Coupled fire-atmosphere simulations of three Australian fires where unusual fire behaviour occurred.

Predicting where and how a fire will burn is critical information for mitigating the impacts of bushfires and minimising risk at fuel reduction burns. Firefighter entrapments and fatalities occur mostly at fires that display rapid changes or fluctuations in fire activity. In this thesis, I explore several of the factors that lead to rapid changes in fire behaviour. Understanding these factors is necessary in order to produce accurate fire predictions, which are critical for fire-fighter safety and effective operations. Weather is a primary driver of fire activity; consequently, meteorological information is a key input for anticipating fire behaviour. At present, weather forecasts focus on near-surface conditions; but fires and the atmosphere are three dimensional, and dynamical interactions occur that can have a dramatic influence on fire behaviour. However, these fire-atmosphere interactions are poorly understood due to their complex nature and the difficulty of collecting observational data from a bushfire. In order to further understanding of dynamical interactions between a fire and the surrounding atmosphere, we have simulated three Australian fires where unexpected fire activity occurred, using the coupled fire-atmosphere model WRF and SFIRE. The coupled simulations have been run in feedback on and feedback off mode in order to assess the impact that the fires have on their surrounding atmosphere. The results show significant changes to the mesoscale atmospheric structure as result of the energy released by the fire. Computational fire behaviour models are being used by fire managers in real time and this use will grow in the future. The question is, given we know that fires affect the surrounding atmospheric flow; what weather inputs should the fire models of the future use? The Australian fire science community is currently presented with the opportunity and the challenge to design, develop, and implement fire behaviour simulation models that contain appropriate and comprehensive meteorological inputs. The results presented in this thesis are thought provoking for the current approach to fire weather forecasts and for the use and development of computational fire simulations in the future.Thesis (Ph.D.) -- University of Adelaide, School of Mathematical Sciences, 201