Surface water is a valuable resource. Variability in its availability can seriously affect communities that depend upon that resource for health and economic well-being. The analysis of rainfall, streamflow and other data assists in assessing the variability of the resource, and in determining the resulting reliability of supplies derived from that source. Streamflow records are rarely long enough to confidently assess supply reliability, so methods such as rainfall-runoff modelling and stochastic data generation have been developed to extend streamflow records.

SMEC has been at the forefront of development and application of such techniques within Australia and internationally. The models used for rainfall-runoff modelling include the Sacramento model, the Monash model and AWBM. Stochastic data generation has adopted a range of models from simple Markov chain models to multi-lag auto-regressive moving average (ARMA) models.

Application of such models requires a review of data quality, application of suitable models taking account of seasonality and spatial trends. It is also necessary to understand the proposed use of the data so that methods are adopted that address the critical flow characteristics.

Typical recent projects include: 

• application of rainfall-runoff models and regression models to extend streamflow records for the Darwin water supply headworks system in tropical north Australia 
• stochastic generation of 10,000 replicates of streamflow, each of 100 years, at 9 sites for the investigation of the Sydney water supply headworks system 
• use of regression models to extend five streamflow records to 40 years length for modelling of the five irrigation headworks storages for the Kukadi Irrigation system, Maharashtra State, India