Field study 01 · Crew pre-positioning
A season written
in two acts.
In 2025, 354 wildfires burned 1.93 million hectares of Manitoba and drove more than 31,000 people from their homes. We rebuilt the season from public data, then re-solved it week by week.
01 · The season
Two waves, a
thousand km apart.
A May wave in the west and centre, around Lac du Bonnet, Flin Flon, Lynn Lake. Then a larger July wave in the far north and east, around Island Lake and Thompson. A fixed footprint can be right for one act. It cannot be right for both.
The old way
Eight bases,
fixed all season.
Spread 20 crews across 8 established attack bases and leave them. When the fires move a thousand kilometres, most of the season ignites out of reach.
38.4%
of ignitions within 100 km of a crew
02 · The loop, running
Move the crews
before the fires do.
Each Monday the model re-solves: where should 20 crews stand this week, given only the trailing two weeks of hotspots? Bronze rings are the optimized positions. No hindsight, just a weekly re-solve.
03 · The result
A third more
reachable.
Same crews. Same fires. Same information a duty officer had that Monday morning. The gap compounds all season, and it held up under stress.
71.5%
within 100 km, re-solved weekly
38.4%
within 100 km, fixed footprint
+31 pts
minimum advantage across 500 stress-tested seasons
04 · How it works
Optimized by math.
Pressure-tested by simulation.
Discover the decision → optimize it to provably best → pressure-test it against 500 alternate seasons → re-solve when Monday moves the world. The same loop Ourus runs on fleets, plants, and grids.
ourus.ai