Tornado intensity prediction based on environment elements at tornado events starting points

In this research, we analyze the behavior of tornado events which occurred in the State of Missouri between 1950-2013. The formation of tornadoes is a complex meteorological phenomenon which is still not fully understood to the level of accurate predictability based on current atmospheric sensor networks, e.g., Doppler Radar. Environmental conditions at tornado start points may play a key role in determining the intensity; and therefore we explore the potential of using environment features as tornado intensity predictors. We analyze both the spatial and temporal features of historical Missouri tornado events. Environmental features at tornado starting point; such as ground slope, elevation, temperature and precipitation are explored as the predictors of tornado intensity through application of Decision Tree and support vector machines (SVM) models