Fire Probability for Carbon Accounting
The Fire and Resource Assessment Program (FRAP) at CAL FIRE is engaged in investigating various methods for estimating the likelihood of wildfire occurrence across the State of California in the coming decades. In collaboration with researchers from the University of California and The George Washington University, FRAP has developed a map of annual wildfire probability for the period 2021-2050. This map is being provided to assist in the quantification of GHG benefits associated with fuel reduction activities funded under the CAL FIRE-California Climate Investments (CCI) Forest Health Grants Program.
This data represents mean projected annual probability (%) of wildfire occurrence for the period 2021-2050, developed by researchers at the University of California and The George Washington University with support from CAL FIRE and California Climate Investments (CAL FIRE contract #8CA06938). Probability was projected annually through 2099 using each of four climate models (CanESM2, HadGEM2-ES, CNRM-CM5, and MIROC5) under the RCP 8.5 emissions scenario at 1 km resolution. Predictive models were trained using a GAM framework on a spatially randomized subset of data from 1980-2019, and incorporate predictions of actual evapotranspiration, climatic water deficit, and local housing density, as well as distance to roads and electrical infrastructure, area of cultivated land, and historical fire perimeters. Mean annual probability of wildfire was derived from annual projections over the thirty-year period 2021 - 2050 and across all four climate models. 1 km grid cells with >=50% cultivated land, urban cover, water, or barren/rock were excluded in the model or ex post facto for purposes of this dataset.