Estimation of one-dimensional energy density spectrum of surface waves using Genetic Algorithm

One-dimensional energy density spectrum was estimated for a grid point in the Indian Ocean (IO) region. Empirical orthogonal function (EOF) analysis and genetic algorithm (GA) was applied for the prediction. EOF analysis was implemented separately for the southwest and northeast monsoon months, using a 5-years model generated energy density spectrum data. The first eigenmode accounted for 50.4% and 64% of the total variability of the above parameter for the two seasons respectively. Then GA was applied separately to the time series of the corresponding first principal component, for both the seasons, with a lead time of 06 hours. It was found that the performance of the combined approach was better for the northeast monsoon than the southwest monsoon and the root mean square error (RMSE) of GA forecasted spectra was found to be less when it was calculated with respect to model computed spectra than when it was calculated with respect to buoy observed spectra.