In order to apply the snowmelt-runoff model (SRM) or any snowmelt model operationally, a logical progression is required from utilization in the pure simulation mode when all data are known to the pure forecasting mode when no future data are known. Significant progress has been made and results are presented which include pure simulation, simulation when the actual output data are unknown, simulation when estimated or forecasted snow cover input data are employed, simulations or forecasts with updating using observed streamflow, and first attempts at true forecasts. Based on these results, an objective method for forecasting with SRM is being developed.