Register      Login
International Journal of Wildland Fire International Journal of Wildland Fire Society
Journal of the International Association of Wildland Fire
RESEARCH ARTICLE

A simple model for wind effects of burning structures and topography on wildland–urban interface surface-fire propagation

Ronald G. Rehm A C and William (Ruddy) Mell B
+ Author Affiliations
- Author Affiliations

A RGR Consulting, Limited Liability Corporation and National Institute of Standards and Technology, Building 224, Room B260, 100 Bureau Drive, Gaithersburg, MD 20899, USA.

B National Institute of Standards and Technology, Building 224, Room B246, 100 Bureau Drive, Gaithersburg, MD 20899, USA.

C Corresponding author. Email: ronald.rehm@nist.gov

International Journal of Wildland Fire 18(3) 290-301 https://doi.org/10.1071/WF08087
Submitted: 29 May 2007  Accepted: 1 July 2008   Published: 28 May 2009

Abstract

The present paper presents a simple model to demonstrate the effect on grass-fire propagation of the winds induced by structural fires in a wildland–urban interface setting. The model combines an empirical formula for wind-driven grass-fire spread and a physics-based analytical solution to the Euler equations to determine the ground-level wind produced by the burning structure. The scaling of the wind is based on the heat release rate of the structural fire as well as other parameters. Also considered are an ambient wind and a topographical wind, assumed to be proportional to the ground slope. Data on grass and structure fires required by the model are discussed. Fire front propagation predicted by this model is illustrated by three examples: a front passing a single burning structure on flat terrain, a front passing a burning structure on a hill, and a front passing several burning structures. The model predicts that a fire front will be accelerated toward the burning structure upon approach and decelerated after passing the structure. Several burning structures multiply the effects of an individual burning structure.


Acknowledgements

The present research (R.G.R) was supported in part by NIST Contract Number SB1341–05-W-1003, Dr Anthony Hamins, Contract Monitor.


References


Albini FA (1984) Wildland fires. American Scientist  72, 590–597.
Baum HR, McCaffrey B (1989) Fire induced flow field – theory and experiment. In ‘Fire Safety Science: Proceedings of the Second International Symposium’, 13–17 June 1988, Tokyo, Japan. (Eds T Wakamatsu, Y Hasemi, A Sekizawa, PG Seeger) pp. 129–148. (International Association for Fire Safety Science and Hemisphere Publishing Corp.: New York)

Baum HR, McGrattan KB, Rehm RG (1994) Simulation of smoke plumes from large oil fires. In ‘Twenty-Fifth Symposium (International) on Combustion’, 31 July–5 August 1994, Irvine, CA. Vol. 25, pp. 1463–1469. (The Combustion Institute: Pittsburgh, PA)

Berry A (2007) Forest policy up in smoke: fire suppression in the United States. Property and Environmental Research Center, pp. 1–21. (Bozeman, MN) Available at http://www.perc.org/articles/article1004.php [Verified 4 May 2009]

Carlton D, Andrews P, Bevins C (2004) BehavePlus fire modeling system, user’s guide, quick start tutorial. USDA Forest Service, Rocky Mountain Research Station, Systems for Environmental Management, Online Help development. Available at http://firemodels.org/content/view/12/26/ [Verified 18 May 2009]

Cheney NP, Gould JS , Catchpole WR (1998) Prediction of fire spread in grasslands. International Journal of Wildland Fire  8, 1–13.
Crossref | GoogleScholarGoogle Scholar | Dold J, Zinoviev A, Weber R (2006) Non-local flow effects in bushfire spread rates. In ‘V International Conference on Forest Fire Research’, 27–30 November 2006, Coimbra, Portugal. (Ed. DX Viegas) (Centro de Estudos sobre Incendios Florestais: Coimbra, Portugal)

Environmental Literacy Council (2007) Forest fires. Environmental Literacy Council. (Washington, DC) Available at http://www.enviroliteracy.org/article.php/46.html [Verified 5 May 2009]

Fendell FE, Wolff MF (2001) Wind-aided fire spread. In ‘Forest Fires, Behavior and Ecological Effects’. (Eds EA Johnson, K Miyanishi) Ch. 6, pp. 171–223. (Academic Press: San Diego)

Finney MA (2004) FARSITE: Fire Area Simulator – model development and evaluation. USDA Forest Service, Rocky Mountain Research Station, Research Paper RMRS-RP-4, March 1998, Revised February 2004. (Missoula, MT) Available at http://www.firemodels.org/downloads/farsite/publications/fireareaall.pdf [Verified 5 May 2009]

GAO (2005) Technology assessment; protecting structures and improving communications during wildland fires. General Accountability Office, Report GAO-05–380. (Washington, DC)

Gustafson EJ, Hammer RB, Radeloff VC , Potts RS (2005) The relationship between environmental amenities and changing human settlement patterns between 1980 and 2000 in the Midwestern USA. Landscape Ecology  20, 773–789.
Crossref | GoogleScholarGoogle Scholar | Jones WW, Peacock RD, Forney GP, Reneke PA (2004) CFAST – consolidated model of fire growth and smoke transport (version 5) technical reference guide. National Institute of Standards and Technology, Special Publication 1030. (Gaithersburg, MD) Available at http://fire.nist.gov/bfrlpubs/fire04/PDF/f04046.pdf [Verified 5 May 2009]

Linn RR (1997) Transport model for prediction of wildfire behavior. Los Alamos National Laboratory, Scientific Report LA13334-T. (Los Alamos, NM)

McGrattan KB, Baum HR , Rehm RG (1996) Numerical simulation of smoke plumes from large oil fires. Atmospheric Environment  30, 4125–4136.
Crossref | GoogleScholarGoogle Scholar | CAS | McGrattan KB, Hostikka S, Floyd JE, Baum HR, Rehm RG (2007) Fire dynamics simulator (version 5): technical reference guide. National Institute of Standards and Technology, NIST Special Publication 1018-5. (Gaithersburg, MD) Available at http://fire.nist.gov/bfrlpubs/fire07/PDF/f07048.pdf [Verified 19 May 2009]

Mell W, Jenkins MA, Gould J , Cheney P (2007) A physics-based approach to modeling grassland fires. International Journal of Wildland Fire  16, 1–22.
Crossref | GoogleScholarGoogle Scholar | Murphy K, Rich T, Sexton T (2007) An assessment of fuel treatment effects on fire behavior, suppression effectiveness, structure ignition on the Angora Fire. USDA Forest Service, Pacific Southwest Research Station, Report R5-TP-025. (Vallejo, CA)

National Interagency Fire Center (2006a) Historically significant wildland fires. (Boise, ID) Available at http://www.nifc.gov/fire_info/historical_stats.htm [Verified 7 May 2009]

National Interagency Fire Center (2006b) Total wildland fires and acres (1960–2007). (Boise, ID) Available at http://www.nifc.gov/fire_info/fires_acres.htm [Verified 7 May 2009]

Ohlemiller T , Corley D (1994) Heat release rate and induced wind field in a large scale fire. Combustion Science and Technology  97, 315–330.
Crossref | GoogleScholarGoogle Scholar | Quintiere J (1997) ‘Principles of Fire Behavior.’ (Delmar Publishers: Albany, NY)

Radeloff VC, Hammer RB , Stewart SI (2005) Rural and suburban sprawl in the US Midwest from 1940 to 2000 and its relation to forest fragmentation. Conservation Biology  19(3), 793–805.
Crossref | GoogleScholarGoogle Scholar | Rehm RG (2006) The effects of winds from burning structures on ground-fire propagation at the wildland–urban interface. National Institute of Standards and Technology, Report GCR 06–892, pp. 1–31. (Gaithersburg, MD)

Rehm RG, Hamins A, Baum HR, Mcgrattan KB, Evans DD (2002) Community-scale fire spread. In ‘Proceedings of the California’s 2001 Wildfire Conference: 10 Years after the 1991 East Bay Hills Fire’, 10–12 October 2001. (Eds KS Blonski, ME Morales, TJ Morales) Oakland California Technical Report 35.01.462, pp. 126–139. (University of California Forest Products Laboratory: Richmond CA)

Rothermel RC (1972) A mathematical model for predicting fire spread in wildland fuels. USDA Forest Service, Rocky Mountain Research Station, Research Paper INT-115. (Missoula, MT)

Sethian JA (1999) ‘Level Set Methods and Fast Marching Methods, Evolving Interfaces in Computational Geometry, Fluid Mechanics, Computer Vision, and Materials Science.’ (Cambridge University Press: Cambridge, UK)

Trelles J, Pagni PJ (1997) Fire-induced winds in the 20 October 1991 Oakland Hills Fire. In ‘Proceedings of the Fifth International Symposium for Fire Safety Sciences’, 3–7 March 1997, Melbourne. (Ed. Y Hasemi) pp. 911–922. (International Association for Fire Safety Science: Boston, MA)

Wolfram S (1999) ‘The Mathematica Book, 4th edn.’ (Wolfram Media and Cambridge University Press: Cambridge, UK)




Appendix

In this Appendix, Lagrangian equations are presented for determining the fire-front propagation on a specified surface or topography in the presence of burning structures. The governing equations are the ODEs describing the propagation of an element of the fire front along the surface:

E5

The equations are given in vector form WF08087_IE2.gif, where WF08087_IE3.gif are unit vectors in the x, y, z directions. WF08087_IE6.gif is the rate of spread (ROS) vector of the fire front at the location (x, y, z), and are the components of the unit normal to the fire front directed toward the unburnt fuel. s is the arc length of the curve.

The linear relation given in Eqn 1 between the ROS and the local wind velocity is assumed. For a line of sufficient head width W and for very low moisture content of the grass, this reduces to:

E6

where WF08087_IE8.gif and WF08087_IE9.gif with ROS0 = 0.165 [m s–1] and cf = 3.24. Then

E7

For notational simplicity, we will not explicitly show the time dependency of these functions as we obtain the tangent and the normal vectors to the fire front. The tangent to this curve is determined by WF08087_IE10.gif and can be written as:

E8

Define the normal vector to the fire front by the parameter n as follows:

E9

WF08087_IE11.gif is determined by the requirement that the dot product of the normal and tangential vectors be zero: WF08087_IE12.gif

Define the following quantities:

E10

and

E11

Then, we can write the unit normal vector to the fire front on the prescribed surface as follows:

E12

At the wildland–urban interface, houses as well as wildland fuels can be burning. In this model, we account for three elementary wind fields, the ambient wind, a topographically induced wind and the entrainment wind produced by all burning structures. We assume that the total velocity at any location can then be determined by adding linearly the contributions of all the elementary wind fields. The total velocity locally is taken to be the sum of these velocity contributions:

E13

We take for simplicity the horizontal components of the local ambient wind (Va,x, Va,y) to be uniform in space. Then, the z-component of the ambient wind is given by WF08087_IE13.gif.

In the absence of any ambient wind and over flat terrain, a fire front is found to propagate with the uniform ROS as discussed above. When there are topographical features, Z(x,y), but no ambient wind, it is found that the ROS of the fire front increases uphill and decreases downhill because of buoyancy effects. This observed behavior can be treated by defining a topographically induced horizontal velocity that is proportional to the gradient of the hill (Rothermel 1972). Here, we take this equivalent horizontal velocity to be given by the relations WF08087_IE14.gif and WF08087_IE15.gif, where α is a proportionality constant (Rothermel 1972). The z-component of the topographically induced velocity is given by WF08087_IE16.gif.

For a fire front exposed to the velocity field generated by a single burning structure of HRR Q0, the characteristic length and velocity scales are D* and V* as discussed earlier. Let WF08087_IE17.gif denote the vector distance from the center of the structure to the element of the fire front. The velocity at this point will be where is the dimensionless velocity and is the length scale defined above, and the dimensionless vector distance, WF08087_IE18.gif, is WF08087_IE19.gif.

The detailed solution for the dimensionless velocity function at ground level, G(r) was obtained analytically by Baum and McCaffrey in terms of special functions. For computational purposes, however, this solution was replaced in the example calculations presented here by the functional form given below, which closely approximates the analytical solution:

E14

where r0 = 0.8; r1 = 1.0; f0 = 0.407199; f1 = 0.045029; a = –2.39441; b = 11.2283; c = –13.6154; and d = 4.9468.

Therefore, for a single burning structure at (x = h, y = H), the induced horizontal entrainment velocity components at any point (x, y) are

E15

The vertical component of the entrainment velocity is taken to be

E16

The total entrainment from all of the burning structures is obtained by simple summation (Rehm 2006). If, for example, the jth structure has an entrainment velocity and a characteristic length scale determined by the heat release rate of the burning structure (as described in Rehm 2006), and if the location of the structure is given by x = hj, y = Hj, then the total entrainment velocity is given by

E17

where we have assumed that the entrainment velocity from each burning structure is only dependent on the vector distance in a horizontal plane between the observational location (x, y) and the location of the burning structure.

Only the first two of the three vector-component equations for the fire-front spread need to be solved, as the front is constrained to the surface z = Z(x, y) (assuming that we can describe the surface explicitly in this form). The functional form for the surface is used to eliminate z in the component equations for x and y, yielding two ODEs for x(s, t) and y(s, t). These are solved as described below, and z is determined at (x, y) from the equation for the surface.

At each point, the fire front is advanced in the direction normal to the front at a speed determined by the local ROS for the fire. This ROS, in turn, depends on the wind speed at that location. For computational purposes, the fire front is discretized and then moved incrementally to its new location. We consider the fireline initially to be a straight line along the x-axis, running between –L and L and divide this interval into 2I panels each of length δ, where δ = 1/I. We start with an approximation to the normal ROS, and then numerically solve the governing equations. We use the MOL and a centered difference scheme for the spatial discretization at all interior nodes of the fire front. For the end nodes, we use a one-sided difference scheme with the neighboring interior node.