Analysis methods for numerical weather prediction

摘要: SUMMARY Bayesian probabilistic arguments are used to derive idealized equations for finding the best analysis numerical weather prediction. These compared with those from other published methods in light of physical characteristics thc NWP problem; namely predetermined nature basis analysis, need approximation because large-order systems, undcrdctcrminacy problem when using observations alone, and availability prior relationships resolve underdeterminacy . Prior result (1) knowledge time evolution model (which together use a distribution constitutes four-dimensional data assimilation); (2) that atmosphere varies slowly (leading balance relationships); (3) nonlinear coupling parameters scales atmosphere. Methods discussed include variational techniques, smoothing splines, Kriging, optimal interpolation, successive corrections, constrained initialization, Kalman-Bucy filter, adjoint assimilation. They all shown relate hence each other. Opinions given on particular might he more appropriate. By comparison method some insight is gained into appropriate choices practical methods.