A High-Resolution Distributed Snowmelt Model
in an Alpine Catchment
Michael Colee1, 2, Robert Harrington3, Thomas Painter1, 2, and Jeff Dozier1, 2, 4
Spatially distributed estimates of snow cover and snowmelt and understanding the processes govern-
ing them are increasingly important for a variety of applications from hydrologic and land use
change models to landscape evolution and climate change models. A physically based snowmelt
model, SNTHERM, is run for each 30-m pixel in the Tokopah Valley, an alpine catchment in the
Sierra Nevada of California. The catchment is approximately 1900 ha, roughly 75% of which is
above timberline. Soil and vegetation classifications of two sub-basins within the study area were
done; the soil classification was field based and the vegetation classification was based on aerial
photo interpretation. These surfaces were then used to select spectral endmembers for classification
of soil type and vegetation cover using a snow-free AVIRIS scene of the area. Three field surveys
during April, May, and June of the 1997 melt season obtained depth and density measurements of the
snowpack suitable for estimating the distribution and ablation of snow water equivalence (SWE) for
the basin. The April and May datasets are used here for model initialization and validation, respec-
tively. The initial snow surface was interpolated from a gridded field survey of 429 depth and 8
density samples distributed over the area with the krigging method. Initial snow-covered area (SCA)
was derived from TM data and albedo was derived from AVIRIS data. Clouds obscured the study
area in the April AVIRIS scene. Therefore, albedo values from proximal cloud-free areas outside the
study area were mapped into the basin. Temperature, relative humidity, and wind speed were distrib-
uted over the basin based on observed lapse rates using a least squares fit calculated at hourly
timesteps from three stations within the basin. Hourly radiation inputs to the model were calculated
with the topographically corrected clear-sky radiation model, TOPORAD, and adjusted for cloud
cover based on measured solar and longwave radiation from two sites in the basin. Model perfor-
mance was evaluated by comparing outputs at one month with the May field data set of 317 depth
measurements and 76 density measurements, AVIRIS-derived SCA and grain size, and TM-derived
SCA for that period. Temporal accuracy of the model was also coarsely assessed by comparing
model outputs to measured basin hydrographs over the course of the one-month run. The snowmelt
model was not coupled with a hydrological model, therefore routing of snowmelt out of the basin
was instantaneous. Understanding the spatial and temporal dynamics of a snowpack at fine resolu-
tions will be increasingly important for the study of climate change and its impact on the ecology and
meteorology of alpine areas.
1
Department of Geography, University of California, Santa Barbara, California 93106, USA
2
Institute for Computational Earth System Science, University of California, Santa Barbara, California
93106, USA
3 Department of Hydrology and Water Resources, University of Arizona, Tucson, Arizona 85721, USA
4 School of Environmental Science and Management, University of California, Santa Barbara, California
93106, USA
93
Previous Page
