Water Resources Research Act Program

Details for Project ID 2008MT183B

Student Fellowship: Predictive Modeling of Snowmelt and the Hydrologic Response:

Institute: Montana
Year Established: 2008 Start Date: 2008-03-01 End Date: 2008-12-01
Total Federal Funds: $1,000 Total Non-Federal Funds: Not available

Principal Investigators: Tyler Smith

Project Summary: In Montana and much of the Rocky Mountain West, the single most important parameter in forecasting the controls on regional water resources is snowpack (Williams, Cline et al. 1999). In the mountainous areas of Montana, in particular, snowmelt is the key driving force in downstream supplies of water; water is deposited at much higher volumes in the form of snow in mountainous areas, which results in greater accumulation and storage and diminished evaporation (Hauer, Stanford et al. 2007).

Much work has been done in the past to help characterize and identify the controls on snowpack dynamics, from how a snowpack undergoes ripening to the eventual release and hydrologic pathways of the water associated with the snowpack itself. However, the current models are still largely lacking in their ability to relate the first-order controls on snowmelt runoff to individual watershed characteristics. Additionally, little work has been performed considering uncertainty representation in coupled snowmelt/hydrologic models.

Perhaps equally important to any predictive modeling approach is creating a model that transcends beyond the empirical and reaches into the physical. By recognizing the balance between necessary physical inputs and data limitations, it is my hope to better describe the amount of physical reality that is required to create a reliable model. While empirical models are often sufficient in some systems, they cannot predict a watersheds response to change (land use, vegetation, variable source area pollution, forestry practices, etc.) because they lack the physical basis.

The dilemma of balancing data input complexity and model output sophistication is as true as ever when considering the snowmelt models presently available. While current methods range from simple to complex, all tend to place focus on validation against measured watershed discharge. While discharge is no doubt an important parameter, little has been done in the realm of validation against other watershed properties/variables that also have value in the ultimate determination of a models predictive abilities. Because of the complexities associated with snowmelt mechanisms in this region, single parameter models (temperature-index approaches) do not adequately describe the intricacies of the process, whereas full-blown physics based models (energy balance approaches) often require more inputs than can be widely obtained.