Study on Saturation Estimation of Initial Perturbations in Ensemble Marine Forecasting based on BGM Method

Abstract

Because the ocean is a highly complex nonlinear chaotic system, its numerical prediction results are strongly uncertain due to the influence of the initial values and the model itself. Ensemble forecasting is an effective way to reduce the uncertainty of forecasting. The initial perturbations quality, especially the saturation estimation, is of practical significance for improving ensemble forecasting techniques. In this paper, the factors affecting initial perturbation saturation of ensemble forecasting are analyzed by BGM based on ROMS model, and the perturbation saturation velocities of the two methods are compared. The results show that the saturation velocity of the initial perturbation is faster than that of the initial perturbation space structure. The larger the initial perturbation is, the higher the saturation degree and velocity are, and the saturation time of the same variable varies at different layers. The saturation of different variables is also different; the perturbation saturation speed of the regional adjust perturbation method is faster than that of the classical static perturbation method.