Coupled Conditional Nonlinear Optimal Perturbations and its applications to ENSO ensemble forecasts

Considering the limitations of current ensemble forecasting initial perturbation methods in describing the interactions among various spheres of the Earth system, this study proposes a new method called the Coupled Conditional Nonlinear Optimal Perturbation (C-CNOP), which incorporates the effect of muti-sphere coupling uncertainties. By applying the C-CNOP method to ensemble forecasting of El Niño-Southern Oscillation (ENSO), which is a typical tropical Pacific ocean-atmosphere coupling phenomenon, the study demonstrates that the C-CNOP method can generate coupled initial perturbations (CPs) that much appropriately consider the effect of initial uncertainties of coupled ocean-atmosphere system. It is revealed that the CPs significantly improve ENSO ensemble-mean forecast skill of the time variability of Niño3.4 sea surface temperature anomalies (SSTAs) and the spatial variability of ENSO mature-phase SSTA. Particularly, despite the weakest ocean-atmosphere coupling in the tropical Pacific during the boreal spring and summer, CPs can still capture the uncertainties of the weak coupling when predicting from these seasons, which suppresses the rapidly growing of ENSO prediction errors caused by the strongest ocean-atmosphere coupling instability during these seasons, and thus effectively extends the lead time of ENSO forecasting. Hence, the C-CNOP method is an approach to produce ensemble forecasting initial perturbation that can consider the effect of initial coupling uncertainties much appropriately, which is expected to play an important role in future Earth climate system predictions with further research.