A NEW NUMERICAL WEATHER PREDICTION MODEL, UTILIZING MULTIPLE-INSTANT OBSERVATION DATA
摘要
In this paper, assuming that the output signal is not correlated with the noise, a multiple-instant prediction model is proposed. This model combines dynamical method with statistic method, utilizing multiple-instant observation data. Thus, the numerical weather prediction is formulated as an evolutional problem. The multiple-instant observation data together with the prognostic ones are expanded into empirical orthogonal functions.The dynamical equation used in the present work is the barotropic vorticity equation. However, the model proposed here may apply to any other dynamical equation as well, for example, the primitive equation.In order to solve the statistic-dynamical forecasting equation numerically, a triple iteration method has been proposed. The results of practical forecasts are also given.