New AI program to help stave off wildfire devastation
New AI program to help stave off wildfire devastation
August 8, 2017
Metro Toronto
Article by: Kevin Maimann

University of Alberta professor helps develop computational tool that recognizes extreme fire weather patterns

A helicopter battles a wildfire in Fort McMurray in May 2016.

JASON FRANSON / CANADIAN PRESS FILE - A helicopter battles a wildfire in Fort McMurray in May 2016.

A new tool that mimics human thought could save Albertans from disaster when future wildfires break out.
University of Alberta professor Mike Flannigan, in the department of renewable resources, has helped develop a computational model using Artificial Intelligence to better predict conditions that lead to rapid spreading of wildfires.
“When we see a person’s face, we recognize it. Well, this is recognizing patterns that are significant with severe fire weather,” Flannigan said.
“So you see a face that’s smiling, you infer something. If it’s frowning or angry you get that information.”
Forest researchers at the U of A and the University of Oklahoma designed the model, referred to as a “self-organizing map,” to use atmospheric pressure variables and predict where and when extreme fire weather is expected.
The map inputs meteorological data in real time, learns from the data, and updates its maps and predictions, mimicking neurons in a human brain.Flannigan said the technology, which is not yet in use, could serve as an early warning system to potentially get resources in place before a fire causes significant damage.
“If you wanted to get a firefighter in Ontario to a fire in Alberta and you place a call to Ontario, it takes at least two days before that firefighter is on the fire in Alberta,” he said.
“So having advance warning can really help, during those critical times, to say, ‘We need more helicopters, we need more boots on the ground, start making the calls now.’ ”
Alberta is hit hard by wildfires every summer.

About 210,000 hectares burn each year – roughly three times the size of Edmonton – from about 1,500 separate fires.
Last year, the Fort McMurray wildfire forced the evacuation of close to 90,000 people and caused $3.77 billion in insurable losses, making it the costliest natural disaster in Canadian history.
Canada spends up to $1 billion on fire management every year, and those costs are on the rise.
“I’m not saying (the model) would have prevented Fort McMurray. No one can say that,” Flannigan said. “But overall, it will help fire management do a better job and reduce the likelihood of disastrous fires.”
The researchers describe the computational model in a study published Tuesday in the Canadian Journal of Forest Research.
Wildfire stats, according to Alberta Wildfire
208,698: A 10-year average of the number of hectares burned in Alberta each year, from 2006-2015
1,547: A 10-year average of the number of wildfires in Alberta each year
64: The percentage of Alberta wildfires caused by humans in the same time span
36: The percentage caused by lightning