中文摘要：

借助伴随同化方法，利用实测水位资料，对空间分布的风应力拖曳系数做了反演研究。假定风应力拖曳系数具有空间分布特征，即在模拟海区中均匀选取一些独立点，利用这些独立点的风应力拖曳系数线性插值得到全场的风应力拖曳系数。同化实验结果表明，采用空间分布的风应力拖曳系数得到的模拟结果，明显优于将风应力拖曳系数取为常数和依照经验公式计算风应力拖曳系数时的模拟结果，其中8509号台风风暴潮四个实验的绝均差分别为25cm、20cm、19cm和14cm，E4的误差明显小于其余三个实验。这说明采用空间分布的风应力拖曳系数，可以有效减小模拟水位与实测水位之间的误差，提高风暴潮的数值模拟精度。对反演得到的风应力拖曳系数的初步分析表明，风应力拖曳系数与风速的关系不明确，无法得出确定的结论。

英文摘要：

Based on the data from the observed sea level elevation, numerical experiments are performed to study the spatially varying drag coefficient （DC） with the adjoint assimilation method. We assume the DC has spatially varying character： in the concerned sea, some grid points are uniformly selected as the independent DC, while the DC at other grid points can be obtained through linear interpolation with the independent DC. The results demonstrate that the simulation of the mentioned method is much better than those when DC is treated as a constant or calculated by an experimental formula. Extraordinarily, the mean absolute deviation of four experiments of storm surge caused by Typhoon 8509 are 25cm, 20cm, 19cm and 14cm, the deviation of Exp.4 is much lower than other three experiments. These data indicates the spatially varying DC is capable of reducing the deviations between elevations of simulated values and observed data, with the simulated precision of storm surge increased. A preliminary analysis on the inversion results of DC illustrates that the relation between DC and wind velocity is uncertain, and we may not able to conclude the relationshila formula between DC and wind velocity.