Initial value small perturbation method suitable for regional numerical mode set simulation forecast

Abstract

The invention discloses an initial value small perturbation method suitable for regional numerical mode set simulation forecast. The method comprises the following steps: 1, calculating an average value and a total disturbance field of mode variables at a certain forecasting moment t by using historical reanalysis data statistics; 2, standardizing mode variables related to a total disturbance field in a mode area at a certain forecasting moment; 3, carrying out joint orthogonal decomposition on the standardized mode variable total disturbance field in the mode area at a certain forecasting moment; 4, sequentially reconstructing and converting the orthogonal vector and the coefficient after joint orthogonal decomposition; and 5, constructing initial value disturbance of a mode region at a certain forecasting moment. A historical data set is fully utilized to obtain mutually orthogonal initial value disturbances, so that the initial value disturbances not only can fully reflect disturbances of different scales and different structures closely related to occurrence and development of specific weather in a numerical mode initial value field, but also can reflect the possibility of actual observation error distribution.

Description

technical field [0001] The invention relates to the technical field of meteorology, in particular to an initial value small disturbance method suitable for ensemble simulation forecast of regional numerical models. Background technique [0002] The quality of the initial value conditions of the numerical model plays an important role in the forecast performance. In order to improve the initial value of the model, meteorologists around the world have successively developed the optimal statistical difference, three-dimensional variational assimilation, four-dimensional variational assimilation and ensemble Kalman filter. However, due to the uncertainty of observations, it is difficult for the initial value of the model to reach the "true value", and it is also difficult for the physical process of the model to "truly" describe the actual atmosphere. In order for the model to be able to predict the possibility of some extreme weather processes, meteorologists have proposed the ...

AI Technical Summary

Problems solved by technology

The above-mentioned initial value perturbation techniques have the following common characteristics: (1) need the model to carry out cyclic integral forecasting to count the forecast error covariance matrix; (2) need to solve the eigenvector or singular vector of the error covariance matrix; (3 ) Whether the initial value disturbance can grow rapidly, diverge and uniformly discrete in the model forecast is used as the criterion for judging the quality of the ensemble forecast; (4) It is difficult to express the actual observation error characteristics

Images