Sensitivity of the simulated East Asian summer monsoon climatology to four convective parameterization schemes

Abstract

This study attempts to clarify the effects of parameterized convection on a simulated East Asian summer monsoon climatology from 1997 to 2006 within a regional climate model (RCM) framework. Four different convective parameterized schemes (CPSs) were chosen in the National Centers for the Environmental Prediction regional spectral model, the Simplified Arakawa‐Schubert (SAS), Relaxed Arakawa‐Schubert (RAS), NCAR Community Climate Model version 3 (CCM), and new Kain‐Fritsch (KF2) schemes. On the whole, there is no single scheme that outperforms the other in all aspects of the simulated climatology. The results show that the SAS and KF2 schemes are capable of reproducing the 10‐year June–July–August climatology with a certain wet bias over land and dry bias over the oceans. The SAS scheme is capable of resolving the interannual variations of the monsoon precipitation, and the KF2 is able to capture the intraseasonal variation. The RAS scheme significantly reduces precipitation over land compared to other schemes, but interannual variabilities are reasonably reproduced. On the contrary, the CCM scheme significantly overestimates the precipitation over land as well as ocean. Discernable differences among four CPSs are found in the distribution of the spatial spectrum, in which the KF2 scheme is the most effective in adding regional details to an external large‐scale forcing. The feasibility of a physics ensemble using different convection schemes and the impacts of domain size are also discussed.