The Bushfire Behaviour and Management Group University of Melbourne

The Bushfire Behaviour and Management Group at the University of Melbourne aims to provide a strong

scientific basis for the prediction of wildfire behaviour and the impact of wildfires in order to identify

management strategies aimed at reducing the risk of damage to human values while maintaining

environmental values in the landscape. To do this our work covers three primary research areas: fire

behaviour; landscape flammability; and risk modelling.

In natural and Wildland-Urban Interface environments, the science of fire behaviour and fire dynamics is

used to predict when and how wildfires spread through the landscape and the associated impacts to

communities. This science is used to develop computer models to simulate wildfire spread, providing

information to fire fighters and the general public about wildfires in their area. Our research focuses on

understanding how fuel, weather, terrain and spatial scale interact and influence ignition, spread and fire

intensity. We use a range of methods including retrospective analysis of past wildfires and computer

modelling. Development of physics-based and operational models requires the knowledge of many

characteristics of wildfires at different scales. Fire experiments are the only way to do this in a safe and

quantitative way. We specialise in undertaking laboratory experiments, field observations of fires and

prescribed burns to improve our understanding of fire behaviour, the impact on structures and provision

of essential information for future fire modelling.

The nature of fire behaviour varies greatly through space and time; some years high levels of rainfall may

result in few wildfires with little impact on human values. During other years, in particularly during periods

of prolonged drought, the likelihood of uncontrollable wildfire is high. Therefore, an understanding of the

processes which drive spatial and temporal variation in landscape flammability is critical to for assessing

fire risk. We study landscape flammability from two different perspectives: the development of methods

to extend laboratory research to large landscape scales; and the development of methods to extract

information from fire observations and large scale measurements to determine the roles of key

environmental drivers (i.e. dryness, vegetation type). This enables the processes that drive fire behaviour

at large scales to be better understood, and identification of potential levers that managers can use to

influence future fire risk.

Risk modelling provides a methodology for investigating alternative burning strategies, management

scenarios and decisions. We examine scenarios such as the impact of individual fires; to the effect of

climate change on fire regimes. Our risk modelling approach provides a systematic method for assessing

trade-offs between different management strategies, with estimates of residual risk and cost-

effectiveness across a range of values and assets types. Our work in this field aims to provide robust

calculations of risk (at the local and landscape scale) to assist decision making and inform on-ground

management for a range of assets including people, property, economic, environmental, cultural and

infrastructure. Through new software developed by our team we also have the capacity to integrate

future climate scenarios, enabling the assessment of risk under a changing climate. We use a range of

modelling approaches to quantify risk, specialising in fire simulation, spatial data analysis, Bayesian

Network Modelling and fire regime simulations.