Spatial Variations in Seamless Predictability of Subseasonal Precipitation Over Asian Summer Monsoon Region in S2S Models

Abstract

AbstractThis study assesses the seamless predictability of subseasonal precipitation over the Asian summer monsoon (ASM) region. The prediction skill of 12 models from the subseasonal‐to‐seasonal (S2S) hindcast database varies considerably, suggesting a large uncertainty in the prediction of ASM precipitation. However, all models show that ASM precipitation is better predicted over ocean than over land, and less reliable in subtropical areas than in tropical areas. The results reveal significant spatial variations in the prediction skill and predictability of ASM precipitation. This study investigates the factors controlling these spatial variations by analyzing the area‐averaged precipitation predictability in three different subregions: the maritime continent (MC), the Indian summer monsoon (ISM) subregion, and the East ASM (EASM) subregion. Precipitation is best predicted over the MC, followed by the ISM and EASM. The distinct disparities in prediction skill among subregions are mainly controlled by differences in potential predictability; that is, the intrinsic limits of predictability among subregions. The high potential predictability of the MC is attributable to small noise due to a strong correlation with the El Niño–Southern Oscillation, whereas the low potential predictability of the EASM is attributable to its small signal. Meanwhile, the comparable signal and noise related to the intraseasonal oscillation in the ISM precipitation, limits the predictability of this subregion. Finally, the evaluation of the ASM precipitation ensemble forecasts in the S2S models demonstrates that the ensemble systems are underdispersive in the three subregions in most S2S models.