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

ENVIRONMENTAL SIGNIFICANCE OF PESTICIDE CONCENTRATIONS

The question of whether the pesticide concentrations measured in samples collected during this study have a significant effect on human or environmental health is difficult to answer. Standards and guidelines have been established by various agencies for a number of the pesticides discussed in this report. Generally, these standards and guidelines are estimates of concentrations in water below which adverse effects on human health or aquatic life are not expected to occur (International Joint Commission, 1977; Canadian Council of Resource and Environment Ministers, 1991; Nowell and Resek, 1994). Regulatory standards have been established by the USEPA for the concentrations of specific compounds in drinking water (the maximum contaminant level, or MCL). Other (nonregulatory) values have been established by the USEPA, the Canadian Council of Resource and Environment Ministers (CCRM), and the International Joint Commission (IJC) for the protection of human health or aquatic life. In this report, the terms criteria and criteria values are used for both regulatory standards and nonregulatory guideline concentrations.

The criteria values selected for comparison are shown in table 8. For pesticides with a USEPA-established criterion value, that value was used. If no USEPA-established value was available, values established by other agencies were used. Of the 46 compounds discussed in this report, 26 have a criterion established for protection of human health and 18 have a criterion established for protection of aquatic life. The criteria have several limitations (discussed below) that must be considered when they are used for comparison with pesticides concentrations measured in streams. The criteria can, however, provide an indication of the potential for adverse environmental effects of the pesticide concentrations measured during this study.

Limitations of Criteria

The criteria established for protection of both human health and aquatic life have limitations to their use in evaluating the potential effects of pesticides in streams (Nowell and Resek, 1994). The limitations discussed here do not address the validity or accuracy of these criteria, an assessment of which is beyond the scope of this report. Rather, these limitations pertain to the relevance or the usefulness, or both, of the criteria for comparisons with concentrations of pesticides measured in rivers and streams during this study.

(1) Criteria have not been established for many of the pesticides. Several of the pesticides that were frequently detected during this study, including alachlor, azinphos-methyl, carbaryl, DCPA, EPTC, prometon, and propargite, lack an established value for protection of either human health or aquatic life, or both.

(2) Criteria have been established for very few pesticide transformation products. DEA, a transformation product of atrazine, was one of the most frequently detected compounds at many of the sites in this study. Recent studies have shown that transformation products of several other commonly detected compounds, including alachlor, cyanazine, and metolachlor, can be present at higher levels and can persist much longer in surface waters than the parent compounds (Goolsby and others, 1993; Kalkhoff and others, 1998).

(3) Human-health and aquatic-life criteria generally are based on toxicity tests conducted with a single compound. The criteria do not take into account the possible additive or synergistic effects of more than one pesticide or combinations of pesticides, pesticide transformation products, or other chemicals that may be present in water. In addition, testing most often involves exposure of an organism to a single compound at a series of concentrations to determine a no-effect level. As shown in previous sections of this report, exposure in actual rivers and streams is more likely to involve a mixture of pesticides and pesticide transformation products, with frequent fluctuations in concentrations and in the types of compounds present. In many locations, seasonal pulses of relatively high concentrations of several pesticides are superimposed on a background of low-level concentrations of many other chemicals.

(4) The aquatic-life criteria do not account for the possible combined effects of pesticides and other potential stressors on aquatic biota, such as high concentrations of suspended sediment, low dissolved oxygen, fluctuations in temperature, or the presence of metals or other inorganic contaminants. It is difficult to predict the effect of individual pesticides on aquatic life in complex natural systems.

(5) Recent concerns about the possible effects of pesticides and other organic compounds on the endocrine systems of humans and aquatic organisms (Colborn and Clement, 1992) are not addressed by the criteria. In general, the criteria do not account for potential long-term effects of pesticides, particularly effects on development and on the reproductive success of future generations of relatively long-lived organisms. There is evidence, however, that some pesticides may affect the endocrine systems of fish (Goodbred and others, 1997).

(6) Most of the sampling sites discussed in this report are on streams that are not used as sources of drinking water, although in many cases the water eventually reaches rivers that are used for drinking water. Some pesticides are removed during normal treatment procedures at water-supply facilities. Studies have shown, however, that several commonly detected compounds, including atrazine, cyanazine, and metolachlor, are only partly removed from water during conventional water treatment (Baker, 1985; Wnuk and others, 1987; Miltner, 1989; Patrick, 1990; Kent and others, 1991).

(7) The human-health criteria are based on long-term (lifetime) exposure to a specific compound. Obviously, the results presented in this report cannot be used for an assessment of lifetime exposure, even if the water were being used directly for drinking water. The aquatic-life criteria established by the USEPA are based on exposure times of 24 to 96 hours. Although sampling was quite frequent at some of the sites in this study, it was not possible to determine the actual exposure times of aquatic organisms to specific compounds using the data from these samples. In addition, the USEPA criteria specify a recurrence interval of 3 years, which implies that an aquatic ecosystem can recover if the concentration of a specific compound does not exceed the criterion value more than once in 3 years (Stephan and others, 1985). For most of the streams sampled, the data were not sufficient to make such a specific determination; thus, the criteria are used in this report as indicators of the potential (or the lack thereof) for adverse effects.

The above limitations are important and should be considered whenever the criteria values are used. Despite these limitations, the criteria are the only nationally consistent, toxicologically derived values that can be used for comparison with the concentrations of many of the pesticides included in this study. In their compilation of standards for pesticides in water, Nowell and Resek (1994) state that the use of these national criteria "facilitate[s] federal and state regulation, as well as consistent comparison and evaluation of water-quality conditions among different hydrologic systems. National standards and guidelines are widely used to assess the potential water-quality significance of pesticide concentrations measured in the aquatic environment ... "

Comparison of Concentrations with Human-Health Criteria Values

All comparisons with human-health criteria values in this report refer to chronic criteria, which are based on exposure over a 70-year lifetime. Criteria values based on acute (short-term) exposure generally are much higher than chronic criteria values. An acute human-health criterion value was exceeded in only one sample during this study. Estimated concentrations of atrazine (120 µg/L) and cyanazine (160 µg/L) exceeded the 1-day health advisory level (HAL) for drinking water (100 µg/L for both compounds) in one sample from Kessinger Ditch in Indiana (whit-kess) in May 1993 (U.S. Geological Survey, 1999); water from this stream is not used for drinking water. For most of the other compounds with acute criteria values, the acute values are 10 to several thousand times higher than the maximum concentration detected in this study.

Of the 46 compounds discussed in this report, 26 have a human-health chronic criterion (table 8). All these criteria apply to an average concentration in finished (treated) drinking water. The criteria values ranged from 1 to 100 µg/L for most of the target herbicides and 0.1 to 50 µg/L for most of the target insecticides. Five compounds--alachlor, atrazine, carbofuran, lindane, and simazine--have an MCL established by the USEPA. The MCL is an enforceable standard for the concentration of a specific compound in drinking water. The MCLs are based primarily on results of toxicity testing but also are influenced by the economic and technological feasibility of water treatment and by analytical detection capability (Nowell and Resek, 1994). Eighteen other compounds have a USEPA-established HAL, which is a nonenforceable guideline derived solely on the basis of toxicity testing. The criteria values for three compounds--p,p¢-DDE, dieldrin, and a-HCH--are in terms of a risk-specific dose (RSD), which is the concentration associated with a specified cancer risk level. The RSD values given for these compounds (table 8) are for a cancer risk level of 10-5, meaning that the excess cancer risk associated with drinking water containing a compound at a concentration of the RSD is estimated to be 1 in 100,000 persons. It also should be noted that the herbicide alachlor is classified by the USEPA as a probable human carcinogen (Nowell and Resek, 1994). The MCL for alachlor is 2 µg/L, but the maximum contaminant level goal (MCLG) is a concentration of zero, as for all compounds classified as known or probable human carcinogens. For a thorough discussion of human-health criteria values for pesticides and for a compilation of these values, see Nowell and Resek (1994).

Of the 26 compounds with human-health criteria, 7 compounds were detected at concentrations greater than their criterion value in one or more samples (table 8). These include the herbicides alachlor, atrazine, cyanazine, and simazine and the insecticides diazinon, dieldrin, and a-HCH. These four herbicides and the insecticide diazinon were detected at concentrations greater than their criteria values at six or more sites. These five compounds will be discussed in more detail below. Dieldrin concentrations exceeded the criterion value of 0.02 µg/L at four sites, but in only 1 to 3 samples at each site. a-HCH was detected in one sample at a concentration slightly higher than the criterion value of 0.06 µg/L. Of the remaining 19 compounds with human-health criteria, 8 compounds were detected at concentrations greater than one-tenth of the criterion value at one or more sites and 11 were not detected at concentrations greater than one-tenth of the criterion value at any site.

In previous sections of this report, concentrations of specific pesticides were described primarily by using aggregated data from groups of sites, often in terms of monthly median concentrations. The human-health criteria, however, are based on long-term exposure to specific chemicals; compliance with regulatory standards for drinking water is based on the annual mean concentration of a specific compound in drinking water. For comparisons with human-health criteria, the most appropriate measure of concentration is the long-term mean concentration of a specific compound at a site. The time-weighted mean (TWM) concentration provides a relatively unbiased estimate of the mean concentration for a period during which sampling frequency varied. At a number of sites in this study, sampling was not sufficient during some parts of the year to determine a reliable TWM concentration for a 1-year period. TWM concentrations were determined, however, for all the target compounds for a 5-month critical period, during which sample collection was most intense at each site. These values were combined with monthly median concentrations for the remaining 7 months to obtain an estimate of the annual mean concentration of each target compound at each site. Specifically, the estimated annual mean concentration was calculated as

where:

tw_mean = the time-weighted mean concentration for the 5-month critical period, which is approximated as a 155 days.

mon_medi = the monthly median concentration for 1 of the 7 remaining months, all of which are approximated as 30 days.

Five compounds (alachlor, atrazine, cyanazine, diazinon, and simazine) were detected at concentrations higher than the human-health criteria values at numerous sites (table 8). Estimates of annual mean concentrations for these five compounds are given in table 9 and figure 33 for all 58 sites. In figure 33, the top of each bar represents the TMW during the 5-month critical period for the compound at a specific site, and the shaded part of the bar represents the estimated annual mean concentration. Annual mean concentrations exceeded criteria values for cyanazine at two sites [Maple Creek in Nebraska (cnbr-maple) and Kessinger Ditch in Indiana (whit-kess)] and for atrazine at one site [Kessinger Ditch (whit-kess)]; concentrations did not exceed criteria values for alachlor, diazinon, and simazine at any site. TWM concentrations during the 5-month critical period exceeded criteria values for atrazine at two sites [Prairie Creek in Nebraska (cnbr-prairie) and Kessinger Ditch (whit-kess)] and for cyanazine at three sites [Maple and Shell Creeks in Nebraska (cnbr-maple and cnbr-shell) and Kessinger Ditch (whit-kess)] all of which are in corn-growing areas. The TWM concentration for diazinon exceeded the criterion value at one urban site [Rush Creek, near Arlington, Texas (trin-rush)]. No human-health criteria values were exceeded at any of the integrator sites. These results show that the long-term mean concentrations of these pesticides rarely exceed human-health criteria even though concentrations in individual samples often exceed the criteria values during seasonal pulses.

Comparison of Concentrations with Aquatic-Life Criteria Values

The aquatic-life criteria values that were selected for comparison with concentrations measured during this study are shown in table 8. The USEPA has established aquatic-life criteria for six of the target compounds (U.S. Environmental Protection Agency, 1991). For 11 compounds with no USEPA-established values, Canadian criteria values were selected (Canadian Council of Resource and Environment Ministers, 1991). The selected aquatic-life criterion value for diazinon is the Great Lakes Water-Quality Objective established by the IJC (International Joint Commission, 1977). No aquatic-life criteria have been established by the United States or Canadian governments for the remaining 28 target compounds.

The aquatic-life criteria values range from 0.1 to 10 µg/L for the target herbicides and from 0.01 to 0.1 µg/L for most of the target insecticides. The criteria values generally are based on the highest concentration at which no adverse effects are observed for the most sensitive aquatic organism tested (plant or animal), multiplied by an appropriate safety factor. The safety factor can range from 0.01 to 0.1, depending on the persistence of the chemical and on the availability and the quality of the toxicity data for the chemical. Information on the derivation of the criteria values can be found in the Canadian water-quality guidelines document (Canadian Council of Resource and Environment Ministers, 1991) for Canadian values, the USEPA water-quality criteria summary (U.S. Environmental Protection Agency, 1991) for criteria established by the USEPA, and the IJC document (International Joint Commission, 1977) for diazinon. In general, the Canadian criteria are somewhat more stringent than the USEPA values. The Canadian criteria "...are set at such levels as to protect all forms of aquatic life and all aspects of the aquatic life cycles. The clear intention is to protect all life stages during indefinite exposure to the water" (Canadian Council of Resource and Environment Ministers, 1991). The USEPA values are based on an average concentration for either a 24-hour or 96-hour period, depending on when the criterion was established, and are set at levels that will protect 95 percent of the organisms for which acceptable chronic-toxicity data are available (Nowell and Resek, 1994). In addition, the USEPA criteria contain a recurrence interval provision which states that if the criterion value is not exceeded more than once in 3 years, aquatic ecosystems are expected to recover (Stephan and others, 1985). The Canadian and IJC criteria do not contain this provision. It should be noted that all the aquatic-life criteria are guidelines--they are not enforceable standards.

Concentrations of one or more compounds exceeded an aquatic-life criterion value in at least one sample from 25 of the 37 agricultural sites, 10 of the 11 urban sites, and 4 of the 10 integrator sites (table 10). No aquatic-life criteria values were exceeded in any sample from the remaining 19 sites throughout the entire sampling period. At most sites, fewer than 5 compounds exceeded aquatic-life criteria values, but at four sites, 6 to 8 compounds were detected at concentrations greater than their aquatic-life criterion value.

Only 9 of the 27 target herbicides have an aquatic-life criterion, all of which are Canadian values (table 8). Concentrations of four of these herbicides--linuron, metribuzin, simazine, and tebuthiuron--did not exceed criteria values in any sample (U.S. Geological Survey, 1999). Triallate concentrations exceeded the criterion value of 0.24
µg/L in six samples from two sites in the Central Columbia Plateau study unit in Washington (ccpt-crab.m and ccpt-palouse). Trifluralin concentrations exceeded the criterion value of 0.1 µg/L in 6 samples, 5 of which were from two agricultural sites in the San oaquin River Basin in California (sanj-orest and sanj-salt). Metolachlor concentrations exceeded the relatively high criterion value of 8 µg/L in 11 samples from five sites. All samples in which metolachlor exceeded its aquatic-life criterion value were collected during seasonal peaks in herbicide concentrations in corn-growing areas. The criteria values for atrazine and cyanazine were exceeded much more frequently; these compounds are discussed in more detail below.

Aquatic-life criteria are established for 9 of the 19 target insecticides, and all of these criteria values were exceeded in at least one sample from one or more sites (table 8). Concentrations of five compounds--carbofuran, dieldrin, a-HCH, lindane, and parathion--exceeded criteria values in fewer than five samples at 1 to 3 sites (U.S. Geological Survey, 1999). Carbofuran concentrations exceeded its relatively high criterion value of 1.75 µg/L only in Zollner Creek in Oregon (will-zollner); the criterion value was exceeded in 2 samples in 1993 and in 1 sample in 1994. Criteria values for four insecticides--azinphos-methyl, chlorpyrifos, diazinon, and malathion--were exceeded much more frequently; these compounds are discussed in more detail below.

Six compounds--atrazine, azinphos-methyl, chlorpyrifos, cyanazine, diazinon, and malathion--exceeded aquatic-life criteria values at 10 or more sites (table 8). To more closely examine the potential effects of these compounds in the sampled streams, it is important to determine whether the concentrations that exceeded the criteria values were isolated cases or whether the concentrations remained higher than these levels for significant periods of time. Because the aquatic-life criteria are based on either a one-time exposure (Canadian values) or an exposure to a 24-hour or a 96-hour average concentration (USEPA values), the monthly median concentrations and 5-month TWM concentrations are not appropriate measures of concentration for comparison with these criteria values. A more useful measure is the number of days the concentration of a specific compound exceeds its criterion value at a site each year. The number of days a compound exceeded its criterion value during a year was estimated by linear interpolation between concentrations in successive samples from a given site. These estimates are given in table 11 for the six compounds that exceeded criteria values at 10 or more sites. The reliability of the estimates varies, depending on the sampling frequency at a particular site. In some cases, a somewhat arbitrary judgement was made as to whether the concentration of a compound remained higher than the criterion value between adjacent samples. For samples in which a concentration exceeded a criterion value but concentrations in adjacent samples were less than the criterion value, only the concentration for that day was counted as greater than the criterion value. For a few sites, data from 1992 also are included in table 11 and in figures 34-40. Sampling at these sites was very intensive during 1992, allowing more reliable estimates of the length of time concentrations exceeded criteria values. Observations for the six compounds that exceed aquatic-life criteria values at 10 or more sites are discussed in the following sections.

Malathion

Concentrations of malathion exceeded its USEPA aquatic-life criterion value of 0.1 µg/L at 13 sites (table 8). Concentrations of malathion greater than this value, however, were rare (table 11). Concentrations greater than 0.1 µg/L were primarily in isolated samples. An example of this is shown in figure 34 for Las Vegas Wash in Nevada (nvbr-lasvegas), which had the highest estimate for days on which malathion exceeded the criterion. Between each instance of a concentration greater than the criterion value, a concentration less than the criterion value was measured in one or more samples. A similar pattern was observed for the other sites at which malathion concentrations exceeded the criterion value.

Azinphos-Methyl

Concentrations of azinphos-methyl exceeded its USEPA aquatic-life criterion value of 0.01 µg/L at 16 sites (table 8). At 11 of these sites, concentrations exceeded the criterion value only 1 day in a given year (table 11), similar to malathion. Concentrations were greater than 0.01 µg/L for many more days at four sites--Crab Creek Lateral (ccpt-crab.rl) and El68 Wasteway (ccpt-el68) in Washington and Orestimba Creek (sanj-orest) and the San Joaquin River (sanj-sanj) in California. Concentrations at Orestimba Creek in 1992 and 1993 and at Crab Creek Lateral in 1993 are shown in figure 35. Concentrations at Orestimba Creek were much higher than the criterion value during the summers of 1992 and 1993. Concentrations at Crab Creek Lateral ranged from 0.05 to 0.2 from mid-May through July 1993. A similar pattern was observed at the ccpt-e168 and sanj-sanj sites, with concentrations exceeding the criterion value primarily during early to midsummer. It should be noted that the analytical recovery of azinphos-methyl was low (table 2); therefore, the concentrations reported for samples in which azinphos-methyl was detected probably are low estimates.

Chlorpyrifos

Concentrations of chlorpyrifos exceeded its USEPA aquatic-life criterion value of 0.041 µg/L at 20 sites (table 8); 6 sites were urban indicator sites (table 11). At most sites, concentrations greater than the criterion value occurred only in isolated samples (table 11). Concentrations were greater than 0.041 µg/L for longer periods at several sites, including Crab Creek Lateral (ccpt-crab.rl) in Washington, the Merced River (sanj-merced) and Orestimba Creek (sanj-orest) in California, and Rush Creek (trin-rush), an urban site in Texas. Concentrations of chlorpyrifos in Orestimba Creek in 1992 and 1993 and in Rush Creek in 1993 are shown in figure 36. In Orestimba Creek, chlorpyrifos was detected in most samples from early spring through the summer during both years, with concentrations in some samples much higher than the criterion value. In Rush Creek, chlorpyrifos was detected in samples for much of the year, with concentrations exceeding 0.041 µg/L in 5 of 16 samples collected during the summer of 1993.

Diazinon

Concentrations of diazinon exceeded the IJC aquatic-life criterion value of 0.08 µg/L at 18 sites (table 8). Concentrations exceeded the aquatic-life criterion value for extended periods at 12 sites, including 4 agricultural sites, 1 integrator site, and 7 urban sites (table 11). Concentration patterns for several of the agricultural indicator sites and the integrator site are shown in figure 37. These five sites are in areas where diazinon is used extensively for agriculture--the San Joaquin Valley in California (sanj-sanj) and the Willamette River Basin in Oregon. Figure 37 shows distinct peaks in concentrations for the Merced (sanj-merced), San Joaquin (sanj-sanj), and the Orestimba (sanj-orest) sites during the months of January or February, or both; these peaks are the result of the application of diazinon to orchards during the dormant season. Concentrations of diazinon in Orestimba Creek were much higher than the criterion value in many of the samples collected during the summer of 1992 and in several samples collected during the summer of 1993. Concentrations exceeded the criterion value in Salt Slough (sanj-salt) and Zollner Creek (will-zollner) mainly during the summer. Concentrations for three urban sites--Rush Creek near Dallas, Texas (trin-rush), Accotink Creek near Washington, D.C. (poto-acco), and Cherry Creek in Denver, Colorado (splt-cherry)--are shown in figure 38. In Rush Creek, concentrations were much higher than the criterion value in nearly every sample from March through September of 1993. In Accotink and Cherry Creeks, concentrations also were greater than the criterion value for much of the summer.

Atrazine

Concentrations of atrazine exceeded the Canadian aquatic-life criterion value of 2 µg/L at 17 sites (table 8). Nearly all of these sites are agricultural indicator sites and integrator sites in corn-growing areas. The criterion value also was frequently exceeded in Little Buck Creek (whit-little), an urban stream in Indiana with a substantial amount of cropland planted in corn in its drainage basin. At sites where the atrazine criterion value was exceeded, concentrations often remained higher than this value for extended periods (table 11). Temporal patterns of herbicide concentrations at these sites were discussed previously. Concentrations of atrazine at integrator sites on the White River in Indiana (whit-white) and the Platte River in Nebraska (cnbr-platte) and at the agricultural site on Chambers Creek in Texas (trin-chamb) are shown in figure 39. At all three sites, concentrations were much higher than the criterion value of 2 µg/L after application in the spring, remained elevated for a period of several weeks or more, and then declined to lower levels for the remainder of the year. The plots for the White and Platte rivers show that this pattern is repeated annually.

Cyanazine

Concentrations of cyanazine exceeded the Canadian aquatic-life criterion value of 2 µg/L at 10 sites (table 8). The aquatic-life criterion value for atrazine also was exceeded at these 10 sites (table 11). The estimated period of time that concentrations of cyanazine were greater than the criterion value, however, was shorter than the period for atrazine at all sites. Concentrations of cyanazine at the integrator sites on the White (whit-white) and the Platte (cnbr-platte) rivers and at the agricultural site on Maple Creek in Nebraska (cnbr-maple) are shown in figure 40. The temporal pattern of concentrations at these three sites was very similar to the patterns observed for atrazine.

Co-Occurrence of Pesticides at Concentrations Exceeding Criteria Values

At many sites, concentrations of more than one compound exceeded an aquatic-life criterion value (table 10). In some cases, this occurred in the same sample or during the same period. For example, the concentrations of atrazine and cyanazine in the White River in Indiana [whit-white (figs. 39 and 40)] were much higher than their respective aquatic-life criteria values during the same period in both 1992 and 1993. Using the data from all 58 sites for the entire sampling period, one or more aquatic-life criteria values were exceeded in 410 samples: 131 samples had 2 compounds with concentrations greater than their criteria values, 17 samples had 3 compounds with concentrations greater than their criteria values, and 7 samples had 4 compounds with concentrations greater than their criteria values. The presence of multiple compounds with concentrations greater than their aquatic-life criteria values was widespread, occurring at 29 sites, including 8 of the 11 urban sites, in 12 of the 19 study units.

At some sites, the occurrence of two or more compounds with concentrations greater than the criteria values was particularly evident. Figure 41 shows a variety of situations where concentrations of two or more compounds exceeded aquatic-life criteria values at some of these sites. In these plots, concentrations of each compound have been normalized by dividing the concentrations by the aquatic-life criterion value for that compound. Normalizing the concentrations accounts for differences among the criteria values for the various compounds. In the plots, a normalized concentration greater than 1 indicates a concentration greater than the criterion value. Note that the concentration scales in these plots are logarithmic; thus, a normalized concentration of 10 indicates that the concentration in that sample was 10 times higher than the criterion value for that compound.

At several the agricultural sites, criteria values for atrazine and cyanazine, and occasionally metolachlor, were exceeded in the same sample or during the same period (fig. 41A,B). This also was true at the Platte and White river sites (cnbr-platte and whit-white), the two integrator sites where corn is the major crop grown in the basin (figs. 39, 40, 41C). Occasionally, concentrations of the insecticide chlorpyrifos also exceeded its criterion value in the same samples from these sites.

At agricultural sites in the San Joaquin River (sanj-salt, sanj-merced, and sanj-orest) and the Willamette River (will-pudding and will-zollner) basins, concentrations of two or more insecticides often exceeded criteria values in the same samples or during the same period. Concentrations of diazinon, chlorpyrifos, and azinphos-methyl were higher than their criteria values in many of the samples from Orestimba Creek in California [sanj-orest (fig. 41D)]. To a lesser extent, diazinon and chlorpyrifos concentrations exceeded criteria values in samples from the Merced River in California [sanj-merced (fig. 41E)]. In Zollner Creek in Oregon [will-zollner (fig. 41F)], concentrations of several insecticides, including azinphos-methyl, carbofuran, chlorpyrifos, diazinon, and malathion, as well as the herbicide atrazine, were higher than or near criteria values at various times throughout 1993 and 1994, often in the same samples.

Example plots for four of the urban sites are shown in figure 41G-J. In Rush Creek, Cherry Creek, and Las Vegas Wash (trin-rush, splt-cherry, and nvbr-lasvegas), diazinon concentrations greater than the criterion value often coincided with concentrations of other insecticides greater than or near their criteria values, including the insecticides azinphos-methyl, chlorpyrifos, and malathion. Concentrations of several insecticides were greater than the criteria values for much of the year in Rush Creek (fig. 41G) and Las Vegas Wash (fig. 41H). Finally, Little Buck Creek in Indiana [whit-little (fig. 41J)] shows characteristics of both agricultural and urban basins. Atrazine concentrations in samples from this site exceeded the criterion value in spring and early summer similar to concentrations in samples from agricultural sites in corn-growing areas, whereas diazinon and malathion concentrations were greater than or near their criteria values for much of the summer and autumn, which is typical of many of the urban sites.

The plots in figure 41 and the data in table 11 illustrate another important point concerning the concentrations measured during this study. Throughout this report, it has been stated that herbicide concentrations generally were greater than insecticide concentrations at most sites, often by a factor of 10 or more. The aquatic-life criteria values for herbicides, however, often are substantially greater than the values for insecticides. When concentrations of compounds are normalized to their criteria values (fig. 41), it is apparent that at most sites insecticides may be more important than herbicides in terms of potential effects on aquatic life. This is evident in agricultural areas with high use of insecticides, such as the Orestimba (sanj-orest), Merced (sanj-merced), and Zollner (will-zollner) sites, and at most urban sites, where insecticides commonly are detected at concentrations greater than their criteria values. In addition, two of the agricultural sites in the Central Columbia Plateau (ccpt-crab.m and ccpt-el68) study unit had concentrations of azinphos-methyl greater than the criterion value for extended periods (table 11, figure 35). Even where insecticide levels are much lower than herbicide levels, insecticides may be more important in terms of potential effects on aquatic life. At many of the agricultural sites usually associated with high levels of herbicides (for example, sites in corn-growing areas) concentrations of azinphos-methyl, chlorpyrifos, diazinon, and malathion were occasionally greater than the criteria values, often during the same period when herbicide levels were high. Whether this combination of herbicides and insecticides has an effect on the toxicity to aquatic organisms is unknown, but it is not accounted for in the derivation of the criteria values for either group of compounds.