Local data assimilation scheme for wave predictions close to the Portuguese ports

Abstract

The present work describes developments in improving the wave predictions close to the most important Portuguese ports. A wave modelling system was implemented and focused on the Portuguese continental nearshore. It is based on WAM, for the wave generation and on SWAN, for the coastal transformation. The system has been operational since October 2008 and delivers daily forecast products corresponding to four days ahead. Validations of the results were performed against both buoys and satellite data. In order to improve the wave predictions in the nearshore areas neighbouring the Portuguese harbours targeted, two new steps were performed and they are presented in this work. Firstly, high resolution computational domains covering the vicinities of the ports were connected to the wave prediction system. As a further step, a data assimilation scheme was also implemented. The procedure considered uses the data provided by the buoys that are operating in the vicinity of the main Portuguese ports and gives real time estimations of the significant wave height. A recursive successive correction algorithm was developed to improve the model predictions close to the locations of the buoys. The assimilation of the buoy data was performed on three levels - time, geographical and spectral spaces. This procedure allows corrections to the boundary conditions considered for the high resolution computational domains implemented in the vicinity of the harbour areas. The results show that data assimilation makes the wave prediction system more accurate, particularly in conditions where waves are usually less well modelled such as during storms.