Every Warning Light Was On: What Most Likely Happened at CFB Wainwright

A fire burning since midday. An Extreme federal fire-danger rating. A wind warning naming Wainwright almost two hours before 4 p.m. Prime Rogue’s reconstruction of the Oct. 17 fire that killed four soldiers, and why National Defence’s account does not hold.

By Kevin Duska Jr. | Prime Rogue Inc. | Published 6 a.m. MT, Monday, Oct. 19, 2026. All times Mountain Time.

The Department of National Defence Is Not Being Transparent About CFB Wainwright

At 12:17 p.m. on Saturday, Oct. 17, a NASA satellite detected fire on the CFB Wainwright training area. At 2:14 p.m., Environment Canada extended a wind warning to the Wainwright region: a line of showers and weak thunderstorms was moving east, with gusts up to 100 km/h expected. Four were killed and 39 injured.

The fire was burning. The military says Range Control knew. The fire danger was Extreme. The wind change was forecast and then formally warned. Training continued.

Our assessment:

On the evidence of the government’s own records, it was not safe to continue live training on the afternoon of Oct. 17 while a known fire burned on the range. This was not an unforeseeable act of weather. It was a decision, and the decision to keep training is the central question National Defence has not answered: who made it, on what information, and under what rules.

National Defence told the public something else. Its news release said a fire “ignited” at about 4 p.m. An internal CAF memo, reported by Global News, said the fire index was low and blamed an “unexpected and sudden” wind change. Each of those claims is contradicted, or left badly incomplete, by records kept by the Government of Canada, NASA and NOAA. Together they gave Canadians a misleading account of how four soldiers died: a sudden fire, low danger, a wind no one saw coming.

The record does not show whether that was deliberate, and we do not allege it was. The first release was written within hours of a mass-casualty event. What the record does show is that by Sunday afternoon National Defence’s own commander had described a fire known for hours, and National Defence still has not corrected a word. We name no individual as responsible; that is for investigators. The families of four soldiers are owed a public account that matches the record.

Every record in this story is public. The raw data is linked at the end so readers can check the work.

A Public Account That Misleads, and Still Stands

Both of National Defence’s news releases, and the memo, point the same way: toward a sudden, unforeseeable event. Here is what each one said, and what it left out.

What National Defence told the public	What it left out
A fire “ignited within the training area” at “approximately 4:00 p.m.” – release, Oct. 18	Fire was detected on the range at 12:17 p.m. Its own commander says Range Control knew of a fire for several hours.
“While conducting training a fire ignited and high wind speeds enabled the fire to grow and spread very rapidly” (second release, Oct. 18	That the wind was forecast at 5 a.m. and warned of at 2:14 p.m., and that the fire danger was Extreme.
The fire index was low (memo, as reported by Global News	The federal government’s own Fire Weather Index for the base was Extreme.
An “unexpected and sudden” wind change (memo, same report)	Environment Canada had warned of strong gusts with high confidence, and narrowed the warning to Wainwright’s region at 3:05 p.m.

Three things make this more than a first-day error.

• It was repeated. A second release the same day again says the fire “ignited.” The internal memo, as reported by Global News, also says a grass fire “ignited” during training.
• It was contradicted by National Defence’s own side, and left standing. On Sunday afternoon, Rutland told reporters Range Control had known of a fire for hours. As of Monday morning, neither release had been corrected, updated or clarified.
• Questions went unanswered. When Prime Rogue asked on Sunday what “ignited” at 4 p.m. referred to, National Defence’s media office pointed to the briefing. It has not said whether the fire Range Control monitored is the fire that reached C Company despite repeated requests for comment.

An account that tells the public a fire started at the moment it turned deadly, in conditions no one could foresee, shifts the question away from the decisions made in the hours before. That is the effect of what National Defence has said, whatever the intent.

What most likely happened

This is our reconstruction. Each step carries a confidence rating: “certain” where it comes straight from an official record, “very likely” or “likely” where it is our assessment from several independent records. It is not a finding of fact, and National Defence has not confirmed it.

1. A fire is burning on the training area by late morning (likely). No satellite detected fire at the site overnight. How it started is unknown. National Defence’s own release says the fire ignited “while conducting training.”
2. By 12:17 p.m. it is burning in a band almost 2 km long (very likely). Landsat 8 recorded 59 fire detections between Jamieson Lake and Peterson Pond. Maj.-Gen. Wade Rutland, commander of 1st Canadian Division, later said Range Control knew of a fire for several hours and was monitoring it, the Edmonton Journal reported.
3. It is monitored, not put out (likely). NOAA-20 still saw it burning at 2:49 p.m. Until about 3 p.m., the southwest wind pushed it northeast, so it grew long and narrow, with a long flank facing southeast.
4. At 2:14 p.m., Environment Canada warns Wainwright (certain). Environment Canada extended its wind warning to the Lloydminster – Wainwright – Vermilion – Provost region, marked “=NEW=”: a line of showers and weak thunderstorms “moving east,” gusts up to 100 km/h, forecast confidence high.
5. At 3:05 p.m., the warning narrows to Wainwright’s region alone; at 3:16 p.m., the front arrives (certain). The base station recorded an 82 km/h gust from the west.
6. Between 3:10 and 3:40 p.m., the fire grows fast (likely). NOAA’s GOES-18 satellite flagged fire near the site, rising from about 77 MW to about 1,200 MW. A Wainwright resident saw “a wall of smoke” at about 3:30 p.m., The Globe and Mail reported
7. At about 4 p.m., the wind turns northwest (likely). The fire’s long southeast flank, 2 to 3 km of it, becomes its front, and it runs southeast. The area burned by Oct. 18 runs southeast from the detections.
8. It reaches C Company (likely, not established). Where the soldiers were has not been disclosed.

It was not safe to keep training

This is our judgment, and here is what it rests on. While training continued that afternoon, every measure the federal government publishes pointed the same way:

• A fire was burning on the range, and had been since at least 12:17 p.m. The military says Range Control knew.
• Fire danger was Extreme. Natural Resources Canada’s Fire Weather Index for the base’s own station was 30.8; Alberta’s Extreme class starts at 29.5. It had nearly doubled from the day before.
• The air was dry. Humidity at the base was 16 per cent at 2 p.m. No rain had fallen since Oct. 12.
• A wind warning was in force from 2:14 p.m., naming Wainwright and warning of the strong gusts that arrived, with high confidence.
• The forecast had been raised at 2:14 p.m. to “west 60 gusting to 90 late this afternoon.”
• The front was measured at the base at 3:16 p.m.: 82 km/h. That is about 45 minutes before the time National Defence gives.

Any one of these might be managed. Together, around a fire still burning at 2:49 p.m., they describe conditions in which a wind-driven grass fire can move faster than people on foot. Canadian Army ranges do tie training to fire danger. At Garrison Petawawa, for example, the 2019 range standing orders restrict which ammunition may be fired when the fire danger is Extreme, and allow “no firing of any ammunition of any type” at Extreme Plus. What Wainwright’s standing orders required that afternoon has not been disclosed. We have asked, and we are filing for them under the Access to Information Act.

The weather did not fail. The forecasts did not fail. On the record available, what failed was the decision to keep troops training while a known fire burned on the range, the fire danger was Extreme and a wind warning was in force. Who made that decision, and what they were told, is now the question.

National Defence’s account, against the record

National Defence’s written account rests on three claims. Each is examined below.

1. “Ignited” at 4 p.m.

What National Defence said about CFB Wainwright

National Defence’s Oct. 18 news release, “Incident at Canadian Forces Base Wainwright,” says: “A fire ignited within the training area of Canadian Forces Base (CFB) Wainwright, Alberta,” “at approximately 4:00 p.m. (local time) on October 17, 2026,” and “The fire spread rapidly beyond the boundaries of the training area.” A second release the same day, naming the soldiers, says: “While conducting training a fire ignited and high wind speeds enabled the fire to grow and spread very rapidly overtaking the unit administrative area.” As of Oct. 19, neither release had been corrected or updated.

At a briefing in Edmonton on Sunday, Maj.-Gen. Wade Rutland gave a different account. Range control was aware there was a fire for several hours and was monitoring it, Rutland said, according to the Edmonton Journal. “The fire was a kilometre away when the wind shifted,” he said, as reported by CTV News.

What The Records Show.

Fire was burning on the range for hours before 4 p.m.

• 12:17 p.m., Landsat 8. NASA’s Fire Information for Resource Management System recorded 59 fire detections in a band about 1.9 km long, between Jamieson Lake and Peterson Pond. FIRMS rates them low or medium confidence, depending on the file.
• 2:49 p.m., NOAA-20. The VIIRS sensor recorded eight fire detections at the same site, with a fire radiative power of 5.8 to 11.5 megawatts each.
• 3:10 to 3:40 p.m., GOES-18. NOAA’s geostationary satellite images the area every 10 minutes. Its fire product flagged a fire pixel near the site at 3:10 p.m. (about 77 MW). At 3:30 p.m. it flagged three pixels totalling about 1,200 MW, and one remained at 3:40 p.m. Cloud obscured the site for much of midday and again from 4 p.m. NASA’s FIRMS archive also lists the 3:10 p.m. GOES detection. This is Prime Rogue’s analysis of NOAA’s raw files. NOAA’s Hazard Mapping System did not include any GOES-18 detection at the base, so we rate the GOES-18 record as likely rather than certain.
• Before that, nothing. Apart from one detection 2.7 km to the south-southwest on Oct. 14, no satellite detected fire here from Oct. 12 to 16, or overnight into Oct. 17. The fire most likely started on the morning of Oct. 17, before 12:17 p.m.

Around 3:30 p.m., Wainwright resident Tanya Porter saw “a wall of smoke,” The Globe and Mail reported

Other satellite passes that afternoon did not record fire at the site, including Suomi-NPP at about 2:29 p.m., NASA’s Aqua at about 3:15 p.m. and NOAA-21 at about 3:33 p.m. Cloud covered parts of the area. NOAA-20 detected the fire 20 minutes after the Suomi-NPP pass, so a miss under cloud is weak evidence that no fire was burning.

Why It Matters For Understanding What Really Happened at CFB Wainwright

Read alone, “ignited” at 4 p.m. tells the public the fire started at the moment it became deadly. The satellite record shows a fire already burning for hours. The military’s own briefing says it was already known. The release has not been changed to say so.

It is likely, but not established, that the fire NASA detected is the fire that reached C Company. National Defence has not said.

2. The memo: a low fire index

What the military said.

A Strategic Joint Staff memo obtained by Global News said the fire index was low before the fire, according to Global’s report (https://globalnews.ca/news/12107414/fire-risk-was-low-before-sudden-wind-shift-sparked-deadly-base-fire-caf-memo/, follow). The same memo says: “During training, a grass fire ignited, and strong winds caused the fire to spread rapidly and shift direction.” The memo has not been published in full. National Defence has not said which index it meant, what its value was, or who set it.

What the records show.

Natural Resources Canada’s Canadian Wildland Fire Information System (CWFIS) publishes a daily Fire Weather Index for the weather station at CFB Wainwright. The Globe and Mail first reported the gap between the federal fire-danger data and the memo.

• Oct. 17: 30.8. On Alberta’s scale, as listed by CWFIS, anything from 29.5 up is Extreme. The day before it was 16.3, which is High.
• The ground was dry. The Drought Code was 461.5 and the Buildup Index was 77.6. No rain had fallen at the station since Oct. 12.
• The neighbours were Extreme too. Vermilion was 31.9, Lloydminster 39.2, Stettler 55.9, Vegreville 58.1 and Coronation 67.2.
• The afternoon was worse than noon. The published index is computed from weather at noon standard time (1 p.m. MDT). Humidity at the base fell to 16 per cent at 2 p.m. and 17 per cent at 3 p.m. Using Environment Canada’s hourly readings, Prime Rogue calculated the index at about 54 at 2 p.m., 59 at 3 p.m. and 91 at 4 p.m. This is our calculation, not a published figure. It uses the standard Van Wagner (1977) method and reproduces the CWFIS noon value exactly.

The daily index has limits: it is computed from noon weather and misses the afternoon peak. But on no reading of the federal data was the fire danger low.

3. An “Unexpected and Sudden” Wind Change

What the military said. The CAF memo reported by Global News described an “unexpected and sudden” wind change. At Sunday’s briefing, Rutland said winds were initially “well below 50 kilometres” an hour, according to The Epoch Times. Reuters reported he said winds were under 50 km/h before the fire. Rutland declined to say whether the incident was preventable, the Edmonton Journal reported. “And if our protocols need to change, then they’ll be updated,” he said, according to The Epoch Times.

What the records show.

Environment and Climate Change Canada (ECCC) forecast strong wind that afternoon, warned of it before 4 p.m., and recorded it at the base. All forecast and warning text below is from the MSC Datamart.

Time (Oct. 17)	Record	What it said
5 a.m.	ECCC forecast, Lloydminster – Wainwright – Vermilion – Provost	Wind “southwest 30 km/h gusting to 50 this morning then west 50 gusting to 80 this afternoon,” with a risk of a thunderstorm.
11 a.m.	ECCC forecast update	Similar strong west wind in the afternoon.
2:14 p.m.	ECCC wind warning, extended to the Wainwright region (marked =NEW=)	“A line of showers and weak thunderstorms is located from east of Edmonton to near Red Deer and is moving east. Strong wind gusts can be expected as this line of showers passes through the area.” Gusts up to 100 km/h. Forecast confidence: high.
2:48 p.m.	CFB Wainwright weather station	Gust of 51 km/h.
3 p.m.	CFB Wainwright weather station	23.3 C, humidity 17 per cent, southwest wind 32 km/h.
3:05 p.m.	ECCC warning update	Warning ended for every other region and kept only for Lloydminster – Wainwright – Vermilion – Provost. Its text now read: “Wind gusts to 90 km/h are expected this evening.”
3:16 p.m.	CFB Wainwright weather station	Peak gust of 82 km/h, from the west.
4 p.m.	CFB Wainwright weather station	12.9 C, humidity 43 per cent, northwest wind 48 km/h.

Lloydminster’s station, about 80 km northeast of the fire site, recorded wind to 98 km/h between 3 and 4 p.m.

The forecasts did not give the minute the wind would change, and they called for west wind rather than northwest. But strong west wind and a risk of thunderstorms were forecast at 5 a.m. The line of showers and weak thunderstorms that brought the gusts was the subject of a warning nearly two hours before 4 p.m., and the gusts were measured at the base’s own station by 3:16 p.m.

The Burned Area

A Sentinel-3 image from Oct. 18, compared with one from Oct. 4, shows a newly darkened area of about 7.5 to 8.3 square kilometres. It begins at the Oct. 17 detections and runs southeast for about 5 km, roughly 2.5 to 3 km wide. That is the direction a northwest wind would drive a fire.

This is preliminary. Sentinel-3 pixels are 300 m across, and some darkening may be seasonal. We will publish a sharper measurement when a clear Sentinel-2 or Landsat image is available.

Fire On This Part of The Range is Recurring

NASA’s archive shows fire detected within 3 km of the Oct. 17 site every year from 2012 to 2026 except 2015, on 45 separate days before Oct. 17, mostly in April and from September to November. Several were within half a kilometre, including Oct. 6, 2012, Oct. 18, 2014, and Oct. 5, 2017. On Oct. 14 this year, three days before the fire, NOAA-21 detected fire on the west shore of Jamieson Lake.

Satellites detect heat, not cause. Many of these are likely prescribed burns or range fires, and the records don’t say which. The pattern raises a question for investigators: was a fire burning on this part of the range treated as normal?

A Question about Speed

Rutland said the fire was about a kilometre away when the wind shifted. “It is not normal for a fire to move a kilometre in 30 to 60 seconds,” he said, as reported by CTV News.

A kilometre in 30 to 60 seconds is 60 to 120 km/h. That is between about three-quarters of and one and a half times the strongest gust recorded at the base, 82 km/h. A 2022 study of 58 grassland wildfires in southern Australia under critical burning conditions found that “the forward rate of fire spread is approximately 20% of the average 10-m open wind speed” (Cruz, Alexander and Kilinc, Fire, 2022). At the 4 p.m. average of 48 km/h, that is about 10 km/h, or six minutes per kilometre. Even using the 82 km/h peak gust instead of the average, it is about 16 km/h, or roughly four minutes per kilometre.

This does not tell us what happened. It means at least one part of the account needs explaining. The fire may have been closer than a kilometre when the wind changed. The soldiers may have been in its path for longer than a minute. Or the distance and the time describe something else, such as spot fires thrown ahead by the wind. National Defence has not said which.

What The Records Cannot Show

• What started the fire.
• Where C Company was, or how far the fire was from the soldiers at any moment.
• What Range Control knew, when, and what it decided.
• Whether the fire NASA detected at midday is the fire that reached the soldiers. It is likely, but not established.

Those answers sit in records National Defence holds: the Range Control log, range bookings and danger-area maps, the Wainwright range standing orders, and the weather briefings the exercise received.

National Defence’s Response

Prime Rogue sent National Defence a summary of these findings and five questions at 0216 AM, after several previous unanswered requests for comment, with a deadline of 8 a.m. ET (6 a.m. MT) today. National Defence did not respond by the deadline. We will add any later response to this story.]

We asked:

1. Is the fire NASA detected at 12:17 and 2:49 p.m. the fire that reached C Company? If so, why does the release say a fire “ignited” at about 4 p.m., and will it be corrected?
2. What index did the CAF memo describe as low, what was its value, and when was it set? How does National Defence reconcile it with Natural Resources Canada’s Extreme rating?
3. Was Range Control aware of Environment Canada’s forecasts and its 2:14 p.m. warning before the fire reached C Company? Does National Defence stand by the memo’s description of the wind change as “unexpected and sudden”?
4. At what time did the fire reach C Company, and how far from the fire were the soldiers when the wind changed?
5. Is the area where the fire was detected within a permanent danger area? Are fires there routinely monitored rather than suppressed, and what do the Wainwright range standing orders require when the fire danger is Extreme and a fire is burning on the range?

Prime Rogue is also filing requests under the Access to Information Act for the Range Control log, the range standing orders, the daily fire-danger ratings used on the range, the weather briefings received on Oct. 16 and 17, and the range bookings for Oct. 17.

According to the CAF memo reported by Global News, Canadian Forces National Investigation Service investigators were sent to Wainwright, and the RCMP offered support. Asked about a board of inquiry, Rutland said, “I would come back in at least a month,” CTV News reported.

Clarifications to Our Earlier Coverage

• Earlier stories put the wind shift at 4 p.m. The base’s station records it between 3 and 4 p.m.: a westerly gust of 82 km/h at 3:16 p.m., and northwest wind by 4 p.m.
• Earlier stories said the wind warning was continued at 3:03 p.m. ECCC’s records show the warning was first issued for the Wainwright region at 2:14 p.m. and updated at 3:05 p.m.
• Earlier stories described the Oct. 18 detections about 5 km southeast as near the base boundary. Satellite imagery suggests they fall on farmland rather than range land. The base’s legal boundary is not confirmed.

The Raw Imagery

These are the unprocessed satellite images behind the analysis, as delivered.

The full set, including every GOES-18 frame, the Sentinel-2 bands and NASA’s original files, is in the raw-data package linked below.

How we did this

Every record in this story is public. Fire detections come from NASA FIRMS (Landsat 8 OLI, NOAA-20 VIIRS) and from NOAA’s GOES-18 fire product, which Prime Rogue read from NOAA’s public archive. Weather records are ECCC forecasts and warnings and hourly observations for the CFB Wainwright station (climate ID 301S001; CZWN). Fire-danger data is Natural Resources Canada’s CWFIS daily station files. Imagery is Copernicus Sentinel-2 and Sentinel-3. Times are converted to Mountain Time.

We use estimative language. “Certain” means the fact comes directly from an official record. “Very likely” means several independent records agree. “Likely” means the evidence points one way, but rests partly on our own analysis or on fewer sources. “High confidence” in our overall assessment means the conclusion holds across those ratings. The raw NASA data, with checksums, is in our raw data package. The full analysis is here.

The soldiers

Sgt. Morgan Wong of the Loyal Edmonton Regiment, serving with 3 PPCLI, and Pte. Shaun Farrell, Pte. Aidan Goulding and Pte. Hayden Lykkemark of 3 PPCLI were killed. National Defence said. 39 Canadian Armed Forces members were injured. Prime Rogue’s thoughts are with their families and with the people of Wainwright.

CAF Family Information Line (24/7): 1-800-866-4546. Canadian Forces Member Assistance Program: 1-800-268-7708.

Related:

• DND said the fire “ignited” at 4 p.m. The military says it was known for hours
• Full satellite analysis
• Live updates

Sources

• NASA FIRMS active fire data, Oct. 12–18, 2026, and archive 2012–2026
• NOAA GOES-18 ABI L2 Fire/Hot Spot Characterization, full disk (https://noaa-goes18.s3.amazonaws.com/index.html, follow)
• ECCC forecasts and warnings (FPCN16, WWCN16), Oct. 17, MSC Datamart
• ECCC hourly observations, Wainwright CFB Airfield 21
• Natural Resources Canada, CWFIS daily fire weather
• Copernicus Sentinel-2 and Sentinel-3 imagery
• National Defence, “Incident at Canadian Forces Base Wainwright”
• National Defence, “Four Canadian Armed Forces members killed during training in CFB Wainwright”
• Global News on the CAF memo
• CTV/CP24 on the Oct. 18 briefing
• Edmonton Journal (via Yahoo) on the Oct. 18 briefing
• The Epoch Times on the Oct. 18 briefing
• Cruz, Alexander and Kilinc, “Wildfire Rates of Spread in Grasslands under Critical Burning Conditions,” Fire 5(2):55, 2022
• Garrison Petawawa Range Standing Orders, 2019
• The Globe and Mail