A Simple, Computationally Efficient Distributed
Snowmelt Runoff Model for Use on Large Basins
Albert Rango1 and Kaye Brubaker2
There is a strong trend toward development of physically based, fine-grid distributed models for
estimates of snowmelt runoff under varying hydrological conditions. The physical basis of these
models is an improvement over simple degree-day models; however, the complexity necessary in
their design has also led to major disadvantages in their practical application. These models have
difficulty in producing continuous streamflow hydrographs for entire years or snowmelt seasons
because of complex linkages between various hydrological processes that are amplified by the fine
grids used and the need for large basins, e.g., >1000 km2. It is easier to produce model-generated
snow water equivalent maps for a basin, and these are sometimes used as the end product. As a result,
the model produced output is often compared with other simulated data and not with actual stream-
flow. Because a number of authors have shown that complex models do not outperform simple
hydrological models, it seems that some range of snowmelt runoff models is necessary for various
applications. That range should include simple degree-day models on one end and complex fully
distributed physical models at the other extreme. In between, there may be a valuable niche to be
filled by a simple, computationally efficient but distributed snowmelt runoff model able to produce
continuous hydrographs with a daily timestep on large river basins. Such a model has been designed
as a modification to the widely used degree-day-based Snowmelt Runoff Model (SRM). The so-
called radiation version of SRM is a distributed model using hydrological response units (HRU). It
requires input of temperature, precipitation, snow-covered area, and net radiation (or cloudiness,
humidity, and pressure to estimate net radiation) to HRUs based on elevation and aspect. It can use
remotely sensed snow-cover extent data directly from existing World Wide Web sites. As with the
original SRM, the model produces continuous hydrographs and operates very rapidly on a personal
computer. Although more complex in a number of ways than the original SRM, it is also somewhat
easier to apply because some elements of user hydrological judgment have been eliminated. Exam-
ples of input requirements, output on a variety of basins, and resulting hydrological responses to
climate change scenarios are provided for the radiation version of SRM. This new version of SRM
provides an intermediate type of snowmelt runoff model that may find significant applications in
snowmelt streamflow forecasting and evaluations of hydrological response to climate changes. It
would be useful to perform an intercomparison of snowmelt runoff models covering a full range of
complexities for streamflow forecasting or climate change evaluations.
1 Hydrology Laboratory, USDA/ARS/BARC-W, Building 007, Room 104, Beltsville, Maryland 20705, USA
2 Department of Civil Engineering, University of Maryland, College Park, Maryland 20742, USA
90
Previous Page
