Home › Methodology

Methodology: how we detect and verify fires

Transparency about what data we use, how we process it and what its limits are.

34Hotspots processed (12 months)
20Fires published (12 months)
0Discarded as false alarms
0Verified by the editorial team

1. Sources

Thermal detections from NASA FIRMS (VIIRS and MODIS), burned-area perimeters from EFFIS (Copernicus), road status from the DGT, and wind and humidity from Open-Meteo. Each data item states its source and capture time.

2. From hotspots to fires

A hotspot is a heat anomaly detected by a satellite in a pixel between 375 m and 1 km. We group hotspots that are close in space and time into a single fire and calculate its FRP (radiative power, in MW), number of detections and dates of first and last detection.

3. Confidence levels

Each detection includes the satellite's confidence. You can filter the map to show only high-confidence ones. Isolated, low-power or non-repeating hotspots are treated with caution; they may be agricultural burns, industry or reflections.

4. Discards

We mark as "discarded" what looks like a non-forest source (for example, industrial facilities or controlled burns). The number of discards in the last 12 months is shown above.

5. "Verified" badge

A fire only gets the badge when the editorial team has checked the detection against an official source or images. No badge means "detected by satellite, not cross-checked".

6. Wind and spread

The wind shown is the current Open-Meteo forecast. The indicated direction of spread is a simple guide (fire tends to advance in the direction the wind blows); it is not a spread model or a smoke forecast.

7. Limitations

This website does not replace the emergency services. In case of danger, call 112.