VIIRS vs MODIS active fire detection: what the pixel size tells you

If you've pulled hotspot data into your GIS during an active incident, you've probably loaded VIIRS and MODIS fire detections in the same afternoon and watched them disagree about where the fire actually is. Two different sensors, built at two different pixel sizes and running on two different satellite programs, produce two different pictures of the same fire. Neither one was designed to hand you a perimeter.

Two sensors, two pixel sizes

MODIS flies on Terra and Aqua, has been collecting fire data since the early 2000s, and its active fire product works off roughly 1 km pixels. That's the legacy standard a lot of older fire detection workflows were built around, and it's still useful for catching the overall shape of a large, actively spreading fire from a distance.

VIIRS is newer, carried on Suomi NPP and the NOAA-20 and NOAA-21 satellites, and its active fire product runs on the 375 m I-band, roughly one-seventh the pixel area of MODIS. Across the three platforms you get more overpasses a day than MODIS alone provides, and the finer VIIRS fire pixel resolution picks up smaller, cooler, or more isolated heat sources, the kind of spot fire or slopover that a 1 km MODIS cell can miss or blur into a much bigger-looking footprint than what's burning on the ground.

What the pixel size changes on the ground

A MODIS pixel flags as "fire" if even a corner of that 1 km cell is hot enough, which means a single detection can represent anything from a wind-driven run to a smoldering log pile, and you can't tell which from the point alone. Drop that same fire into a VIIRS grid and the tighter 375 m pixel narrows the box considerably. A two-acre flare-up in a drainage shows up closer to where it sits instead of getting smeared across a kilometer of terrain that mostly isn't burning.

Both sensors are polar-orbiting, so revisit timing depends on where your incident falls relative to each satellite's swath that day, not a fixed schedule. Thick smoke and cloud cover can mask heat from both. Gaps in active fire detection satellite data trace back to orbital geometry and atmospheric conditions, not a dropped feed on your end, and that's worth knowing before you flag a quiet period on the map as something more than it is.

Where this leaves the IC desk

VIIRS and MODIS hotspots are good for exactly what they're built for: confirming that something is burning right now, with a timestamp for when the satellite last looked. Neither one draws a perimeter. A hotspot is a point, and a point isn't a boundary. For suppression planning, division assignments, and the morning briefing, someone still has to translate scattered heat points into an actual fire edge on a map, which is the gap a SWIR-based burn severity layer is built to close, separating the active front from ground that's already burned and updating that line daily while the incident is live. That's the role Fire Perimeter Map fills once the hotspot feed has told you something's moving.

It's also worth remembering what happens when the fire goes quiet. Once an area stops producing heat, VIIRS and MODIS stop detecting it, full stop, there's nothing left for them to see. But the burn severity picture doesn't disappear with the heat. BAER teams and recovery planners still need to know how hot different parts of the footprint burned, and that classification keeps working well after the last hotspot has cooled.

If your morning routine involves reconciling two hotspot layers by hand before you can even start on the actual perimeter, it might be worth seeing what a daily, severity-graded fire line looks like for your next incident.