B.C. researchers develop prediction system for human-caused wildfires
UBC researchers used satellite imagery to predict human-caused fire frequency in Alberta test zone
A team of B.C. researchers has developed a system it says can be used to predict the likelihood of human-caused wildfires in boreal forests, allowing fire officials to focus their preparations and prevention efforts.
The system uses satellite imagery to track the growth of new leaves in forest undergrowth during the spring, which gives the researchers an idea of how flammable a given tract of forest is going to be over the rest of the season.
"We essentially just go download that imagery and process it and create a prediction on about a week-by-week basis," said Paul Pickell, a post-doctoral fellow in UBC's Faculty of Forestry.
"We can know by the end of March, before the start of the fire season, essentially when is going to be the most flammable part of the fire season for human-caused fires."
- Increasing wind speeds will stoke more Alberta wildfires
- Interior B.C. needs better air quality monitoring during wildfires: Kamloops doctors
Accurate within 10 days
Pickell said the system came about when the researchers noticed that human-caused wildfires tended to spike just before the season's new leaves appeared in early spring.
"This is a time of the year when there's a lot of snow cover that's going away and that's exposing these dead and living fuels on the forest floor," Pickell said.
"A lot of those fuels are pretty dry, especially the living fuels that don't yet have leaves."
The researchers figured that, if they tracked new leaf growth, they would be able to extrapolate and predict when human-caused fires would be most likely to start.
The team tested the system over a 400,000 square-kilometre area of boreal forest in Alberta. Pickell said the system was able to accurately predict peak times for human-caused fires for about 80 per cent of the test area within a 10-day window.
More testing required
Of course, the system is not without its limitations. For starters, it's only useful for predicting the likelihood of human-caused fires.
In B.C., only about 40 per cent of wildfires are started by humans; the remaining 60 per cent are started by lightning.
Additionally, the system was only tested in Alberta's boreal forest, where the researchers know new leaf growth to be a major factor in the spread of wildfires.
"In British Columbia and other forest types, it's a little trickier, and we're not quite sure yet how the system would play out in other areas," Pickell said.
He said the system could also be made more robust by including information from the current fire danger system, such as day-of weather observations.
With files from CBC Radio One's Daybreak South.