A Stochastic Model for Strong Wind Events in Central Europe based on Historical Wind Data

For the generation of synthetic storms in order to estimate claims due to storm hazards there exist a couple of commercial tools. They are based on expert's knowledge and physical models of wind appearance or work with distortions of historical hazards. In contrast to these approaches we present a purely data driven model. From the historical data we extracted the hourly wind peaks over the last two decades and derived by algebraic methods a stochastic storm model in form of a polynomial chaos expansion.