A team of forest management specialists affiliated with various institutions across Canada has found that fuel aridity has been the most influential driver of burn severity in wildfires in Canada over the past several decades.
In their paper published in the journal Science, the group describes how they collected and analyzed 40 years of spatiotemporal wildfire data for Canada and used the findings to build a multinomial logistic regression model as a means for better understanding the increase in number and severity of wildfires in Canada over the past several decades.
Jianbang Gan, with Texas A&M University, has published a Perspective piece in the same journal, outlining the work done by the team and others working to understand the major drivers of wildfires in other parts of the world.
As the planet grows warmer, wildfires have become more numerous and more severe. They rage harder, grow bigger, and cause more damage. Such wildfires, called forest fires when they burn large swaths of forests, have become more commonplace in parts of Russia, North America, and Australia over the past several decades. Studies into these fires have found that in addition to tree and human structure loss, they are causing more CO2 emissions to be released into the atmosphere as the carbon trapped in trees is set free.
It has also been found that such fires are becoming faster in places like the U.S. where they release embers into the air ahead of rapidly encroaching flames, igniting human structures before emergency responders can intervene, all because of stronger winds. In this new effort, the research team in Canada wanted to know more about the major drivers behind the increase in the number and size of wildfires over the past several decades in their country.
To that end, they collected as much data as they could find surrounding wildfires occurring in Canada over the past 40 years. They then used that data to create what they describe as a multinomial logistic regression model—a type of model that allows for multiple categories of dependent or outcome variables. It does so by making estimates among associations between sets of predictors and multi-category, unordered outcomes.
In looking at what their model showed, the researchers found fuel aridity (dryness of the trees) to be the most influential driver of burn severity. They also found that summer wildfires tended to be more severe and conditions for such fires have grown worse over the past two decades. The researchers conclude by noting that variations in drivers led to different outcomes in different parts of the country.
Journal information: Science
