Burn Severity Mapper
App: Burn Severity Mapper
Manual: Burn Severity Mapper user manual
Burn Severity Mapper is a Google Earth Engine app developed within SeverusPT to facilitate the production of burn severity maps from optical satellite imagery. The app implements a pre- versus post-fire comparison based on the Normalized Burn Ratio (NBR), reducing the amount of manual work usually required to search, subset, filter, composite, process, and export satellite data.
The app is optimized for mainland Portugal, where ancillary burned area datasets from EFFIS and ICNF can be used to support fire selection and interpretation. The core workflow, however, can be applied to any user-defined area where suitable pre- and post-fire imagery is available.
Main functionality
- Define an Area of Interest using interactive geometries.
- Select Sentinel-2, Landsat 8, or Landsat 9 imagery.
- Configure pre- and post-fire periods using a fire date, explicit date ranges, or a same-season comparison with the previous year.
- Filter satellite images by cloud cover and inspect the selected image list.
- Generate dNBR, RdNBR, RBR, and categorical burn severity layers.
- Visualize continuous and classified severity outputs in the map interface.
- Segment the burned perimeter using a threshold-based procedure that combines dNBR and Change Vector Analysis magnitude.
- Calculate total burned area and burned area by severity class.
- Download outputs as GeoTIFF or KML where Earth Engine export limits allow it.
Burn severity indicators
The app uses NBR, calculated from near-infrared and shortwave-infrared reflectance, as the basis for comparing pre- and post-fire conditions. It then derives three related burn severity indicators:
- dNBR: the difference between pre-fire and post-fire NBR.
- RdNBR: a relativized dNBR indicator that accounts for pre-fire NBR conditions.
- RBR: a relative burn ratio designed to normalize dNBR by pre-fire conditions.
The categorical dNBR output uses five severity classes: unburnt/very-low, low, moderate, high, and very-high. These classes provide a rapid way to interpret the spatial pattern of fire effects and to calculate area summaries by severity level.