Obtaining the correct sea surface temperature: bias correction of regional climate model data for the Mediterranean Sea

Here, the COSMO Climate Limited-area Modelling (CCLM) regional climate model (RCM) is used as external forcing for a Mediterranean basin-scale ocean model based on the general estuarine transport model (GETM). CCLM is forced by different global climate models (GCMs) (MPI and EcEarth) and by the ERA-interim (ERAin) reanalysis. Sea surface temperatures (SST) simulated by the different runs of the ocean model are compared with satellite measurements. As a substantial cold bias in simulated SST is found, a simple bias-correction methodology is applied to the RCM atmospheric variables, namely (1) air temperature (which is usually underestimated for the region by RCMs runs), (2) cloud cover (typically overestimated) and (3) wind intensity (as zonal wind intensity is usually overestimated). The performed analysis identifies wind velocity as the most important variable to correct in order to satisfactorily obtain Mediterranean SST. For many of the RCM realizations significant improvement in the simulated SST are only achieved when wind intensity values were bias-corrected towards observed values. Air temperature and cloud cover had a more marginal importance in reducing the SST bias observed in RCM-forced runs of the ocean model. By comparing the ERAin-driven run and the GCMs driven runs, our analyses suggest that the main source of observed bias is related with the GCMs being used as lateral boundary condition for the RCM realizations. However, a relative simple bias-correction methodology is sufficient to reduce a large part of the induced bias in SST and to improve the vertical water stratification characteristics within the Mediterranean basin that will allow to analyze current and future biogeochemical conditions of the studied basin.