论文部分内容阅读
以一维海洋水温模型为例，利用伴随法进行海洋观测数据同化试验，以便为水温的数值预报提供较准确的初始场．文中利用泛函的Gateaux微分和Hilbert空间上伴随算子的概念讨论了连续的伴随模型的建立，并通过选择适当的差分格式离散伴随模型，使其保持连续时的伴随关系，同时给出了水温初始场最优化过程及相应的同化试验数值结果．
Taking the one - dimensional ocean water temperature model as an example, the ocean observation data assimilation test is conducted by using the adjoint method to provide a more accurate initial field for the numerical prediction of water temperature. In this paper, we discuss the establishment of continuous adjoint models using the Gateaux differential of functions and the concomitant operator in Hilbert space. By means of selecting proper differential schemes, the adjoint model is discretized to keep the concomitant relationship in continuous time. At the same time, Initial field optimization process and corresponding assimilation test numerical results.