A study on theory of second-order adjoint model

The Hessian matrix, which is formed by the second-order partial derivatives of the cost function with respect to control variables, plays an important role in the procedure of variational data as similation (VDA), sensitivity analysis, etc., and it can be obtained by establishing the first-order ad joint (FOA) and the second-order adjoint (SOA) models for direct model. The derivations of the FOA and SOA models of shallow water equations model are given in detail, which is based upon the Gateaux differential of functional and the concepts of the adjoint operators in Hilbert space. The result for SOA model of the shallow water equations model is obtained, which improves the theory established in the pa per of Wang et al. (1992).