WATER SUPPLY FORECASTING IN REAL-TIME1

ABSTRACT: A strategy is developed for making seasonal water supply forecasts in real time. It links the traditional regression based forecasting techniques to real-time data acquired by systems such as the Soil Conservation Service SNOTEL and the U.S. Bureau of Reclamation Hydromet networks. The concept is based on interpolating between the forecast values obtained by using real-time data in the forecast equations that bracket the real time. Different interpolation procedures were examined and the procedure for calculating confidence intervals about the forecast estimates is developed. The entire conceptual procedure is demonstrated using data from the Reynolds Creek, Idaho, experimental watershed maintained by the USDA-ARS Northwest Watershed Research Center in Boise, Idaho.