How well do the CMIP6 HighResMIP models simulate precipitation over the Tibetan Plateau?

Abstract

The simulation of precipitation over the Tibetan Plateau (TP) is still a challenge in climate modeling community, although there have been a number of improvements in recent global climate models. In this study, we evaluate the differences in the simulation of precipitation over the TP using multi–resolution outputs from the Coupled Model Intercomparison Project Phase 6 (CMIP6) High Resolution Model Intercomparison Project (HighResMIP). Our results show that both the high resolution (HR) and low resolution (LR) models could generally simulate the annual precipitation well over the TP during the period of 1998–2014. With their finer resolution, the HR models, give some improvements in the simulation of annual precipitation, showing a significantly reduced “wet bias” over the TP compared with the LR models. In addition, the increased resolution performs a better simulation of the precipitation extremes, especially over the southern and southeastern TP, revealing the added value of HR models. We further investigate the differences in the simulation of precipitation between the LR and HR models using a linearized moisture budget equation, which suggests that the underlying differences can be largely explained by the contribution of the dynamic effect (δDY). This effect, dominated by the differences in the simulation of the atmospheric circulation, could induce local moisture convergence, generating excessive precipitation in the LR models compared with the HR models. This implies that the large–scale circulation and moisture conditions can be resolved more realistically in HR models, which is an important factor in determining the simulation of precipitation over the TP.