North American Experience

On average about 7 firefighters die each year in the western United States, (Thoele 1994). Most are single fatality incidents but in 1994, 25 ground and aviation firefighter deaths occurred - one of the worst seasons in 61 years (Thoele 1994). Since 1985 the number of people deploying fire shelters has dramatically increased. This has been attributed variously to: drought conditions; unpredictable weather; low experience levels of crews, supervisors and management; and, to increased reliance on fire shelters (Putnam 1991). There have been at least 25 fatalities involving fire shelters. In many cases firefighters have had insufficient time to deploy their tents before being engulfed in flames. There is little doubt that a protective shelter will not provide protection at extreme levels of fire intensity, but advocates argue that the tent is a last resort that gives the firefighters a chance to survive when there are no other options for escape. The question is then raised: does the provision of safety gear promote more aggressive fire suppression and more risk-taking by firefighters?

Those questions will be examined on two recent incidents in the US.

The Butte Fire August, 27 - 29 1985

On the Butte fire, protective shelters have been attributed to saving the lives of at least 60 of 73 firefighters who were overrun by a fast moving crown fire in 18 to 20 m high lodgepole pine (P contorta) (Jukkala and Putnam 1986, Rothermel and Mutch 1986). This incident certainly boosted the statistics in favour of fire shelter effectiveness but reports on the fire avoided the question as to why the firefighters were placed in danger in the first place. The fact that firefighters must carry shelters and that substantial safety zones are constructed for the possible deployment of these shelters implies a greater level of aggression, and may inevitably lead to supervisors ignoring early warnings and unconsciously placing firefighters at additional risk. Certainly my analysis of the situation, seen from an Australian context, raises questions about basic fire-fighting strategies employed on this fire and which may reflect a different cultural attitude towards fire suppression and the safety of firefighters.

The following strategies and tactics employed on the Butte fire can be compared with normal tactics used on the fire line in Australia:

The fire had reached 8000 ha.

On the day of the incident crews were attempting to burn out a 1200 ha catchment using an indirect fire line established along the ridge of the catchment.

Around 11 km of fire line had been bulldozed along the ridges. These fire lines included several locations where safety zones (where a firefighter in normal clothing can survive) of approximately 90 x 120 m had been cleared in dense 20 m high lodgepole pine on the ridge top The drainage was south-facing (equivalent to north-facing in Australia) with south facing aspects having slopes in excess of 20 leading up to the fire line on the ridge top.

Forecast weather conditions were not provided but the following were considered typical of late afternoon conditions in the area: Temp 25 C, RH 12%, Wind Speed 20 km h-1 from the south. (Rothermel and Mutch, 1986). These conditions convert to a McArthur Forest Fire Danger Index (FFDI) of 30 (Very-High) using a drought factor of 10 to reflect the substantial drought conditions that were reported in the area. The actual conditions at the time of the incident 14 km south of the fire area were - Temp 30 C, RH 6%, Wind Speed 25 km h-1(Alexander 1991); this converts to FFDI 58, Extreme. Considering that wind speed was expected to increase in the afternoon when the inversion broke-up the increase in fire danger index from 30 to 55 should not have been surprising.

Although highly active fire behaviour was observed when the fire was 3 km away and at least 2 hours before the fire reached the ridge line, it appears that no efforts were made to remove firefighters from the ridge and that the planned escape route was to the safety zones.

While there are certainly specific differences in the behaviour of fires in north American and Australian fuel types, the generalised condition forecast by fire danger systems is quite similar. For example, the rate of spread of 1.44 km h-1 forecast by the McArthur Forest Fire Danger Index for a crown fire in a 20 t ha-1 eucalypt fuels in undulating country is quite comparable with the average rate of spread of 1.33 km h-1 during the major run of the fire. Much faster rates of spread would be predicted for, and were recorded on, the steep up-slope runs towards the ridge line. Given a similar scenario in Australia with forecast northerly wind on a north-facing aspect with slopes up to 20 , I consider a fire controller would be criminally negligent to place men on a ridge-top fire line under conditions of increasing fire danger. Besides which there would be little point until it was clear that conditions had abated and were suitable to carry out burning-out operations in the evening at a low to moderate fire danger.

The low humidity of 12% together with substantial drought meant the potential for extreme fire weather always existed and only required a small increase in wind speed to be realised.

The large safety zones turned out to be inadequate for the conditions and the firefighters needed the additional protection of the shelters. I consider that these firefighters were unnecessarily exposed to high-intensity fire for around 70 minutes (undoubtedly complicated by the large wind-rows of piled debris on the perimeters of the safety zones) and could have been avoided by sound planning of expected fire behaviour under forecast weather. Although it was recognised that firefighters should not be placed in such a threatening situation (Rothermel and Mutch 1986) there appears to be little published information on the decision making which allowed this situation to develop.

South Canyon Fire 4-7 July 1994

Fourteen firefighters were killed in a single incident on the South Canyon fire on July 6, 1994. The investigation of the fatalities on the South Canyon fire (see Rosenkrance et al. 1994) found that 12 of the firefighters who perished did not have enough time to open their shelters and get under them. Two of the firefighters who perished had fully deployed their shelters but still succumbed to smoke inhalation and heat. Twenty six firefighters were cut-off by flames from the designated safety zones and survived by fleeing across the ridge-top to a lee-slope. Eight firefighters who were in a different location when the incident occurred managed to reach designated safety areas on burnt ground and deployed their shelters without burns or smoke inhalation. One of this group survived unscathed without deploying a fire shelter.

The investigation team found that attitudes of both firefighters and fire management contributed significantly to the incident. Specific warnings of high winds and extreme fire behaviour, although known, were not conveyed to the firefighters and, as a result, firefighters remained on a dangerous section of the fire line right up until the fire broke away beneath them on steep slopes under the influence of strong winds associated with a frontal passage.

The investigation into the incident found that there was no evidence that fire shelters encouraged tactical risk taking. However there is no doubt that many critical safety factors were over-looked and this may reflect an organisational attitude promoting over-aggressive fire suppression. I believe that there is circumstantial evidence at least that the carrying of fire shelters will encourage greater risk taking although this proposition is strongly denied by advocates of fire shelters ( R Mangan pers. comm.) .