Performance of an Energy-Budget and Temperature Index-based SNOW-17 Snowpack Models at SNOTEL Stations in the Carson River Basin

Research is being conducted at the Hydrology Laboratory of the NOAA National Weather service to determine the optimal transition pathway from conceptual to advanced energy-budget snow models for operational water resources forecasting. Research has shown that energy-budget snow models can generate more accurate simulations of snow variables than conceptual models in certain cases. However, energy-budget models are very sensitive to errors in the input forcing data. The purpose of our current work is to evaluate the performance of an energy-budget snowpack model driven by meteorological input data with various uncertainty levels to recommend acceptable and available meteorological data sources for lumped and distributed model applications. An energy-budget snowpack model driven by meteorological fluxes and the NWS snow accumulation and ablation model, SNOW-17, driven by temperature from the same data source are run at the same time period at SNOTEL stations in the Carson River Basin. The outputs from these models are compared to each other and to ground measurements. Sensitivity and error propagation tests on input and parametric data are conducted.

The humidity and wind field data are extracted from NCEP NARR. The operational surface shortwave radiation fluxes from Geostationary Operational Environmental Satellites (GOES) generated as part of the Global Energy and Water Cycle Experiment (GEWEX) Continental Scale International Project (GCIP) are used as incoming solar radiation. Two sources of surface long wave radiation are considered. One is the NCEP NARR data set. To supplement this, surface long wave radiation will be calculated using emissivity, air temperature and cloud cover. Two sets of surface temperature and precipitation data are used in the models. One set is extracted from NCEP NARR data, the other is the SNOTEL station measurements. Land surface parameters in the energy-based model are from LDAS. Results of the model comparisons and sensitivity analyses will be presented.