Operations research for decision support in wildfire management

The February 2009 ‘Black Saturday’ bushfires resulted in 173 fatalities, caused AUD$4 billion in damage and provided a stark reminder of the destructive potential of wildfire. Globally, wildfire-related destruction appears to be worsening with observed increases in fire occurrence and severity. Wildfire management is a difficult undertaking and involves a complex mix of interrelated components operating across varying temporal and spatial scales. This thesis explores how operations research methods may be employed to provide decision support to wildfire managers so as to reduce the harmful impacts of wildfires on people, communities and natural resources. Some defining challenges of wildfire management are identified, namely complexity, multiple conflicting objectives and uncertainty. A range of operations research methods that can resolve these difficulties are then presented together with illustrative examples from the wildfire and disaster operations research literature. Three mixed integer programming models are then proposed to address specific real-world wildfire management problems. The first model incorporates the complementray effects of fuel treatment and supression preparedness decisions within an integrated framework. The second model schedules fuel treatments across multiple time periods to maintain fire resistant landscape patterns while satisfying various ecological and operational requirements. The third model aggregates fuel treatment units to minimise total perimeter requiring management.