Elsevier

Forest Ecology and Management

Volume 424, 15 September 2018, Pages 236-245
Forest Ecology and Management

Effect of woody debris on the rate of spread of surface fires in forest fuels in a combustion wind tunnel

https://doi.org/10.1016/j.foreco.2018.04.039Get rights and content

Highlights

  • Influence of woody debris (WD, >6–50 mm) not consistent in forest fire spread models.

  • The effect of WD on fires in forest litter was studied in a combustion wind tunnel.

  • Three amounts of WD (0.2, 0.6 and 1.2 kg/m2) and a control were investigated.

  • Wind was fixed at one speed and the range of fuel moisture content restricted.

  • The presence of any WD halved the rate of spread of heading fires relative to no WD.

Abstract

The treatment of the contribution of woody debris (WD, such as branches or small logs >6–50 mm diameter) to the rate of forward spread of a fire in current operational forest fire spread models is inconsistent. Some models do not take into account this fuel at all (i.e. only consider the combustion of fine fuels (6 mm diameter)), while others incorporate effects based on little or no data. An experimental programme utilising a large combustion wind tunnel investigated the effect of WD on the spread of fires burning through forest litter (surface fuel) beds of 1.0 kg m-2. Fires spreading with (heading) and against (backing) the wind were investigated. Three treatments of WD load (0.2, 0.6 and 1.2 kg m-2) and a control (0 kg m-2) were studied using a single constant wind speed (1.0 m s-1) and a narrow range of fine and woody fuel moisture contents (10.0–12.7% and 9.2–11.6% oven-dry weight, respectively) determined by ambient conditions. Presence of WD was found to approximately halve the overall rate of spread of heading fires relative to when no WD was present, regardless of the level of treatment. No effect of WD on rate of spread was found for backing fires. Potential explanations of these findings and implications for the use of operational forest fire spread models are explored, as are future research needs.

Introduction

A critical requirement in the operational management of a wildland fire (i.e. a forest or grassland fire) is knowledge of its rate of spread. Knowledge of the fire’s speed is essential for effective decisions related to fire suppression planning and execution, firefighter safety, the assessment of the threat to life and property, and for the issuance of appropriate warnings (Sullivan et al., 2012, Cruz et al., 2015). Despite the extensive research that has been, and continues to be, devoted to this topic (e.g. Anderson, 1964, Byram, 1959 Curry and Fons, 1938, Rossa, 2017 Plucinski et al., 2017b), complete fundamental understanding of fire spread is still lacking (Finney et al., 2013). This situation stems from the complexity of the interactions and ranges of spatial and temporal scales involved in wildland fire (Sullivan, 2017a, Sullivan, 2017b) and the fact that many physical mechanisms underpinning fire spread are still contested (e.g. Sullivan and Cruz, 2015). Empirical and quasi-empirical models (Sullivan, 2009a) derived from small and large scale experiments (e.g. Rothermel, 1972, Forestry Canada Fire Danger Group, 1992 Cheney et al., 1998) form the basis of all current operational systems due to their overall robustness and reliability (Sullivan, 2009b). Many of these models have undergone steady improvements or been replaced with improved versions over time, some have not (Cruz et al., 2018a).

The layers of forest vegetation and debris through which fire propagates is collectively called ‘fuel’ and can be classified according to various attributes in regard to their morphology, location, moisture loss or how they contribute to the behaviour of the fire via combustion (Byram, 1959, Agee, 1996, Scott and Burgan, 2005, Keane, 2016, Cruz et al., 2018b). Fine dead fuels, mostly comprising fallen leaf, bark and twig material 6 mm or less in diameter or living leaf and fine stems (3 mm diameter) of shrubs and trees, are those fuels that respond rapidly to changes in their thermal environment due to their high surface area relative their mass (Rothermel, 1972), generally with a comparatively high surface area-to-volume ratio (SAVR) often > 50 cm-1 (Brown, 1970). These fuels are often described as 1-h fuels as this is their moisture response time to changes in the environment (Fosberg and Deeming, 1971). They ignite readily and are rapidly consumed by fire (Chandler et al., 1983, p. 10). Coarser dead fuel elements (i.e., those >6 mm diameter), such as fallen branches, boughs and toppled stems (collectively known as down woody material or woody debris (WD)), have been shown to contribute significantly to total fuel load in most tropical, temperate and boreal forests (Hollis et al., 2011, van Leeuwen et al., 2014, Volkova and Weston, 2015). These fuel have a much reduced SAVR compared to fine fuels (generally < 50 cm-1 are often described as 10-h (>6–25.4 mm diameter) or 100-h (>25.4–75 mm diameter) fuels (Fosberg and Deeming, 1971)). However, the larger physical dimensions and masses of these coarser fuel elements mean that they take longer to ignite and combust.

Due to their relative ease of ignition and combustion, fine fuels have long been held to be the primary source of energy driving the behaviour of the flame front in a forest fire (McArthur, 1967, Rothermel, 1972, Rothermel, 1993), and changes in the way these fuels combust (particularly in regard to efficiency of energy transfer (Anderson and Rothermel, 1965)) affect the behaviour and spread of the fire front. In contrast, WD and larger live fuel elements generally ignite during or very soon after the passage of the flame front and can play a significant role in aspects of a forest fire other than the speed of the fire, particularly behind the fire front. For example, the contributions from WD elements have been found to significantly affect the total radiant heat flux behind the fire front and thus firefighter safety (Sullivan et al., 2002), the intensity, power, total energy release, severity, and burning depth (Agee, 1996, Hollis et al., 2011, Cruz et al., 2012), smoke production (Ward, 2001) and subsequent effects on human health (Weinhold, 2011, Reisen et al., 2011), and gaseous emissions (Bertschi et al., 2003) but also suppression difficulty (Proctor and McCarthy, 2015) and physical impact on soil and vegetation (Smith et al., 2017).

Operational forest fire spread prediction systems (e.g. Behave (Andrews, 1986) and BehavePlus (Andrews, 2014) based on the model of (Rothermel, 1972), the McArthur Forest Fire Danger Meter (McArthur, 1967), the Western Australia Forest Fire Behaviour Tables (FFBT) (Sneeuwjagt and Peet, 1985), the Canadian Forest Fire Behavior Prediction System (CFFBPS) (Forestry Canada Fire Danger Group, 1992) and the Dry Eucalypt Forest Fire Model (DEFFM) (Cheney et al., 2012)) are generally designed to predict the quasi-steady rate of forward spread of a fire using estimates of prevailing fuel, weather and topographic conditions. Some models, such as the McArthur, FFBT and DEFFM systems are purely empirical and directly consider only the contributions to fire spread from the combustion of fine dead fuels on the forest floor. Others, such as BehavePlus and the CFFBPS, integrate various degrees of quasi-empirical modelling based on physical considerations and allow the incorporation of the contribution of the combustion of other fuels, such as coarser dead fuels and fine live fuels, to total heat release and thereby fire propagation.

In the Rothermel model, mass of WD, weighted non-linearly by the SAVR of each fuel size class (the higher the SAVR, the greater the contribution), contributes to the overall fuel load and thus reaction intensity and propagating flux ratio but also to bulk density and total heat required to ignite the fuel (the heatsink). As a result, the influence of WD on rate of spread is quite complicated and not readily interpretable (Burgan and Rothermel, 1984). Increases to fuel bulk density and the heatsink generally result in decreases to rate of spread that overwhelm potential increases as a result of increased reaction intensity. In the CFFBPS, WD mass directly contributes to the rate of forward spread calculation through the Build-up Effect related to a Build-up Index as a function of surface fuel (both coarse and fine) consumption. In neither case does it appear that the methods by which the effects of WD on rate of spread are incorporated are based on data. Both McAlpine (1995) and Van Wagner (1998) point out that in a dataset of over 400 experimental fires no statistical evidence could be found to support an effect of Build-up Index on rate of spread, perhaps due to limitations of the data; instead a conceptually satisfying arbitrary function was incorporated into the system. As a result, the manner in which the effect of WD on rate of spread is included in operational systems around the world is inconsistent, ranging from not at all (e.g. McArthur and DEFFM) to arbitrarily, not based on data (Rothermel and CFFBPS).

Little is known about the characteristics of the particulate and gas emissions from the combustion of many types of WD, particularly WD found in the dry eucalypt forests of south-eastern Australia. As part of a larger research project attempting to characterise emissions released by combustion of WD under prescribed burning conditions (Cope et al., 2018), the opportunity was taken to investigate the effect WD on the forest floor has on fire behaviour, in particular rate of spread and flame dimensions, in the controlled conditions of a large combustion wind tunnel. Fallen branch material in the range 6–50 mm in diameter laid over a continuous fine litter fuel bed was used to test the hypothesis that the rate of forward spread of a fire is unaffected by the presence of WD in the fuel bed. Such an approach enables the ignition and combustion of WD via free-burning fire spreading through a surface fuel layer as would be found in nature, rather than burning pieces of WD in isolation from an external heat source as has been done in other studies (e.g. Albini and Reinhardt, 1995, Burrows, 2001).

Section snippets

Experimental design

The design of this experiment focused on two variables: the amount (or load) of WD and the fire spread mode (i.e. whether the fire burns with the wind (heading fire) or against the wind (backing fire)). All other burning conditions (wind speed, fine fuel moisture content (MC) and surface fuel load) were either non-varying (wind speed and surface fuel load) or variation was limited as much as was practical (e.g. surface fuel MC). The values of these experimental factors were chosen to represent

Results

A summary of the experimental results is presented in Table 2. Median fuel bed composition by mass was: 51.7% twigs, 40.5% leaves and 7.3% bark. The coefficient of variation for these fuel bed components was 9.0, 11.2 and 12.1% respectively, indicating that litter fuel composition did not vary greatly between fuel beds. Leaves had a higher carbon content (53.4%) than all other fuel components (48.7–49.1%). The molar N/C ratio for litter was 0.590% and 0.269% for woody fuel.

Fine and WD fuel MCs

Discussion

The results of the experiments disproved the experiment’s working hypothesis in the case of fires burning with the wind (i.e. heading fires) that the presence of WD would not affect fire rate of spread. Indeed, the presence of any WD was shown to halve the rate of spread compared to a fire burning in litter only fuels. There was, however, little significant effect of the amount of WD on the rate of spread for fires with WD. In contrast, the hypothesis was confirmed for fires burning into the

Conclusions

Experiments were conducted in the CSIRO Pyrotron combustion wind tunnel under controlled conditions (fixed wind speed, fine fuel load and fine fuel moisture content) to investigate the effect of the presence of branch and small log material (< 50 mm diameter, collectively known as woody debris (WD)) had on the rate of spread of fires burning with and against the wind in litter fuels. The amount of WD range from 0 kg m-2 (control) to 1.2 kg m-2. It was found that the presence of WD had a direct

Acknowledgements

Funding for this work was made available through the Victorian Department of Environment, Land, Water and Planning. The authors thank Martin Cope of CSIRO Oceans and Atmosphere for leading the overall project and Fabienne Reisen, Camila Trindade, Grant Edwards, Dean Howard, Stephen Wilson, Clare Murphy (née Paton-Walsh), Elise-Andrée Guérette, Maximilien Desservettaz, and Valentina Wheeler for their assistance during the experimental programme. We also thank Miguel Cruz and Lachie McCaw for

References (75)

  • J.E. Smith et al.

    Does the presence of large down wood at the time of a forest fire impact soil recovery?

    For. Ecol. Manage.

    (2017)
  • A. Sullivan et al.

    A contractionless, low-turbulence wind tunnel for the study of free-burning fires

    Exp. Thermal Fluid Sci.

    (2013)
  • L. Volkova et al.

    Redistribution and emission of forest carbon by planned burning in Eucalyptus obliqua (L. Hérit.) forest of south-eastern Australia

    For. Ecol. Manage.

    (2013)
  • Agee, J.K., 1996. The influence of forest structure on fire behavior. In: Proceedings of the 17th Annual Forest...
  • F.A. Albini et al.

    Modeling ignition and burning rate of large woody natural fuels

    Int. J. Wildland Fire

    (1995)
  • Anderson, H.E., 1964. Mechanisms of fire spread. Research Paper INT-8, USDA Forest Service, Intermountain Forest &...
  • W.R. Anderson et al.

    Convective heat transfer in fire spread through fine fuel beds

    Int. J. Wildland Fire

    (2010)
  • Andrews, P., 1986. BEHAVE: fire behaviour prediction and fuel modelling system – BURN subsystem, Part 1. General...
  • P.L. Andrews

    Current status and future needs of the BehavePlus Fire Modeling System

    Int. J. Wildland Fire

    (2014)
  • A. Banerjee et al.

    Cointegration, Error Correction, and the Econometric Analysis of Non-Stationary Data

    (1993)
  • I. Bertschi et al.

    Trace gas and particle emissions from fires in large diameter and below ground biomass fuels

    J. Geophys. Res.

    (2003)
  • Boundy, B., Diegel, S.W., Wright, L., Davis, S., 2011. Biomass energy data book: Edition 4. Oak Ridge National...
  • G.E.P. Box et al.

    An analysis of transformations

    J. Roy. Stat. Soc. Series B (Methodological)

    (1964)
  • J.K. Brown

    Ratios of surface area to volume for common fine fuels

    For. Sci.

    (1970)
  • Burgan, R.E., Rothermel, R.C., 1984. BEHAVE: fire behaviour and fuel modelling system – FUEL subsystem. General...
  • N.D. Burrows

    Flame residence time and rates of weight loss of eucalypt forest fuel particles

    Int. J. Wildland Fire

    (2001)
  • G.M. Byram

    Combustion of forest fuels

  • C. Chandler et al.

    Fire in Forestry 1: Forest Fire Behaviour and Effects

    (1983)
  • Cheney, N.P., 1981. Fire behaviour. In: Gill, A.M., Groves, R.H., Noble, I.R. (Eds.), Fire and the Australian biota....
  • N.P. Cheney et al.

    Prediction of fire spread in grasslands

    Int. J. Wildland Fire

    (1998)
  • R.D. Cook

    Detection of influential observation in linear regression

    Technometrics

    (1977)
  • Cope, M., Lee, S., Meyer, M., Sullivan, A., Surawski, N., Trindade, C., Reisen, F., Wain, A., Weston, C., Volkova, L.,...
  • M.G. Cruz et al.

    Empirical-based models for predicting head-fire rate of spread in Australian fuel types

    Aust. Forest.

    (2015)
  • M.G. Cruz et al.

    A hierarchical classification of wildland fire fuels for Australian vegetation types

    Fire

    (2018)
  • J.R. Curry et al.

    Rate of spread of surface fires in the Ponderosa pine type of California

    J. Agric. Res.

    (1938)
  • Davies, J.B., Oates, A.M., Trumbull-Ward, A.V., 2002. Ecological vegetation class mapping at 1:25 000 in Gippsland....
  • M.A. Finney et al.

    On the need for a theory of wildland fire spread

    Int. J. Wildland Fire

    (2013)
  • Cited by (20)

    • Potential for forest thinning to reduce risk and increase resilience to wildfire in Australian temperate Eucalyptus forests

      2021, Current Opinion in Environmental Science and Health
      Citation Excerpt :

      On the one hand, thinning usually results in greater amounts of woody debris [25], making it more difficult to completely extinguish wildfires after the fire front has passed through. On the other hand, a wind tunnel simulation found the presence of logs halved the rate of spread of heading fires compared with when no logs were present [29]. A recent study [30] of area burnt in the 2009 Black Saturday fires in Victoria found no difference between thinned and unthinned areas in crowns burnt in two ages of Eucalyptus regnans forests but significantly lower probability of fire impact in thinned younger-aged mixed species stands (at a stage when forests are most vulnerable).

    • Production of low oxygen bio-oil via catalytic pyrolysis of forest residues in a kilogram-scale rotary kiln reactor

      2020, Journal of Cleaner Production
      Citation Excerpt :

      It was reported that forestry residues were the biomass resource with the greatest potential to increase bio-energy production worldwide (Junginger et al., 2014), making them an ideal feedstock for pyrolysis. Forest residues have a low economic value and constitute a fire hazard (Sullivan et al., 2018). Using them for energy production would unlock the energy potential of these feedstocks (Junginger et al., 2014).

    • Exploring how fire spread mode shapes the composition of pyrogenic carbon from burning forest litter fuels in a combustion wind tunnel

      2020, Science of the Total Environment
      Citation Excerpt :

      This conclusion is supported by results from Santin et al. (2017) who suggest that the charring intensity metric (i.e. a measure of temperature residence time) plays a critical role in PyC production. Inclusion of larger fuel elements could potentially change the results we found since coarse woody debris is quite often not fully consumed during the passage of a fire front, leaving PyC in the form of charcoal after fire (Sullivan et al., 2018). Fig. 6 shows that the flanking FSM presented the smallest changes in overall 13C NMR composition due to fire.

    View all citing articles on Scopus
    1

    Current address: School of Civil and Environmental Engineering, University of Technology Sydney, Ultimo 2007, Australia.

    View full text