Estimation of Significant Wave Heights Using Numerical and Neural Techniques and Comparison with Buoy and Satellite Observations

Abstract

Short term significant wave heights have been evaluated using the numerical WAM model and the results are compared with satellite (Jason-2) and moored buoy measurements. Additionally artificial neural network (ANN) is used to predict significant wave heights over a future time step and such predictions are also compared with corresponding satellite and buoy measurements. The observations from moored buoys monitored by India's National Institute of Ocean Technology for a period of about four and a half years at three locations in the Arabian Sea and three in Bay of Bengal covering the west as well as east coast of India are involved. The buoy recorded largest waves during the cyclonic condition, and this was confirmed at a location code named: BD11 by the satellite measurements as well as the WAM and ANN based evaluations. The evaluation accuracy reflected in the coefficient of correlation is found to be high in the Arabian Sea than the Bay of Bengal.