Date of Award

1996

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Engineering Science (Interdepartmental Program)

First Advisor

Richard L. Bengtson

Abstract

Geographic Information System (GIS) technology was used to develop a local scale screening procedure for assessing shallow ground water's vulnerability to nitrate contamination. The procedure was based on the spatial distribution of ten key physical and chemical properties of soils. Both flow characteristics and attenuation capabilities of soils were incorporated into the assessment. Land surface areas were classified by the GIS to show the level of vulnerability present at a site as a result of existing natural conditions. The procedure produced a Vulnerability Index (VI), and resulted in VI ranging from 4.98 to 6.72 for the study area (with ten indicating the highest vulnerability). Procedure evaluation involved shallow water well sampling for nitrates, and the correlation of VI values with nitrate concentrations. Four GIS-based approaches were used to evaluate the procedure's reliability. Vulnerability was examined as a point value (at the wellhead), and within a wellhead buffer area of three different sizes: 400- and 800-meter radius, and a custom-shaped area upgradient of the wellhead of equal area to the 800-meter area. Sampling was conducted in the Big Creek sub-basin of the Tangipahoa River in Southeastern Louisiana. Wells located on sixteen dairies and ten residential (non-dairy) sites were tested for nitrates and phosphates. Bi-weekly sampling was conducted during the summer, 1993 and winter, 1994. Water quality results indicated that 84.7 percent of all samples had ground water nitrate concentrations less than 1 mg-N/L. Nineteen of 26 sites had concentrations that remained below this level. Higher concentrations were observed at seven sites. The highest concentrations observed were 12.95 Mg-N/L (dairy site) and 8.57 mg-N/L (residential site). Phosphates were not detected at any sampling site. Comparisons of VI to nitrate concentrations by linear regression analysis indicated the procedure was moderately reliable. Area assessments consistently provided better correlation between vulnerability and nitrate occurrence than did the point assessment. The best correlation between VI and maximum nitrate concentrations (R of 0.454) was obtained for the custom buffered approach.

Pages

310

Share

COinS