C-DRONE GUIDE · 10 AUGUST 2026
Forest fire burn severity mapping by drone: multispectral method, prioritising reforestation, price
Once firefighting resources have left and the blaze is declared out, one question remains open for the forest owner, the municipality or the national forestry office (ONF): which plots are burnt to the ground, which only scorched at the understorey with crowns still green, and where does the untouched edge begin? A satellite map answers that question at the scale of the whole forest, with a ten-to-thirty-metre pixel that blurs the nuances within a single plot. A drone, flying low with a multispectral camera, renders that severity tree by tree — enough to target priority reforestation areas, separate timber still worth salvaging from timber that is lost, and document a loss for the insurer. Here is the method, what it adds over satellite imagery, the applicable flight framework and the prices observed in 2026.
Published on 10 August 2026, reviewed on 11 August 2026 — regulations in force as of August 2026.
What satellite imagery misses at plot scale
Reference burn-severity maps rely on the NBR (Normalized Burn Ratio) index and its differenced version, dNBR, computed from Landsat or Sentinel satellite images taken before and after the fire — a proven method at the scale of an entire forest, used in particular by INRAE teams to track large Mediterranean wildfires. But the resolution of these satellites — ten metres for Sentinel-2, thirty for Landsat — blends a fully charred tree and its still-green neighbour into the same pixel as soon as the damage mosaic becomes fine-grained, which is the rule rather than the exception on a fire that crept through the understorey in places and climbed into the crowns elsewhere.
For an owner or a municipality that has to decide, plot by plot, where to cut, where to let natural regrowth take over and where to replant, this coarse resolution is not enough: it locates the damage, it does not detail it. That is exactly the gap a low-altitude drone flight fills, with a ground pixel of a few centimetres instead of several metres.
The method: multispectral imagery and severity classification
The drone flies over the burnt area with a multispectral camera, capturing near-infrared and red-edge bands alongside the visible spectrum to reflect the state of the vegetation: green, well-hydrated foliage strongly reflects near-infrared light, while a charred crown or bare ash-covered ground reflects almost none in that band. From these images, a locally computed NBR-type index, or a simple machine-learning supervised classification, separates several severity classes: charred ground with vegetation fully destroyed, understorey burn with the crown preserved and still green, partial foliage scorching, and untouched areas.
A study by Shin, Seo, Kim, Park and Woo published in 2019 in the journal Forests (MDPI) tested this approach on the Gangneung forest fire in South Korea, which had destroyed 700 hectares in April 2019: using multispectral drone imagery, the authors compared three classifiers (maximum likelihood, spectral angle mapper, NDVI thresholding) and obtained high overall accuracy for all three in separating severity classes, though with some residual confusion between spectrally similar unburnt pine and deciduous stands (see the study on Google Scholar). The principle matches the one detailed in our guide to water-stress mapping by multispectral drone: the same vegetation-index logic, applied here to a damage diagnosis rather than a crop-monitoring one.
Prioritising reforestation, securing salvage logging, documenting the loss
The severity map delivered to the owner or manager serves three concrete purposes. First, prioritising reforestation: fully burnt zones, with no surviving timber stock or nearby seed trees, justify active replanting, whereas an understorey burn with a green crown can often regenerate naturally — a distinction that avoids replanting where the forest is coming back on its own, and leaving unattended the areas that genuinely need it. Second, securing salvage logging: a tree left standing but weakened by fire (a scorched crown, bark split at the base) poses a falling risk to logging crews, which the severity map helps flag before ground work begins — an issue that echoes our guide to storm windthrow mapping, with the difference that burnt timber loses value very fast and forces a tight logging schedule.
Finally, the map serves as a documentary record for the insurance loss adjuster or for a public reforestation grant application: it objectifies the area actually destroyed, broken down by severity class, rather than a broad visual estimate of the loss. On a forest already monitored otherwise, this post-fire mapping complements the routine stand inventory described in our guide to forestry management by drone, and can also be compared over time with an earlier LiDAR forest carbon inventory to quantify the biomass loss.
Flight framework and 2026 pricing
One safety point comes first: as long as active firefighting continues, no private drone should fly near a forest fire — aerial firefighting assets (air tankers, water-bombing helicopters) operate at very low altitude, and an uncoordinated drone can force them to halt drops, with a collision risk and heavy criminal penalties. Drone mapping only takes place once the fire is officially declared out and the area reopened by the fire service. The flight then usually falls under the open category, subcategory A3 for a forest area far from third parties, after checking restricted zones on the Géoportail map — a temporary restriction linked to the fire can still be active in the days following extinction, so check before any take-off.
Ranges observed in France in 2026 (excl. VAT):
| Service | Observed price (excl. VAT) |
|---|---|
| Multispectral mapping up to 20 ha (flight, processing, classified severity map) | €800 to €1,800 |
| Forest area from 20 to 100 ha | €1,800 to €5,000 depending on access and terrain |
| Beyond 100 ha | lower per-hectare rate, quoted |
| Expert report for an insurer (destroyed area quantified by class) | €400 to €900 in addition to the flight |
Our guide to how much a drone service costs covers the parameters that move these amounts. This service falls under our drone surveying and photogrammetry offering; request a quote stating the burnt area, the extinction date and the purpose of the survey (reforestation, salvage logging, insurance).
Frequently asked questions
Can a drone fly while the fire is still active?
No, never. As long as aerial firefighting assets are operating, any private drone in the area poses a collision risk that can force air tankers and helicopters to stop dropping water, and exposes the operator to criminal penalties. Drone mapping only happens once the fire is officially declared out and the area reopened.
Does drone mapping replace a forester's expertise?
No, it makes it objective. The drone produces a classified severity map that is measurable and comparable over time; it is the forester or the national forestry office's expert who interprets that map to decide where to replant, where to leave natural regeneration, and which trees to salvage first before they lose value.
How soon after the fire should the flight take place for a reliable result?
Ideally within the weeks following extinction, before rain washes away the ash and pioneer vegetation starts recolonising the ground, which would blur the spectral reading of severity. A second pass several months later tracks vegetation recovery, whether spontaneous or after replanting.