Impacts of SST Configuration on Monthly Prediction of Western North Pacific Summer Monsoon in Coupled and Uncoupled Models

Zhu Xueyan,Liu Xiangwen,Huang Anning,Yuan Jian,Deng Weitao
DOI: https://doi.org/10.1007/s00382-021-06063-6
IF: 4.901
2022-01-01
Climate Dynamics
Abstract:This study examines the impacts of sea surface temperature (SST) configuration on the monthly prediction of summer monsoon over the western North Pacific (WNP) by conducting several sets of hindcast experiments using the Beijing Climate Center Climate System Model and its atmospheric component model. The results show that the atmosphere-only model exhibits limited skill in predicting the WNP monsoon rainfall and circulation, and this skill can hardly be improved by simply increasing the frequency of prescribed SST observation. Compared to the atmosphere-only model, the coupled model shows much better performance in predicting the WNP monsoon rainfall and circulation, which can be further improved by adopting the observed SST with relatively higher frequency in the model initialization. This indicates that the high frequency of observed SST used is much more important in the coupled model than in the uncoupled model. In addition, the uncoupled model forced by the SST predicted by coupled model tends to produce better prediction of WNP monsoon rainfall and circulation than that forced by the observed SST. Both the coupled model and the atmosphere-only model forced by the coupled model predicted SST can well reproduce the surface latent heat flux and shortwave radiation flux over the WNP, leading to a reasonable SST-monsoon relationship and thus skillful predictions of WNP monsoon. Therefore, although the Tier-1 approach based on coupled model is increasingly popular, the Tier-2 approach based on atmosphere-only model is still feasible for the monthly prediction of WNP summer monsoon despite the lack of air-sea interaction. To obtain more skillful Tier-2 prediction, we recommend seeking for SST forcing that is unrealistic but consistent with the atmospheric model rather than SST forcing with very high accuracy.
What problem does this paper attempt to address?