Addressing the Issue of Systematic Errors in a Regional Climate Model

Abstract

A methodology is proposed in which a few prognostic variables of a regional climate model (RCM) are strongly constrained at certain wavelengths to what is prescribed from the bias-corrected atmospheric general circulation model (AGCM; driver model) integrations. The goal of this strategy is to reduce the systematic errors in a RCM that mainly arise from two sources: the lateral boundary conditions and the RCM errors. Bias correction (which essentially corrects the climatology) of the forcing from the driving model addresses the former source while constraining the solution of the RCM beyond certain relatively large wavelengths in the regional domain [also termed as scale-selective bias correction (SSBC)] addresses the latter source of systematic errors in RCM. This methodology is applied to experiments over the South American monsoon region. It is found that the combination of bias correction and SSBC on the nested variables of divergence, vorticity, and the log of surface pressure of an RCM yields a major improvement in the simulation of the regional climate variability over South America from interannual to intraseasonal time scales. The basis for such a strategy is derived from a systematic empirical approach that involved over 100 regional seasonal climate integrations.