Research topics for Honours and postgraduate students

1.    Optimisation of prescribed burning for the reduction of bushfire hazard

Supervisors: Prof. George Milne (milne@csse.uwa.edu.au), Dr Paul Johnston (paulj@csse.uwa.edu.au), School of Computer Science and Software Engineering

 

Uncontrolled bushfires are an inevitable occurrence in forested regions due to ignitions that occur by natural and anthropogenic means.  The factors that influence the severity of forest fires are the weather and the quantity of fuel.  Fires in extreme weather conditions in heavy fuel can not be directly fought once established, whereas fires under the same conditions with a reduced quantity of forest fuel may be contained.  Prescribed burning reduces the quantity of forest fuel available by deliberate ignition of forest fires under mild weather conditions.  If a fire occurs later under extreme conditions in the same place, it can be more easily contained than if no prescribed burning had occurred.  The aim of prescribed burning is to minimise the risk of large forest fires subject to a number of constraints including the limited manpower of a fire agency.  The other constraints are to minimise the risk that a prescribed fire escapes control or damages assets, to minimise the impact of smoke on cities during prescribed burning and sometimes to achieve a particular intensity during the prescribed fire.  A fire spread simulator can be used to estimate the area burnt and intensity of an individual fire.  Using the fire spread simulator, climate data and maps of forested areas and ignition probabilities, the challenge is to develop prescribed burning policies that will be most effective in reducing the risk of large forest fires.

 

2.    Real-time forecasting of bushfires from satellite imagery

Supervisors: Prof. George Milne, Dr Paul Johnston

 

The UWA bushfire simulator can rapidly run simulations of bushfire spread over the landscape given ignition locations, vegetation maps and weather information.  Landgate provides real-time observations of fire via satellite detection of currently burning bushfires and image processing to determine regions burnt by fire (fire scars).  We also have access to current weather recorded at automatic weather stations.  The combination of real-time data and rapid simulation will be combined to provide forecasts of fire position into the future and simulations can be re-calculated as new data becomes available. The project will develop software to import geo-referenced data into the bushfire simulator and forecast fire position over the next few hours to few days.