Development of a Statistical Validation Methodology for Fire Weather Indices

Abstract

Fire managers and forecasters must have tools, such as fire indices, to summarize large amounts of complex information. These tools allow them to identify and plan for periods of elevated risk and/or wildfire potential. This need was once met using simple measures like relative humidity or maximum daily temperature (e.g., Gisborne, 1936) to describe fire weather, and with increasing decision-support requirements over time, eventually led to more complex systems like the National Fire Danger Rating System (NFDRS). When there was no readily available index for some fire or fire-weather characteristic, managers and forecasters borrowed what they could find from related fields. For example, the K-Index and Lifted Index, designed for prediction of severe weather linked to thunderstorms, were used to indicate stability before the development of the Haines Index. High values of these non-fire indices resulted from high moisture as well as high instability, and as such were not ideal for fire weather forecasting.

Citation

Potter, Brian E.; Goodrick, Scott; Brown, Tim. 2003. Development of a Statistical Validation Methodology for Fire Weather Indices. In: 2d International Wildland Fire Ecology and Fire management Congress: 5th Symposium on Fire and Forest Meteorology; 2003 November 16-20; Orlando, FL. Boston, MA: American Meteorological Society: J11.1.