Forecasts using SNOTEL data on November 1 and December 1 gave the best estimates of spring snowmelt runoff volumes almost as frequently as forecasts on March 1 and April 1 for the Upper Rio Grande basin in Colorado during 1983-2001. Forecasts for 2002, a record drought year, greatly overestimated flow volumes regardless of prediction month. Attempts at improving predictions using the El Nino/La Nina Southern Oscillation Index were not successful