Burned area
satellite only5,225ha
Perimeter published by GWIS (Copernicus).
- Source
- GWIS (Copernicus)
- Date
- 4 Sept 2026 – 16 Sept 2026
+− cross-check
No official record is matched to this page to compare it with.
MG · Madagascar · –
Antsohihy fire, 4 September 2026 Antsohihy, Analalava
Out · last detection 16 Sept 2026, 22:31 UTC
The fire started on 4 September 2026 in Antsohihy, Madagascar, and satellites saw it burning for 13 days. Copernicus’s global GWIS system rebuilt a burnt perimeter of 5,225 hectares from heat detections, spread across 2 municipalities.
Number 11 in Madagascar in 2026, of 110.
GWIS near-real-time perimeter, from the layer that changed in August 2025: this fire is not compared with fires from earlier years, because since then the GWIS layer gives more and larger perimeters.
Use this data
Compare with…
Suggestions: Antsohihy · 2021Antsohihy · 2019Antsohihy · 2025
Use this data
Fire page data · version 2026-10-24matches the official recordwith caveatssatellite onlymodel
When copying, the sentence is
Tick the figures you want: they make up the lede and the CSV.
Duration by satellite
satellite only13days
4 Sept 2026 to 16 Sept 2026: first and last day with fire on the page.
- Source
- GWIS (Copernicus) · NASA FIRMS · VIIRS
- Date
- last detection: 16 Sept 2026, 22:31 UTC
+− cross-check
Satellites pass a few times a day and cloud or smoke can hide the fire: a lack of detections is not an official extinction. Each country’s emergency service declares a fire contained or out.
Start
satellite only4 Sept 2026
First day with fire seen by the satellites or the perimeter.
- Source
- GWIS (Copernicus) · NASA FIRMS · VIIRS
- Date
- 4 Sept 2026
+− cross-check
No official record is matched to this page, so the alert time and the cause are not known here.
Hotspots
satellite only163
VIIRS detections at 375 m. Most intense hotspot: 95 MW.
- Source
- NASA FIRMS · VIIRS (Suomi NPP)
- Date
- 4 Sept 2026 – 16 Sept 2026
+− cross-check
There is no official record to compare with. Each hotspot is a pixel detected in one satellite pass, not a separate fire.
The worst day
satellite only1,007MW
Total radiative power on 10 September 2026; hotspots that day: 45.
- Source
- NASA FIRMS · VIIRS (FRP)
- Date
- 10 Sept 2026
+− cross-check
Fire radiative power (FRP) measures the energy the fire gives off when the satellite passes. Added up by day, it compares intensity, not area.
Danger at the start
model38.9FWI · very high
Fire Weather Index (Canadian system) computed by WorldFires for the day the fire started. Consecutive days without rain before: 19.
- Source
- WorldFires calculation (Canadian FWI system)
- Date
- 4 Sept 2026, local noon
- Model
- ERA5 0,25° (ECMWF · Copernicus C3S) · Open-Meteo.com
+− cross-check
This is an index computed from weather model data, not a measurement of the fire nor an official warning. Danger warnings are issued by each country’s authority. Emergency number and official warnings
Weather during the fire
model32.8°C · 31% RH
High of 32.8 °C and minimum humidity of 31% on 9 September 2026; gusts of 62 km/h on 16 September 2026.
- Source
- weather model at the fire’s grid cell
- Date
- 4 Sept 2026 – 16 Sept 2026
- Model
- ERA5 0,25° (ECMWF · Copernicus C3S) · Open-Meteo.com
+− cross-check
These are model values for a cell of about 25 km, not from a weather station: they may differ from what was measured on the ground.
Living inside
model≈ 1,000people
Estimated residents inside the perimeter and 2.4 ha of built-up area (100 m grid, epoch 2025).
- Source
- GHSL R2023A · JRC
- Date
- calculated on 24 October 2026
+− cross-check
This is an estimate on 100 m cells, made from satellite imagery and census data: it does not mean that homes burned or that anyone was harmed. The perimeter also contains land that never burned.
This page’s change log has been kept since 24 October 2026: every night what it publishes is compared and the difference is logged. Anything earlier is reconstructed from each source’s date and marked as reconstructed.
Show
-
· reconstructed GHSLPopulation and built-up area computed
-
· reconstructed weatherWeather and FWI computed from the ERA5 reanalysis
What was saved on 24 October 2026
Every night a snapshot of what the page shows is saved. If anything changes, it appears in the list as a log entry.
- Area
- 5,225 ha
- Dates
- 4 Sept 2026 – 16 Sept 2026
- Name
- Antsohihy
- Official record
- —
Versioned permanent link
https://worldfires.org/en/fire/antsohihy-2026-09?v=2026-10-24A version keeps the area, dates, name and official record; the area as it was logged (changes under 2% are not logged). The map, hotspots and weather always show today’s data.
Citation
A lede of about 60 words figures only, no adjectives
CSV for this fire antsohihy-2026-09.csv · daily, no individual hotspots
| fecha | focos_viirs | frp_mw | tmax_c | tmin_c | hr_min_pct | viento_max_kmh | racha_max_kmh | precip_mm | fwi | clase_fwi | dentro_del_incendio |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 2026-09-04 | 1 | 1 | 31.4 | 21.2 | 32 | 20.4 | 41 | 0.3 | 38.9 | muy_alto | 1 |
| 2026-09-05 | 5 | 17 | 31.8 | 20.5 | 32 | 21.1 | 38.9 | 0 | 39.8 | muy_alto | 1 |
| 2026-09-06 | 18 | 245 | 30.6 | 19.7 | 31 | 21.3 | 39.6 | 0 | 40.2 | muy_alto | 1 |
| 2026-09-07 | 1 | 1 | 31 | 20.1 | 36 | 15.7 | 28.4 | 0.7 | 29.7 | alto | 1 |
| 2026-09-08 | 14 | 119 | 32.3 | 21.1 | 31 | 26.3 | 55.8 | 0 | 44.3 | muy_alto | 1 |
| 2026-09-09 | 54 | 887 | 32.8 | 20.5 | 31 | 29.4 | 59 | 0 | 48.6 | muy_alto | 1 |
| 2026-09-10 | 45 | 1,007 | 32.3 | 21.4 | 31 | 29.7 | 57.2 | 0 | 49.2 | muy_alto | 1 |
| 2026-09-11 | 4 | 6 | 32.1 | 20.6 | 30 | 25 | 45.7 | 0 | 44.3 | muy_alto | 1 |
| 2026-09-12 | 2 | 11 | 31.6 | 20.3 | 31 | 15.8 | 31.3 | 0 | 35.8 | alto | 1 |
| 2026-09-13 | 1 | 11 | 31.4 | 20.4 | 30 | 13.8 | 23.4 | 0 | 34.3 | alto | 1 |
| 2026-09-14 | 1 | 4 | 32.7 | 20.1 | 28 | 29.1 | 52.9 | 0 | 50.8 | extremo | 1 |
| 2026-09-15 | 2 | 2 | 32.3 | 21.3 | 28 | 30.1 | 58.7 | 0 | 52.3 | extremo | 1 |
| 2026-09-16 | 15 | 82 | 31.2 | 21.5 | 31 | 32.6 | 61.6 | 0 | 55.6 | extremo | 1 |
| 2026-09-17 | 0 | 0 | 31.5 | 19.5 | 30 | 27.4 | 52.6 | 0 | 49.7 | muy_alto | 0 |
| 2026-09-18 | 0 | 0 | 32.4 | 20.7 | 31 | 25 | 47.5 | 0 | 47.2 | muy_alto | 0 |
| 2026-09-19 | 0 | 0 | 32.6 | 20.1 | 27 | 20.2 | 38.2 | 0 | 42.4 | muy_alto | 0 |
| 2026-09-20 | 0 | 0 | 32.2 | 20.1 | 34 | 25.1 | 48.6 | 0 | 47.8 | muy_alto | 0 |
| 2026-09-21 | 0 | 0 | 31.7 | 19.4 | 31 | 24.8 | 45 | 0 | 47.7 | muy_alto | 0 |
| 2026-09-22 | 0 | 0 | 32.6 | 20.3 | 35 | 16.5 | 34.2 | 0 | 39 | muy_alto | 0 |
| 2026-09-23 | 0 | 0 | 31.4 | 22.4 | 37 | 21.6 | 40.3 | 0 | 42.1 | muy_alto | 0 |
| 2026-09-24 | 0 | 0 | 32.3 | 22 | 36 | 18.3 | 38.2 | 0.1 | 39.1 | muy_alto | 0 |
| 2026-09-25 | 0 | 0 | 32.6 | 22.7 | 32 | 15.1 | 31.3 | 0.1 | 36.4 | alto | 0 |
| 2026-09-26 | 0 | 0 | 32.8 | 22 | 30 | 16.5 | 34.6 | 0 | 37.8 | alto | 0 |
It is generated in your browser from this page’s data: one file per fire, aggregated by day and without individual hotspots.
Live widget for your story
It updates itself while the fire is still active and links to this page. Free, with the attribution shown at the bottom.
Sources for this page: GWIS (Copernicus), NASA FIRMS (VIIRS), ERA5 (ECMWF/Copernicus C3S) · Open-Meteo.com (CC BY 4.0), GHSL (JRC). Open-Meteo.com · How we build fire pages: methodology
- Burnt area
- 5,225ha
- GWIS perimeter (Copernicus)
- Duration
- 13days
- 4 Sept 2026 – 16 Sept 2026
- Municipalities hit
- 2
- districts
- Started on
- 4 Sept
- 2026
- Heat detections
- 163
- in a single hotspot: 95 MW
- Danger at start
- 39
- Very high · FWI index
- Average daily power
- 184MW
- detected power per day, on average
Where it burned
HOTSPOTS BY DAY
4 Sept16 Sept
16 Sept 2026
How it evolved, day by day
4 Sept
7 Sept
10 Sept
13 Sept
16 Sept
Fire radiative power detected each day, in megawatts, and number of hotspots. The magenta bar marks the peak day.
The weather it burned in
33°high
62km/h · peak gust
28%lowest humidity
It had gone 19 days without rain (1 mm or more) when it started.
This area had burned before
33% of the perimeter overlaps earlier fires recorded since 2016: 5 ha in 2016, 11 ha in 2017, 232 ha in 2018, 1,014 ha in 2019, 352 ha in 2021, 825 ha in 2024. Part of it burned more than once, so the yearly figures add up to more than the 1,711 ha that had burned before.
How big is it?
London boundary: geoBoundaries · Office for National Statistics (Open Government Licence v3.0).
Paris boundary: geoBoundaries · Institut national de l'information géographique et forestière (IGN-F) (Etalab Open License 2.0).
Madrid boundary: geoBoundaries · Open Data ESRI, IGN Central Registry of Cartography Area (CC BY 4.0).
Berlin boundary: geoBoundaries · Federal Agency for Cartography and Geodesy (dl-de/by-2-0).
This compares size, not location: both shapes are at the same scale, north up, centred on each other.
What happened to each town
-
Antsohihy where it started4,211 ha reached on 4 Sept0.9 %
-
1,049 ha reached on 6 Sept0.1 %
Burnt area inside each municipality, from crossing the GWIS perimeter with administrative boundaries. The distance runs from the point marking the centre of each area to the nearest fire: staying a kilometre away does not mean it reached the houses.
People and buildings inside the perimeter
GHSL 2025Built-up area2.4 ha
Residents≈ 1,000
Inside the perimeter there were about 2.4 hectares of built-up land and about 1,000 residents (GHSL 2025, satellite-based estimate).
This is an estimate on 100 m cells, made from satellite imagery and census data: it does not mean that homes burned or that anyone was harmed. The perimeter also contains land that never burned. The GHSL 2025 epoch is a projection made from 2020 and earlier data.
Source: Global Human Settlement Layer (GHS-BUILT-S and GHS-POP R2023A), European Commission Joint Research Centre (JRC), © European Union, CC BY 4.0. Sum inside the perimeter computed by WorldFires. human-settlement.emergency.copernicus.eu
The weather day by day
ECMWF ERA5 reanalysis at the fire location, the standard reference for past weather, and the Canadian Fire Weather Index (the same system Copernicus uses), computed by WorldFires from that data. Days after the fire run until the first rain.
Open-Meteo.com (CC BY 4.0) · Copernicus Climate Change Service (ERA5)
| Day | High | Min. humidity | Peak gust | Rain | Danger (FWI) |
|---|---|---|---|---|---|
| 4 Sept | 31.4 °C | 32 % | 41 km/h | 0.3 mm | Very high 39 |
| 5 Sept | 31.8 °C | 32 % | 39 km/h | 0.0 mm | Very high 40 |
| 6 Sept | 30.6 °C | 31 % | 40 km/h | 0.0 mm | Very high 40 |
| 7 Sept | 31.0 °C | 36 % | 28 km/h | 0.7 mm | High 30 |
| 8 Sept | 32.3 °C | 31 % | 56 km/h | 0.0 mm | Very high 44 |
| 9 Sept | 32.8 °C | 31 % | 59 km/h | 0.0 mm | Very high 49 |
| 10 Sept | 32.3 °C | 31 % | 57 km/h | 0.0 mm | Very high 49 |
| 11 Sept | 32.1 °C | 30 % | 46 km/h | 0.0 mm | Very high 44 |
| 12 Sept | 31.6 °C | 31 % | 31 km/h | 0.0 mm | High 36 |
| 13 Sept | 31.4 °C | 30 % | 23 km/h | 0.0 mm | High 34 |
| 14 Sept | 32.7 °C | 28 % | 53 km/h | 0.0 mm | Extreme 51 |
| 15 Sept | 32.3 °C | 28 % | 59 km/h | 0.0 mm | Extreme 52 |
| 16 Sept | 31.2 °C | 31 % | 62 km/h | 0.0 mm | Extreme 56 |
| 17 Sept after | 31.5 °C | 30 % | 53 km/h | 0.0 mm | Very high 50 |
| 18 Sept after | 32.4 °C | 31 % | 48 km/h | 0.0 mm | Very high 47 |
| 19 Sept after | 32.6 °C | 27 % | 38 km/h | 0.0 mm | Very high 42 |
| 20 Sept after | 32.2 °C | 34 % | 49 km/h | 0.0 mm | Very high 48 |
| 21 Sept after | 31.7 °C | 31 % | 45 km/h | 0.0 mm | Very high 48 |
| 22 Sept after | 32.6 °C | 35 % | 34 km/h | 0.0 mm | Very high 39 |
| 23 Sept after | 31.4 °C | 37 % | 40 km/h | 0.0 mm | Very high 42 |
| 24 Sept after | 32.3 °C | 36 % | 38 km/h | 0.1 mm | Very high 39 |
| 25 Sept after | 32.6 °C | 32 % | 31 km/h | 0.1 mm | High 36 |
| 26 Sept after | 32.8 °C | 30 % | 35 km/h | 0.0 mm | High 38 |
Before and after, from space
Drag the line to compare the same ground before the fire and after it.
↔
Before · 28 Aug 2026
After · 23 Sept 2026
Sentinel-2 did not pass over on that day. Try another date, or use false colour, which is daily.
If the day is cloudy, change the date: each satellite passes once a day. In false colour the burnt ground looks brown or red.
The smoke
CAMS · CopernicusOn 18 September 2026 the air beside the fire reached 16 µg/m³ of PM2.5, but no daily average passed the WHO guideline (15).
PM2.5 peak16 µg/m³
Days over the guideline0 days
PM1019 µg/m³
Optical depth0.32
WHO 15
2 Sept6 Sept10 Sept14 Sept18 Sept
Good
Fair
Moderate
Poor
Very poor
Extremely poor
Particles from the Copernicus CAMS model at the centre of the perimeter; these are not measurements in each town. The pale bar is the worst hour, the solid one the daily average.
Open-Meteo.com (CC BY 4.0) · Copernicus Atmosphere Monitoring Service (CAMS)
Detections table · 163Every satellite heat detection, one by one, with time, satellite, power and coordinates.
⬇ Download CSV
Times in UTC. Each detection is a 375 m pixel: a large fire yields hundreds per pass.
| Date and time | Satellite | FRP (MW) | Lat | Lon |
|---|
Other fires nearby
Frequently asked questions
- How long did the Antsohihy fire last?
- According to the heat detections seen by satellite inside the perimeter, it burned from 4 September 2026 to 16 September 2026 (13 days). The perimeter dates it from 4 September 2026 to 11 September 2026.
- How much area did it burn?
- The perimeter GWIS (Copernicus) rebuilt from heat detections is 5,225 hectares (52.2 km²). It is an estimate: official national figures may differ.
- Which municipalities burned?
- The fire entered 2 municipalities: Antsohihy (4,211 ha, 0.9 %); Analalava (1,049 ha, 0.1 %).
- What was the weather like?
- During the fire the temperature reached 32.8 °C, humidity dropped to 28 % and gusts hit 62 km/h. The fire danger index (FWI) reached 56, class Extreme.
- Where does this data come from?
- From the GWIS (Copernicus) burnt perimeter, NASA FIRMS VIIRS hotspots and the administrative boundaries of Madagascar (geoBoundaries). It is not an official report.
Where to verify it
A satellite detection is not an official report. Final figures and the fire’s status are published by each country’s emergency service.
Satellite cross-check
History of this page Changes in area, dates, name, official record or merged pages, compared every night since 24 October 2026.
No changes since 24 October 2026.
RSS Changes to this page With cross-checks and a citable version
Does something look wrong on this page?
This page is built from satellite data and official records, and it sometimes gets things wrong: the wrong dates, the wrong town, two fires merged into one, or something that was not a wildfire at all. If you know this fire first-hand (you live nearby, saw it, fought it or reported on it), tell us what really happened and we will check it.
Report an error
How we know this
- Perimeter: GWIS (Copernicus), rebuilt from heat detections rather than mapped on imagery.
- Dates adjusted with the VIIRS detections inside the perimeter, which said 4 September 2026 to 11 September 2026. 163 Suomi-NPP VIIRS detections on those dates.
- We have not matched this page with any official record.
- Weather: Copernicus ERA5 reanalysis at the fire location, in cells of about 25 km.
- Built-up area and population come from the JRC Global Human Settlement Layer (epoch 2025, 100 m cells), summed inside the perimeter: an estimate of what was there, not a damage count.
Where this data comes from
The perimeter comes from the Copernicus Global Wildfire Information System (GWIS), which rebuilds it from MODIS and VIIRS heat detections and publishes it under CC BY 4.0. It is not measured on imagery as in Europe, so it is an estimate. Heat detections come from the VIIRS sensors aboard Suomi-NPP and NOAA-20, distributed by NASA FIRMS at 375 m resolution. Municipal boundaries come from geoBoundaries ADM2 (Creative Commons Attribution 3.0 Intergovernmental Organisations (CC BY 3.0 IGO)). The area per municipality is computed by WorldFires by crossing the perimeter with those boundaries, so it may differ from each country’s official figure.