JASON-2卫星高度计波高数据同化对中国渤黄海海浪数值预报影响研究

利用原国家海洋局北海分局浮标所测有效波高数据对Jason-2卫星高度计所测有效波高数据进行验证,采用50km空间窗和0.5h时间窗,得到219个时空配准点。对配准结果进行统计分析表明,Jason-2卫星高度计测得有效波高与浮标测量结果存在-0.277m的偏差,均方根误差为0.372m。利用最小二乘回归(OLR)对Jason-2有效波高数据进行校正可使其均方根误差下降至0.247m,减少34.5%。基于第三代海浪模式WAVEWATCH III对Jason-2有效波高数据进行最优插值同化试验,其中背景误差相关函数取为指数形式,相关距离尺度选为500km。与浮标观测数据比较表明,同化后模式有效波高均方根误差比未同化时减少11.56%,,能够有效地改善模式精度。以此为初始场进行为期3d的数值预报实验。与未同化实验相比,卫星高度计有效波高数据同化对模式0～72 h预报有不同程度的改善,改善程度随预报时间的增加而降低。

The impact of assimilation of Jason-2 altimeter wave data on wave forecast model in the Bohai Sea and Yellow Sea

The significant wave height data of Jason-2 satellite altimeter are validated using the buoy significant wave height data of the North China Sea Branch,the State Ocean Administration.Using 50 km space window and 0.5 h time window,219 spatiotemporal registration points are obtained.Statistical analysis of registration results shows that there is a deviation of-0.277 m between the significant wave height measured by Jason-2 satellite altimeter and measured by buoy,and the root mean square error is 0.372 m.Using ordinary least regression(OLR)to correct Jason-2 significant wave height data can reduce its root mean square error to 0.247 m and 34.5%.Based on the third generation wave model WAVEWATCH III,the optimal interpolation Assimilation Experiment of Jason-2 significant wave height data is carried out,in which the background error correlation function is taken as an exponential form and the correlation distance scale is chosen as 500 km.The comparison with buoy observation data shows that the root mean square error of the significant wave height of the assimilated model is 11.56%less than that of the non-assimilated model,which can effectively improve the prediction effect of the model.A three-day numerical prediction experiment was carried out using this as the initial field.The results show that the sea wave data assimilation of satellite altimeter improves the prediction of 0-72 h to some degrees,and the degree of the improvement decreases with the increase of prediction time.