An Objective Analysis Method of Multivariate Optimum Interpolation in Mesoscale Model

The forecast results from numerical weather prediction models are closely related to their initial values. Thus, the improvement of the objective analysis would result in the increase of forecast accuracy. In this study, based on the features on the features of the MOMS mesoscale model, an objective analysis method with multivariate optimum interpolation is advanced. The forecast results of T42 model is used as the first guess and the features of mesoscale model have been taken into consideration in the selection of observational stations, correction of the records, and filling up a vacancy of observation data. The objective analysis schemes have provided good initial values for mesoscale numerical weather prediction model and diagnostic analysis, which is favourable for the great improvement of the forecast accuracy of mesoscale model.