A combined prediction of the risk of dry thunderstorm outbreaks

Lightning is a major cause of wildfires, and is disproportionately responsible for igniting major wildfires in the continental US and Alaska. Additionally, dry thunderstorms (those that occur without a significant accumulation of rainfall at the surface) are responsible for outbreaks of wildfires in the western US. We have developed two statistical methodologies which use model-generated data to predict the risk of lightning occurrence. The first uses a maximum likelihood algorithm to predict the probability of dry thunderstorms, and was developed from lightning strike data and upper-air and surface meteorological observations. The second uses lightning strike data and a suite of model-output variables to create a perfect prognosis forecast of the likelihood of occurrence of at least 1 and at least 10 lightning strikes. We have combined these predicted probabilities over the western US (where dry thunderstorms are a particular threat) to produce a single risk prediction of dry thunderstorm outbreaks, using the 12-km NAM model out to 84 hours. We will describe the development and verification of these algorithms, and present examples of the output.