Methodology & Data Sources — Canada 🍁
Canadian FireRisk reports are built entirely on openly-licensed Natural Resources Canada data. Here is exactly what we use, how current it is, how precise it is, and every parameter of the formula — so any expert can reproduce your score.
Methodology ca-v1.0 (2026-08) · last updated August 2026
See the U.S. methodology (shared curves) →How the 0–100 score is calculated
Model ca-v1.0 (2026-08)Your Canadian FireRisk is a transparent weighted blend of four independent federal datasets, each pulled live for your exact coordinates. It uses the same 0–100 scale and the same five tier bands as our U.S. reports, so a High in British Columbia means the same thing as a High in Colorado. Government danger maps are labeled inputs — never re-badged as our score.
Distance-decayed proximity, size & recency of federally recorded fires since 2000 — plus a distance-zero entry if the point sits inside a mapped burn perimeter (point)
Today's official fire-danger class at the point, updated daily — the Canadian score includes current conditions (point)
Nearest live fire within 40 km: distance, stage of control, size and count (point, near-real-time)
Fire spreads up to 8× faster uphill (point)
No silent defaults
If a dataset has no data for your exact point — including no fire-danger polygon over many urban cores — it is dropped from the formula and the remaining weights are renormalized to sum to 100%, never filled with a guessed "Low" or "Moderate." If every input is absent, the report says the score couldn't be computed rather than inventing one.
Confidence shown
Every report shows a confidence level: High (both core hazard signals — fire danger and burn history — answered), Moderate (one), or Limited (neither). A search that resolves only to a city or postal-area centroid is capped at Moderate and flagged, since the point may not be your specific parcel.
Includes today's conditions
Unlike the U.S. score, the Canadian composite blends today's danger class and live fire activity with the durable landscape signals — so Canadian reports are cached for at most one hour. The number can't outlive the conditions it blends.
What we deliberately do not put in the score
Your home's defensible space and structure hardening can't be measured remotely — we never fake a number for them; the guided walkthrough scores them from your own answers and builds a FireSmart-aligned plan. We also exclude datasets we can't honestly use: BC's PSTA and danger-rating layers are "Access Only" licensed (not open), so they are deliberately not inputs; NRCan's legacy cffdrs FWI rasters were frozen stale on our verification probes and its fuels raster was broken — neither is consumed. Every input above was live-probe-verified before it entered the formula.
Every data source, dated and cited
Recorded fire history within 40 km
Whether this exact point sits inside a mapped burn
Today's official fire-danger class at the point
Live fires within 40 km, with stage of control
Slope & elevation
Address → coordinates & province
Contains information licensed under the Open Government Licence — Canada. FireRisk is provided for informational purposes; no score here is insurance-grade or property-inspection-grade, and provincial danger ratings and restrictions should be verified with your provincial wildfire agency.
Full formula detail
Every parameter in every sub-score, exactly as coded, so any expert can reproduce the calculation.
Fire danger today (30% of total)
The CWFIS fdr_current class at your point maps to the midpoint of the same tier band our overall scale assigns that word — transparent and auditable. The class code is 0-indexed (verified against the live legend); an unrecognized code is rejected, never guessed. No polygon at your point (common over urban cores) → the factor is excluded, never assumed "Low."
code 0
Low · Bas
→ 10
code 1
Moderate · Modéré
→ 30
code 2
High · Élevé
→ 50
code 3
Very High · Très élevé
→ 70
code 4
Extreme · Extrême
→ 90
Burn history sub-score (35% of total)
NFDB large-fire records within 40 km (scored since 2000, nearest 14 by distance) run through the exact same continuous curve as the U.S. fire-history score — proximity weight × recency multiplier × 92, plus density and size bumps, capped at 100. Hectares and kilometres convert at the boundary (1 ha = 2.471 acres, 1 mi = 1.609 km), so the identical curve applies; the full anchor tables are published on the U.S. methodology page. Two Canadian specifics: if NBAC shows this exact point inside a mapped burn perimeter (2000+), it enters as a distance-zero fire — the truthful reading, since an NFDB point can sit 10+ km from the perimeter that actually contains the address. And if the NFDB archive is unreachable, the factor is dropped and renormalized — never scored as "no past fires."
Live fire activity sub-score (20% of total)
Scored from the nearest live fire on the CWFIF national feed within 40 km: a continuous distance curve (anchors below, interpolated between), scaled by the fire's official stage of control — Out of Control ×1.00, Being Held ×0.72, Under Control ×0.45, unknown stage ×0.85 (fail toward caution, never above Out of Control). Size adds up to +8 (anchors: 1,000 ha +2 · 10,000 ha +5 · 100,000 ha +8) and multiple live fires add +3 (two) or +6 (three or more), capped at 100. A healthy feed showing nothing nearby scores 2 — a real, provably-quiet reading. A feed outage or truncated pull excludes the factor: absence is unprovable.
100/100
0 km
92/100
1.6 km (1 mi)
84/100
3.2 km (2 mi)
62/100
8 km (5 mi)
40/100
16 km (10 mi)
22/100
24 km (15 mi)
8/100
40 km (25 mi)
Slope & terrain sub-score (15% of total)
Identical to the U.S. slope score: five elevation samples (your point plus four cardinal neighbours ~100 m away) from the Copernicus DEM via Open-Meteo, converted to degrees and mapped through the same continuous curve (0°→8 · 5°→12 · 10°→25 · 20°→48 · 30°→68 · 40°→83 · 60°→96, interpolated between, capped above 60°). If terrain can't be resolved, slope is dropped and the remaining weights renormalize — it is never assumed flat.