Published works

The rate at which grass fires spread over flat terrain can be predicted with reasonable accuracy.  However, estimating the rate of fire-spread through other vegetative types is often more problematic.  In wooded areas, say, the localized wind speed is determined in part by ground cover, bushes, tree trunks and the forest canopy. This presents a significant challenge to the development of physical models of bushfires.

At first sight, forests appear to be quite random. However, the geometries of individual trees and forests share a remarkable similarity. Not only are the diameters of the branches of an individual tree distributed in fixed ratios, but the ratios of the tree trunks in a forest share the same distribution.  This uniformity has important implications for the modelling of bushfires. For example, in this work, we have generated fractal geometries of trees, and these have been used to estimate the velocity profiles of the wind as it flows through a stand of trees. The self-similarity between individual trees and forests provides an exciting new intellectual framework in which to develop the next generation of bushfire models.  Results obtained to date offer considerable promise.

The rate of spread of bushfires can be dominated by embers and firebrands that are conveyed ahead of a fire front. We need to understand how embers are generated and conveyed, and how they ultimately ignite fires. We have designed and constructed an ember generator to help elucidate these mechanisms. A feature of its design is that the firebrands, and the air that conveys them, exit the generator with velocities that are close to being uniform and this helps in our analysis. Results obtained to date clearly demonstrate the uniformity of the air velocity, and methods of measuring the dispersion of embers are underway.

If the rate of ignition of bushfires is to be modeled accurately we need to know the thermo-physical properties of vegetative materials. An extensive battery of experiments aimed at measuring the properties has been developed, and requirements for further testing will be established with end-users.

A conflict exists between modelling the physical details that govern the rate of spread of bushfires and the availability of computing power. As a result, one strand of our research into the next-generation models of bushfires is to develop improved computational methods. One such method has enabled us to model buoyancy-driven flows with great accuracy – it is known as explicit filtering of the discretised Navier-Stokes equation. Having proven the concept, the next stage of our work is to apply it specifically to the rate of spread of bushfires.

Infrastructure must be not only bushfire-resistant, but also aesthetically pleasing and economical to build. To be truly creative, designers benefit from having access to a deep understanding of the mechanisms that determine the rate of heat transfer between a bushfire and structures. Hence, a further strand of our integrated research program aims to develop simple-to-use formulae that will help designers of infrastructure at the urban-bushfire interface.