authority: folia
slug: burn-severity
name: Burn Severity
domain: land
subdomain: biosphere
authority_id: folia:burn-severity
parent: fire
aliases:
  - dnbr
  - rdnbr
external_ids:
  cf_standard_name: null
canonical_units: categorical
external_validations:
  product:
    id: usgs-mtbs
    docs_url: https://www.mtbs.gov/
  library:
    name: FIREMON Landscape Assessment (Key & Benson 2006)
    id: firemon-la-dnbr
    docs_url: https://www.fs.usda.gov/rm/pubs/rmrs_gtr164/rmrs_gtr_164_13_land_assess.pdf
  sentinel_hub:
    processor: burn_severity
    docs_url: https://custom-scripts.sentinel-hub.com/sentinel-2/burn_severity/
description: |
  Classified raster of fire impact derived from the differenced Normalized
  Burn Ratio (dNBR) and its relative form (RdNBR). The canonical MTBS class
  breaks on the scaled (×10³) dNBR axis are:

  | code | class | dNBR range (scaled ×10³) |
  |------|-------|--------------------------|
  | -1 | Increased greenness / regrowth | < -100 |
  | 1 | Unburned to low | -100 to +99 |
  | 2 | Low severity | +100 to +269 |
  | 3 | Moderate severity | +270 to +439 |
  | 4 | High severity | +440 to +659 |
  | 5 | Increased greenness (post-fire) | ≥ +660 |

  dNBR (delta Normalized Burn Ratio) is `NBR_pre - NBR_post`, where
  `NBR = (NIR − SWIR2) / (NIR + SWIR2)`. The RdNBR (relative) variant
  normalizes by pre-fire NBR to reduce the effect of pre-fire vegetation
  cover. MTBS uses USGS analyst review of Landsat scene pairs and ships
  pre-computed annual aggregates for the US; Sentinel Hub and GEE recipes
  compute dNBR automatically per pre/post-fire scene pair. Used by
  post-fire recovery, ecology, insurance, and emissions ops.
example_recipes:
  - title: "Compute dNBR from pre/post-fire Sentinel-2 scenes (Earth Engine)"
    tool: gee
    requires: []
    snippet: |
      var nbr = function(img) {
        return img.normalizedDifference(['B8', 'B12']).rename('NBR');
      };
      var pre  = ee.Image('COPERNICUS/S2_SR/PRE_FIRE_SCENE_ID');
      var post = ee.Image('COPERNICUS/S2_SR/POST_FIRE_SCENE_ID');
      var dnbr = nbr(pre).subtract(nbr(post)).multiply(1000).rename('dNBR');
    docs_url: https://custom-scripts.sentinel-hub.com/sentinel-2/burn_severity/
  - title: "Compute dNBR from pre/post-fire NBR rasters (GDAL)"
    tool: gdal
    requires: []
    command: |
      gdal_calc.py -A nbr_pre.tif -B nbr_post.tif --outfile=dnbr.tif \
        --calc="(A-B)*1000" --type=Int16
    docs_url: https://gdal.org/programs/gdal_calc.html
theme: hazards
consumed_by: []
produced_by: [analysis_dnbr]
demos: []
status: documented
sources:
  - https://www.mtbs.gov/
  - https://www.fs.usda.gov/rm/pubs/rmrs_gtr164/rmrs_gtr_164_13_land_assess.pdf
  - https://custom-scripts.sentinel-hub.com/sentinel-2/burn_severity/
  - Key, C.H. & Benson, N.C. 2006, USDA Forest Service General Technical Report RMRS-GTR-164-CD (FIREMON Landscape Assessment, dNBR methodology)
  - Eidenshink et al. 2007, Fire Ecology 3(1), 3-21 (MTBS project overview)