Predicting Fire Emissions: Fire Behavior

Fire behavior is the manner in which fire reacts to the fuels available for burning (DeBano and others 1998) and is dependent upon the type, condition, and arrangement of fuels, local weather conditions, topography, and in the case of prescribed fire, ignition pattern and rate (fig. 4-3). Important aspects of fire behavior include:

• Fire intensity (rate of energy release per unit area or unit length of fire perimeter, generally during the flaming combustion period).
• Rate of spread (rate of advancement of flaming front, length per unit time), crowning potential (involvement of tree and shrub foliage and spread within the canopy), smoldering potential (smoldering combustion of fuels that have been preheated or dried during the flaming stage).
• Residual smoldering potential (propagation of a smoldering combustion front within porous fuels such as rotten logs or duff, independent of preheating or drying).
• Residence time in the flaming, smoldering, and residual stages of combustion.

These aspects influence combustion efficiency of consuming biomass, as well as the resulting pollutant chemistry and emission factor (fig. 4-4).

The Emissions Production Model (EPM) (Sandberg 2000; Sandberg and Peterson 1984) and FARSITE (Finney 1998) take into account fire behavior and ignition pattern to estimate emission production rates. Fire behavior during the flaming stage of combustion in surface woody fuels and some shrub vegetation is effectively predicted within models such as BEHAVE (Andrews and Bevins 1999) and its spatial application, FARSITE (Finney 1998). However, EPM and other applications do not consider fire intensity or other fire behavior attributes when estimating emissions from flames, and that may result in a reasonable approximation for criteria pollutants but also be a limitation to the estimate of hazardous air pollutants or trace gases. BURNUP (Albini and Reinhardt 1997), FARSITE (Finney 1998), and EPM v2.0 (Sandberg 2002) attempt to model the extent and duration of flaming and smoldering combustion in downed woody fuels and duff. Current capability to model residual combustion, combustion in rotten logs and duff, and fire behavior in the foliage canopies of trees and some shrubs remains inadequate to predict emission rates with any reasonable degree of accuracy.

The Los Alamos and Lawrence Livermore national laboratories offer an approach to predicting fire behavior, plume trajectory, and dispersion, by combining a fire physics model, FIRETEC, with a dynamic atmosphere model, HIGRAD, to produce a highly detailed numerical simulation of fire spread and atmospheric turbulence (Bradley and others 2000). The approach builds on prior experience in predicting the dispersion of hazardous air pollutants from fires such as burning oil fields or "nuclear winter" scenarios. This modeling approach is limited to the propagating front but is unique in its coupling of atmospheric and fire physics.