INTRODUCTION
Lake sediments are excellent archives to reconstruct long-term fluctuations of environmental conditions (Williamson et al., Reference Williamson, Saros, Vincent and Smol2009). This is particularly true in arid and semiarid regions, where a direct relationship exists between climate and watershed hydrology, and water loss is mostly the result of evapotranspiration (Holmes et al., Reference Holmes, Fothergill, Street-Perrott and Perrott1998). This type of environment is particularly sensitive to hydrologic changes and consequently good for assessing evapotranspiration and climatic variability. Environmental fluctuations trigger changes in the biota, the mineralogy, and the geochemistry of lakes, which are recorded in the lake sediments (Battarbee, Reference Battarbee2000; Piovano et al., Reference Piovano, Ariztegui and Moreira2002).
Ostracods and chironomids are part of the biological records of lake sediments. Their occurrence and abundance are defined by biological variables and the physiochemical properties of their host water, which generally reflect the same physical and chemical conditions in the past (Forester, Reference Forester1991). Therefore, both organisms are excellent proxies in paleoenvironmental studies, especially for the Quaternary (Cusminsky et al., Reference Cusminsky, Schwalb, Pérez, Pineda, Viehberg, Whatley, Markgraf, Gilli, Ariztegui and Anselmetti2011; Ramón Mercau et al., Reference Ramón Mercau, Laprida, Massaferro, Rogora, Tartari and Maidana2012). Ostracoda analyses have provided information about salinity, solute composition, temperature, and stream flow conditions (Palacios-Fest et al., Reference Palacios-Fest, Cohen and Anadón1994; Markgraf et al., Reference Markgraf, Bradbury, Schwalb, Burns, Stern, Ariztegui, Gilli, Anselmetti, Stine and Maidana2003; Mezquita et al., Reference Mezquita, Roca, Reed and Wansard2005). Meanwhile, chironomids have been used to identify changes in temperature, salinity, and lake productivity (Walker, Reference Walker2001; Brodersen and Anderson, Reference Brodersen and Anderson2002; Heinrichs and Walker, Reference Heinrichs and Walker2006; Urrutia et al., Reference Urrutia, Araneda, Torres, Cruces, Vivero, Torrejón, Barra, Fagel and Scharf2010; Massaferro et al., Reference Massaferro, Recasens, Larocque-Tobler, Zolitschka and Maidana2013, Reference Massaferro, Larocque-Tobler, Brooks, Vandergoes, Dieffenbacher-Krall and Moreno2014). The combined use of these organisms in paleoenvironmental studies is increasing (Pérez et al., Reference Pérez, Bugja, Massaferro, Steeb, van Geldern, Frenzel, Brenner, Scharf and Schwalb2010; Bunbury and Gajewski, Reference Bunbury and Gajewski2012; Laprida et al., Reference Laprida, Massaferro, Ramón Mercau and Cusminsky2014; Ohlendorf et al., Reference Ohlendorf, Fey, Massaferro, Haberzettl, Laprida, Lücke and Maidana2014; Mayr et al., Reference Mayr, Laprida, Lücke, Martín, Massaferro, Ramón-Mercau and Wissel2015). In this context, these multiproxy studies are very important for a better understanding of past environmental changes because they have great power in not only generating valuable paleoenvironmental information, but also potentially enabling the integrated responses of different biological communities (Battarbee, Reference Battarbee2000).
During recent decades, late Holocene paleoclimatic reconstructions carried out on both sides of the Andes in northern Patagonia have increased, allowing us to identify variations in temperature and precipitation during the last millennium. Based on dendrochronological, glacial analysis, and rainfall reconstructions, researchers have identified a warm interval between the eleventh and thirteenth centuries, coincident approximately with the Medieval Climate Anomaly observed in the Northern Hemisphere (Villalba, Reference Villalba1994; Mann, Reference Mann2002). Consistent with these findings, the analysis of varved lake sediments from Lake Puyehue (40°40′S, 72°28′W; 185 m above sea level [asl]), Chile, suggests the existence of a more arid climate between AD 1400 and 1500 (Boës and Fagel, Reference Boës and Fagel2008). Afterward, the temporal evolution of sediment supply in the same lake examined using geochemistry and sedimentology, as well as dendrochronological studies, shows a cold-moist period between AD 1500 and 1700, probably associated with the onset of the Little Ice Age (LIA), and a drier period during AD 1700–1900 (Lara and Villalba, Reference Lara and Villalba1993; Villalba, Reference Villalba1994; Luckman and Villalba, Reference Luckman and Villalba2001; Bertrand et al., Reference Bertrand, Boës, Castiaux, Charlet, Urrutia, Espinoza, Lepoint, Charlier and Fagel2005; Boës and Fagel, Reference Boës and Fagel2008; De Batist et al., Reference De Batist, Fagel, Loutre and Chapron2008). On the east side of the Andes, paleostudies using lacustrine sedimentary varves, chironomids, and fossil pigment records in Frías Lake (41°04′S, 71°48′W; 790 m asl) and Hess Lake (41°20′S, 71°44′W; 735 m asl) have shown low precipitation around AD 1800, followed by a period of low temperatures between AD 1800 and 1940 (Ariztegui et al., Reference Ariztegui, Bösch and Davaud2007; Guilizzoni et al., Reference Guilizzoni, Massaferro, Lami, Piovano, Ribeiro Guevara, Formica, Daga, Rizzo and Gerli2009). Finally, examination of instrumental and tree-ring records have shown that the twentieth century is the most variable climatically, with extreme intervals of drought and very wet periods (Villalba et al., Reference Villalba, Lara, Boninsegna, Masiokas, Delgado, Aravena, Roig, Schmelter, Wolodarsky and Ripalta2003). Together, these studies describe paleoclimatic variations in northern Patagonia on the both sides of the Andes. However, paleoenvironmental data from the eastern steppe region are not sufficient to allow a detailed reconstruction of climate during last millennium. Paleoclimatological research on the forest-steppe ecotone has focused mainly on the dendrochronology of the Austrocedrus chilensis (Villalba and Veblen, Reference Villalba and Veblen1996, Reference Villalba and Veblen1997). The lacustrine basin Cari-Laufquen (41º35′S, 69º25′W) was the only lake intensely studied in northern Patagonian for paleohydrologic reconstruction. In this environment, paleoshorelines and ostracoda assemblage analysis suggested several high lake levels during the Late Pleistocene (Garleff et al., Reference Garleff, Reichert, Sage, Schäbitz and Stein1994; Bradbury et al., Reference Bradbury, Grosjean, Stine and Sylvestre2001; Cartwright et al. Reference Cartwright, Quade, Stine, Adams, Broecker and Cheng2011). Afterward, a subsequent drop in water level was detected, suggesting a decrease in rainfall and drier conditions in the area during the early Holocene (Whatley and Cusminsky Reference Whatley and Cusminsky1999, Reference Whatley and Cusminsky2000; Cusminsky et al., Reference Cusminsky, Schwalb, Pérez, Pineda, Viehberg, Whatley, Markgraf, Gilli, Ariztegui and Anselmetti2011).
The aim of this study is to provide a hydroclimate reconstruction of the last millennium from El Toro Lake, located in the northern Patagonian steppe, to contribute to the knowledge of late Holocene climate change. For this purpose, we analyzed biological, sedimentological, geochemical, and mineralogical proxies to infer hydrologic changes in El Toro Lake.
SITE LOCATION AND DESCRIPTION
El Toro Lake (40°19′S, 70°25′W; 1025 m asl) is located 50 km southwest of the town of Piedra del Águila (Neuquén province) in northern Patagonia, Argentina (Fig. 1). This body of water is located in the steppe ecoregion, with a low annual precipitation (~300 mm) because of the orographic rain shadow effect of the Andes on the southern westerly winds (SWWs). Patagonia is defined as a temperate or cool-temperate region (Paruelo et al., 1998), and the Piedra del Águila area is characterized by mean annual, summer, and winter temperatures of 11.6, 18.2, and 5.1°C, respectively (http://es.climate-data.org). Midlatitude westerlies dominate the climatic system in this area; hence, annual and seasonal precipitation variability is largely explained by changes in the strength and position of the SWWs. During the austral winter (July), SWW storm tracks migrate toward the equator to a latitude of ~40°S, defining this as the rainy season, whereas precipitation declines as a result of a poleward displacement of the SWWs during the Austral summer (Ariztegui et al., Reference Ariztegui, Bösch and Davaud2007; Iglesias et al., Reference Iglesias, Whitlock, Bianchi, Villarosa and Outes2012). The vegetation surrounding the lake is characterized by a matrix of cushion shrubs (e.g., Mulinum spinosum), tussock grasses (e.g., Stipa speciosa, Festuca pallenscens), and herbs (e.g., Asteraceae sp.). El Toro is an endorheic lake, with a surface area of 0.63 km2 and a drainage area of ~12 km2. The catchment is a closed basin with the water body occupying a deflation depression. Most of the catchment is composed of salty deposits formed by sporadic wetting and drying around the lake, over an undulating plateau landscape that has an average altitude of more than 1000 m asl (with the highest points reaching 1200 m) composed of olivine-rich basalts (Cucchi et al., Reference Cucchi, Espejo and González1998), conforming a restricted catchment. There are no records of carbonate rock exposures in the basin. A bathymetric map of the lake is not available, but direct observations and partial measurements during the sampling and later fieldwork show that this is a more or less concentric basin, with the lake depth varying by a few meters. With the water body at 10 cm maximum depth (austral fall, 2016), temperature was 5.6°C, dissolved oxygen was 13.6 mg/L, conductivity was 11,400 µS/cm, and pH was 9.5, with ion concentrations as follows: Na, 62 mg/L; Ca, 46 mg/L; Mg, 23 mg/L; K, 6.0 mg/L; and Cl, 47 mg/L. In the austral summer of 1997, Tartarotti et al. (Reference Tartarotti, Baffico, Temporetti and Zagarese2004) recorded a maximum water depth of approximately 1 m, a chlorophyll a concentration of 1.98 mg/m3, a dissolved organic matter (OM) concentration of 521 g/cm3, and conductivity between 3950 and 8800 µS/cm. During core recovery, in the austral fall of 2002, El Toro Lake had a maximum water depth of 4 m, pH of 10 (Corning 360i pH Meter), dissolved oxygen concentration of 10.9 mg/L (SensION 6 Dissolved Oxygen Meter), and conductivity of approximately 2510 µS/cm (Thermo Orion Model 135A Conductivity Meter). About 150 km to the west is located the Southern Volcanic Zone (33°–46°S; Stern, Reference Stern2004) of the Andes, a current active volcanic arc including several volcanoes with activity during historical times affecting northern Patagonia.
(color online) Location of El Toro Lake in Neuquén province of northern Patagonia in Argentina. At right, image of El Toro Lake (obtained through Google Earth). Paleoshorelines are observed.
METHODOLOGY
Core extraction
A short sediment core, 32.5 cm long, was extracted in 2002 from the deepest part of the lake, in a water depth of 4 m, using a messenger-activated gravity type corer. The sampling site was selected after confirmation using an echo sounder survey that the lake bottom was flat (Rizzo, Reference Rizzo2007). In the laboratory, the core was longitudinally cut, opened, photographed, visually inspected, and subsampled every 1 cm or at natural boundaries. Each subsampled sediment layer was freeze-dried until constant weight. Physical properties of the sediment as water content and dry density were calculated by weighing wet and freeze-dried subsamples. After inspection of the dry sediment under a binocular magnifying glass, four layers were identified as tephras because of higher volcanic glass concentration relative to adjacent layers. These layers were not seen during visual inspection.
Dating methods
A chronology for the core was determined using 210Pb and 137Cs techniques in the upper portion (Joshi and Shukla, Reference Joshi and Shukla1991; Robbins and Herche, Reference Robbins and Herche1993; Ribeiro Guevara and Arribére, Reference Ribeiro Guevara and Arribére2002; Ribeiro Guevara et al., Reference Ribeiro Guevara, Rizzo, Sánchez and Arribére2003) and accelerator mass spectrometry (AMS) radiocarbon dating of the older sediment. The 210Pb, 226Ra, and 137Cs specific activity profiles were determined by high-resolution gamma ray spectrometry using a well-type HPGe detector (Centro Atómico Bariloche) (Ribeiro Guevara and Arribére, Reference Ribeiro Guevara and Arribére2002; Ribeiro Guevara et al., Reference Ribeiro Guevara, Rizzo, Sánchez and Arribére2003). Radiocarbon dating was conducted on ostracod shells recovered at depths of 30 and 25 cm at the Beta Analytic Radiocarbon Dating Laboratory and DirectAMS, respectively. Two extra samples of modern ostracod shells and an aquatic plant were dated at the DirectAMS laboratory to evaluate any hard-water effect. The calibration and integration of the information in the age-depth model was carried out using OxCal 4.2 (available at https://c14.arch.ox.ac.uk/oxcal/OxCal.html), using the Southern Hemisphere radiocarbon calibration data SHCal13 (Hogg et al., Reference Hogg, Hua, Blackwell, Niu, Buck, Guilderson and Heaton2013). Ages are given as calibrated calendar years AD at a 95.4% (2σ) confidence level (Table 1).
Accelerator mass spectrometry 14C and calibrated ages for El Toro Lake from age-depth model using OxCal 4.2 (Bronk Ramsey, Reference Bronk Ramsey2009) and the Southern Hemisphere radiocarbon calibration data SHCal13 (Hogg et al., Reference Hogg, Hua, Blackwell, Niu, Buck, Guilderson and Heaton2013). pMC, percent of modern carbon.
Mineralogy
X-ray diffraction (XRD) analyses were carried out on finely ground sample material (<20 μm), measured with a PANalytical X’Pert PRO diffractometer, with Cu lamp (kα=1.5403 Å) operated at 40 mÅ and 40 kV at the Centro de Investigaciones Geológicas (La Plata, Argentina). Twenty selected layers were measured along the core. The samples were measured from 2° to 40° 2θ, with a scan speed of 0.04°/s and a time per step of 0.50 s. Whole rock samples were powdered in agate mortar, and data were semiquantified by our own patterns and international empirical factors (Moore and Reynolds, Reference Moore and Reynolds1989). Mineralogical composition throughout the core was analyzed to identify possible input variations from the catchment, as a response to changes in conditions affecting the basin.
Geochemical analyses
These analyses were performed on each subsampled layer (1 cm thick). OM content was determined using 0.5 g of each of the sediment samples by loss on ignition at 550°C for 4 h (Heiri et al., Reference Heiri, Lotter and Lemcke2001). Approximately 20 mg of each of the sediment samples was leached in 1% Na2CO3, and aliquots extracted after 3, 4, and 5 h were analyzed for dissolved silica using a wet-alkali digestion technique and measured with a spectrophotometer (GENESYS 20 Thermo Spectronic). A least squares regression analysis was applied on the increase in Si extracted versus time and extrapolation to the intercept allowed estimation of biogenic silica (BSi) concentration (DeMaster, Reference DeMaster1981). Elemental concentrations for aluminum (Al), calcium (Ca), and rubidium (Rb) were obtained throughout the core for every subsample by instrumental neutron activation analysis following the methods described by Ribeiro Guevara et al. (Reference Ribeiro Guevara, Rizzo, Sánchez and Arribére2005). Sediment samples (approximately 100 mg) were irradiated in the RA-6 nuclear reactor (Bariloche, Argentina) in plastic vials. Five gamma-ray spectra were collected in a coaxial HPGe detector. The absolute parametric method was used to determine the elemental concentrations. Analytical errors depend on the nuclear parameters of each element, irradiation conditions, and composition of the sample varying between 4 and 10%. Certified reference materials (i.e., National Institute of Standards and Technology Buffalo River Sediment and International Atomic Energy Agency SL1 Lake Sediment) were analyzed together with the samples for analytical quality control.
Tephras
Tephra layers were not clearly observed by visual inspection, being first identified in the sediment sequence by changes in sediment dry density and OM concentration. After a detailed inspection of the subsampled material at 1 cm resolution under binocular magnification (10× to 50×), some of the changes in dry density and OM were confirmed as volcanic layers because of the presence of glass particles immersed in a matrix of background lake sediment. XRD analyses supported the visual identification. For a detailed morphological characterization, selected volcanic particles, mainly glass shards, were mounted on a holder and coated with Au for inspection with a Philips 515 scanning electron microscope (SEM) at an acceleration voltage of 20 kV (Centro Atómico Bariloche). Compositional information was obtained by spectrometry dispersive energy coupled to the SEM.
Biological proxies: ostracods and chironomids
Ostracods were analyzed at 1 cm intervals throughout the core. Subsamples of about 1 g of dried sediments were sieved through a mesh with 63 µm pore size, rinsed with deionized water, and air-dried. Ostracods valves, both adults and juvenile stages, were picked and counted under a stereoscopic microscope. Adult abundance was expressed as number of valves per gram of dry weight. Species were identified following the methods of Cusminsky and Whatley (Reference Cusminsky and Whatley1996), Meisch (Reference Meisch2000), and Cusminsky et al. (Reference Cusminsky, Pérez, Schwalb and Whatley2005).
Subsamples from each 1-cm-thick layer were prepared using the methods described by Walker (Reference Walker2001) for chironomid analyses. Subsamples of 1–2 g dry weight sediment were deflocculated in 5% KOH, heated to 70°C for 20 min, and subsequently sieved through 95 µm mesh. Head capsules were picked from the wet residue in a Bogorov sorting tray under a stereomicroscope at 20× to 40× magnification (Olympus SZ4060). Larval head capsules were mounted in Euparal and then identified under stereoscopic microscope (100× to 400× magnification) following Wiederholm (Reference Wiederholm1983), Coffmann and Ferrington (Reference Coffman and Ferrington1996), Cranston (Electronic guide to the chironomidae of Australia, http://apes.skullisland.info/node/3), and Epler (Reference Epler2001).
Data analyses
Proxy diagrams were plotted using TILIA 1.7.16 and TILIAGRAPH 2.0.2 software (Grimm, Reference Grimm1991). To distinguish different assemblages along the sequence, a stratigraphically constrained cluster analysis (CONISS) was applied to the percentage values of ostracod and chironomid assemblages (Grimm, Reference Grimm1987). Zonation significance levels were evaluated with a one-way analysis of similarities (ANOSIM), with zones as factors using the program PRIMER-E v. 6.12 (Clarke and Gorley, Reference Clarke and Gorley2005). This analysis tests for differences among factors using permutation and randomization methods based on the similarity matrix Bray-Curtis (Clarke and Warwick, Reference Clarke and Warwick2001).
To define the autoecology of each ostracod species identified in the El Toro sediments, the species optima (weighted average [WA]) and tolerance (standard deviation) of each taxa were calculated based on data from Cusminsky et al. (Reference Cusminsky, Pérez, Schwalb and Whatley2005), Ramón Mercau et al. (Reference Ramón Mercau, Laprida, Massaferro, Rogora, Tartari and Maidana2012), Ramos et al. (Reference Ramos, Alperin, Pérez, Coviaga, Schwalb and Cusminsky2015), and Coviaga (Reference Coviaga2016) (Table 2). However, the current lack of modern autecological data inhibits the possibility of calculating optima salinity concentrations of identified chironomids in Patagonian steppe systems. Chironomid information can be interpreted in a meaningful in this work only by comparison with other paleolimnological data (e.g., ostracods). Therefore, optima salinity (WA) of the main chironomid taxa was obtained from other lakes worldwide (Walker et al., Reference Walker, Wilson and Smol1995; Eggermont et al., Reference Eggermont, Heiri and Verschuren2006; Zhang et al., Reference Zhang, Jones, Bedford, Langdon and Tang2007) and from Polypedilum from Los Juncos Lake in the Patagonian steppe (Modenutti et al., Reference Modenutti, Diéguez and Segers1998; Fuentes and Donato, Reference Fuentes and Donato2014), from which we estimate ranges (Table 2).
Species optima (weighted average [WA]) and tolerance (standard deviation [SD]) for host waters of ostracods identified in El Toro sequence. Maximum and minimum values are in parentheses. Sources: Cusminsky et al. (Reference Cusminsky, Pérez, Schwalb and Whatley2005), Ramón Mercau et al. (Reference Ramón Mercau, Laprida, Massaferro, Rogora, Tartari and Maidana2012), Ramos et al. (Reference Ramos, Alperin, Pérez, Coviaga, Schwalb and Cusminsky2015), and Coviaga (Reference Coviaga2016). Optima salinity (WA) of chironomid main taxa from other lakes worldwide (Walker et al., Reference Walker, Wilson and Smol1995; Eggermont et al., Reference Eggermont, Heiri and Verschuren2006; Zhang et al., Reference Zhang, Jones, Bedford, Langdon and Tang2007) and regional Polypedilum data from Lake Los Juncos (Patagonian steppe; Modenutti et al., Reference Modenutti, Diéguez and Segers1998; Fuentes and Donato, Reference Fuentes and Donato2014) were used for ranges estimation. The salinity tendency is expressed with a dotted line, corresponding to values measured worldwide. Ostracods are ordered following increasing salinity conditions from left to right, whereas chironomids show a slight decreasing salinity trend in the same direction, always in the salinity conditions range of Limnocythere patagonica, present in more diluted waters. TDS, total dissolved solids.
a Modenutti et al. (Reference Modenutti, Diéguez and Segers1998).
b Daga, R., personal observation, 2011.
c Calculated following Zhang et al. (Reference Zhang, Jones, Bedford, Langdon and Tang2007).
RESULTS
Chronology
The 210Pb and 137Cs dating was performed by evaluation of the specific activity in the upper layers. Unsupported 210Pb specific activity does not exhibit an exponential profile allowing the sequence dating (Fig. 2); therefore, 210Pb dating is not considered for analysis. The 137Cs specific activity profile shows a sharp decrease below a depth of 10 cm (4 g/cm2; Fig. 2). According to the 137Cs deposition sequence in the region, corresponding to the South Pacific nuclear tests that started in 1966 with contributions of Northern Hemisphere emissions with highest values in 1964 (Ribeiro Guevara and Arribére, Reference Ribeiro Guevara and Arribére2002), the mean sediment rate is estimated to be 10.5 mg/cm2/yr (2.56 mm/yr) between AD 1963 and 2002.
Specific activity profiles of 137Cs, 210Pb, and 226Ra (in secular equilibrium with supported 210Pb).
Radiocarbon ages obtained from ostracod shells recovered at depths of 30 and 25 cm and the modern samples are reported in Table 1. To discard a hard-water effect, ostracod shells and the fruit of the aquatic plant Zannichellia palustris (horned pondweed), collected in September 2011 and October 2015, respectively, were sampled from the lake sediments. The 14C dating from the modern ostracods and the aquatic plant samples show 105.30±0.30 and 118.59±0.71 pMC (percent of modern carbon), respectively, and no evidence of old carbon. Furthermore, no records of exposed carbonate rocks are reported in the basin, excluding therefore, a hard-water effect on the ostracod radiocarbon ages. Furthermore, no hard-water effect has been observed in other steppe lakes in Patagonia (Haberzettl et al., Reference Haberzettl, Fey, Lücke, Maidana, Mayr, Ohlendorf, Schäbitz, Schleser, Wille and Zolitschka2005; Cartwright et al., Reference Cartwright, Quade, Stine, Adams, Broecker and Cheng2011; Ohlendorf et al., Reference Ohlendorf, Fey, Massaferro, Haberzettl, Laprida, Lücke and Maidana2014).
For the development of a continuous record, the integration of 137Cs and 14C ages on an age-depth model is needed, applying in this case the Bayesian model OxCal 4.2 (Bronk Ramsey, Reference Bronk Ramsey2009) with the Southern Hemisphere calibration data SHCal13 (Hogg et al., Reference Hogg, Hua, Blackwell, Niu, Buck, Guilderson and Heaton2013) to convert radiocarbon ages into calendar ages. The resultant model calculates the age of a query depth with the resultant 68.2 and 95.4% age probability ranges (1σ and 2σ), also allowing extrapolation of the dating to the bottom of the sequence to obtain the time spanned by the core (Fig. 3a). According to the age-depth model, sedimentation rates showed an increasing tendency, from 10 mg/cm2/yr (0.27 mm/yr) in the lower levels to 105 mg/cm2/yr (2.56 mm/yr) in recent times.
(color online) Age-depth model using OxCal 4.2 (Bronk Ramsey, Reference Bronk Ramsey2009) and Southern Hemisphere radiocarbon calibration data (SHCal13; Hogg et al., Reference Hogg, Hua, Blackwell, Niu, Buck, Guilderson and Heaton2013). (a) Age-depth model considering only the radiocarbon ages and 137Cs dating. (b) Corrected age-depth model including a boundary at 14 cm depth representing the most probable position where changes in the sedimentation conditions occurred, considering the change in lithogenic elements Al and Rb (see Fig. 2). Radiocarbon ages are represented as R_Date with the lab code, and calendar ages are shown as C_Date, like surface (date of core extraction) and 137Cs dating. Bottom represents the extrapolated age of the base of the sequence according the age-depth model. The agreement indices for the model (Amodel) and individual dates (A) are shown for both models, resulting in better agreement in panel b.
However, the age-depth model does not allow at this point analyzing the approximate moment of the change in the sedimentation rate. In this regard, the concentration profiles of lithogenic indicators Al and Rb, usually interpreted to be of detrital origin (Boës et al., 2011), were analyzed and compared. After a period of stable behavior between depths of 29 and 20 cm, the concentration of Al and Rb increases (Fig. 4). The strong increase–decrease from depths of 20 to 14 cm could be affected by the dispersed tephra TO17 (16–18 cm depth), but the increasing trend from 14 cm up to the present is clear (Fig. 4), not recovering the pre–20 cm background values at any point. Other minor Al and Rb peaks are related to the other tephras identified. This change in the baseline of Al and Rb concentrations, as well as other geochemical tracers (Ce, Cr, Eu, Fe, Hf, La, Lu, Sc, and Yb, among others not shown here), could be marking that the change in sedimentation conditions had started before AD 1963 (determined by 137Cs dating), probably associated with changes in erosion processes in the catchment. Considering this, an extra element in the Bayesian model was introduced by fixing a boundary at 14 cm depth, where the change in lithogenic elements was observed. This is inferred as the more reasonable depth where a change in the sedimentation conditions took place. Testing for model consistency, the agreement index for the model (Amodel) is 97.2% (far above the acceptance threshold of 60%) (Fig. 3), and convergence values (C) are above the suggested 95%. As result, the age-depth model shown in Figure 3b was adopted as the most representative for El Toro Lake.
(color online) Schematic sedimentary sequence from El Toro Lake. From left to right: lithologic units, sedimentary dry density, organic matter, biogenic silica (mg/g), calcium (Ca), aluminum (Al), rubidium (Rb), and mineralogical profile (Am, amphibole; Amorph, amorphous; Cal, calcite; Cl, clay; Fsp, feldspar; Px, pyroxene; Qz, quartz).
Lithology and geochemical features
The core sequence consisted of two different lithological sections, unit A from the base of the core to a depth of 10 cm and Unit B from a depth of 10 cm to the top (Fig. 4). Unit A (32.5–10 cm) is composed of sandy-silt sediment, with coarse stratification varying between light and dark gray, an abundance of ostracods, and scattered plant remains. Sedimentary dry density and OM exhibited a negative relationship. The bulk mineralogy is represented by calcite, feldspar (including plagioclase and K-feldspar), and quartz in variable proportions, with low clay development (Fig. 4). Calcite is associated with ostracods in the sediments (calcite follows the ostracod profile); therefore, feldspar and quartz represent the 60–90% of the terrigenous fraction without strong variations throughout the core. Following the ostracod decrease, calcite showed a clear decline revealing conditions that did not allow growing and/or preservation of ostracod shells in the sediment. Highest dry density and lowest OM values correspond to zones with high volcanic particle concentration, represented in the mineralogical profile by amorphous material (volcanic glass), dispersed at 32–30, 18–16, and 13–12.5 cm depths, together with pyroxene and amphibole as accessory phases, representing tephra layers. Between 30 and 20 cm levels, sedimentary dry density decreased until the minimum at 25–26 cm, whereas OM concentration exhibited a marked increase reaching a maximum of 21% at 20 cm. BSi concentration remains low (mean value of 2.0±0.5 mg/g), and % Ca relatively stable at about 10% throughout this section. Unit B (10–0 cm) consists of abundant vegetation remains at the base (10–8 cm), continuing with homogeneous dark grayish-brown clayed material, increasing fine sandy sediment fraction up to core. The mean sedimentary dry density was low and tended to increase in the upper 5–0 cm. OM concentration decreased from the bottom to the top of this section (Fig. 4). After a minimum at a depth of 9–10 cm (0.6 mg/g), BSi concentration notably increases to the top (mean concentration of 4.0±1.6 mg/g), showing a maximum value at 5 cm depth (4.6 mg/g); at the same position, % Ca showed a slight decrease to the top of unit B. Bulk mineralogy does not vary with depth, observing a slight increase of the terrigenous fraction (feldspar + quartz) from a depth of 10 cm up to the core. Amorphous material (volcanic glass) and amphibole at 6–8 cm depth represents another dispersed tephra layer.
Tephra layers
Four zones were identified in the core with increasing glassy shards abundance compared with background sediment. Considering the relative distance between El Toro Lake and the volcanic arc (>150 km), these four zones could be associated with volcanic ash fall layers from widely dispersed explosive eruptions from the Southern Volcanic Zone. Even though increases in dry density and decreases in OM were associated with the tephra layers, they were not clearly visually identified; only a slight darker color was observed during core description, which could be considered as cryptotephras (Lowe, Reference Lowe2011). Furthermore, the mineralogical analyses of the core also provided evidence of the presence of amorphous material, related to volcanic glass, at the same depths. Considering this, four tephra layers named TO31, TO17, TO13, and TO8 (labeled according their depth location; Fig. 4) were separated for further analyses.
Tephra TO31 consists of brown and colorless highly vesicular glass dispersed at a depth of 32–30 cm together with pyroxene and amphibole. Brown glassy shards, more abundant than colorless shards, range from low to highly vesicular particles, with both spherical and elongated fluidal vesicles. Colorless to whitish shards are highly vesicular with similar morphology as brown particles. The individual analyses of selected brown and colorless particles have highly evolved composition, with SiO2 ranging from 71 to 73%, together with a minor glass population with 57–60% of silica. The dispersed particles in 2–3 cm can reflect the wind effect in reworking and transport of volcanic particles from the catchment.
Tephra TO17 is composed mainly of white pumice particles and brown glass shards, also with presence of lithics and crystal fragments (amphibole and pyroxene; possibly from the catchment), dispersed at a depth of 18–16 cm. White pumice fragments are highly vesicular, with both vesicular and fluidal textures. Brown glasses are moderately vesicular. Both particles showed evolved compositions (70–73% of SiO2).
Even though the mineralogical profile showed the highest amorphous concentration at TO13 (13–12.5 cm), it was difficult to identify because of its fine grain size. Volcanic components correspond to fine glassy particles with moderate vesicularity and a minor presence of colorless to whitish fragments. Both kinds of particles showed different composition, with 59–61% SiO2 for the former and 71–72% SiO2 for the latter.
In the same way, TO8 is a layer that has a low abundance of volcanic particles at a depth of 8–7 cm. They are identified as fine, pale-brown glass shards dispersed in the sediment, which showed variable compositions ranging between 57% and 70% SiO2; therefore, this is considered a possible reworked layer.
Bioproxies
Ostracods density show a decreasing trend, with a mean density of 41±16 individuals (ind)/g at a depth of 32.5–25 cm, declining to a mean density of 8.4±8.2 ind/g to the core top (Fig. 5). Four species were identified: Eucypris fontana (Graf, 1931); E. virgata Cusminsky and Whatley, Reference Cusminsky and Whatley1996; Limnocythere rionegroensis Cusminsky and Whatley, Reference Cusminsky and Whatley1996 (variety 1 [var. 1]; Cusminsky et al., Reference Cusminsky, Schwalb, Pérez, Pineda, Viehberg, Whatley, Markgraf, Gilli, Ariztegui and Anselmetti2011); and L. patagonica Cusminsky and Whatley, Reference Cusminsky and Whatley1996. The eucyprids E. fontana and E. virgata were recorded in all samples, whereas L. rionegroensis was absent between depths of 26 and 20 cm, and 13 and 12 cm, and L. patagonica only appeared in samples at depths of 11 and 10 cm. Throughout the core, the assemblages consisted mainly of valves. All samples contained juveniles instars (i.e., earlier and later molt stages), but adult valves were absent in samples at depths of 21, 20, 12, and 1 cm. The material, well preserved with <5% of the valves broken, and the presence of both adults and juveniles of different molt stages, suggests that reworking and size sorting did not occur and that the shells were accumulated in situ. Therefore, it may be considered that the ostracod assemblages represent an autochthonous biocoenoses, reflecting the environmental conditions of the sampling site (Holmes, Reference Holmes2001; Keatings et al., Reference Keatings, Hawkes, Holmes, Flower, Leng, Abu-Zied and Lord2007; Cusminsky et al., Reference Cusminsky, Schwalb, Pérez, Pineda, Viehberg, Whatley, Markgraf, Gilli, Ariztegui and Anselmetti2011).
(color online) Bioproxies (% relative abundance) determined for the El Toro Lake sediment core. Total numbers of ostracods and chironomids (head capsules) per gram of dry sediment are shown as filled curves. Zones were calculated by a CONISS cluster analysis (TILIA v. 1.7.16).
Ostracod species identified in the core vary considerably with respect to their ecological preference (Table 2). E. fontana and E. virgata are euryhaline taxa, with a preference for waters of moderate salinity (species optima=8686 and 2006 µS/cm, respectively). On the other hand, L. rionegroensis and L. patagonica are stenohaline taxa. The former prefers more saline waters (species optima=24,861 µS/cm), whereas L. patagonica is restricted to low salinity (species optima=1072 µS/cm) (Table 2).
Chironomids appear in the upper levels of the core, and their abundance varied from 1 head capsule per gram (hc/g) at a depth of 14 cm to 106 hc/g at a depth of 10 cm, with an average of 43 hc/g. A total of 9 taxa belonging to the subfamilies Chironominae (Chironomini) and Orthocladiinae are present. Chironominae was the most abundant subfamily in the upper 14 cm (91%), and Orthocladiinae registered its maxima abundance at 8 cm (16%). At the generic level, Polypedilum indeterminated (ind.) 1 (83%) was the dominant taxon, followed by Chironomus (6%) and Parapsectrocladius (5%), whereas the rest of the taxa present were relatively rare.
Clustering analysis conducted on discrete biotic groups (i.e., ostracods and chironomids) showed similar results. For this reason, a unique CONISS integrating both biological proxies was presented. This analysis allowed the delimitation of five zones: Z1 (depth of 32.5–29 cm), Z2 (29–20 cm), Z3 (20–14 cm), Z4 (14–10 cm), and Z5 (10–0 cm). The one-way ANOSIM analysis indicated significant differences among the taxa abundances of the zones identified by CONISS (ANOSIM, R global=0.51, P=0.001). Pairwise comparisons between zones were all significant, except the differences among the three lower zones—Z1, Z2, and Z3.
The first zone, Z1 (32.5–29 cm, ~AD 1300–1400), was characterized by elevated ostracod abundance (41±17 ind/g). The assemblage was composed of L. rionegroensis var. 1 and E. fontana with a small contribution of E. virgata (1.8%). In the lower half of the zone, L. rionegroensis var. 1 was the dominant species, whereas in the upper half, it was replaced by E. fontana.
In the next zone, Z2 (29–20 cm, ~AD 1400–1750), the density of ostracods varied greatly, with a high abundance between depths of 29 and 26 cm (45±16 ind/g) and subsequent decreasing abundance toward the upper levels (3.8±3.8 ind/g). E. fontana dominated the ostracod assemblages through all of this zone, with important contributions of E. virgata, mainly at the base and the top of Z2. L. rionegroensis var. 1 was only present at a depth of 28 cm with a very low density (2 ind/g).
The third zone, Z3 (20–14 cm, ~AD 1750–1940), had a low mean ostracod abundance (9±8 ind/g), with a high occurrence at a depth of 18 cm and without adult individuals at 22 cm and 21 cm. The ostracod assemblage consisted principally of E. fontana and L. rionegroensis var. 1, as well as a small proportion of E. virgata. An alternation between L. rionegroensis var. 1 and E. fontana as the dominant species was evidenced along this zone.
The fourth zone, Z4 (14–10 cm, ~AD 1940–1963), had a low ostracod occurrence (6±9 ind/g), with no ostracod adults at a depth of 13 cm. In the lower levels of Z4, ostracod assemblages consisted of E. virgata and L. rionegroensis var. 1, whereas in the upper part, E. fontana was clearly the dominant species, with small contributions of the limnocytherids L. rionegroensis var. 1 and L. patagonica. Chironomids were first recorded in low density in the upper levels represented mainly by Polypedilum ind. 1 and Parapsectrocladius (Orthocladiinae) in the top level, being recorded in this zone with an average density of 3.5±3.8 hc/g.
In the youngest sediment, Z5 (10–0 cm, AD 1963–2002), ostracod abundance remained low (10±9 ind/g). The highest diversity was recorded at a depth of 10 cm, with the presence of the four species. Eucypris fontana and L. rionegroensis var. 1 characterized the ostracod assemblage at the base and the top of the zone, whereas E. virgata was the dominant species in the middle section. In this zone, chironomids were the dominant fauna showing maximum abundance at a depth of 10 cm (106 hc/g) and lowest abundance at a depth of 1 cm (19 hc/g), with an average at intermediate levels of 57±12 hc/g. In this zone, all chironomid taxa determined were recorded, which included Chironominae (Polypedilum ind. 1, Polypedilum ind. 2, Chironomus, and Parachironomus) and Orthocladiinae (Parapsectrocladius, Orthocladius/Cricotopus, and three orthoclads not identified). Taxonomic richness sharply increased at a depth of 10 cm, which is maintained up to the maximum value in 5 cm; from this layer, the richness gradually decreases until the core upper layers. At the top level, a sharp drop in abundance and the disappearance of subfamily Orthocladiinae was observed coincident with the presence of E. fontana and L. rionegroensis var. 1, as mentioned previously.
DISCUSSION
Sedimentation rates
Based on the radiocarbon ages, and according to the age-depth model, a sedimentation rate of 0.27 mm/yr (10 mg/cm2/yr) was calculated for the lower levels of the core. Although with differences in water depth, the rather low sedimentation rate observed for El Toro Lake is comparable to those of deeper semiarid aquatic environments characterized by a cool-temperate dry climate with strong winds, located in the steppe and semidesert low shrublands of southern Patagonia (Massaferro et al., Reference Massaferro, Recasens, Larocque-Tobler, Zolitschka and Maidana2013; Ohlendorf et al., Reference Ohlendorf, Fey, Massaferro, Haberzettl, Laprida, Lücke and Maidana2014; Mayr et al., Reference Mayr, Laprida, Lücke, Martín, Massaferro, Ramón-Mercau and Wissel2015). A remarkable rise in the sedimentation rate of an order of magnitude was calculated (2.56 mm/yr, 105 mg/cm2/yr) for AD 1963–2002, inferring from tracers of detrital mineral material that this change started a few decades earlier. The increment in lake sedimentation rates could be associated with enhanced clastic input, mainly related to more humid conditions. Wetter conditions were reconstructed from instrumental and tree-ring records for AD 1925–1977 (with pronounced drought from 1957 to 1962), linked to pressure changes in the domain of the southeastern Pacific anticyclone (Villalba and Veblen, Reference Villalba and Veblen1997; Villalba et al., Reference Villalba, Grau, Boninsegna, Jacoby and Ripalta1998). Furthermore, the construction of a road in the upper part of the lake catchment (Fig. 1) could have generated availability of erodible material to be easily transported to the lake basin. Although there is no information about the precise moment of the construction of this part of the road, records mention that the construction started during the first decade of the twentieth century, and the road connecting the cities of Bariloche, Piedra del Águila, and Neuquén was built during AD 1935–1943 (http://www.bariloche.org). Increased humidity and available material could have generated the appropriate conditions for sedimentation rate increase in El Toro Lake since the beginning of the twentieth century.
Tephra analyses
The dominant composition of the four tephras corresponded to evolved dacitic-rhyolitic volcanic products generated from highly explosive eruptions of acid magmas from the Southern Volcanic Zone that may have reached this area. Basaltic-andesitic magmas predominate over more silica-rich rocks in the Southern Volcanic Zone, although dacites and rhyolites do occur (Lara et al., Reference Lara, Naranjo and Moreno2004; Stern, Reference Stern2004), limiting the number of possible volcanic sources for the El Toro tephras. One of the largest historical eruptions in the Andes was the plinian eruption in 1932 of Quizapu, more than 500 km to the north, with dispersal across the continent and wide compositional range (52–70% SiO2; Hildreth and Drake, Reference Hildreth and Drake1992; Ruprecht et al., Reference Ruprecht, Bergantz, Cooper and Hildreth2012). Ash reached 40°S, and reported mineralogy included dominant plagioclase, ubiquitous amphibole, and pyroxene as the main phases. However, it is the only eruption reported with extended pyroclastic dispersion, not allowing the correlation of the older tephra layers. Located more than 330 km to the south, Chaitén Volcano had the last explosive rhyolitic eruption in AD 2008, probably reaching the El Toro area (although there were no reports of this). Chaitén Volcano had several similar eruptions during the Holocene (Amigo et al., Reference Amigo, Lara and Smith2013; Lara et al., Reference Lara, Moreno, Amigo, Hoblitt and Pierson2013), but the clear rhyolitic character and presence of typical cristobalite in their products (Daga et al., Reference Daga, Ribeiro Guevara, Poiré and Arribére2014) exclude it as a possible tephra source. The geochemical composition of the tephras also resembles the Cordón Caulle Volcanic Complex (CCVC; located about 150 km to the west) products, with several eruptions during recent centuries. In this regard, a 2 cm tephra was recovered in El Toro Lake after the latest 2011–2012 eruption from CCVC (Daga et al., Reference Daga, Bertrand, Bedert, Ribeiro Guevara and Ghazoui2013), confirming the potential influence of previous CCVC eruptions on the area. Eruptions from this volcanic complex have had a unique abundance of silicic products, relative to other Southern Volcanic Zone volcanoes (Lara et al., Reference Lara, Naranjo and Moreno2004), allowing their wide dispersion. A more basic andesitic population of glasses was measured in some El Toro tephras and observed in some volcanic components from CCVC tephras recovered in lakes more proximal to the volcano (Daga et al., Reference Daga, Ribeiro Guevara, Sánchez and Arribére2006, Reference Daga, Ribeiro Guevara, Sánchez and Arribére2010). The compositional homogeneity among CCVC products during historical events (Gerlach et al., Reference Gerlach, Frey, Moreno-Roa and López-Escobar1988; Daga et al., Reference Daga, Ribeiro Guevara, Sánchez and Arribére2006; Bertrand et al., Reference Bertrand, Daga, Bedert and Fontijn2014) does not allow discrimination between tephras. However, the stratigraphic position in this case could be used to correlate with historical eruptions (possibly at AD 1960, AD 1921–1922; AD 1759?; González-Ferrán, Reference González-Ferrán1995; Petit-Breuilh Sepúlveda, Reference Petit-Breuilh Sepúlveda2004; Lara et al., Reference Lara, Moreno, Naranjo, Matthews and Pérez De Arce2006) because of the agreement with the established age-depth model. Nevertheless, the mineralogical assemblage obtained here for the tephra layers is not the reported for CCVC products, in which pyroxene is a common mineral phase but not amphibole (Gerlach et al., 1988). Consequently, it is not possible to ensure at this point that El Toro tephras correlate with CCVC events, and more information is necessary to elucidate their origin; for this reason, the tephra layers were not included in the age-depth model. However, tephra descriptions and preliminary discussion are presented here with the possibility of using such layers for chronological purposes for this lake when more information becomes available. This might include both the glass and mineral chemistry of tephras and possible volcanic sources, considering the relatively scarce information from several volcanoes from the Southern Volcanic Zone.
The presence of tephra layers does not seem to have affected ostracod and chironomid assemblages, maybe because of tephra layers being too dispersed and mixed with the background sediment to influence the communities in a strong way. Even though an apparent decrease in ostracod abundance is observed above the tephra layers (Fig. 5), the main abundance reductions throughout the core are related to other reasons, as a consequence of the hydrologic dynamics of the lake.
Hydrologic dynamic reconstruction in El Toro Lake and correlations with other environmental records in the northern Patagonian region
El Toro Lake is an endorheic saline basin, hydrographically isolated from the Andes. In this area, evaporation is the dominant process because of the Andes rain barrier effect (i.e., low precipitation) and wind intensity and high solar radiation. However, paleoshorelines observed in the area represent the lake-level variations throughout the Holocene (Fig. 1). Our results, particularly the development of the autochthonous ostracod assemblage throughout the core, show the continuous existence of a lacustrine environment over the past 700 yr. However, the analysis of ostracod and chironomid distribution, BSi concentration, and sedimentary features, inferred to reflect salinity (estimated by total dissolved solids [TDS]) fluctuations, suggested hydrologic changes during the late Holocene. The variation of these parameters can be used to reconstruct the lake-level fluctuations and are associated with environmental parameters among which precipitation-evaporation (P:E) balance seems to be the most important. The ostracod fauna recovered in the core has been previously found in other environments from Patagonian steppe, also with a hydrology dominated by the P:E balance (Schwalb et al., Reference Schwalb, Burns, Cusminsky, Kelts and Markgraf2002; Cusminsky et al., Reference Cusminsky, Schwalb, Pérez, Pineda, Viehberg, Whatley, Markgraf, Gilli, Ariztegui and Anselmetti2011). In El Toro Lake sediments, levels with a predominance of the ostracod L. rionegroensis var. 1 suggest a turbid lake with high-salinity waters, whereas periods with dominance of E. fontana and E. virgata indicate moderate-salinity conditions, probably because of a rise in lake level. L. rionegroensis var. 1 is a stenohaline species, restricted to salinities above 2300 mg/L TDS (Figs. 5 and 6) (Ramón Mercau et al., Reference Ramón Mercau, Laprida, Massaferro, Rogora, Tartari and Maidana2012), with an optima of about 24,861 µS/cm (Table 2). This species prefers sodium- and chlorine-sulfate-dominated waters and had been postulated as a possible indicator of high ionic concentration waters associated with dry climatic conditions (Cusminsky and Whatley, Reference Cusminsky and Whatley1996). Moreover, the presence of a monospecific assemblage of this species in the Laguna Cháltel indicates that the hydrochemistry was driven by evaporative conditions (Ohlendorf et al., Reference Ohlendorf, Fey, Massaferro, Haberzettl, Laprida, Lücke and Maidana2014). On the other hand, E. fontana and E. virgata (Fig. 5) could be considered euryhaline species, capable of living in waters with a wide range of ion concentrations (Table 2) (Cusminsky et al., Reference Cusminsky, Pérez, Schwalb and Whatley2005; Ramón Mercau et al., Reference Ramón Mercau, Laprida, Massaferro, Rogora, Tartari and Maidana2012; Coviaga, Reference Coviaga2016). Nevertheless, both taxa prefer moderate-salinity conditions (optima estimated at about 8686 and 2006 µS/cm, respectively; Table 2). In this way, increase in the abundance of these species could be interpreted as intervals of lower ionic concentration than periods with a high abundance of L. rionegroensis.
(color online) Integration of total ostracods and chironomids and biogenic silica from El Toro Lake, with limits of defined zones. TDS (total dissolved solids; mg/L) was selected as salinity indicator for inference of variations throughout the core. Values for estimations of TDS in each zone and saline limits from Cusminsky et al. (Reference Cusminsky, Pérez, Schwalb and Whatley2005) and Ramón Mercau et al. (Reference Ramón Mercau, Laprida, Massaferro, Rogora, Tartari and Maidana2012). The estimations for each zone were assigned after considering the relative abundances of ostracod species and ostracod-chironomid assemblages. Note that the variation profile is only an inference for salinity tendency along the core rather than a definition of absolute TDS values for each defined zone. However, direct measurements on the lake during 1997 (Tartarotti et al., Reference Tartarotti, Baffico, Temporetti and Zagarese2004) and 2002 (core extraction) reported TDS values in agreement with the paleosalinity inference trend. The direction of lake-level variation (upper axis) is inferred as a consequence of salinity changes. In addition, the El Toro profile is compared with northern Patagonia climatic reconstructions from different proxy records: stream flows from Neuquén River (Mundo et al., Reference Mundo, Masiokas, Villalba, Morales, Neukom, Le Quesne, Urrutia and Lara2012), terrigenous particles mass accumulation rate from Puyehue Lake (Bertrand et al., Reference Bertrand, Boës, Castiaux, Charlet, Urrutia, Espinoza, Lepoint, Charlier and Fagel2005) (note that higher terrigenous particles at 1920–1960 was not related to wet conditions but to high sediment availability), and aridity index and number of trees (Villalba and Veblen, Reference Villalba and Veblen1997).
The absence of adult valves at depths of 21, 20, 12, and 1 cm probably was because of unfavorable ecological conditions for ostracod development. Environmental conditions may have changed before the ostracods could reach maturity, and therefore, the absence of adults may represent a population that did not survive before reaching the reproductive stage (De Deckker, Reference De Deckker2002). The presence of juvenile carapaces in these levels also suggests a rapid change of environmental parameters, becoming unfavorable for ostracod development and leading to a massive death of the juveniles. Likewise, the absence of adults at depths of 21–20 cm and 12 cm agrees with a marked alternation between L. rionegroensis and E. fontana as dominant species, supporting the hypothesis of a notable change in environmental conditions. Both taxa exhibited quite different ecological preferences, so the lack of adult valves could indicate a transition from low to high (21 and 20 cm, E. fontana replaced by L. rionegroensis) and high to low (12 cm, L. rionegroensis replaced by E. fontana) salinity conditions. In the same way, the absence of adults at a depth of 1 cm depth matches with the marked decrease in chironomid abundance and a rise in L. rionegroensis density, also suggesting strong environmental modifications, particularly an increase in the water salinity.
Most chironomids are adapted to freshwater habitats, but some tolerate a wide range of salinities and have also been used to quantify lake salinity (Walker et al., Reference Walker, Wilson and Smol1995; Verschuren et al., Reference Verschuren, Cumming and Laird2004). However, larval numbers have been shown to decline with increasing salinity (Heinrichs and Walker, Reference Heinrichs and Walker2006), and chironomids disappeared completely from polysaline to hypersaline lakes in Africa (Verschuren and Eggermont, Reference Verschuren and Eggermont2006), also evidenced in eastern Bolivian Andean lakes (Williams et al., Reference Williams, Brooks and Gosling2012). Scarce information exists about the chironomid fauna in environments from Patagonian steppe, but a recent work from Laguna Cháltel, located southernmost in the Patagonian steppe, suggests that Polypedilum is a chironomid that resists heavy drought, being present even during ephemeral periods (Ohlendorf et al., Reference Ohlendorf, Fey, Massaferro, Haberzettl, Laprida, Lücke and Maidana2014). Moreover, recent new records of Polypedilum (Tripodura) quinquesetosum (Edwards) extend its distribution toward the Patagonian steppe, being observed in Los Juncos Lake, a nonpermanent system (Fuentes and Donato, Reference Fuentes and Donato2014). In summer 2011, a maximum depth of 50 cm and 835 µS/cm (~500 mg/L) were recorded in this lake (Daga, R., personal observation). Back in the austral spring of 1996 and 1997, this closed and temporary pool recorded a surface of 2 ha and maximum depth of 1 m, with pH 7.8 and conductivity around 500 µS/cm (~300 mg/L) (Modenutti et al., Reference Modenutti, Diéguez and Segers1998). Comparing the taxa identified in El Toro Lake and the optimal salinity found in other lakes in the world, it was observed that Polypedilum has higher optimal salinities (e.g., about 573 mg/L in Canadian lakes and 1428 mg/L in Tibetan lakes; Walker et al., Reference Walker, Wilson and Smol1995; Zhang et al., Reference Zhang, Jones, Bedford, Langdon and Tang2007) than Parachironomus (e.g., about 312 mg/L in Canadian lakes and 203 mg/L in African lakes; Walker et al., Reference Walker, Wilson and Smol1995; Eggermont et al., Reference Eggermont, Heiri and Verschuren2006). Cricotopus/Orthocladius is common but never dominant in lakes with intermediate salinities (150 at 6000 mg/L) (Walker et al., Reference Walker, Wilson and Smol1995).
We correlated Chironomidae genera from Patagonia with the same genera from other regions worldwide, using them to estimate salinity ranges (Table 2), compared with ostracod ranges whose optima in Patagonia are known. As we detailed in the “Methodology” section, the chironomid information can be interpreted in a meaningful way only by comparison with other paleolimnological data (e.g., ostracods), considering that local/regional salinity estimates may differ in the ranges used here.
In addition, BSi is a useful proxy to salinity reconstructions. The BSi content of sediment represents the siliceous skeletal matter from the epilimnion, minus the dissolution that occurs during settling and on the lake floor (Cohen, Reference Cohen2003). BSi concentration would be controlled by ionic concentration as was demonstrated by Ryves et al. (Reference Ryves, Battarbee, Juggins, Fritz and Anderson2006), who proved that dissolution increases with salinity and is independent of pH and water depth. In this regard, we suggest that in El Toro Lake, salinity would be a significant factor controlling the dissolution of BSi, supporting the low BSi content throughout the core, especially below a depth of 10 cm.
The late Holocene climate history inferred from our record of El Toro Lake is discussed in light of other paleoclimate records from the northern Patagonian region.
-
∙ Zone 1 (~AD 1300–1400 yr). The unit was characterized by the highest ostracod abundance, suggesting favorable environmental conditions for their development and preservation. In the initial levels, L. rionegroensis var. 1 dominated the ostracod assemblages. This species inhabits high-evaporation environments, and their presence would indicate a period characterized by a low water level and elevated ionic concentrations, associated with a negative P:E balance (Fig. 6). The low BSi concentration probably also suggests high-salinity conditions (Lent and Lyons, Reference Lent and Lyons2001; Ryves et al., Reference Ryves, Juggins, Fritz and Battarbee2001). From the ostracod optima (WA) estimated (Table 2) could be inferred more saline conditions during this period, together with lower lake level (Fig. 6). Our results partially overlap with the evidence of a low-precipitation period registered at similar latitudes, on the west Andean side, by low accumulation of terrigenous particles prior to AD 1490 and thin varves at ~AD 1400 (Bertrand et al., Reference Bertrand, Boës, Castiaux, Charlet, Urrutia, Espinoza, Lepoint, Charlier and Fagel2005; Böes and Fagel, 2008). Moreover, on the east side, extreme dry years around AD 1380 were reconstructed from the stream flow values in the Neuquén River (38°S) (Mundo et al., Reference Mundo, Masiokas, Villalba, Morales, Neukom, Le Quesne, Urrutia and Lara2012) (Fig. 6).
-
∙ Zone 2 (~AD 1400–1750). A decrease in sedimentary density coupled with a rise in OM content of sediments was recorded, probably because of the return to normal lake conditions after the tephra deposition (Fig. 4). The mean sedimentary BSi concentration was higher than the previous zone and could be interpreted as lower-salinity conditions (Ryves et al., Reference Ryves, Battarbee, Juggins, Fritz and Anderson2006). This period was characterized by an assemblage dominated by the euryhaline species E. fontana and E. virgata, also suggesting a lower-salinity concentration. Altogether, the proxies indicate that El Toro Lake went through a phase of higher-level and more diluted waters, associated with humid and cold (following regional records) weather conditions (Fig. 6), likely the consequence of a strengthening of the midlatitude westerlies (Jenny et al., Reference Jenny, Valero-Garcés, Urrutia, Kelts, Veit, Appleby and Geyh2002). Around the middle of the seventeenth century, humid and colder conditions were recorded in the region, and although the precise timing of the LIA in Patagonia is still a matter of debate (Villalba, Reference Villalba1994; Ariztegui et al., Reference Ariztegui, Bösch and Davaud2007), our results match with a second pulse of the LIA. The same regional climate evidence was collected on both sides of the Andes in northern Patagonia. Bertrand et al. (Reference Bertrand, Boës, Castiaux, Charlet, Urrutia, Espinoza, Lepoint, Charlier and Fagel2005) recorded a wet period during AD 1490–1700, resulting in higher catchment erosion and an increasing lake terrigenous supply (measured by mass accumulation rate) in Lake Puyehue (Fig. 6). In the same environment, Boës and Fagel (Reference Boës and Fagel2008) identified an increase in the precipitation from AD 1510 to 1630 based on a varve-thickness index. To the eastern side of the northern Patagonian Andes, Mundo et al. (Reference Mundo, Masiokas, Villalba, Morales, Neukom, Le Quesne, Urrutia and Lara2012) identified higher stream flows in Neuquén River during AD 1584–1593 and 1634–1643 (Fig. 6). Moreover, based on tree-ring reconstructions, a negative summer temperature departure between AD 1520 and 1670 was identified at 41°S associated with cold-humid conditions (Villalba, Reference Villalba1990), and a long cold interval extending from AD 1500 to 1660 was reconstructed by Villalba et al. (Reference Villalba, D’Arrigo, Cook, Jacoby and Wiles2001) in the same region.
-
∙ Zone 3 (~AD 1750–1940). The reappearance of the mesohaline L. rionegroensis var. 1 suggests the return to more saline and lower lake-level conditions (Fig. 6), also suggested by a slight increasing tendency in sedimentary density and a decrease in OM and BSi contents, even previous to the tephra deposition. These trends point to an arid period, which agrees with the dry and cold climate reconstructed by Lara and Villalba (Reference Lara and Villalba1993) and Bertrand et al. (Reference Bertrand, Boës, Castiaux, Charlet, Urrutia, Espinoza, Lepoint, Charlier and Fagel2005) around AD 1700–1900 in northern Patagonia (Fig. 6). Although during this time El Toro Lake had a higher ion concentration than in the previous zone (Fig. 6), the alternation between L. rionegroensis var. 1 and E. fontana reflected fluctuations in P:E balance, probably in response to the summer temperature deviations measured between AD 1700 and 1800 (Lara and Villalba, Reference Lara and Villalba1993). Climatic fluctuations were also inferred for the past 200 yr from tree-ring records in northern Patagonia (Villalba and Veblen, Reference Villalba and Veblen1997) (Fig. 6). Both peaks of L. rionegroensis var. 1 registered in this zone are coincident with the driest periods. The first period, beginning in the nineteenth century (~19 cm; Figs. 5 and 6), coincides with a low tree-ring index (Villalba and Veblen, Reference Villalba and Veblen1997), while the second peak, in the late nineteenth century (~16 cm; Figs. 5 and 6), matches with the driest period that occurred in northeastern Patagonia from AD 1895 to 1919 (Villalba et al., Reference Villalba, Grau, Boninsegna, Jacoby and Ripalta1998). According to stream flow reconstructions done by Mundo et al. (Reference Mundo, Masiokas, Villalba, Morales, Neukom, Le Quesne, Urrutia and Lara2012), the longest dry period in Neuquén River (38°S) occurred between AD 1888 and 1897, supporting our hypothesis of a lake dominated by evaporative conditions. This drought was followed by a cold-wet period around AD 1920, recorded in the east Andean side of northern Patagonia by Lara and Villalba (Reference Lara and Villalba1993), Villalba et al. (Reference Villalba, Grau, Boninsegna, Jacoby and Ripalta1998), and Mundo et al. (Reference Mundo, Masiokas, Villalba, Morales, Neukom, Le Quesne, Urrutia and Lara2012). Notably, the same tendency was observed in the Chilean lake Puyehue between 1920 and 1950 (Böes and Fagel, 2008). During the end of this period, the construction of a road in the upper part of the catchment (Fig. 1) could have generated available erodible material to be transported to the lake by the increased rains, starting the period with increased sedimentation rate.
-
∙ Zone 4 (~AD 1940 to 1963). The lowest mean ostracod density and the start of the occurrence of chironomids characterize this zone. The ostracod assemblage was composed of E. virgata and L. rionegroensis var. 1 at the base of the zone and gradually replaced by E. fontana and L. patagonica in the upper levels. Our results indicated that this zone was a transitional period from a low lake level with a negative hydrologic balance to a deep permanent lake with more diluted water. Villalba and Veblen (Reference Villalba and Veblen1996) observed that the cold-wet conditions from AD 1938 to 1941 are abruptly interrupted by the extremely hot-dry summers of AD 1942 and 1943, covered by the instrumental climate record from AD 1910 to the present. The presence of L. rionegroensis var. 1 around the AD 1940 decade, ~13 cm (Figs. 5 and 6), could identify this hot-dry period in northern Patagonia. After that, the appearance of the freshwater L. patagonica, characteristic of alkaline, bicarbonate-dominated waters with low salinities (Ramón Mercau et al., Reference Ramón Mercau, Laprida, Massaferro, Rogora, Tartari and Maidana2012), agrees with the climatic reports in Mundo et al. (Reference Mundo, Masiokas, Villalba, Morales, Neukom, Le Quesne, Urrutia and Lara2012; Fig. 6). In the same way, the first appearance of the chironomid Polypedilum ind. 1 and isolated levels with the ostracods E. fontana and E virgata might be related to the episodic pulse of meso-oligohaline waters (Ohlendorf et al., Reference Ohlendorf, Fey, Massaferro, Haberzettl, Laprida, Lücke and Maidana2014). We suggest that in the first half of the twentieth century, El Toro Lake rose to its previous water levels because of an increase in the precipitation and less evaporation in the region (Fig. 6).
-
∙ Zone 5 (~AD 1963 to 2002). The limit from Z4 to Z5 is in agreement with the transition between the lithological units, highlighting differences in the limnological conditions of the lake. Moreover, compared with the previous zones, during this period El Toro Lake displayed a notable increase in chironomid abundance coupled with the continuation of a high sedimentation rate. This could be associated with the higher sediment availability in the lake catchment area since the first decades of the twentieth century, linked to increased precipitation resulting in high erosion of the lake watershed.
Consistently, the remarkable increase in BSi concentration could be associated with lower-salinity water. As was previously discussed (see Zone 2), diatom preservation (i.e., BSi deposition) is higher under low-salinity conditions (Flower and Ryves, Reference Flower and Ryves2009). Additionally, increasing species richness and diversity of chironomids in sediments was observed, correlating with the development of less saline conditions (Heinrichs and Walker, Reference Heinrichs and Walker2006). Previous works have suggested that midges respond indirectly to both salinity and water depth, but mainly to habitat distribution, such as substrate type and macrophytes (Zhang et al., Reference Zhang, Jones, Bedford, Langdon and Tang2007; Luoto, Reference Luoto2012). In this way, the change observed in lithology suggests that chironomids could have responded to other environmental variables and not only to a physiological tolerance limit to osmotic stress in the chironomids themselves. The presence of vegetation remains at that depth could have stabilized the lake bottom and prevented further erosion of fine sediments with reduction of inorganic turbidity in the water (maybe conditions absent in the deeper zones), favoring the development of chironomids, not just related to lower-salinity conditions (Wolfram et al., Reference Wolfram, Donabaum, Schargerl and Kowarc1999).
The changes observed in the biota density, the rise in BSi content, together with an increase in the sedimentation rates suggested that El Toro Lake went through a period with more diluted and deeper waters, with the presence of submerged vegetation. The rise in the lake level could be related to the beginning of a wetter period, with elevated precipitation around AD 1970, 1980, and mid-1990s (Masiokas et al., Reference Masiokas, Villalba, Luckman, Lascano, Delgado and Stepanek2008) and the cooling trend identified in the region from AD 1950s to mid-1970s (Lara and Villalba, Reference Lara and Villalba1993; Villalba et al., Reference Villalba, Lara, Boninsegna, Masiokas, Delgado, Aravena, Roig, Schmelter, Wolodarsky and Ripalta2003). Furthermore, the construction of several dams along the Limay River (Fig. 1) that occurred from 1973 could have affected the local climate, specifically through an increase in moisture, as was observed in North America and mainly in Mediterranean and semiarid climates (Degu et al., Reference Degu, Hossain, Nigoyi, Pielke, Shepherd, Voisin and Chronis2011). However, the evidence is not conclusive because temperature and precipitation records from the Alicura Reservoir weather station do not show a clear pattern (Autoridad Interjurisdiccional de las Cuencas de los ríos Limay, Neuquén y Negro), and a warmer tendency from 1976 was observed in the Bariloche area, located 150 km southwest from El Toro (Villalba et al., Reference Villalba, Lara, Boninsegna, Masiokas, Delgado, Aravena, Roig, Schmelter, Wolodarsky and Ripalta2003).
In contrast, central and southern Chile (35–50°S) underwent a general decrease in annual precipitation since AD 1950 (Aravena and Luckman, Reference Aravena and Luckman2009; Garreaud et al., Reference Garreaud, Lopez, Minvielle and Rojas2013). This record was associated with the continued strengthening of the meridional gradient of sea-level pressure, favoring a poleward shift of the westerly winds and associated fronts, related to a persistent trend toward the positive Southern Annular Mode (Quintana and Aceituno, Reference Quintana and Aceituno2012; Moreno et al., Reference Moreno, Vilanova, Villa-Martinez, Garreaud, Rojas and De Pol-Holz2014). Generally, the influence of the westerlies decreases toward the east across the Andes, and the low-level zonal wind in Patagonia is negatively correlated with monthly precipitation to the east of the Andes (Villalba, Reference Villalba1990; Berman et al., Reference Berman, Silvestri and Compagnucci2012). On the other hand, the center-north region of eastern Patagonia has a significant positive correlation with precipitation over the surrounding Atlantic and the center of Argentina, and the enhancement of seasonal precipitation is associated with a weakened westerly flow over the region (Berman et al., Reference Berman, Silvestri and Compagnucci2012). In this case, the southerly contraction of the midlatitude westerlies could increase precipitation over central eastern Patagonia by enhancing moist air mass inflow from the north and the Atlantic (Berman et al., Reference Berman, Silvestri and Compagnucci2012; Agosta et al., Reference Agosta, Compagnucci and Ariztegui2015), possibly accounting for the different precipitation pattern with western Andean records, even though both zones are under the influence of the westerlies. Although more information is needed to reach clearer conclusions, the present work provides independent information in accordance with the climatic records.
Toward the youngest level of Z5, an increase in ostracod occurrence, joined with a reduction in chironomid abundance, and the disappearance of some taxa were recorded, suggesting a rise in the lake salinity, probably because of a new decrease in lake level (Fig. 6). This is in agreement with studies developed in the region on both Andean sides, which have identified low precipitation and warm temperatures in the last decades of the twentieth century (Villalba et al., Reference Villalba, Lara, Boninsegna, Masiokas, Delgado, Aravena, Roig, Schmelter, Wolodarsky and Ripalta2003). Also, the endorheic Cari-Laufquen system, located in the Patagonian steppe, about 200 km southeast of El Toro Lake (Fig. 1), recorded a strong drought period during the 1970s, moisture during 1980s, and drought again in current times in agreement with the precipitation recorded from three meteorological stations located in the area (Departamento Provincial de Aguas, 2012).
CONCLUSIONS
The clustering analysis based on ostracod and chironomid distribution, interpreted as high- or relative low-salinity conditions, allowed the delimitation of the transitions from dry to wet or wet to dry intervals throughout the sequence (Fig. 6). From the bottom to the top of the core, five zones were distinguished: low lake level, elevated salinity conditions, and dry climate period (Z1); lake-level rise, moderate-salinity conditions, and cold climate (Z2); two short alternations of dry-wet periods (Z3); transition from low to higher water level associated with dry and with cold and moist conditions, respectively (Z4); and higher to lower lake level with salinity fluctuations (Z5).
Our observations showed the usefulness of lacustrine ostracods, chironomids, and sedimentary parameters as proxies for hydrologic condition changes. This study provides information that can be used to improve our understanding of paleoclimate in northern Patagonia under a regional-global atmospheric dynamic context. More detailed studies of chironomid community ecology in a wide gradient of salinity values in the Patagonian region are necessary in order to maximize the freshwater-saline threshold response, and other ecological parameters, of these chironomid communities in different environments. A more complete database will allow us a more detailed interpretation of subfossil assemblages in terms of salinity, ionic composition, and other habitat variables, such as permanence of the water body.
ACKNOWLEDGMENTS
The authors wish to express their gratitude to Ricardo Sánchez for his collaboration in sampling and sample conditioning and to the reactor RA6 operation staff for their assistance in sample analysis. We especially thank Silvia Dutrús for support during the analysis of BSi, to the Grupo de Caracterización de Materiales del Centro Atómico Bariloche for the support in SEM analyses and the XRD personnel from Centro de Investigaciones Geológicas (La Plata University). The authors also thank Autoridad Interjuridiccional de las Cuencas de los ríos Limay, Neuquén y Negro for the meteorological information and Mr. Zingoni for allowing us unrestricted access to El Toro Lake. We are also grateful to the associate editor, Peter Langdon, and the reviewers, Jonathan Holmes and two anonymous reviewers, for their valuable comments at the revision stage. This study is a contribution to the projects PICT 2010-0082, PICT 2012-280, PICT 2014-1272, PIP 00819, PIP 00021 587/12, and Comahue University Project B166.
