A New Probability Prediction Theory Of Typhoon Disasters And Its Application In Flood Protection

In recent years, the frequency and intensity of the typhoon is gradually increasing on account of the global warming and sea level rise. The disasters induced by typhoon are complex, such as strong wind, huge waves, storm surge, heavy rain, floods, and so on. In 2006, typhoon Saomai and Bilis attacked the coasts of China, and caused typical wind disaster and flood disaster respectively. In this paper, considering typhoon characteristics such as the typhoon occurring frequency (λ), drop of central pressure (ΔP), radius of maximum wind speed (Rmax), typhoon moving speed (S), minimum distance between typhoon center and certain area (δ), typhoon moving angle (θ) and duration from typhoon land to dissipation (t), a new probability prediction theory of typhoon disasters is proposed and the probability analysis of typhoon induced heavy rain in Zhejing Province is carried out as example by PGMCED and PNLTCEVD model. It is shown that the new model can give more reasonable and comprehensive predicted result than traditional methods.In this paper, the Poisson-Nested Logistic Trivariate Compound Extreme Value Distribution (PNLTCEVD) is applied to design flood of the Three Gorges dam for the first time. The most likely combination (the largest probability of occurrence) and the most adverse combination of major tributaries flood are obtained and corresponding design flood of different joint return periods are calculated. The results can be used to risk analysis and flood control of this project.