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，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Hilber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程及相应的同化试验数值结果．“,”ln order to provide more aoeurate initial field for numerical predict of the sea temperature the adjointrnmethod is used to 0rolmlilate the sea temperature observations into one - dimensional sea temperature model in thisrnpaper. How to establish the adjoint model is discussed by using the Gateaux differential of function and the eoneeptsrnof the adjoint operators in Hibert space. At the same time it is verified that selecting proper finite difference schemerncan ensure discrete form remaining the same adjoint relationship to continuous form of the model. The numerical results describing the varying sea temperature field ate illustrated.rn