An investigation of CMIP5 model biases in simulating the impacts of central Pacific El Niño on the East Asian summer monsoon

Abstract

The delayed impacts of the central Pacific (CP) El Nino on the East Asian summer monsoon (EASM) are evaluated by comparing historical runs from Coupled Model Intercomparison Project Phase 5 models against reanalysis data. In observations, an anomalous western North Pacific anticyclone (WNPAC), linking CP El Nino to the EASM, forms due to the transition of sea surface temperature (SST) warming into SST cooling over the CP, which generates a WNPAC through a Gill–Matsuno response. In comparison with the observational result, only one-third of the models (i.e., the type-I models) capture a weaker and smaller WNPAC, whereas the other two-thirds (i.e., the type-II models) fail to reproduce a WNPAC. The simulation biases in both of type-I models and type-II models mainly arise from an unrealistic, long-lasting CP El Nino warming, which causes a north Indian Ocean SST warming bias in models through air–sea interaction process. This north Indian Ocean SST warming generates the WNPAC through capacitor effects, which is different from the WNPAC formation mechanism in observations. This discrepancy leads to simulation biases in type-I models. In type-II models, the unrealistic CP El Nino warming persists into summer, which produces an anomalous cyclone over the central-western Pacific. The opposite effect of the CP and north Indian Ocean SST warming on the WNP atmospheric circulation leads to disappearance of the WNPAC. Hence, large simulation biases are produced in type-II models. Further analysis demonstrates the slow decay of CP El Nino is caused by the unrealistically simulated climatological SST, which creates strong warm meridional oceanic advection and results in a sustained CP El Nino warming.