Data Guide
Wildfire
Tristan Ballard
Created: Jan 1, 2023 - Updated: Aug 20, 2023
Wildfire is an uncontrolled fire, caused by a combination of combustible fuels (such as dead, dry wood) and ignition sources (such as human activity or lightning). It is strongly influenced by prevailing weather conditions, and typically occurs where human activity meets an accumulation of wild fuel along the Wildland Urban Interface (WUI).
Methodologies
Unified Wildfire
Metric: Annual probability of wildfire at an asset location.
This indicator combines global, high-resolution historic fire observations from a mulit-decadal satellite record with historical data on a variety of factors that contribute to fire risk, including precipitation, temperature, topography, land cover, ignition sources, and ecology. Using peer-reviewed and patented AI technology, our model is able to determine, for every location on earth, the component of fire risk that is weather-driven. This lets us pinpoint the forests of the world where fire risk is most affected by weather conditions which will see sharp increases in fire risk under warming scenarios. We then run model inferences using ensembled simulations of future weather under a variety of scenarios, which lets us provide validated, asset-level estimates of future fire risk.
Observational Dataset: We use satellite derived observations to provide high-resolution exposure to wildfires. The observations are made using NASA MODIS satellites and processed using an active fire algorithm to produce the MCD64A. This covers the historical period of 2001 to present. This indicator shows the years in which a wildfire occurred within 1 kilometer of an asset. For point-based assets, this indicator is either 0 or 1, representing a binary detection of fire. For polygon-based regions, this indicator is a floating point value between 0 and 1, representing the area detected as burned in any given year within the region.
Known Limitations
Sust Global’s fire model relies on static landcover into the future. This assumption is unlikely to hold given human development and changing forest extents, especially later into the 21st century.
Sample Assessment
United States Counties
Projected mean annual wildfire risk exposure 1980-2010 (left) and 2022-2052 (right) with SSP5-RCP8.5 scenario.
| US Counties with top wildfire exposure over 2022-2052 | ||||
|---|---|---|---|---|
| State | County Name | Score (1980-2020) | Score (2022-2052) | Percent Change |
| Idaho | Custer | 0.03 | 0.08 | 143 |
| California | Colusa | 0.07 | 0.08 | 11 |
| California | Glenn | 0.06 | 0.07 | 15 |
| Deer Lodge | Montana | 0.02 | 0.06 | 170 |
| Valley | Idaho | 0.03 | 0.05 | 83 |
Indian Towns
Projected mean annual wildfire risk exposure over 1980-2010 (left) and 2022-2052 (right) with SSP5-RCP8.5 scenario.
| Indian Towns with top wildfire exposure over 2022-2052 | |||
|---|---|---|---|
| Region Name | Score (1980-2020) | Score (2022-2052) | Percent Change |
| Bastar | 0.15 | 0.08 | -45 |
| Bijapur | 0.12 | 0.07 | -43 |
| Dantewada | 0.1 | 0.08 | -25 |
| Garhchiroli | 0.15 | 0.08 | -48 |
| Kondagaon | 0.1 | 0.07 | -31 |