Lawrence Livermore National Laboratory UCRL-ID-126643 Holocene Paleohydrology of the Tropical Andes from Lake Records M.B. Abbott G.O. Seltzer K.R. Kelts J. Southon March 3, 1997 This is an informal report intended primarily for internal or limited external distribution. The opinions and conclusions stated are those of the author and may or may not be those of the Laboratory. Work performed under the auspices of the U.S. Department of Energy by the Lawrence Livermore National Laboratory under Contract W-7405-ENG-48.
Transcript
Lawrence�
Livermore�
National�
Laboratory
UCRL-ID-126643
Holocene Paleohydrology of the Tropical Andes from Lake Records
M.B. AbbottG.O. SeltzerK.R. KeltsJ. Southon
March 3, 1997
This is an informal report intended primarily for internal or limited external distribution. The opinions and conclusions stated are those of the author and may or may not be those of the Laboratory.Work performed under the auspices of the U.S. Department of Energy by the Lawrence Livermore National Laboratory under Contract W-7405-ENG-48.
DISCLAIMER
This document was prepared as an account of work sponsored by an agency of the United StatesGovernment. Neither the United States Government nor the University of California nor any of theiremployees, makes any warranty, express or implied, or assumes any legal liability or responsibility forthe accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed,or represents that its use would not infringe privately owned rights. Reference herein to any specificcommercial product, process, or service by trade name, trademark, manufacturer, or otherwise, doesnot necessarily constitute or imply its endorsement, recommendation, or favoring by the United StatesGovernment or the University of California. The views and opinions of authors expressed herein donot necessarily state or reflect those of the United States Government or the University of California,and shall not be used for advertising or product endorsement purposes.
This report has been reproduceddirectly from the best available copy.
Available to DOE and DOE contractors from theOffice of Scientific and Technical Information
P.O. Box 62, Oak Ridge, TN 37831Prices available from (423) 576-8401
Available to the public from theNational Technical Information Service
U.S. Department of Commerce5285 Port Royal Rd.,
Springfield, VA 22161
Holocene Paleohydrology of the Tropical Andes from Lake Records
Mm B. ABBOTT
Department of Geosciences, Merrill Science Center, University of Massachusetts, Box 35820, Amherst, Massachusetts 01003-5820
GEOFFREY O. Samm
Department of Earth Sciences, Heroy Geology Luboratoty, Syracuse University, Syracuse, New York 13244-1070
KmutY R. KELTS
Limnologicai Research Center, University of Minnesota, 220 Pillsbwy Hall, 310 Pillsbury Dr SE, Minneapolis, Minnesota 55455
AND
Jom SOWON
Center for Accelerator Mass Spectrommy, Lawrence Livemtore National Labomtoq, P.O. Box 808, L-397, Livemsore, Cai~omia 94551-99W
Two century-scale time series in northern Bolivia constrain theages of abrupt changes in the physia gmchemicd, and biologicalcharacteristics of sediments obtained fkom Iakes that formed dur-ing deglaciation from the late %istocene glacial maximum. Thewatersheds of Lagum Vichani (16012’S, 68%7’ W, 3780 m) andhgo Taypi Chaka Kkota (16013’S, 68’’21’W,4300 m), located onthe eastern and western slopesof the Cordillera Real, respectively,contain small cirque glaciers. A high-resolution chronology of thelake sediments is provided by 23 AMS “C dates of discrete macro-fossils. Late Pleistocene glaciers retreated rapidly, exposing thelake basins between 10,700 and 9700 *’C y B.P. The sedimentaryfacies suggest that after 8900 “C yr B.P. glaciers were absent fromthe watersheds and remained so during the middle Holocene.Anincrease in the precipitation-evaporation balance is indicatedabove unconformities dated to -2300 l’C yr B.P. in both LagoTaypi Chaka Kkota and bguna Vihani. An abrupt increasein sediment accumulation rates after 1400 “C yr B.P. signals theonset of Neoglaciation. A possibleIii exists betweenthe observedmillennial-scale shifts in the regional precipitation--evaporationbalance and seasonal shifts in tropicat insolation. raIswu~vxw ~wMhin@nL
INTRODUCTION
Lake-sediment records from both sides of the CordilleraReal, Bolivia. were analyzed to investigate the Holoceneenvironmental dynamics of the Bolivian Andes (Fig. 1).These time series also extend the 3500-yr record of changesin the precipitation-evaporation balance for the region es-tablished by dating water-level fluctuations in Lake Titicaca(Abbott et af., inpress; Binford er al., in press). The shallow
southern basin of Lake Titicaca (Lago Wliiaymarka) con-tains an incomplete Holocene record because it was almostcompletely desiccated prior to about 3500 14Cyr B.P. (Wirr-mann et al., 1990; Wirrmann and Mourguiart, 1995; Abbott,1995). Also, all of the closed basin lakes in the cordilleraare ephemerrd lakes lacking a complete Holocene recordthat extends through the proposed tnid-Holocene dry phase.Therefore, we targeted glacier-fed lakes with a positive water ,balance to obtain continuous Holocene paleodimate se- ~quences. Although the glacial lakes in this study maintainoverflowing stages throughout the dry season (May-No-vember), the erosion and deposition histories of Lago TaypiChaka Kkota and Laguna Viscachani provide strong evi-dence for a dry middle Holocene. We suggest that cirquegIaciers were absent from the watersheds during the middleHolocene, and without ice to provide a meltwater bufferduring the dry season, a lake would have been seasonallylower or become desiccated. The absence of a meltwaterbuffer also rendered these high-altitude lake systems moresusceptible to annual to century-scale droughts during themiddle Holocene.
Previous water-level studies in Lake Titicaca imply a neg-ative trend in the regional precipitation-evaporation balancefrom the latest Pleistocene to the middle Holocene (Servantand Fontes, 1978; Wirrmann and De Oliveira Almeia 1987;Mourguiart, 1990). Studies of glacier retreat and highersnowlines for this period provide further evidence supportinga negative water balance in the region (Gouze et al., 1986;Seltzer, 1992).
Here we present new proxy climate data from two sitesin tropical South America indicating a middle Holocene dry
..
iI
‘L\
\\
\\
\//
\3
I1
II
/-/
-\\
phase, resulting in the nearly complete melting of glaciersbetween 8900 and 2300 14Cyr B.P. in watersheds with head-walls of 5500 m altitude and lower in the Cordlllera Real,Bolivia. Terrestrial records of Holocene deglaciation basedon radiocarbon dates of peat downvalley from terminal mo-raines indicate that the glaciers had receded to within 200m of their modem extent by 9700-8600 “C yr B.P. (Seltzer,1992; Seltzer et al., 1995). Late Holocene glacier advancesappear to have obliterated evidence of any moraines thatmay have formed during the middle Holocene. Therefore,only by investigating Iacustrine sediment records can wedetermine if glaciers melted completely during the early andmiddle Holocene. We also discuss a possible link betweenobserved millennial-scale shifts in the regional precipita-tion-evapomtion balance and seasonal shifts in insolationincident in the tropics of the Southern Hemisphere.
CLIMATE SETTING OF STUDY AREA
Pronounced seasonal contrasts in precipitation am charac-teristic of the Andean Altiplano (14-21°S). During the Aus-tral summer, a heat-induced low-pressure cell forms overthe center of South America that draws moismre from theSouth Atlantic. At the same time, convective activity overthe Altiplano associated with the “Bolivian High” (Acei-tuno and Montecinos, 1993) produces precipitation in theAndes. As a result, most precipitation falls between Decem-ber and March in this region. Periodc blocking of the east-erly atmospheric circulation by the westerlies (Aceittmo andMontecinos, 1993: Kessler, 1988) and perturbations causedby El Niiio Southern Oscillation (ENSO) events (Thompsonet al., 1984: Ribstein et al., 1995; Francou et al., 1995) canproduce dry phases on the Altiplano during the normal wetseason. We contend that the synoptic climatic condhionsthat lead to seasonal and intra-seasonal variations in effectivemoisture today on the Altiplano serve as reasonable analogsfor changes that may have occurred over longer periodsduring the Holocene.
This study uses sediment records collected from both sidesof the Cordillera Real, Bolivia, to investigate changes ineffective moisture. The two study sites are sepamted by 25km: Lago Taypi Chaka Kkota (16013’S, 68°21 ‘W, 4300 maltitude) is located in the Rio Palcoco valley on the westernside of the cordillera (Fig. 1A). Laguna Viscachani (16012’S,68”07’W, 3740 m) is situated in the Rio Zongo valley onthe eastern side of the divide (Fig. lB). Both watershedscontain small cirque glaciers, with headwalls at 5500 maltitude.
The eastern cordillera of the northern Bolivian Andes isa zone of steep climatic gradients resulting from a relief of7600 m above the Amazon Basin. The result is a pro-nounced rain shadow, with precipitation decreasing east towest from >1400 mndyr in the lowlands to c700 rnrn/yr
in the Altiplano (Hoffman, 1975). Roche etal.(1992)esti-mated that the highest peaks of the Cordillera Real receive>800 mm/yr precipitation; 50 km to the west over the Alti-plano this decreases to <500 rnm/yr. Ribstein et al. (1995)measured -900 mm of precipitation during the 1992-1993hydrological year on Zongo Glacier, located on the continen-tal divide 10 km from Laguna Viscachani and 20 km fromLago Taypi Chaka K.kota. Summer precipitation (Decem-ber-March) accounts for 65 to 78% of the annual total, andwinter precipitation accounts (June-September) for only 3to 8% of the annual total (Roche et al., 1992).
METHODS
Cores were taken with a square-rod piston corer (Wright et
al., 1984) and a piston corer designed to collect undisturbedsediment-water interface profiles (Fisher et al., 1992). Totalcarbon (TC) and total inorganic carbon (TIC) were measuredwith a UIC Coulomernc System. Total organic carbon(TOC) was calculated by the difference of TC-TIC. Bio-genie-silica was analyzed by a time-series dissolution experi-ment (DeMaster, 1979; DeMaster, 198 1). Core Iithology wasdetermined from smear-slide mineralogy and detailed in-spection of sediments noting Munsel color, texture, sedimen-tary structures, and biogenic features. Magnetic susceptibil-ity was measured with a Barrington Susceptibility Bridge atlow frequency. Values were corrected for mass differenceswith bulk density measurements.
Temestrkd macrofossils were not present in sufficientquantities for AMS 14Cmeasurements at most stratigraphiclevels. Therefore, we used Isoetes macrofossils for AMS 14Cmeasurements from Lago Taypi Chaka Kkota and Myriophy- .hum macrofossils from Laguna Viscachani. The contempo- - ‘rary radiocarbon reservoir was assessed by measuring the14Cactivity of live samples of these submerged macrophytesfrom each lake. The results of radiocarbon measurements of .living aquatic vegetation are 114 and 111% Modem for theyears A.D. 1992 and 1994, respectively, indicating that thelake reservoir effects are minimal in these systems; however,it is possible that the reservoir age has changed with time.
The radiocarbon ages were measured at the Center forAccelerator Mass Spectrometry (CAMS), Lawrence Liv-ermore National Laboratory. Radiocarbon ages are reportedeither as 14C yr B.P. (uncalibrated) or cd ‘~ yr B.P. ifcorrected and calibrated according to the methods outlinedfor CALIB 3.0 by Stuiver and Reimer (1993). Accumulationrates (g/cmz/lOO yr) werecalculated using crdibrated datesas products of sediment accumulation rates (cmlyr) and bulk&nsity (g/cm3) without considering errors associated withradiocarbon measurements or calibration.
RESULTS
Lago Taypi Chaka Kkota
AMS “C dates indicate that the transition from glacialsilt to organic-rich lacustrine sediments occurred at 10,790
05 10 15 5~5 oOrgwlkc (Wt%) cm WtRatio Dry& Densilf
.-
Radiocarbon ages and sediment pro~rnes, including organic carbon, biogenic silica, C/N ratio, mass magnetic susceptibility, and dry bulkdensity, from core E collected from Lago Tayp~Chaka Kkota. - -
* 60 l~C y B.P. (CAMS-19243) in the Lago Taypi ChakaKkota lake basin (4300 m), which is located more than 10km downvalley and 1200 m below the catchment headwall(5500 m) (Fig. 2. Table 1). Seltzer (1992) estimated thatthe time of maximum late Pleistocene glaciation in Bolivia(Choqueyapu II Glaciation) was about 14,000 “C yr 13.P.based on glacial-geologic studies in the Bolivian-peNviattAndes. Gouze er al. (1986) obtained a maximum limitingage for the Choque yapu 11glaciation of 16,600 3 130 14Cyr B.P. on the western side of the Cordillera Real. In theRio Palcoco valley, the terminal moraine assumed to haveformed during the Choqueyapu II glaciation is -5 km down-valley from Lago Taypi Chaka Kkota. Seltzer (1991) de-scribed rapid deglaciation of the Rio Palcoco valley on thebasis of bulk radiocarbon dates obtained from the base ofthe organic-rich lacustrine deposits in Lago Taypi ChakaKkota and Lago Allka Kkota. situated less than 3 km fromthe headwall. These dates are 10,460 ~ 140 (Beta-35071)and 9980 t 90 “C yr (Beta-35069), respectively. Further-more, the glaciers receded to withht their Neoglacial limitby 8640 z 80 “C yr (Beta-35052) as indicated by a date
from the transition of glacial silt to fibrous peat in a bog200 m from the present glacier front at 4670 m (Seltzer,1991). Glacier advances during the late Holocene extendedless than 200 m from the present glacier fronts in the Cordil-Iera Real (Gouze et al., 1986; Seltzer, 1990).
The sediments that accumulated in Lago Taypi Chaka
. .
Kkota between 9300 t 100 (CAMS- 1OO64)and 8810 x 110l*C yr (CAMS- 10065) have higher concentrations of organicmatter and blogenic silic% and a fourfold increase in accu-mulation rate compared to sedimentation between 10,790”.and 9300 t 100 “C yr B.P. (Figs. 2 and 3). Between 10,7$@;* 60 and 9300 t 100 14Cyr B.P. the inorganic content ofthe sediments decreased from >80 to <40%, indicating thatthe glacial sediment source was greatly reduced or absent.Biogenic silica increased from <20 to >60% just prior to9300 t 100 14Cyr. This shift signals a marked increase inlake productivity and corresponds with a continued increasein accumulation rate. Well-preserved Isoetes plant fragmentsoccur above the 1080 cm level, suggesting a reduction inthe concentration of glacial silt in the water column, allowinglight transmission to the lake bottom. The higher lake pro-ductivity is consistent with increased nutrient input into thelake system during the initial stages of soil stabilization ina recently deglaciated watershed.
A marked shift in the sedimentation pattern occurred after8810 f 110 yr B.P., based on a change from mm-scalelaminations to cm-scale bands. The cm-scale bands thatformed between 88103 110 and 5320 t 110 “C yr B.P.(CAMS-4979) rdtemate between dark-brown laminae con-taining aquatic macrophytes and megaspores and light-grayIaminae with lower organic carbon content and lacking well-preserved macrofossils. Two hypotheses are proposed to ex-plain the cm-scale bands. One is that variations in glacier
TABLE 1AMSRadiocarbonDates from Lago Taypi ChakaKkotaand LagunaVkachani
Measured‘Wage Median calibrated ‘~ ageLake Depth (cm) Materiat (“C Y B.P.) (Cat “C yr B.P.) cAMs-#
Lago Taypi Chaka Kkota 860.5 aquatic macrofossil 1470 t 80 1310Lago Taypi Chaka Kkota 879.5
aquatic macrofossil 9590 z 70 10,760Laguna Vkcachani 737
16,064aquatic macrofossil 97902 70 10,970 17,016
“ Not used in sedimentation rate calculations.
activity drove century-scale changes in sediment input to thelake basin; the other is that variable lake level occurredduring the middle Holocene, caused by the disappearanceof glaciers from the watershed and the loss of glacial meltwa-ter as a buffer during the dry season.
In the first hypothesis, increased glacier activity may ex-plain the formation of the Iigbt-gray cm-scale bands. Al-though a small increase in the mineral content of the lightbands is noted, the inorganic accumulation rate is decreasedin these intervals. This is not consistent with the inferenceof increased glacier activity. Furthermore, the blogenic-silicacontent of the light cm-scale bands retins >40%, in con-trast to values of <5% at present and <2% during the earlyHolocene. Therefore, this hypothesis is considered unlikely.
In the second hypothesis, modern analogs are offered toaid the interpretation of the past system. lsoetes is foundonly in very shallow water (< 1 m deep) in the contemporarysystem. Assuming this was also true in the paSL the darkcm-scale kmtinae that contain well-presexved fsoetes plantmatter could only be formed during century-scale periodswhen the lake level was seasonally below its overflowingstage. This implies that glaciers were greatly reduced orabsent from the watershed after 8810 t 110 14C yr B.P.and is consistent with sedimentological evidence providedabove. Low lake levels are unlikely in the modem systembecause glacial meltwater maintains the water level at its
overflowing stage even during the dry season. A modemexample of a sirniiar lake system without glaciers in it’swatershed is Laguna Koilpa Kkota (17’’26’S, 6788’W, 4400 -m). In this system kike level is seasonally lower and the-lake often is desiccated during the dry season because thecatchrrtentIacks glaciers to buffer water level (Seltzer, 1994). .
A marked decrease in the organic carbon content occurs .at 5320 ~ 110 ‘~ yr B.P. This suggests lower lake produc-tivity or increased decomposition of organic matter duringthe period between 5320 t 110 and 2880 ~ 60 yr B.P.(CAMS-1 1068). Furthermore, increases in the inorganicmatter conten~ mass magnetic susceptibility, and bulk den-sity occur with little change in the rate of inorganic sedimen-tation. Again, two hypotheses could explain this interval.Either the lake system returned to ovetllowing conditionsafter 5320 t 110 “C yr B.P. or the lake was seasonallylower and occasionally desiccated after this time.
If the first hypothesis is comec~ decreased lake productiv-ity signals the return of year-rmmd overflowing conditionsdriven by the return of cirque glaciers to the Lago TaypiChaka Kkota watershed. However, this is inconsistent withthe erosion history of the catchmenc as recorded in the lakesediments. Increased glacier activity would have resultedin greater input of fine-grained glacial sediment, producinghigher accumulation rates and lower lake productivity. Al-though trends in sediment analyses are consistent with lower
bgo TaypiChaka Kkota AmumulationRates(g/cm2/l00 yr)Core E
Measured
Total Sediment SiogenicSilicaRadiocarbonAW
S30
1=—-----lOYrT J ,4:::
Rm ‘---
E“A~30 ,::), .,’
1080
.~-r tb io 40 80 100 0.01 0.1 i 0:1 i io
Weight PercentOrganic Malter Inorganic Matter
FfG. 3. Comparison of or.ga”icmatter.biogenic-siticaand inorganicmattercontentand theirrespectiveacc.mulatim ratesplomed on a logscalefrom core E in Lwo Taypi C~aka KkoIa.
productivity. the total accumulation rste decreased in theintervalfrom5320t 1IO to 2290 t 60 14Cyr B.P. (CAMS-11067). The average inorganic accumulation rste for thisperiod (0.4 @cm’/ 100 yr) dropped to less than hslf the ratecalculated for the mndem system (0.9 g/cmJ/lOO yr). Thissuggests that if glacien were present, they were less exten-sive than now.
If the second hypothesis is correct, the lake would havebeen seasonally lower or desiccated during the dry seasonfor the period from 5320 * 110 to 2290 * 60 ‘“C yrB.P. Accumulation rates reach the lowest Holocene valuesduring this phase. suggesting that sediments could havebeen partially lost by deflation when the lake was desic-cated. Seasonal exposure of organic sediments may haveresulted in higher rates of organic matter oxidation (KWlops and Killops. 1993) and increased hulk density. Pro-gressive decreases in organic carbon and biogenic-silicacombined with increases in bulk density, inorganic mattercontent. and mass magnetic susceptibility are consistentwith desiccation surfaces. The decreases in accumulationrates described above and the apparent unconformity from2880 t 60 to 2290<60 “C yr B.P. (880 cm) suggest [batthe lake was seasonally dry or desiccated for an extendedperiod prior to 2290 ~ 60 “C yr B.P. Peaks in massmagnetic susceptibility and ARM magnetization occur atthe 887 cm level. Water levels in Lake Titicaca dropped
10 to 12 m between 2450 and 2250 “C yr B.P. beforerising abruptly after 2250 “C yr B. P., supporting the hy-pothesis of a dry phsse at this time (Abbott, 1995).
Above the 229il 2 60 “C yr B.P. level in the sediment ~core, abmpt increases in organic carbon snd bingenic-silica,combined with decreases in mass magnetic susceptibility and.bulk density, indicate a return to more humid environmentalconditions and higher Iske productivity. These conditionswere temporay, as orgsnic carbon and biogenic-silica con-tent decreased rapidly to near late Pleistocene values by 1470~ 80 “C yr B.P. (CAMS-4980). After this time, inorganiccontent, mass magnetic susceptibility, and bulk density in-cressed rnpidly. Organic cnrbon accumulation rates fellabruptly to tbe Inwest Holncene values (<0.1 g/cm2/ 100 yr)and the inorganic accumulation rstes increased to the highestvslues since the early Holocene (>0.9 g/cmz/lOO yr), whenthe glaciers were in tbe final sfsges nf recession.
.faguna Viscachani
The transition fmm glacial silts to organic+icb Iacustrinesediments nccurred at 9790 * 70 yr B.P. (CAMS-17016) inLaguna Viscachsni (3740 m). This contact is lLMO“C yryounger than the 10,790 z 60 “C yr B.P. age obtained fmmLago Taypi Cbaka Kkota (4300 m) on the western side ofthe range. The latest Pleistocene ice limit in tbe Lago Taypi
FIG.
LagunaVkcachaniCore B
BiogenicSi (wt %)Magnetisms=-ptibility
(Slxlo%kg)
o 10 20 30 0 0.5 1 1.5 28mrl ——
MeasuredRadiocarbonAge
(w y B.P.)
115.47% Mcdm7 (CAMS’
)’ ‘?a
#
I. --t) 5 10 5~1 o 0.5 1 1.5
OrganicC(wt‘%.) C/N Ratio DtyBulkDensity(g/cm3)
13ZOMO(CAM3-17011)ls50t30 (CM4S-17012)
2440iso (CWa-17010]
moi30 (CM4S-17013)
~ (CAMS-1504314oeois0 (CAMS17014)
S%oHO (OAMS-W015)
4530i70(CAMS-MOWsmcam(cmts-lm15j
4. Radiocwbon ages and sediment properties.including organic carbon, biogenic silica. UN ratio, mass magnetic susceptibility, and dry bulkdensity, from core B collected from Laguna Viscachsmi.
Chaka Kkota watershed is at -4200 m altitude, indicatingthat glaciers on the eastern side of the mountains advancedat least 460 m lower than those on the west. This supportsmodem observations of a pronounced rain shadow acrossthe Cordillera Real (Seltzer, 1990). Therefore, the LagunaViscachani watershed should receive higher average annualprecipitation and be more prone to the regeneration of cirqueglaciers after the middle Holocene dry phase thau the LagoTaypi Chaka Kkota catchment. Furthermore, the higher pre-cipitation on the eastern side of the Cordillera Real rendersLaguna Viscachani less susceptible to periods of prolongeddry-season desiccation.
After 9790 z 70 “C yr B.P., abrupt increases in organicmatter and biogenic-silica content signal higher lake produc-tivity. Figures 4 and 5 show that organic matter (>0.2 g/cmz/l OOyr) and biogenic-silica (>0.7 g/cm2/100 yr) accu-mulation rates were at their highest Holocene levels. Organicmatter sedimentation decreased from 9790 x 70 to 5450 t90 “C yr B.P. (CAMS-16063), when it reached it’s lowestHolocene level of 0.03 g/cmz/lOO yr. lle biogenic-silicaaccumulation rate decreased before stabilizing after 8550 *70 “C yr B.P. (CAMS- 17015). The inorganic accumulationrate was stable after 9590270 “C yr B.P. (CAMS- 16064)until 5450 t 90 ‘JC yr B.P. The subfacies formed during themiddle Holocene vary from organic-rich kuninae containing
well-preserved macrophytes to lower organic sedimentswithout macrofossils. These deposits are consistent with a .”lake system that shifted between overflowing conditions and”seasonally lower water levels.
Decreased accumulation rates between 5450 * 90 and2510 ~ 80 “C yr B.P. (CAMS-17013) were synchronous .with a similar trend in Lago Taypi Chaka Kkota. This sup-ports the hypothesis that water levels were seasonally lower.Furthermore, the inorganic accumulation rate is the lowestof any period during the Holocene (0.5 g/cm2/100 yr), sug-gesting that cirque glaciers were not present in these water-sheds. An initial increase in the biogenic-silica content fol-lowed by a steady decline from 5450 t 90 to 2510580‘~ yr B.P. is the same trend observed in the Lago TaypiChaka WOW core during thiS period.
After 2510 k 80 “C yr B. P., accumulationratesof organicmatter, biogenic-silic~ and inorganic matter increased. Thelargest increase was in the inorganic matter accumulationrate from -1.0 to 2.2 g/cm2/100 yr. An abrupt increase inthe mass magnetic susceptibility after 2480 t 60 “C yr B.P.(CAMS- 17010) and inorganic matter content are consistentwith the return of cirque glaciers in the Laguna Viscachanicatchment. After 1220 t 80 “C yr B.P. (CAMS-17011),the accumulation rate of inorganic matter doubled. This iscoincident with the stabilization of increased values of mass
bguna ViscachaniCore B
500
660
71a
760
Accumulation Rates (gkmzll 00 Y)
ToIal Sediment Biogenic SiIics
0.1 1 10 0.01 0.1 1——
4-4 —
MeasuredRadiocarbon Age
(’w y B.P.)
, 15.47% W. (CAMSwcc+)
,22e6n,cAw-170v),- (CAM$-17012)
I I I I
1—
I—
Q 26 4Q 60 60 100 0.01 0.1 1 0.1 1 10
Weight Precent Organic Matter Inorganic Matter
PtG. 5. Comparison of organic matter, biogenic-silica, and inorganic marter content and their respective accumulation fares pbrred on a log satefmrn cm-e B in Lagma Viscachani.
magnetic susceptibilityy and bulk densiry, suggesting Neogla-ciation. Furthemrow organic matter and blogenic-silica con-tent rctumed to pre-Holocene vshtes.
CLIMATIC CONTROLS ON REGIONAL GLACIATION
Little information exists on the mid-HoIocene activity ofglaciers in the tropical Andes. In northern Bolivia we can saywith cenainty that deglaciation at the end of the Pleistoceneproceeded rapidly, so that by -10,000 “C yr B.P. manyglaciers were near or within their modem limits. In the lateHolocene Neoglaciation is evidenced by extensive, sparselyvegetated moraines only within 200 m of modem glaciers.However, there is little geomorphic record of mid.Holoceneglacier activity in this area.
We suggest that glaciers are critical for maintaining theperennial lakes in the Cordillera Real. During the exten-sive dry period that occurs every year in the austral winter,meltwater generation from glaciers in these watershedsaPPears to maintain the overflowing condition of the lakes.This is supported by the observation that ephemeral lakesin the cordillera occur in basins that are not linked tomodem glaciers. Thus, factors that influence glacier massbalance may also indirectly influence the hydrologic massbalance of tbe lakes,
Francou e! cd. ( 1995) assessed the mass balance of Zongo
Glacier (16017’S, 68”09’W, 6000-4890 m) and measuredwater discharge, tempefsmm, mrd pfecipitstion from Sep-tember 1991 through August 1993 to determine the response ~of modem rropical glaciers to climatic variability. ZongoGlacier is on the eastern side of the Huayna Potosi massif ~‘<10 km from Laguna Viscacbani. The study period spanneda strong ENSO event that occurred during the summer of1991– 1992. Frmrcou et al. (1995) compared the “normal”1992-1993 hydrological year with the 1991-1992 ENSOevent and concluded that (1) there was a marked negativewater balance during the ENSO event year, with water loss
stnounting to twice the precipitation gain, (2) the equilib-rium-line altitude of the glacier rose 200 m, (3) the glacieraccumulation-area rstio (AAR) decreased fmm 86 to 58%,and (4) a very low accumulation rate occurred at high eleva-tions. These effects are directly related to increased radiation
receipt at the glacier smface, which causes wsrmer summertemperatures and results in a higher rste of ablation. Further-more, tbe summer wet season lasted only 2 months duringthe 1991-1992 ENSO event instead of the 4 months of a“normal” hydrological year.
Fmncou et al. (1995) concluded that factots controlling themass bslance of glaciers in tie tropics include the duration ofthe wet season, and rbe tempcrarorc dudng the pcricds thatprecede and follow the rainy season. Temperature and pmcipi-wion during the momlrs of October-December and Msrch–
May appear to be critical for determining the mass balance oftropical glaciers. The association of ENSO events with lowaccumulation or negative mass balance qmted by Fmncou etal. is paralleled by ice-tote evidence of decmsed precipitationon the Quelccaya ice cap (Thompson et d., 1985).
Ribs@in et al. (1995) observed that during the 1991-1992 ENSO event, Zongo Glacier produced increased runoffdespite lower precipitation values compared with the 1992–.1993 “normal” hydrological year. The decrease in precipi-
tation corresponded to decreased cloud cover, which allowedmore solar radiation to reach the glacier surface. Thus, ex-tended periods of low precipitation would first lead to in-creased mekwater production by the glaciem followed bydesiccation of these valleys after the glaciers had completelyablated.
Wetter conditions on the northern Altiplano are associatedwith the Bolivian High, which develops as a result of convec-tive precipitation over this region in the austral summer (J.Lenters and K. Cook, unpub. data). We assume here thatany decrease in summer insolation would be related to adecrease in precipitation on the Altipkmo, thereby havingthe effect of decreasing cloudiness and increasing the receiptof solar radiation at the glacier surface. Given the conceptualmodel above, the glaciers would ablate and eventually dkap-pear if these phases lasted long enough. Also, an increasein insolation in the austrrd winter would lead to increasedglacier ablation during that season.
GCM simulations by Kutzbach and Guetter (1986)showed that at 9000 yr B.P. (-8450 14C yr B.P.) whenperihelion occurred in July the seasonality of solar radiationwas decreased in the Southern Hemisphere. Climate modelsimulations. predicted that incident solar radiation was de-creased in January by 28 W/mz (6%) and increased in Julyby 11 W/mz (5%) relative to the A.D. 1950 value (Fig. 6A).These changes caused (1) decreased seasonality, resulting incooler summers and warmer winters, (2) decreased sea-levelpressure over the ocean and increased sea-level pressure overland, resulting in a lower pressure gradient and decreasedtransport of water vapor over the continenL and (3) a de-crease in the net precipitation-evaporation balance, withthe decreased precipitation being more significant than thedecreased evaporation. The summer and winter average in-solation values returned to near-modem levels and the rateof change slowed about 3000 yr B.P. (-2900 ‘~C yr B.P.)which generally corresponds with wetter conditions on theAltiplano (Fig. 6A). Figure 6B illustrates a plot of monthlyinsolation at 20°S for the Austral summer and transitionalmonths for the past 20,000 yr. For the months of December-February which are the peak of the wet season, insolationvalues were at their lowest between 12,000 and 8000 yr B.P.(- 10.000-7600 “C yr B.P.) (Edwards et al., 1993), whichcorresponds with deglaciation and the beginning of the mid-Holocene dry phase observed at the sites in this stu~dy.Insol-
~ 4~
g :;;- :
z -14. :i=
-16-:
-18
-20, ...,...,...,..4 -6 4 -2 0
~8
Insolation(% Dtierenee from A.D. 1950)
O,_
a.-Dsi? I-8 -6 -4 -2 0 2V 4 6 8-
1 1
Insolation(’?’o Dffferenee from A.D. 1950)
0.
-2-jc
-8-64-20246Insolation(% Differencefrom A.D. 1950)
FtG. 6. InsolationchanEcsfor ttw PSI X3,(XMyr B.P. as cotcutatedfrom@~@wby~(197tik ~(1~), d~~h(lWl).(A) Summer snd winter insolationvalues for 15%and 2Y showing seasomdinsolationvatucs math present levels after 4(MXIyr B,P. (3700 ‘y yr B.P.).Summer andwintcf vsluessre dcfincdasthc 90dayperiod centercdon thewtsticc. (B) Monthty insolation for the summer Wetsesson and(c) fortl’lewinterdry season.
ation values for the months of June-August reached theirhighest between 13,000 and 7000 yr B.P. (1 1,000-6600 14Cyr B.P.), likely resulting in increased ablation during thewinter months (Fig. 6C). Thus, seasonality in the SouthernHemisphere was decreased over the past 20,000 yr, peakingat the Pleistocene/Holocene transition. Given the scenarioabove, a decrease in summer insolation may have resultedin lower summer precipitation; an increase in winter insola-tion may have resuhed in enhanced ablation of the glaciersduring the winter months. The combination of decreasedsummer precipitation and increased melting in the winterwould contribute to the disappearance of cirque glaciers fromwatersheds with headwalls lower than 5500 m. Both factorswould contribute to the prolonged dry phase in the middleHolocene documented in this study.
CONCLUSIONS
The sediment records from Lago Taypi Chakrt Klcota andLaguna Viscachani contain striking similarities in the patternand timing of geochemicrd, biological, and sedimentologicalchanges, supporting the hypothesis that these systems arerecording a regional climatic signal. Twenty-three AMS 14Cdates constrain the timing of the climatic shifts and yieldsufficient time resolution to determine the Holocene erosionand deposition history of these catchments. Sediment coreanalyses showing abrupt shifts from nearly pure silt to in-creasingly organic-rich silt, suggest that glaciers recededabove 4300 m altitude on the western side of the CordilleraReal by 10,700 lqC yr B.P. and above 3740 m on the easternside of the range by 9700 14Cyr B.P. Increased accumulationrates of organic matter and biogenic silica coupled with anabrupt decrease in the inorganic content of the sediments inboth watersheds. suggests that the cique glaciers disap-peared between 9700 and 8900 14Cyr B.P. High accumula-tion rates of organic carbon and biogenic silica fmm 9700to 5400 14Cyr indicate more favorable environmental condi-tions and higher lake productivity than today. Cm-scalebands of sediments deposited during this period bear shal-low-water macrophytes indicating water levels of <1 m andalternate with laminae containing lower organic content andno macrofossils >0.5 pm in size. This suggests that therewere century-scale periods of seasonally lower water levels.After 5400 “C yr B.P., macrophytes are absent from thesediments, but biogenic silica accumulation rates increaseuntil the interval between 2900 and 2500 14Cyr B.P. TheLago Taypi Chaka Kkota core shows an abrupt increase ininorganic content, bulk density, and mass magnetic suscepti-bility afler 4000 lJC yr B.P., which is most likely the resultof prolonged droughts leading to low-water stands duringthis interval. No such changes are observed in the LagunaViscachani core during this time, but accumulation rates inboth lakes reached their lowest level between 5400 and 2300
*4Cw B.P. A desiccated surface forming an uncontortmtyin the Lago Taypi Chaka Kkota core has an age of 2300 14Cyr B.P., indicating that higher water levels were reachedafter this time. After 2300 14Cyr B.P. sediment characteris-
tics became increasingly similar to those of latest Pleistocene
glacial sediments, indicating the return of cirque glaciers to
watersheds with headwalls at 5500 m and intense Neoglacial
activity after 1400 14C yr B.P.
ACKNOWLEDGMENTS
We thank Herbert Wrigh~ Alexander Wolfe, and Ramon Aravena forproviding criticat comments on an earlier version of the manuscript. TheFamilia Ando of La ~ Tbe Servicio t3eologico de Bolivia andORSTOM-Bolivia all provided valuable logistical assistance. We also thankChalmersCtappettonand Eric Leonard fortheirreviews of this manuscript.The workwas supportedby a gmduate fellowship and grants of the NSF-RTG at theUniversity of Minnesota and the L~ologicat Research Center and fieldgrantafrom Geological Society of Atneric~ Sigma Xi, and the Depamnentof Geology and Geophysics, Unive.mityof Minnesota.
REFERENCES
AbboG M. B. (1995). Holocene climatic vsriabUity for lake sites in theBolivian Andes andinteriorAlaska basedon sedimentcdogyandradkmr-bon dating by acceleratormass specrrometty.Unpublished Ph.D. disserta-tion, University of Minnesota.
Abbott, M. B., BinfomLM. W., Bmrtster,M., and Kelts, K. R (in press). A3500 “C yr high-resolutionrecordof lake level changes in Lake Thicaca,South America QuarternaryResearch.
keiNno, P.,andMontecinos,A. (1993). Circulation aoomahes associatedwith dry and wet periods in the South American Altiplano. In “Fourth
InternationalConferenceon SouthernHemisphere Meteorology andOceanography, American Meteorological Society,” pp. 330-331.
Berger, A. ( 1978b). A simple atgorithm to compute long term variations.of daily or monthly insolation. Cormibution No. 18, Univetsit6 CathoI-ique de Louviam htstitutd’Astronornieet de Geophysique, G. Lemaitre,Louvian4a-Neuve, B- 134S Belgique.
Berger. A., and LoutrcAM. F. (1991). Insolation values for the climate ofthe last 10 milticaryears. QssaternmyScience Reviews 10,297-3 t7.
Binford, M. W., Kolata, A. L., Brcnner,M., Janusek,J., Abbott, M. B., andCurtis, J. H. (in ptess). Climate variation aad the rise and fat] of anAndean civilization. QuatemaW Research.
DeMaster, D. J. ( 1979). The marinebudgets of silica and 3*Si.Unpublishedph.D. dissertation. Yale University, New Haven.
DeMaster, D. J. (198I). The supply and accumulationof silica in the marineenvironment. Geachemica et Cosnsachimica Acra 45, 1715-1732.
Edwards, R. L., Beck, W. J.. Burr.G. S., Donahue, D. J., Chappell.J. M. A.,Bloom, A: L.. Dmffel, E. R. M., and Taylor, F. W. (1993). A large dropin atmospheric “CJ’% and reduced melting in the Younger Dryas, docu-mented with % ages of corats. Science 260, %2-968.
t%sher,M. M.. Brenner,M., andReddy, K. R. ( 1992). A simple, inexpensivepiston corer for collecting undisturbedsediment/water interface pmftles.Journal of Paleoliwusoiogy 7, 157-161.
Francou, B., Ribstein, P., Saravi& R., and Ttiau+ E., (1995). Monthlybalance and water dkcharge of an inter-tmpicat giacien Zongo Glacier.Cordillers Real, Bolivia l@S. Journal of Glaciology 41,61-67.
Gouze, P., Argollo. J.. Safiege. J.-F.. and Servant, M. (1986). hrterpr6tationpaf60climatique des oscillations des glaciers au tours dcs 20 demiersmillhaires darts les tigions tropicales: exemple des Andes Boliviennes.Comptes Rendus de L’,4cadAttie des Sciences Paris, .%rie II 303, 219–223.
Hoffman, J. A. J. ( 1975). “Climatic Adas of South America.” World Mete-orological Organization. Ginebta. Hungary.
Kessler, A. ( 1988). “Die Schwankungen des Wasserhaushaltes de sitdameri-kanischen Ahiplano und das Weltklima.’” v. Jahrbuch der Geograph-,ischen Gesellschaft zu Harmover. 139-159.
Killops. S. D., and KNops, V. J. (1993). “An Introduction to Organic Geo-chemistry.” Longman Scientific & Technical, New York.
Kutzabach, J. E.. and Guetter, P. J. ( 1986).The influence of changing orbitalparameters and surface boundary conditions on climate simulations forthe p@ 18,000 year% Journal of the Atmospheric Sciences 43, 1726-1759.
Mourguiart, P. (1990). Une approche nouvelle du problkme pow.$par lesreconsrmtions des pa160niveaux lacusrres: utilisation d’une fonction detransfert brke sur Ies faunes d’osrracodes. Giodwwrnique 5, 151-166.
Ribstein, P., Tkiau. E., Francou, B.. and Satavia, R. ( 1995). Tropical climateand glacier hydrology: a case study in Bolivia. Journal of Hydrofogy
165, 221-234.
Roche, M. A., Bourges. J. C., and Mattes, R. (1992). Climatology andhydrology of the Lake Tlticaca basin. In ‘lake Titicaca: a synthesis ofIimnologicai knowledge”- (C. Dejoux and A. Iltis, Eds.), pp. 63-83.Khswer, Boston.
Seltzer. G. O. ( 1990). Recent glacial history and pafeoclimate of the Peru-viars-Bolivian Andes. Quarerwan Science Reviews 9, 137--152.
Seltzer. G. O. ( 199I). Glacial history and climate change in the Peruvian-Bolivian Andes. Unpublished Ph.D. dissertation. University of Minne-sota.
This work was performed in part underthe auspices of the U.S. Departmentof Energy at Lawermce LivermoreNational Laboratory under contractW-7405-ENG-48 .
Seltzer, G. O. (1992). Late Quatemary glaciation of the Corddlera Real,Bolivia. Journal of @atemary Science 7,87-98.
Seltzer, G. O. ( 1994). A lacustrine record of late Pleistocene climatic changein the subtropical Andes. Boreas 23, 105-111.
Seltzer, G. O., Rodbell, D. T., and Abbott, M. B. (1995). Andean GlaciafLakes and Climate Variaiility Since the Last Glacial Maximum. Bufletinde l’Institut Francais d’lftudes Andines, 24,539-550.
Servant. M., and Fontes, J.-C. (1978). L.es Lacs quatemaires des hamsplateaux des Andes bolivierms premieres interpretations pal60climat-iques. Cahiers de 1‘ORSTOM, Sine Gt!ologique 10, 9-23.
Stuiver. M., and Reimer, P. J. (1993). Extended “C data base and revisedCALIB 3.0 “C age crdibrationprogram. Radiocarbon 35,215-230.
Thompson, L. G., Mosley-Thompson, E., Bolzan, J. F., and Koci, B. R.(1985). A 1500-yr record of tropicrd precipitation in ice cores from theQuelccaya Ice Cap, Peru. Science 229,971-973.
Thompson, L. G., Mosley-Thompson, E., and Morales, B. (1984). El NifioSouthern Oscillation events recorded in the stratigmphy of the tropicalQuelccaya ice cap. Pem. Science 226,50-53.
Wkrrnatm, D., and Fernando De Oliveira Almeiti L. (1987). Low Holo-cene level (7700 to 3650 years ago) of Lake Th.icaca (Bolivia). Pafaeoge-ography, Palaeoclimatology, Palaeoecology 59, 315-323.
Wirrmatm, D., and Mourguiti P. (1995). Late Quatematy spatio-temporallimnologicaf variations in the AMphmo of Bolivia and Peru. QuatemaryResearch 43, 344-354.
Wirrmarsn, D., Mourguiart, P., and Fernando De Oliveira Afmeida. L.(1990). Holocene sedimentology and ostracods distribution in Lake Titi-caca-pafeohydrologicaf interpretations. Jn’ ‘Quatemary of SourhAmericaand Antarctic Peninsula. Vol. 6“ (J. Rabassa. Ed), pp. 89-129. A. A.Balkema, Rorterdam.
wright, H. E.. Mann, D. H., and G1aser. P. H. (1984). Piston corers for Peatand lake sediments. Ecology 65, 657-659.
