Artificial Intelligence Tools to Forecast Ocean Waves in Real Time

Abstract

Prediction of wind generated ocean waves over shortlead times of the order of some hours or days ishelpful in carrying out any operation in the sea suchas repairs of structures or laying of submarinepipelines. This paper discusses an application ofdifferent artificial intelligent tools for thispurpose. The physical domain where the wave forecastingis made belongs to the western part of the Indiancoastline in Arabian Sea. The tools used are artificialneural networks, genetic programming and model trees.Station specific forecasts are made at those locationswhere wave data are continuously observed. A timeseries forecasting scheme is employed. Based on asequence of preceding observations forecasts are madeover lead times of 3 hr to 72 hr. Large differences inthe accuracy of the forecasts were not seen whenalternative forecasting tools were employed and hencethe user is free to use any one of them as per herconvenience and confidence. A graphical user interfacehas been developed that operates on the received waveheight data from the field and produces the forecastsand further makes them accessible to any user locatedanywhere in the world.