How artificial intelligence can help identify and fight wildfires in Canada

Open this photo in gallery:

After the devastating wildfire season in Canada, scientists are looking to artificial intelligence to play a bigger role in detecting and fighting fires.

The wildfires, which burned from British Columbia to Nova Scotia, temporarily displaced thousands of Canadians across the country, blew smoke as far as Europe and burned through 3.3 million hectares of forest — about half the size of New Brunswick. All this despite the fact that the federal government spends nearly $1-billion a year fighting the fires.

Global experts have warned in recent years that warming climate change will lead to longer wildfire seasons and longer exposure to smog in Canada. In the year In February 2022, the United Nations predicted that wildfires would be the most devastating of the year and called reducing the risk of wildfires “more pressing than ever.”

These global alarm bells are leading experts to predict that Canada will make greater use of AI technologies in drones, sensors and high-tech satellites to fight the fires.

“It’s changing in terms of climate and technology,” said Joshua Johnson, forest fire research scientist at Natural Resources Canada (NRC). “I think that’s scary at times, but it forces us to be creative.”

Canada has seen 100,000 square kilometers burned by this wildfire.

Comment: Canada’s wildfire approach should change from reactive to proactive.

The current AI revolution, fueled by the advent of powerful data processing machines like ChatGPT, has revolutionized many industries, including computing, customer service, and more in recent months. American computing software called RADRFIRE has enabled the creation of software that processes imagery from drones and satellites nationwide to detect, predict, assess early fires and communicate with fire stations.

The same technology, according to Mr. Johnson, will ultimately help Canada make critical decisions about how to allocate resources during a major wildfire.

“If we have 100 fires in the country and 30 crews, we need to quickly decide where to focus our efforts – especially now that we have fires from coast to coast,” he said. “AI will be a powerful tool to help us make those calls.”

Mr Johnson said the branch of government had recently been supported by a number of start-ups, all offering AI-powered solutions to wildfires. Sensiotech, a Brazilian company also incorporated in Canada, uses localized ground-based sensors that monitor 14 variables — including humidity, soil temperature, salinity and pH — to get a minute-by-minute reading of the potential for local ignition.

Sensio Tech founder Joao Lopes says beta tests in Brazil suggest his invention can predict wildfires five days before they occur.

“With satellites, you’re seeing the current; you can’t necessarily predict where the fires will burn in the future,” he said, adding that he expects forecasting technology could eventually predict fires months in advance, giving more preparation time. Mr. Lopes hopes Canadian investors will embrace the technology, despite Canada’s size. Despite the inherent challenge, each sensor covers five hectares of forest — a tiny fraction of the country’s 362 million hectares, most of which are north of the power grid and far beyond connectivity.

Mr. Lopes isn’t the only one using AI to create firefighting technology in recent years. The World Economic Forum developed FireAid: A Live Hazard Map to update fire data in real time and better allocate resources. Similarly, US startup Pano AI uses computer vision to quickly and accurately identify new wildfires in five western states and parts of Australia.

Other companies focus on reducing unforced errors. Booth Solutions, a California startup, uses machine learning to continuously scan power-grid infrastructure for faults — such as insulation damage and overheating — to identify fire hazards. In the year In 2021, the company tested the technology in Newfoundland and reported quickly repairing power lines after a hurricane.

“Climate change, the age of the grid, and as our cities deforest, we’ve seen a greater appetite for our products,” said Caitlin Albertoli, CEO of Booth Solutions.

Mike Flanigan, British Columbia’s research chair in predictive services, emergency management and fire science, is wary of thinking about machine learning and new technologies as the only answer to the country’s wildfire fight. An easy and quick way to affect change is by changing our own behavior, because half of all fires in the country are human-caused. For example, implementing more forest closures – banning activities in the forest in an area prone to fire – would be a good start.

“We need to look at AI and machine learning as another tool in the toolkit, not a panacea to solve all our problems,” Professor Flanigan said. “We’re not there yet; you’d need thousands of drones to replicate a single water bomb attack.”

However, he is bullish on AI tools that can better predict the location of intense fire weather and cut the time it takes to respond to it. Current forecasting and modeling technology, Professor Flannigan said, could delay firefighting efforts by up to three days, giving the fire enough time to spread quickly.

Mr. Johnson sees the promise of AI-driven sensor technologies, as they could be particularly useful for monitoring tourist areas like Banff and Jasper around the clock. The NRC said it hopes to strengthen firefighting technology in the coming years and be prepared if fires threaten those high-traffic areas.

For example, the Canadian government plans to launch RADRFIRE in 2029: WildfireSat, a $170 million satellite designed to deliver vital information twice daily to fire stations and use thermal imaging to facilitate decision-making. But until then, Mr. Johnson said, he saw the value of entertaining new technologies that could reduce the damage of a fire from anywhere.

“I firmly believe we’ve taken some good steps in the right direction over the last 10 years,” he said. “Going forward, there will be times when we feel like the climate is moving faster than we are, but I think we’re on the right track.”