The Investigation In Quantitative Applications Of Tropical Cyclone Prediction Using Satellite Data

The Northwestern Pacific Ocean is one of the most regions bursting out Tropical Cyclone (TC), our country is close to it and influenced severely by TC. The rainstorm, which come with TC usually bring us too much inconvenience and serious loss, so it is significant for preventing and abating disaster that improving TC prediction.For TC, its intensity decides to its influence in a way. The standard method for estimating TC intensity is based on satellite observation, called Dvorak technique, and utilized operationally by tropical analysis centers around the world. It has been applied widely for a long time, but in a word, it is a subjective technique and its results depend on analyst's experience highly, so using it has too much troublesome and error. In order to improve accuracy of intensity estimates and eliminate this subjectivity, an purely objective technique for estimating TC intensity is studied using GMS-5 Black-body temperature (TBB) data of 15 TCs over the northwestern Pacific Ocean during 2001~2002. After a lot of experimentations, an objective formula for intensity estimation has been got based on the multiple linear regression framework and the regression factors include several kinds of TBB characteristic values related to the intensity features of TC cloud system and the latitude of TC center location. Using the minimum sea level pressure (MSLP) in The Tropical Cyclone Annual to validate the satellite-based estimates, statistical analysis indicates the multiple correlation coefficient is more than 0.8, and which reaches 0.89 through 24 hours time-running average scheme. This result is quite similar to that showed by American scientists such as Velden and Olander recently.The intensity of TC is highly correlated with its structure. The thermal structure of TC that appeared in 12 TCs over the Northwestern Pacific Ocean during June 10 ~ September 10, 2002 are studied using the retrieved data from NOAA-16 AMSU . The results show that AMSU can detect sensitively many kinds of tropical cyclone parameters, which belong to the different intensities of tropical cyclones. And AMSU is able to exhibit the important thermal structure characteristics of TC and the internal relationship between the mid-,upper-tropospheric temperature anomaly and intensity change through combining the structure characteristics of relevant humidity field, and also reveal how TC will develop in future.It is very important and difficult in the operation that predicting TC moving direction. An objective technique of Tropical Cyclone (TC) moving direction predicting over the Northwestern Pacific Ocean during 2001-2002 is studied using GMS-5 Outing-Longwave Radiation (OLR) data. Three formulae used to predict TC moving direction in 6, 12 and 24 hours , respectively have been gotten based on the multiple linear regression framework and some regression factors include several kinds of OLR characteristic values and the minimum sea level pressure of tropical cyclone center. The multiple correlation coefficients of the statistical analysis using the information of TCs in The Tropical Cyclone Annual to validate the satellite-based prediction are 0.954, 0.945 , 0.914 respectively. It is significant to better predict unusual TC moving direction.All of these above aim to enhancing TC prediction in the operation. After relevant checkout, they could be applied in future operation and would contribute to TC predicting and improving the forecasting and warning capability.