"A Coupled Atmosphere-Fire Model:
Convective Feedback on Fire-Line Dynamics"
written by NCAR scientists Terry Clark and Janice Coen with Mary Ann
Jenkins (York University, Canada), and David Packham
(Monash University, Australia).

Above: Fingers of flame about a kilometer apart
are evident in this photo of the Owens Valley, California,
sagebrush fire of July 1987.
(Photo by Charles George,
courtesy International Fire Science Laboratory.)
This summer
proved to be a dangerous one in the tinder-dry forests of the
southwest
United States and Alaska. Winds play a critical role in fire spread,
but a fire itself can modify local winds, helping it grow even more quickly.
Now a senior scientist at the National Center for Atmospheric Research (NCAR) in Boulder, Colorado, has created one of the world's first computer models
that traces the interplay over time between fire behavior and winds,
pointing the way toward future models that might aid in fire prediction
and management. First results from this model were published in the
May 1996 issue of the Journal of Applied Meteorology.
Clark specializes in using supercomputers to model small-scale atmospheric phenomena. His work has analyzed severe thunderstorms, downslope windstorms, and the dynamics near fronts. For the fire-atmosphere study, one of Clark's atmospheric models was coupled, or connected, with a model of dry eucalyptus forest fires (a major threat in Australia). Although forests vary in how they burn, the authors expect that their main findings will translate to a variety of settings.
Here are two animation sequences, each representing twelve minutes of model simulation. The first is an animation sequence of vertical cross section of buoyancy and the second of horizontal cross section of vertical velocity.
Most previous studies on fire and wind have assumed a straightforward relationship between large-scale winds and fire behavior. However, the authors note, "Forest fires are very complex phenomena. . . . Interactions between forest fires and airflow are highly nonlinear [unstable], and radiation and combustion properties are not fully understood." Using the coupled model, the scientists were able to examine a variety of wind speeds and observe--at resolutions as fine as 20 meters--how a fire's development can alter the circulation around it. Among their findings:
A fire's
pattern of growth depends not only on
large-scale winds but on the balance between those winds and a fire's
heat output. If the winds relative to an advancing fire line are weak,
and the heating is particularly strong, a fire can force its own
circulations, possibly resulting in unstable, "blow-up" fire conditions.
(It was a sudden
blow-up that may have killed 14 firefighters near Glenwood Springs,
Colorado, in 1994.) On the other hand, strong winds relative to the
fire line--though literally fanning the flames--tend to produce a more
stable regime in which the fire is less likely to create its own
circulation pattern. Thus, the fire's spread may be more predictable.
Air
temperatures near a fire are lower than
one might normally think. In the first several minutes of a new fire,
the model shows surface temperatures soaring, which creates a chimney-like
plume of rising air. Shortly thereafter, the atmosphere establishes a
balance between the updraft (blowing at near-hurricane speeds as high as
30 meters per second) and the heat provided by the fire. In the model,
the updraft strengthens and pulls in surrounding cooler air as a fire's
heat output increases. This keeps air temperatures near the fire in the
range of 60 to 100 degrees C, even as the fire itself burns at more than
800 degrees C.
The model
helps to explain a
commonly observed
trait of wind-driven fires: the growth of fingers of flame, spaced about a
kilometer apart, that form the main fire line. Previous researchers had
proposed that the fingering was due to variations in either the fire's fuel
or the local geography. However, the coupled model suggests that,
when winds are weak, a fire line several kilometers or more in length is
inherently unstable and very likely to break up into fingers.
The calculations for the coupled model were performed on NCAR's CRAY Y-MP supercomputer with support from the National Science Foundation (NSF). Clark and his colleagues plan to continue their fire-atmosphere modeling. They are now investigating a second, smaller-scale type of fire fingering that occurs through a process roughly similar to the one that causes supercell thunderstorms to rotate. Preliminary model results show the development of a tornado-like vortex within a fire, much like the vortices sometimes observed in actual fires.
NCAR is operated by the University Corporation for Atmospheric Research under sponsorship of NSF.