HR: 1340h
AN: H23D-1458    [Abstracts]
TI: Regional Assessment of the Relationship Between Landscape Attributes and Water Quality in Five National Parks of the Rocky Mountains
AU: * Nanus, L
EM: lnanus@usgs.gov
AF: United States Geological Survey, PO Box 25046, Denver, CO 80225 United States
AU: * Nanus, L
EM: lnanus@usgs.gov
AF: University of Colorado at Boulder, 450 UCB, Boulder, CO 80309 United States
AU: Williams, M W
EM: markw@snobear.colorado.edu
AF: University of Colorado at Boulder, 450 UCB, Boulder, CO 80309 United States
AU: Campbell, D H
EM: dhcampbe@usgs.gov
AF: United States Geological Survey, PO Box 25046, Denver, CO 80225 United States
AB: Atmospheric deposition of pollutants threatens pristine environments around the world. However, scientifically-based decisions regarding management of these environments has been confounded by spatial variability of atmospheric deposition, particularly across regional scales at which resource management is typically considered. A statistically based methodology coupled within GIS is presented that builds on small alpine lake and sub-alpine catchments scale to identify deposition-sensitive lakes across larger watershed and regional scales. The sensitivity of 874 alpine and subalpine lakes to acidification from atmospheric deposition of nitrogen and sulfur was estimated using statistical models relating water quality and landscape attributes in Glacier National Park, Yellowstone National Park, Grand Teton National Park, Rocky Mountain National Park and Great Sand Dunes National Park and Preserve. Water-quality data measured during synoptic lake surveys were used to calibrate statistical models of lake sensitivity. In the case of nitrogen deposition, water quality data were supplemented with dual isotopic measurements of d15N and d18O of nitrate. Landscape attributes for the lake basins were derived from GIS including the following explanatory variables; topography (basin slope, basin aspect, basin elevation), bedrock type, vegetation type, and soil type. Using multivariate logistic regression analysis, probability estimates were developed for acid-neutralizing capacity, nitrate, sulfate and DOC concentrations, and lakes with a high probability of being sensitive to atmospheric deposition were identified. Water-quality data collected at 60 lakes during fall 2004 were used to validate statistical models. Relationships between landscape attributes and water quality vary by constituent, due to spatial variability in landscape attributes and spatial variation in the atmospheric deposition of pollutants within and among the five National Parks. Predictive ability, model fit and sensitivity were first assessed for each of the five National Parks individually, to evaluate the utility of this methodology for prediction of alpine and sub-alpine lake sensitivity across the catchment scale. A similar assessment was then performed, treating the five parks as a group. Validation results showed that 85 percent of lakes sampled were accurately identified by the model as having a greater than 60 percent probability of acid-neutralizing capacity concentrations less than 200 microequivalents per liter. Preliminary findings indicate good predictive ability and reasonable model fit and sensitivity, suggesting that logistic regression modeling coupled within a GIS framework is an appropriate approach for remote identification of deposition-sensitive lakes across the Rocky Mountain region. To assist resource management decisions regarding alpine and sub-alpine lakes across this region, screening procedures were developed based on terrain and landscape attribute information available to all participating parks. Since the screening procedure is based on publicly available data, our methodology and similar screening procedures may be applicable to other National Parks with deposition-sensitive surface waters.
DE: 1800 HYDROLOGY
DE: 1803 Anthropogenic effects (4802, 4902)
DE: 1819 Geographic Information Systems (GIS)
SC: Hydrology [H]
MN: Fall Meeting 2005