부산대학교 도서관

Reliable prediction of rainfall in advance of the rainy season would have enormous social and economic benefits to countries such as Ethiopia that depend on rain fed agriculture. This thesis is trying to do two things. Firstly it tries to develop a statistical seasonal forecasts of the Ethiopian rains which could be used operationally. Secondly, it also tries to understand the mechanisms by which the predictors (remote SSTs) are linked to rainfall using observational and modelling studies. Because of the spatial variation in both the interannual variability and the annual cycle of rainfall, Ethiopia was divided into homogeneous rainfall zones, and separate forecasts were developed for each zone. Two techniques (multiple linear regression (MLR) and linear discriminant analysis (LOA» were applied to four sets of predictors (selected by either stepwise regression or discriminant analysis either including or excluding the contemporaneous season). All the forecasts had more skill than either a random or climatological forecast. For all forecasts, the extreme years (very rainy and very dry) were more reliably forecast than average years. The mechanisms for the link between SSTs and rainfall were studied using observational analysis and modelling experiments. In the observational analysis, two sets of composites were analysed. The first set of composites was composites of atmospheric fields based on excess and deficit rainfall years which were meant to identify the large scale atmospheric variables responsible for rainfall variability. The second set was also composites of global atmospheric fields but based on warm and cold SST of the oceanic regions used as predictors in the forecasting models. This analysis aimed to identify the link between the remote SST and the rainfall via large scale atmospheric features. In the Modelling experiments, an ensemble of atmosphere only GeM (HadAM3) integrations were done by forcing with observed and idealised SST anomaly patterns. The observed SST forced run is intended to examine whether the model captures some of the observed interannual variability of rainfall related to fluctuation in the SST. The idealised SST pattern forced experiment is done in order to understand the sensitivity of atmospheric circulations to various locations of SST forcing. The result from the idealised experiment shows that warming of equatorial Pacific directly affects Ethiopian rainfall by among other things, weakening the tropical easterly jet and reducing the influx of moisture from the Atlantic and Indian oceans.