A new prediction model of daily weather elements in Hainan province under the typhoon weather

This paper proposes a new prediction model for severe natural disasters, especially typhoon using daily weather analysis. Hainan province in China is selected to be a typical application region, where natural disasters, especially typhoons take place frequently. These disasters have great impacts on the life and property safety of the residents, and therefore are in specific need of accurate prediction. A new prediction model of daily weather in Hainan province under the typhoon weather is proposed in this paper based on the best track datasets of typhoons and the corresponding daily weather data. This model utilizes the statistical methods and data mining technology in combination with the dynamic migration information of tropical cyclones and can provide the dynamic prediction of daily weather elements in any designated location. Three surface meteorological observation stations of Hainan province during the years 1951–1920 are used to test the model. Test results show that the prediction equations established for the vast majority of daily weather elements have passed the significant test. Besides, Typhoon Damrey is used as a case to illustrate the whole daily weather prediction model in detail and comparisons between the model and other official forecast (such as JTWC, UKMO and CMA) are performed thoroughly. It is worth noting that the model proposed in this paper is not limited to Hainan province and can be generalized to other areas in the world.