| Summary: | Extreme rainfall is one of the disastrous events that occurred due to massive rainfall overcome time beyond the regular rainfall rate. The catastrophic effects of extreme rainfall on human, environment, and economy are enormous as most of the events are unpredictable. Modelling the extreme rainfall patterns is a challenge since the extreme rainfall patterns are infrequent. In this study, a model based on descriptive indices to forecast extreme rainfall is proposed. The indices that are calculated every month are used to develop a Back Propagation Neural Network model in forecasting extreme rainfall. Experiments were conducted using different combinations of indices and results were compared with actual data based on mean absolute error. The results showed that the combination of six indices achieved the best performance, and this proved that indices could be used for forecasting extreme rainfall values.
|
|---|