Methods to Predict Thunderstorm Development 1-6 Hours into the Future Using Machine Learning and Other Approaches Docket: UAH-P-14010 Technology Despite extensive research on nowcasting convective storm initiation (CI), the key factors involved in the CI process are not well understood, making 0-6 hour forecasting challenging for socially relevant space and time scales (10's of km, 30 min). Forecasting thunderstorms in such short timeframes by numerical weather prediction (NWP) models is hindered by a heavy reliance on extrapolation of real-time observations, which does not work well in the 2-6 hour timeframe. Researchers at UAH have developed software for forecasting first-time CI in the 1-6 hour timeframe. This innovation involves building a "CI footprint" composed of gridded maps of convective storm development over the next 1-6 hours. The maps are created using machine learning and statistical methods that incorporate many input parameters, fields, and association rules to arrive at the best predictors for a given day and time. This "CI footprint" concept significantly improves the accuracy and reliability of CI nowcasts compared to previous methods that relied on NWP models. This capacity is unique in the current marketplace, giving the invention an additional competitive advantage. Applications Aviation weather forecasting Short-term weather prediction Numerical weather prediction (NWP) of convective storm development, timing, and location Early detection of severe storms Cell phone weather forecasting applications Advantages Improves NWP model forecasts Status State of Development: Proof of concept Licensing Status: Not available for licensing Patent Status: Patented