Wildfire Theme Services

Example of fire severity in a burned area. Legend: darker shades indicate higher severity of the fire. Credits: Terradue

The Burned Area Mapping (BAM) service identifies active fires and quantifies burned areas using Copernicus Sentinel-3 SLSTR thermal data and Sentinel-2 MSI optical imagery.

Developed by Terradue, the service processes thermal anomalies daily, generating 10-metre resolution maps of burned extent, probability, and severity. The FireCCISFD20 algorithm is adapted to integrate spectral indices (MIRBI and NBR2) with VIIRS active fire data for the extraction of the Copernicus Sentinel-2 based products.

Once hotspots are detected, the service automatically triggers Sentinel-2-based burned area delineation. It produces thematic layers, including burned/unburned zones, severity (using Relativised Burn Ratio), and per-pixel burn probabilities. This supports rapid post-fire assessments and emergency planning.

Outputs include aggregated detected hotspots, burned area classification layer, burned area probability, and burned area severity calculation.

It is executed systematically but can support on-demand processing. The service is open source and available at 10 m resolution every 5 days (for Sentinel-2 based outputs) and daily (for Sentinel-3 based outputs).

Diagram of the fire danger service. Example of Fire Danger Map over Perú. Legend: warmer colors indicate higher fire danger. Credits: Indra

Developed by Indra, the Fire Danger Mapping (FDM) service integrates EO-based static and dynamic inputs to estimate daily fire danger levels across Latin America and the Caribbean.

It combines land cover, vegetation height, topography, and fire history with daily Fire Weather Index (FWI) from the Global Wildfire Information Service (GWIS). Outputs are 100 m resolution raster maps classifying fire danger into five levels, from low to extreme.

This open-source service aids in anticipating high-risk areas and supports proactive fire prevention strategies. It enables regional customisation by adjusting variable weights. Execution is currently systematic but can also be activated on-demand, with forecasts available up to 7 days in advance.

Diagram of the fire recovery service. Example of a two-year regeneration map. Legend: warmer, red colors indicate null or minor recovery; greener colors indicate higher recovery (in percentage).Credits: Indra

The Fire Recovery Mapping (FRM) service monitors vegetation recovery using Copernicus Sentinel-2 NDVI data.

Developed by Indra, it begins once hotspots from the BAM service cease. By comparing post-fire NDVI to pre-event baselines, the service calculates a recovery index expressed as percentage per pixel (0% to 100%).

Delivered at 10 m resolution, the outputs include monthly NDVI composites and a cumulative recovery raster, highlighting spatial regeneration rates. The service supports post-fire ecological restoration and resource planning, with data updates ranging from 15 to 380 days post-event. The service is open source and systematically executed, but also supports on-demand use.

It employs the Normalized Burn Ratio (NBR), delta NBR (dNBR), and the Relativised Burn Ratio (RBR) to produce maps reflecting fire impact intensity.

The user provides pre- and post-event imagery (RED, NIR, SWIR2 bands). The output includes NBR indices, severity maps, and RGB overviews. It supports fire impact assessments and guides restoration efforts, delivering results at the native resolution of input imagery.

It computes coherence and backscatter to create visual composites that highlight surface changes. Developed by Terradue, it produces high-resolution geophysical assets and RGB visualisations.

Key outputs include Coherence maps, average Sigma Nought layers and Red-Cyan backscatter composites.

Used in wildfire contexts, SAR-COIN detects changes such as burnt infrastructure or altered vegetation structure. The service is open source and executed on-demand.

It uses a sieve algorithm to eliminate small pixel clusters and creates vector layers for selected classes. This tool is especially useful for smoothing classified wildfire outputs before further analysis.

Modes include filtering only, vectorising only, and combined filtering and vectorising.

It is executed on-demand and supports raster-to-vector conversion for burned area and severity maps.

It computes NDVI differences, applies thresholding, filters noise, and vectorises results. Developed by Terradue, the service works with co-registered multi-mission data (e.g., Landsat, Copernicus Sentinel-2).

The outputs include NDVI loss rasters, polygon layers, and overviews for quick interpretation. This on-demand service is open source and effective for delineating fire damage zones.

It supports inputs from different sensors or processors (e.g., NDVI, burned area, land cover) and enables generation of derived products using symbolic expressions.

This open-source, on-demand service facilitates post-processing workflows, such as combining burned area and vegetation recovery data into integrated fire impact layers.