Assessment of prediction and predictability of short rains over equatorial East Africa using a multi-model ensemble

In this study, the performance of dynamical seasonal forecast systems is evaluated for the prediction of short rain anomalies over equatorial East Africa. The evaluation is based on observational datasets and the Asia-Pacific Climate Center (APCC) Ocean–Atmosphere coupled multi-model ensemble (MME) retrospective forecasts (hindcasts). These forecast systems have different hindcast periods; here, we have selected common years from 1982 to 2005. The ensembles of individual models and their MME mean are evaluated. Hindcasts initialized on the 1st of August from each year alone are considered, as these are the most relevant to short rain predictions. The coupled climate model ensemble reproduces the spatial distribution of mean September-October-November (SON) rainfall and seasonal climate variations over equatorial East Africa with further improvement in MME mean. Individual coupled models and MME mean also show statistically significant skill in forecasting sea surface temperatures anomalies (SSTAs) over the western and eastern parts of the equatorial Indian Ocean, giving significant correlation at 99 % confidence level for Indian Ocean dipole (IOD). Moreover, five out of ten coupled models and MME mean show statistically significant skill in predicting equatorial East Africa short rains. The fidelity of hindcasts is further measured by anomaly correlation coefficient (ACC) and four models as well as MME mean show significant skill over East Africa. It is shown that the reproduction of the observed variability in the East African region is mainly due to a realistic relationship of East African rainfall with the Indian Ocean dipole. Overall, the skill of the dynamical models is attributed to the fact that slowly evolving SSTs are the primary source of predictability and to the fact that coupled climate models produce skillful predictions of SON SST anomalies over the tropical Indian Ocean. This information opens the possibility of using readily available seasonal forecasts as skillful predictions of equatorial East Africa short rains.