Direct Flame Contact
Direct flame contact is one of the three structure ignition pathways, together with firebrands and radiant heat. Direct flame contact refers to flames impinging on building systems and materials. It may come either from the main wildfire flames, from burning elements and ornamental vegetation surrounding structures, or from neighboring structures.
Flame contact is the most hazardous structure ignition mechanism at the Wildland-Urban Interface (WUI) as it provides the highest heat fluxes (radiant and convective). Flame contact might be originated from nearby burning wildland fuels, burning ornamental vegetation, and nonnatural burning items close to structures and neighboring structures.
Flame contact from the main wildland fire perimeter may play an important role in igniting peripheral structures of a community, in those cases where wildland fuels are too close to settlements. This situation involves usually high heat loads. However, flame contact is not generally considered a source of ignition in managed, dispersed, and cleared WUI communities.
Despite the existence of several codes, standards, and guidelines issued to minimize the risk of WUI structure ignitions, immediate home surroundings at the WUI are too often characterized by the presence of all sorts of combustible elements. Ground fuels, stored material, ornamental trees or hedges, fences, and outbuildings (e.g., garages, garden, or storage sheds) might cause severe impact in case of ignition and significant flame impingement on home’s exterior vulnerable surfaces. Regardless of building designs and practices, houses always have weak elements to fire exposure (e.g., openings, glazing and flooring systems, decks and verandas, or eaves and gutters). This type of elements is responsible for houses’ vulnerability, either because they are combustible or made of materials sensitive to fire or because their geometry enhances heat transfer.
Direct flame contact between neighboring structures is also a driver of fire propagation across densely populated WUI communities (structure-to-structure spread). In these scenarios, flame exposure times are generally higher than those of wildfire flame impingement, since structures may burn for longer periods. When decreasing separation between houses, home-to-home spread of fire is more significant than wildland-to-home spread (Cohen 2008). This situation has been evidenced worldwide in several past fires, leading to multistructure involvement with a large number of home losses.
Direct Flame Contact Ignition Pathway
Flames impinging solid surfaces at the WUI are usually turbulent, highly nonsteady, hence presenting temperature and geometry fluctuations. In contact with structure’s combustible components, flames may heat those up to piloted ignition provided they have sufficient temperature, impinge during sufficient time, and there is enough heat release rate delivered over a certain area (Babrauskas 2003). When there is flame contact, a combination of radiation and convection heat transfer takes place from the fire to the exposed object. Heat transfer is in this situation highly efficient. On the one hand, flames and hot gases flowing over the combustible surfaces are responsible for convective heating causing a surface temperature increase of the exposed object. On the other hand, radiation heat transfer is also at its maximum level, since there is no space between flames and the target. Therefore, losses due to atmospheric absorptivity within the radiation path are null, and view factors can be considered close to unity, particularly in vertical surfaces with total flame impingement.
The dominance between convection and thermal radiation heat transfer during direct flame contact may depend on flames and target characteristics. Flame composition (particularly soot particles concentration), thickness, and temperature will be responsible for the amount of heat emitted by radiation, whereas flow configuration, solid size and shape, and temperature difference between the flame and the solid will govern convection heat transfer. Driven by one or another heat transfer mode or both, after a certain amount of heat dosage, combustible materials will reach ignition temperature; solids will begin to pyrolyze emitting vapors that will be immediately pilot-ignited by the presence of the impinging flame.
Flames Impinging WUI Structures: Physical Characteristics and Impact on Building Systems
Flames impinging WUI structures can be highly variable in terms of thermal and geometric characteristics, according to the burning fuel types they come from: wildland fuels (e.g., grasslands, shrublands, and forests), ornamental vegetation (trees, hedgerows, short grasses), and different types of nonnatural combustible elements (plastics, fabrics, hydrocarbons, etc.) or adjacent structures. Moreover, wind and local turbulences may exacerbate flames size and intensity, hence increasing heat fluxes and exposed areas.
While a comprehensive characterization and categorization of the physical features of these different types of fire exposures would be necessary for prevention and suppression purposes at the WUI, there is no exhaustive survey nor database available in the literature gathering this information. Research community has produced summary of flames physical characterization. Experimental studies have reported radiative, convective, and total energy emitted by wildland-fueled flames, flame geometry, and temperatures in different types of fuel complexes. As a selected recent example, Frankman et al. (2012) monitored 25-m flame height crown fires with irradiance peaks nearly 300 kW/m2 and convective heat with maximum values of 40 kW/m2. Regarding temperatures, Butler et al. (2004) reported maximum values of around 1300 ∘C in crown fire experiments. Other quantitative data with wildfire flame metrics can be found in Cruz et al. (2011) and Wotton et al. (2012).
Some studies are available in the scientific literature regarding the flame contact effect on WUI structures and building systems. They are mainly experimental, although some modelling efforts can also be found. Tran et al. (1992) performed laboratory experiments involving flame impingement over wood materials. Mock wall sections were exposed to a propane flame from a burner with two burning programs (40 kW during 10 min and 160 kW during 2 min, respectively). Results showed that after 1 min of flame impingement, hardboard siding material ignited and self-extinguished after the burner flame was turned off, hence pointing to the duration of flame impingement as a dominant factor for continued burning. Two sets of field experiments were also used to evaluate exterior walls performance under crown fires exposure (Cohen 2000, 2004) involving flames of around 20 m height. Although experimental design contemplated radiant heat exposure as the main ignition pathway under inspection, Cohen (2000) reported flame contact being responsible for the ignition of 10-m wood wall sections located 10 m apart from the flaming edge. In the first experimental campaign (Alexander et al. 1998), in two of the five burns, flames extended into the clearing beyond the forest edge igniting the wooden panels. Without flame contact, only scorch occurred with flames being at 10 m distance. The second set of experiments (Stocks et al. 2004) provided similar type of results. In this case, three out of seven wall sections located 10 m apart from the fire edge ignited with one of them experiencing brief flame contact by a turbulent crown fire flame extension. Grishin et al. (2014) experimentally investigated wooden shield ignition by grass fire flame contact. Flames up to 8.4 kW/m2 and within a range of 0.4–1.2 m height impinged over uniform wooden shields and fences. Results evidenced that fence shield was less damaged due to lesser exposure time (around 50 s) to radiation and convection heat fluxes and to higher convection cooling effect due to gaps in the fence. In contrast, uniform wood shields experienced ignition after 95 s of flame impingement. The thermal impact of structural elements exposed to wildland fires was studied by physics-based modelling in Porterie et al. (2005). The authors of this work validated their model showing it was capable of simulating the transient thermal response of structures exposed to the combined effect of radiation and convection due to flame impingement. Moreover, they showed evidence of how the presence of structures can modify the flow pattern leading to fire plume enlargement.
Some effort has been devoted to ornamental vegetation-fueled flames’ characterization. Although single trees and hedgerows have been evidenced to involve large and intense flames, the contribution of ornamental fuels to the WUI fire problem is generally tackled through experimental studies with the aim of classifying species flammability (e.g., White and Zipperer 2010; Ganteaume et al. 2013). Other fuels that might be present close to WUI structures can involve a wide variety of materials (plastics, fabrics, liquid fuels, etc.). Manzello et al. (2017) studied flame contact ignition in reentrant corners exposed to three different types of flaming mulch beds through realistic-scale experiments. Corners constructed from wood studs and oriented strand board (with and without sidings) were tested under different wind speeds. Except for one test, wall ignition was observed in all experiments.
Neighboring structure fires can also provide flame contact in high housing density WUI areas. Structure fires physical characteristics in the WUI have been estimated in terms of heat release rate and duration. As an example, according to data collected in the Oakland Hills Fire (California, USA, October 1991), Trelles and Pagni (1997) modelled a house fire considering 15-m height flames and an energy release rate with a peak of 45 MW for 1 h and a two-step decaying phase (10 MW for 3 h followed by 5 MW for 3 more h). Himoto et al. (2018) experimentally investigated flame spread through mock two-story wooden houses (built at 1/3 scale, i.e., 3.15-m height models) emulating an urban area. The estimated heat release rate of a model house was approximately 16 MW in a fully developed phase fire. Observed maximum flame heights reached 7.8 m. In experiments evaluating residential structure separation distances (Maranghides and Johnsson 2008), heat release rate of a fire-resistant construction (3.7 m by 4.3 m room with a 2.4 m height ceiling) reached maximum values of around 10 MW, with flames escaping from the window during 9 min.
The role of flame contact of structure-to-structure ignition has been studied within the framework of urban fires and WUI fires. In Himoto et al. (2018), a field experiment on fire spread within a group of model houses is reported. Separation distances ranged from 0.45 m to 3.6 m. With average wind velocities between 1.1 and 1.6 m/s, fire plume tilting was barely observed, hence being thermal radiation heat transfer the major contributing factor of fire spread rather than direct flame contact. In a full-scale laboratory experiment, Maranghides and Johnsson (2008) report structure-to-structure ignition by flame impingement. Tests showed how an adjacent structure separated 1.8 m from the fire source can be ignited if flames from a fire inside a house exit through window openings. In Rehm (2008) a fire propagation model is developed considering, apart from ambient wind, induced wind generated by burning structures. In the model, structure-to-structure ignition is envisioned by direct flame impingement, considering structures, when ignited, part of the fuel system.
Direct Flame Contact Ignition Occurrence in WUI Real Fires
Investigations of WUI fires provide empirical data on the relative importance of the different structure ignition pathways. Despite the inherent complexity of fire forensic studies (due to the general disappearance of evidences and the difficulty in stablishing common surveying criteria), some historical analyses have revealed that in a WUI fire, structures ignite mainly through firebrands and embers that come in direct contact with combustible elements. The impact of nearby flames (either due to radiation or due to direct contact) from the main fire front or other adjacent fuels is difficult to distinguish without close witnesses. However, in cases where houses are particularly adjoining the wildland, or where structures within a development are close to each other or close to ornamental vegetation, flame impingement to structures is expected.
Some recent studies of WUI fires in Europe, North America, and Australia support this information. The Pedrógão Grande fire (Portugal, June 2017) inspection of affected homes allowed the identification of ignition pathways in 1003 structures (Viegas et al. 2017). In 63% of the cases, firebrands were directly responsible for structure ignition, while the rest of homes were impacted by nearby flames, coming either from the main fire (23%) or from burning fuels (13%) or structures (1%) in the vicinity. Data in terms of flames impact (either direct flame contact or radiation) could not be extracted due to the lack of reliable information.
The Fort McMurray fire (Alberta, Canada, May 2016) destroyed more than 2400 structures. In those, few ignitions could confidently be attributed to direct flame contact or radiant heat from the forest. Firebrands were mainly responsible (either directly or indirectly) for house losses (Westhaver 2017). Investigations of “The Trails Community” largely affected by the Witch-Guejito fires (California, USA, October 2007) provided some data on the ignition pathways of 74 destroyed homes (Maranghides and Mell 2011). Those located in the interior of the development (49%) were lost as a result of exposure to firebrands generated from burning wildland, residential vegetation, and structural fuels. Direct flame impingement from structure-to-structure ignition was not identified as a significant contributor to fire spread within the community. In structures at the perimeter (i.e., lots that had direct contact with the wildlands), residential vegetation carried the fire to the structure. However, no more details were provided regarding the primary ignition cause of those homes. Waldo Canyon fire (Colorado, USA, June 2012) investigation evidenced direct flame contact as ignition mechanism from wildfire-to-home spread (Quarles et al. 2013). Insufficient separation from homes to the trees and overhanging decks over slopes facilitated flame contact in some of the surveyed homes. Moreover, 3–6-m spacing was detected in places where home-to-home ignition occurred being direct flame contact hence plausible.
A survey conducted after the Canberra fire (ACT, Australia, January 2003) identified a high percentage of damaged and destroyed houses without direct flame or radiation attack from the main fire front (Leonard and Blanchi, 2005). Nevertheless, direct flame contact was observed as a prevailing mechanism in some house-to-house ignitions. Comparatively, the “Black Saturday” fires (Victoria, Australia, February 2009) investigation reported more flame involvement from bush vegetation. In the Victorian fires, 2118 houses were destroyed. It is estimated that 13% ignitions were due to bush-fueled flame contact (Leonard et al. 2009).
Direct Flame Contact in WUI Codes and Standards
In some WUI fire-prone areas, the building construction follows codes and standards with particular provisions aimed at hardening structures to WUI fire exposures (e.g., ICC 2018; CBSC 2016; NFPA 2013; SAC 2009). Those are mainly focused on thermal radiation and flame contact exposure.
Regarding the flame contact component, there are some available standards particularly devoted to evaluate building materials and assemblies’ performance to this type of exposure through experimental test procedures. As an example, the California Building Code (CBSC 2016) details the compliance in terms of materials and construction methods for exterior wildfire exposure of new buildings located within a WUI fire area in Chapter 7A. Standards of quality gathered in this section (i.e., State Fire Marshal standards) detail fire resistance tests consisting on direct exposure of different material assemblies to diffusion flames coming from a gas burner. Heat outputs from the burner and times of exposure vary depending on the tested elements (e.g., exterior wall sidings and sheathings, windows, decks, roof eaves, etc.) and range between 80 to 300 kW of heat rate and 3 to 10 min of exposure.
Similar standards issued by other agencies evaluate the performance of WUI building materials following similar test procedures as the abovementioned SFM standards. Several ASTM testing methods (e.g., ASTM E108, E2632, E2707, E2726, E2886, and E2957) involve direct flame impingement simulating external wildfire exposure to analyze fire performance of exterior wall assemblies, deck materials, eaves, roof coverings, vents, etc.
Australian standard (SAC 2009) considers direct flame contact in the worse (i.e., extreme) level of bushfire attack categories (i.e., Flame Zone). The nominal radiation threshold above which direct flame contact is imminent is set at 40 kW/m2. According to this standard, building elements being threatened at this level such as roofs, decks, or exterior walls have to be large flame approved. The method for determining the performance of these external construction elements when exposed to direct flame impingement simulates exposure from the fire front or large burning items such as other buildings or adjacent isolated trees and shrubs. It involves standard heating regimes for about 30 min (as specified in AS 1530.4 (2005) and ISO 834-1), rather than a transient high-temperature exposure lasting a few minutes.
Field surveys of WUI fires have led to empirical proofs of the most vulnerable elements to direct flame contact. The most significant accepted evidences are that siding materials often ignite to either direct flame contact or radiant heat exposures (Hakes et al. 2017). Although it cannot be said that there are specific mitigation strategies for one or other ignition pathways, it is clear that flames exposure (either radiation or contact components) to structures can be minimized designing noncombustible area hence defensible spaces around homes. In addition, ignition-resistant construction in terms of configurations, materials properties, fire-retardant treatments, etc. will reduce ignition occurrence in case of flame contact exposure.
- Alexander ME, Stocks BJ, Wotton BM, Flanningan MD, Todd JB, Butler BW, Lanoville RA (1998) The international crown fire modelling experiment: an overview and progress report. In: Proceedings of the second symposium on fir and Forest meteorology. American Meteorological Society, Boston, pp 20–23Google Scholar
- Babrauskas V (2003) Ignition handbook: principles and applications to fire safety engineering, fire investigation, risk management and forensic science. Washington Fire Science Publishers, IssaquahGoogle Scholar
- California Building Standards Commission, CBSC (2016) California existing building code. California code of regulations, title 24, part 10. International code council, WashingtonGoogle Scholar
- Cohen JD (2000) Preventing disaster: home ignitability in the wildland-urban Interface. J For 98:15–21Google Scholar
- Cohen JD (2008) The wildland-urban interface fire problem: A consequence of the fire exclusion paradigm. Forest History Today, fall: 20–26.Google Scholar
- International Code Council, ICC (2018) International wildland–urban Interface code, 2018Google Scholar
- Leonard J, Blanchi R (2005) Investigation of bushfire attack mechanisms involved in house loss in the ACT Bushfire 2003. CSIRO Manufacturing & Infrastructure Technology, HighettGoogle Scholar
- Leonard J, Blanchi R, Lipkin F, Newnham G, Siggins A, Opie K, Culvenor D, Cechet B, Corby N, Thomas C, Habili N, Jakab M, Coghlan R, Lorenzin G, Campbell D, Barwick M (2009) Victorian 2009 bushfire research response – final report. Bushfire CRC, Aspendale, 195~ppGoogle Scholar
- Maranghides A, Johnsson E (2008) Residential structure separation fire experiments. NIST TN, Gaithersburg, p 1600Google Scholar
- National Fire Protection Association, NFPA (2013) NFPA 1144- standard for reducing structure ignition hazards from wildland fireGoogle Scholar
- Porterie B, Nicolas S, Consalvi JL, Loraud C, Giroud F, Picard C (2005) Modeling thermal impact of wildland fires on structures in the urban interface. Part 1: radiative and convective components of flames representative of vegetation fires. Numer Heat Transfer, Part A 47:471–489CrossRefGoogle Scholar
- Quarles S, Leschak P, Worley K, Brown R, Iskowitz C (2013) Lessons learned from Waldo canyon. Insurance Institute for Business & Home Safety, Richburg, South California, 48 pp. https://fireadapted.org/wp-content/uploads/2018/06/waldo-canyon-report.pdf
- Standards Australia Committee, SAC (2009) Australian standard–construction of buildings in bushfire-prone areas AS 3959-2009. Standards Australia, 110 ppGoogle Scholar
- Tran HC, Cohen JD, Chase RC (1992) Modeling ignition of structures in wildland/urban interface fires. In: Proceedings of the 1st international fire and materials conference, 24–25 Sept 1992, Arlington. Inter Science Communications Limited, London, pp 253–262Google Scholar
- Trelles J, Pagni PJ (1997) Fire-induced winds in the 20 October 1991 Oakland hills fire. In: Hasemi Y (ed) Fire safety science–proceedings of the fifth international symposium, Melbourne, pp 911–922Google Scholar
- Viegas DX, Almeida MF, Ribeiro LM (2017) O complexo de incendios de Pedrógão Grande e concelhos limítrofes, iniciado a 17 de junho de 2017. Centro de Estudos sobre Incêndios Florestais. Universidade de Coimbra, 238 ppGoogle Scholar
- Westhaver A (2017) Why some homes survived: learning from the Fort McMurray wildland/urban interface fire disaster. Institute for Catastrophic Loss Reduction, Toronto, 81 ppGoogle Scholar