California Fire Severity Prediction Maps from 1984 to 2022

A spatial database of potential fire severity raster datasets for California, produced by the California Fire Severity Prediction Mapping Project. The data uses Random Forest models to classify burnable landscapes into low, moderate, or high severity potential for eight distinct regions. The project was authored by Stacy A. Drury and uses data from 1984 to 2022.

Use Cases

• Modeling potential vegetation mortality based on spatial, temporal, and environmental predictors.
• Planning mechanical fuels treatments based on regional fire severity potential maps.
• Assessing post-fire landscape conditions for areas that have burned since 1984 using the reburn model.
• Informing wildfire risk mitigation strategies using combined potential fire severity raster layers.

Strengths

• Covers eight distinct California regions: Klamath, North Coast, Northeast, North Sierra, East Sierra, South Sierra, South Coast, and Central Coast.
• Includes 16 total Random Forest models (non-reburn and reburn models for each region) based on data from 1984 to 2022.
• Provides source R and Google Earth Engine scripts for reproducibility of the raster production.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

Stacy A. Drury via figshare.

Collection Method

Modeled using the Random Forest platform on spatial, temporal, and environmental information.

Time Range

1984 to 2022

Freshness

Last updated 2026-05-22 05:45:36; freshness should be verified.

Geography

California, USA, divided into eight regions.