Forecasting Streamflow and Water Demand to Improve the Efficiency of Water Supply in South Australia

The South Australian Water Corporation supplies water to metropolitan Adelaide and rural areas through an extensive water distribution system that includes diversions from the River Murray. An Integrated Demand Management SCADA combines real-time field data with historical water demand, reservoir inflow, prevailing electricity tariffs and operator instructions, to optimise the operation of this system. To optimise further the headworks operation, SA Water required real-time forecasts of water demand and reservoir inflow to replace the historical statistics that were being used. This paper reports on the process of developing forecasting models and the lessons learnt in that process. Advanced statistical methods and recent research findings were applied to develop probabilistic models of water demand and reservoir inflow at daily, weekly and monthly time-steps. The models have been transferred into Windows-based software, providing SA Water with a highly advanced tool to manage bulk water supply in South Australia.