High-impact weather including tornados, hail, aircraft turbulence, and severe wind events, cause significant loss of life and property.  In this talk, I present our recent work on developing and applying data mining techniques to a variety of high-impact weather phenomena.  I will present the Spatiotemporal Relational Random Forest and its applications in depth and overview additional machine learning methods and related applications to weather phenomena.  Although weather is a continuous and dynamic process, meteorologists often study it through discrete high-level features and relationships, which makes it an excellent application domain for spatiotemporal relational data mining. In recent years, there has been a dramatic increase in data available for meteorological study.  This data includes both observations and simulations, which have grown finer in temporal and spatial scales.  Data mining provides an approach to understanding these data and to guiding decisions for prediction of the events.