How well do climate models simulate atmospheric teleconnctions over the North Pacific and East Asia associated with ENSO?

Abstract

During the El Nio and La Nia mature phase, atmospheric teleconnections over the North Pacific and East Asia vary considerably on sub-seasonal time scales, and are strongly phase-locked to the sub-seasonal evolution. In this study, we investigate how well climate models that participated in the Coupled Model Intercomparison Project Phase 5 (CMIP5) simulate the sub-seasonal evolution of teleconnections over the North Pacific and East Asia associated with El Nio-Southern Oscillation (ENSO). In the observations, there is a prominent anticyclone anomaly over the Kuroshio extension region (i.e. Kuroshio anticyclone), which significantly affects East Asian climate in the early winter (November-December) of El Nio years. However, in January, the Kuroshio anticyclone suddenly disappears, and a cyclonic flow dominates over the North Pacific. It is found here that the CMIP5 models simulate the overall extratropical teleconnection patterns, but they fail to reproduce some of these sub-seasonally-varying features in atmospheric circulation. For example, the models tend to simulate a weaker Kuroshio anticyclone in the early winter during El Nio phases, and fail to capture the abrupt decay of the Kuroshio anticyclone in the late winter. We demonstrate here that these systematic errors in ENSO teleconnection can be explained by systematic errors in tropical precipitation associated with ENSO. That is, negative precipitation anomalies over the western North Pacific (WNP) are too weak in the models compared to that in the observations, and their amplitude tends to be strengthened from December to the following January, while they are weakened in the observations. In addition, analyses on the inter-model diversity strongly support that relative magnitudes of WNP and central Pacific precipitation anomalies are critical for determining sub-seasonal evolution of ENSO teleconnections over the North Pacific and East Asia.