Regional Rainfall Forecasting using Large Scale Climate Teleconnections and Evolutionary Algorithms

The continuous occurrence of changes in the global climate causes significant variability in the seasonal and intra-seasonal rainfall pattern, which often leads to frequent floods and droughts in India. To reduce the magnitude of effect of such natural calamities and for better management of water resources, it is essential to predict the rainfall, well in advance. In this study the possible relationship between the large scale climate indices like, El-Niño Southern Oscillation (ENSO), EQUitorial INdian Ocean Oscillation (EQUINOO) and a local climate index of Ocean-Land Temperature Contrast (OLTC) are studied and used to forecast regional monsoon rainfall of Orissa state in India. To handle the highly non-linear and complex behavior of the climatic variables in forecasting the rainfall, this study employs Artificial Neural Networks (ANNs) methodology. In this study, to optimize the ANN architecture, Genetic Optimizer is used. After identifying the lagged relationship between climate indices and monthly rainfall, the rainfall values are forecasted for summer monsoon months of June, July, August and September (JJAS) individually as well as total monsoon rainfall. The models are trained individually for monthly and for seasonal rainfall forecasting. Then the trained models are tested to evaluate the performance of the model. The results show reasonably good accuracy for monthly and seasonal rainfall forecasting. The study demonstrates the possible teleconnection between large scale climate indices and regional rainfall over Orissa and also shows the usefulness of ANNs in rainfall forecasting.