Evaluation of Model Enhancements and Probabilistic Forecasting Techniques for The Snowmelt Runoff Model

Accurate streamflow forecasts are critical for the responsive management of water resource systems, which are designed and operated for the purposes of irrigation, flood control, and hydroelectric power generation. In the Western United States, water supplies are often derived form runoff due to snowmelt. The objective of this project is to develop an ensemble (probabilistic) prediction system for short to medium range streamflow forecasts (1 to 15 days). The hydrologic model used in this study is the Snowmelt Runoff Model (SRM), which is a conceptually-based model designed to simulate and forecast daily streamflow in mountain basins where snowmelt is a major contributor to runoff (Martinec et al., 1994). To optimize model efficiency and aid in its operational implementation, three enhancements have been made to the model. These enhancements are: 1) the use of an antecedent temperature index method to track snowpack cold-content and account for the delay in melt associated with diurnal refreezing, 2) the use of both maximum and minimum critical temperatures to partition precipitation into rain, snow, or rain/snow mixed, and 3) the stochastic modeling of forecast errors to generate streamflow ensembles, from which exceedence probabilities can be obtained. Results from retrospective model runs, using temperature and precipitation forecasts from NCEP GFS model (2000-2004) will be examined and used to identify the error in the streamflow forecasts. (Abstract only)