Dynamic development of the 2013 Aberfeldy fire

Abstract

The 2013 Aberfeldy fire, in eastern Victoria, exhibited unexpectedly rapid fire spread after its ignition on 17 January. In the first 16 hours after ignition at around 11:30am, the fire spread approximately 30 kilometres towards the southeast and burned 20,000 hectares. Much of this rapid fire spread occurred at night. Also of significance was that the fire burnt in complex topography, with slopes greater than 20?, and that in a number of instances the fire exhibited lateral spread; that is, spread in a direction almost perpendicular to the prevailing wind direction. In this paper we consider the development of the 2013 Aberfeldy fire in light of a number of recent insights into the dynamic behaviour of fires burning in rugged terrain under strong winds. In particular, we draw on findings that initially related to the 2003 Canberra fires to explain the occurrence of the lateral spread. More recent numerical and experimental work will also be discussed in this context. The observed development of the fire is also considered in the context of the traditional fire spread modelling approach, whereby the fire is assumed to propagate at a quasi-steady rate of spread modulated by the combined effects of wind and terrain. In particular, vector fields representing the expected rate and direction of fire spread are derived using traditional methods and are applied across the fire affected landscape. Infrared linescans, showing the development of the fire at several junctures in time, exhibit fire propagation patterns that are difficult to reconcile with the rate of spread vector fields at several key locations. Specifically, the development of the fire into narrow fingers propagating laterally across the tops of slopes in the immediate lee of a ridge line do not match with the predictions derived from traditional quasi-steady fire propagation models. This pattern of fire spread does however match quite well with that shown to occur when fires are affected by vorticity-driven lateral spread (VLS). The VLS phenomenon arises due to a three-way interaction between strong winds, steep terrain and a significant fire in the landscape. Analysis of the Aberfeldy fire showed that the observed spread adhered to VLS occurrence thresholds established through consideration of other notable fires, and through numerical modelling and experimental analysis. The strong pyroconvection and pattern of smoke observed in association with lateral spread events in the Aberfeldy fire were also consistent with that observed in other confirmed VLS events. The findings of this case study indicate that there were readily identifiable dynamic processes that drove the development of the Aberfeldy fire. These processes were essentially the same as those that drove the development of the 2003 Canberra fires and a number of significant fires in the ensuing years. The findings further suggest that research into the dynamic drivers of extreme bushfires, which now spans about 10 years, can offer significant improvements in the way such fires are modelled operationally. Some suggestions about how these improvements could be implemented within the working environment of a Fire Behaviour Analyst are offered.