Why wind shifts make fires spread faster

Reprinted with permission from CSIRO's "Onwood" Publication 15 - Summer 1996-97

Fire-fighters often find they have to cope with a big increase in a fire's rate of spread when the wind changes direction, even though wind speed stays the same. Recent research has explained this puzzling phenomenon.

Contrary to accepted fire spread theory, Phil Cheney and his fire research colleagues at the Division have found that the width of the fire front has a big influence on rate of spread. Previously it was thought that, under a given set of conditions, established fires of any size will move forward at the same pace, which will vary directly with wind speed. What a wind shift does is immediately widen the forward edge of the fire. According to Cheney, a typical outcome for a fire with a fairly narrow head and much longer flank is a three-fold increase in rate of spread.

Experiments that the CSIRO team conducted in open grasslands and woodlands in the Northern Territory revealed the importance of fire head width. After their initial work showed no clear relationship between fire spread and wind speed, they lit fires simultaneously at single points and along 100-metre and 200-metre lines. The widest fires always spread fastest, Cheney says. "And if you had a sudden shift that increased the width of the fire, it immediately went faster."

Experiments were then conducted in forests, with similar results. The research in both grass and forest indicates that rate of spread increases with fire head width up to a limit which varies with wind speed.

Under calm conditions, all fires move at the same slow speed regardless of width. Under light winds, typical of conditions for prescribed burning for fuel reduction, a fire only 20 metres wide might move at its "potential rate of spread". However, with strong winds, fire speed will keep rising up to a width of perhaps 200 metres.

Because they show there is no simple relationship between wind speed and fire spread during the early stages of fire growth, the findings pose problems for fire investigators trying to work out when a blaze started. "If you get three or four wind shifts in the first 10 minutes, a grass fire will probably reach its potential rate of spread in that time", Cheney says. "But if the wind is not very variable, like a sea breeze, then the fire might spread for an hour before it reaches its maximum speed."

Fires can take even longer to build up to their potential rate of spread in forests. Cheney says the CSIRO findings reinforce the importance of tackling a fire early; at that stage the target is not only small, it should also be spreading at well below its maximum rate for the wind speed.

The research has also confirmed the value of efficient attacks on the flanks of fires. "If you can keep the head of the fire narrow, you can keep it below its potential rate of spread even if you can't put it out", Cheney says. "Eventually it will run into a natural barrier or something that might allow you to catch up with it."

However, the findings highlight the need for care in lighting lines of fire along a break or fire trail when burning out in an attempt to secure a perimeter. "If the resulting fires can spread with the wind, or upslope, they will immediately travel at the potential rate of spread for the conditions", Cheney points out. "This could be very much faster than the behaviour that has been observed in a fire developing from a single point, and could cause problems for firefighters elsewhere along the fire trail."

Phil Cheney

Acknowledgment
This research required experiments with large fires (some more than 100 ha) burning under dangerous fire weather. It could not have been conducted without the collaboration and support of the Bush Fires Council of the Northern Territory and the Department of Conservation and Land Management, Western Australia. Our collaborators prepared experimental sites and kept the experimental fires under control.

"Onwood" is published by CSIRO Forestry and Forest Products
Editor: Mick Crowe
© Copyright - CSIRO Australia