Comparison of Firebrand Propagation Prediction by a Plume Model and a Coupled–Fire/Atmosphere Large–Eddy Simulator

Abstract

Firebrand spotting is one of the most vexing problems associated with wildland fires, challenging the lives and efforts of fire-fighting planners. This work is an effort to model numerically the event of firebrand spotting for the purposes of reviewing past modelling approaches and of demonstrating a more current coupled fire/atmosphere approach. A simple, two-dimensional treatment of the process of firebrand lofting is examined under the restrictive conditions typical of a classical plume modelling approach. Using this approach, the differences in trajectories of combusting and non-combusting particles are investigated. Next, firebrand spotting is examined using a coupled fire/atmosphere LES (Large Eddy Simulator) in which the processes of firebrand lofting, propagation, and deposition are connected. The behaviour of combusting and non-combusting firebrands released from a moving grassfire into three-dimensional time-varying coupled atmosphere-wildfire induced circulations is examined. When these results are compared to the results of a classical plume model for firebrand spotting, it is found that firebrand propagation in the coupled LES simulated flow is significantly different from that obtained by the two-dimensional empirically-derived plume model approach. The coupled atmosphere-wildfire LES results are explorative and need to be subjected to direct testing.

The Spotting Distribution of Wildfires

“…As mentioned above, coupled fire-atmosphere models [2,17,18] may provide an improvement to static fluid dynamic computations currently employed in operational front prediction software (for North American examples see [19,20]). Indeed, Large Eddy Simulations (LES) coupled with firebrand dispersal have shown that the fire plume may be quite different from the plume used in standard models like the Baum and McCaffery plume [21], leading to different firebrand trajectories [22][23][24]. In addition, there is an extensive literature covering empirical measurement of plume characteristics; extensive measurement of plume heights and characteristics in [25] compared a variety of plume models (different from Baum and McCaffery) against measurements for approximately 2000 wildfire plumes.…”

“…We summarize all models and their references in Table A1. Temperature T(t) T1, Newton's Law of Cooling, T2, Stefan-Boltzmann law [22].…”

“…Any fire plume is more turbulent than laminar, and our knowledge about plumes is mostly experimental [61]. There is a complicated interaction between the atmosphere and the fire, hereafter referred to as fire-atmosphere interactions, which has been extensively studied from experiments, and physical modelling [1,2,17,18,[22][23][24]31,[62][63][64][65][66][67].…”

Physical parametrisation of fire-spotting for operational fire spread models: response analysis with a model based on the Level Set Method

“…Besides these statistical approaches, few numerical models based on Large Eddy Simulation (LES) (Himoto and 25 Tanaka, 2005;Thurston et al, 2017;Tohidi and Kaye, 2017) or Computational Fluid Dynamics (CFD) (Wadhwani et al, 2017), small world networks , cellular automata models (Perryman et al, 2013) also exist in the literature. Bhutia et al (2010) present one such study based on coupled fire/atmosphere LES for predicting the short range fire-spotting. They simulate multiple firebrand trajectories for analyzing the sensitivity of the flight path to different particle sizes, release heights and wind conditions but also mention the limited applicability of such models to operational use due to the computational 30 demands.…”

Modelling wildland fire propagation by tracking random fronts

“…In fact, extremely important phenomena in wildland fire propagation are turbulent hot-air transport due to the turbulent nature of the atmospheric boundary layer that can consequently affect fire-atmosphere interactions (Clark et al, 1996;Potter, 2002Potter, , 2012aLinn and Cunningham, 2005;Sun et al, 2006;Filippi et al, , 2011Sun et al, 2009;Mandel et al, 2011;Forthofer and Goodrick, 2011), as well as the fire spotting phenomenon (Sardoy et al, 2007(Sardoy et al, , 2008; Kortas et al, 2009;Perryman, 2009;Bhutia et al, 2010;Koo et al, 2010;Wang, 2011;Morgante, 2011;Perryman et al, 2013). Both processes have a random character; therefore, the fire front motion turns out to be random.…”