Improved seasonal predictive skill and enhanced predictability of the Asian summer monsoon rainfall following ENSO events in NCEP CFSv2 hindcasts

Abstract

The dominant modes of the Asian summer monsoon (ASM) rainfall variability, as well as their seasonal predictive skill and predictability, are investigated using two sets of seasonal hindcasts made with the NCEP Climate Forecast System (CFSv2): one from the NCEP CFS Reanalysis and Reforecast Project (CFS_RR) and the other using a Multi-ocean Analyses Ensemble initialization scheme (CFS_MAE). The 1st and 2nd empirical orthogonal function (EOF) modes of the observed ASM rainfall anomalies correspond respectively to the contemporaneous and delayed responses to El Nino and the Southern Oscillation (ENSO) in its developing and decaying years. In general, CFSv2 is capable of skillfully predicting these two dominant ASM modes on the seasonal time scale up to 5 months in advance. Moreover, the predictive skill of the ASM rainfall in CFS_MAE is much higher with respect to the delayed ENSO mode than the contemporaneous one. The predicted principal component of the former maintains high correlation skill and small ensemble spread about two seasons ahead while the latter is significantly degraded in both measures after one season. A maximized signal-to-noise EOF analysis further shows that the delayed ASM response to ENSO is also the most predictable pattern at long leads in CFS_RR. The improved predictive skill of the ASM rainfall following ENSO events originates from the enhanced predictability associated with the active air-sea feedback in the Indo-northwestern Pacific domain from the ENSO peak to the ENSO demise phase, which are well captured in the CFSv2 hindcasts.