B.C. wildfire forecasters rely on outdated 1990s data as fires intensify
The models used to predict fire spread and intensity have not had a major update since 1992, officials say.
Wildfire forecasters in British Columbia are using computer models based on fire behaviour data that has not had a major update since 1992 to predict increasingly volatile blazes.
The forecasting programs are used as soon as a wildfire is found to estimate "rates of spread, fire intensities and other metrics of fire behaviour," yet the models are still fed by observations gathered from fires decades ago.
"The last time it had a major update was 1992," said CBC British Columbia reporter Georgie Smyth in a recent report.
READ MORE: Bristol-led drone swarm wins $500,000 wildfire detection prize
READ MORE: Los Angeles wildfires show growing health threat of wildfire smoke
Smyth said the reason is that "it's no longer safe to go in and collect new metrics" from the intense, life-threatening fires crews now face.
The gap is already showing up in the results. On some of British Columbia's toughest fires, the models "have underperformed," including a blaze that "predicted half the spread distance that was observed that day."
A long stretch of dry weather has made forests in British Columbia's southern Interior especially flammable.
Smyth said the area is "the driest place in Canada" and added that "nowhere else in the country is experiencing as bad a drought."
Smyth said the region is seeing such explosive fires, ones that can totally incinerate forests from the ground to the tops of 30 meter trees.
Some blazes become even more severe. "Some get so intense they create fire clouds," Smyth said.
That means the original database was built for different kinds of wildfire conditions than the ones crews face. As Smyth said, "Those were not the conditions that we collected data for."
In a written report by Smyth on the CBC website, forecasters explained that collecting data from life-threatening fires remains a challenge. Experiential burns are also not feasible amid heavy drought in this part of Canada.
Other ways to gather data include prescribed, or preventative, burns. The CBC also mentioned a collaboration with NASA, where the U.S. agency obtains data by flying through firestorms.
Lastly, the outlet noted artificial intelligence might be a predictive tool in the future.
Neal McLoughlin, superintendent of predictive services with the B.C. Wildfire Service, said real-time observations are still the gold standard.
"If we're not making an effort to collect that data on the ground to inform these models, they're not going to make good predictions," he told the CBC. "It's as simple as that."
Wildfire forecasting is getting harder as fires become more extreme, making the tools even more critical.
With files from The Cool Down and Yahoo News