Fires Are Escaping Our Ability to Predict Their Behavior

Like a bad disco dancer, fires move with an unusual amount of unpredictability. While computers can model wind, land a rover on Mars, and predict the size of T-rex, the behavior of wild fires is hard to forecast. And these fires are only getting more unpredictable.

According to The Atlantic:

Since the 1970s, modeling programs such as Farsite, FlamMap, and FSPro have become an essential part of fighting wildfires. The models, which are calibrated against how past fires have typically progressed, consider vegetation type; to­pog­raphy (flames prefer to travel uphill); a fire’s perimeter; and air temperature, wind, and humidity. They then predict where a fire will go, and when.

But they don’t make fires like they used to. The average size of a wildfire has tripled since 1980, and the fires are burning totally different kinds of land. Fire fighting has kept forests continuous, which means a new fire can sweep across without finding and gaps or breaks in its fuel source. Beetles have decimated trees in some areas, making them brittle and dry—perfect for fires. These things totally wreck the standard model. Fires move uphill instead of downhill. They burn land in unexpected ways. “Timber stands that models say will burn slowly erupt as if doused with kerosene,” writes The Atlantic.

Here, for example, is how one model would predict a wildfire to jump:

But in these new fires, that break in the burn pattern might never happen. NASA has tracked data on wildfires for the past 10 years, but the future will probably look quite different, in ways we can’t really predict.

More from Smithsonian.com:

Climate Change Means More Wildfires in the West
The Legacy of America’s Largest Forest Fire

Rose Eveleth | | READ MORE

Rose Eveleth was a writer for Smart News and a producer/designer/ science writer/ animator based in Brooklyn. Her work has appeared in the New York Times, Scientific American, Story Collider, TED-Ed and OnEarth.