FSim burn probability
One way to map the likelihood of wildfires is to use a stochastic model of realistic wildfire simulations. The Large Scale Fire Simulation Project1, conducted by the USDA Forest Service Rocky Mountain Research Station, does exactly this. Resulting spatial information is included in the Frequency group as a supporting model of future wildfire probability.
Historic fire perimeters
The best way to predict what may happen in the future is to evaluate the pattern of wildfires from the past. The Geospatial Multi-Agency Coordination Group of the USGS, recently replaced by the National Interagency Fire Center (NIFC) Wildland Open Fire Data website, provides spatial data for wildfire locations since 2000. This data consist of historic ignition points and wildfire perimeters.
Wildfire perimeters are used to create a regional map of where wildfires occurred in the past, and therefore, where they are likely to occur in the future. Wildfire perimeters are aggregated to create a heat map showing past fire cluster density.
Probability of ignition
Based on climate (temperature and humidity), vegetation cover, aspect, and elevation, the probability of ignition is a measure of how likely an ember is to ignite in fine vegetation and start a fire. Places that are historically drier, unshaded, and with more of a southern exposure are more likely to ignite than wetter, shaded, north-facing areas. This varies over a finer scale than other frequency factors due to high-resolution assessment of topography across the landscape.
Distance to nearest transmission line
Electrical transmission lines have a long history of causing wildfires. Areas closer to transmission lines, therefore, are more likely to experience a fire than those that are farther away. Electrical transmission line data from the HIFLD dataset was buffered and the distance to the nearest transmission line included in the risk rating. Beyond the maximum buffer distance of one mile, point values return to background levels.