quaternary science reviews · stable isotope values (d18o&d13c) of multiple ostracode species...

at SciVerse ScienceDirect

Quaternary Science Reviews xxx (2012) 1e16

Contents lists available

Quaternary Science Reviews

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

Stable isotope values (d18O & d13C) of multiple ostracode species in a largeNeotropical lake as indicators of past changes in hydrology

Liseth Pérez a,f,*, Jason Curtis b, Mark Brenner b, David Hodell c, Jaime Escobar d,e, Socorro Lozano a,Antje Schwalb f

a Instituto de Geología, Universidad Nacional Autónoma de México (UNAM), Ciudad Universitaria, 04510 Distrito Federal, MexicobDepartment of Geological Sciences and Land Use and Environmental Change Institute, University of Florida, Gainesville 32611, Florida, USAcGodwin Laboratory for Palaeoclimate Research, Department of Earth Sciences, University of Cambridge, Downing Street, Cambridge CB2 3EQ, UKdDepartamento de Ingenieria Civil y Ambiental, Universidad del Norte, ColombiaeCenter for Tropical Paleoecology and Archaeology, Smithsonian Tropical Research Institute, Panamaf Institut für Geosysteme und Bioindikation, Technische Universität Braunschweig, Langer Kamp 19c, 38106 Braunschweig, Germany

a r t i c l e i n f o

Article history:Received 22 June 2012Received in revised form13 October 2012Accepted 17 October 2012Available online xxx

Keywords:Stable isotopesAutecologyOstracodaNeotropicsLate Pleistoceneeearly HoloceneEnvironmental changeLake hydrodynamicsLake level change

* Corresponding author. Instituto de Geología, Unive4290x221; fax: þ52 55 5622 4281.

E-mail address: [email protected] (L. Pé

0277-3791/$ e see front matter � 2012 Elsevier Ltd.http://dx.doi.org/10.1016/j.quascirev.2012.10.044

Please cite this article in press as: Pérez, L., eindicators of past changes in hydrology, Qua

a b s t r a c t

Modern lake hydrodynamics, ostracode species autecology, stable isotopes (d18O and d13C) of multipleostracode species, ostracode taphonomy and sediment geochemistry were studied to improve inter-pretation of the late Pleistoceneeearly Holocene (w24e10 ka) stable isotope record of ostracodes insediment core PI-6 from Lago Petén Itzá, northern Guatemala. Oxygen and carbon stable isotopes inmodern and fossil species assemblages of Lago Petén Itzá were used as indicators of changes in thebalance between evaporation and precipitation, past lake level and carbon source. Ostracodetaphonomy was used to detect past periods of strong currents, high-energy environments, and possiblepartial or full mixing of the lake. The modern lake water isotopic composition displays clear seasonaldifferences that are independent of lake level fluctuations. Modern benthic species displayed lowerd18O and d13C values than nektobenthic species, with differences of 3.0& and 5.3&, respectively. Valvesof nektobenthic species display higher values of d13C because these ostracodes live in shallowerenvironments among abundant algae and aquatic plants, where productivity is high. The benthicspecies Limnocythere opesta Brehm, 1939 displayed the smallest average offset from d18O water (þ0.3&)and the largest offset from d13CDIC values (�4.1&) among studied ostracode species. Nektobenthicspecies Heterocypris punctata Keyser, 1975 displayed the smallest difference relative to the d13CDIC

values (�0.1&).Late Pleistoceneeearly Holocene climate conditions and water levels in Lago Petén Itzá can

be summarized as follows: 1) high lake levels and cold conditions (Last Glacial Maximum [LGM],w24e19 ka), 2) fluctuating lake levels and cold conditions (Heinrich Stadial 1 [HS1], w19e15 ka), 3)high lake levels and warm and wetter conditions (Bølling-Allerød [BA], w15e13 ka), 4) low lakelevels and dry conditions (Younger Dryas [YD], w13e11.5 ka) and 5) high lake levels and warm andwetter conditions (early Holocene, w11.5e10.0 ka). Average lake level fluctuation in Lago Petén Itzáduring the late Pleistoceneeearly Holocene was as much as w25 m. Ostracode analyses suggest thatthe LGM was characterized by relatively low d18O (þ4.7 to þ6.0&), and d13C values (�7.1 to �6.4&)in ostracode valves, high inferred water depths and high percentages of broken adult and juvenilevalves (>66%), suggesting a high-energy environment, strong currents, partial to full mixing,downslope transport, colder water temperatures and wetter conditions. An increase in the relativeabundance of the benthic species L. opesta and higher numbers of broken valves suggest heavyprecipitation events during the LGM (w23.7, 21.7, 20.8 and 20.1 ka). HS1 was predominantly dry, butwe identified times when lake levels were slightly higher, at the onset of the deglacial and a briefperiod (w17e16 ka) between HS1b and HS1a. All studied climate proxies indicate wetter and warmerconditions and lake system stability during the BA. Lake levels dropped during the YD and graduallyincreased during the Preboreal and early Holocene. We demonstrate that modern and fossil

rsidad Nacional Autónoma de México (UNAM), Ciudad Universitaria, 04510 Distrito Federal, Mexico. Tel.: þ52 55 5622

rez).

All rights reserved.

t al., Stable isotope values (d18O & d13C) of multiple ostracode species in a large Neotropical lake asternary Science Reviews (2012), http://dx.doi.org/10.1016/j.quascirev.2012.10.044

Delta:1_given name

Delta:1_given name

Delta:1_surname

Delta:1_given name

Delta:1_given name

Delta:1_surname

Delta:1_given name

Delta:1_given name

Delta:1_surname

Delta:1_given name

Delta:1_given name

Delta:1_surname

Delta:1_given name

Delta:1_given name

mailto:[email protected]

mailto:imprint_logo

www.sciencedirect.com/science/journal/02773791

http://www.elsevier.com/locate/quascirev

mailto:journal_logo

http://dx.doi.org/10.1016/j.quascirev.2012.10.044



L. Pérez et al. / Quaternary Science Reviews xxx (2012) 1e162

Please cite this article in press as: Pérez, L., eindicators of past changes in hydrology, Qu

ostracode isotopic signatures, species assemblages and taphonomy can be used together withphysical and geochemical variables in Lago Petén Itzá sediments to make high-resolution inferencesabout late Pleistoceneeearly Holocene environmental changes in the lowland Neotropics.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Stable isotope analysis of lacustrine ostracode shells is usedroutinely in paleoenvironmental studies (Lister, 1988; Hodell et al.,1991; von Grafenstein et al., 1992; Holmes, 1996; Schwalb, 2003).There are remote regions, however, where even the modernostracode fauna remains largely unstudied. For instance, the firstdetailed taxonomic investigation of the ostracode fauna from 63aquatic ecosystems on the Yucatán Peninsula (Guatemala, Belizeand Mexico) was undertaken in 2005 (Pérez et al., 2011). Never-theless, previous paleolimnological studies in some waterbodiesfrom the region used the oxygen isotopic signature of ostracodevalves to infer past shifts in the ratio of evaporation to precipitation(E/P) (Hodell et al., 1995; Curtis et al., 1996, 1998; Escobar et al.,2012). A firm understanding of the ecology of the local ostracodefauna is necessary for interpreting stratigraphies generated fromstable isotope analyses of the calcite valves.

The first studies to use isotopic measurements on lacustrineshell material from the Yucatán Peninsula were carried out byCovich and Stuiver (1974) at Lake Chichancanab, Mexico. Theymeasured d18O in gastropod shells. Among the records reportedfrom the Yucatan Peninsula were those of Hodell et al. (1995),Curtis et al. (1996, 1998) and Rosenmeier et al. (2002). Hodell et al.(1995) illustrated the great potential of ostracodes as indicators ofHolocene climate change, using oxygen isotope signatures in Cypriaophthalmica (Jurine,1820) and Cyprinotus cf. salinus (Brady,1868) insediment cores from Lake Chichancanab.

Recent advances in lake drilling technologies (Dean, 2010)enabled recovery in 2006 of high-quality, long sediment cores fromthe deepest (w160 m) and possibly the oldest (w200 ka, Muelleret al., 2010) lake in the northern lowland Neotropics, Lago PeténItzá, Guatemala. This sediment record is probably unique in theregion. Initial paleoenvironmental inferences used sedimentology(Mueller et al., 2010), geochemistry (Hodell et al., 2008), relativeabundances of ostracode taxa with known ecological preferences,ostracode-based water depth and conductivity transfer functions(Pérez et al., 2011), and the stable isotope signature of shells fromthe ostracode Limnocythere opesta (Escobar et al., 2012). The latterstudy focused on the paleoclimate of northern Guatemala ratherthan past lake hydrology per se. Few studies in the Neotropics haveused modern lake hydrodynamics and related environmentalvariables, i.e. calibration, to interpret fossil stable isotope records.Escobar et al. (2012) interpreted the d18O record as a response tofluctuating E/P and temperature. Interpretation of shifts in carbonisotope values is more complicated. The authors used fluctuationsin the d13C of ostracode shells as an indicator of the position of thethermocline, and speculated that such variations are controlled bythe nature of organic carbon delivered to the sediment. Elsewhere,investigators have used carbon stable isotopes in ostracode shells toinfer past catchment vegetation, modes of organic decay, origin ofCO2 and lake productivity (Schwalb, 2003; Bright et al., 2006). d13Cof ostracode shells in Lago Petén Itzá might serve as an indicator ofproductivity and/or sources of carbon in the sediment. There isgreat potential for using other, unstudied proxies for paleoenvir-onment in this lake, such as chironomids and diatoms. Otherostracode attributes (taphonomy), used in conjunction with stableisotopes, can provide important information for better interpreta-tion of the isotopic record in carbonate fossils. For instance,

t al., Stable isotope values (d1

aternary Science Reviews (20

ostracode taphonomy and valve preservation shed light on theenergy of the environment and sediment conditions, and can beused to reconstruct past lake levels or determine past water-column circulation patterns (Park et al., 2003; Park and Cohen,2011).

Paleolimnological studies carried out in the northern Neotropicsgenerally measured stable isotopes in a single ostracode taxon toavoid vital effects. When this was not possible, several species fromdifferent sediment depths, and in some cases parallel cores, wereused to obtain a continuous record. Vital effects should be evalu-ated by comparing isotope values frommultiple species collected atthe same sediment depth (this study). Recent studies highlightedthe influence of species autecology and environmental factors onostracode stable isotope values (von Grafenstein et al., 1999;Decrouy et al., 2011; Van der Meeren et al., 2011). Such informationfrom remote areas is often missing. Pérez et al. (2010a, 2010b)provided the first autecological information for modern ostracodespecies of Lago Petén Itzá, the lowlands and highlands of Guatemala(2012) and the Yucatán Peninsula and surrounding areas (2011).Modern stable isotope analyses of ostracodes are important forcalibrating paleoclimatic and paleolimnological reconstructions,however, such analyses were still lacking prior to this study.

In summary, few studies have usedmodern lake hydrodynamicsand ostracode species autecology and stable isotopes in paleo-environmental and paleoclimate reconstructions. This study aimsto improve the knowledge of ostracode species autecology, theisotopic composition of modern and fossil ostracodes and present-day and late Pleistoceneeearly Holocene lake hydrodynamics ofLago Petén Itzá. Here we used 1) autecological information for eightextant ostracode species in Lago Petén Itzá, 2) oxygen and carbonisotope values for modern ostracode species collected across a NeSwater depth transect (0e160 m), 3) isotope composition of lakewaters and related environmental variables determined duringsummer and winter, 4) isotope values in fossil ostracode assem-blages (Last Glacial Maximum [LGM]eearly Holocene) of LagoPetén Itzá, 5) complementary ostracode information (taphonomy,inferred water depth), as well as geochemical analysis, to improveprevious paleoenvironmental reconstructions and to infer past lakehydrodynamics during the late Pleistocene to early Holocene.

2. Study area

Lago Petén Itzá is located in the Central Petén Lake District,northern Guatemala, on the southern Yucatán Peninsula (Fig. 1).Nearby waterbodies include Lakes Salpetén, Quexil, Sacnab, Yaxháand Perdida, among others (Brenner et al., 2002). It is the deepest(zmax > 160 m) and one of the largest lakes (w112 km2) in the karstlowland Neotropics (w110 m asl). The lake is oligo-mesotrophicand population growth around the lake is changing its trophicstate, at least in some locations (Rosenmeier et al., 2004; Pérezet al., 2010c). Surface waters of the large north basin, however,display a total phosphorus concentration of only 9 mg L�1 andtransparency is relatively high (Secchi disk¼ 7.5 m). Lago Petén Itzáwaters are dominated by calcium, magnesium, bicarbonate andsulfate (Pérez et al., 2010b).

The thermocline is located between 20 and 40 m during thewarm period of the year, but appears to nearly breakdown duringthe colder winter months. Average monthly air temperature ranges

8O & d13C) of multiple ostracode species in a large Neotropical lake as12), http://dx.doi.org/10.1016/j.quascirev.2012.10.044

Fig. 1. Study area. A. Map of Mexico (MEX) and Central America showing the locationof Lago Petén Itzá, northern Guatemala (GUA). B. Bathymetric map of Lago Petén Itzáshowing drill site PI-6 and the NeS water depth transect where the long core andsurface sediments were collected for ostracode analyses, respectively. Most importantlimnological variables of Lago Petén Itzá are listed below.

L. Pérez et al. / Quaternary Science Reviews xxx (2012) 1e16 3

from 23.8 to 27.7 �C and the annual average relative humidity is67.3% (Pérez et al., 2010b). Average annual precipitation in Petén isw1665 mm. Two main seasons characterize northern Guatemala,a rainy season from June to October and a dry season from Januaryto May. Modest rainfall occurs during November and December.Rainfall in the region is mainly controlled by the Azores-Bermudahigh-pressure system and the seasonal migration of the Inter-tropical Convergence Zone (ITCZ). The ITCZ is found farthestnorthward in the summer rainy season and farthest southwardduring the dry season, when the Azores-Bermuda high in the Gulfof Mexico and the Caribbean dominate (Hastenrath, 1991).

Lago Petén Itzá is an appropriate water body for study of lakehydrodynamics because it lacks surface inflows or outflows,making it sensitive to changes in the ratio of evaporation toprecipitation. Such closed-basin lakes respond faster to changes inE/P than do open systems. In years 1938, the early 1990s and 2008the lake experienced high stages and flooding that destroyedlakeside infrastructure and forced some residents to abandon theirhomes (Pérez et al., 2010c). Rice (1997) reported a lake level riseafter 1976 that was related to increased annual rainfall.

3. Materials and methods

3.1. Field methods

Surface sediment samples (n ¼ 27) for ostracode analysis(species assemblages and stable isotopes) were collected along

Please cite this article in press as: Pérez, L., et al., Stable isotope values (d1

indicators of past changes in hydrology, Quaternary Science Reviews (20

a NeS transect in Lago Petén Itzá in November 2005, and Februaryand March 2008, extending from the littoral zone to a maximumwater depth of w160 m (Fig. 1). Environmental variables temper-ature, pH, conductivity, dissolved oxygen and alkalinity as well asthewater chemical (SO4, Cl, Ca, Mg, Na, K) and isotopic composition(d18O, d13C) were determined throughout the water column abovethe lake’s deepest point. Further information on field methods isfound in Pérez et al. (2010b, 2011). Long core PI-6, taken in w71 mof water (Fig. 1), was retrieved in 2006 by the Petén Itzá ScientificDrilling Project (PISDP). For further information regarding corerecovery, see Mueller et al. (2010) and Pérez et al. (2011).

3.2. Laboratory methods

3.2.1. Modern ostracodesOstracodes were extracted from 50-ml, wet surface sediment

samples by sieving samples with a 63-mm mesh and using a LeicaMZ 7.5 stereo microscope and fine brushes. At least 100 ostracodevalves were removed from each sample, including adults andjuveniles as well as ostracodes with well preserved soft parts(living) and specimens without soft parts (dead). Ostracode cara-paces with soft parts were used for identification to species level. Ifvalves of living ostracodes were found in abundance, only suchvalves were used for isotope analysis. If such valves were scarce, butpresent, valves of both living and dead ostracodes were analyzed.Ostracode identification followed Furtos (1933, 1936a, 1936b),Brehm (1939), Ferguson et al. (1964), Keyser (1976, 1977) and Pérezet al. (2010a, 2012). Autecological information for ostracode speciesin Lago Petén Itzá was taken from Pérez et al. (2010a, 2010b).

3.2.2. Fossil ostracodes (LGMeearly Holocene)Samples of 1-cm thickness were taken for fossil ostracode

analysis primarily at 20-cm intervals from LGMeearly Holocenesediments in long sediment core PI-6. LGM sediments containedlower numbers of ostracode valves and counts had to be made on5 g dry sediment aliquots. Long-core sediment samples were wet-sieved using a 63-mm mesh. Ostracodes were extracted, identifiedand enumerated with respect to numbers per g dry sediment. Adultand juvenile intact and broken valves were differentiated. Brokenvalves were counted if >50% was encountered and when identifi-cation was still possible. Details of fossil ostracode analysis can befound in Pérez et al. (2011).

3.2.3. Measurement of stable isotopes in lake watersWater samples for stable isotope analysis were preserved in the

field with CuSO4 (500 mg L�1) and analyzed in the Department ofGeological Sciences, University of Florida. Samples were analyzedusing a VG/Micromass PRISM Series II isotope ratio mass spec-trometer with a multi-prep preparation device (oxygen isotopes ofwater) and a Finningan-MAT DeltaPlus XL isotope ratio massspectrometer with a GasBench II universal on-line gas preparationdevice (carbon isotopes of DIC).

3.2.4. Measurement of stable isotopes in ostracode carbonatesOnly modern and fossil ostracode valves without evidence of

alteration, i.e. dissolution or development of calcite overgrowth,were selected for oxygen and carbon isotope analysis. Thiswas doneto avoid diagenetic effects on measured values (Holmes and Chivas,2002). All ostracode valves for isotope analysis were 1) soaked in15% H2O2 for 15 min to remove organic material, 2) rinsed withdistilled water and then with methanol and 3) dried in an oven at50 �C overnight. If valves were not completely clean after theprocedure, fine brushes were used to remove remaining sedimentsor other impurities. The number of valves used for each measure-ment depended on the species. Average adult valveweights were as



follows: 1 mg for Darwinula stevensoni, 2.9 mg for Cypridopsis okee-chobei, 4.6 mg for Cytheridella ilosvayi, 19.5 mg for Heterocyprispunctata, 2.6mg for L. opesta,1.2mg for Physocypria globula, and2.5mgfor Pseudocandona sp. We used a total weight of 40e60 mg for eachsample from surface and long-core sediments. Stable isotopes weredetermined with a Kiel III carbonate preparation device attached toa Finnigan-MAT 252 isotope ratio mass spectrometer in theDepartment of Geological Sciences, University of Florida.

We collected surface sediments from 0 to 160 m across the NeSwater depth transect, but we only used measurements in valvescollected from each species water-depth range (tolerance) forinterpretation. Water depth ranges of species were taken fromPérez et al. (2010b) and ranged from 0 to 60 m. We used theequation of Kim and O’Neil (1997) to determine the oxygen isotopiccomposition of calcite precipitated under equilibrium conditions inlake waters.

3.2.5. Statistical analysisSimple correlation (Spearman’s correlation) was used to explore

correlations between environmental variables and stable isotopesin the water column profile at the lake’s deepest point. Data werefirst normalized (x � mean/stdev) and analyses were done usingthe program PAST, version 1.89 (Hammer et al., 2001).

Here we provide a water depth reconstruction for Lago PeténItzá that is an improvement on previously published inferences(Pérez et al., 2011). The newly inferred lake level record was cali-brated by including statistical analysis of ostracode counts insurface sediment samples from the NeS water depth transect inLago Petén Itzá as well as ostracode data from nearby lakes in theCentral Petén Lake District. Sediment samples from water depths>60 m were excluded in the development of a Weighted AveragePartial Least Squares (WA-PLS) transfer function because recentstudies showed that ostracode species in Lago Petén Itzá only live atwater depths to a maximum of 60 m (Pérez et al., 2010a, 2010b).Analysis was done using the software C2 (Juggins, 2003). Our newapproach prevents over-estimation of past water depth fromostracodes in core samples, a problem that characterized ourprevious inferences (Pérez et al., 2011).

4. Results

4.1. Water column temperature profiles

Fig. 2 displays vertical temperature profiles for years 2008, 2005and 2002 at the lake’s deepest point, and several measurementsfrom May to August 1980 at a station near San Andrés, a town onthe lake’s north shore. Measurements in the 1980s were deter-mined from just before the beginning, to the end of the rainyseason. The highest surface water temperatures (32 �C) weredetermined at the end ofMay and the lowest values weremeasured(27.7 �C) at the end of August 1980. Highest temperature difference(7.5 �C) between surface waters and waters below the thermoclinewere determined in May, at the end of the dry season. By June, withthe onset of the rainy season, surface water temperatures haddecreased by 3.5 �C and surface temperatures in July and Augustwere only 28 �C and 27.7 �C, respectively.

A vertical temperature profile from the deep-water site (160 m),determined in August 2002, displayed temperatures that rangedfrom 30 �C at the surface to 25.4 �C in the hypolimnion. Tempera-tures recorded in November 2005 and February 2008 were slightlycooler than temperatures reported for the rest of the year. TheNovember 2005 topebottom temperature profile ranged from 27.6to 25.8 �C and the February 2008 profile from 27.9 to 25.6 �C.Temperature differences between surface and deep waters weresmaller than those reported from May to August.



Aweak late fallewinter thermocline (NovembereFebruary) wasdetermined at the lake’s deepest point. During late springesummer(MayeAugust), however, the position of the thermocline at theshallower San Andrés station varied, with the top of the metal-imnion as shallow as 5 m and the bottom of the metalimnion asdeep as 30 m. Lake thermal stratification is more stable duringsummer months than during NovembereFebruary. In November2005, water at 20 m depth was slightly warmer than overlyingwaters, perhaps reflecting loss of heat from surface waters and theonset of thermocline breakdown.

4.2. Environmental variables and isotopic composition of lakewaters

Fig. 3 displays variables in a water column profile from LagoPetén Itzá, including dissolved oxygen concentration, dissolvedinorganic carbon (DIC), chlorophyll a and total phosphorus plottedwith the stable oxygen and carbon isotopes. Dissolved oxygenprofiles from November 2005 and February 2008 overlie oneanother. Values were near saturation in surface waters, declinedthrough the thermocline and displayed hypoxia in the lower half ofthewater column. d18OSMOW inwaterswere slightlymore positive inFebruary 2008 than in August 2002, but overall there was littlevertical or inter-annual variation, with all values between þ2.7and þ3.4&. d13CDIC values in February 2008 were between �5.1and �0.1& and the hypolimnion displayed more negative values.DIC values in February 2008 ranged from 1.3 to 8.7 mg L�1. Chloro-phyll a concentrations in summerwere consistently low throughoutthe vertical profile (�0.25 mg L�1), but displayed a small peak(1.17 mg L�1) at w42.8 m water depth, just below the base of thethermocline. The top 20 m displayed concentrations ranging fromonly 0.11 to 0.25 mg L�1. Total phosphorus ranged from6 to 21 mg L�1.The highest concentration in the epilimnionwas 11 mg L�1, whereasthe highest value in the hypolimnion was 21 mg L�1. In February2008, d18O and d13C were positively correlated with oxygenconcentration, temperature, pH and chlorophyll a (r values > 0.5)and the two isotopes were highly correlated with one another(r¼ 0.98, Table 1). Temperature, pH and dissolved oxygen displayedhigh correlations with one another as well (r � 0.88).

4.3. Ostracode autecology

Table 2 contains a list of the ostracode taxa that have inhabitedLago Petén Itzá since the late Pleistocene, along with autecologicalinformation for each taxon. Candonocypris serratomarginata (Furtos,1936), Cypretta sp. and Stenocypris major (Baird, 1859) wereexcluded because we lacked sufficient material for stable isotopeanalyses. Eight species were collected in modern surface sedimentsof Lago Petén Itzá, five of which were also found in sedimentsdeposited during the LGM and seven of which were also found indeposits of the deglacial to early Holocene.

4.4. Stable isotope values in modern ostracodes from the NeSwater-depth transect

Ostracodes found in Lago Petén Itzá from 0 to 60 mwater depthdisplayed d18O values from þ0.1 to þ3.5& during winter (Figs. 4and 5). Highest values were determined in shells of benthic Pseu-docandona sp. (þ1.3 to þ3.5&), the nektobenthic species C. okee-chobei Furtos, 1936 (þ1.2 to þ3.0&) and in shells of benthic D.stevensoni Brady & Robertson, 1870 (þ1.6 to þ1.9&). Limnocythereopesta, a benthic species, displayed the lowest d18O values at allsampled water depths, except 15 m. Oxygen isotope values ofL. opesta varied from þ0.1 to þ1.7&. Similar to L. opesta, C. ilosvayiDaday, 1905 displayed lower values than other ostracode species,


Fig. 2. Vertical temperature profile of Lago Petén Itzá during spring (red), summer (orange, yellow and green curves), fall (light blue) and winter (dark blue curve) determined at theSan Andrés station (Fig. 1, 1980 data) and at the lake’s deepest point. Winter data indicate a breakdown of the thermocline. Summer measures indicate the thermocline is locatedfrom about 5 to 30 m or between 20 and 40 m, depending on the site and perhaps year of measurement. Dashed lines indicate previously published data. Temperature data forAugust 2002 (green) are from Hillesheim et al. (2005), and data for November 2005 (light blue) and February 2008 (dark blue) are from Pérez et al. (2010b). (For interpretation ofthe references to color in this figure legend, the reader is referred to the web version of this article.)


with 18O values ranging fromþ0.4 toþ1.5&. The d18O values of thenektobenthic species P. globula Furtos, 1933 were slightly morepositive and varied from þ1.1 to þ1.9&. Heterocypris punctata,a littoral ostracode species, displayed a value of þ1.4&. Of allostracode taxa, L. opesta displayed d18O values closest to the equi-librium d18Ocalcite value, with an offset ranging fromþ0.03 toþ1.1&across the water depth transect. Pseudocandona sp. displayed thelargest offset, ranging from þ1.0 to þ2.9&. Offsets for the otherostracode species were as follows: C. ilosvayi (þ0.1 to þ1.0&),P. globula (þ0.7 to þ1.0&), Strandesia intrepida Furtos, 1936 (þ0.6toþ1.7&), C. okeechobei (þ0.9 to þ2.4&), H. punctata (þ1.0&), andD. stevensoni (þ1.2 to þ1.4&).

Ostracode d13C values across the NeS water depth transectranged from �6.3 to �1.0&, a broader range than found for d18O



values (Fig. 4). Nektobenthic species C. okeechobei, H. punctataand S. intrepida generally displayed the highest d13C values.Values in shells of C. okeechobei and S. intrepida ranged from �3.4to �1.0& and from �1.7 to �1.4&, respectively. Heterocyprispunctata displayed d13C values of �1.4& at a water depth of0.1 m. Physocypria globula, another nektobenthic species, but onethat prefers deeper waters, displayed more negative values (�5.4to �2.5&). Benthic species displayed more negative values thannektobenthic species. Carbon isotopic composition of C. ilosvayiranged from �5.2 to �2.1&, and those of D. stevensoni rangedfrom �4.6 to �3.4&. Benthic species L. opesta and Pseudocandonasp. displayed the most negative values among all species, withvalues ranging from �6.3 to �2.9& and from �6.1 to �2.2&,respectively.


Fig. 3. Main environmental variables of ostracode host waters determined during August 2002, November 2005 and February 2008 at Lago Petén Itzá’s deepest point. Dashed linesindicate previously published data. August data (green) are from Hillesheim et al. (2005), and November 2005 (light blue) and February 2008 (dark blue) data are from Pérez et al.(2010b). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)


Ostracode d13C values were compared with the d13CDIC of lakewaters (Fig. 4). Nektobenthic species displayed smaller differencesthan did benthic species (C. okeechobei [�3.3 to �0.02&],S. intrepida [�1.5 to �0.1&], H. punctata [�0.1&]), except forP. globula, which displayed offsets similar to those of benthicspecies (�5.3 to �1.2&). Benthic species L. opesta displayed thelargest difference (�6.2 to �2.7&), followed by Pseudocandona sp.(�6.1 to�1.0&), D. stevensoni (�5.4 to �2.1&) and C. ilosvayi (�5.0to �1.7&).

Fig. 5 shows a cross plot of the oxygen and carbon isotope valuesof ostracode valves collected across the NeSwater depth transect inLago Petén Itzá. Nektobenthic species are easily distinguished frombenthic species, except for P. globula. d13C values are more positivein nektobenthic than in benthic ostracodes (>3.5&). Limnocythereopesta displayed the lowest d18O and d13C values among all ostra-code species, whereas C. okeechobei displayed the highest values.

4.5. Fossil ostracode assemblages in sediment core PI-6

Cytheridella ilosvayi was the only species analyzed for stableisotopes that was absent in late Pleistocene sediments (Table 2).

Table 1Correlation coefficients between variables in awater column profile from Lago PeténItzá collected in February 2008. *indicates correlations significant at P < 0.05 and**indicates correlations significant at P < 0.01.

February 2008

Variable O2 Temp. pH d18O d13C DIC HCO3 Chl a

O2 1.00Temp. 0.90** 1.00pH 0.93** 0.88** 1.00d18O 0.74** 0.76** 0.68** 1.00d13C 0.75** 0.72** 0.69** 0.98** 1.00DIC �0.08 0.10 �0.09 0.424 0.70 1.00HCO3 �0.39 �0.43 �0.34 �0.31 �0.23 0.43 1.00Chl a 0.76** 0.69** 0.67** 0.56* 0.55* 0.04 0.31 1.00



Unit III in core PI-6 (LGM, 24e19 ka) was characterized by thepresence of benthic species D. stevensoni, L. opesta and Pseudo-candona sp., and nektobenthic species P. globula and C. okeechobei(Fig. 6). Nektobenthic species were more common during Unit II(deglacialeearly Holocene, w19e10 ka). They includedC. okeechobei, H. punctata, P. globula and S. intrepida. Threebenthic ostracode species, D. stevensoni, L. opesta and Pseudo-candona sp., were found in these sediments. Limnocythere opestaand P. globula dominated late Pleistocene species assemblages.Few or no valves of P. globula were found from w13 to11 cal ka BP (subunit I, Younger Dryas [YD]), whereas this specieswas very abundant (up to 6114 valves g�1) from w15 to13 cal ka BP (subunit II, Bølling-Allerød [BA]).

Total numbers of adult valves were lower during Unit III(�396 g�1) than during Unit II (�6479 g�1). Juveniles of P. globulaexceeded the number of adults, whereas the number of adultL. opesta was sometimes much greater than the number of juve-niles. A high number of broken adult and juvenile valves charac-terized Unit III. Higher numbers of valves were found in gypsum-rich deposits (subunits V, Heinrich Stadial b [HS1b], III HeinrichStadial a [HS1a] and I, YD, Fig. 6) and during subunit II (BA). Thehighest numbers of broken adult (�2646 valves g�1) and juvenile(�7409 valves g�1) valves during the deglacial were encountered at14.6, 13.6e13.4 and at 10.4e10.7 cal ka BP. For all ostracode species,numbers of juvenile valves exceeded adult valves. This wasgenerally the case for all sediment samples, except those markedwith an asterisk in Fig. 6.

4.6. Geochemical proxies and ostracode taphonomy

Fig. 7 shows the relative abundance of L. opesta, magneticsusceptibility, total sulfur (i.e. gypsum), C/N ratio andpercent brokenadult and juvenile ostracode valves. Fluctuations in relative abun-dance of L. opesta are inversely related to shifts in magnetic suscep-tibility. Relatively greater magnetic susceptibility (<80 SI*E�6) was


Table 2Overview and ecology of the ostracode fauna of Lago Petén Itzá, Guatemala since the late Pleistocene.

Species SEM Modernlakea

DeglacialeearlyHolocene

LGM Lifestyle

Depthrange (m)a

Autecologya

Cypridopsis okeechobei þ þ þ NB �40 Prefers littoral zones with rich vegetation.

Cytheridella ilosvayi þ � � B �40 Prefers warm shallow waters (>20 �C), tolerateshigh conductivities (up to 5960 mS cm�1) and sulfateconcentrations (2300 mg L�1).

Darwinula stevensoni þ þ þ B �15 Prefers shallow waters, slow currents, can toleratehigh salinities (13.5&).

Heterocypris punctata þ þ � NB �1 Prefers littoral zones with abundant macrophytes,calm water and tolerates salinities of up to 10&).

Limnocythere opesta þ þ þ B <40 Marks the thermocline, abundant in littoral zoneswith abundant macrophytes. Tolerates highconductivities and sulfate concentrations.

Physocypria globula þ þ þ NB <50e60 Displays high tolerance, prefers deep waters,can tolerate water with low dissolved oxygen(w3 mg L�1).

Pseudocandona sp. þ þ þ B <40 Prefers shallow water (<15 m), prefers sedimentswith high organic content.

Strandesia intrepida þ þ � NB <15 Prefers shallow, well-oxygenated warm waters.

Abbreviations: Nektobenthic (NB), Benthic (B).a Information from Pérez et al. (2010b, 2010c).


measured in unit III, subunit IV and II. Greater concentrations of totalsulfur (TS<20%) andC/N ratios (<40) characterize subunits V, III andI. Percentages of broken adult and juvenile valves were high (>66%and>39%, respectively) in unit III but lowduring subunit II and at the



end of unit II (<41% and<33%). Highest percentages of broken adultand juvenile valves were typical of subunits V, III and I. Highpercentages of broken ostracode valves characterized the beginningof subunit V and end of subunit III.


Fig. 4. Oxygen (right) and carbon (left) isotope signatures (& VPDB) of eight ostracode species inhabiting Lago Petén Itzá today. The left black line indicates the winter temperatureprofile determined at the time of ostracode collection. Right black line shows the d18O values of the theoretical calcite precipitated in equilibrium and the dashed black line displaysd13CDIC values of lake water. The dashed line indicates previously published data from Pérez et al. (2010b).


4.7. Water depth transfer function

Fig. 7 displays the inferred past water depth in Lago PeténItzá. Here we show the average water depths. The water depthreconstruction suggests that lake stage at the core site duringunit III averaged w26 m, and during subunits V, IV, III, II, I ofUnit II, lake level averaged 24, 31, 10, 40, 4 and 58 m, respec-tively. Highest lake levels were inferred for subunits II and I, andlowest lake levels occurred in subunits V, III and I (Fig. 7). Theimproved transfer function, using the WA-PLS model, has an r2

of 0.74 and a RMSE of only 9 m after excluding ostracodesamples from waters deeper than 60 m, where ostracodes arescarce or absent.

4.8. Late Pleistocene ostracode stable isotope record

Fig. 8 displays the late Pleistoceneeearly Holocene oxygen andcarbon isotope records developed using multiple ostracode taxa.Oxygen and carbon isotope values uncorrected for vital effectfluctuated between þ3.1 and þ7.3& and from �8.1 to þ3.7&,respectively. Highest d18O and d13C values were determined invalves of H. punctata. Nektobenthic species H. punctata andC. okeechobei (þ2.6&) displayed higher d13C values than otherspecies. Generally, L. opesta and P. globula showed more positived18O values (�þ7.2& and þ6.9&, respectively) than other



ostracode species. In deposits corresponding to the LGM, wefound sufficient ostracode valves for measurement in only twosediment samples. The d18O values ranged from þ5.12 to þ5.3&and d13C values from�7.1 to�6.4&. d18O values display variabilitybetween w19 and 15 ka BP. HS1b and HS1a are characterized byhigh d18O values, up to þ6.9& (P. globula) and þ7.3&(H. punctata), respectively. Between HS1b and HS1a, valuesdecrease to þ4.9& (P. globula), however slightly higher d18Ovalues of þ6.4& (L. opesta) were determined at w16.6 cal ka BP.d13C from early HS1b to late HS1a shows an upward trend,from �8.1 to þ3.7&. Highest d13C values in HS1b and HS1a werereported forH. punctata. The BA is characterized by lower d18O andd13C values in L. opesta (þ3.8& and �5.9&) and P. globula (þ3.8&and �5.7&). Stable isotope values increase after the BA and d18Ovalues in L. opesta range from þ5.2 to þ5.7& and d13C from �1.9to �0.2& during the YD. At w11.7 cal ka BP values start todecrease.

Fig. 9 displays d18O and d13C values for valves of multipleostracode species in long core PI-6. Isotope values for periods ofhigh and low lake level, as well as for periods of wetter and drierconditions during the late Pleistoceneeearly Holocene are dis-played. The LGM, early deglacial and early Holocene were char-acterized by d13C values between �8.0 and �6.3& and d18Ovalues between þ3.1 and þ6.1&. Relatively high d18O values,from þ5.1 to þ5.3&, are typical of valves from the LGM, whereas


Fig. 5. Cross plot of d18O and d13C values in modern ostracode valves from eightnektobenthic and benthic ostracodes species living in Lago Petén Itzá. Nektobenthicspecies, except for Physocypria globula, generally displayed higher stable isotope values(oxygen and carbon) than benthic species.


d18O values from þ3.10 to þ3.6& characterized valves depositedin the early Holocene. Valves from the BA displayed less negatived13C values (�5.9 to �3.4&) than those from the LGM, earlydeglacial and early Holocene. In general, greater d18O (�þ5.0&)and d13C (��3.0&) values were measured when lake levels werelow.

Fig. 6. Late Pleistocene to early Holocene (24e10 ka) ostracode species assemblages in Lagoconditions (red) characterized the BA. Lithological units are in Mueller et al. (2010) and can(gypsum sands) and subunits IV, II (carbonate-rich clays). Ostracode counts are presented (i(TAV, black), total adult broken valves (TABV, white), total juvenile valves (TJV, black) and toexceeded the number of juvenile valves as an indicator of an environment of high energy levor other sites. Arrows indicate when numbers of broken adult and juvenile valves were almosOther abbreviations: Last Glacial Maximum (LGM), Heinrich Stadial 1 (HS1), Bølling Allerødlegend, the reader is referred to the web version of this article.)



5. Discussion

5.1. Modern lake hydrodynamics

Paleolimnological reconstructions are more robust whenmodern lake dynamics are well understood. Here we presentlimnological variables in Lago Petén Itzá, which provide informa-tion about the position of the thermocline during the rainy(summer) and dry (winter) seasons, and report data on dissolvedoxygen, d18O, d13C, dissolved inorganic carbon, chlorophyll a andtotal phosphorus, which can be used as indicators of lake produc-tivity and the balance between precipitation and evaporation.

Lake surface waters reach a maximum temperature (32 �C) inMay, but temperatures begin to decrease with the onset of summerrains (Fig. 2). Coldest surface water temperature (27.7 �C) was re-ported in August 1980. Lake water temperatures, however, candiffer year-to-year, because of changes in solar radiation (e.g. cloudcover), duration of the rainy season, and climate change (Cohen,2003). For instance, surface temperature in August 2002 was 30 �C.

Temperature measurements in summer 2002 (August) indi-cated the thermocline was between 20 and 40 m. Thermal datacollected in 1980 from a station near the town of San Andrésindicate the summer thermocline position was shallower, between5 and w30 m water depth. Temperature profiles in summermonths, especially August, show a well-defined epilimnion andhypolimnion. Whereas the change in temperature across thethermocline in August 2002 was not great (<5 �C), the densitydifference at such high temperatures is evidently sufficient toimpart stable thermal stratification. Thewinter profile, on the otherhand, shows only about 2 �C difference between surface and deepwaters. This small temperature difference, coupled with strongnortes, strong cold northeasterly winds along the Gulf of Mexico,can lead to winter mixing. This probably explains the lowertemperatures from 10 to 20 m in the November vertical profile(Fig. 2). With the onset of winter, surface waters begin to cool,particularly at night, and because they possess greater density than

Petén Itzá. Cold conditions (blue) dominated the LGM, HS1, and the YD while warmerbe summarized as follows: Unit III (montmorillonite clays), Unit II: subunit V, III and In black, adult valves g�1 and in white, juvenile valves g�1) as follows: total adult valvestal juvenile broken valves (TJBV, white). Asterisks indicate the times when adult valvesel when juvenile ostracodes might have been transported by currents to deeper waterst as high as the total number of ostracode valves, indicating possible mixing of the lake.(BA), Younger Dryas (YD). (For interpretation of the references to color in this figure


Fig. 7. Ostracode-inferred water depth and modern lake level (vertical line) at core site, relative abundance of Limnocythere opesta (LOP rel Ab.), geochemical proxies and ostracodetaphonomic proxies (adult broken valves (ABV), and juvenile broken valves (JBV)) during the late Pleistoceneeearly Holocene in Lago Petén Itzá, Guatemala. Lithological units are inMueller et al. (2010). Magnetic susceptibility reflects lithologic changes in long core PI-6 during the LGM and deglacialeearly Holocene. High total sulfur (TS) and C/N ratios indicatelow lake levels and dry climate conditions. White bars indicate sections where no ostracodes were found. Cold conditions are indicated with blue and warmer conditions with red.The dashed lines indicate previously published data. Magnetic susceptibility data are from Mueller et al. (2010) and the abundance of L. opesta, TS concentrations and C/N ratios arefrom Pérez et al. (2010b). Abbreviations: Last Glacial Maximum (LGM), Heinrich Stadial 1 (HS1), Bølling Allerød (BA), Younger Dryas (YD). (For interpretation of the references tocolor in this figure legend, the reader is referred to the web version of this article.)


the underlying strata, they begin to stream downward. On the daywe sampled, there had probably been some warming and windmixing of the uppermost water column, but the profile certainlyreflects the breakdown of thermal stratification.

In November 2005, oxygen concentrations remained constantthroughout the top 10 m (8.9 mg L�1), and reflect slight supersat-uration, despite low chlorophyll a concentrations. d13CDIC values in

Fig. 8. Late Pleistoceneeearly Holocene stable isotope record of multiple ostracode species.globula, Cypridopsis okeechobei, Heterocypris punctata. Lower values characterized periods wdrier, lake levels were lower in Lago Petén Itzá, and d18O and d13C were higher. Cold conditstable isotopes determined at a higher resolution by Escobar et al. (2012). Abbreviations: Las(YD). (For interpretation of the references to color in this figure legend, the reader is referr



the epilimnionwere slightly more positive than in the hypolimnionin February 2008. Basterrechea (1988) defined the circulationpattern of Lago Petén Itzá as warm monomictic, and indicated thatweak mixing takes place during winter. High-temporal-resolutionsampling of the Lago Petén Itzá water column should be carriedout over several years to determine if circulation is complete andexactly when it occurs.

Lines with symbols show the d18O and d13C values of Limnocythere opesta, Physocypriaith high lake levels such as the LGM, BA and the early Holocene. The HS1 and YD wereions are indicated with blue and warmer conditions with red. Blue dashed line showst Glacial Maximum (LGM), Heinrich Stadial 1 (HS1), Bølling-Allerød (BA), Younger Dryased to the web version of this article.)


Fig. 9. Cross plot between late Pleistocene d18O and d13C values (symbols) of fossil valves from multiple ostracode species. Colors indicate the values that correspond to the LGM(gray), lithologic subunits I, III, V (yellow) and subunits IIeIV (brown) and the Holocene (green) (Mueller et al., 2010). Lowest d18O and d13C values indicate times when lake levelswere higher (wet) and climate warmer (e.g. Bølling-Allerød and the early Holocene) and when lake levels were higher and cold (LGM and early deglacial). Higher stable isotopevalues correspond to periods with fluctuating (intermediate-low) lake levels and wetedry conditions during the deglacial. Numbers indicate the age (cal ka BP). (For interpretationof the references to color in this figure legend, the reader is referred to the web version of this article.)


We only have measurements of lake water d13CDIC for February2008, but we suspect that the difference between values in winterand summer are much larger than for d18O values (wþ0.4&),which should be taken into account when interpreting the fossilostracode isotope record. It should be easy to identify fluctuationsin the fossil d13CDIC profile of ostracodes. d18O values mainly indi-cate wet and dry periods (high and low lake levels), whereas d13Cvalues provide information on primary productivity and carbonsources to the sediment. Low total phosphorus concentration inLago Petén Itzá suggests low productivity in the lake. Higher TPvalues in deeper waters probably reflect the presence of sestondescending through the water column, but might also reflectpresence of zooplankton or non-photosynthetic bacteria at greaterdepths. Chlorophyll a was measured only in August 2002.Concentrations are low in the epilimnion, near zero in the metal-imnion, but display a small peak at the metalimneticehypolimneticboundary, at w43 m.

It was necessary to understand how environmental variablescorrelate with oxygen and carbon stable isotopes in modern lakewaters of Lago Petén Itzá to infer paleohydrology and lake levelchanges from stable isotope signatures recorded in multipleostracode species in the lake (Table 1). This was done to discernwhich limnological variables should be taken into account wheninterpreting the oxygen and carbon stable isotopes in fossil ostra-code shells. February 2008 data suggest that d18O and d13C in thewater column are positively correlated with dissolved oxygen,temperature and pH. Variations in the d18O values in the epilimnionmight be explained by changes in E/P. Higher evaporation couldexplain more positive d18O values measured in February 2008. The



weighted mean oxygen isotopic composition of precipitation in theregion is w�4.0& (Hillesheim et al., 2005), which could explainlower d18O values in lake surface water during the rainy season. Thefact that d13C values are positively correlated with dissolved oxygenand pH in February probably reflects several processes: 1) oxygenconcentration is higher in the epilimnion as a consequence ofprimary production, but declines in deeper waters of the troph-olytic zone, 2) pH is higher in upper waters as a consequence of CO2withdrawal for photosynthesis, but is greater in bottom waters,where intense respiration prevails, and 3) DIC of bottom waters isdominated by light carbon from microbial breakdown of organicmatter with relatively negative d13C values. d13C values, concen-tration of dissolved oxygen and pH in upper waters of Lago PeténItzá were slightly higher than in hypolimnetic waters, reflectingproductivity in the epilimnion.

5.2. Modern isotopic composition and autecology of ostracodespecies

Clear differences between the oxygen and carbon isotopiccomposition of the studied ostracode species in Lago Petén Itzá canbe seen in Fig. 5. Stable isotopeswere determinedmainly in valves ofliving ostracodes, e.g. ostracodes with well-preserved soft parts.Such isotope values reflect the isotopic compositionof the lakewaterwhen they were living. Nektobenthic species display more positived13C and d18O values than do benthic species, at least during winter.The benthic species L. opestadisplayed themost negative d18O valuesand the lowest offset from expected equilibrium values amongostracode species. This might be expected for a benthic species that



lives on the bottom of the lake, unaffected by evaporation andtemperature shifts in the epilimnion. We could not compare ourLimnocythere datawith data from otherwaterbodies in the northernNeotropics because such information is lacking, analyses were runon ostracode valves of other species, or multiple species were notanalyzed in other studies. von Grafenstein et al. (1999) studied thestable isotopic values of multiple ostracode species in LakesAmmersee and Starnberger See, southern Germany. They found thatLimnocythere inopinata and D. stevensoniwere the ostracode specieswith the smallest offset (w0.8&) from equilibrium d18Ocalcite,whereas species of the Candona group displayed higher offsets(�2&). L. opesta was the ostracode species in our study with thelargest offset relative to d13CDIC. The average d13C offset was,however, much larger (�4.0&) than that for d18O (þ0.3&).

Similar to L. opesta, C. ilosvayi is a benthic species and displaysonly slightly more positive d18O values, reflecting the fact that bothspecies live at similar water depths (�40m) and temperatures. d13Cvalues of L. opesta, however, are much more negative than those ofC. ilosvayi. This might suggest that L. opesta is an infaunal speciesthat builds its shell from relatively negative DIC from the interstitialwaters. The more positive values of C. ilosvayi and D. stevensonisuggest they may live above the sediment surface (epifaunal). Inaddition to the influence that themicrohabitat has on d18O values ofostracode valves, vital effects (metabolic differences) may play animportant role as well.

Physocypria globula was the only species found living down toa water depth of 60 m. d18O values of P. globula fluctuate little alongthe NeS water depth transect (þ1.1 toþ1.9&), whereas d13C variedfrom �5.4 to �2.5&, probably suggesting that this species moultsin a specific water depth or water depth range. The verticald18OSMOW profile at the deepest point of Lago Petén Itzá varied littlecompared to the d13CDIC profile. Physocypria globula is a nekto-benthic species that lives in the littoral zone and in the watercolumn. Previous studies revealed that this species prefers deeperwaters, w50e60 m (Pérez et al., 2011), because it tolerates waterswith low concentrations of dissolved oxygen (Curry and Filippelli,2010). Its broad tolerance with respect to a combination of envi-ronmental variables explains why it displays more positive d13Cvalues in shallower water where productivity is high and lakewaters are 13C-enriched, and why it displays more negative valuesin deeper waters, where there is decay of organic matter.

The nektobenthic species H. punctata was found only in waters<1 m deep. This suggests that it is restricted to littoral zones. Itmight therefore be expected to display more positive d18O becauseshallow waters are more affected by evaporation. However, thislittoral nektobenthic species displayed d18O values slightly lower(þ1.4&) than the benthic species Pseudocandona sp. (þ1.5&).These results are similar to those of Bridgwater et al. (1999) for LakePátzcuaro, central México. That study reported lower d18O valuesfor H. punctata than for the benthic species Candona patzcuaro.Bridgwater et al. explained this difference as a consequence of thespecies occupying different micro-habitats. They suggested thatreduced mixing of the water column restricted transport of fresh-water with low d18O ratio, into the bottom waters. Warming of theshallow littoral zone down to the surface sediments and the factthat species belonging to Candonidae generally prefer the phytalzone (Meisch, 2000) might account for the small differencebetween d18O values of H. punctata and Pseudocandona sp., Cypri-dopsis okeechobei, S. intrepida and Pseudocandona sp. displayed thehighest d18O values among ostracode species (�þ3.5&) in LagoPetén Itzá. Cypridopsis okeechobei and S. intrepida are nektobenthicspecies that prefer littoral zones with abundant macrophyteswhere productivity is high, temperatures are warm and evapora-tion is high (Pérez et al., 2010b), and this could well explain theirvery positive d18O values, similar to H. punctata.



5.3. Late Pleistoceneeearly Holocene hydrodynamics and climaticconditions

Lago Petén Itzá displayed dramatic fluctuations in water level,hydrodynamics and climate conditions during the latePleistoceneeearly Holocene. Such environmental and climaticfluctuations can be summarized as follows:

5.3.1. High lake levels (w26 m) and cold conditions (LGM,w24e19 ka)

The LGM was characterized by low numbers of adult specimens(�396 valves g�1) and by dominance of benthic species over nek-tobenthic species (Fig. 6). This suggests that species such asD. stevensoni, L. opesta, Pseudocandona sp. and P. globula, toleratedslightly lower water temperatures. Using pollen, Bush et al. (2009)suggested a cooling of 3e5 �C relative to present. Because thetemperature difference between the dry (winter) and wet(summer) season in the northern Neotropics is small, lake watertemperatures for Petén Itzá during the LGM have not been inferred.Distributions of ostracodes and other bioindicators (cladocerans,diatoms and chironomids) of the Yucatán Peninsula are determinedby conductivity and HCO3 rather than temperature, one reasonwhytransfer functions for temperature have yet to be developed (Pérezet al., in press). Limnocythere opesta and P. globula and C. okeechobei,however, seem to prefer warmer temperatures, in contrast toPseudocandona sp. and D. stevensoni, which were rare during thedeglacialeearly Holocene, when temperatures increased. Thiswould suggest that these latter taxa could be winter species, i.e.reaching adulthood during wet winter months. We found few adultspecimens of these species, suggestingminimal winter rains duringthe sampling months.

The water depth reconstruction based on ostracodes is consis-tent with geochemical proxies and lithology. It is important tocompare this water depth reconstruction with inferences fromother proxies because ostracode distribution is not directly affectedby water depth. It is, instead, the combination of factors at a givenwater depth that affect ostracode distribution. Those factorsinclude: macrophyte distribution, suspended load of streams, dustinflux, nutrients, water transparency, waves, among others. Ourinferred water depth suggests that during the LGM lake levelaveraged w26 m (Fig. 7). Ostracode taphonomic analysis alsosuggests that there were strong currents and lake mixing duringthe LGM, which accounted for high numbers of broken adult andjuvenile valves and times when adult ostracode valves exceededjuvenile valves (Figs. 6 and 7). In the absence of transport, juvenilevalves should exceed adult valves, because each individual moultsabout eight times to reach adulthood (Meisch, 2000). Coldertemperatures and a smaller difference between surface and lakebottom water temperatures, might have led to an unstratified orweakly stratified lake, and lake circulation. There are, however,other factors that might have accounted for the high number ofbroken valves, such as sediment redeposition, shifting wave base,an oscillating thermocline level, periods of flash floods or down-slope transport of shells during storm events. LGM sediments arecharacterized by the presence of clay with interbedded turbidites,suggesting individual heavy rain events (Mueller et al., 2010) anddown-slope transport. White bars in Fig. 7 indicate times whenadult ostracodes were absent and times when few or no juvenileswere present. All ostracode species that lived during the LGMdisplay sporadic increases in the number of adult valves. Theseresults indicate movement by currents and down-slope transport.d18O values were relatively low (þ5.1& toþ5.3&), but not as low asduring the BA and early Holocene (�þ3.4&). Escobar et al. (2012),however, determined stable isotopes at a higher resolution andreported d18O values fromþ4.7 toþ6.0& and d13C values from�8.5



to �4.6& (Fig. 8). This suggests colder temperatures during theLGM and lower precipitation than during the BA and early Holo-cene, when lake level averaged w40 m. d13C values were morenegative (�7.1& to �6.4&) than values during the BA and earlyHolocene. The low values during the LGM and BA and early Holo-cene indicate high lake levels. The slightlymore positive d13C valuesduring the BA and early Holocene suggest higher temperatures andtherefore more productivity. For instance, Hodell et al. (2012)inferred a water temperature of �24 �C and �26 �C, respectively.The temperature inference was made by measuring the oxygenstable isotopic composition in shells of L. opesta. Alternatively thelow carbon isotope values may reflect the fact that L. opesta lived in13C-depleted sediments, suggesting that it might be an infaunalspecies. Porewaters can be depleted or enriched owing to oxidationof sediment organic matter or methanogenesis. Escobar et al.(2012) indicated that methanogenesis in Lake Petén Itzá is pre-vented by abundant dissolved sulfate in the water column andporewaters.

The water depth transfer function derived using ostracodessuggests maximum late Pleistocene lake levels during the LGM,about 20.6 cal ka BP, when d18O and d13C values show a smalldecrease. Lower lake levels were inferred at 23.7, 21.7, 20.8 and20.1 cal ka BP (Fig. 7). Oxygen, and especially carbon stable isotopevalues measured in shells of L. opesta are slightly higher duringthese short periods (Fig. 8). This indicates that even during theLGM, lake level in Lago Petén Itzá fluctuated, but not as much asduring the deglacial. Another possibility is that ostracode valveswere transported to deeper waters by heavy precipitation events,suggested by the presence of several turbidites in LGM sediments.Because application of water depth transfer functions can beconfounded by changes in lake productivity and water chemistry,we compared our inferences of past water depth with geochemicaland taphonomic proxies and obtained consistent results. Fluctua-tions in the relative abundance of the ostracode L. opesta generallycorrelate with fluctuations in magnetic susceptibility (Fig. 7).Although the two analyses were carried out at different samplingresolutions, slightly lower values of magnetic susceptibility duringthe LGM coincide with peaks in the relative abundance of L. opesta.Such increases in the abundance of L. opesta might suggest thatvalves were transported from shallower to deeper waters. Previousstudies by Pérez et al. (2010b) indicate that L. opesta is an indicatorof shallow waters, because higher numbers of living specimenswere collected at water depths <20 m. Percentages of broken adultvalves were greater during periods of high lake levels in the LGM. Atseveral times, the number of broken adult and juvenile valves wasalmost as high as the total number of ostracode valves, indicatingpossible mixing of the lake (see arrows, Fig. 6) and down-slopetransport, because of colder temperatures and higher precipita-tion typical of the LGM (Hodell et al., 2008). From w23.0 to22.0 cal ka BP, the increase in P. globula, along with a decline in L.opesta, reflects a continuous increase in lake level. C/N ratios werelow and stable, indicating that deposited organic matter came fromautochthonous production. Ostracode analyses, combined withlithologic and magnetic susceptibility measurements (Muelleret al., 2010), and TS and C/N ratios, indicate that wetter andcolder conditions characterized the LGM.

5.3.2. Fluctuating lake levels (w10e31 m) and cold conditions (HS1,w19e15 ka)

The HS1 is characterized by dominance of L. opesta and P. globula(Fig. 6) suggesting colder temperatures. Inferred past water depths,changes in the relative abundance of L. opesta, TS concentration, C/N ratios and percentage of adult and juvenile broken valves indicatelake level fluctuations, especially fromw19 to 16.7 cal ka BP (Fig. 7).Both low and high lake levels characterized HS1b (subunit V), but



lower lake levels, gypsum deposition and dry conditions weremorecommon. Relative abundance of L. opesta during HS1 displaystrends that correspond to C/N ratios and magnetic susceptibility.High L. opesta percentages are associated with high C/N ratios andlow magnetic susceptibility, validating previous autecologicalstudies (Pérez et al., 2010b) that found this species is more abun-dant in shallow waters that are rich in aquatic vegetation.

Hodell et al. (2012) described the HS1 in northern Guatemala asan arid period displaying hypolimnetic water temperatures as lowas 16e20 �C, estimated from coupled measurements of oxygenisotopes in gypsum hydration water and ostracode shellcarbonate. Although drier climate conditions prevailed, therewere short wet periods during HS1. Our independent water depthreconstruction provides a more detailed picture and suggestshigher lake levels during the early deglacial (w19e18.5 ka), at17.5, 17 (subunit V), and from 16.5 to 16.0 ka (subunit IV, Fig. 7),coinciding with slightly lower d13C and d18O values in ostracodevalves (Fig. 9). Similar to the LGM conditions, the low stable isotopevalues are attributed mainly to 1) 13C-depleted deep waters, owingto decay of organic matter and 2) 18O-depleted waters owing tohigher precipitation and lower evaporation.

Benthic species Pseudocandona sp. and D. stevensoni are stillpresent during the early deglacial, but valve numbers begin todecrease w18.0 cal ka BP, indicating increasing aridity that char-acterizes most of HS1. Both species were present during the LGM,because they prefer slightly colder temperatures and higher waterdepths, contrary to the climatic conditions during HS1. Themaximum number of adult and juvenile valves during HS1,however, occurs at 18.0 cal ka BP (Fig. 6), when lake levels droppedafter the LGM. Cypridopsis okeechobei, a nektobenthic species thatprefers littoral zones with abundant macrophytes and lives ata maximum water depth of 40 m, appears at the onset of thedeglacial, when lake level starts to decrease (Fig. 7). The d13C valuesof C. okeechobei are more positive than those of benthic L. opestaand nektobenthic P. globula, reflecting high photosyntheticproductivity in their littoral habitat (Fig. 8). Vital effects, however,could have played an important role. The d18O values ofC. okeechobei are generally slightly more negative than those ofL. opesta and P. globula, different from ourmodernmeasurements inwhich C. okeechobei displays higher d18O values (Fig. 4). This mightbe a consequence of the thermal stratification of the water column.Today, the lake is thermally stratified during summer and mixesweakly in winter months when temperatures in the water columndecrease and become nearly isothermic. Similar d18O values ofbenthic and nektobenthic species could suggest lower lake levels,colder temperatures, and lack of lake thermal stratification duringHS1b. Hodell et al. (2012) pointed out that the temperaturereconstruction using d18O in benthic L. opesta valves reflects pasthypolimnetic temperatures, most likely an estimate of winter watercolumn temperatures if the lake was deep and stratified. Suchtemperature inferences may apply more generally to the wholewater column if lake temperatures were cooler and circulation wasmore regular.

Following higher lake levels at the onset of the deglacial, lowerlake levels were common during HS1b, and themainly dry HS1 waspunctuated by warmer and/or wetter periods (subunit IV) whenclay was deposited in the lake. Low d18O values at 17.0 and16.1 cal ka BP coincide with clay deposition (subunit IV), indicatingan increase in precipitation and therefore higher lake levels. Thenegative d18O peak at 16.1 cal ka BP in subunit IV indicates most ofthe period was warmer and/or wetter, but at 16.6 cal ka BP, d18Ovalues increase slightly, suggesting there was a short period of drierconditions.

Total sulfur concentration remains relatively high throughoutHS1 even in subunit IV when calcite-rich clay accumulated in the



lake. Even if warmer and/or wetter conditions dominated duringthe phase between HS1b and HS1a, slight increases in TS concen-tration, C/N ratio and a higher relative abundance of L. opestasuggest lower lake levels during most of HS1. The number of valvesof C. okeechobei dramatically decreased at w17.0 cal ka BP whenlake level increased (subunit IV). This species and L. opesta recov-ered when gypsum deposition began again about 15.7 cal ka BP(Fig. 6), indicating lake level lowering.

Littoral zone indicators H. punctata and S. intrepida appearsporadically during HS1, coinciding with gypsum deposition andindicating much lower lake levels (HS1a, b, Fig. 6). NektobenthicH. punctata displayed highest d13C values during HS1b and HS1aand highest d18O values during HS1a (Fig. 8). Similar isotope valueswere seen in modern ostracodes restricted to shallow waters,indicating that this nektobenthic species can be used as an indi-cator of shallow waters.

During HS1a, arid conditions negatively affected species such asL. opesta, P. globula and C. okeechobei. This marked the first longperiod (15.5e15.3 cal ka BP) when P. globulawas completely absentand suggests drier conditions persisted for amuch longer time thanpreviously. This was apparently a stable phase of HS1, characterizedby low lake level, high TS concentration, generally high C/N ratiosand high percentages of broken ostracode valves, attesting to highenergy levels. Highest d18O values during HS1 were determinedduring HS1a, as well as slightly higher d13C values. Higher d18Ovalues during HS1a might be a result of greater and longer aridityandhigher evaporation thanduringHS1b. Higher d13C values duringHS1a could suggest lower lake levels and higher productivity.

High-energy environments and lake mixing likely occurredduring HS1b (subunit V) and HS1a (subunit III) when conditionswere cold and lake levels low (Fig. 6). C/N ratios and the percent ofbroken adult and juvenile valves (Fig. 7) fluctuate strongly duringtimes of gypsum deposition (subunits V and III), coinciding withtimes of valve transport. Colder temperatures and shallower watermay have permitted greater lake mixing, which could explain thehigh percentage of broken valves.

5.3.3. High lake levels (w40 m), warmer and wetter conditions(BA, w15e13 ka)

The BA is dominated by P. globula, followed by L. opesta. Otherostracode species are absent or scarce. Highest numbers of adultand juvenile valves during the deglacial were recorded for the BAand consist mostly of P. globula. Thus, the BA assemblage closelyresembles the modern assemblage at the core site (71 m waterdepth). Physocypria globula is a species known to prefer deeperwaters (Pérez et al., 2010b). The BA displayed relatively stable highlake level, low abundance of L. opesta, high magnetic susceptibility,low TS and C/N ratios, and low ostracode d18O and d13C values. Themaximum inferred water depth at the core site was w76 m, sug-gesting deeper water than today (71 m at core site). The appliedtransfer function, however, has a RMSE of 9 m and other paleo-climate proxies (Hodell et al., 2008; Mueller et al., 2010) do notsuggest such deep waters.

Carbon isotope values were slightly higher during the BA thanduring the LGM, although inferred lake levels are higher thanduring the LGM. It might have been that water temperatures duringthe BA were much higher (20e24 �C, Hodell et al., 2012), whichcould have increased lake productivity. Similar d13C values(w�6&) were determined in modern ostracode shells of L. opestacollected at water depths >20 m, suggesting that the source ofcarbon andwater temperatures could be similar to those during theBA. d18O values are lower than those measured during the LGM,indicating wetter and warmer conditions in the BA. The d13C recordof L. opesta and P. globula display similar values, whereas their d18Ovalues differ from one another. The d18O values of P. globula during



the BA are generally more negative than those for L. opesta, exceptat 14.1 cal ka BP (Fig. 8). We found similar d13C values in bothspecies in the deep, modern lake (Figs. 4 and 5). This was not thecase for d18O differences between the species. In our modern dataset for winter months, benthic L. opesta display more negative d18Ovalues than nektobenthic P. globula. The BA may resemble themodern summer months, in that nektobenthic species displaymore negative d18O values as a consequence of higher temperaturesand summer precipitation (�4.0&) (Hillesheim et al., 2005),leading to more negative d18O values in lake waters.

Lower percentages of broken adult and juvenile valves in the BA,compared to the rest of the Lateglacial and little fluctuation in thestable isotope values, indicate a stable, thermally stratified lakewith little mixing or water currents. At w14.6 cal ka BP there isa peak in the relative abundance of L. opesta, and slightly lower lakelevels and magnetic susceptibility, indicating a short, drier period.

5.3.4. Low lake levels (w4 m) and dry conditions (YD,w13e11.5 ka)

The YD is characterized by the dominance of L. opesta, whereasP. globula is rare, similar to the situation during HS1a. Littoral zoneindicator H. punctata is also present. Total numbers of adult andjuvenile valves are much lower than during the BA, despite the factthat lake levels were lower and more nektobenthic and benthicspecies might be expected. The relative abundance of L. opestawashigh, TS concentrations and C/N ratios remained high, and thepercentage of broken valves was again high, similar to all previousperiods of low lake level characterized bygypsumdeposition. Stableisotope values of L. opesta fluctuated slightly, however a decrease isseen from w12.8 to 12.5 cal ka BP. d13C values of C. okeechobei at12.5 ka are much higher than those of L. opesta, similar to ourmodern data. The oxygen and carbon stable isotope values firstdecrease and then increase, correlating with our water depthreconstruction. A slight increase in lake level (<35 m) occurred atthe onset of the YD, followed by a later lake level drop (<10 m).Other sediment variables do not displaymajor changes at that time,although the TS profile shows a decrease, and inferred lake levelsdid not drop between 12.8 and 12.5 cal ka BP (Fig. 7). This suggeststhat cold temperatures drove a decrease in lake productivity anda short wet period might explain the lower d18O values.

5.3.5. High lake levels (w58 m), warm and wetter conditions (earlyHolocene, 11.5e10.0 ka)

The onset of the early Holocene is well defined by an increase inostracodes (L. opesta and P. globula) and inferred lake level. LowerTS concentrations and C/N ratios, higher percentages of brokenostracode valves and declines in d13C and d18O values of ostracodesindicate wetter conditions, rising lake level and stronger currents.

6. Conclusions

Modern limnological information and species autecology arehelpful for understanding stable isotope signatures of ostracodespecies that live in Lago Petén Itzá. Such information is essential forinferring past lake hydrodynamics from isotope measure insubfossil ostracode shells in lake sediment cores. Lake watertemperatures, together with environmental variables dissolvedoxygen, pH, and water depth, chemistry, substrate, macrophytecover, habitat structure, and food availability determine the spatialdistribution of ostracodes in Lago Petén Itzá. Generally, d18O andd13C values of littoral and shallow-water species were higher thanthose inhabiting deeper waters, i.e. �20 m. d13C values generallydecline moving from the epilimnion to the deeper thermoclinewaters (20e40 m). Limnocythere opesta displayed the lowest d18Oand d13C values and the lowest d18O offset from calcite equilibrium



among modern ostracode species. Nektobenthic species C. okee-chobei, H. punctata and S. intrepida showed highest d13C values,reflecting higher productivity. Low isotope values in L. opestasuggest it might be an infaunal species, whereas other benthic taxa,displaying higher d13C values such as D. stevensoni and C. ilosvayi,might be epifaunal.

The late Pleistocene stable isotope record (d13C and d18O) ofostracodes in Lago Petén Itzá shows fluctuations that are explainedmainly by changes in the balance between evaporation andprecipitation, lake stage, and carbon source. We identified fiveperiods in the late Pleistoceneeearly Holocene record from LagoPetén Itzá, with respect to lake level and evaporation/precipitation(E/P), using ostracode analysis in combination with lithology,magnetic susceptibility, TS and C/N ratios: 1) high lake levels andcold conditions (LGM), 2) fluctuating lake levels and cold condi-tions (HS1), 3) high lake levels and warm and wetter conditions(BA), 4) low lake levels and dry conditions (YD) and 5) high lakelevels and warm and wetter conditions (early Holocene). Thecomposition of fossil species assemblages changed at the LGMe

early deglacial transition. Benthic species and P. globula character-ized the LGM, and nektobenthic species and L. opesta dominatedduring the deglacial. Deep-water species P. globula dominatedduring the early Holocene.

Future studies should carry out monthly sampling of limno-logical variables throughout the year over several years in this largeNeotropical lake. The isotopic composition of multiple livingostracode species, as well as their life cycles, should be analyzed athigh temporal resolution throughout the year as well. Interpreta-tion of long stable isotope records might be improved if suchinformationwere acquired, along with monthly fluctuations in lakevariables. Our study demonstrates that use of isotopic measure-ments from multiple ostracode species to generate reliable paleo-environmental inferences, requires a strong understanding ofspecies ecological preferences and lake hydrodynamics.

Acknowledgments

We thank all who assisted us during field trips to Lago Petén Itzáand other regional waterbodies. We are grateful to the followingpeople and agencies: Roberto Moreno, Margarita Palmieri, Mar-garet Dix, Eleonor de Tott (Universidad del Valle de Guatemala),Consejo Nacional de Áreas Protegidas (CONAP), Asociación para elManejo y Desarrollo Sostenible de la cuenca del Lago Petén Itzá(AMPI), Michael Hillesheim, Burkhard Scharf, Julia Lorenschat, RitaBugja, Susanne Krüger,Wolfgang Riss and Evgenia Vinogradova andthe team of the Petén Itzá scientific sampling party as well as thestaff of the National Lacustrine Core Facility (LacCore, Minnesota,USA) for core curation and members of the Drilling, Observationand Sampling of the Earth’s Continental Crust (DOSECC) for theirsupport during drilling operations in Lago Petén Itzá in 2006. TheDeutsche Forschungsgemeinschaft (DFG, grant Schw 671-3),Technische Universität Braunschweig (TU-BS, Germany), NationalScience Foundation (NSF, USA), International Continental ScientificDrilling Program (ICDP), Swiss Federal Institute of Technology andSwiss National Science Foundation provided financial support forcore recovery and analysis.

References

Basterrechea, M., 1988. Limnología del Lago Petén itzá, Guatemala. Rev. Biol. Trop.36, 123e127.

Brehm, V., 1939. La Fauna microscopica del Lago Petén, Guatemala. An. Esc. Nac.Cienc. Biol. 1, 173e203.

Brenner, M., Rosenmeier, M., Hodell, D.A., Curtis, J.H., Anselmetti, F., Ariztegui, D.,2002. Limnología y paleolimnología de Petén, Guatemala. Revista Universidaddel Valle de Guatemala 12, 2e9.



Bridgwater, N.D., Heaton, T.H.E., O’Hara, S.L., 1999. A late Holocene paleolimno-logical record from central Mexico, based on faunal and stable-isotope analysisof ostracod shells. J. Paleolimnol. 22, 383e397.

Bright, J., Kaufman, D.S., Forester, R.M., Dean, W.E., 2006. A continuous 250,000 yrrecord of oxygen and carbon isotopes in ostracode and bulk-sedimentcarbonate from Bear Lake, Utah-Idaho. Quat. Sci. Rev. 26, 2258e2270.

Bush, M.B., Correa-Metrio, A.Y., Hodell, D.A., Brenner, M., Anselmetti, F.S.,Ariztegui, D., Mueller, A.D., Curtis, J.H., Grzesik, D.A., Burton, C., Gilli, A., 2009.Re-evaluation of climate change in lowland Central America during the LastGlacial Maximum using new sediment cores from Lake Petén Itzá, Guatemala.In: Vimeux, F., Sylvestre, F., Khodri, M. (Eds.), Past Climate Variability in SouthAmerica and Surrounding Regions from the Last Glacial Maximum to theHolocene. Springer, pp. 113e128.

Cohen, A., 2003. Paleolimnology. Oxford University Press, New York.Covich, A.P., Stuiver, M., 1974. Changes in oxygen 18 as a measure of long-term

fluctuations in tropical lake levels and molluscan populations. Limnol. Ocean-ogr. 19, 682e691.

Curry, B.B., Filippelli, G.M., 2010. Episodes of low dissolved oxygen indicated byostracodes and sediment geochemistry at Crystal Lake, Illinois, USA. Limnol.Oceanogr. 55, 2403e2423.

Curtis, J., Brenner, M., Hodell, D., Balser, R., Islebe, G., Hoghiemstra, H., 1998. A multi-proxy study of Holocene environmental change in the Maya lowlands of Peten,Guatemala. J. Paleolimnol. 19, 139e159.

Curtis, J., Hodell, D., Brenner, M., 1996. Climate variability on the Yucatan Peninsula(Mexico) during the past 3500 Years, and implications for Maya culturalevolution. Quat. Res. 46, 37e47.

Dean, W., 2010. Recent advances in global lake coring hold promise for globalchange research in paleolimnology. J. Paleolimnol. 44, 741e743.

Decrouy, L., Vennemann, T.W., Ariztegui, D., 2011. Controls on ostracod valvegeochemistry: part 2. Carbon and oxygen isotope compositions. Geochim.Cosmochim. Acta 75, 7380e7399.

Escobar, J., Hodell, D.A., Brenner, M., Curtis, J.H., Gilli, A., Mueller, A.D.,Anselmetti, F.S., Ariztegui, D., Grzesik, D.A., Pérez, L., Schwalb, A.,Guilderson, T.P., 2012. A w43-ka record of paleoenvironmental change in theCentral American lowlands inferred from stable isotopes of lacustrine ostra-cods. Quat. Sci. Rev. 37, 92e104.

Ferguson, E.G., Hutchinson, G.E., Goulden, C.E., 1964. Cypria petenensis, a new namefor the ostracod Cypria pelagica Brehm 1932. Postilla 80, 1e4.

Furtos, N., 1933. The ostracoda of Ohio. In: Ohio Biological Survey. The Ohio StateUniversity Press, Columbus, pp. 413e524.

Furtos, N., 1936a. Fresh-water ostracoda from Florida and North Carolina. Am. Midl.Nat. 17, 491e522.

Furtos, N., 1936b. On the Ostracoda from the cenotes of Yucatán and vicinity. In: TheCenotes of Yucatan, a Zoological and Hydrographic Survey. Carnegie Institutionof Washington, Washington, pp. 89e115.

Hammer, Ø., Harper, D.A.T., Ryan, P.D., 2001. PAST: palaeontological statistics soft-ware package for education and data analysis. Palaeontol. Electr. 4, 9.

Hastenrath, S.L., 1991. Climate Dynamic of the Tropics. Kluwer Academic Publishers,Dordrecht, Netherlands.

Hillesheim, M., Hodell, D., Leyden, B., Brenner, M., Curtis, J., Anselmetti, F.,Ariztegui, D., Buck, D., Guilderson, T., Rosenmeier, M., Schnurrenberger, D.,2005. Climate change in lowland Central America during the late deglacial andearly Holocene. J. Quat. Sci. 20, 363e376.

Hodell, D.A., Turchyn, A.V., Wiseman, C.J., Escobar, J., Curtis, J.H., Brenner, M.,Gilli, A., Mueller, A.D., Anselmetti, F., Ariztegui, D., Brown, E.T., 2012. Late Glacialtemperature and precipitation changes in the lowland Neotropics by tandemmeasurement of d18O in biogenic carbonate and gypsum hydration water.Geochim. Cosmochim. Acta 77, 352e368.

Hodell, D.A., Anselmetti, F.S., Ariztegui, D., Brenner, M., Curtis, J.H., Gilli, A.,Grzesik, D.A., Guilderson, T., Mueller, A.D., Bush, M.B., Correa-Metrio, A.,Escobar, J., Kutterolf, S., 2008. An 85-ka record of climate change in lowlandCentral America. Quat. Sci. Rev. 27, 1152e1165.

Hodell, D.A., Curtis, J.H., Brenner, M., 1995. Possible role of climate in the collapse ofClassic Maya civilization. Nature 375, 391e394.

Hodell, D.A., Curtis, J.H., Jones, G., Higuera-Gundy, A., Brenner, M., Binford, M.,Dorsey, K., 1991. Reconstruction of Caribbean climate change over the past10,500 years. Nature 352, 790e793.

Holmes, J., 1996. Trace-element and stable-isotope geochemistry of non-marineostracod shells in Quaternary paleoenvironmental reconstruction.J. Paleolimnol. 15, 223e235.

Holmes, J.A., Chivas, A.R., 2002. The Ostracoda: Applications in QuaternaryResearch. In: Geophysical Monograph Series, vol. 131. Washington, D.C.

Juggins, S., 2003. C2. 1.5 ed. Newcastle University, Newcastle upon Tyne.Keyser, D., 1976. Zur Kenntnis der brackigen mangrovebewachsenen Weichböden

Südwest-Floridas unter besonderer Berücksichtigung ihrer Ostracodenfauna.Ph.D. thesis, Universität Hamburg, Germany.

Keyser, D., 1977. Brackwasser-Cytheracea aus Süd-Florida (Crust.: Ostracoda:Podocopa). Abh. Verh. Naturwiss. Ver. Hamburg 20, 43e85.

Kim, S.-T., O’Neil, J.R., 1997. Equilibrium and nonequilibrium oxygen isotope effectsin synthetic carbonates. Geochim. Cosmochim. Acta 16, 3461e3475.

Lister, G.S., 1988. A 15,000-year isotopic record from Lake Zürich of deglaciation andclimatic change in Switzerland. Quat. Res. 29, 129e141.

Meisch, C., 2000. Freshwater Ostracoda of Western and Central Europe.Süßwasserfauna von Mitteleuropa. Spektrum Akademischer Verlag GmbH,Heidelberg, Berlin, Germany.



Mueller, A.D., Anselmetti, F.S., Ariztegui, D., Brenner, M., Hodell, D.A., Curtis, J.H.,Escobar, J., Gilli, A.,Grzesik, D.A.,Guilderson, T.P., Kutterolf, S., Plötze,M.L., 2010. LateQuaternary palaeoenvironment of northern Guatemala: evidence from deep drillcores and seismic stratigraphy of Lake Petén Itzá. Sedimentology 57, 1220e1245.

Park, L., Cohen, A.S., Martens, K., Balek, R., 2003. The impact of taphonomicprocesses on interpreting paleoecologic changes in large lake ecosystems:ostracodes in Lake Tanganyika and Malawi. J. Paleolimnol. 30, 127e138.

Park, L.E., Cohen, A.S., 2011. Paleoecological response of ostracods to early LatePleistocene lake-level changes in Lake Malawi, East Africa. Palaeogeogr. Palae-oclimatol. Palaeoecol. 303, 71e80.

Pérez, L., Lorenschat, J., Massaferro, J., Pailles, C., Sylvestre, F., Hollwedel, W.,Brandorff, G.-O., Brenner, M., Islebe, G., Lozano, M.S., Scharf, B., Schwalb, A.Bioindicators of climate and trophic state in lowland and highland aquaticecosystems of the northern Neotropics. Rev. Biol. Trop., in press.

Pérez, L., Lorenschat, J., Brenner, M., Scharf, B., Schwalb, A., 2012. Non-marineostracodes (Crustacea) of Guatemala. In: Cano, E.B., Schuster, J.C. (Eds.), Bio-diversidad de Guatemala, vol. 2. Universidad del Valle de Guatemala,Guatemala, pp. 121e131.

Pérez, L., Frenzel, P., Brenner, M., Escobar, E., Hoelzmann, P., Scharf, B., Schwalb, A.,2011. Late Quaternary (24e10 ka BP) environmental history of the Neotropicallowlands inferred from ostracodes in sediments of Lago Petén Itzá, Guatemala.J. Paleolimnol. 46, 59e74.

Pérez, L., Lorenschat, J., Brenner, M., Scharf, B., Schwalb, A., 2010a. Extant freshwaterostracodes (Crustacea: Ostracoda) from Lago Petén Itzá, Guatemala. Rev. Biol.Trop. 58, 871e895.

Pérez, L., Lorenschat, J., Bugja, R., Brenner,M., Scharf, B., Schwalb, A., 2010b.Distribution,diversity and ecology of modern freshwater ostracodes (Crustacea), and hydro-chemical characteristics of Lago Petén Itzá, Guatemala. J. Limnol. 69, 146e159.



Pérez, L., Bugja, R., Massaferro, J., Steeb, P., van Geldern, R., Frenzel, P., Brenner, M.,Scharf, B., Schwalb, A., 2010c. Post-Columbian environmental history of LagoPetén Itzá, Guatemala. Rev. Mex. Cienc. Geol. 27, 490e507.

Rice, D.S., 1997. Ingeniería Hidráulica en el centro de Petén, Guatemala. In:Laporte, J.P., Escobedo, H. (Eds.), X Simposio de Investigaciones Arqueológicas yEtnológicas, Guatemala, 1996. Museo Nacional de Arqueología y Etnología,Guatemala, pp. 581e594.

Rosenmeier, M., Hodell, D., Brenner, M., Curtis, J., Martin, J., Anselmetti, F.,Ariztegui, D., Guilderson, T., 2002. Influence of vegetation change on watershedhydrology: implications for paleoclimatic interpretation of lacustrine d18Orecords. J. Paleolimnol. 27, 117e131.

Rosenmeier, M.F., Brenner, M., Kenney, W.F., Whitmore, T.J., Taylor, C.M., 2004.Recent eutrophication in the southern basin of Lake Petén Itzá, Guatemala:human impact on a large tropical lake. Hydrobiologia 511, 161e172.

Schwalb, A., 2003. Lacustrine ostracodes as stable isotope recorders of late-glacial and Holocene environmental dynamics and climate. J. Paleolimnol.29, 256e351.

Van der Meeren, T., Ito, E., Verschuren, D., Almendinger, J.E., Martens, K., 2011. Valvechemistry of Limnocythere inopinata (Ostracoda) in a cold arid environment eimplications for paleolimnological interpretation. Palaeogeogr. Palaeoclimatol.Palaeoecol. 306, 116e126.

von Grafenstein, U., Erlenkeuser, H., Müller, J., Kleinmann-Eisenmann, A., 1992.Oxygen isotope records of benthic ostracods in Bavarian lake sediments.Naturwissenschaften 79, 145e152.

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. Palaeogeogr. Palae-oclimatol. Palaeoecol. 148, 133e152.


quaternary science reviews · stable isotope values (d18o&d13c) of multiple ostracode species...

Documents