Data assimilation of typhoon-generated ocean waves using a drifting buoy observation network: case study of typhoon NANMADOL in 2022

Recent deployments of dense drifting wave buoy networks provide opportunities to improve wave modeling under extreme tropical cyclones, where observations are scarce and model skill remains limited. We developed a data assimilation (DA) system that incorporates drifting buoy observations into a spectral wave model and evaluated its performance for Typhoon NANMADOL over the western North Pacific. Two DA experiments were conducted using drifting buoy observations: one assimilating significant wave heights and the other assimilating frequency spectrum data. The DA results were evaluated using coastal wave observations. When high waves reached the Japanese coast, both DA systems reduced coastal wave errors, with significant wave height RMSE decreasing by about 0.6 m and mean period RMSE by about 1 s. The influence of DA extended beyond the assimilation region through wave propagation. Considering its simplicity, significant wave height DA performed better than spectral DA. The parameter controlling the spatial range of assimilation (decorrelation scale) must be tuned according to buoy location to avoid coastal error growth. Experiments excluding selected buoys show that coastal buoys mainly improve significant wave height, whereas buoys near the typhoon core improve wave period. These results demonstrate that assimilating drifting buoy observations can significantly improve typhoon-induced wave predictions.