Landing A Strong Typhoon Moving Grapes Simulation Mode Initial Value Test Research

Four landing typhoons in 2005 are numerically simulated to forecast byGRAPES model in this paper, a primary estimate of GRAPES model's predictioncapability about typhoon track,landing time,location and intensity etc. is obtained,and the reasons about errors are analyzed. A few of significant evidence can be offeredfor the application and improvement of GRAPES model on typhoon analysis andforecast。It is showed that the average distance deviation of typhoon center forecasted byGRAPES model is 131km for 24 hour predication and is 252km for 48 hourpredication. The forecasts of landing time and locations are relatively accurate, Butthat of landing intensity is weaker. The average 24 hour model forecast of landinglocation is a slightly better than that of 48 hour.The forecast errors are relatively high at their early and ending stages. The reasonis that, at these stages, as well as being far from land, deviations of these typhoons'location and intensity in the original field are large enough to bring on forecast errorsgreatly. The analysis indicates that the assimilation of unconventional observationdata is essential in the model, which can fetch up the lack of original information. Inaddition, high model resolution may be needed to improve the model performance.Aimed to the condition that model initial fields in model is weaker, a scheme isapplied in this paper to improve the model initial fields. Using observation data, anon-symmetry simulation typhoon is constructed that meets with static equilibrium,gradient wind equilibrium, quasi- thermodynamic equilibrium, coinciding withenvironment field and convection parameterization scheme, then the data of bogustyphoon is assimilated by GRAPES three-dimensional variational data assimilationsystem, and the assimilated typhoon original field is rebuilt. Comparative experimentsbased on typhoon Maisha simulation show that, the model with assimilation schemewell simulates typhoon Maisha's track, intensity, landing time and location. Inparticular, the typhoon track prediction is markedly improved. Further analysisindicates that the assimilation scheme significantly improve the model initial field andtyphoon structure, strengthen the typhoon vortex and improve its vertical structure. After assimilated, the structure of the eye of typhoon is much more concentrated, thewarm center structure is much more clear, the vortex structure in growing stage iswell simulated. The optimized typhoon original fields can reduce the prediction errorswhich are caused by shortage of data over the sea, Some improvements have beenmade in the track and landing forecast.