Use of AI helping fight wildfires by predicting where they might start

The Kiskatinaw River wildfire along the B.C.-Alberta border can be seen from the sky in early June. (Photo: Alberta Wildfire)

Wildfire season has become a common occurrence over the last few years with large-scale fires destroying large swaths of cities – and uprooting lives.

To combat this, scientists are developing new technology using artificial intelligence (AI) to predict where fires will take place beforehand so that they can be prevented.

• The information you need to know, sent directly to you: Download the CTV News App

The rising number of wildfires that have become a part of spring and summer in Alberta over the past decade is pushing several Canadian companies to develop ways to prevent the kinds seen in Fort McMurray in 2016 and in Jasper National Park last year.

“A combination of statistical modeling, probabilistic modeling, physical modeling to capture the fire behaviour,” Gio Roberti of AISIX Solutions Inc. said.

Some use drone mapping, others such as AISIX Solutions Inc., use complicated computer systems.

Millions of simulations measure minute data such as temperature, rain levels, wind patterns and dryness.

Software sifts through the results, creating a map that predicts the likeliest spots where a wildfire could start and spread, Roberti added.

“We are able to capture to understand what the probability of a fire occurring today is and in 10, 20, 50 years, according to different climate change scenarios,” Roberti told CTV News Edmonton.

“There might be cases where mitigation measures might be taken where, for example, preventive burns or fuel thinning and things like that.”

The technology is new, but Roberti says it’s already in use, identifying potential hot spots to municipalities and insurance companies.

“People take preventative measures when they see their community on the map,” he said.

Roberti said the technology could also work beyond the Canadian border.