|KUMAR, MUKESH - Duke University
|DOZIER, JEFF - University Of California
Submitted to: Advances in Water Resources
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/13/2013
Publication Date: 3/26/2013
Citation: Kumar, M., Marks, D.G., Dozier, J., Reba, M.L., Winstral, A.H. 2013. Evaluation of distributed hydrologic impacts of temperature-index and energy-based snow models. Advances in Water Resources. 56:77-89. DOI: 10.1016/j.advwatres.2013.03.006.
Interpretive Summary: Temperature index snow model is evaluated against a full energy balance snow model to determine the hydrologic necessity for more complex representation of snow. Both snow models are coupled to a fully distributed, physics based hydrology model (Penn St. Integrated Hydrology Model – PIHM). Neither the over-calibrated or standard calibrated T-index models are effective.
Technical Abstract: Proper characterizations of snow melt and accumulation processes in the snow-dominated mountain environment are needed to understand and predict spatiotemporal distribution of water cycle components. Two commonly used strategies in modeling of snow accumulation and melt are the full energy based and temperature-index based methods. Here we evaluate the distributed hydrologic impacts of considering these modeling strategies in conjunction with a physics-based hydrologic model (PIHM). Specifically we answer following four questions: a) Are predictions of snow accumulation, distribution and melt in different land covers sensitive to the choice of temperature index or energy-based snow modeling approach, b) How does consideration of different snow melt and accumulation strategies affect streamflow prediction at the watershed outlet, c) What is the consequent impact of different snow cover and duration predictions on distribution and timing of soil moisture, groundwater, and evapotranspiration, and d) How is the net catchment water balance affected by different predictions of snow cover extent and rates of depletion from the three model configurations? Results illustrate marked differences in prediction of snow accumulation and melt from the two modeling strategies, even when they performed reasonably well at the observation site. Comparison of how different snowcover representations simulated by the three model configurations affect watershed response illustrate the sensitivity of streamflow, spatio-temporal distribution of soil moisture, groundwater, evapotranspiration, and integrated catchment water balance to patterns of snow deposition and melt. Though both the Is+P and calTi+P models adequately predict streamflow in the calibration year (WY06), the uncalTi+P model was unable to predict streamflow in either year. Is+P model performed the best in prediction of streamflow in both years. Both Ti models were unable to simulate the known spatial detail in patterns of SWE and hydrologic states. This experiment demonstrates the hydrologic benefit of coupling a physics-based snow and hydrologic model and suggests that if a temperature index snow model is selected, it must be carefully calibrated in both time and space.