Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Assessment of Groundwater Contamination Problems in Agricultural Areas LEE A. MULKEY U.S. Environmental Protection Agency ASSESSMENT NEEDS Groundwater contamination from human activities is an increas- ingly important environmental concern worldwide. The sources of contamination are many and have included chemical spills in rail- way or highway accidents; chemical leaks from underground storage tanks, wastewater treatment ponds, and sewage disposal systems; chemical leachate from land disposal of industrial wastes; and pesti- cide movement from the soil surface after application. Contaminants include synthetic organic chemicals and inorganic pollutants, par- ticularly nitrates and heavy metals. Contamination impacts range from water quality impairment of irrigation sources to human health risks from drinking water supplies. In some cases, contaminated groundwater can also contaminate surface water, especially under conditions of shallow return flows to streams, artificially drained soil systems, and surface water interception of contaminant plumes. Agricultural production as a source of groundwater contamina- tion and the associated public health risks are the focus of this paper. The major concern over the last two decades has been elevated ni- trate levels in water supply wells. More recently, however, the role of pesticides in groundwater contamination has become the focus of intensive investigation. Pesticides can present many possible human health risks, and each risk is chemical-specific. The evidence of groundwater contamination from agricultural use of pesticides is unequivocal. In some cases, concern for health 65
66 risks posed by concentrations measured in groundwater has con- tributed to the decision to ban specific chemicals such as EDB and DBCP, and to restrict the use of others such as aldicarb. These ex- amples suggest the need for continued monitoring for other specific pesticides and development of control strategies. The widespread use of pesticides presents a number of difficul- ties. First, pesticides are used over large areas, at different times during the year, and for a variety of farming needs. These variables are further complicated by the well-known variations in soil proper- ties and climate. Such complicating factors dramatically increase the sampling requirements for complete characterization of the system. Actions to clean up pesticide contamination once discovered are not feasible because of the extent of contamination and the associated costs for removal and treatment. A more prudent approach to min- imizing groundwater contamination is to evaluate the potential for contamination before a pesticide is released for general use or before an existing chemical is used for a new purpose. Recent trends in certain pesticide properties are also a cause for concern. For insecticides, the historical trend in chemical solubili- ties has shown a dramatic increase, ranging from relatively insoluble organochlorines (< 1.0 mg/1 in water) developed in the period 1940- 1960 to highly soluble carbamates (> 10,000 mg/1 in water) intro- duced since 1970. Persistence for more recently introduced chemicals has decreased, but mobility appears to have been enhanced. Percolation to groundwater is the mechanism for transport of dissolved pesticides. Agricultural practices that increase infiltration and hence percolation have the potential for increasing chemical leaching. Most notable among such practices is the use of conserva- tion or reduced tillage. By reducing tillage, soil is apparently more amenable to infiltration. Ironically, the dramatic increase in the use of conservation tillage has in part been promoted by the water quality benefits expected. Rainfall runoff, soil erosion, and hence pol- lutant loadings from conservation tillage systems to surface waters are decreased, but groundwater contamination may increase. Assessment capabilities are needed to evaluate groundwater con- tamination by pesticides. The assessment procedures desired depend upon the approaches available to control the problem. Available control options include: 1. monitoring and water treatment; 2. monitoring and pesticide use restrictions;
87 3. monitoring, pesticide use restrictions, and best management practices; and 4. pre-release evaluation of potential pesticide groundwater con- tamination. Note that monitoring is a key element in three of the four control options. Measurements are an essential part of the fourth option but not necessarily directed toward ambient monitoring of pesticide residues in soils or groundwater systems. The basic elements of each of the above listed options are briefly summarized. In the first option, contaminated drinking water supplies can be treated for removal of pesticide residues. An effective monitoring program must be maintained to identify treatment needs and confirm continued treatment effectiveness. Once pesticides are identified, the appropriate removal operations must be employed. For trace-level concentrations that can create chronic health problems, such removal technology is costly, must be closely maintained, and can be unre- liable. Large treatment systems for municipal water supplies exist but the more rural, less populated areas often will not have such systems. Widespread use of this approach requires extensive moni- toring assessments and subsequent design, construction, operation, and maintenance of treatment technology. The second option involves restrictions on the further use of pesticides after they are discovered as contaminants in groundwater systems. This is an effective way to control current problems and can help prevent future problems. Restrictions can vary from total bans on specific chemicals to reductions in allowable application rates. Assessment methods are focused on sampling procedures, statistical designs, and appropriate interpretation of observed data. A poten- tially difficult aspect of this approach is the problem of extrapolation, due to the diffuse, highly variable, and widespread nature of pesticide use as well as specific soil/plant/climate conditions. Extrapolation across such a wide range of variables is difficult at best and most often restricts the set of control options available to simple bans or rate-of-use reductions over large geographical areas. Restrictions based on monitoring can prevent some problems in cases where the pesticide has not yet been widely used, but it is essentially a post facto approach that requires evidence of contamination before action is taken. In addition, if extrapolation is to be accurate and risks effectively managed, scientific understanding of pesticide interactions with the environment is essential. A notable example is the ban placed on
68 aldicarb on potato production on Long Island, New York. Con- tamination of local water supply wells led to an intensive monitoring program for soils, aquifers, and water supply systems. Based on these results, a total aldicarb ban for the region was enacted to prevent further contamination. The third option for controlling contamination is a refinement of the second option by adding the use of best management practices (BMPs). BMPs are agronomical, engineering, and pesticide use practices that either singly or in combination can prevent or limit groundwater contamination. BMPs for soil erosion have proven to be effective in preventing soil loss and are known to be beneficial in controlling surface water pollution from agricultural runoff. BMPs for pesticides are not as well developed but include methods such as application timing, slow release formulations, chemical substitution, and integrated pest management. This approach presents several technical and regulatory difficul- ties. The cause/effect relationships among the BMPs and ground- water contamination must be clearly understood. More importantly, the efficiency of regulation is low because of the dispersed and tech- nically complex set of procedures required. Monitoring requirements are likely to be higher, and the extrapolation problem is further com- pounded by the uncertain role of BMPs. Successful implementation of BMPs offers a key advantage, however, in that a wider array of safe pesticide use practices can result. One notable example of BMP implementation is the strategy for aldicarb use on citrus in the state of Florida. The application rate was reduced, the timing of applica- tion was restricted to avoid excessive percolation from supplemental irrigation, and a total ban was placed on areas closer than 1,000 feet to the nearest well. The assessment methodology used in devel- oping this strategy was comprehensive and included mathematical modeling. Further use of BMPs is elaborated below. The observance of groundwater contamination is required for the three control options summarized thus far. A more prudent strategy, however, is the prevention of contamination, or at least the prevention of unacceptably high levels of contamination. Ac- ceptable contamination occurs at a level judged to pose acceptable risks to impacted populations. Pesticide regulation as practiced in the United States offers an opportunity to prevent contamination because each new chemical must be evaluated for its environmental and human health risks before it is released for general use. Increas- ingly, assessments of potential groundwater contamination are being
09 used to specify use restrictions and management practices designed to prevent future problems. Assessment methodologies for groundwater contamination are actively being developed by research laboratories and are already partially in use by the U.S. Environmental Protection Agency (EPA). These assessments are predictive and must rely on knowledge of spe- cific chemical properties and chemical interactions with the soil/ plant/groundwater system as influenced by hydrological and mete- orological cycles. Predictive techniques are of great benefit for all of the control options described above. Extrapolation is essentially a prediction problem, and knowledge of pesticide interaction with a wide range of agronomical and environmental conditions can lead to more informed regulatory programs. A comprehensive control strategy for reducing risks posed by pesticide contamination of groundwater can include a combination of all the options discussed above. Prevention is preferred where possible. Treatment of contaminated water is costly, may be unre- liable, and must be maintained as long as the pesticides are in use. Monitoring is a key element in identifying existing problems and in confirming the effectiveness of control strategies. Assessment meth- ods are essential to implement the desired options. The research approach adopted for their development will now be discussed. RESEARCH APPROACHES Assessment methodologies that can predict the behavior of pesti- cides in the environment is the goal of the research approach adopted by EPA. The major phases of the approach include: ⢠identification and mathematical description of physical, chemi- cal, biological, and transport processes that transform pesticides within soils and groundwater; ⢠characterization of the environmental properties that influence pesticide behavior in soil/groundwater systems; ⢠characterization of pesticide use patterns within the environ- ment; ⢠development of simulation models for predicting leaching and groundwater transport; ⢠conduct of field model validation studies; ⢠demonstration of assessment methodologies. The assessment procedure addresses both commercial chemicals and
70 new chemicals submitted for evaluation and approval by the regula- tory authority. The transformation and transport of pesticides in soils and groundwater can be viewed as the result of the interaction of the hydrological cycle and biological/chemical processes. Processes re- ceiving the most attention to date include: ⢠Hydrolysis: The reaction of chemicals with water can transform pesticides into by-products that are less harmful than the parent compound. Hydrolysis reactions are commonly catalyzed by hydrogen or hydroxide ions. The reactions can occur in both the sorbed and dissolved phases and are pH dependent. ⢠Reduction-oxidation: Redox reactions for pesticides involve the subtraction or addition of electrons. Many such reactions are microbially mediated. Redox reactions are strongly influenced by the oxidation status of soil/groundwater systems. ⢠Biodegradation: Pesticide transformation through biological deg- radation is perhaps the most significant process. Rates may be influenced by microbial growth rates and the availability of nutri- ents. In surface soils, this process may be dominant; groundwater systems may be much less biologically active. ⢠Sorption: Pesticide interaction with soil and aquifer materi- als can retard movement by partitioning the chemical between dissolved and sorbed states. In some cases, kinetic limitations influence sorption. The magnitude of sorption will often depend on sorbent properties. ⢠Plant uptake: Some pesticides are removed from soil solutions by plant roots. Very soluble, highly mobile chemicals can be removed from soil-water to a significant degree. ⢠Advection: Pesticide movement in bulk water flow is the ma- jor mode of transport into and through groundwater systems. Recharge and subsequent advection through surface soil layers are determined by the water balance imposed by local hydrol- ogy. Movement within groundwater aquifers may be dominated by regional flow fields responding to gradients including recharge. ⢠Dispersion: Pesticide movement also occurs because of disper- sive mixing within the porous media. Molecular diffusion and advective mixing combine to spread contaminant plumes. ⢠Volatilization: Pesticide escape to the atmosphere is important in surface soil layers and may play a dominant role in determining the mass of pesticide available for leaching.
71 The processes just described are dependent upon both chemical and environmental properties. Chemical properties and specific rate constants can be determined within the laboratory, but the problem remains to characterize the environments into which the pesticide is placed. Soil and aquifer properties influencing transport include porosity, bulk density, hydraulic conductivity, and dispersiveness. For sorption and transformation processes properties such as organic carbon content, microbial population density, pH, redox status, and temperature must be known. Such properties must be measured for site-specific assessments. For assessments over regions or other large areas, soil and subsurface data bases must be analyzed statistically and maps must be generated. Pesticide use patterns and application technology determine the location and manner in which chemicals become part of the envi- ronment. This information is key to identifying potential problems for chemicals already in use and provides a focus for analysis of new chemicals not yet in use. Techniques such as map overlays showing the alignment of high recharge areas, e.g., sandy soils in high rain- fall areas, are useful in assessments. Another important factor is the mode and spatial distribution of pesticide applications within the tar- get areas. Surface applied chemicals are subject to removal through volatilization and photo-reactions. Soil-injected chemicals may be more amenable to leaching, depending upon chemical properties. Predictive needs in assessment require the development of math- ematical models that systematically combine the transport and fate processes influencing movement to groundwater. The hydrological cycle must be represented in the form of water balance equations and coupled to mass transport or advection/dispersion equations. Often the equations are highly nonlinear and require numerical solu- tions. Models developed from process descriptions can be laboratory- validated, but their performance under field conditions must be demonstrated. Detailed field monitoring studies that enable com- plete characterization of the soil environment and subsequent mea- surement of the fate and movement of pesticides are necessary. Com- parison of these data with the predicted values produced by models demonstrate the uncertainties inherent in the modeling process and may identify model errors or incomplete understanding of the under- lying processes. Models that have been field-validated can be combined with environmental data bases, and assessments at several geographical
72 scales can be completed for specific chemicals. Specific sites may be of some interest, but most often the concern is the risk of groundwater contamination over large areas. Thus, application to a full range of conditions is desirable. Finally, these demonstrations serve as the means to communicate and transfer research results to those charged with chemical regulation or to those interested in more refined development of environmentally acceptable chemicals. All phases of these research approaches are currently underway within EPA. The multidisciplinary approach of combining elements of laboratory, field, and mathematical analyses has produced ground- water assessment tools now in use for regulating pesticides. CURRENTLY USED ASSESSMENT TOOLS EPA requires that each chemical submitted for registration be tested for its physical and chemical properties and its fate charac- teristics. Typically, laboratory and small-plot experiments to mea- sure rate constants, sorption properties, and transformation prod- ucts identified above are required. Guidelines have been issued that summarize the experimental protocols as currently defined by the research process. In cases where commercial chemicals are known to be a problem, monitoring studies are required to better define the problem. Re- cently, a nationwide statistical survey was initiated to evaluate more fully the extent of contamination in well water supply systems. Chemical-specific data are used in models to estimate risks posed by the proposed pesticide use. The Pesticide Root Zone Model (PRZM) is used to predict leaching to groundwater for a wide range of geograpical and climatic conditions. The PRZM computes the daily percolation of rainfall or irrigation that moves below the root zone and the associated quantity of pesticides. Modeled processes include runoff, evapotranspiration, plant uptake, sorption, degra- dation, advection, and dispersion. Detailed model documentation, user's manuals, and the computer code for implementation on either micro-, mini-, or mainframe computers are available. A complete assessment methodology that includes the fate of leached chemicals in groundwater is not yet in general use. Such a system is under development, however, and research has shown that the integrated approach is quite feasible.
73 AN ASSESSMENT EXAMPLE Aldicarb is used as a systemic pesticide for the control of ne- matodes and other insects in citrus production. A dominant citrus production area is the state of Florida where the subsurface system is characterized by shallow, porous sands overlying three different aquifer systems. The assessment approach adopted for aldicarb in- cluded a monitoring and modeling study. Monitoring was completed to determine the current extent of drinking water contamination of municipal water supplies pumping groundwater. Large systems serving highly populated areas in regions of high aldicarb use were sampled. No widespread contamination was found and the majority of the population was not at risk. Previous monitoring studies had discovered well water contamination in shallow wells in rural areas serving only one or two households. An estimate of the risk to exist- ing, small systems and an evaluation of the best policy to reduce such risks were accomplished through a mathematical modeling study. Process studies of aldicarb demonstrated the degradation path- ways. Laboratory and field-plot studies were conducted to evaluate each rate constant. The environmental properties that influence aldicarb leaching and transport included precipitation, evaporation, soil water content, hydraulic conductivity, soil organic matter, pH, and gradient. Statistical characterization of these properties was combined with citrus production and chemical use data to delineate systems requiring analysis as shown in Figure 1. The system depicted in Figure 1 combines the surface soil zones where aldicarb is applied and the groundwater aquifer systems im- portant as water supply sources. Each of the areas was modeled by combining the PRZM with a groundwater transport model to predict the risk to individual wells downgradient from treated areas. Typical results are given in Figure 2. These results were used to compare expected concentrations of aldicarb at different distances from the treated area to a health-based standard shown on the figure as 42 ug/1. Using this information, a regulation was developed that allows aldicarb use only on citrus production located more than 1000 feet from the nearest water supply well. The aldicarb assessment for Florida citrus production demon- strates how quantitative assessment methods are used to identify, control, and prevent pesticide contamination of groundwater. Mon- itoring and modeling were combined to assess current and future risks. The process-based behavior of aldicarb was investigated using
74 ALABAMA Jacksonville Atlantic Ocean Explanation 1 - Roridan Aquifer Alone - Unconfined 2 - Leaky Two-Aquifer System - Surficial Aquifer - Roridan Aquifer 3 - Leaky Three-Aquifer System - Surficial Aquifer - 1ntermediate Aquifer - Roridan Aquifer 4 - Surficial Aquifer Alone - Unconfined 5 - Two-Aquifer System - Surficial Aquifer - 1ntermediate Aquifer FIGURE 1 Overlap of idealized unsaturated and saturated areas in Florida. laboratory methods developed by researchers. The relevant environ- mental and chemical use data were developed from soil data bases and census data. Finally, the results were communicated to regula- tors who developed the control policies designed to permit continued use of aldicarb while preventing unacceptable groundwater contami- nation.
75 iooo r < EC LLl o 42 ug/l No Decay 200 400 600 D1STANCE (feet) 800 1000 FIGURE 2 Concentration in downgradient well from citrus groves for the worst case year in ten years.
