Wildfire detection is a race against exponential growth. A fire that takes an hour to reach one hectare can, under wind and dry fuel, cover ten times that area in the following hour. Traditional detection — watchtowers, phone reports, patrol aircraft — depends on someone seeing smoke, being in range of a phone signal, and reporting it accurately. Satellite-based wildfire detection removes that dependency: sensors in orbit scan the same ground repeatedly, day and night, whether or not anyone is watching.
How satellite-based wildfire detection works
Active-fire detection relies on thermal anomaly sensing: a fire radiates strongly in the mid-infrared band, producing a sharp contrast against the surrounding background temperature. Polar-orbiting sensors such as VIIRS (carried on the Suomi NPP and NOAA-20/21 satellites) resolve this anomaly down to roughly 375 metres and publish detections through NASA FIRMS within one to three hours of overpass. Geostationary instruments such as the Flexible Combined Imager on Meteosat Third Generation trade spatial resolution for cadence, scanning the full disk every ten minutes and catching ignitions several hours earlier than a polar pass alone would allow.
Why detection speed matters
Fire spread is not linear. Early suppression, while a fire is still small and access is still possible, is disproportionately more effective than intervention after a fire has established multiple fronts. Every additional hour between ignition and the first verified alert is an hour of uncontrolled growth, particularly under the wind and low-humidity conditions typical of Mediterranean fire seasons. This is why OrbiVigil treats time-to-detection as a primary engineering metric, not an afterthought.
From single hotspot to confirmed event
A raw thermal detection is not the same as a confirmed wildfire. Refineries, steel plants, cement works, gas flaring and even large solar installations all produce persistent thermal signatures that resemble a fire pixel in isolation. A detection system that reports every hotspot as a fire quickly earns false-alarm fatigue among the people who need to trust it. Reliable wildfire detection therefore requires cross-checking each candidate against a registry of known industrial heat sources, against independent sensors, and against short-term persistence — a real ignition typically grows and moves; an industrial source does not.
How OrbiVigil does this
OrbiVigil's fusion engine ingests Tier 1 minutes-scale signal — NASA FIRMS/VIIRS, Meteosat Third Generation active-fire products and Sentinel-3 SLSTR fire radiative power — the moment it is published, and cross-references every candidate against a persistent-hotspot registry built from historical detections at known industrial sites. Confirmed detections are promoted through a five-level confidence ladder, from a first observation through to a critical, human-validated event, using Sentinel-2 optical imagery and Copernicus EFFIS event data for corroboration. Every promotion above the initial observation level requires analyst validation before an alert is issued. OrbiVigil never issues a public evacuation instruction directly: confirmed events are delivered as CAP 1.2 messages to the authorized emergency authorities responsible for public warning.
Detection is only the first step. What happens after a fire is confirmed — how its risk is monitored, how its spread is forecast, and how that forecast becomes an alert — is covered in our related pages on satellite wildfire monitoring, the sensor physics behind thermal satellite fire detection, and the broader natural disaster early warning platform that OrbiVigil Fire belongs to.