U.S. Geological Survey Toxic Substances Hydrology Program--Proceedings
of the Technical Meeting, Colorado Springs, Colorado, September 20-24, 1993,
Water-Resources Investigations Report 94-4015
Modeling Statistical Relations Among Shallow Ground-Water Quality,
Human Activities, Land Use, and Thickness of the Unsaturated Zone on Long
Island, New York
by
Paul E. Stackelberg (U.S. Geological Survey, West Trenton, New
Jersey) and David A.V. Eckhardt (U.S. Geological Survey, Ithaca, N.Y.)
Abstract
Water-quality data from 90 shallow observation wells in five areas of
differing land use in Nassau and Suffolk Counties were used to develop statistical
models that relate shallow ground-water quality to variables representing
human activities and hydrogeologic conditions. The five study areas represent
the full range of land uses in the two-county area and lie along the regional
ground-water divide, where ground-water has a vertically downward component
and thus recharges deep zones of the aquifer system.
Maximum-likelihood logistic-regression analysis of explanatory variables
that describe the degree and type of human activities and hydrogeologic
conditions at the 90 well sites was used to develop statistical models that
can predict the probability of the presence of contaminants in shallow ground
water within the two-county area. Variables found useful for representing
human activities and hydrogeologic conditions include (1) population density
within a 1/2-mile radius of each well, (2) percentages of land-use categories
within the same area, and (3) thickness of the unsaturated zone at each
well site. One model that relates the probability of the presence of volatile
organic compounds (VOC's) at concentrations of at least 1 microgram per
liter to commercial and high- and medium-density residential land was selected
to illustrate model application and testing procedures. Resulting predictions
were compared with an independent set of water-quality data from 378 wells
in Nassau and Suffolk Counties. The comparison indicates that VOC's are
more prevalent in areas with relatively little commercial and dense residential
land than predicted by data from the five study areas.
|
|