Impact of hydroclimate parameter uncertainty on system yield

Abstract

Water resource system behaviour is commonly evaluated using Monte Carlo simulation of synthetic streamflow and rainfall data. Fundamental to this process is the estimation of stochastic model parameters by calibration to short observed records. The uncertainty in key hydroclimate parameters, such as the mean, variance and serial autocorrelation, flows through to the simulated hydroclimate data. This study considers the subsequent influence of model parameter uncertainty on the calculated yield of a case study water supply system. A procedure is developed to estimate the system yield for individual samples of model parameters, providing a probability distribution of plausible system yields. The uncertainty in hydroclimate parameters is found to have an appreciable impact on the uncertainty in system yield – of the order of +/−10%. This uncertainty is likely to surpass, potentially by an order of magnitude or more, the influence of many choices made by water managers in both the modelling and operation of water resource systems.