Canada's most advanced wildfire risk dataset: physically-based burn probability, intensity, and footprint data built on 30M+ simulated fires, helping insurers, governments, and industry assess and price wildfire risk with confidence.
resolution, down to 1 m in the wildland-urban interface
SSP1 / 2 / 5
climate pathways, bi-decadal to 2100
WHY WILDFIRE 3.0
Industry-leading wildfire intelligence for comprehensive risk assessment
Five reasons risk teams choose a physically-based model over a statistical curve-fit.
01
Physically-based fire science
Built on the Canadian Forest Fire Behaviour Prediction System, not a statistical curve-fit to past fire seasons, so results stay stable under changing climatic conditions.
02
30M+ simulated fires across 15,000+ fire-seasons
A Monte Carlo framework that captures the full range of plausible wildfire activity, not a single deterministic forecast.
03
250 m resolution, down to 1 m in the WUI
Granular wildfire insight for any location, urban, rural, or industrial, with resolution down to 1 m in the wildland-urban interface. Built-in exposure and vulnerability modeling estimates building-level loss.
04
Climate-adjusted projections to 2100
Historical baseline plus SSP1, SSP2, and SSP5 climate scenarios across four bi-decadal intervals, built for adaptation and long-range risk planning.
05
Seamless data integration
Available via API and data exports for underwriting, planning, and enterprise risk-modeling workflows.
See how Wildfire 3.0 fits your underwriting, planning, or portfolio workflow.
Standard cellular-automata fire models systematically underestimate the size and shape of fast-moving, high-wind fires. Our proprietary Cell2Fire2 engine corrects this bias without sacrificing the simulation speed needed to run tens of millions of fires.
Corrects the arrow-shaped fire fronts and directional bias found in standard grid-based fire models
Approaches the accuracy of legacy vector-based engines at a fraction of the computational cost
Purpose-built to scale to national, Monte Carlo-based simulation
Simulated fire footprints on a flat C1 fuel grid under uniform weather, with varying wind direction.Canopy height around a building footprint (red outline), used to define defensible space.
WILDLAND-URBAN INTERFACE
Defensible space, defined at 1 m
In the wildland-urban interface, Wildfire 3.0 works down to 1 m, showing how close fuel sits to each building. Defensible space is derived from ultra-high-resolution canopy height data, or you can supply measured values for individual buildings.
High-resolution wildfire risk mapping
Risk assessments at 250 m resolution, down to 1 m in the wildland-urban interface, delivering:
Burn probability and fire intensity metrics for any location
Historical fire occurrence dating back to the 1970s, combined with a 2010–2020 climate baseline and climate-driven future projections
Granular wildfire footprints for urban, rural, and industrial zones
Hazard, loss, and live-fire data from one consistent model
Hazard and loss products are available for the historical baseline and for each climate pathway where applicable. Daily products reflect current conditions.
HAZARD & RISK SCORES
Local burn probability and risk scores
Burn probability and risk scores for individual locations
Regional burn probability and risk scores
Burn probability and risk scores aggregated by region
Burn intensity
Fire intensity metrics for any location
Fire footprints
Granular wildfire footprints for urban, rural, and industrial zones
Historical fire information
Historical fire occurrence since the 1970s
LOSS & REINSURANCE METRICSRolling out
AAL and PML
Average Annual Loss and Probable Maximum Loss for portfolio pricing and reinsurance
Ground-up and gross loss
Loss reported on a ground-up and a gross basis
Net lossComing soon
Loss reported on a net basis
Occurrence exceedance probability
Probability that the largest single-event loss in a year exceeds a given amount
Aggregate exceedance probability
Probability that total losses in a year exceed a given amount
Event loss table and year loss table
Event-level and year-level losses for your own analysis
NEAR REAL TIME
Daily fire danger and 24h hotspots
Current fire danger and hotspots detected in the last 24 hours
Fire alerts
Daily fire alerts for the locations you care about
MODEL VALIDATION
How we know Wildfire 3.0 is right
We keep calibration and independent validation separate, because they answer different questions.
CALIBRATION
Ecozone-level calibration
Fire Weather Index and Spread Event Day parameters are tuned against 11 years (2010–2020) of satellite-observed burn area from Canada's National Burned Area Composite (NBAC).
INDEPENDENT VALIDATION
The 2025 wildfire season
In the 2025 Canadian wildfire season, the locations Wildfire 3.0 ranked as highest-risk were among those subsequently affected by fire.
INTERPRETATION
Built for the long run, not one season
Wildfire is a stochastic process. We recommend evaluating model performance across multiple seasons rather than any single year's fire activity.
A yearly-updated historical layer incorporating the most recent wildfire and climate data
Higher-resolution fuel maps derived from recent remote sensing
Improved local-scale ignition modeling to reduce simulation-boundary artifacts
Expanded vulnerability factors, including FireSmart-based risk characteristics
Smoke and secondary-peril modeling (evacuation, business interruption)
CLIENT TESTIMONIAL
“This collaboration with AISIX enables us to assist our clients with wildfire risk intelligence that supports asset-level exposure analysis and scenario-based planning.”
