Next generation models for predicting the behaviour of bushfires: Challenges and prospects — ASN Events

Next generation models for predicting the behaviour of bushfires: Challenges and prospects (#22)

Graham R Thorpe 1 , Khalid Moinuddin 1 , Andrew Ooi 2 , Daniel Chung 2
  1. Victoria University, Melbourne, VIC, Australia
  2. Department of Mechanical Engineering, University of Melbourne, Melbourne, VIC, Australia

The application of physics-based models of the rate of spread of bushfires has been restricted, in part, by the availability of computing resources. New developments in computer hardware and software are lifting these restrictions. The computer games industry demands a very high rate of data processing by graphical processing units, and this paper discusses how this computer hardware can be exploited to develop massively parallel processing. Computer programming protocols are being developed to make use of the new hardware.

When we observe the behaviour of bushfires we notice several features that are apparently random and do not repeat themselves. Flames, for example, are unsteady. The physical phenomena that engender this unsteadiness are captured by the equations that govern the behaviour of bushfires. This makes the solution of the equations both difficult and approximate. This paper will report on opportunities to make use of the most contemporary approaches that are available to solve the equations.

The phenomena that govern the rate of spread of bushfires occur over a very wide range of length scales. The methods used to model bushfires account for scales that vary from those associated with the smallest eddies - a fraction of a millimetre, up to several kilometres that are associated with the terrain and atmosphere. This paper will discuss how this wide range of length scales can be accounted for without compromising accuracy or requiring excessively long computing times.

The outcome of the research will be an increased capacity to accurately predict the behaviour of bushfires under a very wide range of conditions.