Data & Science

Wildfire Tracking Explained: How Satellites, Agencies and Maps Follow a Fire From Ignition to Containment

📅 September 2026 ⏱ 50 min read ✍ US Wildfire Tracker team

Wildfire tracking is the chain that turns a column of smoke into a record on a map: someone or something detects heat, a dispatch center opens an incident, agencies publish the record and the perimeter, and satellites keep measuring the fire until it is declared out. This guide walks that chain end to end.

What does wildfire tracking actually mean?

Wildfire tracking is the process of detecting a fire, opening a numbered incident record for it, mapping its edge, and updating both the record and the map until agencies declare the fire out. Four separate systems do that work, and none of them talks to the public directly.

The first system is detection. A fire becomes known when a person calls 911, a lookout takes a compass bearing, a camera flags smoke, an air attack aircraft radios a location, or a satellite sensor measures a thermal anomaly. Detection produces a point, not a story, and wildfire tracking begins there.

The second system is the incident record. A dispatch center creates a record with a name, a discovery time, a point of origin and a jurisdiction. That record is the spine everything else hangs on, and it carries a unique identifier for the rest of the fire's life.

The third system is mapping. Firefighters walk or fly GPS lines, infrared aircraft scan the fire at night, and analysts turn those inputs into a polygon. The fourth system is measurement: satellites keep reporting heat whether or not anyone has updated the paperwork.

Why one fire produces four different truths

A single fire can be described four ways at the same moment, and all four can be correct. The incident record says 3,400 acres because that was the last approved figure. The perimeter polygon says 3,910 acres because it was flown six hours later.

The satellite feed says heat is still detectable a mile outside that polygon, because a spot fire crossed a ridge after the flight. The weather feed says a Red Flag Warning starts at noon. Good wildfire tracking shows all four and timestamps each one instead of averaging them into a single number.

That is the difference between a map and a report. A report picks one number. A tracking map keeps the disagreement visible so you can judge which source is fresh enough to act on, which is the same discipline described in our complete fire tracker guide.

Who does the tracking, and who publishes it

Federal, state, tribal and local agencies do the tracking. The National Interagency Fire Center coordinates the interagency data; NASA and NOAA operate the satellites; the National Weather Service issues fire-weather products. Public maps are consumers of those feeds, not sources of them.

That distinction matters when a number looks wrong. If the acreage on a map is stale, the fix is not the map; it is the agency record the map is reading. US Wildfire Tracker shows the "Last NIFC update" timestamp for exactly that reason.

How does wildfire detection start with a 911 call?

Most US wildfires are still discovered by people. NIFC's own data field for the earliest size estimate defines it as the acreage "when the fire is first reported by the first person that calls in the fire," which tells you how central the public phone call remains to wildfire detection in 2026.

The numbers support that. In 2025, NIFC recorded 69,556 human-caused wildfires and 8,294 lightning-caused wildfires. Humans therefore started about 89 percent of all US wildfires that year, and human-caused fires start disproportionately near roads, homes and workplaces where somebody is present to see them.

Acreage tells the opposite story. Those 69,556 human-caused fires burned 2,274,333 acres in 2025, while the 8,294 lightning fires burned 2,857,141 acres. Lightning caused roughly one fire in ten but about 56 percent of the acres, because lightning starts fires in places nobody is watching.

Share figures above are our own arithmetic on the NIFC cause tables published at nifc.gov; the fire and acre counts are NIFC's.

What happens in the first ten minutes

A 911 dispatcher takes the report and passes it to the wildland dispatch center that covers that jurisdiction. The center assigns resources, and an incident commander gives a size-up on arrival. NIFC's data standard records that separately as the "initial response acres," measured when the incident commander performs the first size-up.

Only after those steps does a machine-readable record exist for wildfire tracking systems to read. That gap — minutes to an hour, sometimes longer in remote country — is why a satellite hotspot can appear on a public map before any official record exists for the same fire.

1. A caller reports smoke; the 911 center routes the call to the wildland dispatch center with jurisdiction.
2. Dispatch opens an incident, assigns a local identifier, and orders initial attack resources.
3. The incident commander arrives, performs a size-up, and reports acres, fuels and threats by radio.
4. Dispatch enters the fire discovery date and time, point of origin, and cause category into the computer-aided dispatch system.
5. That system pushes the record into the interagency integration service, where public data services can read it.

Why the caller's acreage is almost always wrong

Discovery acres are an estimate made under pressure, often from a moving vehicle or a distant ridge. NIFC's standard says the estimate should include unburned and unburnable islands inside the perimeter, which most callers do not attempt. The minimum recordable size is 0.1 acres.

Treat the first acreage on any fire as an order of magnitude, not a measurement. The number usually changes twice in the first day: once at size-up, once after the first mapping product. Tracking fires well means expecting those revisions rather than reading them as errors.

Do fire lookouts still matter for tracking fires?

Staffed fire lookouts remain part of US wildfire detection in 2026, and the Forest Service still describes them as playing an essential role alongside camera stations and satellites. A lookout is slow to scale but still very hard to beat at one wildfire tracking job: judging whether a smoke column is a wildfire.

The tool that made lookouts precise is the Osborne Firefinder, a map mounted on a rotating steel disc with brass sighting mechanisms that lets an observer read an azimuth and a vertical angle to a smoke, in the manner of an engineer's transit.

Range depends entirely on air clarity. A Forest Service technology bulletin on the instrument reports that under the most favorable visibility a lookout observer could spot a fire 15 miles away, under average summer conditions 8 miles, and under smoked-in emergency conditions only 5 miles.

How two lookouts turn a bearing into a point

One azimuth gives a line, not a location. Two lookouts reporting azimuths to the same smoke give an intersection, and the dispatch center takes that crossing point as the fire's coordinates. This is triangulation, and it predates satellite fire detection by roughly seventy years.

The method still has a practical advantage over machine detection. A lookout can report smoke color, drift direction and column behavior in one radio call, which tells an incident commander something about fuels and fire behavior that a heat pixel cannot.

Where lookouts fail

Lookouts see smoke, so they cannot see through smoke. During a busy lightning bust or a regional smoke event, the same conditions that produce fires also blind the people watching for them. That is precisely the situation where satellite wildfire tracking keeps working.

Lookouts also cover a fixed viewshed. The 5,131,474 acres that burned across the US in 2025 were spread over terrain no fixed network of towers could cover, which is why detection has become layered rather than replaced.

How do camera networks speed up wildfire detection?

Camera networks are the modern industrial version of a lookout: fixed pan-tilt-zoom cameras on ridges and towers, streaming continuously, with software watching every frame for the visual signature of a new smoke column. In wildfire tracking terms they shorten the gap between ignition and the first confirmed report.

The largest US example is the network operated by the University of California, San Diego with CAL FIRE. The university reports a network of more than 1,050 cameras monitoring California's highest fire-threat regions around the clock, generating on the order of two billion data points per second.

The detection software was rolled out to 21 CAL FIRE units statewide in September 2023. The same program has mapped roughly 34,000 square miles of terrain with LiDAR and hyperspectral imaging, which supports fuel and terrain analysis rather than detection alone.

Why cameras and satellites answer different questions

A camera answers "is that smoke?" in seconds, at a resolution no orbital sensor matches, but only inside its own viewshed. A satellite answers "where is heat?" everywhere, but at 375 meters or coarser and only when it is overhead. Layered wildfire tracking uses both, and neither substitutes for the other.

Camera confirmation also matters for dispatch confidence. A dispatcher who can pull up an image before ordering aircraft is making a different decision than one working from a single unconfirmed phone call, and confirmation time directly affects how large the fire is at initial attack.

What do aircraft add to early wildfire detection?

Aircraft occupy the middle rung of the wildfire tracking ladder: slower to arrive than a satellite pass but far more precise, and able to stay over a fire while it changes. On a new fire, the air attack platform often supplies the first reliable geometry anyone has.

Detection flights matter most in remote country after lightning. A single storm can leave dozens of holdover fires smoldering for days before they surface, and a patrol aircraft flying a grid over the storm track finds them faster than waiting for each one to grow enough for a satellite pixel.

Once a fire is established, aircraft shift from detection to measurement. The Forest Service's national infrared program flies fixed-wing aircraft with thermal line scanners at night, and the resulting imagery is what most large-fire perimeters are actually built from.

Why the air picture rarely reaches the public directly

Air attack traffic is operational, not published. What reaches a public map is the product built from it: a perimeter polygon whose mapping method field records how the shape was derived. On the current national perimeter service, that field is where aircraft work becomes visible to everyone else.

Unmanned aircraft have started to fill gaps that crewed platforms cannot, particularly for night infrared work over smaller incidents. The data path is the same: imagery goes to a fire GIS specialist, who produces the polygon that eventually appears on a live fire tracker map.

How does satellite fire detection actually work?

Satellite fire detection does not look for flames. It looks for pixels that are anomalously hot in the middle infrared compared with their neighbors. An algorithm compares each candidate pixel against a background window of surrounding pixels and flags the ones that stand out.

The physics is straightforward. A burning flame front radiates strongly near 4 micrometers, a wavelength where ordinary sunlit ground is comparatively dim. Sensors therefore carry a dedicated fire channel tuned to that region, plus longer-wave channels used to characterize the background and reject false alarms.

Two outputs of satellite wildfire tracking matter to anyone reading a public map. The first is a confidence value, expressed as a percentage from 0 to 100 in the standard products. The second is fire radiative power, reported in megawatts, where higher values indicate higher fire intensity or a larger burning area.

What confidence actually means

Confidence is not a probability that a fire exists in the ordinary sense. It is the algorithm's own summary of how cleanly the pixel passed its tests: how far above background it sat, how the neighboring pixels behaved, and whether known contaminants such as sun glint or cloud edges were present.

Standard products group those values into low, nominal and high classes. A low-confidence detection is not automatically wrong; it is often a small or partly obscured fire. A high-confidence detection is not automatically a wildfire either, because a steel mill also passes the test cleanly.

Why fire radiative power is the more useful number

Fire radiative power correlates with fire intensity and size, so a rising sequence of values over successive overpasses is a signal that a fire is growing. Analysts use it to rank simultaneous incidents when resources are short, and it feeds smoke-emission models that drive air-quality forecasts.

For a member of the public, the practical use is comparative. Two detections in the same county with wildly different radiative power are not equivalent events, even though both appear as a dot. That difference is one reason our live fire tracker guide spends so long on timestamps and intensity rather than on dot counts.

What do MODIS and VIIRS add to fire tracking?

MODIS and VIIRS are the two polar-orbiting instrument families that produce most US active fire detections. They fly in sun-synchronous orbits, cross the same latitude at roughly the same local time each day, and see any given place only during an overpass.

MODIS flies on NASA's Terra and Aqua satellites at 705 km altitude with a 2,330 km cross-track swath and 36 spectral bands. Terra crosses the equator at 10:30 a.m. on a descending node; Aqua crosses at 1:30 p.m. ascending. Its fire product resolution is 1,000 meters.

VIIRS flies on Suomi NPP, NOAA-20 and NOAA-21. Its active fire product resolves at 375 meters, which NASA describes as observations about three times more detailed than MODIS, and which lets it detect many small and low-temperature fires that MODIS misses.

Why the overpass schedule dictates what you see

Because each VIIRS satellite gives roughly two CONUS overpasses per day, a three-satellite VIIRS constellation plus two MODIS platforms produces a handful of looks per day, not continuous coverage. A fire that starts at 3 p.m. and blows up by 6 p.m. may have no polar detection until the evening pass.

Night passes are unusually valuable. Without solar reflection contaminating the shortwave infrared, the contrast between a fire and its background is much stronger, so nighttime detections tend to be cleaner and catch smaller fires than daytime ones at the same resolution.

How quickly detections reach the public

NASA has stated that active fire locations are available to users within minutes of the satellite overpass, and its Fire Information for Resource Management System publishes ultra real-time detections for much of the US and Canada in less than 60 seconds of the satellite flying over, using direct-readout ground stations.

That speed applies to the detection, not to the interpretation. The pixel arrives fast; deciding whether it is a wildfire, a prescribed burn or a gas flare takes context that no orbital sensor supplies on its own.

How do GOES satellites make wildfire tracking constant?

Geostationary satellites solve the overpass problem. NOAA's GOES East and GOES West sit about 22,236 miles above the equator and stare at the same hemisphere continuously, so their fire detection algorithm runs on a schedule rather than waiting for an orbit to come around.

The instrument is the Advanced Baseline Imager. It carries 16 spectral bands — two visible, four near-infrared and ten infrared — at 0.5 km for the red visible band, 1.0 km for other visible and near-infrared bands, and 2 km for the infrared bands used in fire work.

Scan cadence is the payoff. The ABI scans the contiguous US every 5 minutes, produces full-disk imagery every 5 to 15 minutes depending on mode, and can run a mesoscale sector over a specific incident every 30 or 60 seconds.

What geostationary coverage buys you

Constant staring means detection of ignition rather than detection of an established fire. NOAA's Next Generation Fire System uses this to flag new fires early; Mike Pavolonis, who developed the system and manages NOAA's wildland fire program, has said geostationary satellites "enable detection of fires within 5 to 30 minutes of ignition."

The Morrill Fire in Nebraska is the documented example. NOAA reports it was first detected at 1:40 p.m. Mountain Time on March 12, 2026, with an alert posted 13 minutes after the initial detection. The fire traveled more than 65 miles by 9:00 p.m. and eventually burned 642,029 acres before reaching 100 percent containment on March 24, 2026.

The trade-off you are accepting

Two kilometer pixels are large. A quarter-acre fire under a forest canopy will not register on GOES until it produces enough radiant energy to move the average temperature of a four square kilometer box. Geostationary sensors are early, not sensitive.

The current GOES East spacecraft, GOES-19, became operational on April 7, 2025 at 75.2 degrees west longitude after launching on June 25, 2024. Its predecessor moved to backup status, which is why archived detections from before that date carry a different satellite identifier.

How does an IRWIN record power fire tracking?

IRWIN — the Integrated Reporting of Wildland Fire Information service — is the interagency exchange that keeps one fire from becoming five different records. It was created as a central repository for basic fire information, to cut redundant data entry and improve consistency across agencies, with initial implementation in May 2014.

IRWIN does not replace agency systems. It connects them. Computer-aided dispatch systems, incident status reporting, decision support tools and geospatial services all read from and write to the same record, so an update entered once propagates instead of being retyped.

The service also detects conflicts. When two dispatch centers open records for what turns out to be the same ignition, IRWIN supports resolution so a single unique record survives. Records with unresolved conflicts are marked as quarantined and are excluded from public data services until the conflict clears.

The fields that end up on your screen

Most of what a public map displays per fire is an IRWIN field with a specific definition. Learning these wildfire tracking definitions removes most of the confusion people have when two sources disagree about the same fire.

Why "percent contained" is not "percent extinguished"

The official definition is narrow: percent contained indicates the percent of incident area that is no longer active. It is a statement about control line and fire activity around the edge, not about how much fuel inside the perimeter has stopped burning.

This is the single most misread number in all of wildfire tracking. A fire at 60 percent containment can still have thousands of acres of interior heat, and satellites will keep detecting that heat for days. The two facts are not contradictory.

Strategy percentages most people never see

IRWIN also records the share of the incident area under each suppression strategy: full suppression, confine, point or zone protection, and monitor. A fire that is 100 percent under a monitor strategy is being watched deliberately, not neglected, and its containment figure may stay at zero for weeks.

Those fields explain why a large fire in wilderness can sit at 0 percent containment all season without anything going wrong. The strategy, not the containment number, is what tells you how the fire is being managed.

How does WFIGS publish the data behind fire tracking?

WFIGS — Wildland Fire Interagency Geospatial Services — is the group that turns IRWIN records into public geospatial services. NIFC describes it as providing authoritative geospatial data products under the interagency Wildland Fire Data Program, hosted in the National Interagency Fire Center's ArcGIS Online organization.

Two services carry most of the public load: current incident locations (points) and current interagency perimeters (polygons). On September 10, 2026, we queried both directly and found 579 current incident locations and 197 current perimeters nationwide.

That ratio is the central fact of public wildfire tracking. Roughly one current incident in three had a published perimeter that day. Most fires never get a mapped polygon at all, because they are caught small and controlled before anyone flies or walks a line around them.

What qualifies for the current services

The inclusion rules are explicit, and they explain most disappearances people notice. NIFC's dataset description states that a fire appears in the current perimeter service only if it meets every one of the following conditions.

1. The incident is categorized in IRWIN as a wildfire or a prescribed fire.
2. It has not been declared contained, controlled, nor out.
3. Its fire report records have not been completed and certified.
4. It is valid and not quarantined in IRWIN because of potential conflicts with other records.
5. The source polygon is set to public feature access, approved feature status, and visible.

Fall-off rules: why fires vanish from the map

NIFC applies explicit rules so stale records are not retained. Those rules are enforced hourly, and they are age-based rather than status-based, which means a fire can drop off simply because nobody updated it.

Fires that started before December of the previous calendar year are excluded outright. Anything that falls off the current service moves to the year-to-date service or the full history service, so the geometry is not lost — only its "current" status changes.

Why the criteria were set this way

NIFC states that the inclusion and fall-off criteria were determined by a task group of the NWCG Geospatial Subcommittee, and that perimeters are pulled from multiple sources with rules in place to ensure the most current or most authoritative shape is used. That last clause is why a polygon can change shape without the acreage changing.

How do perimeter maps make tracking fires precise?

A perimeter is the only wildfire tracking layer that states where the fire actually is rather than where heat was measured. Each polygon carries a mapping method field describing how the shape was derived, and that field is the honest label on the data's precision.

We queried the national current perimeter service on September 10, 2026 and counted how the 197 active polygons were produced. The distribution shows how mixed the real workflow is: most large fires are assembled from several sources rather than one clean survey.

Reading the mapping method like a precision label

A GPS-walked line is accurate to a few meters but slow and only feasible on small fires or short line segments. An infrared interpretation covers a hundred thousand acres in one pass but represents the heat perimeter at the moment of the flight, several hours before you see it.

A hand sketch is a professional's estimate made in the field. An auto-generated shape is not a survey at all — it is a system-created placeholder, and on our sample day 15 of 197 current polygons were exactly that. Treat the acreage on those as provisional.

The timestamp that matters most

Every polygon carries a capture time representing when the shape was actually collected, which is distinct from when the record was created or last edited in the database. A perimeter uploaded at 9:00 a.m. may describe the fire as it looked at 2:00 a.m.

That six or seven hour gap is normal and expected on large fires. It is also why a satellite hotspot appearing outside the polygon is not automatically an error in either source: the polygon is older than the pixel.

What do infrared mapping flights add to fire tracking?

Infrared mapping flights are how wildfire tracking produces large US fire perimeters. The Forest Service's National Infrared Operations program flies fixed-wing aircraft equipped with thermal line scanners at night, when the thermal contrast between fire and cooled terrain is strongest.

The sensitivity is remarkable. NOAA reports that NIROPS aircraft sensors detect hotspots as small as 6 inches from 10,000 feet, and can map 300,000 acres per hour. That combination — fine detail at industrial scale — is what no satellite currently matches over a single incident.

The current fleet includes a Citation and a Super King Air B200, the latter in service since the mid-1980s. Output from the line scanner is a geo-referenced raster image, which is then handed to a specialist rather than published directly.

What an infrared interpreter actually does

Infrared interpreters locate hotspots, hand lines, dozer lines and rocks in the imagery, and identify terrain features such as roads, streams, ridges and buildings with high accuracy. They transfer the ortho-rectified line scanner data onto local terrain maps so the product lines up with everything else the incident is using.

Their deliverables are practical: GIS-ready files, GIS layers, geo-referenced PDF maps and KMZ files. NIFC states plainly that these maps guide incident personnel on where to concentrate resources, and also help inform local people and media about fire status.

Why the flight happens at night

Daytime solar heating warms rock faces, bare soil and south-facing slopes enough to confuse a thermal scanner. At night those surfaces cool while the fire does not, so the heat perimeter stands out cleanly. Night flying also puts the product in interpreters' hands in time for the morning briefing.

That schedule creates the daily rhythm most large-fire perimeters follow: fly after dark, interpret overnight, publish in the early morning, then hold that shape for roughly 24 hours while the fire keeps moving. Everything between flights is inference.

Contained, controlled, out: fire tracking status terms

These three words are the official wildfire tracking milestones of a fire's life, and they are not synonyms. The interagency definitions are precise, and every public fire tracking map inherits them from the incident record whether or not it explains them.

To contain a fire is to reach the point where its spread is stopped by and within specific features, constructed or natural, so that no further spread is expected under foreseeable conditions. For reporting purposes, containment has a date and time.

To control a fire is to construct fireline, or use natural features, to surround the fire and any spot fires from it, and reduce its burning potential to a point where it no longer threatens further spread or resource damage under foreseeable conditions. Control also carries a date and time.

Out is the final state. The incident record has a fire out date and time, and only after that does the fire stop being an active incident in the interagency system. On many large fires that date arrives after a season-ending precipitation event, not after suppression.

Why 100 percent contained is not the end

Containment describes the edge. Interior fuel can burn for weeks inside a fully contained perimeter, and mop-up — the systematic extinguishing of remaining hot spots inside the line — is what closes the gap between contained and out.

This is why a fire reported as 100 percent contained can still generate satellite detections. It is also why our guide to following current wildfires treats containment percentage and heat detections as two independent signals rather than one.

Initial attack, extended attack and what they signal

Initial attack is an aggressive suppression action consistent with firefighter and public safety and the values to be protected. Extended attack is the phase where initial attack resources can no longer achieve the objectives set for the incident and additional resources and overhead are committed, generally pushing the incident into another operational period.

When you see an incident acquire a management team and a daily update rhythm, it has moved into extended attack. That transition is usually visible in the data before it is visible in the news, because the incident management organization field changes first.

How do incident teams change wildfire tracking data?

When a fire outgrows local resources an incident management team takes over, and the wildfire tracking data changes character. NIFC describes an incident management team as mobilized to provide command and coordination during complex emergencies, with members drawn from federal, state, local, tribal and territorial agencies who volunteer for the duty on top of their regular jobs.

The interagency data standard records which organization holds the incident: a Type 1, Type 2 or Type 3 incident management team, a Unified Command, a Unified Command with a team, or a National Incident Management Organization. The field is null when no team is assigned, which is the case for the large majority of fires.

Complex Incident Management Teams are the national tier. NIFC states they can be deployed anywhere in the United States, and that while they most often manage large wildfires, they also respond to high-profile events, search and rescue operations, and natural disasters such as hurricanes and floods.

Why a team assignment improves your data

Teams bring a situation unit, a GIS specialist and a public information officer. That means a daily perimeter, a daily narrative, and an approved incident status summary. Before a team arrives, updates depend on whoever at the local unit has time.

Reporting thresholds formalize this. The national situation report defines large wildfires as those exceeding 100 acres in timber fuel models or 300 acres in grass fuel models, and also includes any wildfire with a nationally recognized incident management team assigned regardless of size.

Unified Command and why it matters to the record

Unified Command applies when more than one agency has incident jurisdiction or the incident crosses political jurisdictions. Under it, agencies work through their designated incident commanders at a single incident command post, establishing common objectives and issuing one incident action plan.

For wildfire tracking, that unification is what prevents two competing acreage figures. Without it, a fire crossing from federal to state land can generate two records that then have to be reconciled through the conflict-resolution process.

How often the national picture is published

The National Incident Management Situation Report is compiled by Predictive Services staff at the National Interagency Coordination Center, and its publication frequency scales with the National Preparedness Level: weekly on Fridays at Preparedness Level 1, Monday through Friday at Level 2, and daily at Level 3 and above.

That cadence is a useful proxy for national fire activity. If the national report is publishing daily, the system is stretched, and individual incident updates are more likely to lag because the same people produce both.

Why does wildfire tracking show false hotspots?

Satellite wildfire tracking measures heat, and heat has many sources. NOAA lists the recurring causes of commission errors in its fire products explicitly: ambiguity with fresh burn scars and sandy soils, sun glint over optically bright surfaces, and thermal anomalies linked to industrial activities such as steel mills and gas flaring.

Daytime adds its own contaminants. NOAA training material for VIIRS active fires notes that potential daytime false alarms include reflected solar radiation from solar panels, hot and bright surfaces, and cloud glint. Structure fires in urban environments also register as thermal anomalies.

The scale of the problem is smaller than most people assume. In its April 2026 notice modifying the Hazard Mapping System, NOAA stated that false alarms are typically under 1 percent of detections — but 1 percent of a national feed is still a lot of dots on a busy day.

Prescribed burns: real fire, wrong conclusion

Prescribed burning is the deliberate application of fire under predetermined conditions to meet management objectives, and it is genuinely burning vegetation. A satellite cannot distinguish it from a wildfire, because there is no physical difference to detect.

The volume is substantial. NIFC's ten-year average for national prescribed fire reporting is 13,857 prescribed fires treating 2,187,642 acres per year across the federal agencies and state and other cooperators that report. Much of that burning happens in the Southeast in late winter and early spring.

Fortunately, the record system does distinguish them. IRWIN categorizes incidents as wildfire or prescribed fire, and the national perimeter service includes both categories. A tracker that surfaces the incident type lets you separate them; one that only plots heat cannot.

The 2026 change to human quality control

For years, NOAA analysts manually reviewed automated detections in the Hazard Mapping System, removing false ones and adding hotspots the algorithms missed. Effective April 15, 2026 at 0000 UTC, NOAA retired that standard practice of manually verifying automated fire detections.

NOAA expects fire pixel counts to decrease by roughly 5 percent as a result, with low-intensity fires and fires obscured by cloud or smoke most affected locally, and false alarms that were previously filtered now passing through. It also warns that pyrocumulus cloud artifacts may occasionally produce erroneous fire growth data.

The practical consequence for anyone doing wild fire tracking in 2026 is that the raw feeds are slightly noisier and slightly less complete than they were in 2025. Cross-checking a detection against an official incident record matters more now, not less.

How a probability score handles the ambiguity

US Wildfire Tracker assigns each detection a fire probability tier that fuses three things: whether an official NIFC record is linked to it, how fresh the NIFC perimeter is, and how recent the satellite heat is from VIIRS, MODIS or GOES. Weak signals below 25 percent probability are hidden by default.

That default exists because of everything above. An unlinked, isolated, repeating heat source in an industrial county is far more likely to be a factory than a fire, and putting it on the map with the same weight as a 3,000-acre incident would be misleading.

How fresh is the data when you are tracking fires?

Every wildfire tracking layer has its own clock, and the differences are large enough to change what you should do. A single map showing all four feeds is showing you four different moments in time simultaneously.

Geostationary detections are the fastest. The GOES imager scans the contiguous US every 5 minutes, and NOAA's system has flagged ignitions and issued alerts within 13 minutes of first detection in documented cases. Polar-orbiting detections can reach users in under 60 seconds of overpass through direct readout, but only when a satellite is overhead.

Official records are the slowest and the most authoritative. A perimeter flown at 2:00 a.m. may publish at 8:00 a.m.; an acreage figure may be a day old if the incident is not on a daily reporting cycle.

The trap of a fast map on slow data

A map that refreshes every three minutes is not showing you a fire that is three minutes old. It is showing you the most recent version of records that themselves update on agency schedules. Fast polling improves nothing if the upstream record is 14 hours old.

The fix is per-layer timestamps. US Wildfire Tracker shows "Last NIFC update," "Last satellite heat" and "Last GOES heat" separately for each fire so you can see which clock you are reading. Our deep dive on real-time wildfire data works through how to use those three fields together.

How does a public wild fire tracking map fuse feeds?

A public wildfire tracking map's job is fusion: reconciling four government feeds that were never designed to be read together. The work is unglamorous — matching identifiers, aligning timestamps, deciding what to hide — and it is the entire difference between a data dump and a usable picture.

Wildfire tracking fusion starts with the join key. Every incident carries a unique IRWIN identifier, so incident attributes and perimeter polygons can be matched reliably. Satellite detections carry no such key, so they must be associated with incidents by distance and time instead.

That association is inherently probabilistic. A VIIRS pixel 400 meters outside a perimeter that was flown eight hours ago probably belongs to that fire. The same pixel 40 miles away, with no incident nearby, may be a new start, a prescribed burn or a factory.

What US Wildfire Tracker actually combines

The map at wildfire-map.org reads official NIFC incident records and WFIGS interagency perimeter services, NASA satellite thermal detections from MODIS and VIIRS, NOAA Hazard Mapping System GOES heat detections, and National Weather Service fire-weather alerts from the public api.weather.gov feed, on an OpenStreetMap basemap.

Each detection receives a fire probability tier built from the three signals described earlier. Weak signals below 25 percent are hidden by default and can be shown deliberately; outdated records can be toggled on; and the app states plainly that hotspots can reflect prescribed burns or industrial heat.

What a per-fire panel should show you

A complete fire panel answers three questions: what do agencies say, what do satellites say, and how old is each answer. US Wildfire Tracker's per-fire view lists name, acres, percent contained, discovery age, incident type, managing organization and total personnel when NIFC provides them.

It then adds the three freshness stamps — last NIFC update, last satellite heat, last GOES heat — plus the official perimeter outline, an optional historical perimeter, official links, an InciWeb search, driving directions away from the fire, and a shareable link to that specific incident.

Data sources fused

NIFC and WFIGS records and perimeters, NASA MODIS and VIIRS, NOAA GOES heat, NWS fire-weather alerts

Basemap

OpenStreetMap

Refresh

Automatic every few minutes while the tab or side panel is open; last saved results shown if a refresh fails

Coverage

National view plus all 50 states and the District of Columbia

Price

Free, no account required, no ads or third-party ad trackers in the extension

Where the extension fits

The same map is available as a Chrome extension listed as Fire Tracker in the Chrome Web Store, where it is rated 5 stars. It opens the map in the Chrome side panel or a full tab, which suits monitoring during a wind event without giving up a browser window. You can add it to Chrome for free or use the web map with no install at all.

Location handling is deliberately narrow. Browser geolocation is used only to find fires within a set distance in miles, and location is never stored, sold or shared. Our nationwide wildfire tracker guide covers how the state filters and travel workflows are meant to be used.

Where is wildfire tracking technology heading next?

Three changes are reshaping wildfire tracking and wild fire tracking tools in the second half of this decade: purpose-built fire satellites, automated detection replacing manual review, and government portals opening operational feeds to the public. All three are already in motion in 2026.

The first is dedicated constellations. The FireSat program, a collaboration involving Muon Space, Earth Fire Alliance and Google Research, flies a six-channel multispectral infrared payload designed to detect fires as small as 5 by 5 meters — roughly the size of a classroom rather than a city block.

A prototype launched on March 14, 2025, and three operational satellites reached orbit on July 7, 2026. Published mission planning describes about 50 satellites in the full constellation, a swath of roughly 1,500 km, an average ground sample distance of 80 meters, and a stated false positive rate below five percent.

Automated detection at national scale

NOAA's Next Generation Fire System is the second change. It applies automated analysis to geostationary imagery so that ignitions surface without a human reviewing every frame, and NOAA expects it to transition to operational status later in 2026.

The same logic drove the retirement of manual verification in the Hazard Mapping System in April 2026. The direction of travel is clear: algorithms produce the detections, and human expertise moves upstream into algorithm design and downstream into incident decisions.

Public access to operational feeds

The third change is distribution. NOAA launched a Wildland Fire Data Portal on January 26, 2026, expanding public access to satellite fire information that previously reached mainly operational users, hosted on its cloud framework and drawing on GOES East and GOES West.

Cost is the argument behind the investment. NOAA has cited an annual US wildfire cost on the order of $424 billion, which is the frame agencies use when justifying satellite programs whose value is measured in fires that never grow large.

What is not changing

None of this removes wildfire tracking's perimeter problem. Detection is getting faster and finer; the authoritative shape of a fire still depends on an aircraft flying at night and a specialist interpreting the imagery. Expect detection latency to keep falling and perimeter latency to stay roughly where it is.

What can wild fire tracking never tell you?

Wild fire tracking is a monitoring system and not a warning system, and the distinction is safety-critical. No public map issues evacuation orders. Those come from your county sheriff, emergency management office or state agency, and they are the only instruction that matters.

Detection also has physical limits. Satellites cannot see through thick cloud in the shortwave infrared. Small, cool, smoldering fires under a closed canopy may not register at all. A fire can exist for hours with no detection and no record, and the map will look empty because it is honest, not because the county is safe.

Timing is the second limit. A polar-orbiting satellite that is not overhead cannot detect anything, and a geostationary satellite with 2 km pixels needs a fire large or hot enough to move a four square kilometer average. The gap between those two conditions is where most new fires live for their first hour.

What the acreage does not describe

Acreage is a scalar. It says nothing about which side of the fire is active, where the head is, what the terrain does to the wind at 4:00 p.m., or whether a fire is spotting a mile ahead of its own perimeter. A 900-acre fire in steep terrain above a canyon community is a more serious situation than a 9,000-acre fire in wet meadow.

Percent contained has the same weakness. Both numbers describe past accomplishment; neither forecasts behavior. Fire-weather products from the National Weather Service — Red Flag Warnings and Fire Weather Watches — are the forward-looking layer, and they belong in the same view as the fires themselves.

Smoke is a separate problem

The fire nearest you is often not the source of the smoke you are breathing. Smoke travels hundreds of miles and arrives without any local fire at all, which is why smoke analysis is a separate NOAA product with its own schedule and its own density categories. Our smoke and air-quality guide covers that layer properly.

If your question is "should my kids play outside today," the answer comes from air-quality data and not from a fire perimeter. If your question is "do I need to leave," the answer comes from your county, informed by the fire data.

Wildfire tracking glossary: terms you will see

These are the wildfire tracking terms that appear on incident records, perimeter attributes and satellite products. Definitions follow the interagency wildland fire standards published through NIFC and the National Weather Service fire-weather glossary, condensed for readers who are not fire practitioners.

Incident and command terms

Incident commander

The individual responsible for direct management of all incident operations, and the leader of an incident command team.

Incident command system

The combination of facilities, equipment, personnel, procedures and communications operating within a common organization structure to manage resources against incident objectives.

Operational period

A defined block of time, for example 0600 to 1800, for which specific objectives are set and specific resources assigned; also called a shift.

Unified Command

A structure used when more than one agency has jurisdiction, allowing joint management through a common set of objectives and a single incident action plan.

Area command

An organization led by an area commander that manages multiple incidents, each with its own incident commander and team, setting overall strategies and priorities.

Overhead

Personnel assigned to supervisory or management functions, including command and general staff, directors, supervisors and unit leaders.

Agency administrator

The official responsible for administering policy on an area of land, with full authority to direct the incident management organization.

Suppression and tactics terms

Initial attack

An aggressive suppression action consistent with firefighter and public safety and the values to be protected.

Extended attack

The phase when initial attack resources can no longer achieve the objectives set, and additional resources and overhead are committed.

Direct attack

Suppression in which control lines are constructed next to or very close to the main fire edge.

Indirect attack

Suppression in which control lines are located some distance from the fire edge and the intervening fuel is burned out.

Confine

Tactical actions to manage a fire within a predetermined area or perimeter, usually defined by geographic features.

Suppression

A management action intended to extinguish a fire or alter its direction of spread.

Control line

An inclusive term for all constructed or natural fire barriers and treated fire edges used to control a fire.

Fire line

The part of a control line that is scraped or dug to mineral soil.

Backfiring

Intentionally setting fire to fuels inside the control line to slow, knock down or contain a rapidly spreading fire.

Spot fire

A fire set outside the perimeter of the main fire by flying or rolling sparks and embers.

Hotspotting

Checking the spread of fire at points of more rapid spread or special threat, usually as an early step in prompt control.

Mapping and data terms

Fire perimeter

The entire outer edge or boundary of a fire.

Map method

The attribute recording how a perimeter polygon was derived, from GPS-walked through infrared image interpretation to hand sketch.

IRWIN identifier

The unique identifier assigned to each incident record in the interagency integration service, used to join records across systems.

Quarantined record

An incident record held out of public services because of potential conflicts with another record for the same ignition.

Fall-off rule

The age-based rule that removes stale records from the current services, enforced hourly.

ICS-209

The incident status summary report; its versions are recorded as initial, update or final.

Size class

The letter classification of fire size, where A and B are under 10 acres, C is 10 to 100 acres, and D through L are larger.

Satellite terms

Thermal anomaly

A pixel measured as anomalously hot relative to its neighbors, which is what an active fire product actually detects.

Confidence

A 0 to 100 percent value summarizing how cleanly a pixel passed the detection algorithm's tests; grouped into low, nominal and high classes.

Fire radiative power

Radiant energy released by the fire, reported in megawatts; higher values indicate greater intensity or a larger burning area.

Commission error

A false detection, caused by sources such as sun glint, bright surfaces, gas flaring, steel mills or urban structure fires.

Ultra real-time

Detections delivered in under 60 seconds of satellite overpass through direct-readout ground stations.

Full disk, CONUS and mesoscale scans

The three geostationary scan modes, refreshing every 5 to 15 minutes, every 5 minutes, and every 30 or 60 seconds respectively.

Weather terms that change fire behavior

Red Flag Warning

A fire-weather product calling attention to weather of particular importance to fire behavior, issued for conditions of limited duration that may produce extreme burning conditions.

Extreme fire behavior

Behavior that ordinarily precludes direct control, involving high rates of spread, crowning, spotting, fire whirls or a strong convection column.

Prescribed burning

Burning of forest or range fuels on a specific area under predetermined conditions to fulfill management objectives.

Blackline concept

Containment using a fire line with all adjacent fuels inside the line burned out for a safe distance.

If a term on a map is not in this list, the incident record's own field definitions are usually the fastest way to resolve it, and the state-by-state context in our US wildfire map guide explains how the same fields are used differently by region.

Frequently Asked Questions About Wildfire Tracking