Improvement of an extended ensemble coupled data Assimilation–Forecast system and its application in El Niño diversity predictions

Abstract

The El Niño–Southern Oscillation (ENSO) can cause climate anomalies on a global scale, and further affect human life and activities in coastal zones. Therefore, its forecast is of great significance for early disaster warning and coastal management. However, the frequent occurrence of central Pacific (CP) El Niño events increases the diversity and complexity of ENSO, severely reducing its prediction efficiency. In this study, an extended ensemble coupled data assimilation–forecast system was employed to investigate the prediction of different types of El Niño events, including eastern Pacific (EP) and CP events. The extended system was based on the fifth-generation Lamont–Doherty Earth Observation (LDEO5) model, in which an advanced ensemble Kalman filter was used to construct a multisource data assimilation system, and a stochastic optimal method was used to measure the influence of atmospheric stochastic processes on model prediction errors. The extended system was used to predict two types of El Niño events that occurred between January 1950 and December 2018. The results showed that the extended system was generally able to predict EP events with a higher accuracy than CP events for all lead times. The extended system successfully predicted the mature phase of EP events up to 12 months in advance but could only predict the mature phase of CP events up to 6 months in advance. The extended system was also able to depict the evolution of both EP and CP events, although the sea surface temperature anomalies were underestimated. The extended system not only provides a useful platform for improving ENSO prediction accuracies in association with El Niño diversity but also provides an important tool for disaster early warning and coastal management.