Observed and simulated air-sea feedbacks associated with ENSO and monsoon

Abstract

Associated with the double ITCZ (Inter-tropical convergence zone) problem, a dipole SST bias pattern (cold in the equatorial central Pacific and warm in the southeast tropical Pacific) remains a common problem in current coupled models. Based on a newly-developed model, we demonstrated that a serious consequence of this SST bias is to suppress the thermocline feedback in El Niño/Southern Oscillation (ENSO) simulation. Firstly, the excessive cold tongue extension pushes the anomalous convection far westward, diminishing the convection-low level wind feedback and thus the air-sea coupling strength. Secondly, the equatorial surface wind anomaly exhibits weak meridional gradient, leading to a weakened wind-thermocline feedback. Thirdly, the equatorial cold SST bias induces a weakened upper-ocean stratification, yielding the underestimation of the thermocline-subsurface temperature feedback. Finally, the dipole SST bias underestimates the mean upwelling through both dynamic and thermodynamic effects. In recent decades, El Niño events have occurred more frequently over the equatorial central Pacific (CP Warming, CPW). Here, we ascribe this predominance of the CPW to a dramatic decadal change in the Pacific mean state and annual cycle. The mean state change characterized by a decadal La Niña-like pattern tends to anchor convection and surface zonal wind anomalies to the vicinity of the dateline, facilitating surface warming to occur in the CP. The annual cycle change, with the trade winds intensifying during boreal winter and spring, prevents the warming development but helps the warming decay in the EP. More CPW events are expected in the coming decade if the La-Niña-like pattern persists. The western North Pacific (WNP) Subtropical High (SH) has profound impacts on Asian summer monsoon, North Pacific storms. The cause of the interannual variability of WNPSH, however, remains controversial. Here we show that the anomalous WNPSH is primarily determined by a remote cooling/warming in the equatorial central Pacific and a positive thermodynamic feedback between the local circulation and a dipole sea surface temperature in the Indo-Pacific warm pool. We demonstrate that a physical-empirical prediction model built on these physical understandings has comparable performance with those of three state-of-the-art coupled climate models in re-forecast of the strength of the WNPSH.