Multigrid methods for improving the variational data assimilation in numerical weather prediction

Abstract

There are two necessary conditions to solve NumericalWeather Prediction models, initial and boundary conditions. Especially, the initial condition has an important bearing on the model performance. To get a good initial condition, many data assimilation technique have been developed for meteorology and oceanography. Currently, the most commonly used for operational applications are 3 or 4 dimensional variational data assimilation method. In this paper, we study the minimization iteration procedure in 3 dimensional variational data assimilation method. The numerical methods applied to the cost function minimizing process is usually iterative methods, for example the conjugate gradient. In this paper, we apply the multigrid methods based on cell centered finite difference (CCFD) on the variational data assimilation to improve the iterative minimization procedure.