National Water-Quality Assessment (NAWQA) Program Home Page

National Map with
Links to Study Units

Project Publications

Project Main Page

USGS National Water Quality Assessment Data Warehouse

MA-100 --- Abstract

Pesticides in Surface Water in the Connecticut,
Housatonic, and Thames River Basins, 1992–95

Water-Resources Investigations Report 98-4247

By Marc J. Zimmerman

From March 1993 through September 1995, surface-water-quality samples were collected routinely from streams in the Connecticut, Housatonic, and Thames River Basins study unit of the U.S. Geological Survey’s National Water Quality Assessment Program (NAWQA). The streams sampled in this study were selected to reflect typical water-quality conditions in urban, agricultural, and forested settings. One hundred thirty-nine of these samples were analyzed for a wide array of pesticides. The length of time during which sample and data collection occurred ranged from several days for intensive studies of the interactions of ground water with surface water to several weeks for high-flow and low-flow investigations. A longer-term study was conducted at a single urban site that was sampled weekly in the spring and summer of 1993 and 1994 and less frequently in autumn and winter of those years. The relatively large number of samples collected at this single site is the likely reason for the detection there of 22 different pesticides or their metabolites, usually at low concentrations.

Although some herbicides and insecticides were found in streams draining both urban and agricultural settings, different groups of pesticides were usually associated with these settings; in particular, insecticides were more commonly detected in urban than in agricultural samples. Pesticides were rarely detected in streams draining forested settings.

The most commonly detected pesticide, atrazine, was virtually ubiquitous; it was found in samples from all land-use and basin categories. Atrazine was detected most frequently in streams draining agricultural basins. Metolachlor was also detected at more agricultural than urban sites. Most of the samples in which carbaryl, diazinon, and prometon were detected came from urban streams.

Concentrations of pesticides determined using a solid phase-extraction methodology did not exceed Maximum Contaminant Levels (MCL) or Health Advisory Levels (HAL) as defined by the U.S. Environmental Protection Agency. Commonly detected pesticides and their highest concentrations were: atrazine (1.10 micrograms per liter), carbaryl (3.2 micrograms per liter), diazinon (0.210 micrograms per liter), metolachlor (0.910 micrograms per liter), prometon (0.140 micrograms per liter), and simazine (0.690 micrograms per liter). The highest concentrations of atrazine, metolachlor, and simazine were detected in samples collected in agricultural basins and the highest concentrations of carbaryl, diazinon, and prometon were detected in samples collected in urban basins. A single atrazine concentration (4.5 micrograms per liter) exceeding the MCL was detected in a sample analyzed using an enzyme-linked immunosorbent assay (ELISA). (It should be noted that the McLand HAL are set for finished drinking water and exceeding them does not mean that a standard was violated.)

The highest estimated total daily loads of pesticides were associated with elevated streamflow in storm runoff during the late spring to early summer period, shortly following pesticide application. Some high loads, however, were also found later in the growing season. Estimated loads in excess of 4 kilograms per day were determined for the Connecticut River at Thompsonville, Conn., and the Naugatuck River at Beacon Falls, Conn.

Detection of pesticides in streams throughout the summer months during base flow periods in urban and agricultural basins suggests a ground-water transport mechanism, although atmospheric transport may also play a role. The repeated application of pesticides (especially insecticides) during the growing season in urban areas, however, may contribute to the detection of these compounds.

Sampling focused on annual periods of normal high and low streamflow, which may have affected data interpretation; additional sampling of stormwater runoff during normal low-flow periods would provide valuable data, as would frequent or repeated sampling of more sites. Use of carefully designed, automated sampling programs accompanied by a sample screening method, such as ELISA, should result in the collection of additional important information while keeping costs down. Sampling for pesticides in rainfall also would further contribute to our understanding of pesticide distribution.

For additional information contact: