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.
8 Using Models to Solve Ground Water Quality Problems JAMES M. DAVIDSON AND P. S. C. RAG A comprehensive national survey to assess ground water con- tamination has not been undertaken to date; yet the contamination of portions of various aquifers in different states is well documented (Holder, 1986; Pye et al., 1983~. The reason for this contam~na- tion, as well as the potential for further contamination, is rooted in the fact that nearly all facets of modern life (urban, industrial, agricultural, etc.) use and/or discard chemicals on a daily basis. Because of the ubiquity of man-made chemicals in our society, the potential for ground water contamination is real, and the problem is one of national, regional, state, and local interest. Not only are regulatory and policymaking agencies faced with monitoring the presence of ground water contaminants, but they are also responsi- ble for developing policies that will prevent further contamination of this valuable and limited natural resource. The use of mathematical models to simulate the behavior and movement of water and toxic chemicals in water-unsaturated and saturated porous media for ground water management is a con- troversial issue within the scientific community as well as among users of these models. Yet many of those responsible for ground water quality protection consider models to be the most pragmatic approach to a complex problem. To complicate the issue further, there is urgency in the problem facing regulatory and policy agen- cies and their need to respond in a manner that is environmentally and fiscally objective. Some of those developing models, as well as those who promote the use of models for ground water man- agement, speak positively about their ability to accomplish this task. 139
140 HAZARDOUS WASTE SITE MANAGEMENT This manuscript briefly discusses the general types of models that are available, their potential role in managing and protecting ground water, the concerns of those who use models for policy issues, and risk assessment. These topics are raised in hopes of helping the reader focus on and understand specific issues and alternatives and the need for a response to the problem of ground water protection. MODEIS FOR P=DICT~G MO~MENT AND FATE AND/OR RAN11~G RISK OF C=MICAlS ~ GROW WATER There is no mode} that will adequately describe all ground water quality problems because the assumptions and simplifica- tions generally associated with models do not adequately mimic all the processes that influence the movement and behavior of the water and/or the chemicals of interest. This is especially true for the chemical and biological processes that influence the movement and fate of chemicals in porous media. Although major advances have been made in recent years in our understanding of the behav- ior of chemicals in water-unsaturated and saturated porous media, research in this area is still in its infancy. This is especially true for those cases in which water-miscible organic solvents may enhance the mobility of selected organic chemicals, in which immiscible sol- vents exist, and in which chemical movement occurs in fractured rock and well-structured soils. At least two distinct modeling approaches can be identified. In the first, models are needed to provide site-specific predictions of the behavior of a particular chemical. This approach requires the use of sophisticated mathematical models that explicitly ac- commodate site-specific characteristics (e.g., hydrogeology, soils, chemical loading) in sufficient detail so as to provide predictions about the spatial distribution of a chemical's concentration or flux, or both, in porous media. Thus, there are numerous input coeffi- cients required for this type of mode! to function. An example of a situation in which such a mode! might be used is the prescription of remediation actions for a hazardous waste disposal site covered under the Comprehensive Environmental Response, Compensa- tion, and Liability Act (CERCLA). Because of the large financial and technical resources generally available for remediation of a Superfund site and the fact that the problem is narrowly focused, one may be justified in using a large, complex model. Also, this
USING MODELS 141 application makes feasible the possibility of continued refinement and calibration of the model using site monitoring data. In contrast, the second modeling approach seeks to establish regulatory policies that will prevent or minimize ground water contamination on a regional, state, or national level. For this case the use of a complex mode! is unrealistic, primarily because of the large quantity of input data that must be provided and the spatial variability of the area under consideration. Also, for these cases, one is generally more interested in evaluating the potential of a particular aquifer to become contaminated whether certain chemicals pose a greater threat than others to a ground water supply. For such applications (e.g., development of regulations and permitting policies), the simulation of a chemical's distribution over time in porous media may not be essential. Rather, it is the relative behavior of the chemical that is of interest. For such applications, simplifications of the more complex models may be adequate. When considering which of the above modeling approaches are better suited to addressing a potential ground water contamination problem, three questions should be asked: 1. How likely is it that a particular aquifer (or a portion of an aquifer) may be contaminated? To answer this question, it is necessary to assess the site vuinerabitity of an aquifer to become contaminated. 2. Given the use patterns of chemicals and their physical, chemical, and biological properties, which chemicals are most likely to intrude ground water? This answer will require deter- mining the contamination potential of a group of chemicals. 3. If a specific chemical or group of chemicals have already contaminated an aquifer, are the concentrations of sufficient mag- nitude to pose an adverse health risk? This question is answered after evaluating the toxicological potency of the chemical. Aquifer vulnerability may be assessed on the basis of the physiographic setting and the hydrogeologic characteristics of a site. The contamination potential may be estimated from the chemodynamic properties of the chemicals of interest. Finally, toxicological potency may be deterrn~ned by comparing the action level set by the Office of Safe Drinking Water with that measured or predicted through the use of a model. Various schemes for
142 HAZARDOUS WASTE SITE MANAGEMENT dealing with these three factors are reviewed in the paragraphs that follow. Empirical approaches rather than simulation models are fre- quently used to rank a site's vulnerability or contamination poten- tial. Aller et al. (1985) have proposed a numerical rating technique called DRASTIC for evaluating the likelihood of ground water con- tamination at a specific site, given the site's geohydrologic setting. The acronym for this rating technique is derived from the seven factors considered in the rating scheme: (1) depth to ground water, (2) recharge rate, (3) aquifer media, (4) soil media, (5) topography, (6) impact of vadose zone, and (7) conductivity of the aquifer. A combination of weights and rating is assigned to each of these factors, and a numerical rating called the DRASTIC index is calculated for a site or area of interest. The DRASTIC scheme currently is being used to design and guide EPA's national survey for ground water contamination (Alexander et al., 1986~. The Arizona Department of Health Services (1982) and the Florida Department of Agriculture and Consumer Services (1986) have developed numerical rating schemes to establish lists of "pri- ority pesticides" that pose a threat to ground water. The proce- dures are based on assigning numerical values to pesticide proper- ties (solubility and persistence), the quantities of pesticides used in the state or local region, and the human health effects. Other approaches based on the numerical ratings of several factors have been proposed for evaluating the suitability of sites for land dis- posal of hazardous wastes and/or the application of contaminants to a site (Seller and Canter, 1980; Gibb et al., 1983; LeGrand, 1983; Michigan Department of Natural Resources, 1983; U.S. EPA, 1983~. These ranking schemes, in the strictest sense, are not descriptive models. Comprehensive mathematical models that include procedures for describing each process influencing the movement, sorption, degradation, and transformation of a specific chemical are more complex than ranking schemes. A primary question of concern in the use of complex simulation models is how chemical processes that occur at the interstitial level of porous media are represented on a field scale. Cherry et al. (1984) reviewed this subject for both organic and inorganic chemicals in ground water systems. Currently, there are insufficient field data available on the behav- ior of chemicals to test the available simulation models properly. Most information regarding processes has been collected under
USING MODELS 143 controlled laboratory conditions and, in general, under equilib- rium or steady-state conditions. Thus, any attempted simulation encounters the problem of the reliability of extending laboratory- scale behavior to field conditions. Of all the processes responsible for organic chemical attenua- tion within the unsaturated and saturated zones, only biologically mediated transformations may lead to complete degradation of organic chemicals. Although the capacity for microbial degrada- tion in surface soils has been studied extensively, the character- ization of microbial activity in the vadose zone and in aquifers has received considerable attention only recently. Given the olig- otrophic conditions of the vadose zone and deep aquifers, these areas were believed not to support or sustain significant microbial populations. As evidence gradually accumulates to suggest that diverse and active microbial populations can survive and func- tion in aquifers, it has been proposed that aquifers might have a certain "assimilatory capacity," that is, the ability to degrade chemicals to some acceptable concentration. Thus, the feasibility of biodegradation as an in situ aquifer restoration technique is currently being explored. Comprehensive process-level models are being formulated, but their validation and integration into larger ground water models are far from complete. We must also note that even through biodegradation the zero concentration of a con- taminant is approached asymptotically. Thus, this technique may only be acceptable for noncarcinogenic chemicals and then only for those with a large maximum contaminant level (MCL) or max- imum contaminant level goal (MC[G). As long as the legislative history of the Safe Drinking Water Act justifies zero concentra- tions for carcinogens in ground water, biodegradation as an in situ remediation technique may not be a viable alternative. The foregoing discussion of simulation models raises some significant questions about their reliability for predicting and/or ranking site vulnerability or contamination potential. Given the seriousness of the problem and the risk to human health that is involved, Wagenet (1986) concluded the following: Our current understanding of the basic principles that determine pesticide fate in the field is incomplete, yet we must make decisions now considering pesticide regulation and management. Current pesticide models used by regulators and academics represent the best tools we have to estimate pesticide fate as a function of soil, climate and management factors. Yet, we have every indication
144 HAZARDOUS WASTE SITE MANAGEMENT that their predictions are not universally reliable, and almost no proof of their credibility in the field. The question Is whether we can feel comfortable about the predictions produced by these models, or whether we should abstain from their Me as predictive tools until their credibility is better established. A healthy and continuing intellectual argument is in progress on this issue, and will probably persist for some time. During this debate, the use of existing models for regulatory and management purposes will continue, and will result In some good decisions, and probably some mistakes. Several points are clear. First, no pesticide model exists that has been proven to estimate consistently and accurately the spatial and temporal distribution of pesticide concentrations in the unsaturated zone. This is true regardless of the resolution used to represent basic principles in the model, and whether the model falls into the research or management category. Second, it follows that current models should be used only to compare the relative, not absolute, behavior of pesticides in field soils. Third, the first two points indicate that our approach to modeling pesticide fate in unsaturated field soils must change if we are to develop a new generation of pesticide models that do not suffer from the limitations of the current models. (pp. 339-340) Although this statement pertains to mathematical models that describe the fate and transport of pesticides, it eloquently summa- rizes the issues pertinent to the application of simulation models for most organic pollutants. Quality Control/Quality Assurance for Models Before any mode} is used to describe or simulate the behavior of a chemical in porous media, its validity should be indepen- dently established by some individual or institution other than its developer. Code testing is generally considered to encompass verification and validation of the mode! (Adrion et al., 1982~. To evaluate ground water models in a systematic and consistent man- ner, some institutions have developed model review, verification, and validation procedures (Morgan and Mezga, 1982; van der HeijUe et al., 19853. Generally, the review process is qualitative in nature, whereas code testing results are evaluated by quantitative performance standards. In April 1984, EPA Order 5360.1, "Policy and Program Re- quirements to Implement the Mandatory Quality Assurance Pro- gram," was issued and for the first time provided a regulatory basis for the agency's quality assurance program. Quality assurance is
USING MODELS 145 the procedural and operating structure required in mode} develop- ment to ensure technical execution of all aspects of the model. The primary goal of the EPA ground water modeling quality assurance program is to ensure that all modeling efforts supported by EPA are of a known and scientifically acceptable quality in terms of computer code, documentation, and operation. Concerns of Those Using Water Quality Models In a recent report by van der HeijUe and Park (1986), the fol- Towing concerns and needs were identified by national and regional EPA staff using water quality models: ~ a limited knowledge of what types of water quality models are available; . the need for assistance in selecting and using available models for specific sites; . guidance in model reliability and interpretation of simula- tions; ~ the need for additional models for multiphase flow and contaminant behavior in the vadose zone; . improved interaction and communication with technical staff located in other regional offices, headquarters, and EPA laW oratories; ~ training in basic processes (e.g., geology, hydrology, fate and transport, geochemistry) for the project officers as well as modeling training for technical experts in the region; and ~ the hiring and retention of technical staffwho have received special training in modeling. Role of Ground Water Quality Models in Regulatory or Policy Issues A policy for resource protection based on monitoring is by nature reactive, not preventive. Policies and regulations based on models, however, can be both preventive and reactive. A discus- sion of some of the principal areas in which mathematical models can and are being used to assist in managing EPA state and/or local government ground water protection programs follows.
146 HAZARDOUS WASTE SITE MANAGEMENT Development of Regulations and Policy Evaluation of the impacts (economic, health risk, etc.) of regulations on policy scenarios requires process-oriented generic models. Some specific uses of such models in the evaluation of ex- isting or proposed policies and regulations include: (1) developing standards for well setbacks with respect to pesticide applications and waste disposal sites, (2) evaluating the potential impact of various types of ~failures" of injection wells, (3) providing techni- cal justification for restricting land disposal of hazardous wastes at specific sites, and (4) evaluating the need and effectiveness of ground water monitoring programs for hazardous waste injection wells. Permitting Ground water models are being used on a site-specific basis by owners/operators of hazardous waste facilities to show compli- ance with permit requirements; they are being used by regulatory agencies to validate information provided for permitting purposes. Models are also being used to evaluate hazardous waste site char- acteristics to determine the optimal locations for monitoring wells, to estimate the transport and fate of contaminants below a waste disposal site, and to assess corrective action should a failure occur. EPA's Office of Waste Programs Enforcement (OWPE) is currently using models to evaluate the following source types: sanitary landfi~Is; municipal, industrial, and mining surface im- poundments; underground storage tanks; septic tanks; agricul- tural feedIots; road de-icing chemicals; hazardous waste landfi~Is; and hazardous waste surface impoundments. OWPE is also in- vestigating the use of ground water modeling for fund-financed CERCLA actions, with an emphasis on using simple, desktop fate and transport calculations to predict leaching to ground water from residual soils at Superfund sites. EPA's Office of Pesticide Programs (OPP) is using models to assess the leaching potential of registered pesticides as well as to evaluate new pesticides prior to registration. Past and present modeling efforts have focused on predicting whether various pes- ticides are likely to leach to ground water under normal usage. This focus results from the fact that pesticides are considered a
USING MODELS 147 nonpoint loading problem by OPP as opposed to a point source (usually the case for other EPA program offices). Remedial Action Ground water models are being used increasingly in the CERCLA response process for remediation of the potential re- lease of hazardous substances. A typical model application for a Superfund-financed or enforcement-related remedial response action includes both site investigation to assist in problem defini- tion and system conceptualization to identify the contamination source and to predict future contamination and health risk. Mod- els are also being used to develop and evaluate remedial alterna- tives during the remedial investigation/feasibility study stages and to analyze design specifications for remedial action alternatives. In addition, models are frequently being used to assess required cleanup levels, set the level of required source removal, and project performance characteristics for remedial action design as well as formulate postoperation and closure requirements. Risk Assessment Risk to human health and the environment owing to the pres- ence of trace concentrations of toxic chemicals in ground water used for domestic purposes is a matter of major concern to modern society. However, the detection of a hazardous chemical in ground water is not necessarily a cause for immediate alarm. Modern chemical detection techniques are so sensitive that it is possible to detect concentrations at levels below federal or state action levels (MCEs or MC[Gs). In fact, chemical detection techniques are constantly improving, causing detection levels to be pushed lower and lower. For chemicals that pose acute toxicity problems, the no observable adverse effect level (NOAEL) may be established; for chemicals that induce chronic effects, however, NOAEI, values are less well defined and generally contain a safety factor of two to three orders of magnitude. Thus, concentrations like these that induce chronic effects may be quite low, and at times they may fall below current analytical detection levels. For those chemicals that have been shown to be carcinogenic, there is no acceptable concentration level; however, an exposure risk may be calculated using various exposure models.
148 HAZARDOUS WASTE SITE MANAGEMENT SUMMARY AND CONCLUSIONS Mathematical models for describing the fate and transport of chemicals in porous media and numerical ranking schemes for assessing site vulnerability and contamination potential, although not satisfactorily validated to date, appear to have a guarded role in policymaking, the development of environmental regulations, and the establishment of remedial actions for regulatory agencies responsible for ground water. The models should not, however, be used without some monitoring effort for the purpose of validation and/or calibration. The extent to which such efforts are conducted will depend on the purpose of the model as well as the areal extent to which the model is expected to be representative. The alterna- tive to modeling is one of reaction through an extensive soil and ground water monitoring program, a position that is not realistic if pristine ground water conditions are the anticipated goal. The release of a chemical to the soil surface will eventually result in some portion of it reaching the ground water, be it a large con- centration or a very Tow concentration. Such attenuation results from degradation, sorption, and volatilization. Complete chemical containment or stabilization is the only waste disposal procedure currently available that provides ground water protection; yet even these procedures are subject to engineering failures. REFERENCES Adrion, W. R., M. A. Branstad, and J. C. Cherniasky. 1982. Validation, verification and testing of computer software. ACM Computing Surveys 14:159-192. Alexander, W. J., S. K. Liddle, R. E. Mason, and W. B. Yeager. 1986. Groundwater Vulnerability Assessment in Support of the First Stage of the National Pesticide Survey. Washington, D.C.: EPA. Aller, L., T. Bennett, J. Lehr, and R. Petty. 1985. DRASTIC: A Stan- dardized System for Evaluating Groundwater Pollution Potential Using Hydrogeologic Settings. Washington, D.C.: EPA Office of Research and Development. Arizona Department of Health Services. 1982. Pesticides with groundwa- ter pollution potential in Arizona. Prepared by Rich and Associates, Phoenix, Arizona. Cherry, J. A., R. W. Gillham, and J. F. Barker. 1984. Contaminants in Groundwater: Chemical Processes. Pp. 46-64 in Groundwater Contamination, J. D. Bredehoeft, panel chairman. Washington, D.C.: National Academy Press.
USING MODELS 149 Florida Department of Agriculture and Consumer Services. 1986. Pesticide assessment procedure: An assessment procedure bared on pesticide and hydrogeologic factors developed for selective pesticide monitoring of groundwater in Florida. Prepared by the Pesticide Review Council. November. Gibb, J. P., M. J. Barcelona, S. C. Schock, and M. W. Hampton. 1983. Hazardous Waste in Ogle and Winnebago Counties: Potential Risk Via Ground Water Due to Past and Present Activities. Document No. 83/26. Illinois Department of Energy and Natural Resources. Holden, P. W. 1986. Pesticides and Groundwater Quality: Issues and Problems in Four States. Washington, D.C.: National Academy Press. LeGrand, H. E. 1983. A standardized system for evaluating waste-disposal sites. National Water Well Association, Worthington, Ohio. Michigan Department of Natural Resources. 1983. Site assessment (SAS) for the Michigan priority ranking system under the Michigan Environmental Response Act. Ann Arbor. Morgan, M. S., and L. J. Mezga. 1982. Evaluation factors for verification and validation of low-level waste disposal site models. DOC/OR/21400- T119. Oak Ridge, Tenn.: Oak Ridge National Laboratory. Pye, V. I., R. Patrick, and J. Quarels. 1983. Groundwater Contamination in the United States. Philadelphia: University of Pennsylvania Press. Seller, L. E., and L. W. Canter. 1980. Summary of selected groundwater quality impact assessment models. Report No. NCGWR 80-3. National Center for Ground Water Quality Research, Norman, Oklahoma. U.S. EPA. 1983. Surface Impoundment Assessment National Report. EPA- 570/9-84-002. Washington, D.C. Van der Heijde, P. K. M., P. S. Huyakorn, and J. W. Mercer. 1985. Testing and validation of groundwater models. In: Practical Applications of Groundwater Modeling. Proceedings of the NWWA/IGWMC Confer- ence, Columbus, Ohio, August 19-20. Van der Hedge, P. K. M., and R. A. Park. 1986. U.S.E.P.A. ground- water modeling policy study group: Report of findings and discussion of selected ground-water modeling issues. International Ground Wa- ter Modeling Center, Holcomb Research Institute, Butler University, Indianapolis, Indiana. Wagenet, R. J. 1986. Principles of modeling pesticide movement in the unsaturated zone. Pp. 33~341 in Evaluation of Pesticides in Ground- water, ed. W. Y. Garner et al. Symposium Series 315. Washington, D.C.: American Chemical Society. PROVOCATEUR'S COMMENTS Ishwar P. Murarka ~ have known Jim for a while. Some of my comments are motivated by that. The rest are motivated by the paper itself. Let me make several comments to raise underlying questions for
150 HAZARDOUS WASTE SITE MANAGEMENT models, modelers, and mode! users. I really think the problem is that of predicting or assessing ground water quality and not quantity. So can ground water quality problems be addressed by models? To answer this question, we need to address the following: (1) Who are the users of these models? (2) Who expects what from modem' and those uses? (3) How good should the models' performance be? Let me interject here that ground water quality problems are not going to be solved by models. I,iabilities are not going to be assigned by models. What models will do is provide some skills and analyses and some answers to "what if" questions that can be used for discussion regarding the nature and extent of ground water quality problems. If a mode} is incomplete, the corresponding uncertainties will be reflected in its predictions, and the discussions will have to recognize that. If the.models are complete but the users do not know how to use them, then the ground water quality problems are neither defined nor solved. The next issue raised by the paper is that of discarding mate- rials that contain chemicals. Is the question one of discarding or not discarding, or is the real question one of proper or improper management of the discarded materials? These distinctions have different implications for ground water quality problems and the use of models. Models used to simulate ground water contami- nation for improperly disposed wastes will give a very different answer than when the same models are used to simulate ground water quality changes that result from a well-managed disposal facility. The issue is not one of using models or not using models but rather the accuracy, precision, and reliability of models. Con- tributing to the problem is a lack of objective performance re- quirements for models and mode! users. Is the problem the accu- racy/completeness of the models, or is the issue that of availability of data to use models? ~ will be brave here and state that ~ can predict with and without models. But why should you or ~ believe any one of my predictions and for what use? This is the area in which the importance of models, modelers, and mode! users cannot be overemphasized. If we want proper answers to our Questions. we must ask those questions clearly and make known a priori the degree of confidence we require in those answers. Let me conclude my comments by stating that models are also used for summarizing and organizing data. Indeed, models
USING MODELS 151 are nothing more than mathematical descriptions used by those who know how to use them. These same models, however, are also available for use by those who do not quite know how to use them. As a result, we develop perceptions or labels such as toxic substances or hazardous materials. Scientifically though, ~ have to leave you with this question: Are we all of a sudden concerned about a "toxic or hazardous" chemical or is the concern really that of "quantities" of a chemical that causes adverse and unacceptable biological effects?