Wildfire

terraPulse Wildfire is a set of global geospatial data products that estimate where wildfires are likely to occur and what they put at risk. The core Wildfire layer gives the probability of fire occurrence as a function of vegetation cover and type, terrain, plant phenology and climate. Its algorithm is trained by ecoregion on more than 20 years of fire history, using terraPulse's data cube, mined from satellite measurements with machine learning, as predictors. Terms follow Hardy (2001).

The Wildfire layer is available at up to 10 m resolution with global extent, up to daily from 2017, with a latency of about two weeks. Wildfire History records fire occurrence and burned area at up to 500 m, daily from 2001. Wildfire Hazard, Danger, Exposure, Vulnerability (Risk) and Severity are available at up to 10 m, global, up to daily from 1984, updated for the previous year. Hazard is the probability of ignition given fuel and terrain; Danger adds weather and is also offered as future projections from predicted meteorological data. Exposure and Risk express what is at stake as biomass or other asset values such as habitat or tree cover, and Severity measures the change in those values or a burn index such as dNBR.

Layers can be explored on an online map with point and polygon queries and CSV download, and are delivered by SFTP, through the terraServe API (GeoJSON, CSV, SHP or TIFF) or in the terraView dashboard. Orders are placed by contacting terraPulse.