|Institution:||University of New South Wales|
|Department:||Biological, Earth & Environmental Sciences|
|Keywords:||Modelling; Bushfire; Climate change|
|Full text PDF:||http://handle.unsw.edu.au/1959.4/54460|
Despite considerable environmental and social impacts of bushfire, there are no high resolution, spatially continuous projections of bushfire risk across Australia under climate change which take into account the interplay between rising carbon dioxide levels and vegetation growth. This thesis aims to address this gap. First, observations are analysed for the presence of trends in fire weather conditions across Australia. Significant increases in average (90th percentile) fire weather were observed at 16 (24) of 38 stations across Australia, with none recording a significant decrease. Second, future fire weather is projected in eastern Australia using skill- selected global climate models. Global climate models project strong increases in fire weather conditions in the southeast, including a lengthened fire season, but little change or decreases in the northeast. A regional climate model is then evaluated for its ability to simulate historical fire weather in southeast Australia. It simulates observed spatial patterns of fire weather well, but has an average positive bias in annual cumulative FFDI of 381. A simple model of fuel load is next developed for use in a land surface model, using net primary productivity as a proxy. This model accounts for the effects of climate and carbon dioxide fertilisation. No trends in fuel load are evident in simulations of historical fuel load over Australia. Finally, these models are used to project both fire weather and fuel load in Australia under climate change and associated increasing carbon dioxide. Fuel load is consistently projected to increase in temperate, grassland and subtropical areas of Australia. The sign of change in fire weather projections is sensitive to model selection. However, the magnitude of increases is much larger than that of decreases and all models suggest a lengthening of the fire season. Overall this research suggests bushfire risk is likely to increase in Australia under climate change, with increased load likely to have a greater impact in load-limited areas such as grasslands. In contrast, fire weather increases are likely to be of greater significance in temperate forested areas in the southeast and southwest. Two key uncertainties in the evolution of Australian fire regimes are the trajectory of regional rainfall change and the impact of carbon dioxide fertilisation on vegetation growth.