Pesticides in Streams of the United States--Initial Results from the National Water-Quality Assessment Program

U.S. GEOLOGICAL SURVEY
Water-Resources Investigations Report 98-4222
Sacramento, California, 1999

SITE SELECTION AND CHARACTERISTICS

Water samples were collected for pesticide analysis at 58 stream sites throughout the conterminous United States (fig. 2). These sites are a subset of the NAWQA surface-water monitoring sites. Eight to 12 stream sites were selected for regular monitoring of streamflow and general water chemistry for each of the study units shown in figure 2. One to 5 of these stream sites were chosen for more intensive sampling and pesticide analysis.

Two general types of sites were selected for intensive sampling--indicator and integrator sites. Indicator sites were chosen to represent water-quality conditions of streams in relatively homogeneous basins associated with specific environmental settings (land use and natural characteristics) that were targeted for study. Water quality at the indicator sites is influenced primarily by the targeted environmental setting, and in most cases, the targeted setting accounts for more than 50 percent of the drainage area. In contrast, integrator sites were chosen to represent water-quality conditions of streams with relatively large basins that are influenced by complex combinations of land-use settings, point sources, and natural influences typical of the region. Integrator sites generally are downstream from indicator sites and are located at key nodes in the drainage network. Results from the integrator sites provide a general check on the persistence of water-quality influences evident at the indicator sites; the results also can be used for water-budget and contaminant transport assessments. For a detailed description of the criteria and procedures used for site selection, see Gilliom and others (1995).

The 58 sites covered in this report consist of 37 agricultural indicator sites, 11 urban indicator sites, and 10 integrator sites. Sampling site locations are shown in figure 2 and site characteristics are summarized in and figure 3. In most of the agricultural basins, cropland and orchard-vineyard land account for more than 40 percent of the basin area and urban land accounts for less than 5 percent. The Merced River in the San Joaquin studyfied as cropland or orchard-vineyard land. Agricultural activity in the farmed area of this basin is very intense, however, with a variety of orchards, vineyards, and row crops; much of the streamflow during the growing season is from agricultural return flows. Agricultural indicator sites and integrator sites are classified according to the major crops grown within the drainage basins (table 1). About three-fourths of the agricultural indicator basins have major crops of corn, soybeans, alfalfa, and wheat and other grains, or some combination of these crops. This is consistent with the national distribution of row crops, with these four crops accounting for about 85 percent of the total row-crop area in the United States (Gilliom and Thelin, 1997). Other crops, such as peanuts, cotton, vegetables, field and grass seed, and sorghum, are represented by fewer agricultural sites. Thus, aggregated results from the agricultural sites are influenced most by pesticide use associated with the four major crops.

Water-quality conditions at urban indicator sites are affected primarily by urban, suburban, commercial, and industrial sources. Urban land uses account for more than 50 percent of the basin area at all but one of the urban indicator sites. The one exception is the Las Vegas Wash site, where more than 90 percent of the drainage basin consists of rangeland. For most of the year, water in the Las Vegas Wash consists almost entirely of effluent from a sewage treatment plant and of runoff from the urban area; the primary influences on water quality in the Las Vegas Wash, therefore, are from urban sources.

Whenever possible, the order of the sites in table 1 is retained in the figures and the tables throughout this report to aid in cross-referencing site information between tables and figures. Sites are identified in some figures and tables by a site code which consists of the study unit abbreviation and a part of the site name. For example, in the White River Basin study unit, the site on Sugar Creek is designated as "whit-sugar." The code for each site is shown in figure 2 and listed in table 1.

TARGET COMPOUND SELECTION AND CHARACTERISTICS

This report includes results of the analysis of water samples for 46 compounds--25 herbicides, 17 insecticides, 2 herbicide transformation products, and 2 insecticide transformation products. The target compounds are listed in table 2, with estimates of their national agricultural use and their primary uses. Compounds were selected for analysis on the basis of national agricultural and nonagricultural use, potential environmental significance, and chemical properties that allow analysis by gas chromatography/mass spectrometry (GC/MS).

The pesticides included in this report account for approximately 72 and 66 percent of national use of herbicides and insecticides, respectively, in terms of the mass used annually in agricultural applications during 1990-93 (Gianessi and Anderson, 1996). These pesticide-use totals do not include the use of inorganic pesticides, such as sulfur and copper, or biological pesticides or the use of oil as an insecticide. The target compounds include 15 of the top 25 herbicides and 15 of the top 25 insecticides used in agriculture in the United States during this period. No fungicides are included among the target compounds. Fungicides constituted approximately 6.5 percent of total agricultural use of pesticides in the United States during 1990-93 in terms of the mass applied annually (Gianessi and Anderson, 1996).

Pesticide use in nonagricultural applications in the United States is not well documented, but data are available that provide information on the relative amounts of different pesticides used on lawns and gardens during 1989-90 (Whitmore and others, 1992). In general, the pesticides discussed in this current report account for a lower proportion of nonagricultural use than agricultural use with 6 of the top 20 insecticides, 6 of the top 20 herbicides, and none of the top 12 fungicides used on lawns and gardens included in the target compounds. The relatively low coverage of pesticide use in nonagricultural settings should be kept in mind in the discussions of results from urban indicator sites. Several other pesticides with high nonagricultural use were measured at these sites using a different analytical method (high-performance liquid chromatography, or HPLC); the results will be analyzed when the data are available.

For most of the agricultural indicator and integrator sites, the target pesticides account for a major portion of agricultural herbicide and insecticide use within the drainage basin (table 3). For example, for nearly all agricultural basins in which corn is the major crop, the target pesticides account for more than 80 percent of the total amount of herbicides and insecticides applied. The target pesticides accounted for a smaller part of the total pesticide use in some basins, particularly in wheat-growing areas and in the two agricultural basins in the Willamette River Basin in Oregon. Pesticides with high use in wheat growing areas and in the Williamette River Basin include the herbicides 2,4-D, bentazon, diuron, and MPCA, the insecticides aldicarb and acephate, and the fungicide chlorothalonil (Giannessi and Anderson, 1996), none of which are included in the target compounds in this report. The target pesticides account for a lower percentage of total pesticide use in most basins when compared with the total amount of all pesticides applied only for agricultural use, including herbicides, insecticides, fungicides, and soil fumigants (table 3). However, several of the pesticides not targeted in this study, particularly the commonly used soil fumigants methyl bromide, 1,3-D, and metam sodium, have physical and chemical properties or application techniques that result in a low potential for transport to streams (Goss and Wauchope, 1990; Draper and Wakeham, 1993). In addition, the volatility of the soil fumigants would be expected to result in relatively fast removal from streams (Gentile and others, 1989, 1992; del Rosario and others, 1994; Rathbun, 1998). Thus, although a considerable amount of pesticide use is not accounted for by the compounds discussed in this report, most of the major agricultural herbicides and insecticides and many of the compounds most likely to be found in streams are included.

In addition, several pesticide transformation products are included in the target compounds in the report. Deethylatrazine (DEA), a transformation product of the herbicide atrazine, frequently has been detected in streams draining areas in which atrazine is used (Larson and others, 1997). A transformation product of the herbicide alachlor, 2,6-diethylacetanilide, also is included in the target compounds. Hexachlorocyclohexane (a-HCH) is a transformation product of the insecticide lindane (y-HCH) and a component of technical grade HCH, an insecticide that is no longer used in the United States. A transformation product of the insecticide DDT, p,p¢-DDE, is another target compound included in this report. This compound is the transformation product of DDT which is most commonly detected in bed sediments of streams and in tissues of aquatic organisms.