Role of buoyant flame dynamics in wildfire spread

Abstract

Large wildfires of increasing frequency and severity threaten local populations and natural resources and contribute carbon emissions into the earth-climate system. Although wildfires have been researched and modeled for decades, no verifiable physical theory of spread is available to form the basis for the precise predictions needed to manage fires more effectively and reduce their environmental, economic, ecological, and climate impacts. Here, we report new experiments conducted at multiple scales that appear to reveal how wildfire spread derives from the tight coupling between flame dynamics induced by buoyancy and fine-particle response to convection. Convective cooling of the fine-sized fuel particles in wildland vegetation is observed to efficiently offset heating by thermal radiation until convective heating by contact with flames and hot gasses occurs. The structure and intermittency of flames that ignite fuel particles were found to correlate with instabilities induced by the strong buoyancy of the flame zone itself. Discovery that ignition in wildfires is critically dependent on nonsteady flame convection governed by buoyant and inertial interaction advances both theory and the physical basis for practical modeling.

Keywords: buoyant instability; convective heating; flame spread; wildfires.

Fig. 1.

Images of saw-tooth geometry and stream-wise streaks in flame fronts at laboratory and field scales. Arrows indicate wind or spread direction. (A) Top view of stationary flames from an ethylene gas burner. (B) Laboratory fire with 0.40-m flames (C) Same as in B but with 1.5-m flames. (D) Front view of 3-m-tall flames from an approaching heading fire in Texas grassland. (E) Top view of burn streaks behind flame zone. (F) Stationary fire with 6-m flames on 16-m wooden crib. (G) Numerical simulation of wildfire with an ∼6-m flame length (reprinted from ref. , with permission from Elsevier; www.sciencedirect.com/science/journal/01681923). (H) Front view of experimental crown fire in Canada with an ∼15-m flame length (image courtesy of the US Department of Agriculture, Forest Service; related to crown fires in ref. 30). (I) Plot of separation between flame peaks (same as stream-wise streak spacing) in relation to flame length.

Fig. 2.

Illustrations of flame dynamics and related flow instabilities observed in experimental burns. Flame-zone buoyancy creates stream-wise vortex pairs that alternately push flames up into peaks and down into troughs. Streaks of smoldering combustion aligned with flame peaks extend back behind the front. Concave flame parcels travel through the flame zone and burst intermittently forward through the troughs to heat unignited fuels.

Fig. 3.

Images from behind the flame zone illustrate buoyant instabilities forming as transverse waves (brackets) that advect forward in (A) a stationary ethylene gas burner and (B) a wind-tunnel experimental fire spreading in cardboard fuel (Movie S1). Stream-wise (longitudinal) vortices (C) induce flame peaks and troughs at up-wash and down-wash convergence zones. Movie images (Movies S2 and S4) at the leading edge of spreading fires (D) show flame vortex circulations and forward flame bursts through flame troughs after flow-tracking analysis (E).

Fig. 4.

Overhead view of a laboratory fire spreading away from camera through a cardboard fuel bed shows (A) coherent parcels of flame surface (bracket) generated by transverse instabilities in the flame zone and approximate locations of thermocouples, which (B) record temperature fluctuations before ignition when coherent flame structures burst forward within the intermittent flame zone of the fuel bed (spline curves distinguish flaming to intermittent transition). Temperature fluctuations recorded by thermocouples in spreading grass fires (C) and a stationary crib burn (D) were plotted with time-average frequency of flame intermittency from all sources (E) to reveal a strong relationship between Strouhal number (St = fL U⁻¹) and Froude number [Fr = U² (gL)^-1] scaled by wind speed U and flame length L.

Fig. 5.

Time series of irradiance, air temperature, and 1-mm particle surface temperature during flame spread in cardboard fuel bed (A). Fine particles heat and cool rapidly as intermittent flame contacts produce a stair-step temperature rise to ignition. Radiation preheating indicated when particle surface is warmer than air. Thermal infrared image sequence (B) from above and behind the fire show heating patterns on rows of 6-mm-wide cardboard particles as fire approaches. Higher temperature and ignition of corners and edges indicate convection as the principal heating mechanism.