Twenty-five years of satellite fire detections from MODIS and VIIRS, merged into one consistent record. Pick any area on Earth and see when it burns, which years were worst, and whether this month is out of the ordinary.
Every glow is a real detection: NASA FIRMS, 27 Oct to 3 Nov 2026. Drag the globe.
Each glow is a 0.1° cell, the grid our record lives on, where VIIRS detected burning: savanna fires across Africa, farm burning in Asia and South America, forest clearing. Real data, 27 Oct to 3 Nov 2026.
MODIS flags a pixel when a fire heats enough of a square kilometre. Small fires slip through.
VIIRS sees fires MODIS never could. Same planet, same week, far more detections.
The world didn't start burning more when VIIRS launched. We calibrate MODIS against VIIRS over the years both flew, so 2005 and 2025 speak the same units.
After the harvest, farmers burn the straw. Small, short fires: this week VIIRS found 88 in the Delta, MODIS 7. A record that mixes the two can't tell you if this season is heavy.
From 700 km up to the field.
Sketch a polygon, drag a box, pick a province, search a place, or type a latitude and longitude.
Every month since 2001 in one grid. Bright cells are heavy burning; the pattern is your fire season at a glance. Above: the Nile Delta.
See how a month ranks against history (percentile, z-score), then download CSV, GeoJSON or GeoTIFF for your GIS.
We calibrate MODIS against VIIRS over the years both flew, then check the seam at 2012 is gone. Raw join: 2.9× more hotspots after 2012. Our record: 1.03×.
Winter fires in northern Patagonia made the news. The calendar puts that month against every July since 2001, on one record that doesn't change meaning when the satellites do.
Open Río Negro, July 2026Every October and November, farmers burn rice straw across the Delta. This week VIIRS found 88 hotspots there and MODIS 7. Is this year's season heavier than usual?
Open the Nile Delta, October 2026
Generated portraitIs today’s activity routine for this season, or a surge?
Generated portraitOne sensor-agnostic series from 2001 to now, ready for trend work, with its validation published.
Generated portraitKnow the critical months before they arrive, and how this season compares.
Not always. A hotspot is a pixel that looked hot from orbit. One fire can make many hotspots, and a gas flare can make one too. We drop known industrial and volcanic sources.
VIIRS sees smaller and cooler fires, so it reports several times more hotspots for the same fires. Glued together, the record jumps in 2012 even though the fires didn't change. We convert MODIS to VIIRS-equivalent values first.
The record runs to last month from NASA's standard archive, with near-real-time data for the newest weeks. The current month is compared with the same days in past years.
Yes. Any area as CSV or GeoJSON (individual hotspots), the monthly calendar as CSV, and the 0.1° grid as a Cloud-Optimized GeoTIFF.
No. Clouds, thick smoke, small fires and burning between satellite passes can all hide a fire. Read zero as nothing detected.
Every number, the globe's fires and the maps are NASA data. The films of the Delta and Patagonia, and the portraits, are generated scenes (Higgsfield) and are marked as such.
Any polygon on Earth, every month since 2001, on one consistent record.