Artificial intelligence could be used to better predict extreme wildfire weather in northern Alberta: study

Click to play video: 'Using artificial intelligence to predict wildfires' Using artificial intelligence to predict wildfires
WATCH ABOVE: A University of Alberta researcher has found a way to use artificial intelligence to help predict wildfires. Fletcher Kent has more on the technology and when this research will be tested – Aug 8, 2017

A new computer model inspired by the human brain could be used to predict the occurrence of extreme wildfire weather in northern Alberta, according to a new study.

The research, a joint effort between forest science researchers from the University of Alberta and University of Oklahoma, was published in the Canadian Journal of Forest Research on Tuesday.

Using atmospheric pressure variables known to affect weather conditions, the scientists built and tested a computer model referred to as a “self-organized map” (SOM).

Real-time meteorological data collected throughout Alberta’s wildfire season – from May to August – can be inputted into the SOM model, which then learns from the data to generate maps that predict where and when extreme weather is expected.

READ MORE: How the Fort McMurray wildfire created its own weather

SOMs are similar to the human brain, in that they are trained to find patterns in data and model complex relationships, according to the study.

Story continues below advertisement

The maps are a type of machine-learning system – an application of artificial intelligence – that can learn by themselves and not by direct programming.

The researchers believe the artificial intelligence could help fire management agencies be better prepared for extreme conditions and be used as part of an early-warning system.

“Most of the impacts from wildland fire happen during a short period of extreme fire weather that is hot, dry and windy,” said Mike Flannigan, co-author of the study and professor at the U of A.

“You need an early warning. If the firefighters are coming from the United States or Mexico or Australia, you need that week because if someone comes from Australia, it generally takes a week from the time they get the call to the time they have boots on the ground.”

READ MORE: Final Fort McMurray wildfire report indicates misunderstanding over seriousness of threat

Wildfires burn an average of two million hectares per year in Canada, most of which can be attributed to only a few days of severe fire weather, the research reads. These “spread days” are often associated with large-scale weather systems.

A SOM may provide a better and more robust method to flag potentially extreme fire-weather events, the study said.

READ MORE: Report into Fort McMurray wildfire cites communications breakdown in early days

However, Flannigan is quick to point out that this is research.

Story continues below advertisement

“To get into daily operational use it takes a bit of time and effort, so this is one of many tools,” he said.

“This isn’t a panacea that’s going to make all of the problems go away. We’ll always have fire on the landscape because we always have fuel, ignition and at times hot, dry, windy weather.”

READ MORE: B.C. In the midst of worst wildfire season since 1958

Flannigan is hopeful they can work with fire management agencies across Canada in order to provide “one more tool in their toolkit” to help better manage fire.

SOMs have been used in the past to predict other weather events, such as monsoons. However, this is the first study to use the technique for predicting extreme fire weather in real time.

Want your weather on the go? Download the Global News Skytracker weather app for Android and IOS.

Sponsored content