Prediction of Sea Level Anomaly in the Arabian Sea Using Genetic Algorithm

Prediction of satellite altimeter derived fields of sea level anomaly (SLA) has been carried out in the Arabian Sea using genetic algorithm (GA). For compressing the spatial variability into a few eigenmodes, a preliminary empirical orthogonal function (EOF) analysis has been carried out on the training set of fields of SLA and GA has been applied to the time series of the principal components (PC). The forecast equations have been applied to an independent validation data set. It has been found that GA is able to improve upon persistence forecast in all the cases and the improvement varies from 6% to 23%. Correlation between actual and predicted PCs exceeds 0.9. Forecast SLA fields have been computed by weighting the spatial eigenmodes by the corresponding forecast PCs. Forecast quality has been evaluated by computing the RMS error of forecast. It has been found that the RMS forecast error is much less than the natural variability of the data represented by its standard deviation. It can thus be concluded that the performance of the forecast is quite satisfactory