The role of extreme value analysis to enhance defendable space for construction practice and planning in bushfire prone environments

Preparation of defendable space for buildings of both existing and planned urban development in bushfire prone areas relies on critical assessment of potential fire conditions. This paper offers new insights into the application of three extreme value assessment methods to the input parameters of both MacArthur and Project Vesta fire behaviour models for the determination of the defendable space and other fire safety measures. Weather data for various durations from 21 New South Wales fire weather districts are processed to derive the fire behaviour model input parameters such as the forest fire danger index and the fuel moisture content. These parameters are then subjected to probabilistic and recurrence modellings using three extreme value analysis methods. The results are evaluated to determine the most appropriate approach. It is found that the recurrence trends can be modelled with log functions with good correlation coefficients. The models are used to predict 1:50 year recurrence values of forest fire danger index which are compared with the existing policy settings. The results indicate the need to revise the policy settings for some of the weather districts.