Hybrid assimilation system and method based on ensemble Kalman filter framework

Abstract

The invention discloses a hybrid assimilation system and method based on an ensemble Kalman filter framework, including an input unit, a climatic state ensemble disturbance unit, a cyclic ensemble disturbance unit, a climate state background error covariance unit, an augmented ensemble disturbance unit, and a background error Covariance unit, climatic state background error covariance unit, mixed background error covariance unit, posterior ensemble average update unit, posterior ensemble disturbance update unit, posterior cyclic ensemble unit, estimate background error covariance by climatic state disturbance set Approximate static background error covariance, which can implement hybrid assimilation method in the framework of ensemble Kalman filter, and use mixed background error covariance to update ensemble mean and ensemble disturbance. The invention overcomes the inconsistency problem existing in the separate hybrid assimilation system and the ensemble Kalman filter system, and can further reduce the error after assimilation under the condition that the hybrid weight and other assimilation parameters are properly selected.

Description

technical field [0001] The invention relates to a numerical weather forecast assimilation method, which belongs to a mixed assimilation method based on variational assimilation and ensemble assimilation. The mixed assimilation method can combine the advantages of variational assimilation and ensemble assimilation to better improve weather forecast results. Background technique [0002] Data assimilation is a technique that combines prior information and observations of a dynamical system to find the optimal estimate of the state of the system. [0003] Commonly used data assimilation methods include variational assimilation and set assimilation. In order to combine the advantages of ensemble assimilation and variational assimilation while minimizing the disadvantages of these two methods, a hybrid ensemble-variational assimilation method is proposed. Previous studies have shown that hybrid ensemble-variational assimilation methods are superior to individual variational or e...

AI Technical Summary

Problems solved by technology

[0006] Therefore, the dilemma that researchers are currently facing is how to develop a hybrid assimilation method that can achieve hybrid assimilation under the framework of a single ensemble Kalman filter to solve the inconsistency caused by separate assimilation systems

Images