Smoldering combustion plays a dangerous role in many prescribed burns and wildfires. Smoldering burns can last much longer (i.e., days to weeks) than flaming combustion and provide relatively more smoke and other pollution. For example, Figure 1 shows the smoke being released during a prescribed burn in Silver Lake, OR. An increased ability to predict the spread rate and emission release rate of smoldering burns for a wide range of fuels and conditions is needed to help reduce the detrimental impacts of smoldering burns on humans (e.g., smoke in Figure 1) and the environment. Unfortunately, the efforts to understand and predict smoldering behavior has been limited to just a few fuels (e.g., peat) and conditions. The overall goal of this effort was to obtain the scientific understanding needed to develop tools for assessing risks of ignition, spread, and emissions from smoldering combustion for fuels and conditions applicable to the Department of Defense.

Figure 1: Illustration of negative impact of smoldering combustion: smoke emissions from burn. The burn was part of a prescribed burn by the Nature Conservancy of Oregon near Silver Lake, OR.

Technical Approach

To accomplish the overall goal, this work focused on identifying physical and chemical processes that control smoldering combustion. A series of laboratory studies were conducted where the chemical and physical composition of the fuel beds was systematically changed and the impact on the smoldering behavior quantified. Corresponding calculations were performed using a detailed physics model to obtain additional insights into the key physical processes. Smoldering burns were also conducted in the field and by burning field samples in the laboratory. Results from these studies were used to help characterize realistic smoldering behavior expected in nature. Finally, emission samples were collected from burns and correlated to smoldering behavior.


Notable advances were made with respect to the influence of chemical composition and physical structure of the fuel bed on smoldering behavior. When the density is fixed, propagation velocities increase with decreasing cellulose (and increasing hemicellulose) content. This sensitivity is caused by the earlier pyrolysis and higher temperatures that occur when hemicellulose is present in the fuel. When the fuel is loosely packed and the density is not controlled, horizontal propagation velocities increase with increasing cellulose content. This sensitivity is due to the lower densities with higher cellulose contents. Fuels with higher lignin content smolder at slower horizontal and vertical propagation speeds. This behavior is caused by slower pyrolysis and higher activation energy of lignin. Once the fuel is composed of 20-30% lignin, lignin is the dominant chemical component in influencing smoldering propagation behavior due to its slower reaction rates. 

Burning of surrogate fuels composed of lignin, hemicellulose, and cellulose can reasonably match horizontal propagation velocities of finely ground Douglas-fir and wheat straw. However, downward propagation velocities are only similar if the inorganic content is minimal (among potentially other factors). The differences in downward smoldering may be due to differences in ash layer development. 

The physical expansion of fuels caused by moisture causes the propagation speed of smoldering to increase, due to the reduction in density. If the fuel does not expand with the addition of water (i.e., water fills the fuel pores), the propagation speed drops primarily due to increase in wet bulk density. In both cases, additional moisture content slightly reduces the mean peak temperature. 

Smaller particles are more prone to self-sustained smoldering and have higher horizontal and downward smoldering spread rates in porous beds with homogenous particle sizes. Limits in self-sustained smoldering based on particles being sufficiently small are not evident within the ranges evaluated in this study (i.e., diameter > 0.5 mm). 

Physical properties of duff and peat soils, in particular moisture content, packing density, and mineral content, have been identified as important factors affecting the ignition and propagation of smoldering combustion. 

It was found that the emissions factor for methane (CH4), non-methane hydrocarbons (NMHC), and fine particulate matter (PM2.5) increased with increasing surface temperature and spread rate. However, with respect to emission fluxes (EM) the only significant relationship observed was a muted positive correlation of EMCH4 and EMNMHC with spread rate.


At least four major benefits are anticipated as a result of this project. First, this study provides strong evidence that fire managers may be overestimating PM2.5 emissions when smoldering duff is expected. If true, fire managers could be unnecessarily limiting much needed use of prescribed fire because of emissions concerns, and the emission factors quantified in this study should enable more accurate prediction of potential emissions from smoldering duff. The second impact of this study is that a new methodology and conceptual framework for considering smoldering behavior of natural fuels has been established with the potential to be transformative. Specifically, this work provides a model for considering smoldering of general real fuels has been developed and made available. The model can be used by the community to examine smoldering of particular fuels based on composition, in contrast to using a model validated for a specific fuel type. A third major benefit of this work was pioneering the use of surrogate fuels to emulate key characteristics of natural fuels. Prior research has considered surrogate fuels for other complicated fuels (e.g., jet fuels, diesel); however, the use of surrogate fuels to understand and predict smoldering combustion has not been widely considered. An advantage of using surrogate fuels is that calculations are more tractable and can arguably better identify key physical and chemical processes. Equally important, the use of surrogate fuel has the potential to allow smoldering behavior to be predicted a priori if the constituents and properties of the fuel bed are known. Lastly, a major implication of this project is further identifying the significant role of heat transfer in controlling/influencing smoldering behavior through porous materials. It is expected that this new fundamental insight will be used by the smoldering community to greatly enhance how we explain and understand smoldering.