Season statistics 2026
The season of the working area in figures: satellite detections, recorded burn scars, the course of fire danger and the emissions estimated from them. Every figure carries its source, timestamp and uncertainty.
As of 18:03 (Sarajevo) · observed since July 27
How often the air is above the limit: air statistics by station
54,236
Detections · season 2026
As of 18:03 (Sarajevo) · VIIRS · SLSTR · FIRMS · GWIS
1,709
Burn scars recorded · 2026
As of 18:03 (Sarajevo) · EFFIS
141,468
Burned area · 2026ha
As of 18:03 (Sarajevo) · EFFIS
12,707
Largest single event · 2026ha
As of 18:03 (Sarajevo) · EFFIS
Season timeline
Satellite detections per week and state since the start of the season. One fire produces many detections: the series shows observation pressure, not a fire count; counted scars are in the country comparison below.
VIIRS · SLSTR · FIRMS · GWIS
The taller the bar, the more satellite detections in that week. One large fire produces many detections, what is counted is detections, not fires.
Observed from July 27, 2026. For 30 of the 41 weeks in this season no retrieval exists, they are marked as not observed, not as a zero week.
Where the detections come from
The curve counts detections, and detections are the sum across several satellites. If one spacecraft drops out, the number falls without anything burning less. So this is where you see how much hangs on which one, not after the outage, but before it.
- NOAA-20 (VIIRS)32.0%24,598 detectionslast Oct 7, 02:08 PM
- NOAA-21 (VIIRS)30.2%23,214 detectionslast Oct 7, 02:53 PM
- Suomi-NPP (VIIRS)31.6%24,303 detectionslast Oct 7, 01:49 PM
- Sentinel-3 (SLSTR)6.1%4,707 detectionslast Oct 6, 10:56 PM
- Aqua (MODIS)0.0%0 detectionsnot in the current fetch plan · shutdown announced from 2026-10
- Terra (MODIS)0.0%0 detectionsnot in the current fetch plan · shutdown announced from 2027-02
If one of these sources falls silent, a kink in the curve is a change of source, not a decline in fire activity. For MODIS Aqua the date is known: NASA begins passivation in autumn 2026. The VIIRS share and Sentinel-3 carry on unaffected.
Country comparison
Recorded burn scars and their total per state. Recording starts at roughly one hectare; small and quickly extinguished fires are systematically missing.
EFFIS
| State | Area | Count | Share | |
|---|---|---|---|---|
| Bosnia and Herzegovina | 63,303ha | 605 | 44.7% | |
| Montenegro | 22,544ha | 383 | 15.9% | |
| North Macedonia | 19,824ha | 155 | 14% | |
| Croatia | 16,215ha | 152 | 11.5% | |
| Serbia | 16,106ha | 304 | 11.4% | |
| Kosovo | 3,077ha | 102 | 2.2% | |
| Slovenia | 399ha | 8 | 0.3% | |
| Total | 141,468ha | 1,709 | 100% |
Fire danger over the season
Daily mean of the Canadian Fire Weather Index per state, coloured by the six EFFIS danger classes. The basis is the country-level record: an approximation over the state envelope, not an area statistic.
EFFIS mf010 · GWIS ecmwf
Fire danger: colour ramp (FWI)no unit
- low< 11.2
- moderate11.2-21.3
- high21.3-38.0
- very high38.0-50.0
- extreme50.0-70.0
- very extreme> 70.0
Daily FWI mean per state (season 2026); classification of the daily mean by the EFFIS thresholds, own processing. Driving model: 413 daily values from EFFIS (Meteo-France 0.1°), 0 from GWIS; exactly one model per day and state, never mixed. 1,547 daily values are missing (blank cells): no country-level run exists for those days.
Each cell is one day. The darker the colour, the higher the fire danger; a hot summer stands there as one continuous dark stripe.
Largest single events
The ten largest contiguous burn scars of the season. Area and place come from the EFFIS perimeter; perimeters arrive days after the fire.
EFFIS
| Area | Municipality | State | First recorded | |
|---|---|---|---|---|
| 1 | 12,707ha | Tomislavgrad municipality | Bosnia and Herzegovina | Sep 4, 2026 |
| 2 | 5,100ha | Prilep | North Macedonia | Aug 17, 2026 |
| 3 | 3,992ha | Nerežišća | Croatia | Sep 11, 2026 |
| 4 | 2,854ha | Nevesinje municipality | Bosnia and Herzegovina | Aug 16, 2026 |
| 5 | 2,675ha | City of Trebinje | Bosnia and Herzegovina | Aug 20, 2026 |
| 6 | 2,536ha | Bileca municipality | Bosnia and Herzegovina | Sep 3, 2026 |
| 7 | 1,990ha | Glamoč municipality | Bosnia and Herzegovina | Sep 5, 2026 |
| 8 | 1,879ha | Gadžin Han | Serbia | Sep 4, 2026 |
| 9 | 1,832ha | City of Livno | Croatia | Sep 5, 2026 |
| 10 | 1,744ha | City of Trebinje | Bosnia and Herzegovina | Aug 20, 2026 |
Biomass and CO₂ balance
Burned biomass and greenhouse gases estimated from the burn scars following IPCC 2006. Every quantity is a range: total uncertainty is 55 to 70 percent; the dot marks the mean, the band the range.
EFFIS · ESA CCI Biomass · IPCC 2006
Every figure is an estimate, not a measurement. The dot is the mean, the band shows how far the true value may lie from it.
140,000haBurned area · 2026
110,000 to 180,000
Calculation
Burned area recorded by satellite. Uncertainty applied 25 percent; 20 to 30 percent are reported for this method. Detection starts at roughly one hectare.
2,000,000t dry matterBiomass consumed · 2026
1,300,000 to 2,600,000
Calculation
Sum over 7 vegetation groups. The group uncertainties are added in quadrature, which assumes their errors are independent of one another.
3,400,000t CO₂Greenhouse gases · 2026
2,200,000 to 4,600,000
Calculation
Sum over 7 vegetation groups. The group uncertainties are added in quadrature, which assumes their errors are independent of one another.
3,400,000t CO₂eAs CO₂ equivalent · 2026
2,300,000 to 4,500,000
Calculation
Sum over 7 vegetation groups. The group uncertainties are added in quadrature, which assumes their errors are independent of one another.
Vegetation affected
| Vegetation affected | Burned area | Share | Greenhouse gases |
|---|---|---|---|
| Grassland | 50,000ha | 35.1% | 0 to 1,700,000 |
| Transitional woodland and shrub | 32,000ha | 22.3% | 250,000 to 1,300,000 |
| Agricultural land | 26,000ha | 18.6% | 0 to 910,000 |
| Broadleaved forest | 21,000ha | 15.1% | 440,000 to 1,000,000 |
| Sclerophyllous vegetation and maquis | 11,000ha | 7.6% | 340,000 to 640,000 |
| Mixed forest | 1,100ha | 0.8% | 23,000 to 55,000 |
| Coniferous forest | 810ha | 0.6% | 18,000 to 42,000 |
No biomass map is available for 7 vegetation groups. Those are calculated from the IPCC Tier 1 table values, coarser, but the route the guidelines provide for this case.
No vegetation data is available for 34 ha from 20 burn scars. That area does not enter the emission estimate: assuming a class would not be an estimate but an assertion.
Basis of the balance: 1,709 recorded burn scars across the seven states.
Uncertainty
Four uncertain quantities enter the estimate: the satellite-derived burned area, the biomass density per pixel, the combustion factor of the vegetation class and the emission factor. Combined, this gives an overall uncertainty of 55 to 70 percent, roughly a factor of 1.6 to 2.
All derived quantities are rounded to two significant figures. Further digits could not be substantiated. Inputs: burned areas from EFFIS, land cover from ESA WorldCover, above-ground biomass from ESA CCI Biomass, methodology after IPCC 2006.
Observation quality
Every data fetch is logged, including those with zero hits. An empty map reads two ways: “observed, nothing detected” is an all-clear, “not observed” is a blind spot. The table shows how reliably each source delivered this season.
Season 2026
| Source | Fetches | Failed | Success rate | Last success | State |
|---|---|---|---|---|---|
| Detections · NASA FIRMS (area API)FIRMS_AREA_API | 1 | 0 | 100.0% | Aug 2, 06:53 PM | overdue |
| Detections · NASA FIRMS (regional file)FIRMS_STATIC_EUROPE | 8,988 | 0 | 100.0% | Oct 7, 06:00 PM | observed |
| Detections · GWIS (includes Sentinel-3)GWIS_HOTSPOT | 8,987 | 40 | 99.6% | Oct 7, 06:00 PM | observed |
| Burn scars · EFFISEFFIS_BURNT_AREA | 66 | 11 | 83.3% | Oct 7, 03:17 AM | observed |
| Fire danger · EFFIS (Meteo-France 0.1°)EFFIS_FWI | 66 | 6 | 90.9% | Oct 7, 03:17 AM | observed |
| Fire danger · GWIS (ECMWF)GWIS_FWI | 0 | 0 | no data | never | not observed |