Wildfire risks around the world are likely to change dramatically during the latter half of this century, with some types of terrain projected to see dramatic increases in likelihood and in the expanse of fire-prone areas during the next 30 years, according to a new study.
The areas at highest risk for increased fires in the short term include landscapes familiar to residents of the western United States – Mediterranean forests, woods, and scrub, such as those that blanket the coastal ranges of California, alpine grass and shrublands, desert shrublands, and forests dominated by conifers.
But over the longer term, significant increases in fire frequency are expected over a much wider range of ecosystems – as are significant decreases.
Indeed, decreases in fire frequency often take a back seat to efforts to project places where fires are likely to increase. Yet the declines can be just as important, suggests Max Moritz, a fire ecologist at the University of California, Berkeley and the lead author of the study, which was published this month in the journal Ecosphere.
"You're worried about the right kind of fire being maintained or restored in these systems," he says. "Big decrease in fires … could be just as disruptive ecologically as a big increase."
For instance, the decreases could allow vegetation types alien to the existing landscape to move in and alter the region's ecology in unpredictable ways. Or fewer fires would allow fuels to build up to levels that allow the fires that finally occur to reach catastrophic levels.
One type of ecosystem projected for significant decreases by century's end are subtropical savannas in Africa, Australia, and elsewhere. These are projected to see fewer fires than they currently experience, the research suggests. This would lead to major ecological changes there, since frequent fires keep savannas savanna-like.
Previous studies largely have produced projections of changing fire risks for individual regions of the globe, such as the western US. This new study is a rare attempt at trying to gauge future changes to wildfire patterns globally as the climate warms, the research team says. The warming has been triggered and sustained by rising carbon dioxide concentrations in the atmosphere, increases researchers have traced to burning fossil fuels as well as by land-use changes.
The study applies to fires once the province of ecologists trying to gauge the effect of climate change on the future distribution of plants and animals, Dr. Moritz explains. It treats fire as an entity that has its own form of ecological requirements: fuel, fire-favorable climate patterns, and ready sources of ignition.
The researchers first gathered global wildfire data gathered by two satellites and spanning 11 years – from 1997 to 2007. They used the information to estimate wildfire probabilities for each of 14 different broad vegetations types, or biomes, around the world under climate conditions that existed at the time.
They then turned to 16 climate models to develop projections of the climate's likely trajectory. They applied the results from each climate model to their fire model separately to see how much convergence they might or might not show in the final results.
They also used an existing model for plant production as a way to see how the abundance of fuel for these biomes could change in a warming climate.
As for ignition sources, the third piece of a fire's "ecosystem," humans supply the latter with enough regularity to suggest ignition sources won't be a limiting factor for fires.
From these steps they estimated changes to fire risks for each of the biomes for two periods: a 30-year span starting in 2010, and another starting in 2070.
After crunching the numbers, the team found that for the period 2010 to 2039, model runs most consistently pointed to an increase in fire risk for biomes already in regions with warm climates. A handful of biomes in the tropics and subtropics showed small decreases in fire risks. But for this period fewer than half of the models agreed on the direction of change across the 14 biomes.
The team notes that this points to major uncertainties in the results. But it also falls in line with a feature of global warming many scientists have highlighted – its effects become more pronounced toward the latter half of the century than they are today. For many of the more-localized effects of climate change – from severe storms to heat waves to wildfire patterns – the climate's natural variability can still make it hard to tease out global warming's fingerprints.
But the team's model runs also indicate significant increases and decreases in risk during the final 30 years of the century. Fire risks increase for up to 62 percent of biomes in mid and high altitudes, while the risk drops by 20 percent, largely in the tropics. And the models are in much greater agreement.
During the 2010 to 2039 period, for instance, boreal forests, which cover a broad swath of Canada and northern Russia, stood about an equal chance of seeing increases or decreases in wildfire risk over relatively small expanses of the territory they cover. Most of the model results, however, in effect shrugged their digital shoulders.
During the 2070 to 2099 period, however, the vast majority of models agree that boreal forests will see an increased risk of wildfire over nearly all of its extent. The likelihood of increased risk runs from 66.7 percent for most of the boreal zone to 90 percent for a small but still significant fraction.
The results held a couple of surprises for the team, Moritz says.
"One is that in the near-term future, despite the fact that the global climate models have huge variations amongst them, about half the planet shows model agreement for disruptions" in fire regimes, he says. Those disruptions include decreases as well as increases in risk.
"As we crank the models forward, we saw much stronger agreement" and the patterns of increased and decreased risk, as well as consistency among models, solidifies, he says.
The other is projected decreases in fire risk for much of the tropics during the latter half of the century. "It's pretty striking," he says.
One potentially important drawback to the team’s approach – it doesn't capture year-to-year climate variability or features such as the El Nino-La Nina cycle – which plays a role in fire risks in some tropical rain forests.
Indeed, in most regions of the world, "fires are determined by more-nuanced aspects of climate variability, the sorts of climate and weather conditions that vary from between years or seasons," writes Cathy Whitlock, a researcher who focuses on climate and wildfire response at Montana State University in Bozeman.
Mortiz acknowledges the point, but adds that at this stage, global climate models do a poor job reproducing the cycles. The approach the team took – basing its analysis on a pair of 30-year intervals and the climate "normals" each would represent – smooths out those year-to-year changes.
Still, he says, if one could include such patterns, a future direction for the work, the results for the tropics might be different.
Another feature the approach lacks at this stage is the ability to feed in changes global warming could bring to the amount of landscape each of the 14 biomes covers as time progresses.
Despite the approach's limitations, it provides "insight into the complex fire responses that lie ahead, especially in the near term," says Dr. Whitlock.
And while the results don't attempt to predict changes to the expanse and distribution of the 14 biomes included in the study, the results could be set beside those of models that do predict those biome shifts to get a sense for the shift in fire risks those altered landscapes could face, she agrees.