This project developed a dynamic GIS model of agricultural nitrate contamination of Sussex County, Delaware, groundwater. The large geographic scale (3 major watersheds) and long time-frame (50 years) of this analysis precluded any effective use of conventional event-driven groundwater simulation models.
Poultry houses were digitized from successive updates of USGS 7.5-minute quad maps since 1945. Cropland was extracted from a 1984 land-use/land cover mapping of Delaware derived from Landsat MSS satellite imagery. STATSGO digital soils data were obtained from the U.S. Soil Conservation Service. 80-meter resolution digital elevation data for the entire county were obtained from USGS. Additional GIS layers were derived from USGS DLG and Census TIGER data sources.
These data were integrated in GRASS to generate a 50-year dynamic simulation of agricultural nitrate delivery, movement, diffusion and removal from Sussex County's groundwater. The simulation model yielded a plausible mapping of nitrate contamination plumes for the entire county.
1992-93 well test results from approximately 1,300 private wells in Sussex County were georeferenced through the Delaware Electric Cooperative's customer database. These data were used to validate the results of the groundwater contamination simulation. Approximately 25 percent of all private wells in the county exhibit N-NO3 contamination in excess of the EPA's maximum safe standard of 10 mg/l. A spatial regression model, accounting for spatial lag pathologies and heteroscedasticity with respect to well depth, explains over 40 percent of the total variation in nitrate levels in the well test data.
A survey-based conjoint analysis of individuals' preferences for alternative water quality and cost scenarios yielded plausible benefit valuations for various levels of nitrate contamination abatement. Costs of a centrally-located poultry manure pelletizing plant in Sussex County were determined from existing engineering data. The economic costs of various reductions in field applications of poultry manure were compared against the economic benefits, by Census block group, of nitrate contamination abatement as predicted from the GIS groundwater simulation model.
The principal conclusions from this analysis are: Diversion of poultry manure from land application to a centrally-located pelletizing plant in the county will generate positive net social benefits for Sussex County as a whole. However, both the costs and the benefits of any groundwater quality improvements resulting from reductions in land applications of poultry manure would be unevenly distributed across the county. Even under a total cessation of land applications of manure across the county, the groundwater simulation model indicates that some Census block groups would see little or no reduction in nitrate concentrations in groundwater, even over a 50-year time horizon.
Mackenzie, J. 1995. 30-meter digital elevation models for Delaware: 51 7.5-minute quads. GIS data product distributed by the Spatial Analysis Lab, College of Agricultural Sciences, University of Delaware (http://bluehen.ags.udel.edu/spatlab).
Sparco, J. and J. Mackenzie. 1994. Agricultural land uses and groundwater contamination: a spatial regression analysis. Paper presented at the Soil and Water Conservation Society of America meetings, Fort Worth, TX, August, 1995.
____. 1995. Marginal valuation of health-related attributes of groundwater using conjoint analysis. Paper presented at the Northeast Agricultural and Resource Economics Association meetings, Burlington, VT, June, 1995. Revision in preparation for the Journal of Environmental Economics and Management.
____. 1995. A GIS analysis of the effects of agricultural land uses on nitrate contamination of groundwater. Submitted for selected paper session, American Agricultural Economics Association meetings, August, 1996. In preparation for the Journal of Soil and Water Conservation.