Ostracodes (Ostracoda, Crustacea) are aquatic micro-crustaceans with a significant representation in the fossil record. If the environmental influence on the species composition of their communities is robustly quantified, past changes in ostracode communities reflected in fossil assemblages can be used for paleo-environmental reconstruction. We analyzed ostracode assemblages in recently deposited surface sediments from 56 lakes in western and central Mongolia, and simultaneously recorded local water chemistry and solute concentration in order to elucidate the distribution of individual ostracode species in relation to these broad environmental gradients. Multivariate analysis indicated that the species variation in ostracode assemblages could be mainly attributed to variations in percent calcium (%Ca) relative to total cation content, mean annual precipitation, calcium concentration, alkalinity, percent bicarbonate relative to total anion content, and mean July temperature. This matches well with the results of a similar analysis on presence/absence data of living ostracodes in nearshore samples, even though some differences exist between the faunal composition of both datasets. The documented response of ostracode species to environmental variation tracks the typical solute evolutionary pathway for surface waters in this region, characterized by calcite precipitation and consequent depletion in dissolved calcium. Hence, the best quantitative inference model (WA-PLS model with Rjack2=0.70, RMSEP = 0.40) for paleolimnological application was obtained for %Ca. Comparison between this model and a specific conductance (SC) inference model based on the same dataset, and based on ostracode datasets from different regions, indicated that the %Ca inference model suffers less than the SC inference model from a step-change in reconstructed values. The statistical power of different inference models based on Mongolian ostracodes are variously affected by the common dominance of a single euryhaline species (Limnocythere inopinata), limited faunal turnover in the freshwater portion of the salinity gradient, and the bimodal frequency distribution of SC among regional lakes. The latter probably represents true scarcity of lakes with intermediate salinity rather than a biased representation in our dataset. In a broader context of ostracode ecology, and with respect to regional paleolimnological applications, we highlight the potential of fossil Mongolian ostracode assemblages to trace past hydrological shifts associated with changes in groundwater inflow.
We thank Yondon Khand (Mongolian Academy of Sciences) and Soninkhishig Nergui (National University of Mongolia) for generous support in the logistic planning of the fieldwork. We also thank Brecht De Meulenaer, Bukhchuluun Tsegmid, Zorig and all other campaign participants for practical assistance in the field. James E. Almendinger (Science Museum of Minnesota) provided valuable advice for the sampling and analysis of water chemistry. We thank two anonymous referees for constructive comments. Fieldwork was partly supported by travel grants from FWO-Vlaanderen (Fund for Scientific Research in Flanders), King Leopold-III Fund and VLIR-UOS (Flemish Interuniversity Council - University Development Cooperation). The research was funded by a PhD grant of the Institute for the Promotion of Innovation through Science and Technology in Flanders ( IWT-Vlaanderen ) to the first author.