The terraPulse Wildfire dataset estimates the probability of fire occurrence as a function of land cover, terrain, plant phenology, and climate. The algorithm is trained on over 20 years of fire history, by ecoregion, in correlation to the suite of terraPulse datasets as predictors. Certainty is represented by Root Mean Square Error.
Data products are available at up to daily, 10-meter resolution or as longer-term (e.g. annual or multi-decadal) summaries at pixel or coarser resolutions.