A new simple fire weather index

Fire behavior analysts and fire managers regularly employ meteorological data to determine how weather may affect the potential for hazardous fire behavior to occur during a wildland fire. Conventional wisdom and experience have shown that fires are more likely to exhibit large growth or erratic behavior when conditions become especially hot, dry, and/or windy. However, while every fire manager wants to know how temperature, relative humidity, and winds will vary over the course of the day such that they could affect a fire, no weather-based fire index that accounts for the potential variability of all these components is widely available for operational use. We will examine whether our simple fire weather index, based only on temperature, humidity, and wind speed, has predictive power for determining days or times of day where weather could significantly enhance the potential for hazardous fire behavior.

We will first describe our rationale for selecting the relative contributions of each component (i.e., an increase in wind speed from 10 kt to 20 kt should have a greater effect on our final index value than a temperature increase from 10 C to 20 C). After defining the index, we will examine its probability of detection and its false alarm rate to establish its predictive power. To determine the probability of detection, we will calculate our index for a selection of historic large or blow-up fire events on the day when the fire growth was observed to be largest. For the false alarm rate, we will consider days before and after the main blow-up day, to verify that our index “lights up” on the days when the largest fire spread was observed. Finally, we will compare the performance of the new index against existing fire weather indices to determine whether it demonstrates more predictive power for this selection of historic large and blow-up fires than the indices that are currently available to the fire community.