Preconditions for CMIP6 models to reproduce the relationship between wintertime ENSO and subsequent East Asian summer rainfall

Abstract

The El Nino-Southern Oscillation (ENSO) in the preceding winter (December-January-February) is one of the key factors affecting subsequent East Asian summer (June-July-August) rainfall (EASR). However, current models face great challenges in reproducing ENSO's impact on the EASR. This study attempts to reveal the factors that determine whether a model in phase 6 of the Coupled Model Intercomparison Project (CMIP6) can successfully reproduce this relationship by analyzing the outputs of historical climate simulation in 20 CMIP6 models. The results show that most of the models that overestimated ENSO interannual variability reproduced significant ENSO-EASR relationships, whereas all models that underestimated ENSO variability failed to reproduce this relationship. Further analyses show that models with stronger ENSO variability tended to simulate more realistic physical processes linking ENSO and EASR, i.e. the connections between ENSO and the tropical Indian Ocean (TIO) sea surface temperature (SST), between TIO SST and the Philippine Sea convection (PSC), and between PSC and EASR. Moreover, among the models that overestimated ENSO variability, only those that successfully captured significant TIO SST-PSC connections reproduced the observed ENSO-EASR relationship, although all these models captured ENSO-TIO SST and PSC-EASR teleconnections well. Therefore, simulating stronger ENSO interannual variability is the first necessary precondition for a CMIP6 model to capture the delayed effect of ENSO on EASR; reproducing a realistic TIO SST-PSC teleconnection is the second necessary precondition. This study will help models to improve their skills in simulation and prediction of EASR.