Ensemble-based assimilation of wave model predictions: Contrasting the impact of assimilation in nearshore and offshore forecasting at different distances from assimilated data

Abstract

Accurate wave predictions are crucial for many marine-related activities, including offshore operations, navigation, and marine renewable energy. To improve the forecast performance, this study presents a data assimilation system that utilizes the WAVEWATCH III wave model and the local Ensemble Optimal Interpolation scheme to assimilate significant wave heights data from HY-2B altimeter in the China Seas. The effects of assimilation are analyzed in both coastal and offshore regions at different distances from assimilated satellite data points during a forecast period. The results show that systematic bias in wave forecasting for both offshore and nearshore locations can be effectively reduced through the use of data assimilation, with localization employed to minimize the influence of random errors. Nearshore forecasting benefits from assimilation to a greater extent than offshore forecasting, with a more enduring improvement observed. Additionally, the decline in forecast improvement with increasing distance from assimilated data highlights the necessity of assimilating a greater quantity of observational satellite data. Furthermore, careful consideration should be given to the size of the ensemble when using a specific localization radius, as the improvement in skill may stagnate with larger ensembles. Finally, assimilating additional nearshore satellite data does not negatively impact the predictive precision.