This study analyses a Sentinel-2 Level-2A image acquired over Monte Faeta, Tuscany, on 29 April 2026.

The objective is to identify the earliest detectable fire-related signal in the satellite scene using spectral indices.

This analysis does not determine the cause or exact ignition point of the wildfire. It only evaluates spatial patterns of fire-related spectral signals at the time of satellite acquisition.

Data source and processing

  • Satellite: Sentinel-2 L2A (Copernicus Data Space)
  • Sensor: MSI (Multispectral Instrument)
  • Resolution: 10 m (bands resampled where needed)
  • Processing: OpenEO Python workflow
  • Area: 2 km by 2 km region over Monte Faeta

All spectral bands were resampled to a common 10 m grid before index calculation.

True colour image

Sentinel-2 true colour composite

The true colour composite (B04, B03, B02) shows vegetation, smoke presence, and surface disturbance. Active fire is not clearly visible in this representation due to atmospheric effects and band sensitivity.

SWIR response

Sentinel-2 SWIR composite

Shortwave infrared bands (especially B12) are sensitive to high temperature surfaces and burned vegetation.

Elevated SWIR reflectance in the study area indicates fire-related thermal disturbance or recent burning.

Burn detection index

The Normalized Burn Ratio (NBR) is used to detect burned or fire-affected vegetation:

Sentinel-2 NBR severity

Lower values of NBR indicate burned or heat-affected surfaces.

Earliest detected fire signal

A single pixel shows the strongest and most consistent fire-related spectral response in the scene.

Location: 43.7714°N, 10.4765°E
Spatial resolution: 10 m (Sentinel-2 MSI)

At this location:

  • SWIR reflectance is elevated, consistent with thermal disturbance
  • NBR is strongly negative, consistent with burned or fire-affected vegetation
  • MIRBI is positive, consistent with fire-related surface conditions

These indicators are spatially aligned within the same pixel.

Interpretation

This pixel represents:

The earliest detectable fire-related signal in the Sentinel-2 observation of this area.

This does not mean it is the ignition point.

The following limitations apply:

  • Sentinel-2 does not measure temperature directly
  • Each pixel represents a 10 m ground area, which may contain mixed land cover
  • Fire spread dynamics can shift the location of maximum signal after ignition
  • The observation is limited to a single satellite overpass in time

Conclusion

Within the analysed Sentinel-2 scene:

  • One pixel shows a consistent and strong fire-related spectral signature
  • Multiple independent indices agree on this location
  • This pixel is the highest-confidence fire signal in the dataset

Therefore, it is the earliest detectable fire signal at satellite resolution in this image.