The storm line, then the fire: every satellite record of the CFB Wainwright fire, on one map

NOAA’s weather satellite shows a line of storms crossing the training area minutes before an 82 km/h gust, and the fire flaring right behind it. A CAF memo called the wind change “unexpected and sudden.”

Kevin Duska Jr. | Prime Rogue Inc. | Oct. 19, 2026

At 2:50 p.m. on Oct. 17, a line of storm clouds sat just west of Canadian Forces Base Wainwright. By 3:10 p.m. its leading edge had reached the training area. At 3:16 p.m., the base’s weather station recorded a gust of 82 km/h. By 3:30 p.m., a fire on the range was hot enough to show up from 36,000 km away.

All of it was recorded by NOAA’s GOES-18 weather satellite, which photographs North America every 10 minutes, and all of it is public. Four soldiers died in the fire that followed. A CAF memo reported by Global News described the wind change as “unexpected and sudden.”

Prime Rogue has now put every satellite record of the fire on a single map, in the order it happened. Each layer was made independently, by a different satellite, agency or sensor. They tell one story.

What the records show, in order

1. 12:17 p.m. NASA’s Landsat 8 detects fire at 59 points between Jamieson Lake and Peterson Pond, on the training area.
2. 2:49 p.m. NOAA-20 detects it again, at 8 points.
3. 2:50 p.m. On GOES-18, the storm line is visible just west of the base.
4. About 3:10 p.m. The line’s leading edge reaches the site. NOAA’s automated fire product first flags heat there.
5. 3:16 p.m. The base’s station records the 82 km/h gust, from the west. By 4 p.m. the wind is from the northwest and the temperature has dropped 10 degrees.
6. 3:30 to 3:50 p.m. The fire’s heat is plainly visible on GOES-18, on the site, as the line passes.
7. Oct. 18. Sentinel-3 shows a new burned area of roughly 7.5 to 8.3 km² running southeast from the detections. That is the direction a northwest wind pushes a fire.

National Defence’s news release says a fire “ignited” at approximately 4 p.m. As of 10:00 AM MT Monday, it had not been corrected.

A storm you could watch coming

The six panels above cover about 290 by 210 km around the base. The bright, lumpy band is the storm line: the “line of showers and weak thunderstorms” that Environment Canada warned at 2:14 p.m. was moving east, with gusts up to 100 km/h.

You can watch it come. At 2:50 p.m. it is west of the base. At 3:10 p.m. it is over the training area. Six minutes later the base station logs its strongest gust of the day. At 3:30 p.m., an orange signal appears where the fire was burning.

Weather observer Ryan Purciel spotted this first. On Oct. 18 he posted a GOES loop noting that the incident appeared to happen, or intensify, just after a strong outflow boundary crossed the area. An outflow boundary is the gust front a line of storms pushes ahead of itself. Our processing of the raw data agrees with his reading.

Two ways of drawing the same scans

Satellite heat signals can be argued over, so we drew the scans two ways. On the right is our view: natural colour, with orange marking pixels at least 10°C hotter than their surroundings in the infrared. On the left is the standard “fire temperature” colouring used by forecasters, which applies no threshold at all. In that view, intense heat simply shows up red.

Neither shows a clear signal on the site through 3:20 p.m. Both show the fire at 3:30 and 3:40 p.m. The left view looks blue because that is how the recipe draws cloud and land.

The absence of a signal before 3:30 does not mean there was no fire. GOES pixels are 2 to 5 km wide at this latitude. A small grass fire does not register until it grows. Landsat, with 30 m pixels, saw it at 12:17.

Every record, one map

The map puts each record where it belongs, on a cloud-free Sentinel-2 image from Oct. 4. The yellow Landsat points and blue NOAA-20 boxes sit between the two lakes. The wind arrows show southwest air until about 3 p.m., then northwest after the gust. The slanted outline is the hottest GOES-18 pixel at 3:30 p.m., drawn at its true size on the ground. The red patch is the area that darkened between Oct. 4 and Oct. 18.

The GOES pixel sits partly northeast of the burned area. That is expected: GOES positions are accurate to about 1 to 2 km, and heat spills into neighbouring pixels. The burned area itself starts at the detections and runs southeast, which fits a fire turned by a northwest wind.

What this means for National Defence’s account

The CAF memo, as reported, described the wind change as “unexpected and sudden.” On the government’s own records, it was neither.

• It was forecast. Environment Canada’s 5 a.m. forecast called for strong winds. Its 2:14 p.m. update forecast “west 60 gusting to 90 late this afternoon.”
• It was warned of. At 2:14 p.m. Environment Canada extended a wind warning to the Wainwright region, describing a line of storms moving east with gusts up to 100 km/h, high confidence. At 3:05 p.m. it ended the warning everywhere else and kept it for Wainwright’s region alone.
• It was visible. On public satellite imagery updated every 10 minutes, the storm line can be seen bearing down on the base for at least 20 minutes before the 3:16 gust.

The release’s account of a fire that “ignited” at about 4 p.m. fares no better. The fire was on the range by 12:17 p.m. and still burning at 2:49 p.m. Maj.-Gen. Wade Rutland said Range Control had known of a fire for several hours, the Edmonton Journal reported. GOES-18 shows fire flaring on the site by 3:30 p.m., half an hour before the release’s time.

We do not allege the account was written to mislead. It is contradicted by the record, and it is still on the government’s website.

The strongest counterpoint

There is one piece of the record National Defence could point to. Environment Canada’s 3:05 p.m. bulletin said gusts to 90 km/h were expected “this evening.” The gust came at 3:16 p.m.

If the people running the range relied on that line, they may have believed they had hours, not minutes. That would make the timing a surprise. It would not make the wind unexpected. The 2:14 p.m. forecast said “late this afternoon,” the warning described the storm line itself, and the line was visibly approaching.

We also do not know whether Range Control was watching satellite or radar imagery, or what weather briefing it had received. Those are records National Defence holds, and has not chosen to disclose.

What the satellites cannot show

• What started the fire.
• Where C Company was, or how far the soldiers were from the fire when the wind turned.
• Whether the fire Landsat saw at 12:17 p.m. is the fire that reached the soldiers. We assess that as likely, not established.
• What Range Control knew, when, and what it decided.
• The burned area is preliminary. It comes from 300 m Sentinel-3 pixels and will be checked against a sharper image.

Questions for National Defence

Prime Rogue has put these questions to National Defence and will update this story with any response.

1. Did Range Control receive Environment Canada’s 2:14 p.m. wind warning, or the 3:05 p.m. update?
2. What weather support was the range receiving on Oct. 17, and did it include radar or satellite imagery?
3. What was the status of the fire on the range at 3 p.m., and who decided training would continue?
4. What do Wainwright’s range standing orders require when fire danger is Extreme?
5. When will the Oct. 18 news release be corrected?

How we did this

The GOES-18 images were made from NOAA’s raw Level 1b full-disk scans, which are openly available. We used bands 1, 2 and 3 for natural colour, bands 5, 6 and 7 for the fire-temperature view, and band 7 (3.9 microns) for heat. Anyone can download the same files and check our work.

Landsat 8 and NOAA-20 detections are from NASA FIRMS. Wind and temperature are from Environment Canada’s CFB Wainwright station. The burned area compares Copernicus Sentinel-3 images from Oct. 4 and Oct. 18. The base map is Copernicus Sentinel-2 from Oct. 4.

Our full reconstruction of Oct. 17, with 13 graphics and the raw records, is here: Every warning light was on. Follow developments on our live updates page

Sgt. Morgan Wong, Pte. Shaun Farrell, Pte. Aidan Goulding and Pte. Hayden Lykkemark were killed in the fire. 39 Canadian Armed Forces members were injured. CAF Family Information Line (24/7): 1-800-866-4546. Canadian Forces Member Assistance Program: 1-800-268-7708.