Applying an isotope-enabled regional climate model over the Greenland ice sheet: effect of spatial resolution on model bias

Abstract. In order to investigate the impact of spatial resolution on discrepancy between simulated δ18O and observed in Greenland ice cores, regional climate simulations are performed with isotope-enabled model (RCM) COSMO_iso. For this purpose, general circulation (GCM) ECHAM5-wiso under present-day conditions MPI-ESM-wiso GCM mid-Holocene dynamically downscaled COSMO_iso for Arctic region. The capability reproduce isotopic ratios cores these two periods is investigated by comparing simulation results measured from snow pit samples, Global Network Isotopes Precipitation (GNIP) stations cores. To our knowledge, first time that a RCM Under conditions, dynamical downscaling (1.1∘×1.1∘) 50 km improves agreement 14 19 observational data sets. A further increase 7 does not yield substantial improvements except coastal areas its complex terrain. mid-Holocene, fully coupled slice km. already agrees well observations improve model–data agreement. Despite lack improvement biases, study shows both periods, values at measurement sites constitute isotope which mainly within subgrid-scale variability global results. correct consequently resolved need be extracted refinement an RCM. context, provide distribution effects local uncertainties can taken into account comparison point measurements outputs. Thus, GCM–RCM chain realistically implemented fractionating processes constitutes useful supplement reconstruct paleo-climate during Greenland. Such chains might also applied reveal full potential GCMs other regions large deviations relative simulated.