The great escapes

By Dr Neil Burrows. This article first appeared in Issue Three 2017 of Fire Australia.

Preventing burns from escaping and becoming a damaging bushfire is of great concern to  fire managers. Escapes, as with most accidents, are usually the culmination of many factors, but in this article I take a look at two significant burns that escaped and identify three primary causal factors.

Bushfires develop speed, energy and killing power from the amount and structure of live and dead vegetation that burns, which builds up and becomes increasingly hazardous over time. Done at the right spatial and temporal scales, prescribed burning reduces the overall flammability and quantity of fuel across a region, thereby reducing the intensity and speed of bushfires. This makes them easier and  safer to suppress, and less damaging to the built and natural environments. Even under severe fire weather conditions, if fuel load and fuel flammability are low to moderate as a result of prescribed burning at the right scale, there will almost always be somewhere on the perimeter that the fire can be safely and effectively attacked. If fuel loads are low, the portion of attackable perimeter can increase significantly when fire weather conditions abate. 

Rotational prescribed burning has been implemented over much of the 2.5 M ha of the south-west Australian forests since the late 1950s. Over that time, several thousand burns have been carried out, amounting to some 14.4 M ha of prescribed burning. Over the same period, about 1.2 M ha was burnt by bushfire. Even with this extensive burning experience, and even though the program has a firm scientific underpinning, the practice is not without risk. Ensuring a prescribed burn does not escape is paramount to firefighter safety, community safety and community confidence in the fire management agency. In Western Australia, successive agencies responsible for carrying out prescribed burning on public lands have a good track record, with only about 2% of burns escaping. In most cases, the escaped fires have been quickly contained, with little damage done. However, in the 2011–12 fire season, several prescribed burns escaped. One (the Margaret River fire) caused significant property damage (32 homes, nine chalets and four sheds) and another (the Milyeannup fire) was at the time the largest bushfire in the southwest region since 1961. 

Because of the high property losses, the Margaret River fire was the subject of a special inquiry commissioned by the Western Australian Government and led by MJ Keelty AO. In handing down his findings, Mr Keelty identified a range of factors, acting together or alone, that contributed to a) the prescribed burn escaping and b) the consequent damage caused by the escape. Because there was little harm done, the Milyeannup fire was not the subject of a special inquiry, but was the subject of an internal report by the Department of Environment and Conservation, the responsible management agency. 

The Margaret River and Milyeannup fires were quite contrasting—not only in size, but in their location and community impact. The Margaret River fire was relatively small (about 3,400 ha) but resulted in significant property losses, whereas the Milyeannup fire was large (approximately 52,000 ha) but caused no significant losses. Fortunately, there was no loss of human life in either incident. The environmental damage of bushfires is less tangible, but is usually commensurate with the scale and intensity (severity) of the fire. These ecosystems are well adapted to fire, and will recover in time.

Inherent risk

Prescribed burning is inherently risky because of uncertainty surrounding interactions between weather, topography, fuels, fire behaviour and human behaviour. The postincident investigations of these escapes list many strategic, tactical, planning and procedural factors that contributed to burn security being compromised and the fires escaping. The reports and inquiries made sensible recommendations, most of which have or are being adopted by the fire and land management agencies concerned. While there were differences between the two prescribed burns in terms of the vegetation and fuel types involved, the complexity of the burns, the size of the escaped fires, the terrain and the consequent damage done, three primary causal factors interacted to compromise burn security in both cases. These were: 

• burning long-unburnt vegetation
• diverse vegetation (hence fuel types) within the burn
• burning in a warming, drying weather phase (spring).

No prescribed burning is without risk of fire escape, but the risk escalates with increasing vegetation age. This is because fuel load and fuel depth increase, fuel structure becomes more flammable, and in forests, ember sources (such as the amount of loose, flammable bark on trees) increases. Prior to core ignition of aerial burns, the usual practice is to create a surface fuel-free edge of at least 50–100 m deep around the inside perimeter of the burn, effectively increasing the width of the fuel break. This is usually done by ground crews walking the burn perimeter, which is a mineral earth track or road, with drip torches (flamethrowers are used in some cases), lighting up according to a pre-determined ignition pattern (‘edging’). It is important that edging is done under relatively high fuel moisture and mild weather conditions conducive to mild, containable fire behaviour. 

So that fire behaviour does not escalate, old fuels need to be burnt under much milder conditions than young fuels, that is under a lower fire danger index and soil dryness index (higher fuel moisture). By design these conditions reduce the amount of available fuel, thereby reducing the potential fire behaviour and risk of fire escape. However, it results in burn patchiness, or incomplete combustion of the fuel complex in multiple dimensions. 

The residual unburnt fuel can occur both in the horizontal and vertical dimensions of the fuel complex. In  the vertical plane, long-unburnt, deep-surface litter fuels with moist profiles may not burn down to mineral earth under mild prescribed burn conditions, with only the drier top centimetre or so of the fuel bed burning. Edging is designed to extend the width of the fuel break around the burn, so improving burn security. However, if the surface (litter) fuels are not mostly burnt to mineral earth, clearly the edge is not going to function as a fuel break when the residual duff fuel dries later on. In the horizontal plane, low-lying vegetation, such as along creek lines, vegetation on south-facing slopes and areas of dense canopy cover, may not be sufficiently dry to burn at all or will burn patchily. 

In addition to variation in fuel moisture, variation in fuel structure can result in fuel patchiness. Less flammable vegetation types, such as sparse low heaths or open canopy woodlands, may not burn under mild prescribed burning conditions. The extent and nature of the flammability differential across the landscape—due to variability in fuel moisture content and fuel structure—will largely be a function of the burn season and the fire danger index. In a Mediterranean-type climate, stronger differentials exist in spring, so burns are more likely to be patchy. 

The nature of the vegetation mosaic itself will reflect variability in fuel properties, which is more pronounced under mild burning conditions (i.e. low fire danger index) but dissipates with the onset of summer and a rising fire danger index. Patchy or mosaic burning  is often ecologically desirable, because it provides habitat diversity, refugia and soil protection, thereby reducing the acute impacts of fire on wildlife and the environment and enhancing post-fire recovery. 

Patchy burning of old fuels under mild conditions in spring will more often than not result in the later re-ignition of the landscape as residual unburnt fuels dry out with the progression of spring to summer. In a large burn, there is almost always an ignition source, such as a smouldering log, spar, stump, peat or duff (‘sleepers’), which may go undetected for days or weeks but then flare up on a warm windy day, reigniting the unburnt, residual fuels. Fuel that didn’t burn during the initial prescribed burn completed days or weeks earlier is now flammable, because it has dried and the fire danger index has increased. 

The severity of fire behaviour and likelihood of the fire escaping will depend on the size, age and fuel properties of the unburnt patch, its proximity to the burn boundary, the fuel condition of the edge, the weather conditions and the detection and suppression response. In longunburnt vegetation, the fire behaviour on re-ignition, including spotting potential, will be significantly greater for the same weather conditions and patch size than if the unburnt patch was younger fuel. Not only is the burn more likely to escape by spotting over the burn boundary, but because of the residual surface fuel on the edge of the burn—which has now dried—the fire can run across the ‘burnt’ edge and break containment lines. The scenario is made a lot more dangerous and difficult for firefighters if the fire breaks out into adjoining long-unburnt fuels, as was the case for both the Margaret River and Milyeannup fires. 

Reducing the risk

The single greatest strategic measure that can be taken to reduce the risk of well-planned prescribed burns escaping, and reducing damage potential should they escape, is to maintain much of the landscape in a young fuel condition by maintaining an active prescribed burn program. The more prescribed burning that is done, the easier and safer it is to do prescribed burning, and the reverse is true. It is easy to take prescribed fire out of a landscape, but very difficult to put it back. In south-west jarrah forests, for example, this means keeping about 45—50% of the landscape at a fuel age of less than about six years and about 80% less than about eight to ten years. 

Attempting to burn old fuels surrounded by old fuels is costly and high risk, especially in spring. If burning old fuels is unavoidable, and you have to start somewhere, then it may be preferable to burn in autumn after some rain. You should take extra measures to  reduce the size and location of unburnt patches within the burn, and take measures to ensure the edging is of a sufficiently high standard to stop flame spread, such as mechanical treatment of edge fuels or multiple ignitions. On the other hand, young fuels can be burnt in spring with much a lower risk of re-ignition and escape than older fuels.

As mentioned above, burn patchiness can be desirable for ecological reasons. For burn security, bushfire mitigation and ecological benefits, many scattered, small, unburnt patches are preferable to a few large, clumped, unburnt patches, which are likely to be lost following re-ignition and create headaches for fire managers. In south-west Australian forests, achieving and maintaining fine-scale patchiness is more likely when burning younger fuels, and less likely as fuels age. Relatively cheap technology—such as drones mounted with infrared cameras—can be used to locate unburnt patches and hot spots to assist with further risk assessment and burn security measures.

There is no evidence that the managed fire regime outlined above will have deleterious impacts on the forest biota. In fact, by the nature of fire behaviour, it is not only safer to carry out prescribed burning regularly (so burning younger fuels), but it enhances the likelihood of the development of a fine-scale mosaic of vegetation at different
structural stages. On the other hand, infrequent burning (in older fuels) is risky, costly, dangerous and will result in more large summer bushfires—leading to homogenous vegetation structures and less habitat diversity.