IAHR Document Library

« Back to Library Homepage « Proceedings of the 18th IAHR APD Congress (Jeju, 2012)

Forecasting Ocean Waves Using Support Vector Regression

Author(s): Shreenivas Londhe; Pradnya Dixit; Shreekant Charhate

Linked Author(s):

Keywords: Wave forecasting; Data driven modeling; Artificial Neural Networks; Support Vector Regression

Abstract: Real-time forecasting of waves over a time step of a few hours or days at a specific site is required in operational planning of any engineering activity in the ocean. Traditionally this is done using numerical models for forecasting over regional domain. For point forecasts various data driven schemes have been tried out of which Artificial Neural Networks seem to be superior compared with other approaches. However ANNs also have limitations in that they are unable to predict the extreme events accurately as well as the forecasting accuracy decreases with increasing lead time of forecast. In the present work another data driven tool of Support Vector Regression is employed to forecast wave height at 4 locations, 2 around USA coast and 2 around Indian coast for the lead times of 3 hr to 96 hours. It is observed that for Indian locations the forecasts are very satisfactory up to 96 hours ahead.


Year: 2012

Copyright © 2024 International Association for Hydro-Environment Engineering and Research. All rights reserved. | Terms and Conditions