USGS Home
Contact USGS
Search USGS

U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2008–5025

Back to Table of Contents

Summary

Logistic regression was used to relate anthropogenic and natural factors to the occurrence of elevated nitrate concentrations in the State of Washington and develop models that estimate the probability of nitrate concentrations in ground water exceeding 2 milligrams per liter. Explanatory variables that were considered for inclusion in the models included hydrogeomorphic regions, soil characteristics, ground-water recharge rates, precipitation, land use, well depth, population density, and the amount of fertilizer applied annually near the well.

Two models that best estimate the probability of elevated nitrate concentrations were selected: one with and one without hydrogeomorphic regions. Other variables that were included in these models were well depth, percentage of agricultural land use within 4 kilometers of a well, average annual precipitation, population density, and soil drainage class. Relations between the estimated probability of detecting elevated nitrate and the actual percentage of wells with elevated nitrate concentrations were good for both models with R² values greater than 0.9.

Maps showing the estimated probability of detecting elevated nitrate concentrations indicate that the agricultural regions are most at risk followed by urban regions. Areas of the State that are not affected by agricultural or urban areas had much lower probabilities of detecting elevated nitrate concentrations. Maps also were generated that show the estimated depths to which wells would need to be drilled in order to have a 90-percent probability of obtaining water with nitrate concentrations less than 2 milligrams per liter. These depths frequently exceed 1,000 feet in the agricultural regions, although wells in urban areas generally would need to be drilled to depths in excess of 400 feet.

Back to Table of Contents