Roadmap / 0.1.2.6.1
Bears and wildlife
v0.1.2.6.1, commit e78d479. File docs/RESULTS_v0.1.2.6.1_wildlife.md.
Written at this version and not revised since. Later versions may change these numbers; see the roadmap.
Switch: wildlife.enabled (off by default, neutral when off); wildlife.north_troop_share sets the occupier presence in polar-bear range. Uses the complex weather model's regions and calendar.
Runs: 2,000 per variant, seed 20261105, complex weather on. Backtests pass; 15 tests pass.
What the module does (aggregate only)
- Wildlife-vehicle collisions from occupier vehicle-km on rural and night routes, peaking in summer and fall.
- Canada sees 4–8 large-animal collisions an hour (Transport Canada).
- 570 people died in them in 2000–2020; over half the collisions involved moose, 42% happened in Jun–Aug and 33% in Sep–Nov (TIRF).
- ~80% of fleet collisions happen in darkness, with the Trans-Canada Highway the top corridor (Samsara 2026).
- Bear encounters in bear country, April to November (polar bears year-round in the north). Armed troops usually shoot first, so casualties are rare. Polar bears: 73 attacks and 20 deaths across all range states in 1870–2014, mostly predatory (Wilder et al. 2017).
- Polar-bear watch: 3–10% of occupier troops in polar-bear range are tied up on bear watch and camp security.
- Boreal insects (Jun–Aug): a 2–10% effectiveness penalty for occupiers who are not used to them.
- "Iconic incident": US troops kill a polar bear in a widely shared incident. A narrative event with a small, uncertain effect on US support.
- Counting: wildlife deaths count toward US deaths for public opinion. A death is a death.
Results
| Scenario | Variant | Wildlife deaths | Vehicle collisions | Injuries | Bear casualties | Polar-bear incidents | US quits |
|---|---|---|---|---|---|---|---|
| Full annexation | off, Jan | – | – | – | – | – | 64.0% |
| on, Jan | 3.6 | 952 | 78 | 17 | 1.3 | 65.6% | |
| on, Jul | 3.8 | 1,001 | 82 | 18 | 1.3 | 62.8% | |
| on, Jul, bad bear year | 9.1 | 998 | 111 | 53 | 1.3 | 61.8% | |
| Alberta detachment | off, Jan | – | – | – | – | – | 49.4% |
| on, Jan | 0.7 | 159 | 13 | 3 | 1.6 | 48.0% | |
| Arctic seizure | off, Jul | – | – | – | – | – | 28.2% |
| on, Jul | 0.03 | 4 | 0.5 | 0.2 | 2.2 | 31.2% | |
| on, Jul, bad bear year + stronger incident effect | 0.09 | 4 | 0.8 | 0.6 | 2.2 | 34.0% |
All figures are means over a war.
Findings
- Moose matter more than bears. A full occupation produces about a thousand large-animal collisions and dozens of injuries, against roughly 17 bear casualties. Both are tiny next to combat: wildlife is 0.2–0.5% of US deaths.
- Wildlife costs strength more than it kills. Bear watch in the north and summer insects in the boreal belt take a few percent off occupier effectiveness. That pressure is small but constant.
- Polar bears matter as a story, not as casualties. In the Arctic seizure, where combat deaths are ~10, a couple of incidents of US troops killing a polar bear move the chance the US quits from 28% to 31–34%. That swing comes entirely from an uncertain narrative parameter. Like blackout blame, it belongs in the narrative module (0.1.21b), not here.
- Outcomes elsewhere don't move. In full annexation and the Alberta detachment the changes are within noise.
Weakest parameters
Everything except the collision base rates is a modeller estimate: bear_encounter_rate, bear_casualty_share, polar_bear_guard_overhead, insect_penalty and the iconic-incident pair. The bear-attack baseline comes from low-confidence secondary compilations.
Not modelled (by design or for later)
- Bird strikes on aircraft.
- Wildlife as surveillance clutter for thermal sensors and drones.
- Ticks and Lyme disease.
- Subsistence hunting as a resistance food source.
- Wildlife effects on Canadian forces and civilians.
Sources
Transport Canada via Canadian Underwriter (2011); BC Wildlife Collision Prevention Program; Traffic Injury Research Foundation via eDriving (2024); Samsara (April 2026); Wilder et al. (2017), Wildlife Society Bulletin 41(3):537–547 (USGS); secondary bear-attack compilations (low confidence).