Quantitative climate modeling based on multi-proxy data of precisely dated speleothems

Speleothem d18O values depend on a complex interplay of various processes occurring in the atmosphere, the soil and karst above the cave as well as during precipitation of speleothem calcite. Thus, the relationship between speleothem d18O values and climate parameters, such as local and supra-regional temperature and precipitation patterns, is in many cases not straightforward. Here we present a set of global simulations for different time periods (Modern Day, LGM, MIS3, …) with a mean duration of 20 years, performed with the EMAC atmospheric chemistry general circulation model, being enhanced with stable water isotope diagnostics. The model resolution comprises of a horizontal grid with a maximum resolution of 0.88° × 0.88° (T106) and a vertical resolution of 31 model levels (top level at 0.01 hPa). The simulated d18O values in precipitation were converted into synthetic speleothem d18O records by accounting for the effects of evapotranspiration and infiltration using model-internal empirical evapotranspiration submodels as well as the drip&#2;water-to-calcite fractionation equation by Tremaine et al. (2011). We then calculated mean speleothem d18O values on an annual and seasonal basis, weighted by the amount of precipitation and infiltration (pw-d 18O, iw-d 18O), respectively. The synthetic speleothem d18O records were compared with speleothem d18O records from the Speleothem Isotopes Synthesis and Analysis Working group (SISAL) database. In general, the model setup is capable of adequately simulating the global pattern of d18O values in precipitation. The synthetic speleothem d18O records clearly benefit from weighting by the amount of precipitation and, partly, even more from weighting by infiltration (i.e., the application of an evapotranspiration model). This is obvious by the better agreement of iw-d18O speleothem records with the SISAL data for regions with an infiltration-to-precipitation ratio <0.7 (e.g., the Mediterranean Region). In temperate regions, such as Central Europe, the iw-d18O is approximately similar to the corresponding pw-d18O. Thus, infiltration in these regions seem to be less influenced by evapotranspiration and surface runoff. The evapotranspiration model but limitations when calculating evapotranspiration in extremely arid and/or cold areas, possibly resulting in biased iw-d18O values. Commonly, the climate of these regions is too harsh for speleothem growth, and therefore currently not subject of this study.