In this paper, we propose a parallel data assimilation module based on ensemble optimal interpolation （EnOI）. We embedded the method into the full-spectral third-generation wind-wave model, WAVEWATCH III Version 3.1...In this paper, we propose a parallel data assimilation module based on ensemble optimal interpolation （EnOI）. We embedded the method into the full-spectral third-generation wind-wave model, WAVEWATCH III Version 3.14, producing a wave data assimilation system. We present our preliminary experiments assimilating altimeter significant wave heights （SWH） using the EnOI-based wave assimilation system. Waters north of 15°S in the Indian Ocean and South China Sea were chosen as the target computational domain, which was two-way nested into the global implementation of the WAVEWATCH III. The wave model was forced by six-hourly ocean surface wind velocities from the cross-calibrated multi-platform wind vector dataset. The assimilation used along-track SWH data from the Jason-2 altimeter. We evaluated the effect of the assimilation on the analyses and hindcasts, and found that our technique was effective. Although there was a considerable mean bias in the control SWHs, a month-long consecutive assimilation reduced the bias by approximately 84% and the root mean-square error （RMSE） by approximately 65%. Improvements in the SWH RMSE for both the analysis and hindcast periods were more significant in July than January, because of the monsoon climate. The improvement in model skill persisted for up to 48 h in July. Furthermore, the SWH data assimilation had the greatest impact in areas and seasons where and when the sea-states were dominated by swells.展开更多
基金Supported by the National Special Research Fund for Non-Profit Marine Sector(Nos.201005033,201105002)the National High Technology Research and Development Program of China(863 Program)(No.2012AA091801)+1 种基金the National Natural Science Foundation of China(No.U1133001)the NSFC-Shandong Joint Fund for Marine Science Research Centers(No.U1406401)
文摘In this paper, we propose a parallel data assimilation module based on ensemble optimal interpolation （EnOI）. We embedded the method into the full-spectral third-generation wind-wave model, WAVEWATCH III Version 3.14, producing a wave data assimilation system. We present our preliminary experiments assimilating altimeter significant wave heights （SWH） using the EnOI-based wave assimilation system. Waters north of 15°S in the Indian Ocean and South China Sea were chosen as the target computational domain, which was two-way nested into the global implementation of the WAVEWATCH III. The wave model was forced by six-hourly ocean surface wind velocities from the cross-calibrated multi-platform wind vector dataset. The assimilation used along-track SWH data from the Jason-2 altimeter. We evaluated the effect of the assimilation on the analyses and hindcasts, and found that our technique was effective. Although there was a considerable mean bias in the control SWHs, a month-long consecutive assimilation reduced the bias by approximately 84% and the root mean-square error （RMSE） by approximately 65%. Improvements in the SWH RMSE for both the analysis and hindcast periods were more significant in July than January, because of the monsoon climate. The improvement in model skill persisted for up to 48 h in July. Furthermore, the SWH data assimilation had the greatest impact in areas and seasons where and when the sea-states were dominated by swells.
