HR: 0800h
AN: C21A-1064    [Abstracts]
TI: Spatial structure differences in snow depth distributions between forested and open sites
AU: * Deems, J S
EM: deems@cnr.colostate.edu
AF: Watershed Science, Colorado State University, Fort Collins, CO 80523-1482 United States
AU: Elder, K J
EM: kelder@fs.fed.us
AF: USDA Forest Service, USFS Rocky Mountain Research Station 240 West Prospect Road, Fort Collins, CO 80526-2098 United States
AU: Fassnacht, S R
EM: srf@cnr.colostate.edu
AF: Watershed Science, Colorado State University, Fort Collins, CO 80523-1482 United States
AB: Knowledge of spatial patterns of snow accumulation is required for understanding the hydrology, climatology, and ecology of mountain regions. Spatial structure in snow accumulation patterns changes with the scale of observation, a feature that has been characterized using fractal dimensions derived from LiDAR data. Previous work has shown that the fractal structure of snow depth distributions differs between sites with different vegetation and terrain characteristics. Attempts to predict snowpack structure in forested areas are typically made based on measurements from unforested areas since field measurements are easier to obtain in unforested areas and remotely sensed snow depth retrievals are complicated by the presence of forest canopy. This study examines two subsets of a single site, and demonstrates differences in spatial structure between forested and unforested areas. Forest and alpine tundra subsets were examined from the NASA Cold Land Process Experiment (CLPX) Fraser-Alpine study site. The forested area shows an essentially random spatial distribution, that could be characterized with a probability distribution function (PDF) based on a field estimate that details the statistical distribution of depth. The tundra area, however, shows strong spatial structure, and is the source of the majority of spatial pattern observable in the full study site. Using a PDF to characterize snow amounts in tundra/open areas is likely to encounter significant scale issues, and the choice of resolution (dictated by satellite sensor resolution, DEM resolution, field survey resolution, etc.) is likely to strongly affect the estimates of snow accumulation and amount of spatial variability. These results indicate that snow cover models need to focus on deterministic approaches in unforested areas, while a less complex statistical approach may be adequate in forested areas. These two approaches themselves are scale dependent, but defensible at the hillslope to basin scale.
DE: 0736 Snow (1827, 1863)
DE: 0758 Remote sensing
DE: 4440 Fractals and multifractals
DE: 4475 Scaling: spatial and temporal (1872, 3270, 4277)
SC: Cryosphere [C]
MN: Fall Meeting 2005