CHAPTER 5 Analysis of Key Topics-Sources, Behavior, and Transport 5.1 SEASONAL PAlTERNS OF PESTICIDE OCCURRENCE Most agricultural pesticides, particularly herbicides, are applied during distinct and relatively short seasonal periods. Preemergent herbicides are applied just before planting, for example, and postemergent herbicides are applied a few weeks after the crop begins to sprout. Some crops receive an autumn application of herbicides to kill the plant before the crop is harvested. Some insecticides also are applied at certain times of the year to control specific pests. Sometimes pesticides not routinely used are applied to control an unexpected pest. Seldom in agricultural applications is the same pesticide used continually for long periods of time (i.e., months) during a growing season on the same crop. The seasonal application of a pesticide is the primary source for transport to surface waters, if residues in soil from applications in previous years are minimal when compared to the amount being applied. The first runoff-inducing rain or irrigation event after application of a pesticide can potentially move significant quantities of the pesticide to surface waters. This has been observed for numerous compounds, especially the preemergent herbicides, in many river systems in the midwestern United States. Schottler and others (1994) observed a strong seasonality in the occurrence of herbicides in the Minnesota River (Figure 3.46), as did Larson and others (1995) and Goolsby and Battaglin (1993) for a number of herbicides in a wide range of stream sizes in the Mississippi River Basin. The seasonal pattern of occurrence for herbicides, such as atrazine and alachlor in midwestern rivers, is well known and somewhat predictable. In late winter and early spring, the concentrations of pesticides are low, often below the detection limit. The source of compounds detected during this time is primarily ground water (Squillace and others, 1993), although discharge from reservoirs, surface runoff from fields, and discharge from tile drains also may add low levels of pesticides to streams. Application of herbicides in the Midwest starts in late April to mid-May, depending on weather conditions. Elevated herbicide concentrations are observed in streams draining agricultural areas for a few days to a few weeks, depending on the timing and number of rain events and the size of the drainage basin. During this period, about 0.2 to 2 percent of the applied chemical may be moved to surface waters. As the crops grow and the rains subside, the movement of pesticides to surface waters is diminished and riverine concentrations decline throughout the summer. For some compounds, such as atrazine, a low- level, relatively constant concentration is reached and maintained throughout much of the autumn and winter. For others, such as metribuzin, alachlor, and EPTC, the concentration drops below detection levels and remains there until the chemicals are applied again the following spring. The low-level herbicide concentrations observed during the low-flow period (autumn through winter) may result from inflow of ground water from alluvial aquifers that were filled up during the © 1998 by CRC Press, LLC 236 PESTICIDES IN SURFACE WATERS high-flow period when pesticide concentrations were also relatively high. The cycle then repeats itself the next spring. The seasonal cycle of herbicide concentrations in midwestern reservoirs is somewhat different than in rivers. Many reservoirs in the Midwest receive much of their water from surface water sources during the spring runoff period, when concentrations of herbicides in tributary streams are relatively high. The water is stored for use during the remainder of the year. For compounds that are relatively stable in water, concentrations may remain elevated in reservoirs much longer than in streams, since they are not flushed from the system as quickly. Thus, concentrations of pesticides in reservoirs can remain relatively high long after inputs from agricultural fields have declined or ceased. This effect was observed in the 1992 study of midwestern reservoirs (Goolsby and others, 1993) described earlier (Section 3.3). In Figure 5.1, detection frequencies for herbicides and selected degradation products in reservoirs and streams are compared. The number of reservoirs with detections was nearly constant for most of the analytes from the June-July sampling period through the October-November sampling period. In contrast, the number of streams with detections dropped considerably between the early summer sampling and the late autumn sampling for most analytes. The same contrast was seen in the concentrations of the analytes. In Figure 5.2, concentrations of atrazine, alachlor, and several transformation products in midwestern streams and reservoirs are compared. The stream concentrations follow the pattern described above, with low levels in the preplanting and postharvest periods, and elevated concentrations during the postplanting period. The concentrations in the reservoirs, on the other hand, were much more stable from the early summer period through late autumn, except for alachlor. Alachlor apparently degraded more quickly in the water column of the reservoirs than the other compounds. The seasonal pattern in reservoirs has implications for users of drinking water derived from reservoirs in this region. Compliance with the Safe Drinking Water Act (SDWA) requires that the annual average concentration of a number of pesticides, obtained with quarterly sampling and analysis, remain below a maximum contaminant level (MCL) established for each specific chemical. For most streams supplying drinking water, the normal seasonal pattern in this region results in annual average concentrations below the various MCLs. For reservoirs, the longer period of elevated concentrations increases the likelihood that at least two of the four quarterly samples may have elevated concentrations of some pesticides. The storage of water with relatively high levels of herbicides in reservoirs also can affect the seasonal pattern of herbicide concentrations in rivers downstream from the reservoir. Depending on the timing of releases of water from the reservoir, downstream concentrations of herbicides would be expected to remain elevated for a longer time than in an unregulated stream. In some cases, the low-level concentrations observed during autumn and winter for certain pesticides, such as atrazine, may be partially attributed to release of water from reservoirs filled during the spring runoff period. Peak concentrations in streams downstream from reservoirs, however, would be expected to be lower because of dilution in the large volume of water in the reservoir (Goolsby and others, 1993). For some compounds with relatively short aquatic lifetimes, such as alachlor, both the duration and magnitude of elevated concentrations downstream from reservoirs may be decreased, due to degradation within the reservoir. For the most part, the effect of reservoirs on seasonal pesticide concentration patterns in streams has not been specifically addressed in published studies. Seasonal patterns of pesticides in streams may be different in different parts of the nation, depending on the timing of pesticide application and significant rainfall or irrigation. For example, the streams draining the Central Valley of northern California have a strong seasonal © 1998 by CRC Press, LLC Analysis of Key Topics-Sources, Behavior, and Transport 237 Reservoirs I I I I - A - - - - - - - April-May June-July August- October- September November " March-April May-June October- November atrazine I alachlor metribuzin o deethylatrazine A deisopropylatrazine 0 prometon A metolachlor cyanazine X ESA Figure 5.1. Detection frequencies for herbicides and selected degradation products in 76 midwestern reservoirs in 1992 (A), and in 147 midwestern streams in 1989 (B). Data are from Goolsby and others (1 993) and Goolsby and Battaglin (1 993). appearance of methidathion and diazinon-organophosphorus insecticides (OPs) used on orchards-in January and February during the rainy season (Kuivila and Foe, 1995), as shown in Figure 5.3. Herbicides and insecticides used on rice in California also have a distinct seasonal pattern of occurrence in surface waters because of release of irrigation water at specific times (Crepeau and others, 1996), as shown in Figure 5.4. In the Yakima River in Washington, concentrations of 2,4-D followed a distinct seasonal pattern from 1967 to 1971, with elevated concentrations generally occurring from May to September (Manigold and Schulze, 1969; Schulze and others, 1973), as shown in Figure 5.5. In general, available data show that the seasonal input of pesticides into surface waters is dependent on the combination of the timing of pesticide application and subsequent rainfall or irrigation, or release of water in regulated © 1998 by CRC Press, LLC 238 PESTICIDES IN SURFACE WATERS 0 $ 100. 1 I I I I I I I desisopropylatrazine (atrazine metabolite) Sampling period Figure 5.2. Temporal distribution of concentrations products in 147 midwestern streams in 1989, and in Goolsby and others (1 993). Sampling period EXPLANATION 0 - Maximum concentration -95th percentile -75th percentile - -Median -25th percentile I -5th percentile i -reporting limit or minimum concentration of atrazine, alachlor, and selected degradation 76 midwestern reservoirs in 1992. Redrawn from © 1998 by CRC Press, LLC Analysis of Key Topics-Sources, Behavior, and Transport 239 Diazinon Methidathion 10 20 30 9 19 January February 0 10 20 30 9 19 January February Figure 5.3. Loads (fluxes) of diazinon and methidathion in the Sacramento River at Sacramento (A) and the San Joaquin River at Vernalis (B) in January and February 1993. Redrawn from Kuivila and Foe (1 995). systems. This is probably true for agriculturally applied pesticides throughout the United States, although there is less published data on the seasonal concentration patterns of pesticides in surface waters outside the midwestern and western United States. The seasonal pattern in urban areas differs from that of agricultural areas because of differences in the timing of pesticide application. Urban runoff in Minneapolis, Minnesota, recently has been shown to contain the herbicides 2,4-D, MCPP, and MCPA during April through October (Wotzka and others, 1994), as shown in Figure 5.6. The low-level appearance of the herbicides in early spring and late autumn was attributed to use on lawns and gardens by commercial applicators. During mid-summer, significantly higher concentrations of herbicides were detected in runoff and attributed to applications by individual homeowners. During this © 1998 by CRC Press, LLC 240 PESTICIDES IN SURFACE WATERS :I Molinate * Po-? 0 . t I + 1992 I I #i 30 I I 4 Figure 5.4. Concentrations of three rice pesticides (rnolinate, 1990-1 992; carbofuran, 1991 -1 992; and thiobencarb, 1991-1992) in the Colusa Basin Drain in the Sacramento Valley, California. Modified from Crepeau and others (1 996). 0.4 0.3 ~~II~II~~~~~~~~~~~~I~~~~I~~~~I~~~~I~~~~ - Thiobencarb - - 0.2 - - 0.1 - 0 30 10 20 30 9 19 29 9 19 May June July © 1998 by CRC Press, LLC Analysis of Key Topics-Sources, Behavior, and Transporl 241 Date Figure 5.5. Concentrations of 2,4-D and river discharge in the Yakima River at Kiona, Washington, 19661 971. Data are from Manigold and Schulze (1 969) and Schulze and others (1 973). 70 60 2 50 2 40 0 - e E 30 E C S 20 10 0 7 7 F Storm sampling date Figure 5.6. Concentrations of the herbicides MCPP, MCPA, dicamba, and 2,4-D in storm drains that drain a residential watershed in Minneapolis, Minnesota, from April to October 1993. Data are from Wotzka and others (1 994). © 1998 by CRC Press, LLC 242 PESTICIDES IN SURFACE WATERS period, inputs of the pesticides were spread out over time with no distinct seasonal pattern. The same observations were made for the insecticide diazinon in the study of the Mississippi River and major tributaries (Larson and others, 1995). In the three river basins with the highest population densities and significant urban centers (the White, Illinois, and Ohio River Basins), the observed flux of diazinon was much greater than would be expected, on the basis of known agricultural use, and had a different seasonal pattern than exclusively agricultural pesticides, such as atrazine, in the same rivers (Figure 5.7). The authors attributed this lack of a seasonal pattern to continual urban use throughout the spring, summer, and autumn. These studies indicate that seasonal patterns of occurrence for urban-use pesticides in surface waters are less distinct and occur over a longer time than for agricultural-use pesticides. A study of the Susquehanna River in Pennsylvania examined the concentrations of 2,4-D and atrazine over a 12-month period (Fishel, 1984), as shown in Figure 3.45. In the Susquehanna River Basin, there are a variety of land uses, including urban, forested, and agricultural areas (see Section 3.3). Each of these could provide inputs of 2,4-D to the river at various times of the year. Atrazine, on the other hand, has exclusively agricultural uses, and inputs to the river occur mainly in the spring and early summer. Atrazine concentrations in the river show the typical seasonal pattern observed in agricultural areas, whereas 2,4-D concentrations lack strong seasonal patterns, probably from the multiple sources of this compound in the basin. Resuspension of bed sediments can provide a seasonal source of hydrophobic, recalcitrant pesticides, such as DDT and other organochlorine insecticides (OCs), to surface waters. Bed-sediment particles can be scoured from the bottom and reintroduced into the water column when streamflow is high enough. Pesticides sorbed to these particles may be released to the water column in the dissolved phase before equilibrium is reestablished (see Section 4.2). Resuspension can occur during periods of high flow resulting from spring or autumn rains, extremely large single-storm events, or large releases of irrigation or reservoir waters. In Chesapeake Bay, increases in organochlorine concentrations in the water column (sorbed to suspended sediments) have been attributed to resuspension of bottom sediments by strong currents in parts of the bay (Palmer and others, 1975). Some of these high-energy events in surface waters have a distinct seasonal pattern. Seasonal patterns in surface-water contamination also have been observed in areas where soil still contains residues from past use of OCs. In the Yakima River Basin in Washington, where irrigation is used to support intensive agricultural activity, total DDT (sum of DDT, DDD, and DDE) concentrations in agricultural drains entering the Yakima River have been shown to be proportional to the suspended-sediment concentration (Johnson and others, 1988; Rinella and others, 1993). Suspended sediment and total DDT concentrations in the river increase during the irrigation season as soil contaminated with DDT is washed into the agricultural drains. The same pattern has been observed in the Moon Lake watershed in Mississippi, where increased total DDT concentrations in the water column occurred during the winter and spring rainy seasons (Cooper, 1991). Soil in this watershed contained significant amounts of DDT (as of 1985), and analysis of sediment cores from Moon Lake showed that recently deposited sediment contained higher amounts of DDT than sediments deposited during the time of heavy DDT use. The authors concluded that DDT in the older sediments was slowly degrading, and the DDT in the recent sediments was coming from eroded soil entering the lake each rainy season. The presence of substantial residues of DDT in soil has been documented in a 1985 study in California (Mischke and others, 1985), and it is likely that seasonal inputs of DDT and other recalcitrant pesticides are occurring in other areas with past use of these compounds (see Section 3.4). © 1998 by CRC Press, LLC Discharge E Discharge k White River Ohio River Illinois River a 8 400 a " 24000 v 2000 P 2 e g.! I6000 Discharge " E 200 .E 3 'i 5 8000 .LC 3 a" 0 fiv 0 0" 0 Date I I I Diazinon 7 Date 2 5 2 CU CU ?? 2 .r F gr! 3 6 7 4 Date Figure 5.7. Comparison of river discharge, atrazine concentrations, and diazinon concentrations in the White (Indiana), Ohio, and Illinois Rivers, 1991-1992. Data are from Coupe and others (1995). © 1998 by CRC Press, LLC 244 PESTICIDES IN SURFACE WATERS 5.2 SOURCES AND CONCENTRATIONS OF PESTICIDES IN REMOTE WATER BODIES On a national scale, the dominant source of pesticides to surface waters is agricultural use, with additional inputs from use in urban areas. Sources in more remote areas, such as forests and roadsides, are much more limited in both area and amount of pesticides applied. The compounds currently used for these purposes-such as 2,4-D, picloram, triclopyr, glyphosate, diflubenzuron, and bacterial agents-generally have short environmental lifetimes, and studies suggest that contributions to surface water contamination from these sources are minimal (see discussion in Sections 4.1 and 5.4.). Thus, in remote non-agricultural areas, atmospheric deposition of relatively long-lived pesticides to surface waters is probably more important than local use. The relative contribution of atmospheric pesticides to a specific surface water body depends on how much of the water budget is derived from drainage, runoff, and precipitation, and how close the water body is to the sources of the pesticides. The magnitude of direct aerial deposition to surface waters is directly proportional to the surface area of the body of water. Generally, lakes are more likely to be affected by atmospheric deposition than streams because the surface areas of lakes represent a much greater proportion of their drainage area than do the surface areas of streams. The significance of the atmospheric input of pesticides to remote lakes and streams is not well known, largely because of the lack of available atmospheric concentration data. The best understanding of the atmospheric inputs of pesticides to surface water comes from years of study of OCs in and around the Great Lakes. One of the earliest observations of pesticides and other chlorinated hydrocarbons in surface waters in a remote area was from Siskiwit Lake on Isle Royale in Lake Superior (Swain, 1978). Residues of numerous organochlorine compounds were detected in the water, sediment, biota, and precipitation on this island, which is hundreds of miles from the nearest intensive agricultural or industrial activity. The conclusion was that all the organochlorine residues found in the lake had come from atmospheric deposition. This conclusion was supported by observations of the same compounds in precipitation. This finding provided the impetus for many research projects investigating the atmospheric inputs of pesticides and other organic chemicals into the Great Lakes ecosystem. Strachan and Eisenreich (1990) estimated that atmospheric deposition is the greatest source of DDT into Lakes Superior, Michigan, and Huron, where the concentrations range from subnanogram to nanogram per liter. Murphy (1984) used precipitation concentration data from Strachan and Huneault (1979) to estimate the loadings of eight organochlorine pesticides into four of the Great Lakes from 1975 to 1976. The depositional amounts ranged from 112 kglyr for hexachlorobenzene (HCB) to nearly 1,800 kglyr for a-HCH, roughly the same as reported by Eisenreich and others (1981). Strachan (1985) reported that precipitation at two locations at opposite ends of Lake Superior contained a variety of organochlorine pesticides. The calculated average yearly loadings ranged from 3.7 kglyr for HCB to 860 kglyr for a-HCH. Voldner and Schroeder (1989) estimated that 70 to 80 percent of the toxaphene input to the Great Lakes was derived from long-range atmospheric transport and wet deposition. The OCs also have been observed in remote surface waters other than the Great Lakes. A number of researchers have reported these chemicals in open ocean areas in the Atlantic and Pacific (Risebrough and others, 1968; Tanabe and others, 1982; Krher and Ballschmiter, 1988; Iwata and others, 1993). Duce and others (1991) have reviewed the literature on the atmospheric deposition of trace chemical species, including OCs, to the world's oceans. As an example, they estimated atmospheric deposition of the HCHs at 2 and 30 mg/m21yr for the South Atlantic and © 1998 by CRC Press, LLC [...]... and many models have not been validated against field data Finally, models should not be used without clearly defined objectives, standard procedures, and thoroughly trained personnel In general, most scientists-academic, regulatory, and industrial-would agree that models provide a valuable tool in understanding and predicting the behavior and fate of pesticides in the environment in general, and in. .. Carolina The timing of planting and herbicide applications in these areas corresponds to the start of the increased concentrations in rain in the Chesapeake Bay area This suggests that atrazine and simazine can be transported in the atmosphere as much as 600 mi from the point of application Concentrations of alachlor and metolachlor in rain did not show the same pattern, being present in rain only during... been considerable work in understanding and modeling the processes involved in runoff Leonard (1990) has outlined the conceptual model for runoff (Figure 5. 9) The general factors that need to be considered include the following: 1 Climatic conditions (rainfall duration, amount, and intensity; timing of rainfall after pesticide application; and time to runoff after inception of rainfall); 2 Soil conditions... Analyzer (DiToro and others, 1982), TOXIWASP/WASTOX/WASP4 (Ambrose and others, 1983; Bums 1983), EXAMS 11, or Exposure Analysis Modeling Systems (Bums and Cline, 19 85) , and TOXIC (Schnoor and McAvoy, 1981) The following are a few examples of how pesticides were modeled in lakes and reservoirs Schnoor and others (1982) modeled the time-dependent fate and transport of atrazine into, within, and out of a... nonequilibrium models, provide the only tools for quickly and inexpensively predicting the behavior and fate of new pesticides in the surface water environment USE OF MODELS Future work in the modeling of pesticide transport, behavior, and fate in surface waters lies in two main areas The first is the continued refining and development of the kinds of models described above Each of the types of models... diazinon in the White (Indiana), Illinois, and Ohio © 1998 by CRC Press, LLC 248 PESTICIDES IN SURFACE WATERS Rivers were low, generally in the 0.01 to 0. 05 pg/L range, but measurable throughout the summer and autumn The seasonal pattern was similar to the pattern described for urban-use herbicides in Section 5. 1, and the estimated riverine flux indicated that much of the diazinon observed originated... pertinent to pesticides in surface waters include structure-activity models that predict chemical properties or the equilibrium state of a transfer process, field runoff models for understanding the delivery of pesticides to surface waters, surface water transport models for simulating rivers and lakes, and regional multimedia models for predicting the equilibrium state of pesticides among land, air, and. .. Topics-Sources, Behavior, and Transport f 247 A more recent (1993) study in Minneapolis, Minnesota, analyzed water in storm sewers draining a residential area for 26 pesticides currently used in urban and agricultural areas (Wotzka and others, 1994) While most samples contained very low or undetectable levels of most of the pesticides, storm-runoff water in June contained the herbicides MCPP, MCPA, and. .. compounds degraded more quickly in the atmosphere and that regional transport probably does not occur Buser (1990) quantified atrazine, simazine, and terbuthylazine in rain, snow, and remote Alpine lakes in Switzerland The © 1998 by CRC Press, LLC 246 PESTICIDES IN SURFACE WATERS concentrations of the herbicides in six mountain lakes, far from agricultural activities, were in the subnanogram per liter... Crossland and others (1986) examined the behavior of two pesticides- methyl parathion and pentachlorophenol (PCP) -in an outdoor experimental pond and modeled their compartmental distributions and transformation rates Halfon (1986, 1987) modeled the behavior and transport of mirex into and within Lake Ontario and used the mirex concentration preserved in a sediment core to calibrate a model of in- lake . - A - - - - - - - April-May June-July August- October- September November " March-April May-June October- November atrazine I alachlor metribuzin o deethylatrazine A deisopropylatrazine. Maximum concentration -9 5th percentile -7 5th percentile - -Median -2 5th percentile I -5 th percentile i -reporting limit or minimum concentration of atrazine, alachlor, and selected degradation. atrazine, simazine, and terbuthylazine in rain, snow, and remote Alpine lakes in Switzerland. The © 1998 by CRC Press, LLC 246 PESTICIDES IN SURFACE WATERS concentrations of the herbicides in