A review of operational water supply forecasting techniques in areas of seasonal snowcover

Examples of seasonal and short-term operational water supply forecasting models in use during the last 20 years and the economic relationships between forecast accuracy and monetary benefits are reviewed and discussed. The importance of input variables associated with forecast models such as soil moisture, antecedent streamflow, wind and others, in addition to the common variables of snow-water equivalent and precipitation, is discussed. A comparison of a precipitation regression model, a water balance model, a pattern search model, a multiple linear regression model, and the Tangboran-Rasmussen Hydrometeorological Streamflow Prediction model is made for six forecast periods on the Middle Fork of the Boise River in southwest Idaho. Results indicate that multiple linear regression and pattern search produce the most accurate forecasts, although data requirements are more stringent when compared to the other models.