The Impact of Data Assimilation on ENSO Simulations and Predictions

Abstract

In this study, the impact of oceanic data assimilation on ENSO simulations and predictions is investigated. The authors? main objective is to compare the impact of the assimilation of sea level observations and three-dimensional temperature measurements relative to each other. Three experiments were performed. In a control run the ocean model was forced with observed winds only, and in two assimilation runs three-dimensional temperatures and sea levels were assimilated one by one. The root-mean-square differences between the model solution and observations were computed and heat content anomalies of the upper 275 m compared to each other. Three ensembles of ENSO forecasts were performed additionally to investigate the impact of data assimilation on ENSO predictions. In a control ensemble a hybrid coupled ocean?atmosphere model was initialized with observed winds only, while either three-dimensional temperatures or sea level data were assimilated during the initialization phase in two additional forecast ensembles. The predicted sea surface temperature anomalies were averaged over the eastern equatorial Pacific and compared to observations. Two different objective skill measures were computed to evaluate the impact of data assimilation on ENSO forecasts. The authors? experiments indicate that sea level observations contain useful information and that this information can be inserted successfully into an oceanic general circulation model. It is inferred from the forecast ensembles that the benefit of sea level and temperature assimilation is comparable. However, the positive impact of sea level assimilation could be shown more clearly when the forecasted temperature differences rather than the temperature anomalies themselves were compared with observations.