) is locatedat an elevation of 3670 m, several dozens of meters below the Sajsi,the Tauca and the Coipasa highstands (Fig. 1). If we accept thehypothesis that these large algal bioherms can only be developed inquite shallow water environments (Rouchy et al., 1996), it thusconstitutes a benchmark that has recorded the successive lacus-trine transgression and regression of each lake cycle (Fig. 2).

These new UeTh and 14C ages along with the data of Placzeket al. (2006b) and Sylvestre et al. (1999) allow us to propose thefollowing synthetic lake chronology (Fig. 8):

5.3.1. Sajsi cycle

1.1 - The Sajsi cycle occurred between w25 and w19 ka. Themaximum lake level of w3670 was reached as early as w23 ka.In particular, the new UeTh age of 23.7 � 2.6 ka for sample Jah-

12a suggests that the Lake Sajsi was thus about 15 m deep in theUyuni basin. This water depth is about 10 m higher than the oneproposed by Placzek et al. (2006b).1.2 - The low water level between the Sajsi regression(w20e19 ka) and the Tauca transgression (w18 ka) was prob-ably characterized by shallow water conditions, i.e. a thin waterlayer of about 5 m above the Salar de Uyuni. This lowstandepisode persisted between 1 and 2 thousand years. The termi-nation of this lowstand episode is probably recorded by samplePK (Table 2), an aragonite crust located at 3657 m. This sedi-mentological feature is indicative of shallow water conditionsand its UeTh and 14C ages (19.9 � 0.9 ka and 18.7 � 0.2 ka,respectively) provide a maximum age for the initiation of theLake Tauca transgression (Sylvestre et al., 1999).

5.3.2. Tauca cycle

2.1 - The Lake Tauca transgression started slowly after 18.5 ka.The rate of this initial water rise was very slow: at 17.5 ka, thewater level is only at 3670 m, which denotes a filling rate ofless than 10 m kyr�1 during the first 1 ka. After 17.5 ka, thesuccession of dated shorelines and tufas indicate a significantacceleration of the transgression: the water level reachedabout 3760 m just after 17 ka, which corresponds to a rate ofca 100 m kyr�1 in less than 1000 years. This transgressionscenario, however, seems to be in contradiction with the lakepaleobathymetry history previously proposed (Sylvestre et al.,1999). This study reconstructed the diatom assemblages of theLake Tauca from two sedimentary sections based on twodifferent outcrops at Churacari Bajo (19�720S, 67�310W,3685 m) and Pakollo Jahuira (19�320S, 67�310W, 3657 m). Akey observation of this work is that benthic diatoms weredominant until w16 ka (Fig. 3 in Sylvestre et al., 1999). Then,the abundance of benthic species decreased abruptly, whileplanktonic diatom simultaneously became dominant after16 ka. The dominance of benthic species theoretically impliesshallow water while planktonic species indicate deep water.These diatom evidences thus suggest that the main trans-gression phase could have actually occurred after 16 ka. It ishowever difficult to reconcile these diatom data with thepresence of many lake deposits older than 16 ka above 3700 m(Fig. 8A). Given that this apparent contradiction cannot beresolved without new geochronological data, we decided topropose 2 possible scenarios for the lake Tauca transgression(Fig. 8A). The first one (blue curve) is established by fitting allthe geochronological shoreline data. It is consistent with anearly lake filling: lake level would have reached 3760 m asearly as 17 ka. The second scenario (green curve) is based onthe diatom data. It suggests that the water level stagnated at3690 m between 17 and 15.7 ka. This stillstand episode wasthen followed by a very rapid transgression (>100 m kyr�1) at15.7 ka.2.2 - The Tauca highstand is clearly recorded by deep lakedeposits and several successive shorelines located between3765 and 3775 m, and is well dated between w16 and 14.5 ka(considering the UeTh and 14C ages having the lowest uncer-tainties). This highstand corresponds to a maximum waterdepth of 120 m in the center of the Uyuni basin, for a total lakesurface of about 52,000 km2 (Blodgett et al., 1997; Coudrainet al., 2002; Blard et al., 2009).2.3 - Although the exact timing of the regression is difficult todefine, two 14C ages from two bioherms described by Sylvestreet al. (1999) at w3740 m indicates that the water drop startedbefore 14.2 ka. The following regression was extremely rapid,

Fig. 8. A) Lake chronology of the Altiplano paleolakes during the last 30 ka. The blue curve has been drawn fitting all the geochronological data of the shoreline deposits. The greencurve indicates an alternative transgression scenario based on the diatoms assemblages determined by Sylvestre et al. (1999). B) North Atlantic SST Uk’37 record from ODP coreODP977A (Martrat et al., 2004; Martrat et al., 2007). C) Greenland temperatures from NGRIP d18O record (Andersen et al., 2004) rescaled according to the GICC05 time scale(Rasmussen et al., 2006; Svensson et al., 2006; Vinther et al., 2006). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of thisarticle.)

P.-H. Blard et al. / Quaternary Science Reviews 30 (2011) 3973e3989 3985

P.-H. Blard et al. / Quaternary Science Reviews 30 (2011) 3973e39893986

since the lake surface had dropped at 3660 m at 13.8 ka, whichindicates a rate of decrease of more than 120 m kyr�1 in lessthan 500 years.

5.3.3. Coipasa cycleOur new dates are in good agreement with the published

chronology (Placzek et al., 2006b):

3.1 The Coipasa transgression occurred before 13.3 ka andreached elevations as high as 3700 m at ca 12.5 ka. This corre-sponds to a water depth of about 45 m in the Uyuni basin.3.2 The Lake Coipasa highstand lasted about 1000 years and theregression was almost completed at 11.5 ka.

Although further data would be useful to better constrainthe exact timing of the end of the Coipasa cycle, it seems quiteclear that this humid event existed and that no further trans-gressiveeregressive cycles occurred after this event. The subsequentHolocene period was characterized by persistent dry conditions onthe Altiplano, suggesting that the rainfall amount remained lowerthan 330 mm yr�1 during this period (Condom et al., 2004).

6. Stable isotopes and d18O of paleolake waters

6.1. Results

d18O and d13C data are displayed in Table 6. For the paleocar-bonate from Lakes Coipasa and Tauca, d13C compositions range

Table 6Stable isotope data.

Sample Sample type Elevation (m) Analytical

Tauca lakeJah-12b Columnar micriobialite 3670 Manual linJah-12b Columnar micriobialite 3670 Manual lin

Jah-14a Laminated bioherm 3670 AutomatedJah-14a Laminated bioherm 3670 AutomatedJah-14a Laminated bioherm 3670 AutomatedJah-14a Laminated bioherm 3670 Automated

Piv-Ostra Ostracods 3765 Manual linPiv-Ga Gastropods 3765 Manual linHua-44 Radial tufa 3770 Manual linHua-44 Radial tufa 3770 Manual linHua-44 Radial tufa 3770 AutomatedHua-44 Radial tufa 3770 AutomatedHua-44 Radial tufa 3770 AutomatedHua-44 Radial tufa 3770 Automated

Coipasa lakeBo-93.26 Bioherm 3680 AutomatedBo-93.26 Bioherm 3680 AutomatedBo-93.26 Bioherm 3680 AutomatedBo-93.26 Bioherm 3680 AutomatedBo-93.26 Bioherm 3680 AutomatedBo-93.26 Bioherm 3680 AutomatedBo-93.26 Bioherm 3680 Automated

Jah-14b Tufa layers 3670 AutomatedJah-14b Tufa layers 3670 AutomatedJah-14b Tufa layers 3670 Automated

Modern gastropodsUru-Uru-Gas Gastropods 3708 AutomatedPoopo-Gas Gastropods 3696 Automated

The carbonateewater fractionation is calculated considering a crystallization temperatua Calculated assuming isotopic equilibrium and using the carbonateewater fractionat

from 1.768 � 0.002 to 5.438 � 0.002 (V-PDB) & and the d18Ofrom �2.186 � 0.005 to 0.357 � 0.005 (V-PDB) & (Table 6; Fig. 9).Replicates of sub-samples all agree within a range of 1& or less,which indicates that the isotopic compositions of these samples arequite homogeneous. The absence of correlation between the d13Cand the d18O values of the Coipasa and Tauca carbonates (R2 < 0.3)(Fig. 9) suggests that carbonate precipitation occurred near theisotopic equilibrium, and that kinetic fractionation or diagenesisoverprint were limited (Veizer et al., 1999). Lowest d13C values mayindicate inputs of soil carbonate into the lake. However, since thisreservoir is generally characterized by very negative values(<�10& V-PDB), the high d13C values of these lake carbonatesargue for a limited contribution of soil carbonate on the C budget ofthe lake.

6.2. d18O of lake paleowaters

d18O value of paleolake water is calculated using the equation of(Kim and O’Neil, 1997), assuming that waterecarbonate fraction-ation of oxygen occurred at equilibrium (Table 6, Fig. 10). Temper-ature of carbonate crystallization used for the calculation is7.5 � 2.5 �C. This estimate of the water paleotemperature is ob-tained by correcting the present-day temperature of Lake Titicaca(13.6 � 1.2 �C) for the 6 � 2 �C cooling determined by modeling thepaleoglacier equilibrium line contemporaneous with the LakeTauca highstand (Blard et al., 2009). This estimate requires theassumption that the airewater temperature difference has notchanged through time.

system d13C (& V-PDB) d18O (& V-PDB) d18O lake watera

(& V-SMOW)

e 5.438 � 0.002 �0.594 � 0.004 �2.03e 5.193 � 0.002 �1.173 � 0.006 �2.63

Average �2.3 � 0.7line 3.643 � 0.002 �0.402 � 0.006 �1.83line 2.777 � 0.002 �0.534 � 0.006 �1.97line 1.763 � 0.002 �1.358 � 0.003 �2.82line 3.271 � 0.014 �1.347 � 0.028 �2.74

Average �2.3 � 0.8e 2.963 � 0.002 �2.186 � 0.005 �3.67e e �1.39 � 0.005 �2.84e 4.755 � 0.003 �1.436 � 0.003 �2.90e 4.708 � 0.003 �1.667 � 0.008 �3.13line 4.893 � 0.002 �1.226 � 0.002 �2.68line 4.725 � 0.003 �1.056 � 0.006 �2.50line 4.819 � 0.003 �1.507 � 0.003 �2.97line 4.856 � 0.037 �1.411 � 0.071 �2.87

Average �2.8 � 0.6

line 2.535 � 0.002 0.052 � 0.003 �1.36line 2.782 � 0.002 �0.557 � 0.004 �1.99line 2.741 � 0.001 �0.313 � 0.004 �1.74line 3.082 � 0.002 0.348 � 0.004 �1.06line 2.504 � 0.003 0.357 � 0.005 �1.05line 2.676 � 0.002 �0.217 � 0.009 �1.64line 2.746 � 0.039 �0.887 � 0.072 �2.33

Average �1.6 � 0.8line 2.553 � 0.001 0.672 � 0.004 �0.72line 2.764 � 0.002 0.123 � 0.004 �1.29line 3.334 � 0.002 �0.924 � 0.003 �2.37

Average �1.5 � 1

line e 0.505 � 0.004 �0.94line e 3.105 � 0.005 1.66

re of 7.5 � 2.5 �C, a value derived from the paleoglacier extents (Blard et al., 2009).ion equation of (Kim and O’Neil, 1997): 1000ln a ¼ 18,030/T � 32.42.

Fig. 9. d18O vs d13C of the Lake Tauca and Lake Coipasa carbonates. The absence ofsignificant correlation between the oxygen and carbon isotopic compositions(R2 ¼ 0.23, p ¼ 0.02) is consistent with a minimal influence of diagenesis or kineticfractionation.

P.-H. Blard et al. / Quaternary Science Reviews 30 (2011) 3973e3989 3987

Calculated d18O for the Lake Tauca paleowater isbetween �2.3 � 0.8& and �3.7 � 0.6& V-SMOW (meanis �2.8 � 0.7&) while the subsequent Lake Coipasa cycle hasa mean d18Owater value of �1.6 � 0.9& (Table 6, Fig. 10). Thus, thewaters of these two paleolake episodes had significantly higherd18O values than present-day precipitation, which rangesbetween �20 and �5& over the Altiplano (Blisniuk and Stern,2005; Gallaire et al., 2010; Placzek et al., 2011). These d18O paleo-water estimates are also significantly above the�25 to�15& rangefor paleoprecipitation inferred from the isotopic record of ancienttropical ice (Thompson et al., 1998).

18O enrichment of lake waters can result from evaporationprocesses. An evaporation model was used to evaluate the sensi-tivity of the lake water isotopic composition to the amount ofevaporation over the lake and to atmospheric conditions. Using

Fig. 10. d18O of paleolake water vs time. Uncertainties in the calculated d18O waterinclude both the analytical reproducibility and the 2.5 �C error associated with thetemperature of 7.5 �C (Blard et al., 2009) used to calculate the carbonateewaterfractionation coefficient (Kim and O’Neil, 1997).

such an evaporation model, Placzek et al. (2011) showed that lakewater d18O is very sensitive to the amount of water that is locallyrecycled. The recycled water represents a closed-loop within thehydrological budget of the lake. Thus, when the proportion ofprecipitation arising from local evaporation is high, the mean d18Ovalue of lake water is closer to the regional rainwater composition(Craig and Gordon, 1965; Placzek et al., 2011). Although there arelarge uncertainties arising from the input parameters of the model(Craig and Gordon, 1965) (mainly humidity and wind speed), ourd18O estimate for Lake Tauca is consistent with a maximum localwater recycling of about 50% for the Lake Tauca episode. Thisestimate assumes that the relative humidity is <60% and usessimilar atmospheric conditions as those used in previous work (i.e.to build Fig. 6 of Placzek et al., 2011). Our data yield a valueof �1.6& for Lake Coipasa water, which is w1& higher than LakeTauca. This observation is consistent with the hypothesis that thefraction of local recycling is smaller when the size of the lakedecreases.

However, it must be emphasized that pH, salinity, as well askinetic and biological effects can also impact d18O (e.g. Carpenterand Lohmann, 1995; Schwalb et al., 1999; von Grafenstein et al.,1999; Arbuszewski et al., 2010), which may complicate the inter-pretations of these isotopic data.

7. Implications for regional climate and possibleteleconnections

This updated lake level chronology for the Altiplano providesstrong evidence that thewetedry fluctuations in the tropical Andesduring the Late Pleistocene are temporally correlated with thecoldewarm events recorded by the Northern Atlantic SST (Martratet al., 2004, 2007) and the Greenland temperatures records(Andersen et al., 2004; Rasmussen et al., 2006; Svensson et al.,2006; Vinther et al., 2006). The Lake Tauca and the subsequentLake Coipasa highstands are in phase with the Heinrich 1 eventsand the Younger Dryas (Fig. 8). However, a careful observation ofthese climatic records indicates that the initiation of the warmingevents recorded in the North Atlantic (Martrat et al., 2004, 2007)and Greenland (Andersen et al., 2004; Rasmussen et al., 2006;Svensson et al., 2006; Vinther et al., 2006) precede the lakedisappearance by w500 years (Fig. 8). Although dating uncer-tainties cannot be ignored in this discussion, such a delay can becompared with the constraints based on hydrological modeling(Condom et al., 2004): after a major change in regional climaticconditions, the total evaporation of Lake Tauca would requirebetween 100 and 500 years. Such delay suggests that the dramaticlateglacial hydrological changes observed on the Altiplano water-shed are not a precursor of the regional climatic shifts, but that theyare more likely a response to large continental changes in atmo-spheric patterns. A plausible mechanism might be the modulationof the ITCZ position due to fluctuations in meridional SST gradientsin the Northern Atlantic (Peterson et al., 2000; Baker et al., 2001;Haug et al., 2001; Kull et al., 2008). The trigger for changes inmeridional gradients and the ITCZ could be the extent of sea icecover in the Northern Atlantic (Denton et al., 1999; Chiang et al.,2003). Variability in tropical Pacific SSTs might also impact andinfluence Atlantic conditions through oceaneatmosphere interac-tions on interannual and decadal timescales (Vuille and Garreaud,2011). More research is required to better characterize these tele-connections over millennial timescales.

8. Conclusion

New 14C and UeTh data obtained fromAltiplano lake carbonateswere combined with published data (Sylvestre et al., 1999; Placzek

P.-H. Blard et al. / Quaternary Science Reviews 30 (2011) 3973e39893988

et al., 2006b) to refine the lake level chronology within the Alti-plano watershed over the last 25 ka:

(i) the Lake Sajsi cycle lasted between 25 and 19 ka and culmi-nated at 3670m at about 22 ka almost synchronously with theglobal LGM;

(ii) the Lake Tauca cycle lasted from 18 to 14.5 ka and is charac-terized by the highest water level, that reached at least 3770mfrom 16.5 to 15 ka;

(iii) the Lake Coipasa cycle lasted between 12.5 and 11.9 andreached an elevation of w3700 m.

These high amplitude lake level fluctuations recorded in tropicalsediments from the Late Pleistocene are notable as they are inphase with the Heinrich 1, BøllingeAllerød, and Younger Dryasclimate oscillations in the North Atlantic and Greenland. Thistemporal coincidence is strong support for the hypothesis that wetand dry events recorded in the Central Altiplano are controlled bynorthesouth displacements of the ITCZ as a result of changes in seasurface temperature gradients in the North Atlantic.

Oxygen and carbon isotopic measurements of these well-datedcarbonates indicate that lake waters had a d18O (V-SMOW)composition of�2.8� 0.7& for the Lake Tauca and�1.6� 0.9& forthe Lake Coipasa. These reconstructions suggest that the proportionof precipitation arising from local water recycling was probably lessthan 50% of the total water input to the lakes.

Acknowledgments

Jean-Marie Rouchy (Museum National d’Histoire Naturelle) iskindly thanked for providing us very well characterized samples.This work was funded by the French Programs INSU “Relief de laTerre” and “EVE-LEFE” and by NSF. Radiocarbon dating was per-formed at the Laboratoire de Mesure du Carbone 14 (LMC14) e

UMS 2572 (CEA/DSM e CNRS e IRD e IRSN e Ministère de laCulture et de la Communication) at Gif sur Yvette, France, on the 14CAMS facility ARTEMIS. John Eiler is thanked for providing us accessto his stable isotope lab at Caltech. Christa Placzek, Jay Quade andJon Patchett are thanked for their constructive reviews that helpedus improving the submitted version of the manuscript. Thiscontribution greatly benefited from the long term scientificcollaborative framework existing between l’Institut de Recherchepour le Développement (IRD) and la Universidad Mayor de SanAndrés, La Paz. This is CRPG contribution n�2139.

References

Andersen, K.K., Azuma, N., Barnola, J.M., Bigler, M., Biscaye, P., Caillon, N.,Chappellaz, J., Clausen, H.B., DahlJensen, D., Fischer, H., Fluckiger, J., Fritzsche, D.,Fujii, Y., Goto-Azuma, K., Gronvold, K., Gundestrup, N.S., Hansson, M., Huber, C.,Hvidberg, C.S., Johnsen, S.J., Jonsell, U., Jouzel, J., Kipfstuhl, S., Landais, A.,Leuenberger, M., Lorrain, R., Masson-Delmotte, V., Miller, H., Motoyama, H.,Narita, H., Popp, T., Rasmussen, S.O., Raynaud, D., Rothlisberger, R., Ruth, U.,Samyn, D., Schwander, J., Shoji, H., Siggard-Andersen, M.L., Steffensen, J.P.,Stocker, T., Sveinbjornsdottir, A.E., Svensson, A., Takata, M., Tison, J.L.,Thorsteinsson, T., Watanabe, O., Wilhelms, F., White, J.W.C., 2004. High-reso-lution record of Northern Hemisphere climate extending into the last inter-glacial period. Nature 431, 147e151.

Arbuszewski, J., deMonacal, P., Kaplan, A., Farmer, E.C., 2010. On the fidelity of shell-derived dO-18 seawater estimates. Earth and Planetary Science Letters 300,185e196.

Baker, P.A., Rigsby, C.A., Seltzer, G.O., Fritz, S.C., Lowenstein, T.K., Bacher, N.P.,Veliz, C., 2001. Tropical climate changes at millennial and orbital timescales onthe Bolivian Altiplano. Nature 409, 698e701.

Bischoff, J.L., Fitzpatrick, J.A., 1991. U-series dating of impure carbonates - anisochron technique using total-sample dissolution. Geochimica et Cosmochi-mica Acta 55, 543e554.

Blard, P.-H., Lavé, J., Farley, K.A., Fornari, M., Jiménez, N., Ramirez, V., 2009. Late localglacial maximum in the Central Altiplano triggered by cold and locally-wet

conditions during the paleolake Tauca episode (17e15 ka, Heinrich 1).Quaternary Science Reviews 28, 3414e3427.

Blard, P.-H., Lave, J., Farley, K.A., Tripati, A., Eiler, J.M., Sylvestre, F., Fornari, M., 2008.Late glacial paleoclimate of the central Altiplano constrained by cosmogenic3He dating and ‘clumped-isotope’ paleothermometry. In: Goldschimdt Confer-ence - July 13e18. Geochimica et Cosmochimica Acta, Vancouver.

Blisniuk, P.M., Stern, L.A., 2005. Stable isotope paleoaltimetry: a critical review.American Journal of Science 305, 1033e1074.

Blodgett, T.A., Lenters, J.D., Isacks, B.L., 1997. Constraints on the Origin of PaleolakeExpansions in the Central Andes. Earth Interactions 1.

Carmouze, J.P., 1991. El balance energético. In: C. D. a. A. I. ORSTOM (Ed.), El LagoTiticaca Paris.

Carpenter, S.J., Lohmann, K.C., 1995. dO-18 and dC-13 values of modern brachiopodshells. Geochimica et Cosmochimica Acta 59, 3749e3764.

Cheng, H., Edwards, R.L., Hoff, J., Gallup, C.D., Richards, D.A., Asmerom, Y., 2000. Thehalf-lives of uranium-234 and thorium-230. Chemical Geology 169, 17e33.

Chiang, J.C.H., 2009. The tropics in paleoclimate. Annual Review of Earth andPlanetary Sciences 37, 263e297.

Chiang, J.C.H., Biasutti, M., Battisti, D.S., 2003. Sensitivity of the Atlantic IntertropicalConvergence Zone to Last Glacial Maximum boundary conditions. Paleo-ceanography 18.

Condom, T., Coudrain, A., Dezetter, A., Brunstein, D., Delclaux, F., Sicart, J.-E., 2004.Transient modelling of lacustrine regressions: two case studies from theAndean Altiplano. Hydrological Processes 18, 2395e2408.

Coudrain, A., Loubet, M., Condom, T., Talbi, A., Ribstein, P., Pouyaud, B.,Quintanilla, J., Dieulin, C., Dupre, B., 2002. Isotopic data (87Sr/86Sr) andhydrological changes during the last 15,000 years on the Andean Altiplano.Hydrological Sciences Journal - Journal des Sciences Hydrologiques 47,293e306.

Craig, H., Gordon, L.I., 1965. Deuterium and oxygen 18 variations in the ocean andmarine atmosphere. In: Tongiogi, E., Lishi, V., Pisa, F. (Eds.), Stable Isotopes inOceanographic Studies and Paleotemperatures, pp. 9e130. Spoleto, Italy.

Cruz, F.W., Burns, S.J., Karmann, I., Sharp, W.D., Vuille, M., Cardoso, A.O., Ferrari, J.A.,Dias, P.L.S., Viana, O., 2005. Insolation-driven changes in atmospheric circula-tion over the past 116,000 years in subtropical Brazil. Nature 434, 63e66.

Delanghe, D., Bard, E., Hamelin, B., 2002. New TIMS constraints on the uranium-238and uranium-234 in seawaters from the main ocean basins and the Mediter-ranean Sea. Marine Chemistry 80, 79e93.

Delclaux, F., Coudrain, A., Condom, T., 2007. Evaporation estimation on Lake Titi-caca: a synthesis review and modelling. Hydrological Processes 21, 1664e1677.

Denton, G.H., Heusser, C.J., Lowell, T.V., Moreno, P.I., Andersen, B.G., Heusser, L.E.,Schluchter, C., Marchant, D.R., 1999. Interhemispheric linkage of paleoclimateduring the last glaciation. Geografiska Annaler Series A-Physical Geography81A, 107e153.

Gallaire, R., Taupin, J.D., Coudrain, A., 2010. Recent developments in the cryo-climatic parameters and isotopic content of precipitation in the BolivianAndes: La Paz and the Zongo Glacier. Hydrological Sciences Journal - Journal desSciences Hydrologiques 55, 467e483.

Garreaud, R.D., Vuille, M., Compagnucci, R., Marengo, J., 2009. Present-day SouthAmerican climate. Palaeogeography, Palaeoclimatology, Palaeoecology 281,180e195.

Ghosh, P., Adkins, J., Affek, H., Balta, B., Guo, W.F., Schauble, E.A., Schrag, D.,Eiler, J.M., 2006. 13Ce18O bonds in carbonate minerals: a new kind of paleo-thermometer. Geochimica et Cosmochimica Acta 70, 1439e1456.

Haase-Schramm, A., Goldstein, S.L., Stein, M., 2004. U-Th dating of Lake Lisan (latePleaistocene Dead-Sea) aragonite and implication for glacial East Mediterra-nean Climate change. Geochimica et Cosmochimica Acta 68, 985e1005.

Hastenrath, S., Kutzbach, J., 1985. Late Pleistocene climate and water-budget of theSouth-American Altiplano. Quaternary Research 24, 249e256.

Haug, G.H., Hughen, K.A., Sigman, D.M., Peterson, L.C., Rohl, U., 2001. Southwardmigration of the intertropical convergence zone through the Holocene. Science293, 1304e1308.

Henderson, G.M., Slowey, N.C., Fleisher, M.Q., 2001. U-Th dating of carbonate plat-form and slope sediments. Geochimica et Cosmochimica Acta 65, 2757e2770.

Huntington, K., Wernicke, B., Eiler, J., 2010. Influence of climate change and uplift onColorado Plateau paleotemperatures from carbonate clumped isotope ther-mometry. Tectonics 29 TC3005.

Huntington, K.W., Eiler, J.M., Affek, H.P., Guo, W., Bonifacie, M., Yeung, L.Y.,Thiagarajan, N., Passey, B., Tripati, A., Daëron, M., Came, R., 2009. Methods andlimitations of ‘clumped’ CO2 isotope (D47) analysis by gas-source isotope ratiomass spectrometry. Journal of Mass Spetrometry 44, 1318e1329.

Innocent, C., Bollinger, C., Chabaux, F., Claude, C., Durand, N., Le Faouder, A., Kiefel,B., Pomies, C., 2004. Intercomparison of new Th isotopic standards: preliminaryresults. In: Springer (Ed.), 14th Annual V M Goldschmidt Conference. Geo-chimica et Cosmochimica Acta, Univ. Copenhagen, Denmark, pp. A321eA321.

Kim, S.T., O’Neil, J.R., 1997. Equilibrium and nonequilibrium oxygen isotope effects insynthetic carbonates. Geochimica et Cosmochimica Acta 61, 3461e3475.

Klein, A.G., Seltzer, G.O., Isacks, B.L., 1999. Modern and last local glacial maximumsnowlines in the Central Andes of Peru, Bolivia, and Northern Chile. QuaternaryScience Reviews 18, 63e84.

Kronfled, J., 1974. Uranium deposition and Th-234 alpha-recoil - explanation forextreme U-234/U-238 fractionation within trinity aquifer. Earth and PlanetaryScience Letters 21, 327e330.

Kull, C., Imhof, S., Grosjean, M., Zech, R., Veit, H., 2008. Late Pleistocene glaciation inthe Central Andes: temperature versus humidity control - a case study from the

P.-H. Blard et al. / Quaternary Science Reviews 30 (2011) 3973e3989 3989

eastern Bolivian Andes (17�S) and regional synthesis. Global and PlanetaryChange 60, 148e164.

Laval, B., Cady, S.L., Pollack, J.C., McKay, C.P., Bird, J.S., Grotzinger, J.P., Ford, D.C.,Bohm, H.R., 2000. Modern freshwater microbialite analogues for ancientdendritic reef structures. Nature 407, 626e629.

Leduc, G., Vidal, L., Tachikawa, K., Rostek, F., Sonzogni, C., Beaufort, L., Bard, E., 2007.Moisture transport across Central America as a positive feedback on abruptclimatic changes. Nature 445, 908e911.

Ludwig, K.R., 2003. Isoplot 3.0 - A Geochronologial Toolkit for Microsoft Excel,Berkeley Geochronological Center Special Publication N�4, pp. 1e71.

Ludwig, K.R., Titterington, D.M., 1994. Calculation of 230Th/U isochrons, ages anderrors. Geochimica et Cosmochimica Acta 58, 5031e5042.

Martrat, B., Grimalt, J.O., Lopez-Martinez, C., Cacho, I., Sierro, F.J., Flores, J.A.,Zahn, R., Canals, M., Curtis, J.H., Hodell, D.A., 2004. Abrupt temperature changesin the Western Mediterranean over the past 250,000 years. Science 306,1762e1765.

Martrat, B., Grimalt, J.O., Shackleton, N.J., de Abreu, L., Hutterli, M.A., Stocker, T.F.,2007. Four climate cycles of recurring deep and surface water destabilizationson the Iberian margin. Science 317, 502e507.

Mook, W.G., van der Plicht, J., 1999. Reporting 14C activities and concentrations.Radiocarbon 41, 227e239.

New, M., Lister, D., Hulme, M., Makin, I., 2002. A high-resolution data set of surfaceclimate over global land areas. Climate Research 21, 1e25.

Osmond, J.K., May, J.P., Tanner, W.F., 1970. Age of Cape-Kennedy barrier and lagooncomplex. Journal of Geophysical Research 75, 5459e5468.

Peterson, L.C., Haug, G.H., Hughen, K.A., Röhl, U., 2000. Rapid changes in thehydrologic cycle of the tropical Atlantic during the Last Glacial. Science 290,1947e1951.

Placzek, C., Patchett, P.J., Quade, J., Wagner, J.D.M., 2006a. Strategies for successfulU-Th dating of paleolake carbonates: an example from the Bolivian Altiplano.Geochemistry Geophysics Geosystems 7.

Placzek, C., Quade, J., Patchett, P.J., 2006b. Geochronology and stratigraphy of latePleistocene lake cycles on the southern Bolivian Altiplano: implications forcauses of tropical climate change. Geological Society of America Bulletin 118,515e532.

Placzek, C.J., Quade, J., Patchett, P.J., 2011. Isotopic tracers of paleohydrologic changein large lakes of the Bolivian Altiplano. Quaternary Research 75, 231e244.

Rasmussen, S.O., Andersen, K.K., Svensson, A.M., Steffensen, J.P., Vinther, B.M.,Clausen, H.B., Siggaard-Andersen, M.L., Johnsen, S.J., Larsen, L.B., Dahl-Jensen, D.,Bigler, M., Rothlisberger, R., Fischer, H., Goto-Azuma, K., Hansson, M.E., Ruth, U.,2006. A new Greenland ice core chronology for the last glacial termination.Journal of Geophysical Research 111. doi:10.1029/2005JD006079.

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.,Weyhenmeyer, C.E., 2004. IntCal04 terrestrial radiocarbon age calibration,0e26 cal kyr BP. Radiocarbon 46, 1029e1058.

Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G.,Ramsey, C.B., Buck, C.E., Burr, G.S., Edwards, R.L., Friedrich, M., Grootes, P.M.,Guilderson, T.P., Hajdas, I., Heaton, T.J., Hogg, A.G., Hughen, K.A., Kaiser, K.F.,Kromer, B., McCormac, F.G., Manning, S.W., Reimer, R.W., Richards, D.A.,Southon, J.R., Talamo, S., Turney, C.S.M., van der Plicht, J., Weyhenmeye, C.E.,2009. IntCal09 and Marine09 radiocarbon age calibration curves, 0e50,000years cal. BP. Radiocarbon 51, 1111e1150.

Richerson, P., Widmer, C., Kittel, T., 1977. The Limnology of Lake Titicaca (Peru-Bolivia), a Large, High Altitude Tropical Lake Institute of Ecology Publication 14.

Rondeau, B., 1990. Géochimie isotopique et géochronologie des stromatoliteslacustres quaternaires de l’Altiplano bolivien.

Rosholt, J.N., 1976. 230Th/234U dating of travertine and caliche rinds. Geol. Soc.America Abs. with Programs 8, 1076.

Rouchy, J.M., Servant, M., Fournier, M., Causse, C., 1996. Extensive carbonate algalbioherms in upper Pleistocene saline lakes of the central Altiplano of Bolivia.Sedimentology 43, 973e993.

Schwalb, A., Bruns, S.J., Kelts, K., 1999. Holocene environments from stable isotopestratigraphy of ostracods and authigenic carbonate in Chilean Altiplano Lakes.Palaeogeography Palaeoclimatology, Palaeoecology 148, 153e168.

Servant, M., Fontes, J.C., 1978. Les lacs quaternaires des hauts plateaux des Andesboliviennes: Premières interprétations paléoclimatiques. Cahiers de l’ORSTOM,Série géologie 10, 5e23.

Svensson, A., Andersen, K.K., Bigler, M., Clausen, H.B., Dahl-Jensen, D., Davies, S.M.,Johnsen, S.J., Muscheler, R., Rasmussen, S.O., Rothlisberger, R., Steffensen, J.P.,Vinther, B.M., 2006. The Greenland Ice Core Chronology 2005, 15e42 ka. Part 2:comparison to other records. Quaternary Science Reviews 25, 3258e3267.

Sylvestre, F., 1997. La dernière transition glaciaireeinterglaciaire (18 000e8000 14Cans B.P.) dans les Andes tropicales sud (Bolivie) d’après l’étude des diatomées.PhD thesis - Museum National d’Histoire Naturelle de Paris.

Sylvestre, F., Servant, M., Servant-Vildary, S., Causse, C., Fournier, M., Ybert, J.P., 1999.Lake-level chronology on the southern Bolivian Altiplano (18�e23� S) duringlate-glacial time and the early Holocene. Quaternary Research 51, 54e66.

Szabo, B.J., Rosholt, J.N., 1982. Surficial continental sediments. In: Ivabovich, M.,Harmon, R.S. (Eds.), Uranium Series Disequilibrium: Applications to Environ-mental Problems. Clarendon Press, pp. 246e267.

Tapia, P., Fritz, S., Baker, P., Seltzer, G., Dunbar, R., 2003. A Late Quaternary diatomrecord of tropical climatic history from Lake Titicaca (Peru and Bolivia).Palaeogeography Palaeoclimatology, Palaeoecology 194, 139e164.

Taylor, S.R., McLennan, S.M., 1995. The geochemical evolution of the continentalcrust. Reviews of Geophysics. 33, 241e265.

Thompson, L.G., Davis, M.E., Mosley-Thompson, E., Sowers, T.A., Henderson, K.A.,Zagorodnov, V.S., Lin, P.N., Mikhalenko, V.N., Campen, R.K., Bolzan, J.F., Cole-Dai, J., Francou, B., 1998. A 25,000-year tropical climate history from Bolivian icecores. Science 282, 1858e1864.

Tripati, A.K., Eagle, R.A., Thiagarajan, N., Gagnon, A.C., Bauch, H., Halloran, P.R.,Eiler, J.M., 2010.13Ce18O isotope signatures and ‘clumped isotope’ thermometryin foraminifera and coccoliths. Geochimica et Cosmochimica Acta 74, 5697e5717.

Veizer, J., Ala, D., Azmy, K., Bruckschen, P., Buhl, D., Bruhn, F., Carden, G.A.F.,Diener, A., Ebneth, S., Godderis, Y., Jasper, T., Korte, C., Pawellek, F., Podlaha, O.G.,Strauss, H., 1999. Sr-87/Sr-86, delta C-13 and delta O-18 evolution of Phanero-zoic seawater. Chemical Geology 161, 59e88.

Vinther, B.M., Clausen, H.B., Johnsen, S.J., Rasmussen, S.O., Andersen, K.K.,Buchardt, S.L., Dahl-Jensen, D., Seierstad, I.K., Siggaard-Andersen, M.L.,Steffensen, J.P., Svensson, A., Olsen, J., Heinemeier, J., 2006. A synchronizeddating of three Greenland ice cores throughout the Holocene. Journal ofGeophysical Research 111. doi:10.1029/2005JD006921.

von Grafenstein, U., Erlernkeuser, H., Trimborn, P., 1999. Oxygen and carbonisotopes in modern fresh-water ostracod valves: assessing vital offsets andautecological effects of interest for palaeoclimate studies. PalaeogeographyPalaeoclimatology, Palaeoecology 148, 133e152.

Vuille, M., Garreaud, R.D., 2011. Ocean-atmosphere interactions on interannual todecadal timescales. In: Matthews, J.A., Bartlein, P.J., Briffa, K.R., Dawson, A.G., DeVernal, A., Denham, T., Fritz, S.C., Oldfield, F. (Eds.), Handbook of EnvironmentalChange. Sage Publications, London, Los Angeles, New Delhi, Singapore.

Wilhelm, S., Hintze, T., Livingstone, D.M., Adrian, R., 2006. Long-term response ofdaily epilimnetic temperature extrema to climate forcing. Canadian Journal ofFisheries and Aquatic Sciences 63, 2467e2477.