Chaos theory helps predict wildfire behavior in highly variable environmental conditions

Wildfires exhibit notoriously unexpected behavior, posing several challenges in predicting their course. While classic approaches employing empirical or multidisciplinary modeling have had some success, they often fall short in emergent situations in which environmental conditions are highly variable.

Sjunneson McDanold et al. have demonstrated a new method for modeling the interplay between atmospheric hydrodynamics and fire behavior to better predict behavior on the margins of wildfires, the fireline. Using outputs from the physics-based FIRETEC model, the group tested more than 2,000 time series based on a variety of parameters. The test, using asymptotic growth rates of shorter time series like those in wildfires, characterized the degree of chaos in front of, within, and behind the leading edge of the fireline.

“This paper is the first step toward broadening our perspective on fire science through nonlinear time-series analysis,” said author Jenna Sjunneson McDanold. “This type of analysis has not been widely used in this field, but it offers a wealth of opportunities to study different aspects of fire behavior that aren’t discernible with alternative methods.”

The group found convective heat transfer and vertical wind velocities were highly chaotic, while the horizontal wind trajectory tended to be relatively stable. The temperature changes in the simulation also exhibited a high level of chaos within the combustion zone but tended to be more stable in front of the fireline. They also found more self-organizing behavior in front of the fireline as the winds were entrained into the heat from the flames.

The group next plans to perform similar analyses across other FIRETEC configurations to confirm the presence of chaos in fire behavior.

Source: “Detecting deterministic chaos in a high-complexity fire model,” by Jenna Sjunneson McDanold, Alex Jonko, Kara Yedinak, Rod Linn, Sophie Bonner, Alexander Josephson, and Nishant Malik, Chaos (2026). The article can be accessed at https://doi.org/10.1063/5.0306306 .