Published works

Large cities around the world source their drinking water from forested catchments, which deliver water that is treatable at minimal costs.

These forests are often prone to wildfire, which tend to increase surface runoff, destabilize soils and trigger increased sediment delivery into water supply systems. The consequences of wildfire for water supply are a significant concern, with the potential for extended periods of water supply disruption, increased water treatment cost, and a demand for catchment restoration and investment in expensive treatment infrastructure. Moreover, with climate change, the likelihood of wildfire is increasing in many forest ecosystems making water supply systems increasingly exposed to post-wildfire contamination.

There are numerous uses for models that predict water quality impacts. On an operational level, models that reveal spatial variation in the erosion susceptibility provide catchment managers with a means to prioritize risk mitigation through fuel reduction or post-wildfire response. In such cases, spatial mapping of erosion risk in a relative sense may be sufficient to provide effective tools for allocating resources and mitigating risk in the areas of the catchment that are most likely to be producing sediment. When developing strategic plans and making decisions about the future management of a water resource, the demands on predictive models increase. For instance, a water supply agency may want to determine if there is a case for upgrading water treatment capability or adjusting the water supply network to reduce the likelihood of water supply interruptions due to wildfires. In this setting, a detailed understanding of risk is required for cost-benefit analysis to inform decisions about such investments.

In this study we are motivated by the need amongst water supply managers to know ‘For how long is it likely that my reservoir will be offline due to contamination by post-wildfire erosion?’. In addressing this question, we seek to provide the means for developing policy and investment strategies within environmental management frameworks that are underpinned by economics and risk. The specific objective of the study is to model how frequency and magnitude of erosion in headwaters translates to probability and duration of water contamination exceeding treatability thresholds at the water offtake. In this study we only consider sediment in our risk model because treatment facilities are often highly vulnerable to turbidity caused by suspended sediment.

Our paper outlines a parameter-reduction approach whereby empirical transfer functions are used to link a probabilistic model of sediment delivery from burned headwaters and a reservoir hydrodynamic model. With these transfer functions we produce a direct measure of risk (i.e. probability of consequence) and discuss practical questions around potential cost of the wildfire threat to water supply agencies and the capacity for catchment managers to mitigate such costs. Furthermore, we use the model to evaluate how different model components contribute to uncertainty in risk.

Download the full non-peer reviewed research proceedings from the Bushfire and Natural Hazards CRC Research Forum here.