Four-dimensional Variational Assimilation Experimentation of Bogus Data for a Typhoon Case

A Bogus data assimilation scheme (BDA)is presented and used to form the initial structure of a typhoon. Then an experiment for No.9608 typhoon is conducted. A mesoscale model (The 5th generation NACA/Pennsyvania State University Mesoscale Model, MM5 for short) is used for both the data assimilation and the prediction. It is found that the scheme is able to recover many of the structural features of a mature typhoon. Obvious improvements are achieved in the prediction of the No.9608 typhoon. First of all, the initial fields of the model variables are well adapted to the forecast model. The spinup phenomenum at the initial time for the model integration can be restrained efficiently. Secondly, the typhoon intensity and track forecast are greatly improved. The errors in the track forecasted at the regular intervals by use of BDA become smaller than those without BDA.