CFB Wainwright: The Flank Became the Front

New satellite imagery shows a 2-km fire line turned into a wall of fire when a forecast cold front hit at CFB Wainwright. It explains how soldiers had 30 seconds to a minute, and why DND’s account doesn’t fit the record.

By Kevin Duska Jr. (Prime Rogue Inc.) · Calgary · Oct. 19, 2026 · All times Mountain

What Really Happened at CFB Wainwright

When the wind swung northwest over CFB Wainwright at 3:16 p.m. on Oct. 17, the fire on the range did not need to travel to become dangerous. Its long side became its front.

New Sentinel-2 satellite imagery from Oct. 19 shows what that meant on the ground. A single run of burned grass, about 2.1 km wide and up to 5.9 km long, stretches southeast from the line where a NASA satellite saw fire at 12:17 p.m. The scar is almost exactly as wide as that fire line was long.

That shape answers a question families and readers have been asking since Maj.-Gen. Wade Rutland relayed what his commander on the ground told him: “within 30 seconds to a minute they saw no fire, and then there was a fire right on top of them.” The soldiers’ account holds up. What does not hold up is the official account of the fire around them.

Sgt. Morgan Wong, Pte. Shaun Farrell, Pte. Aidan Goulding and Pte. Hayden Lykkemark died. Thirty-nine others were taken to hospital. Nothing in this article assigns blame to any person or alleges that anyone acted with intent. It sets public records beside public statements.

What we found

• The fire was on the range for hours. Landsat 8 recorded a fire line about 1.9 km long at 12:17 p.m., and NOAA-20 saw it again at 2:49 p.m.
• The burn scar is one run, southeast. About 11 km² burned along an axis of 131°, the direction a northwest wind pushes a fire. About 85% of it lies outside the area satellites saw burning at 2:49 p.m.
• The wind change was forecast and warned. Environment Canada’s 4:30 a.m. and 10:48 a.m. forecasts called for wind “west 50 gusting to 80 this afternoon.” A wind warning naming Wainwright went out at 2:14 p.m. The base station recorded an 82 km/h gust at 3:16 p.m.
• A 30-to-60-second warning is physically credible. A fire covering a kilometre in that time is not: it would have to move at 60 to 120 km/h, as fast as or faster than the wind driving it.
• Five official statements conflict with the record: on when the fire started, the fire danger, the wind, the distance and the speed.

What the satellites show

The new imagery comes from the European Space Agency’s Sentinel-2, which photographed the base at 10-metre resolution on Oct. 19. We compared it with a cloud-free Sentinel-2 image from Oct. 4 and mapped the change with a standard burn-severity index (dNBR).

The new burn starts at the fire line Landsat 8 recorded at 12:17 p.m., between Jamieson Lake and Peterson Pond, and runs southeast. The lakes bound its southwest side. Its northeast side is a nearly straight edge.

Five measurements matter:

1. Area. About 11 km² burned in the main run. Our earlier estimate from coarser Sentinel-3 imagery, 7.5 to 8.3 km², was too low.
2. Direction. The long axis of the burn runs at 131°, almost exactly downwind of the northwest wind recorded after 3:16 p.m. (306° to 310°).
3. Width. The burn is about 2.1 km wide across that axis. The 12:17 p.m. Landsat fire line was about 1.9 km long.
4. Reach. The furthest burned ground is about 5.9 km from the fire’s starting area. About 1.1 km² burned east of the road that marks the edge of the training area, onto farmland, consistent with the municipal district’s report that the fire “crossed the base fence.”
5. Timing. About 85% of the burned area lies outside the pixels in which NOAA-20 saw fire at 2:49 p.m. Most of the burning happened after that pass.

The burn was most severe 2 to 4 km along the run, where a wind-driven grass fire would be at full speed. Satellites detected heat at the southeast end of the scar again at 2:10 and 2:30 p.m. on Oct. 18, inside the lobe east of the road.

A separate burned patch of about 1 km² sits southwest of the main scar. It matches a NOAA-21 fire detection from Oct. 14 and is not part of this fire.

The weather was forecast

The Canadian Armed Forces memo reported by Global News called the wind change “unexpected and sudden.” Rutland spoke of “uncommonly strong winds” and called what happened “uncommon, rapid.” Environment and Climate Change Canada’s archived bulletins tell a different story.

• 4:30 a.m. and 10:48 a.m. ECCC’s public forecast for Lloydminster – Wainwright – Vermilion – Provost called for wind “becoming west 50 gusting to 80 this afternoon,” with a risk of a thunderstorm.
• 1:30 p.m. ECCC issued a dust storm warning for the counties east of Edmonton (Lamont, Two Hills, Minburn, Leduc-Camrose) as a line of storms moved east: gusts to 100 km/h “producing near zero visibility at times in blowing dust.” That warning did not cover Wainwright. It shows what the storm line was doing upwind.
• 2:14 p.m. ECCC added Lloydminster – Wainwright – Vermilion – Provost to a yellow wind warning, with high confidence: “Wind gusts up to 100 km/h are expected this afternoon. A line of showers and weak thunderstorms is located from east of Edmonton to near Red Deer and is moving east.”
• 3:05 p.m. ECCC ended the warning everywhere else and kept it for the Wainwright region alone.
• 3:16 p.m. The CFB Wainwright weather station recorded a gust of 82 km/h.
• By 4 p.m. The wind had swung from southwest (236°) to northwest (306°), averaging 51 km/h, and the temperature had fallen from 23.3 C to 12.9 C in an hour.

The front crossed Alberta station by station. Edmonton International recorded the shift between 1 and 2 p.m., Vegreville and Camrose between 2 and 3 p.m., and Wainwright between 3 and 4 p.m. NOAA’s GOES-18 weather satellite showed the storm line approaching for at least 20 minutes before the gust hit. CBC meteorologist Johanna Wagstaffe described it as a cold front that shifted winds “from southwesterly to northwesterly,” calling it “a significant change for fire behaviour.”

The fire danger was not low. Natural Resources Canada’s Canadian Wildland Fire Information System recorded a Fire Weather Index of 30.8 at the CFB Wainwright station that day, in the Extreme class. Humidity at the base was 16% to 17% at 2 and 3 p.m. Recomputing the index hour by hour from the station’s own observations, we get about 59 at 3 p.m. and about 91 at 4 p.m., after the front. Those hourly figures are our calculation, using the standard Canadian equations; the 30.8 is NRCan’s.

How the flank became the front

A grass fire pushed by a steady wind grows long and narrow. Its head runs downwind, and its two long sides, the flanks, burn slowly outward.

From noon until about 3:15 p.m., the wind at Wainwright blew from the southwest. A fire on the range under that wind would run northeast and develop a long flank facing southeast. The Landsat line at 12:17 p.m. ran northeast to southwest, at about 55°.

Then the wind turned roughly 70°, from southwest to northwest. The side that had been burning slowly sideways now faced straight downwind, along its entire length.

Australia’s national science agency, CSIRO, has studied this for decades. In their reference text Grassfires, Phil Cheney and Andrew Sullivan write: “An alteration in wind direction can change a flank fire to a head fire, which within 15 seconds can spread at its maximum speed.” A CSIRO review of the 1983 Ash Wednesday fires describes “a 90º wind direction change, turning extended lengths of eastern flank into headfires.” At Wangary, in South Australia, in 2005, a wind change “turned the approximately 55 km long eastern flank of the fire into a headfire.” Nine people died.

The Wainwright burn scar has that shape: a front as wide as the old flank, running downwind of the new wind.

Why 30 seconds is real, and a kilometre in 30 seconds isn’t

Rutland told reporters the fire was “a kilometre away” when the wind shifted. In the Globe and Mail, CTV, Global and CBC versions of his remarks, the fire covered that distance in “30 to 60 seconds.” The Associated Press reported it as “36 seconds,” which appears to be a mishearing of “30, 60.” Prime Minister Mark Carney told the House of Commons the fire “jumped nearly a kilometre in less than a minute.”

As a description of the fire’s speed, that cannot be right. A kilometre in 30 seconds is 120 km/h. A kilometre in 60 seconds is 60 km/h. The strongest gust recorded at the base all afternoon was 82 km/h.

Peer-reviewed research on grass fires finds they spread at roughly 20% of the open wind speed in strong winds (Cruz, Alexander and Kilinc, 2022). At Wainwright’s 51 km/h post-front wind, that is about 10 km/h, or about six minutes per kilometre. At the 82 km/h gust, about 16 km/h, or under four minutes. Some of the fastest grass fires ever measured, at Boonoke in 1987 and in short bursts at Wangary in 2005, reached 27 to 30-plus km/h. Even at those speeds, a fire needs about two minutes to cover a kilometre.

The soldiers’ words point somewhere else. “They saw no fire, and then there was a fire right on top of them” describes how much warning they had, not how fast the fire moved. Four well-documented effects can shrink a fire that is minutes away into seconds of warning. All four were present on Oct. 17.

1. The whole side arrives at once

When a flank becomes the head, there is no single tongue of flame to watch approaching. A 2-km line advances together. There is no end to see coming and no way around it.

2. Smoke and dust run ahead of the flames

Before the change, the southwest wind carried smoke northeast, away from anything southeast of the fire. After the change, the northwest wind pushed smoke and dust southeast, ahead of the front and toward anyone downwind. Upwind on the same storm line, ECCC was already warning of “near zero visibility” in blowing dust. In Wainwright, resident Tanya Porter saw “a wall of smoke” and asked: “Is that a fire or dust?” A front moving behind its own smoke can be close before anyone sees flame.

3. Embers start fires ahead

In strong wind, burning grass and debris are thrown ahead of the front and light spot fires. Researchers warn that spot fires can be mistaken for the main front. For people downwind, fire can appear around them before the main front arrives.

4. Stress distorts time

In controlled studies, people who watch a stressful event consistently overestimate how long it lasted, and more stress produces bigger overestimates (Loftus et al., 1987). “Thirty seconds to a minute” is a human account of warning time, not a stopwatch reading. If anything, the warning may have been shorter.

Put together, a front moving at 10 to 16 km/h, hidden by smoke and dust, throwing embers ahead and arriving along its full width, can leave seconds of warning without breaking any law of physics. That is also why the decisions that mattered most had to come before 3:16 p.m. Once the front arrived, there was no time left to make them.

CSIRO researchers have a name for this: the “dead-man zone.” Their 2001 paper warns that “if the wind direction changes, the fire can advance so rapidly that the firefighters have very little time to seek refuge in the burnt area.” Protection from a flank under a forecast wind change comes from distance and from decisions made in advance.

What was said, and what the record shows

1. When the fire started

Said: DND’s news release put the incident at “approximately 1600 hrs (MDT),” and the CAF’s first statement said a fire was “reported” then. At an Oct. 18 news conference, Rutland said “a grass fire ignited in the training area” at about 4 p.m.

Record: Landsat 8 recorded a fire line about 1.9 km long at 12:17 p.m. NOAA-20 recorded it at 2:49 p.m. GOES-18 recorded heat at the site from about 3:10 p.m., intensifying through 3:40 p.m. Rutland himself said range control “knew there was a fire for several hours.” The Globe and Mail reported that DND said the fire ignited “at about noon.” Global News also reported Rutland saying the fire “did not start within the training area,” which conflicts with his own “ignited in the training area.”

Our reading: “About 4 p.m.” is best read as when the fire was reported or reached the soldiers, not when it started. DND has not said which.

2. The fire danger

Said: An internal CAF memo, reported by Global News and the Globe and Mail, said the “fire index” for CFB Wainwright “was low.” No official has said this on the record.

Record: NRCan’s Fire Weather Index for the base station was 30.8, Extreme. The Globe reported that NRCan’s system showed fire danger “very high or extreme” near Wainwright that afternoon. Humidity was 16% to 17%. No rain had fallen since Oct. 12.

Our reading: We don’t know which index the memo meant. DND has not said.

3. The wind

Said: The memo described an “unexpected and sudden wind change.” Rutland said: “This was uncommon, rapid, and I certainly hadn’t seen anything like it.” He added that the weather for most of the day was “well below 50 kilometres.”

Record: Two ECCC forecasts that morning called for gusts to 80 km/h in the afternoon. A wind warning naming Wainwright went out at 2:14 p.m., an hour before the gust. Rutland is right that the wind before the front was moderate: the station’s strongest gust before the front was about 51 km/h, at 2:48 p.m.. The change was sudden. It was not unexpected.

4. The distance

Said: Chief of the Defence Staff Gen. Jennie Carignan told CTV: “We understand there was a fire that was far removed from the training area, and the fire basically behaved in an unexpected fashion.”

Record: Rutland said the fire was “a kilometre away” when the wind shifted. The satellites place the fire on the range itself from 12:17 p.m., and the burn scar starts there. A kilometre downwind of a 2-km fire line, with a wind change forecast, is not far removed.

5. The speed

Said: “That is not normal for a fire to move a kilometre” in 30 to 60 seconds (Rutland). The fire “jumped nearly a kilometre in less than a minute” (Carney).

Record: That would take a fire speed of 60 to 120 km/h. The published research and the station’s wind data put the likely speed at 10 to 16 km/h, or four to six minutes per kilometre.

Our reading: The soldiers’ 30 to 60 seconds is credible as warning time. Turned into a spread rate, it suggests the fire was too fast for anyone to have anticipated. The record shows the danger was foreseeable for hours. Only the final approach was fast.

What we don’t know

• Where C Company was in relation to the fire, and how far it was from the flank when the wind changed.
• Whether range control or the unit received the ECCC forecasts and the 2:14 p.m. warning, and what the Wainwright Range Standing Orders require when fire danger is Extreme and a fire is burning on the range.
• What index the CAF memo described as low, and when that call was made.
• Whether the fire was being monitored or actively suppressed before 3:16 p.m.
• How visibility actually was where the soldiers were. Smoke and dust are the most likely explanation for the short warning, but only witnesses and DND’s records can confirm it.

The Canadian Forces National Investigation Service is investigating with RCMP support. Carignan has also announced a board of inquiry, which she intends to make public, a technical investigation and an Alberta coroner’s investigation. Rutland declined to tell the Globe and Mail what rules govern training in high winds, and DND could not immediately point to the relevant section of its manual.

Questions for National Defence

Before publishing our Oct. 19 findings, Prime Rogue sent them to National Defence and asked:

1. Is the fire NASA detected at 12:17 and 2:49 p.m. the fire that reached C Company? If so, will DND correct the statement that a fire “ignited” at about 4 p.m.?
2. What index did the CAF memo describe as low, what was its value, and when was it set?
3. Did range control have ECCC’s forecasts and its 2:14 p.m. wind warning before the wind changed?
4. When did the fire reach C Company, and how far was the company from the fire when the wind changed?
5. What do the Wainwright Range Standing Orders require for live-fire training when fire danger is Extreme and a fire is burning on the range?

National Defence had not responded by publication. We will add any response.

How we did this

• Burn scar: Copernicus Sentinel-2 Level-1C imagery from Oct. 19 (raw reflectance) and Oct. 4, resampled to a 10-metre grid. Burn mapped as a dNBR (bands 8A and 12) above 0.2, with water masked. Scar axis and width from a principal-component fit of the mask.
• Fire detections: NASA FIRMS (Landsat 8 and 9, NOAA-20, NOAA-21 and Suomi NPP VIIRS) and NOAA GOES-18 full-disk imagery.
• Weather: ECCC forecasts, warnings and SWOB observations, archived unmodified from the MSC Datamart. ECCC hourly climate data for stations from Edmonton to Coronation.
• Fire danger: NRCan CWFIS daily FWI for the CFB Wainwright station. Hourly values recomputed with Van Wagner’s hourly fine fuel moisture method; the noon recomputation matches CWFIS exactly.
• Fire spread: published peer-reviewed research, cited below. The animation reveals the real burn scar in order of distance from the 12:17 p.m. fire line at a steady 12 km/h. Its timing is illustrative.
• Official statements: quoted from the DND release, the PMO transcript and news reports, linked below. Where reports differ, we say so.

Correction policy: if you see an error, write to kevin at primerogueinc.com or press at primerogueinc.com

Assertion ledger

Every concrete claim in the article, the evidence behind it, and how confident we are. “Primary” means we hold the original record.

Claim	Evidence	Confidence
Fire line about 1.9 km long at 12:17 p.m., oriented about 55°	59 Landsat 8 detections, NASA FIRMS; principal-axis fit	High (primary)
NOAA-20 fire at 2:49 p.m. near the start area	NASA FIRMS VIIRS	High (primary)
GOES-18 heat from about 3:10 p.m., intensifying to 3:40	GOES-18 L1b band 7 anomaly	Medium: thin cloud over the site until about 2:40 p.m. limits earlier frames
About 11 km² burned in one run	Sentinel-2 dNBR above 0.2, water masked	High: 11.1 km² at 0.2 and 10.1 km² at a stricter 0.27; a looser 0.15 threshold also picks up older scars and noise
Burn axis 131°, width 2.1 km, reach 5.9 km	PCA of the burn mask	High
About 85% of burn outside 2:49 p.m. VIIRS pixels	Mask against VIIRS pixel footprints	High
About 1.1 km² burned east of the road at -110.8749	Mask	High; the road as the training area’s edge comes from imagery and the MD’s “crossed the base fence”
Oct. 18 hot spots at the southeast end, 2:10 and 2:30 p.m.	NASA FIRMS VIIRS	High (primary)
Southwest patch is an Oct. 14 fire	NOAA-21 detection at 2:51 p.m. Oct. 14; patch already present before Oct. 17 is unconfirmed	Medium
Forecasts at 4:30 and 10:48 a.m.: “west 50 gusting to 80 this afternoon”	FPCN16 CWWG bulletins, MSC Datamart	High (primary)
1:30 p.m. dust storm warning upwind, “near zero visibility”; did not cover Wainwright	WWCN16 CWWG 171930	High (primary)
2:14 p.m. wind warning added Wainwright; 3:05 p.m. kept for Wainwright only	WWCN16 CWWG 172014 and 172105	High (primary)
82 km/h gust at 3:16 p.m.; NW 51 km/h at 4 p.m.; 23.3 to 12.9 C	CZWN SWOB observations	High (primary)
Pre-front gust about 51 km/h at 2:48 p.m.	CZWN SWOB	High (primary)
Front reached Edmonton 1–2 p.m., Vegreville and Camrose 2–3 p.m., Wainwright 3–4 p.m.	ECCC hourly data	High (primary)
FWI 30.8, Extreme	NRCan CWFIS station data	High (primary)
Hourly FWI about 59 at 3 p.m. and 91 at 4 p.m.	Our calculation (Van Wagner hourly FFMC); labelled as ours	Medium
Grass fires spread at about 20% of wind speed	Cruz, Alexander & Kilinc 2022	High (peer-reviewed)
Flank becomes head “within 15 seconds”	Cheney & Sullivan, Grassfires (CSIRO), quoted via secondary copy	Medium: check the printed edition
Wangary 2005: 55-km flank became head; nine deaths	CSIRO/Bushfire CRC simulation report; SA CFS	High
Stress leads people to overestimate duration	Loftus et al. 1987	High (peer-reviewed)
Rutland: “30 seconds to a minute” quote, “several hours,” “a kilometre away”	CBC live blog, Globe and Mail; secondhand from an unnamed commander	High that he said it
AP’s “36 seconds” is likely a mishearing	Four outlets have “30 to 60”	Medium
CAF memo said fire index “was low”	Global News, Globe and Mail; internal document, not seen by us	Medium
Globe: DND said ignition “at about noon”	Single outlet	Low to medium; attributed
Rutland: fire “did not start within the training area”	Global timeline only	Medium; attributed
Carignan “far removed” quote	CTV/CP24	High
Carney “nearly a kilometre in less than a minute”	PMO transcript	High (primary)
Smoke and dust shortened warning	Inference from the wind shift, the ECCC dust warning upwind and the resident quote	Inference; stated as the most likely explanation

Sources and links

These are the link placements for the article body. News and government sources are follow. Social, mirror and commercial sources are nofollow.

Official statements

• DND news release
• PM Carney told the House of Commons
• Carignan told CTV

News reports

• General Rutland – what his commander on the ground told him
• The Canadian Armed Forces memo reported by Global News
• Global News also reported
• The Globe and Mail reported that DND said
• The Globe reported that NRCan’s system showed
• CBC meteorologist Johanna Wagstaffe
• how the fire spread so quickly

Data

• ECCC’s archived bulletins
• Canadian Wildland Fire Information System
• NASA FIRMS
• Sentinel-2
• GOES-18

Research

• Cruz, Alexander and Kilinc, 2022
• 10% wind speed rule of thumb
• A CSIRO review of the 1983 Ash Wednesday fires
• At Wangary
• “dead-man zone” : Cheney, Gould & McCaw 2001
• Loftus et al., 1987
• Grassfires (Cheney & Sullivan, CSIRO Publishing, 2008)

Prime Rogue, earlier coverage

• storm line approaching
• NASA satellite saw fire at 12:17 p.m.