Quantifying uncertainty in environmental flow assessments

Abstract

Environmental flow proposals are normally justified by predictions of geomorphological and ecological responses. Uncertainty in these predictions is widely acknowledged but rarely quantified. This paper uses a Monte Carlo analysis to quantify the effects of multiple sources of uncertainty on the environmental flow required to flush fine sediments from a gravel bed reach of the Goulburn River, Victoria. Sources of uncertainty are associated with sampling, measurement and selection of model parameters. The volume of water required to mimic the natural sediment flushing regime is estimated to be 160 Gl/year. The uncertainty analysis provides a 90% confidence interval for this volume of 10 Gl/ye..