曾于恒

NTU

Room 1118, New Building, IAP

10:00, Aug 7, 2018

We improve a simple statistical El Ni？o–Southern Oscillation (ENSO) prediction model based on the tropical dynamics and extratropical ocean-atmosphere interaction. The multivariate linear regression model relies mainly on the evolution of pentad thermocline depth anomaly, the zonal tropical surface wind modulation and the extra-tropical sea level pressure. The new model significantly enhances the ENSO prediction in the hindcast experiments in terms of the ENSO occurrences, normalized Root Mean Square Errors and prediction lead time. The dynamics behind the latest statistical model are consistent with the physical processes of ENSO development as follows: the tropical Warm Water Volume (WWV) resulting from the interannually-varying meridional subtropical cell transport provides a sufficient heat source. When the seasonal phase lock of ocean–atmosphere coupling triggers the positive (negative) zonal wind anomaly in boreal summer and fall, an El Ni？o (a La Ni？a) will develop as evidenced by the Kelvin wave propagation. The dynamic may be suppressed or enhanced by the high-frequency influence of extra-tropical ocean-atmosphere teleconnection. The extratropical atmospheric forcing over both the North and South Pacific can further improve the hindcast skill with 6- to 10-months lead time, relaxing the well-known spring prediction barrier. The hindcast skill of ENSO SSTa is generally better than the existing ENSO forecast models in terms of the monthly correlation. The major improvement of the latest ENSO prediction scheme results from the fundamental combination of the low-frequency propagation of warm water volume and the high-frequency modulation of the extra-tropical teleconnection.