2006 Philadelphia Annual Meeting (22–25 October 2006)

Paper No. 10
Presentation Time: 10:55 AM

PRINCIPAL VOLCANIC AQUIFERS OF WASHINGTON, IDAHO, AND HAWAII — WATER QUALITY OVERVIEW AND LOGISTIC REGRESSION MAPPING OF NITRATE AND ATRAZINE


HUNT, Charles D., Jr, U.S. Geological Survey, 677 Ala Moana Blvd Suite 415, Honolulu, HI 96813-5415, RUPERT, Michael G., U.S. Geol Survey, 201 W. 8th Street, Pueblo, CO 81003-308 and FRANS, Lonna M., U.S. Geological Survey, 934 Broadway Ave. Suite 300, Tacoma, WA 98402, cdhunt@usgs.gov

The U.S. Geological Survey National Water Quality Assessment (NAWQA) is conducting “principal aquifer” studies to interpret water quality at a regional scale within distinct geohydrologic settings. The Western Volcanics study includes three NAWQA study units with extensive basalt aquifers: Central Columbia Plateau–Yakima Basin, Washington; Upper Snake River Plain, Idaho; and Oahu, Hawaii. Agricultural lands in all three study units received fertilizers, pesticides, and intensive irrigation for most of the 20th century. The volcanic-rock aquifers are susceptible to contamination because of thin or well-drained soils, largely unconfined aquifer conditions, and high permeability and ground-water velocities analogous to karst limestone. Numerous agricultural and industrial chemicals have been detected in ground water, though most concentrations are below human-health guidelines. Nitrate and the herbicide atrazine were prominent contaminants in all three study units, and solvents and fumigants also were detected widely on Oahu. Logistic regression models were developed to predict the probability of detecting nitrate concentrations above 2 milligrams per liter and atrazine concentrations above 0.001 microgram per liter. Variables differed slightly between models but included the study unit, well depth, percent agricultural land within 500 meters of a well, percent row crop land, percent fallow agricultural land, soil organic matter, soil clay content, and soil slope. Model performance was good, with the chi-squared p-value (calculated from the log-likelihood ratio of the entire model) being less than 0.0001 for both nitrate and atrazine. The regression models can be used as mapping tools to extrapolate sampling results to nearby unsampled areas having similar agricultural and chemical-use practices.