一种求解条件非线性最优扰动的快速算法及其在台风目标观 …（A fast algorithm for solving conditional nonlinear optimal disturbances and its application in typhoon target...）.docVIP

一种求解条件非线性最优扰动的快速算法及其在台风目标观 …（A fast algorithm for solving conditional nonlinear optimal disturbances and its application in typhoon target...） Own collection The error can hardly be avoided For reference exchange If there is an error Please! Thank you Test a fast algorithm for solving CNOP and in the observation of typhoon. Wang Bin Tan Xiaowei China Institute of Atmospheric Physics LASG Beijing One hundred thousand and twenty-nine abstract Conditional nonlinear optimal perturbation (CNOP) is a new method introduced by Mu et al in 2003 It is a linear singular vector in the promotion of the nonlinear case To overcome the defect of linear singular vectors can not represent the fastest growing perturbation of the nonlinear system A nonlinear system for studying predictability and sensitivity of new effective tool however As for the previous CNOP needs to use the adjoint technique The calculation is very large Limit the application of the method In order to overcome the difficulties Based on empirical orthogonal function (EOF) A fast algorithm for solving CNOP The realization of CNOP using GRAPES regional operational forecast model fast calculation And the target observation on Typhoon Matsa in the preliminary test By observing system simulation experiments (OSSE) to test the sensitivity of the method to determine the area (targeting area) the effectiveness and feasibility The test results show that A fast algorithm for solving CNOP The net energy with time and rapid development But development is nonlinear In the target observation of Typhoon Matsa in case The forecast time obtained by the fast algorithm for 24 h CNOP can effectively identify the target area And by targeting the region to improve the initial value Can obviously reduce the target area (inspection area) in the 24 h accumulated rainfall prediction errors especially The accumulated precipitation forecast improvement can be extended to longer time (e.g. 72 h) Although the test time is in twenty-fourth hours