Why Canadian underwriters aren’t ready for wildfire’s true threat

Wildfire is the only major Canadian peril still being underwritten with the broad strokes the industry abandoned for flood years ago. Too often, a property is either “in a wildfire postal code” or it isn’t, with models relying on fuels data that may be a decade old and resolution too coarse to distinguish maintained defensible space from mature, untreated forest. That gap is showing up in loss experience, capacity decisions, and the growing number of properties carriers can’t price with confidence.

“Wildfire has caught up to flood and earthquake as a primary loss driver in Canada,” says Brian Bastian, Head of Product at Property Guardian. “But the tooling hasn’t. We’re asking underwriters to make major capacity decisions on land-hazard data that wasn’t built for insurance and isn’t refreshed often enough to reflect what’s on the ground.”

The data underwriters have today

Most wildfire scores in Canada share the same constraints: they model whether vegetation will burn, not how fire could affect the structure being underwritten. They rely on fuels data that may be years out of date, and they deliver outputs designed for regional planning rather than individual-property pricing.

That leaves carriers with two blunt options: avoid entire postal codes, leaving good risks on the table, or write the business and absorb the volatility. “Neither of those is underwriting,” Bastian says. “That’s guessing.”

A different approach

After hearing repeated feedback from carriers about the Canadian wildfire data gap, Property Guardian developed its Canadian Wildfire Risk Score over two years specifically for the Canadian P&C market. The score is grounded in Canadian wildfire science from research partner ApexRMS and built on physics-based modeling rather than machine-learning proxies.

The score evaluates six risk drivers at the individual-property level: burn probability, fire intensity, suppression difficulty, urban conflagration, wildfire recurrence, and ember exposure. Each is surfaced explicitly so underwriters can see what is driving the rating and explain it to brokers, reinsurers, and clients.

“What an underwriter needs to know is how a wildfire would affect the structure they’re indemnifying, not just whether surrounding vegetation could ignite,” says Paul Brady, Head of Wildfire Insights & Mitigation at Property Guardian. “Embers travel kilometres. A serviced subdivision can still be exposed to a fire that started in the wildland-urban interface. That nuance has been invisible in most Canadian models.”

Brady, a former wildland firefighter, leads Property Guardian’s wildfire insights and mitigation team, which includes Certified Wildfire Mitigation Specialists, meteorologists, data scientists, and property insurance experts. That mix matters, he says, because the model has to reflect both fire behaviour and how underwriters actually evaluate risk.

“You can’t model fire behaviour well if you haven’t stood in front of it, and you can’t build something underwriters will use unless you understand how they evaluate risk,” Brady says. “That is why we built the model with both perspectives.”

What it means in practice

For Canadian underwriters, that creates three immediate opportunities:

• Move beyond postal-code exclusions where property-level risk does not warrant them. Better data lets carriers separate good risks from genuinely uninsurable ones inside the same geography.
• Treat fuels, recurrence, ember exposure, and suppression difficulty as visible inputs, not a single opaque score.
• Refresh the data behind appetite and pricing as wildfire conditions change.

“This is a peril that rewards precision,” Bastian says. “Canadian underwriters deserve a tool built for Canadian conditions, by people who understand both the science and the underwriting problem. That’s what we’ve brought to market.”

The Canadian Wildfire Risk Score is part of a suite of wildfire solutions from Property Guardian. It is available via property lookup, API integration, and bulk portfolio analysis. To learn more or request a demo, contact John Dunn, Head of Enterprise Sales, at john.dunn@greenshieldrisk.com.