Research Letter 21

House versus Bushfire

It may surprise some to learn that the impact of bushfires on houses was first studied in Australia after a fire in the Beaumaris area of metropolitan Melbourne - in 1944 (see Barrow 1945). Beaumaris is now a well-established suburb but then it was semi-rural in character, “a bayside resort”. The main vegetation type was dense tea-tree (Leptospermum spp) with scattered groups of manna gum (Eucalyptus viminalis). Houses occurred in clearings in these thickets but the clearings were “densely covered with dry grass and bracken”. When two fires swept the area on the same day under a hot northerly wind, 58 houses were destroyed.

On Ash Wednesday 1983, a severe bushfire destroyed 234 houses in and around the township of Mount Macedon, not far from the current Melbourne airport. Wilson and Ferguson (1986) studied the probability of house survival there in the face of the fire. First developed as a summer “retreat” for “the wealthy residents of nearby Melbourne”, Mount Macedon spread into the surrounding forest from the 1960s. Eucalypt forest abutted the town on three sides.

In January 1994, bushfires destroyed 206 houses in the wider Sydney area while 119 destroyed houses were examined in the Como-Jannali, Lane Cove and Cottage Point areas in the Sydney metropolitan area by Ramsay et al. (1996). Ramsay and his team had also investigated the effects on houses of the Otways fire in Victoria on Ash Wednesday 1983. There, they surveyed 1148 houses, 623 of which had been destroyed. The main vegetation types carrying the fires were forest, woodland and tall shrubland.

In this Letter, the conclusions of the three research publications reporting the fate of houses in the paths of bushfires are compared.

Wilson and Ferguson’s (1986) results showed the importance of fire intensity (in the surrounding vegetation) to house survival. The higher the intensity of the fire the greater the chance of destruction. This conclusion supported Barrow’s contention that “the nature of the surroundings” was most important because the nature of the surrounding vegetation could be expected to influence the potential fire intensity and extent of ember production. In the Otway Ranges fires the chance of destruction in the presence of trees was 10 times that with grassy vegetation (Ramsay et al. 1996); height of vegetation was also important in the Mount Macedon study (Wilson and Ferguson 1985).

Barrow (1945) emphasized the importance of embers as the cause of fires in houses and Ramsay et al. (1996) concurred:

In the majority of cases, ignition appeared to have been caused by burning debris, although radiant heat and flame played a significant role in cases where the houses were directly abutting dense undeveloped [unmodified, native?] vegetation”.

If embers - that arrive before flames and are present long after (Ramsay et al. 1996) - are a major cause of ignition of houses, then we might expect that an able bodied person present in the house during a fire would be able to put out smouldering materials. Indeed, Ramsay et al. (1996) found that the chance of saving the house in the Otway Ranges fire was 10 times greater when it was occupied. A similar result was recorded in the Macedon fire (Wilson and Ferguson 1985). However, whether or not you should stay in a house during a bush fire depends on the preparations you have made prior to the fire, your personal mobility and judgement, and the risks attending evacuation during a fire event (see Wilson and Ferguson 1984).

Barrow (1945) considered that “the details of construction” of the house were more important than “the materials used in the walls”. Certainly, embers lodging in nooks and crannies – the details – can be envisaged as being important, but other investigators found the material of construction to be a significant factor in house survival. Tiled roofs fared best in Sydney and the Otway Ranges while houses with brick walls survived better than those with timber ones both in these areas (Ramsay et al. 1996) and at Mount Macedon (Wilson and Ferguson 1985). Wooden shingles were the worst roof material encountered in the Mount Macedon study (Wilson and Ferguson 1985).

In the Otway Ranges fires, a house with a flat roof had a 25% greater chance of destruction than one with a pitched roof (Ramsay et al. 1996). However, flat roofs were associated with more secure houses in the Mount Macedon survey (Wilson and Ferguson 1985). This difference in response suggests that there could have been correlations between variables (roof pitch and type of wall cladding for example) that may not have been taken into account or that there were significant case-study differences (significant regional, wind, ember or fire-behaviour differences for example).

Some of the practical outcomes of these studies are obvious: keep fuel loads down, remove or modify ember-producing plants (e.g. stringybark trees), use the most effective materials for building, etc. There are guides to assess and to help. Wilson (1987) has produced a house survival guide[1] based on the Mount Macedon fire experience and is therefore designed for “houses exposed to a forest fire which is blown by strong winds on a very hot day”. Australian Standard AS 3959 is for construction in bushfire prone areas (Ramsay and Dawkins1993).


Barrow, G.J. (1945). A survey of houses affected in the Beaumaris fire, January 14, 1944. Journal of the Council for Scientific and Industrial Research [now CSIRO] 18, 27-37.

Ramsay, G.C. and Dawkins, D. (1993). Building in bushfire-prone areas. Standards Australia and CSIRO, Melbourne.

Ramsay, G.C., McArthur, N.A. and Dowling, V.P. (1996). Building in a fire-prone environment: research on building survival in two major bushfires. Proc. Linn. Soc. 116, 133-140.

Wilson, A.A.G. (1987). House Survival Meter. CSIRO Division of Forest Research, Canberra.

Wilson, A.A.G.and Ferguson, I.S. (1984). Fight or flee? – A case study of the Mount Macedon bushfire. Aust. For. 47, 230-236.

Wilson, A.A.G.and Ferguson, I.S. (1986). Predicting the probability of house survival during bushfires. J. Environmental Management 23, 259-270.

[1] See also the ACT Emergency Services Bureau web site: