Skip to content
OneKitly
Sign up · Sign in

Active Fire Map: Satellite Detections, 24 h

Satellite fire detections around your place over the last 24 hours, with radiative power, the time of the pass and the fire danger index.

Related tools

All Weather & environment tools →
Weather warningsOfficial weather warnings for your place, as the national services publish them: Europe, United States and Canada, with the level, the area and the hours.Weather forecast9-day weather forecast for 170,000 towns or your position: temperature, feels-like, rain, wind, clear-sky UV index, sunrise and sunset.River levelsOfficial river levels and flood classes near you: France, Germany and the United States, with the reading, its time and the scale that applies.Earthquakes todayToday's earthquakes worldwide, merged from USGS, EMSC and INGV without duplicates: magnitude, depth, time, map and the quakes near a place.Air qualityThe official air quality index where you are, on the scale of that country — European 1 to 6, US AQI, or Canadian AQHI — with the pollutant that drives it.Sea temperatureThe sea temperature at your place, read by satellite on a 5 km grid, with the date of the measurement and how far the sampled point sits.

Two polar-orbiting satellites sweep your region several times a day and flag every pixel that is hotter than what surrounds it. Pick your place and this page lists those flags within a hundred kilometres, each with the megawatts it radiated, the minute it was seen and how sure the sensor was — then draws the whole region on a map and, in a third tab, gives the fire-weather index Copernicus publishes for that exact point. What it refuses to do is call those pixels fires. A burning stubble field, a gas flare and a forest front look identical from six hundred kilometres up, and the page says which is which only when it honestly can: never. It also refuses to treat an empty list as good news, because a satellite that has not passed overhead yet produces exactly the same emptiness as a landscape where nothing burns.

How to use it

  1. Choose your place — a town from the list, or the button that uses your position. Coordinates are rounded to about a kilometre before anything leaves your browser.
  2. Read the first tab as a list of hot pixels sorted by distance. The megawatt figure is the one that matters: it measures what is radiating now, while the confidence only says how likely the pixel is to be a real anomaly.
  3. Open the map to see the whole region at once. Circle size follows radiative power on a logarithmic scale, so a two-megawatt stubble fire stays visible next to a four-hundred-megawatt front; the cross marks the place you chose.
  4. Switch to the danger tab for the index. It forecasts rather than observes: it says how readily a fire would spread at that point, and it stays high for days after the last flame has gone out.
  5. Check the reading time under the list before you trust the silence. It is the moment of the most recent detection, not the moment the file was published — on 16 September 2026 those two were six hours apart.

How a hot pixel becomes a line on this page

The instrument is VIIRS, aboard NOAA-21 and NOAA-20. It measures infrared radiance in cells 375 metres across and compares each cell with its neighbours; when one stands out enough, it is written to a file with its coordinates, the minute of the overpass, a confidence category and — the useful number — its fire radiative power in megawatts. Our server downloads the regional file for your part of the world, not the global one: that global file weighed 8.3 MB on 16 September 2026, against 126 KB for Europe.

Two satellites see the same ground, so the same fire appears twice. Merging their files without care inflates the count — measured at plus 30.2 % on the files of 16 September 2026. Detections are therefore collapsed onto a grid of hundredths of a degree, about 1.1 km, and the cell keeps the one with the highest radiative power, because that is the one that best describes what is burning. This is not fire grouping: no source publishes a persistent fire identifier, so nothing here pretends to rebuild one.

When this page answers, and when it cannot

  • You can smell smoke and want to know whether something large is burning within an hour's drive, rather than refreshing a regional news feed.
  • You are travelling in summer and want the fire-weather index at your destination before deciding whether to camp, light a barbecue or take a forest road.
  • You follow a fire season from a distance — a region you come from, a place where someone you know lives — and want the same satellite view for anywhere on Earth rather than one country's map.
  • What it cannot do is warn you. There is no alert, no notification and no watching in your absence: the page answers when you open it, and an evacuation order will always reach you through your civil-protection service first.

Three readings taken on the same day

  • Above Vila de Rei in central Portugal, 16 September 2026: the strongest European detection of the morning radiated 221.05 MW, and the danger index at that same point was 15.07 — moderate. A moderate index and a violent fire coexist without contradiction, because one forecasts and the other observes.
  • Southern Africa on that same day: 54 024 detections in twenty-four hours, against 1 525 for the whole of Europe. Most are agricultural burning at the end of the dry season, which is why the map draws only the two thousand most powerful and says so under the drawing rather than passing a trimmed count off as the total.
  • Above the Rumaila field near Basra, the same twenty-four hours produced 179 detections, sixteen of them at high confidence. Nothing there was burning in the sense anyone means: they are gas flares, alight all year. That square of the map is removed by name, and the list of removals is short and deliberately incomplete — others remain over oil fields elsewhere.

What this page is not, stated plainly

  • It is not an alerting service and never will be from this page. No notification is sent, nothing is watched while the tab is closed, and an evacuation decision belongs to your civil-protection authority, never to a satellite file read by a website.
  • A fire smaller than roughly a hundred square metres, one hidden under a canopy, or one covered by its own smoke can pass unseen. So can any fire that starts and dies between two overpasses — the satellites are not a camera pointed at your valley, they are two orbits crossing it a few times a day.
  • Night passes detect more, because the sensor separates a hot pixel from a sun-warmed one far better in darkness. A night count and a day count are therefore not comparable, and a jump between the two says as much about the hour as about the ground.
  • The danger index carries no date. The service publishes one value per point and returns exactly the same bytes whether a date is asked for or not, so this page shows the hour we read it and claims no forecast day. Its classes are the service's own; the power bands in the first tab are ours, and the page says which is which.
  • Two of the satellites that fed this kind of page are ending. NOAA stops delivering Suomi NPP products on 1 November 2026, Terra's mission ends in February 2027 and Aqua's in September 2027. This page reads NOAA-21 and NOAA-20 only, so that no sensor name is frozen into it.

Privacy

Processed on a server

The place you pick is rounded to about a kilometre in your browser, then sent to our server, which downloads the NASA FIRMS regional detection file and queries GWIS, the Copernicus Global Wildfire Information System, for the danger index at that point. Neither NASA nor Copernicus sees your address — they see our server — and nothing you type reaches them. The last town you chose stays in your browser's local storage under the allin_tools key; a position you share is used for the request and never stored. No lookup is logged or written to any table.

Frequently asked questions

Does a detection mean a fire is burning there?

It means a 375-metre cell was measurably hotter than its neighbours when a satellite looked. That covers wildfires, but equally stubble burning, charcoal production, a steel works, a volcano and a gas flare. The instrument cannot tell them apart and neither can this page, so it writes detection everywhere the data says detection.

The list is empty — does that mean I am safe?

No, and the page says so rather than letting the blank speak. An empty list can mean nothing was detected, or that no satellite has passed over your area recently enough, or that a fire is too small or too smoke-covered to register. Emptiness is the absence of a measurement, which is not the same thing as the measurement of an absence.

Why can the danger index be low while a big fire is burning nearby?

Because they answer different questions. The index is built from weather and dryness and estimates how readily a new fire would spread at that point; it knows nothing about fires already alight. A fire that started on a windy day and is still burning in calm, damp air will sit under a modest index without either figure being wrong.

What does the megawatt figure actually measure?

Fire radiative power: the energy the burning surface is radiating at the instant of the overpass, which tracks how much biomass is being consumed. It is the one figure on this page that grades intensity. Confidence does not — a high-confidence detection may be a modest flare, and a nominal-confidence one a serious front. The four power bands shown are ours; the megawatts are NASA's.

Why is the European EFFIS layer not used here?

Because it returns nothing. Its public active-fire layer was queried on four regions and eleven dates in September 2026 and gave an empty tile for every date after 1 October 2021; ten further queries aimed at the exact coordinates of five real European fires all answered that the search returned no results. The Copernicus GWIS service, asked for the same area on the same day, answers normally — so that is the one this page reads.

How fresh is what I am looking at?

The files cover the last twenty-four hours and are republished through the day; our server keeps one for twenty minutes before fetching it again. The time shown is that of the most recent detection, not of the file — on 16 September 2026 the European file was published at 10:41 while its freshest detection dated from 03:55, and six hours of that gap were not latency but the simple absence of an overpass.

Further reading

All guides →
ExplainerDew Point: What It Is, and How the Calculator Gets ItMagnus–Tetens with the Alduchov–Eskridge coefficients, 17.625 and 243.04. Why an absolute number beats relative humidity for judging comfort, what temperature a window has to reach before it runs, and the one input the calculator does not check.ExplainerWind Chill: What the “Feels Like” Number Actually MeasuresThe calculator runs the 2001 index, not the 1945 one, and refuses to answer above 50 °F or under 3 mph. Here is what its number is a model of — bare skin, in shade, walking — and the two things it says nothing about: a clothed body, and a parked car.GuideYeast Conversion Chart: Fresh, Active Dry and Instant, With the Actual FactorsMultiply fresh yeast by 0.4 for active dry and by 0.33 for instant. Those two numbers are the whole chart, they are King Arthur's professional figures, and they are approximate — here is how far, and what else changes when you swap.ExplainerFive Points of Learning Rate Turn Fifty Repetitions Into Four HundredAt an 85 % learning rate, a task that takes 20 minutes first time reaches 8 minutes after 50 attempts. At 90 % it needs 415 — the same task, five points apart.GuideSchool Holiday Zones: Which One Applies to YouIn France only two of the six holiday periods differ between zones — and those two are two weeks apart from end to end. In Germany the summer break start spans thirty-five days between Länder. Neither can be computed: they are decreed, and a stale copy is the real hazard.ExplainerSow Forty-One Seeds to Get Thirty Plants82 seedlings from 100 seeds is an 82 % germination rate. Wanting 30 plants with a 10 % margin means sowing 41 — and the packet's percentage is not the one your windowsill will deliver.