ORNL ESD personnel are developing a computer fire simulation, EMBYR, an Ecological Model for Burning the Yellowstone Region, to investigate the causes and consequences of large-scale fires like those that burned in Yellowstone National Park during 1988. EMBYR is not designed to make predictions about single fire events, but will be used as a ``what-if'' tool to investigate possible landscape-scale effects of variation in fire frequency, fire management, and global climate regimes over time scales ranging from complete fire seasons to millennia. Nevertheless, EMBYR must be calibrated by comparison with single historical events.
Fire moves by 2 mechanisms in EMBYR. Fire moves to the 8 adjacent cells surrounding a burning cell based on the type of fuel burning, the type of fuel in the adjacent cell, current fuel moisture conditions, and the wind speed and direction. Fire will spread to downwind cells more easily than upwind or lateral cells. Topographic effects of slope affect the spread of fire to adjacent cells, fires being more likely to spread to upslope cells and less likely to spread to downhill cells. Firebrands, burning material that is carried aloft in the convection column, can ignite spatially remote cells. The number of firebrands produced by each burning cell is a function of the type of fuel that is burning in that cell. In the absence of wind, firebrands are distributed in random directions, but, at higher wind speeds, firebrands travel downwind. The likelihood that a remote cell will be ignited by a firebrand depends on the type of fuel in that cell and the current fuel moisture conditions.
On the afternoon of September 16, the monitors put out some spots on the south side of the river, but eventually lost control of the spots. Smokejumpers from West Yellowstone and retardant from Cody were dispatched, but were of no value because the dense smoke prohibited flying. Two major spots and a number of small spots, totaling 40 acres, were extinguished on the south side of the Snake River.
A burnout operation was completed along the north side of the Snake River to eliminate the spotting hazard. During this suppression effort, the main fire, under the influence of the daily up-slope, up-canyon winds, was progressing to the southeast, following the north side of the Snake River. The suppression effort took place along approximately 3 miles of the fire's 27 mile perimeter. A small crew also worked at the extreme western end of the fire to keep it from crossing the Lewis River. The fire was finally stopped by snow on September 25, and was declared out, after burning 7,396 acres, on October 13, 1981.
Click here to see a simulation of the Forest Lake Fire (movie - 1.2 Mb). The simulation will burn and then reverse 3 times in succession to allow closer study. The gray color of the burned areas is proportional to burn severity; lighter grays indicate less intense combustion, while darker grays indicate higher burn intensity. The wind and weather are changing during the 45 day course of the simulation. The effect of shifting wind directions on the fire's progress is evident as the simulation runs. Spots disconnected from the main fire line continually appear downwind of the main fire line. Often these spots coalesce with the fire line as it overruns them. Shortly after the fire begins, there is a noticeable reduction in the burn severity, indicated by a change from dark gray to lighter gray. This is caused by the lighter winds and greater fuel moisture during the stationary high-pressure system mentioned in the Forest Lake narrative above. The rate of spread of the fire is also reduced during this time. The wind speeds then pick up again, and burning conditions improve, as reflected in a subsequent return to more severe burns (darker grays).
Comparison of the actual perimeter of the historical Forest Lake fire with the simulation result indicates fair agreement, both in terms of the area and the shape of the burn. This map shows the progression of the real fire as a series of colored perimeters from red through magenta to cyan. Because EMBYR is probabilistic, another stochastic run would produce slightly different results, but probably a similar overall appearance. Early in the simulation, EMBYR burns more area south of the Snake River than the historical perimeter shows. The Forest Lake narrative above describes how early fire spots south of the Snake were extinguished by fire fighters - would the actual perimeter have resembled the simulation even more had these early spots not been successfully suppressed?
ESD scientists are using EMBYR not to predict behavior or deploy resources for single fire events, but rather as a tool to study fire-induced changes in the long-term structure of the Yellowstone landscape. For more information about EMBYR and this research project, contact: