Abstract

This paper evaluates the performance of a coupled general circulation model FGOALS¡s1.1 developed by LASG/IAP in simulating the annual modes of tropical precipitation. To understand the impacts of air-sea coupling on the annual modes, the result of an o®-line simulation of the atmospheric component of FGOALS¡s1.1, i.e., LASG/IAP atmospheric general circulation model SAMIL, is also analyzed.

FGOALS¡s1.1 can reasonably reproduce major characteristics of the annual mean precipitation. Nonetheless, the coupled model shows overestimation of precipitation over the equatorial Pacifc and tropical South Pacifc, and underestimation of precipitation over the northern equatorial Paci¯c. The monsoon mode simulated by FGOALS¡s1.1 shows an equatorial anti-symmetric structure, which is consistent with the observation. The bias of the coupled model in simulating monsoon mode resembles that of SAMIL, especially over the subtropics. The main de¯ciency of FGOALS¡s1.1 is its failure in simulating the spring-fall asymmetric mode. This is attributed to the false phase of sea surface temperature anomaly (SSTA) annual cycle over the equatorial central-eastern Paci¯c and Indian Ocean, which leads to the bias of the Walker circulation over the equatorial Paci¯c and the anti-Walker circulation over the Indian Ocean in boreal spring and fall. In addition, the domains of the western North Paci¯c monsoon and Indian monsoon simulated by the coupled model are smaller than the observation. The study suggests that the bias of the fully coupled ocean-atmosphere model can only be partly attributed to the bias of the atmospheric component. The performance of FGOALS¡s1.1 in simulating the annual cycle of equatorial SST deserves further improvement.