Are We Cleaning Up? 10 Superfund Case Studies (June 1988)

Home , Cleaning Up (The Wire), List of Superfund sites

Are We Cleaning Up? 10 Superfund Case Studies

June 1988

NTIS order #PB89-139018 Recommended Citation: U.S. Congress, Office of Technology Assessment, Are We Cleaning Up? 10 Super- fund Case Studies–Specia] Report, OTA-ITE-362 (Washington, DC: U.S. Government Printing Office, June 1988).

Library of Congress Catalog Card Number 88-600545

For sale by the Superintendent of Documents U.S. Government Printing Office, Washington, DC 20402-9325 (order form can be found in the back of this report) Foreword

In a few years, Superfund became part of the American vocabulary because so many people feel so strongly about toxic waste. and cleanup of contaminated sites. They remain worried about environmental and health effects, but a new concern has come to the fore: the enormous amount of money and the long times to clean up an ever-growing list of Superfund sites. Yet, even while the public demands effective cleanups, nearly everyone speaking and writing about Superfund seems to feel that serious problems exist. And the focus of public attention has shifted from how much money ought to go to Superfund to how to achieve environmental results and efficiency. Right now there are more questions than answers about diagnosing and f!xing Superfund. Four committees of Congress asked the Office of Technology Assessment to assess how Superfund is being implemented under the 1986 Superfund Amend- ments and Reauthorization Act. They asked OTA to examine a number of technical issues that arise near the beginning of the complex Superfund process. The study was to assess the impacts of statutory provisions and program policies on environmental effectiveness and economic efficiency. The requesting committees were: the House Committee on Public Works and Transportation and its Investigation and Oversight Subcommittee; the House Energy and Commerce Committee and its Oversight and Investigations, and Transportation, Tourism, and Hazardous Ma- terials Subcommittees; the Subcommittee on Superfund and Environmental Over- sight of the Senate Environment and Public Works Committee; and the Subcom- mittee on Environment, Energy, and Natural Resources of the House Government Operations Committee. During our Superfund Implementation assessment we realized that we could learn much by finding out how sites progress through the Superfund program and how—and when—critical decisions about their cleanup are being made. Before we could answer tough but general questions about making Superfund work better, we had to know more about what was actually going on. This special report presents 10 case studies of recent Superfund decisions at sites which OTA believes, from surveying over 100 recent cleanup decisions, to be representative of a broad range of contamination problems and cleanup technologies. We hope that everyone affected by Superfund can learn as much as we have from these case studies. Many people have helped OTA with these case studies, especially Environmental Protection Agency staff around the country who provided us with primary information about the sites. Several companies that are responsible parties at sites also provided key documents. Responsibility for the contents of this document, of course, rests with OTA,

JOHN H. GIBBONS Director

. . . Ill Superfund Implementation Advisory Panel

David Marks, Panel Chairman Massachusetts Institute of Technology Kirk Brown Robert G. Kisse]] Texas A&M University E.I. du Pent de Nemours & Co., Inc. Richard Brownell Steve Lester Malcolm Pirnie, Inc. Citizens Clearinghouse for Hazardous Waste William Child William Librizzi Illinois Environmental Protection &encv“ New Jersey Dqpt. of Environmental Henry Cole Protection Clean Water Action and National Campaign Against Toxic Hazards George Muhlebach Ciba-Geigy Corp. E. William Colglazier Waste Management Institute Robert N. Ogg University of Tennessee CH2M Hill William B, DeVille Robert B. Pojasek Louisiana Department of Environmental ChemCycle Corp. Quality Frank J. Veale, Jr. Deborah Hankins Texas Instruments, Inc. General Electric Corp. James N. Welsh Bob Harris Thermal Kinetics ENVIRON Corp. William N. Hedeman Beveridge & Diamond, P.C.

NOTE: OTA appreciates and is grateful for the valuable assistance and thoughtful critiques provided by the advisory panel members. The panel does not, however, necessarily approve, disapprove, or endorse this report. OTA assumes full responsibility for the report and the accuracy of its contents,

iv OTA Project Staff–Are We Cleaning Up: 10 Superfund Case Studies

Lionel S. Johns, Assistant Director, OTA Energy, Materials, and International Security Division

Audrey B. Buyrn, Manager Industry, Technology, and Employment Program

Joel S. Hirschhorn, Project Director

Kirsten U. Oldenburg, Deputy Project Director David A. Dorau, Analyst Karen L. Jordan, Research Assistant Lynn M. Powers, Editor

Administrative Staff Edna M. Thompson, Administrative Assistant Diane D. White, Administrative Secretary

v Page Summary and Analysis...... I Introduction ...... 1 The Importance of the Record of Decision ...... z The Usefulness of Case Studies ...... 2 Superfund’s Better Side...... 5 Summary of Trends From 10Case Studies ...... 7 IO Case Studies ...... 16 Introduction ...... 16 Methodology ...... 16 Format for Case Studies ...... 16 Case Study l: Chemical Control Corp., Elizabeth, NJ; EPA Region 2 ...... 18 Case Study 2: Compass Industries, Tulsa County, Oklahoma; EPA Region 6 . . 25 Case Study S: Conservation Chemical Company, Kansas City, Missouri, EPA Region 7 ...... 29 Case Study 4: Crystal City Airport, Crystal City, Texas, EPA Region 6 ...... , 34 Case Study5: Industrial Excess Landfill Uniontown, Ohio, EPA Region 5 . . . 41 Case Study6: Pristine, Incv Reading, Ohio, EPA Region5 ...... 44 Case Study7: Renora, Inc; Edison Township, New Jersey, EPA Region 2... . 55 Case Study 8:Sand Springs, Petrochemical Complex, Tulsa County, Oklahoma, EPA Region6 ...... 59 Case Study 9: Schmalz Dump Site, Harrison, Wisconsin, EPA Region 5...... 66 Case Study 10: Tacoma Tar Pits, Tacoma, Washington, EPA Region 10...... 71

Boxes Box Page l. How Does Superfund Operate? ...... 3 2. 10)Case Study Sites With Capsule Findings ...... 9

Table Table Page l. Times for Sites To Reach Points in the Superfund Process ...... 13 SUMMARY AND ANALYSIS Introduction cover what is wrong and fix it. Whether Super- fund will work cost-effectively over the long term depends on how cleanup technologies are Are we cleaning up the mess or messing up evaluated, matched to cleanup goals, selected, the cleanup? In the eighth year of Superfund, and implemented and how permanent the clean- this central question is still being asked. These ups will be. People want their cleanups—the 10 case studies illustrate how the Environmental ones they live near or pay for—to last. Improv- Protection Agency (EPA) is implementing the ing public confidence in Superfund can be ap- Superfund Amendments and Reauthorization proached from different directions, including Act (SARA) of 1986. OTA has examined a great the one taken in this report: making better de- many more sites and believes these case studies cisions about cleanup technology. are representative of what is happening nationwide in the Superfund program. Too much flexibility and lack of central management control are working against an effec- This report examines two fundamental ques- tive, efficient Superfund program. EPA Regions, tions about using technology to cleanup toxic contractor companies, and workers have sub- waste sites. First, is the Superfund program stantial autonomy. In principle, flexibility can consistently selecting permanently effective lead to benefits. But the case studies show the treatment technologies which, according to Superfund program as a loose assembly of dis- SARA, are preferable because they reduce “tox- parate working parts; it is a system of divided icity, mobility, or volume” of hazardous wastes? responsibilities and dispersed operations. There The answer OTA finds is that it is not. is no assurance of consistently high quality Second, are land disposal and containment, studies, decisions, and field work or of active both impermanent technologies, still being fre- information transfer. The need for cleanups, quently used? The answer we find is yes. Fu- the newness of the technological challenge, and ture cleanups are likely for the wastes left in the growth of Superfund mask the inexperience the ground or shipped to landfills. and mobility of the work force. Program man- agers have not offset inexperience in technical The Superfund program promised a lot. Peo- areas and management with tight management ple’s expectations have been high, perhaps too controls and intensive educational programs high for such a new, complicated, large-scale for government and contractor workers. Over- effort. Frustration often makes it difficult to see simplified “bean counting” of results instead real Superfund accomplishments. Since its in- of evaluations of what those results mean tech- ception at the end of 1980, Superfund has re- nically and what they accomplish environ- ceived a great deal of money, over $5 billion mentally provides too little incentive for qual- so far, to clean up the Nation’s worst toxic waste ity work. The current decentralized system also sites. But OTA’S research, analysis, and case does not assure higher levels of program effi- studies support the view shared by most ob- ciency over time, even though some workers servers—including people in affected communi- and offices may become much more effective ties and people in industry paying for cleanups and efficient. —that Superfund remains largely ineffective and inefficient. Technical evidence confirms A widespread belief among Superfund work- that, all too frequently, Superfund is not workers is that “every site is unique.” There is a ker- ing environmentally the way the law directs it nel of truth to this belief. Yet uniqueness has to. This finding challenges all those concerned been carried to an extreme and has blocked un- about human health and the environment to dis- derstanding of common site characteristics,

1 . common cleanup problems, common solutions, lamination will be eliminated or reduced to a I and common experiences with site studies and safe level and environmental protection achieved, decisions. Identifying these commonalities is as well as determining cleanup cost. Technol- necessary to understanding how Superfund is ogy selection is the primary focus of this OTA being implemented nationally and understand- report. But the ROD decision is not everything. ing how to improve the program. At the begin- Just as a map is not the territory, a ROD is not ning, when only a few cleanups were addressed, the cleanup. Future analysis of the environ- sites looked very different from each other. mental results of cleanups is necessary to see Now, with hundreds of cleanups examined, it how the ROD strategic plan is implemented. is easier to see the commonalities and to bene- Because cleanups have been fully implemented fit from the experiences to date. The case studies at so few sites and the data are so sparse, this discuss similar experiences at various Super- study does not fully examine actual cleanup ef- fund sites and help illustrate the link between fectiveness and consistency with ROD goals. identifying commonalities and achieving con- But the case studies examine the entire history sistent cleanups. of the sites. And for some of the sites discussed Cleanup costs are major issues in the case here, the technologies selected have failed or studies. In site cleanup decisions, many peo- early work to clean up immediate threats has ple in government and industry want to keep made matters worse for final cleanup. costs as low as possible. Hence, there is a By examining RODS in detail, the function- tradeoff between environmental protection ing of Superfund comes into focus because goals (How clean is clean?) and the cost of the everything that was done before the ROD must remedy selected (Is it cost-effective?). There is be considered and everything to come later also a tradeoff between effective cleanup at must be anticipated. Analysis of RODS offers some sites versus no action at others. These enormous educational value to improve Super- tradeoffs are getting more difficult as more and fund implementation because they represent more sites requiring cleanup are identified. the critical junction between extensive studies SARA’s preference for permanently effective and expensive remedial cleanups. Cleanup treatment technologies—not a requirement that costs vary widely, from several hundred thou- they always be used—makes these tradeoffs sand dollars to tens of millions of dollars. To even harder; it also places more importance on put cleanup costs in perspective, consider the the accuracy of cost estimates and on evalua- simple concept of acreage. Data on 15 of the tions of the permanency of different cleanup cleanups reviewed in this study indicate that technologies. By understanding the capabilities total cleanup costs can reach $500,000 to $1 of different cleanup technologies, it is easier million per acre, to understand how compromises between cost and environmental performance can lead ei- The Usefulness of Case Studies ther to “gold plated” or “band-aid” cleanups. In Superfund, case studies are particularly The Importance of the Record of Decision important because, even after 8 years, cleanup technology is a new and fast-changing field and A crucial step in the complex process of mov- the work force is relatively young and inexperi- ing a site from discovery to remediation (see enced. Recent college graduates are often put in box 1) is the ROD’s technology selection.1 charge of multimillion-dollar projects at EPA. Cleanup technology determines whether con- These people have had no direct experience and no coursework on cleanup, and they have almost IEpA hag said ‘gThe Record of Decision ., , is the centerpiece no one to learn from, as turnover is high. People of the administrative record against which the Agency’s deci- in contractor firms also lack experience. Research sionmaking maybe judged by the courts.” [U.S. Environmental Protection Agency, “Interim Guidance on Superfund Selection papers and technical manuals have significant of Remedy,” Dec. 24, 1986.] limitations too. They are quickly outdated, are 3

Box 1.-How Does Superfund Operate? The Superfund system is complex. Sites are identified and enter an inventory because they may require a cleanup. At this point, or at any time, a site may receive a Removal Action because of emergency conditions that require fast action or because the site could get a lot worse before a remedial cleanup could be implemented. (Most of SARA’s requirements for remedial cleanups do not apply to removal actions, even though removal actions can cost several million dollars and resemble a cleanup.) In the pre-remedial process, sites receive a Preliminary Assessment (PA); some then go forward to a Site Inspection (S1), with some of those sites scored by the Hazard Ranking System (HRS). If the score is high enough, the site is placed on the National Priorities List (NPL) and becomes eligible for a remedial cleanup paid for by the government, if necessary, or by responsible parties identified as having contributed to creating the uncontrolled toxic waste site. Under current procedures, only about 10 percent of sites which enter the system are likely to be placed on the NPL. Some States have their own lists of sites which require cleanup; these often contain sites not on the NPL. NPL sites receive a Remedial Investigation and Feasibility Study (RIFS) to define contamination and environmental problems and to evaluate cleanup alternatives. The public is given an opportunity to comment on the RIFS and EPA’s preferred cleanup alternative. Then, EPA issues a Record of Deci- sion (ROD) which says what remedy the government has chosen and the reasons for doing so; the decision may be that no cleanup is necessary. A ROD may only deal with part of a site’s cleanup and several RODS may be necessary for a site. The ROD also contains a summary of EPA’s responses to public comments. EPA chooses the cleanup goals and technology in the ROD. In actual fact a number of actions involving different technologies are likely to be chosen for any but the simplest sites. The ROD is like a contract in which the government makes a commitment to actions which will ren- der the site safe. If responsible parties agree to clean up the site, they sign a negotiated consent decree with the government; this stipulates the exact details of how the responsible parties will proceed. If the cleanup uses Superfund money, the State must agree to pay 10 percent of the cleanup cost. In the post-ROD process, the site receives a Remedial Design (RD) study to provide details on how the chosen remedy will be engineered and constructed. The whole process ends with the Reme- dia]Action (RA), the actual implementation of the selected remedy. Many cleanups include long-term monitoring to determine whether the cleanup is effective and if more cleanup is necessary. A ROD may be reopened and amended because of new information discovered or difficulties encountered during the design and remedial action. When a cleanup is deemed complete and effective, the site can be delisted by EPA from the NPL.

too theoretical, assume substantial technical grants. These grants have not been available, how- knowledge, are either too detailed or too general, ever; EPA only began accepting applications in and may be biased to boot. Attending conferences April 1988. where new cleanup technologies are discussed The case studies examine the decisionmak- in detail is difficult because of heavy workloads ingprocess, the quality of the information used and limited funds. Moreover, helping to inform in it, and how well the decision and its techni- the public is also critical, especially because cal support are communicated by EPA to the SARA increases the participation of communi- public. Unlike “bean counting” statistics, which ties in the program through technical assistance give quantitative program results for a large number of sites, case studies show how the 2For example, at EPA’s annual research SYmPOShm in MaY complex Superfund system really functions and 1988 dealing-with treatment of hazardous waste only nine EPA illustrate the quality-of its environmental per- staff people who may be implementing Superfund (i.e., not in the Office of Research and Development) were registered out formance. Case studies cannot totally describe of a total of over 700 people. the extensive site studies (the RIFSS) which pre- 4 cede the ROD. Nor can they go behind the On a broader scale, other questions are im- scenes to investigate all the reasons for deci- portant: Are government policies and EPA’s sions. But the ROD and its supporting RIFS are organization getting in the way of solid, defen- intended to stand alone in making the govern- sible technical work? Is the timing of key pieces ment’s case for the selected remedy and are the of work, such as testing technologies, poor? primary information sources in the 10 case Looking across sites, are there trends for prob- studies. lems in Superfund technology selection? This report does not aim to prove whether The last question is especially important. It a technology is good or bad, or whether a deci- is crucial not to look narrowly at single sites sion is unequivocally right or wrong. Cleaning but across sites. This is key to central, national up toxic waste sites is fraught with technical oversight of Super fund. While individual case uncertainties and surprises which cannot be studies can address technical soundness in a eliminated entirely. The issue of quality of specific ROD, all of them together show how RODS is not a black or white situations Each consistent the program is nationwide in under- one will have good and bad points. Any cleanup standing the advantages and disadvantages of technology can be used effectively for some ap- cleanup technologies and in responding to the plications, and every complex cleanup decision statutory requirements on cleanup technology has strong and weak points. There is no prob- selection. As does other information, RODS lem finding important, correct statements in show that Superfund is being implemented in case study RODS. Indeed, this report often uses a highly decentralized manner. There is incon- statements from one case study RIFS or ROD sistency in ROD format and presentation of in- to illustrate inconsistency or to underscore a formation, examination of cleanup alternatives, point about a problem in another ROD, Gener- and technology selections. In itself, this is not ally speaking, the decisions made in these 10 necessarily bad, but it does mean that central case studies are questionable because, for management oversight and controls by EPA are example: necessary to avoid inconsistency leading to confusion, unnecessary costs and, for some sites, If different and readily available technical ineffective cleanup. Lack of consistency among information had been used, the decision hundreds and, eventually, thousands of sites would have changed significantly, is not an academic issue. Harm to human health The range of cleanup alternatives was too and the environment, loss of public confidence narrow. in government, and wasting money are what’s The analysis was not comprehensive and at stake. was not fair to different technologies. The study work was not internally con- The following case studies also show how a sistent. site moves through the Superfund system. Gen- eral perceptions about delays are documented. Mistakes were made in calculations. and estimates. Rarely has so much information been assem- Critical assumptions were false. bled on individual sites, possible here because Conclusions were stated without analysis EPA has provided OTA with several databases. and documentation. RODS do not contain such comprehensive information, which itself is an important. observation. On the other hand, there are many areas 3An experienced attorney advises responsible parties: “Legal of interest which are not covered in these case issues, scientific and technical findings, plus the all-important policy component all affect EPA decisions, Nowhere is this more studies. Documents on a Superfund site can fill clearly shown than in the context of a Superfund Record of De- file drawers. There are many legal and proce- cision . . . the statute calls on EPA to make decisions based on dural aspects of Superfund; these case studies which remedy is cost effective or which ‘adequately’ protects public health. Applying these terms entails a degree of subjec- focus on technical areas and issues. While le- tive judgment,” [P.H. Hailer, Zfazardous Mderids, January/Feb- gal and liability issues get enormous attention, ruary 1988,] environmental protection is the reason for Superfund and ultimately it is technology which residues that are found toxic through testing must get the cleanup job done. in an onsite RCRA hazardous waste landfill; 3] provision of alternative water for affected Superfund’s Better Side offsite residents; and 4) restoration of groundwater by onsite treatment using air stripping A small fraction of RODS meet SARA’s re- and carbon adsorption. quirements. Six recent well-done RODS are briefly summarized below. While not perfect, The supporting Feasibility Study (FS) was a each ROD sets a good remedial action plan, textbook example of careful analysis, which in- each selects what is likely to be a permanently cluded alternative technologies and citations effective treatment technology, and each proof experiences at other cleanup sites. Most strik- vides adequate data and discussion to justify ing was the early elimination of nontreatment the technology choice. These six RODS contrast options, such as landfilling the hazardous waste, sharply with the 10 case studies which are the because, as stated in the FS, they “do not pro- focus of this report. vide for any treatment of contamination. ” The analysis also reviewed costs for substantial pi- Cooper Road Dump, Voorhees Township, lot treatability studies during the post-ROD de- New Jersey sign phase (the RD) as well as acceptable cancer risk levels as cleanup goals. However, a I EPA Region 2; NPL #473/7704–The ROD of 9/30/87 in 100,000 cancer risk level was used rather decided to take no further action at the site. A than the 1 in 1 million level more frequently detailed technical case, based on substantial site used. Another, and probably related, reason sampling, supported the conclusion that pre- why this ROD is not perfect is that some un- vious removal actions at the site had 1eft it per- treated hazardous material will be landfilled manently clean. The only question this ROD onsite instead of being treated. The higher risk raises is why the site scored so high on the HRS level seems to have been a.compromise made and wound upon the NPL. In hindsight, Cooper to reduce cleanup costs. Also, the delay of the Road Dump illustrates a “false positive,” a site treatability testing until after the ROD is un- that went through the Superfund system undesirable; although for this site there was more necessarily. Indeed, in a survey of EPA Re- information available to justify the technology gional staff, this site was included on a list of selection than in some of the case studies. “sites on NPL that should not be.”s No significant Federal or State money was spent to prove The Davis remedial plan used an excellent that no cleanup was necessary; the responsi- interpretation of cost-effectiveness for making ble party paid for the work. technology choices: “an alternative which has a similar public health and environmental ben- Davis Liquid Waste Site, Smithfield, Rhode Island efit to other alternatives can be screened out due to costs that are higher in order(s)-of-mag- EPA Region 1; NPL #216/770; estimated cost, $28 mil- nitude, ‘e lion.—The ROD of 9/29/87 selected a comprehensive remedial action plan. The plan included: Love Canal, City of Niagara Fails, New York 1) onsite thermal destruction of 25,000 cubic yards of excavated raw waste and contami- EPA Region 2; NPL#142/770; estimated cost, about $30 nated soil with greater than 2 parts per million mifllon.—The ROD of 10/26/87 altered an earlier (ppm) of volatile organic chemicals; 2) placement of incineration ash and pollution control Wompare this to EPA’s guidance which lacks the concept of comparable environmental protection: “[cost-effectiveness] requires ensuring that the results of a particular alternative cannot be achieved by less costly methods. This implies that for any 4Ranking on National Priorities List and total number of ranked specific site there may be more than one cost-effective remedy, sites as of July 1987. with each remedy varying in its environmental and public health W.S. Environmental Protection Agency, unreleased contrac- results.” [U.S. Environmental Protection Agency, “Interim Guid- tor report written by CH2MHill, November 1986. ance on Superfund Selection of Remedy,” Dec. 24, 1986.] 6 decision at Love Canal to use onsite land dis- sibly public concerns. The disadvantage of the posal for dioxin contaminated sewer and creek selected remedy is that the technologies used sediments. Now, a mobile thermal destruction rely almost entirely on separation. Therefore, unit will be used onsite to destroy and remove significant amounts of concentrated hazardous dioxin with an efficiency of 99.9999 percent. residues will have to be moved offsite for dis- The cost for treatment will be twice that for posal or treatment. land disposal, but the ROD selected thermal de- There was some laboratory testing of site struction on the basis of its ability to meet stat- leachate during the FS. Also, the process lead- utory requirements by eliminating toxicity and ing up to the ROD was rigorous, including an mobility. In addition, several site demonstra- extended public comment period with an un- tions elsewhere had successfully destroyed usual opportunity for local citizens to review dioxin-contaminated soil with mobile thermal a draft ROD. (Normally, the public gets a very destruction units. EPA responded to extensive brief statement of EPA’s preferred remedy to community comments against landfilling the review.) Although there was keen community contaminated material onsite and also decided interest, little of it dealt with the selection of not to attempt to separate materials with less technology, but rather with the specific loca- than 1 part per billion dioxin (EPA’s cutoff for tion on which the leachate treatment facility acceptable contamination) because of uncer- would be built. tain reliability in doing so. he-Solve, Inc., North Dartmouth, Massachusetts Operating Industries, Inc., Monterey Park, California EPA Region 1; NPL#206/770; Mimated cost, $19.9 million.—The ROD issued on 9/24/87 is one of the EPA Region 9; NPL #71/770; estimated cost: $4.8 mil- most technically detailed and complete RODS lion.—The ROD of 11/16/87 concerned an in- reviewed for this study. A previous cleanup terim remedial action required to manage con- based on an earlier ROD was stopped when four taminated Ieachate at the site, which had a long, additional hot spots of contamination were complex cleanup history. The ROD selected an found. The newly selected remedy consisted onsite Ieachate treatment system with several of: 1) the source control phase of onsite treat- proven technical steps that can reduce a diverse ment of 25,500 cubic yards of excavated PCB set of organic and inorganic contaminants to contaminated soils and sediments in a mobile levels low enough to permit discharge to a lo- dechlorination facility (volatile organic com- cal water treatment plant. The key steps will pounds will also be reduced); and 2) aquifer res- be gravity separation, coagulant addition, dis- toration by pumping, repeated flushing, and solved air flotation, filtration, air stripping with treatment involving air stripping and carbon vapor phase carbon adsorption, and liquid adsorption, particularly for volatile organic phase granular activated carbon adsorption. compounds. The site will be evaluated every The analysis of alternatives was first rate. five years because some hazardous substances Two constraints were applied that ruled out will remain there; curiously, there are no land more innovative approaches. First, the action use restrictions. had to be implemented easily and rapidly. Sec- While dechlorination was considered an inno- ond, it had to be able to cope with major fluc- vative technology, its selection was based on tuations in the composition of the leachate. positive pilot test results on an actual Super- Thus, some technologies that would actually fund site with similar contamination and cli- destroy organic contaminants, such as plasma matic conditions.r (Other work by EPA shows arc thermal destruction and wet air oxidation, the approach effective in getting residual levels both followed by stabilization of solid residues containing toxic metals, were not considered The technology is sold by six vendors according to U.S. Envi- because they would probably face delays be- ronmental Protection Agency, “A Compendium of Technologies cause of State regulatory requirements and pos- Used In The Treatment of Hazardous Wastes,” September 1987, 7 of PCBS in soils down below 1 ppm.)a Additional the decisionmaking process is clear and there pilot study results will be obtained onsite prior is public accountability. to use, and if dechlorination is unsuccessful, the ROD specified that onsite incineration will Seymour Recycling Corp., Seymour, indiana be used instead. Similar treatability and pilot EPA Region 5; NPL #57/770; ostimateid cost, $18 mil- tests will be performed for the groundwater lion.—The ROD issued on 9/30/87 was the sec- cleanup phase prior to full-scale use, ond one for the site. The selected remedy has Cleanup goals at Re-Solve were based on risk several key components: 1) a full-scale vapor analysis on the basis of possible residential use extraction system to reduce the substantial pres- of the site. A 1 in 100,000 excess (over back- ence of volatile organic compounds; 2) the ex- ground) cancer risk level was chosen for the traction and treatment of contaminated ground- soil and groundwater cleanup instead of the water at and beyond the site boundaries; 3) the more common 1 in 1 million level. Accordingly, application of nutrients to remaining contami- PCBS in the soil will be reduced to 25 ppm, nated soil to stimulate biodegradation; 4) the which is a higher concentration than goals set installation of a multimedia cap to restrict di- at other sites.e For example, 20 ppm was cho- rect contact and limit water intrusion; 5) deed sen at the Ottari and Goss/Great Lakes Con- and access restrictions; and 6) a detailed mon- tainer Corp. site in New Hampshire; 5 ppm, itoring program and technical criteria to de- at the Renora site in New Jersey; 1 ppm, at the tect failure and to plan future action if nec- Tacoma Tar Pits site in Washington; and 1 ppm, essary. at the Liquid Disposal site in Michigan (where A good technical analysis supported the selec- a 1 in 1 million risk was used). A recent EPA tion of this remedy over alternatives such as document refers to cleanup to “the desired incineration and in situ soil washing. Inciner- background levels (1 to 5ppm) or less.”lo In addi- ation would have cost $37 million and in situ tion, an assessment by EPA’s Office of Health soil washing would have cost $17 million, while and Environmental Assessment concluded that the chosen plan will cost $18 million. But tech- a range from 1 to 6 ppm PCBS in soil is equiva- nical impediments—the large size of the site (14 lent to 1 in 100,000 cancer risk.11 The Re-Solve ROD, therefore, illustrates the compromise be- acres), the large quantity of contaminated materials (about 100,000 cubic yards), and the tween level of cleanup and acceptance of cost dangers of excavating soil with large amounts by the government and responsible parties. The of volatile compounds—not cost, were the rea- FS noted that “the voIume of PCB contaminated sons for rejecting alternatives that may have soils increases exponentially as the cleanup levels become more protective.” While the fi- provided more substantial treatment and de- nal decision may be disputed by some people, toxification. In addition, the groundwater treatment is estimated to take from 28 to 42 years, particularly on the issue of residual PCB level, but there is no faster alternative available. Of some concern is that treatability studies were ~A. Kernel et al,, “Field Experience With the KPEG Reagent,” not done before the ROD. But the extraction paper presented at EPA’s Fourteenth Annual Research Sympo- technology is well proven and the final Sey- sium,” May 1988. mour implementation plan is well thought out. Whe PCB concentration level corresponding to the 1 in 100,000 risk level is 30 ppm, but EPA decided that the uncertainty of the approach allowed them to use 25 ppm as being representa- Summary of Trends From 10 Case Studies tive of that risk level. The PCB level for the 1 in 1 million risk level was 3 ppm. Also, it was estimated that onsite groundwater As a rule, RODS are fraught with problems. may contain 10 to 1S ppb PCB after cleanup, which is far in excess of 0.08 ppb, the health-based cleanup level for a 1 in 100,000 The 10 case studies, chosen out of over 100 cancer risk for PCBS, RODS reviewed, illustrate in concrete ways . IOU. S. Environmental Protection Agency, “Report on Decon- some disturbing trends among these problems tamination of PCB-Bearing Sediments,” January 1988, llAs reported by EpA in its ROD for the Liquid Disposal Site —trends that compromise the ultimate protec- in Michigan, Sept. 30, 1987. tion of human health and the environment (see 8 box 2 for capsule findings). These trends are because laboratory tests can neither fully dupli- summarized below. cate field conditions over long periods nor establish what actually is happening to the con- Evaluation and Selection of Permanent taminants. 12 EPA has said that “There is, at Treatment Technologies present, no set protocol for evaluating the ef- ficacy of stabilization technologies.”13 The use Many good, permanently effective waste of stabilization technologies for high levels of treatment technologies are on the market but, 14 organic contamination is particularly unproven. too often, are not fully examined, or are not A recent EPA review of stabilization technol- selected for use. A ROD may simply opt not to ogy said: treat a site at all but rather to bury waste in a landfill or to cap the hazardous area, both im- Although S/S [solidification/stabilization] permanent options. A site’s having too little or technologies have been used for more than 20 too much contaminated material is often cited years, there exists little information on long- as a reason for not choosing a permanent treat- term physical durability and chemical stabil- ment technology. Too little material and too ity of the S/S mass when placed in the ground .., . Generally, S/S technology is recognized much material both mean high cost for treat- effective for inorganic waste, while organic ment relative to costs for nontreatment alter- wastes have the potential to cause problems natives, but cost alone should not guide de- . . . . The long term effects of organics on S/S cisions. performance are important, however, little re- Describing a cleanup technology as a 6’treat- search has been performed. . . . the capability of the technology to perform satisfactorily over ment” can be misleading. SARA sees a treat- long periods of time has yet to be determined ment as a technology “that, in whole or in part, ... , uncontrolled air emissions are a poten- will result in a permanent and significant de- tial problem to workers and the environment.15 crease in the toxicity, mobility, or volume” of hazardous materials “to the maximum extent These EPA views are inconsistent with current practicable,” but SARA’s “treatment” allows EPA decisions that choose stabilization and call much interpretation. Furthermore, EPA has not them permanent remedies. established a hierarchy of preferred results and types of treatment. l~he attractiveness of stabilization type technologies is oftened Not all treatments accomplish the same things. expressed in noncost terms, such as: “Long term effectiveness of incineration, stabilization, and solidification are comparable.” For example, thermal destruction and some bio- IARCO Petroleum Products Co., “Critique of Sand Springs Oper- logical and chemical treatment can irreversibly able Unit Feasibility Study,” Aug. 31, 1987,) destroy or detoxify nearly all of some toxic sub- WLO Weitzman, L,EO Hamel, and E. Barth, “Evaluation of Solid. ification/Stabilization As A Best Demonstrated Available Tech- stances and therefore reduce their mobility and nology,” paper presented at EPA’s Fourteenth Annual Research volume. But a number of physical and chemical Symposium, May 1908. treatments can separate organic and inorganic ItFor example, a recent EPA study found “large losses of organics during the mixing process” [L. Weitzman et al,, op. cit.]. materials and release the hazardous material Another EPA study showed that stabilization was not competi- collected and concentrated to the environment tive with thermal and chemical treatment technologies and soil (e.g., air stripping) or place it in a landfill (e.g., washing for organic contamination [R,C. Thurnau and M.P. Es- posito, “TCLP As A Measure of Treatment Effectiveness: Re- carbon adsorption, precipitation, soil washing, sults of TCLP Work Completed on Different Treatment Tech- solvent extraction). The preferred use of sepa- nologies for CERCLA Soils,” paper presented at EPA’s Fourteenth ration technology uses treatment to destroy the Annual Research Symposium, May 1988]. A demonstration of a stabilization technology under EPA auspices concluded that hazardous material collected. “for theorganics, the leachateconcentrations were approximately equal for the treated and untreated soils” [P.R. de Percin and Chemical fixation, stabilization, and solidifi- S. Sawyer, “SITE Demonstration of Hazcon Solidification/Stabili- cation treatments usually only reduce mobil- zation Process,” paper presented at EPA’s Fourteenth Annual ity, particularly for toxic metals, (but usually Research Symposium, May 1988]. IsC,C. Wiles and H,K. Howard, “U.S. EPA Research in Solidifi- increase volume) and they nearly always leave cation/Stabilization of Waste Material,” paper presented at EPA’s some uncertainty about long-term effectiveness Fourteenth Annual Research Symposium, May 1988. 9

Box 2.-10 Case Study Sites With Capsule Findings Case Study 1 Case Study 6 Chemical Control Corp., Elizabeth, New Jersey Pristine, Inc., Reading, Ohio EPA Region 2; NPL rank: 223 out of 770 EPA Region 5; NPL #531/770 Estimated cost: $7,4 million Estimated cost: $22 million In situ vitrification was developed originally for radio- Unproven solidification (chemical fixation) technology active soils, but its use for chemically contaminated sites was selected to treat in situ highly contaminated subsur- is still unproven. In situ vitrification was selected—without face soil, which previous removal actions had left below treatability test results-chiefiy because its estimated cost the water table and covered up with gravel. No treatability was about half that of onsite incineration. But the estimated study was used. The cost of incineration was overestimated. cost for incineration is probably high by a factor of 2. in- The cleanup will leave untreated contamination onsite. cineration offers more certainty and probably would cost no more than the chosen remedy. Groundwater will be Case Study 2 pumped and treated by air stripping and carbon adsorption. Compass Industries, Tulsa County, Oklahoma Study 7 EPA Region 6; NPL rank: 483/770 Case Renora, Inc., Edison Township, New Jersey Estimated cost: $12 million EPA Region 2; NPL #3781770 Capping (containment) of waste was chosen over inciner- Estimated cost: $1.4 million ation. Capping was called a cost-effective, permanent clean- The selected remedy makes use of offsite landfilling for up even though it does not provide permanent protection soils contaminated with PCBS. Also, biological treatment comparable to incineration. No commitment was made to was selected for soils contaminated with diverse organic treat contaminated groundwater. compounds and toxic metals and for contaminated groundwater, but no treatability study supported its selection. Case Study 3 Case Study 8 Conservation Chemical Co., Kansas City, Missouri Sand Springs Petrochemical Complex EPA Region 7; NPL rank pending Tulsa County, Oklahoma Estimated cost: $21 million EPA Region 6; NPL #761/770 Estimated cost: $45 million Capping of the site and a hydraulic containment system to pump and treat some contaminated groundwater were EPA originally said that solidification technology was chosen over excavating and treating contaminated soil and ineffective for the high organic content wastes and that buried wastes, which was recommended in an EPA study on site incineration was effective. EPA then reversed itself and by the State. Water treatment cannot remove all the and selected solidification for most of the cleanup, which diverse contaminants at the site. The ROD said that no esti- the responsible party had claimed effective based on its mate could be made for the duration of the cleanup. treatability study. Incineration is to he used if solidification technology is not successfully demonstrated or fails after solidified material is landfilled on the floodplain site, Case Study 4 but criteria for failure are unspecified. Crystal City Airport, Crystal City, Texas EPA Region 6; NPL #639/770 Case Study 9 Estimated cost: $1.6 mijlion Schmalz Dump Site, Harrison, Wisconsin EPA Region S; NPL #190/770 Excavation of contaminated soils and wastes (which were Estimated cost: $800,000 buried in a previous removal action) and their disposal in A simple compacted earth cover over the soil contami- an unlined landfill with a cap over it were selected over nated with lead and chromium was selected. Solidifica- incineration. No treatability study supported the conclu- tion/stabilization treatment was rejected, although this was sion that the selected remedy is permanent on the basis a textbook example of appropriate use of the technology. of the adsorption of diverse contaminants to site soil. Ma- Voluntary well abandonment and monitoring was chosen jor failure modes for the landfill were not examined. over pumping and treating contaminated groundwater. Case Study 5 Industrial Excess Landfill, Uniontown, Ohio EPA Region 5; NPL #164/770 Estimated cost: $2 million NO treatabilty study results supported the selection of Providing alternate water to houses that have or are likely chemical stabilization. Significant amounts of untreated to have contaminated wells was a satisfactory interim re- contaminants as well as the treated materials will be left medial action. However, actions to address the source of onsite. The effectiveness of the treatment is uncertain. In- contamination and to stop and treat contaminated ground- cineration was said to offer no better protection and was water are long overdue. reiected because of its higher“ cost. 10

Moreover, a cleanup may consist of many for mobile, onsite incineration in the Chemi- different operations in which treatment may cal Control case in New Jersey and at the Pris- be only a small part. Removal actions may send tine case in Ohio (where the technology was hazardous waste to landfills, perhaps much rejected) was twice the unit cost used at the more than may be treated subsequently. Or ac- Davis Liquid Waste site in Massachusetts (where tion may be taken on contaminated soil but not the technology was selected). At the Chemical on contaminated groundwater or vice versa. Control site, both $500 and $750 per cubic yard Too many RODS assume that any use of any unit costs were quoted for two cleanup alter- technology is a treatment that meets the letter natives using the same onsite incineration. In and spirit of the statutory requirement. Gen- both cases, the material burned was essentially eral Superfund statistics on treatment can be the same and the type of incineration technol- misleading because they do not distinguish ogy was the same (the difference in the options among different technologies used at a site for was where the residuals were disposed). different amounts of material. Such variations make it hard to establish a There is no clear line between sufficient and technology’s cost-effectiveness—or lack of it— insufficient technical and economic data for se- relative to other technologies. Even when a con- lecting among cleanup technologies. A ROD tractor uses the same burden rate (see below) may choose an unproven or inappropriate tech- among ROD cleanup alternatives, inaccurate nology or both with the claim that it is a per- unit costs can distort the comparative analy- manent remedy, or a ROD may eliminate a tech- sis. For example, with Pristine, if direct cost nology because it remains untried on a large had been $186 per cubic yard instead of $730 scale. It is not uncommon to have a multimil- (with the same 83 percent burden), the total cost lion-dollar cleanup decision made without any for incineration would have been $15 million, technical data to support it, either from the tech- not $51 million; Pristine had rejected inciner- nical literature or from tests done on site ma- ation and selected in situ vitrification for $22 terial. million. If total estimated costs have any effect on post-ROD activities, then actual cleanup Information used to compare treatment technologies is often inaccurate and incomplete. costs for clients—and profits to contractors— may vary substantially and some may be much Poor information compromises the RIFS, the greater than they could be. selection of remedy, and public support of certain remedies. Alternative treatment technol- Contractors estimate cleanup costs by adding * ogies that are practical are sometimes ignored to direct costs substantially different levels of or not chosen. Costs for innovative technologies indirect cost (burden or markup). In the Pris- may be unreliable, either too low or too high. tine case, the burden—various contingencies, con- Good or bad experiences at other sites are not struction services, and design costs—amounted studied. An example is the failure, discovered to 83 percent of direct costs, while for Davis in 1985, of chemical stabilization treatment at and Re-Solve, involving the same RIFS contrac- the Conservation Chemical Co. site after only tor, the burden was 35 percent; the Davis and a few years of use; nevertheless, RODS are se- Re-Solve indirect costs explicitly included pi- lecting chemical stabilization for similar prob- lot study work, while the costs for Pristine did lems more than ever before. not. For Seymour Recycling, the burden was 60 percent; for Chemical Control, 56 percent; Contractors may quote a wide range for di- and for Crystal City, 29 percent. The range in rect costs per unit of material treated for any burden rates over different sites and across and given treatment technology. For example, quoted within contractors illustrates an important unit costs of onsite incineration ranged from management problem in Superfund. a low of $186 per cubic yard for Seymour Recy- cling to $730 per cubic yard at Pristine for the RODS cannot always depend on the results same amount of treated material; both sites are of tests done for other sites. Treatability studies in the same EPA Region. 2 Le unit cost quoted refer to tests on site material and are supposed

-J .

11 to bridge the gap between general information the Chemical Control ROD chose a new type about the technology and the more specific in- of in situ stabilization, the Pristine ROD chose formation needed for technology selection in in situ vitrification, and the Sand Springs ROD the ROD. Results of treatability studies on one chose a stabilization technique in the S& E pro- site, particularly for innovative technologies, gram. If, as EPA says, the SITE program exists do not necessarily mean that a given treatment to obtain “sound engineering and cost data” will work or not work for some other waste site, and to “resolve issues standing in the way of unless the conditions are nearly identical or the actual full-scale application, ” then how can technology’s performance is not waste specific. such ROD selections be justified? If they are The problem is that some technologies are very justifiable, are the SITE demonstrations really waste specific, and it is impossible to accurately necessary? . extrapolate positive test results from one waste The chemical character and complexity of site to another, especially because Superfund sites contaminants and how they affect the use of often have very complex, site-specifc wastes. some technologies do not get enough attention. Incineration of organic contaminants is non- A few indicator compounds, used to represent specific, whereas biological treatment is quite . all site contaminants for risk assessment, may . waste specific. Onsite treatment technologies be inappropriate for technology evaluation be- (in which the waste is brought to the technol- cause physical and chemical properties may . ogy) perform more predictably than in situ tech- differ from the way health effects vary. The re- . nologies (in which the technology is brought . sult can be a poor technology choice. Also, site . to the waste) because the latter’s effectiveness . sampling may be insufficient to detect hot spots . depends on site conditions, such as chemical, of contamination that would facilitate using physical, and biological properties of the soil. limited treatment to cut cleanup costs. In ad- . These can vary widely from site to site. dition, groundwater monitoring may not be . When they are done, most treatability studies reliable. are not done early enough. It is critical that they . be done during the RIFS before the ROD, but . Impermanent Technologies . most are done during the design phase after I . the ROD. Treatability studies will improve the When wastes are left in the ground or in . RIFS by providing technical data to support the groundwater or are redisposed in a landfill, a ROD’s analysis of cleanup alternatives and to ROD may claim that the remedy is permanent ensure that the ROD’S cleanup choice is effec-. when, in fact, it is not. Permanence may be . tive and satisfies statutory requirements. How- claimed even when technical factors suggest . ever, EPA now often speeds up RODS, appar- a high probability of failure, that is, of release ently to meet fiscal year goals; thus treatability of hazardous substances, and of another cleanup. tests during the RIFS are sacrificed. This sac- In such cases, the ROD would be more credi- rifice can backfire. Negative test results after ble if it acknowledged the remedy as imperma- the ROD would indicate the wrong technology nent and defended it on its own merits relative choice and the waste of a lot of time and money. to truly permanent ahernatives. Moreover, an Worse, altering a ROD at this point, even for impermanent remedy and a false sense of secu- good reasons, may meet some resistance. Fi- rity could lead! for example, to land use that nally, when responsible parties or technology would only complicate future cleanup and pose companies conduct these tests, EPA may need unacceptable risks. to assure their objectivity by independently verifying the results. Contrary to the law, containment/land disposal decisions seldom analyze the risk of fu- Some RODS choose technologies that are in ture failure, damages, and further cleanup. EPA’s Superfund Innovative Technology Evalu- While some RODS claim that containment/land ation (SITE) program, an indication that a tech- disposal techniques are proven and reliable nology has not yet been proven. For example, technologies with no implementation problems, 12 there is evidence to the contrary. For example, effective remedies and do not satisfy SARA, the RCRA clay cap being installed at the Win- therefore, when permanent technologies are throp Landfill Superfund site in Maine failed practical. The average estimated cost of the in September 1987 before its construction was cleanups in the six good RODS noted earlier completed. The ROD of November 1985 said was $20 million. In contrast, the average esti- the technology was proven, routinely used, and mated cost of not-so-good cleanups in the 10 posed no construction difficulties. There had case studies below was $12 million. (In the 10 been no analysis of potential failure; under the case studies, the average for the five treatment original Superfund statute-the Comprehensive remedies is $16 million and the average for the Emergency Response and Liability Act of 1980 nontreatment remedies is $7.5 million.) It is true —such analysis was not required. Under SARA that a permanent cleanup based on treatment it now is. technology is likely to require a larger initial outlay than an impermanent cleanup based on Sometimes a ROD does not commit to a defi- land disposal. Even a modest cost difference nite outcome even though it appears to have can mean a lot added up over thousands of sites. selected a technology. Contingencies, uncertainties, and multiple future options do not as- EPA is less responsive to community concerns sure the public that there will be a permanent about a remedy being impermanent than to in- remedy and that it will be fully implemented terests which favor a lower cost impermanent in a timely and effective way. Often, the ROD remedy. Thus community concerns about im- does not provide specific technical criteria for permanence are not very likely to lead to a more subsequent decisions, such as for groundwater expensive cleanup technology. There are many cleanup or land use, nor are there necessarily incentives for various parties to keep cleanup assurances of independent validation of data costs low by using onsite containment/land dis- and effective EPA oversight of activities by re- posal or even some relatively inexpensive forms sponsible parties and contractors. Specific of treatment, such as stabilization and separa- groundwater monitoring requirements are par- tion technologies. These parties include poten- ticularly important because recent EPA re- tially responsible parties (PRPs) that may have search has found that “low sampling frequency to pay for the cleanup, States that have to pro- coupled with the generally smaller sampling vide 10 percent of the cost (unless PRPs pay), networks suggest that efforts to characterize and EPA which wants to distribute available groundwater contamination at [Superfund] funds as broadly as possible and which wants sites may be inadequate. to obtain settlement agreements with PRPs to reduce calls on Superfund money. Impermanent remedies, which provide less protection than permanent ones and do not as- In selecting cheap, impermanent remedies, suredly meet cleanup goals, are often selected claims of comparable estimated costs may hide purely because they are cheaper in the short run; the truth that low cost was the key deciding fac- in the long run they are very likely to be more tor. Getting accurate costs to compare cleanup expensive. Regarding cost-effectiveness, when alternatives is crucial. Overestimates or under- two or more cleanup options offer the same estimates may be used to justify a choice or a level of environmental protection and can meet rejection. For example, at the Conservation established cleanup goals (from risk assessment Chemical Co. site in Missouri, where a settle- or existing regulatory standards), everyone will ment with PRPs was involved, an EPA contrac- agree that the lowest cost option should be cho- tor and the State recommended one remedy (re- sen. Impermanent technologies are not cost- jected) which was said to cost $24 million over another remedy (selected) which cost $21 million. But available EPA data suggest that the [email protected]. Plumb, Jr., “A Comparison of Ground Water Monitor- ing Date From CERCLA and RCRA Sites,” Ground Water Mon- rejected remedy would actually cost from $40 itoring Research, fall 1987, pp. 94-100. million to $150 million. 13 —

Program Efficiency is no apparent relationship to the site’s HRS score; a high score does not necessarily speed EPA pushes most RODS to completion by the cleanup (e.g., three sites with similar high HRS end of the fiscal year and this kind of bureau- scores waited 39, 15, and 3 months). For sites cratic pressure can lead to poor cleanup deci- within an EPA Region, however, the HRS score sions. To meet deadlines, EPA may reempha- does seem to matter; this time the waiting size public comments that would otherwise lead period decreased with decreasing score or haz- to reevaluation of facts and technologies; EPA ard level (e.g., in Region 6, the HRS scorehime may make a hasty, technically unsupported de- to RIFS start were 47/39, 32/12, and 29/-3).17 cision as it did at the Sand Springs site in Okla- homa. Typically, there is less than one month That is, the more hazardous the site according to the HRS, the longer it takes to start the RIFS between the end of the public comment period on the site. This seems opposite to what might and the issuance of the ROD. (See table 1 for be desirable; but in Region 6, the French Lim- summary data from the 10 case studies on times ited site ROD said that “The position (rank) of to reach certain stages in the cleanup process.) a site on the [National Priorities] list is incon- The RIFS may also suffer from hurried review sequential.” by EPA because of pressure to issue a ROD by the end of the fiscal year or quarter. The RIFS process, from start of the studies through issuance of the ROD, takes from 2 to The pm-remedial process has received little 3 years. Within this time, early decisions to attention even though sites can be releasing haz- eliminate some technology alternatives and per- ardous substances into the environment and, form treatability studies for others could be, but during the time they are unexamined and un- attended, get worse. The time from site iden- usually are not, made. Studying more technologies than necessary increases the time and cost tification through placement on the NPL is of the RIFS, makes it more difficult to decide about 3 years for the case studies (and often to do treatability testing on the most viable tech- much longer for other sites examined by OTA). The time between a site’s placement on the NPL and the start of the RIFS varies greatly, l~he last scoreltime is an example of a site for which the RIFS averaging about 16 months. Nationwide, there was started 3 months prior to the site’s placement on the NPL,

Table I.-Times for Sites To Reach Points in the Superfund Processa

Average Range From entry into Superfund inventory until: Preliminary Assessment completion ...... 18 months 1-45 Site Inspection completion ...... 21 months 1-44 Placement on National Priorities List ...... 36 months 4-75 Start of RIFS ...... 44 months 20-68 Completion of RIFS ...... 75 months 47-103 Signing of ROD ...... 81 months 50-104 Completion of ROD remedy (ESTIMATED) ...... 10 years 6-20 Between Preliminary Assessment completion until: Site Inspection ...... , . . , ...... 14 months O-39 Placement on NPL ...... 32 months 3-73 Start of RIFS ...... 42 months 13-68 Between placement on NPL and start of RIFS ...... 16 months –3-39 Duration of RIFS: Studies ...... 32 months 21-38 Total period (studies through ROD) ...... 34 months 24-39 Between signing of ROD and ROD estimate of completion of remedial action ...... 38 months 20-120 Duration of public comment period...... 33 days 24-44 Time between end of public comment period and signing of ROD...... 34 days 15-122 aB~sgd on the lo cage studies in this OTA special report. nologies, and sometimes contributes to poor capping a site, or institutional controls, such RODS. as deed restrictions that have uncertain future implementation. After the ROD, actual cleanup action, including remedial design, takes 2 to 3 years, Some- RODS do not consider cumulative exposures times there are repeated RODS and new actions and risks from multiple sources of similar haz- on different parts of the cleanup (called oper- ardous substances. Cleanup levels may look able units) and sometimes on the same part of acceptable on a site basis but might not when the cleanup. two or more Superfund sites are close together. The entire process from site identification An example is the two Superfund sites in Okla- through final (estimated) remedial cleanup can homa on opposite sides of the Arkansas River; frequently take about 10 years. Unexpected neither ROD evaluates risks from the other site. findings sometimes complicate the process. For Environmental risks seem to take a back seat example, remedial cleanup stopped at the Con- to bureaucratic definitions of Superfund sites servation Chemical site in Missouri and at the and to constraints imposed by seeking funds Re-Solve site in Massachusetts when new infor- from responsible parties. mation about the sites’ contamination showed The risks of transporting hazardous materi- a need for more studies, another ROD, and new als offsite for land disposal or even treatment cleanup strategies. Some risks to health and are not considered. Furthermore, SARA’s re- environment are likely during such long requirements to use permanent treatment tech- grouping periods. Contaminants are likely to nologies are not applied by EPA to waste sent migrate from areas of high to low concentra- offsite. The ROD can say that the cleanup will tion, increasing the extent and complexity of be permanent, even though the site was origi- cleanup, particularly for groundwater. nally a land disposal facility, and the wastes are slated for a landfill that itself might become Risk Management and Cleanup Goals a Superfund site. Moving hazardous waste from There are often problems with how risks are one hole in the ground to another is the nonso- assessed and how cleanup goals are met. Differ- lution that was behind SARA’s preference for ent levels of risk maybe used and very differ- permanent cleanup. For the purpose of many ent cleanup technologies may be said to be com- Superfund cleanups, EPA’s assumption seems parable, because EPA allows a broad range to be that hazardous waste sent to a regulated from 1 in 10,000 to 1 in 10 million excess life- landfill will never fail and require cleanup even time cancer risk.la Sometimes compromises are though there is widespread agreement, even made to reduce cleanup cost by allowing a within EPA, that landfill technology will ulti- higher risk than the 1 in I million cancer risk mately fail. There are also many widely recog- commonly used in Superfund. A cleanup can nized uncertainties about regulatory compli- be deemed complete even though significant ance and future corrective action. contamination remains onsite or migrates off- Most RODS seem uncertain about or do not site. Regarding cleanup goals, a cleanup tech- address future land and water use in judging nology can be justified in superficial ways. Haz- whether a selected remedy will be safe and per- ards (the source of the risk) may not be manent. In some cases, there is a lot of interest eliminated through permanent technologies but in reusing the land for productive purposes. For exposures to the hazard—i.e., the risk—may be example, at the Schmalz site in Wisconsin, reduced through impermanent actions, such as where contaminated soil is to remain in place, IWancer risk assessment is not the only way cleanup goals the ROD makes no land use restrictions. Any are established. Current regulatory standards for acceptable levels remedy that leaves hazardous waste in place of contaminants are also used, but these are not available for or caps it suggests the need for explicit atten- many contaminants. When risk assessment is used, probable, worst case, or other levels of risk are calculated. Sometimes pre- tion to future land and perhaps groundwater cleanup risks are also calculated. use. 15

The Record of Decision Document the evaluation focuses on each alternative The technical content and quality of RODS separately with very little comparison. When varies substantially across and within EPA Re- comparative analysis is used, it often is super- gions. Supporting RIFSS generally lack cita- ficial and qualitative or semi-quantitative with tions to the technical literature, important data, only rankings for alternatives. and discussions of actual experiences, good and Even for a technical expert, the basis for a bad, at sites that have used the technologies un- cleanup decision is often hard to understand; der consideration. Multimillion-dollar decisions the public has an even greater problem. RODS are often made without any significant techni- often lack much key information, such as test cal data to support them. A ROD may drop or data, other nearby sources of contamination, choose a cleanup technology with little or no earlier actions, or even an earlier ROD. In hind- discussion or justification. sight, earlier actions are frequently ineffective Probable causes for the meager level of tech- from a longer term perspective and often make nical detail are: enormous public pressure to subsequent attempts to permanently clean sites clean up sites sooner; attempts to compensate more costly and difficult. At the Crystal City for delays; bureaucratic pressures to produce site in Texas, for example, a previous action RODS faster; poor contractor performance; lack buried hazardous materials which must now of central, national oversight; and some at- be excavated and re-buried onsite in a final tempts to carry out activities after the ROD cleanup. The ROD offers an opportunity—not when there is less public scrutiny. Conflicts of yet used—to evaluate past site actions and to interest also may be a problem. Does the RIFS learn from them. contractor own a cleanup technology or will it or some affiliated company stand to profit Sometimes a remedy and its implementation if a particular cleanup technology is selected? constitute a research or demonstration project because there is no treatability study data or Is the RIFS contractor also a responsible party the technology isn’t proven for the site. But the at the site? Does a responsible party own the cleanup technology selected for the cleanup? cleanup is not publicly presented as experimental or highly uncertain. While the technol- EPA Regions are not using a standard format ogy selected may, in some cases, make sense, for RODS. Lack of uniformity makes RODS dif- the public may ultimately think it unfair of the ficult to analyze and compare for oversight and government to hide the uncertainty and risk. quality control purposes. Of particular impor- Moreover, making the claim that a permanent tance is the way alternative cleanups are evalu- remedy has been selected is questionable if the ated. Different criteria are used.lg Sometimes technology is experimental.

IOA july Ig87 directive from EPA’s Assistant Administrator same evaluation criteria, although they were not presented as for Solid Waste and Emergency Response outlined nine “key clearly. [U.S. Environmental Protection Agency, Office of Solid criteria which should be considered in evaluating and compar- Waste and Emergency Response, directives 9355.0-21 (July 24, ing alternatives. ” An earlier directive contained essentially the 1987] and 9355.0-19 (Dec. 24, 1986)]. 10 CASE STUDIES

to EPA, and 2) the Superfund Comprehensive Accomplishments Plan (SCAP). The latter is a Each of the 10 case studies in this report is budget management system; OTA used data drawn from a Record of Decision (ROD), which from SCAP NPL Site Summary reports dated is EPA’s official understanding of the facts from October 27, 1987. site studies and EPA’s explanation of how the facts support its selection of a cleanup technology. Each ROD also includes a summary of Sample Format how EPA responded to public comment and Name, location of site, and EPA region: generally includes a summary of the administrative record (related documents) for the site. Capsule OTA findings: EPA can and sometimes does reject, change, Key dates: and supplement the findings of the contractor(s) who prepare the RIFS and draft the ROD. Entered Superfund system: EPA maintains an inventory of sites called CERCLIS. The date when the region gets notification of a site is Methodology recorded is the site discovery date. Many sites These 10 case studies were selected from re- new to Superfund come with long histories of cent RODS, from September 1987 through De- contamination and cleanup efforts. cember 1987, which EPA has acknowledged Preliminary Assessment: The Preliminary came under SARA. OTA identified issues in Assessment is the first screening step in the pre- its 1985 study Superfund Strategy and in the remedial process; it consists mostly of exami- ongoing Superfund Implementation study of nation of existing records. It is done by EPA which the case studies is just one part. This contractors or by States. Sometimes a PA is other work helped in the selection of represent- done after other actions which are supposed ative case studies. The 10 cases here were cho- to come before it, apparently to satisfy the re- sen to illustrate different technology selection quirement that it be done. problems, none of which are unique to these sites, and to illustrate different types of sites Site Inspection: The Site Inspection involves and hazards. This report discusses about 10 per- some field work and testing to define the na- cent of all recent RODS to which EPA has ap- ture and scope of the hazard. The S1 is the sec- plied SARA. OTA examined nearly all recent ond screening step in the pre-remedial proc- RODs—over a hundred—to verify that the case ess and leads to Hazard Ranking System (HRS) study sites are representative. scoring of the site. The S1 is done by EPA contractors or States; the initial scoring is by EPA Format for Case Studies contractors, EPA regional staff, or States. National Priorities List The case studies are presented in a standard . proposed date: format. Following is a sample of the format with ● final date: explanations, where necessary, of the catego- . site rank: ries and of the terms and sources used in the case studies. Unless noted otherwise, the quo- Sites that get an HRS score of 28.5 or more tations in the case studies are from the site’s go on the NPL and become eligible for reme- ROD. Statements from other RODS, from other dial cleanup. Initially EPA proposes a site for case studies, and other Superfund sites are often the NPL, and, after an opportunity for public used in the case studies to illustrate program comment, the site can become a final NPL site. inconsistencies. To complete each site’s history, Final sites are ranked by their HRS score; the two EPA data management systems were needed: ranks in the case studies are from the NPL as 1) the CERCLIS inventory of all sites reported of July 1987. Then the NPL had 770 ranked sites

16 . 17

and more than 200 proposed sites. EPA revises Removal actions: Removals are site actions on the NPL only periodically, approximately once non-NPL sites and on NPL sites before (or dur- a year. The significance of the rankings for taking) a remedial cleanup. They are usually han- ing action has not been made clear by EPA. A dled by a different office within the Superfund site can receive various Superfund actions with- program than that which handles the remedial out being on the NPL. cleanup. A variety of removal actions can be RIFS start and completion: The Remedial In- taken as emergency or time-critical measures. SARA authorizes more time and money for vestigation/Feasibility Study (RIFS) provides the information base for the ROD. RIFSS are removal actions than did CERCLA. done by private engineering consulting firms, Cleanup remedy selected: paid by EPA, responsible parties, or States. Satisfaction of SARA statutory requirements: Public comment period before ROD: EPA is 1) Selection of permanent cleanup.—The required to make available certain documents Superfund Amendments and Reauthorization for public review for 21 days prior to the ROD; the period can be extended. Act of 1986 (SARA) states that EPA shall: 1) “select a remedial action that . . . utilizes per- Signing of ROD: EPA Regional Administra- manent solutions and alternative treatment tors officially sign RODS, although in a few technologies or resource recovery technologies cases the EPA Headquarters Assistant Admin- to the maximum extent practicable,” and, 2) istrator, Office of Solid Waste and Emergency if this is not done, “publish an explanation as Response, may do so, to why a remedial action involving such reduc- Estimated complete remediation: The ROD tions [in the toxicity, mobility, or volume of the normally estimates when the final action or, hazardous substance, pollutant, or contami- if the action is not final, when the whole nant] was not selected.” (SARA Section 121) cleanup will be done. Z) Accurate assessment of land disposal and containment alternatives.—SARA states that Total time: The total elapsed time of above EPA shall “take into account:” . . . dates. the long- term uncertainties associated with land dis- Brief description of site: posal; . . . short- and long-term potential for adverse health effects from human exposure;. . . Major contamination/environmental threat: the potential for future remedial action costs HRS scores: EPA’s policy is that a site score a if the alternative remedial action in question minimum of 28.5 to be placed on the NPL. were to fail . . . “ (SARA Section 121) (Once only, States can nominate one site for RIFS contractor: Information on time, cost, and the NPL regardless of its score.) The maximum company is normally available from the ROD; subscores for groundwater, surface water, and if not, OTA obtained it from other EPA sources. air are 100, and a formula is used to combine the subscores so that the maximum total score State concurrence: Only information reported by is also 100. (This calculation applies to the ver- EPA in the ROD was used. sion of the HRS used for the case study sites; Community acceptance: The ROD’s responsive- a newer HRS version, required by SARA, may ness summary was chiefly used. It does not nec- change this methodology.) There are many con- essarily reflect the full range of public opinion cerns about the accuracy of HRS scores and about a site because it only describes direct in- their use in ranking NPL sites; an HRS score teractions through the official public comment may not paint an accurate picture of a site’s process between the community and EPA. original or current environmental threat. Sites are not rescored after removal actions or in- Special comments: terim remedial measures. General conclusions: 18

Case Study 1 metals. . . . these contaminants are strongly ad- Chemical Control Corp., Elizabeth, NJ; sorbed to the soil and are present in the ground- EPA Region 2 water in relatively low concentrations. The contaminants found in the [river] sediments ., . are Capsule OTA findings. -Unproven solidification not all attributed to the Chemical Control site.” (chemical fixation)technology was selected to The contaminated layer “averages approxi- treat in situ highly contaminated subsurface mately ten feet thick. . . . some of the more soil, which previous removal actions had left mobile chemicals continue to leach into the below the water table and covered up with groundwater. Significant health threats are gravel. No treatability studywas used.The cost posed by direct contact, fugitive dust emission, of incineration was over estimated. The cleanup and volatilization. Contaminants are only leav- will leave untreated contamination on the site. ing the site via the groundwater. . . . direct con- Key dates: tact with sediments as well as ingestion of contaminated shellfish are both potential exposure Entered Superfund system: 5/1/79 routes. Flooding happens occasionally at the Preliminary Assessment: 5/1/79 site now . . . “ The ROD indicated a volume of Site Inspection: 4/1/79 - 8/1/82 contaminated material of 18,000 cubic yards. National Priorities List –proposed date: 10/1/81 HRS scores.—groundwater O; surface water –final date: 9/1/83 18.18; air 79.49; total 47.13 —site rank: #223 out of 770 Removai actions. —State removal of large quan- RIFS start and completion: 12/31/84 - tities of wastes began in March 1979 and was 6/30/87 (ROD says it began in 11/86, but interrupted by a major fire in April 1980. After SCAP has earlier date) the fire, the State removed more material, Public comment period before Record of including 3 feet of surface soil which was Decision: 7/6/87 - 8/14/87 replaced with gravel. Also, from November Signing of ROD: 9/23/87 1980 until July 1981, the State operated a Estimated complete remediation: 28 to 32 - groundwater recovery and treatment system. months after R OD (around 4/90) This action plus groundwater movement and Totai time.—ll years not just the adsorption of contaminants to soil may explain why the subsequent Remedial In- Briaf description of site.—“The site consists of this vestigation found little contamination and why 2.2-acre property and the portion of the the HRS groundwater score in 1982 was zero. Elizabeth River adjacent to the property, . . . the Overall, the State of New Jersey has spent $25 water table aquifer at the site [is] saline and ti- million on the site. dally influenced. The site is flat and barely above sea level. Chemical Control Corp. oper- After the site became a Superfund site, four ated from 1970 until 1978 hauling, treating, and additional initial remedial measures were car- disposing of a wide variety of industrial wastes. ried out (in 1984, 1985, 1986, and 1987) to re- Throughout its operations, it was cited for vio- move more materials from the site. lations thet included discharging liquids onto Cieanup remedy selected. —Other than contain- the ground adjacent to the Elizabeth River and ment, treatment alternatives considered were accumulating thousands of drums of incompat- soil washing, solvent extraction, and incinera- ible wastes.” tion, The selected remedy was in situ fixation Major contamination/environmontai tlnat.–” . . . soils (chemical fixation, stabilization, and solidifi- beneath those removed by the NJDEP [New Jer- cation often are used to describe similar treat- sey Department of Environmental Protection] ments). Fixation chemicals would be injected are highly contaminated with a variety of or- though an expandable bit drill which would ganic compounds and to a lesser degree with pass through the gravel layer: “A series of over- 19 lapping columns would be formed converting tainment remedies and when only partial de- all of the contaminated soil at the site into a struction treatment technologies are used at a solid mass. This would inhibit water from flow- site. The review has implications for future land ing through the site, thereby preventing the use because use of the land might interfere with production of leachate. In addition, some con- reviews and because results of reviews might taminants may be chemically altered and in- reveal hazards that would block land use. corporated in the solid matrix formed by this A major issue with the selected remedy of action, reducing the toxicity as well as the mo- in situ fixation is that it is not a proven tech- bility of the contaminants. The potential for ex- nology; no data exist to show effectiveness for posing the contaminated soil would be elimi- cleanup of a hazardous waste site comparable nated. The treatment] will create a solid matrix to Chemical Control. No treatability study was that will have extremely low permeability. . . . conducted prior to the ROD to provide evidence because it is implemented primarily below of effectiveness in terms of resistance to long- ground, [the treatment] offers protection against term leaching or actual toxicity reduction. The releases during a flood. ” diverse set of contaminants at the site would The estimated cost for the selected remedy pose a challenge to conventional chemical fix- is $7.4 million, while the cost for excavation, ation techniques. The use of in situ fixation be- onsite treatment, and onsite disposal of residues neath the water table in saline conditions may is $14.5 million for fixation and soil washing exacerbate the difficulty of achieving an effec- and $22.3 million for incineration. The ROD tive cleanup. Various ROD statements on this also commits to some other relatively minor issue include: removal actions and environmental monitor- ● “Although in-situ fixation is not yet a standing, “including an evaluation after five years ard construction practice, several vendors to assess its protectiveness to public health and are available that provide this service. ” the environment. ” ● 64 .*, the in-situ process described in this Satisfaction of SARA statutory requirements. document is currently being studied for use at other hazardous waste sites. ” 1) Selection of permanent clean up.—’’The re- ● “The long-term reliability of this alterna- medial alternative presented in this document tive is especially promising” (emphasis is a permanent solution for closure of the Chem- added). ical Control site. . . . this remedy satisfies the ● “A treatability study and field test will be preference for treatment that reduces toxicity, required during design to prove the tech- mobility, or volume as a principal element. . . . nology . . . “ this remedy utilizes permanent solutions and ● 64 environmental samples will be col- alternate treatment technologies to the maxi- lected to monitor the effectiveness of the mum extent practicable. Based on the input re- remedy.” (No specific technical criteria are ceived during the public comment period, this given.) alternative has been selected by the EPA and ● “Although such an application of this tech- the NJDEP as the final permanent solution for nology is fairly new, promising results have the site. [The selected remedy] also utilizes an been obtained in laboratory tests, and it is alternative treatment technology that offers a being tested at other hazardous waste sites more cost-effective remedy. ” and evaluated under the Superfund Inno- vative Technology Evaluation program. The fact that the ROD met SARA’s require- The Chemical Control project will benefit ment for a five-year review indicates that EPA from the experience gained at these sites.” recognizes that the selected remedy would leave untreated, undestroyed, and toxic wastes on- Despite the last comment, OTA has been only site. EPA normally responds to the statutory able to identify the evaluation within the SITE requirement for review for land disposal/con- program. A vendor in conjunction with Gen- 20 eral Electric has demonstrated its technology The EPA decision to use incineration at the at a GE site in Hialeah, Florida, in April 1988. Southern Crop Services site in Florida at about PCB contaminated soil was treated. A demon- the same time of the Chemical Control deci- stration report is supposed to be available ap- sion undercuts the use of stabilization for proximately four to six months after the dem- Chemical Control. In justifying its selection of onstration. OTA contacted the vendor of the incineration, EPA said: “Solidification and fix- in situ chemical fixation technology for the GE ation of pesticide contaminated soil was found site and was told that no tests have been con- to be technically unacceptable due to the high ducted on materials from Chemical Control but detected concentrations and because organic preliminary laboratory tests have been success- pesticides tend to leach from solidified mate- ful on material from the Hialeah site. GE will rial. This technology was deemed unaccept- be using the technology to actually clean up able.” (U.S. Environmental Protection Agency, the Hialeah site; the cleanup will constitute the Region 4, memorandum, Sept. 8, 1987.) Pesti- SITE demonstration. GE plans to do this clean- cides are among the many different types of up even though EPA has not officially sanc- organic contaminants present at Chemical Con- tioned its use at Hialeah; the agency has prob- trol. The FS for the Re-Solve site in Massachu- lems in issuing regulatory permits for an setts rejected stabilization because “there has innovative treatment technology, necessary be- been limited success in chemically fixing organic cause the cleanup is not at a Superfund site. contaminants such as solvents and PCBS. ” Given the timing of the Hialeah demonstration, it is difficult to understand how the selected The ROD for the Liquid Disposal site in Mich- remedy can be already judged to satisfy require- igan, which also selected stabilization for soil ments for Chemical Control. Officially, there contaminated with organic chemicals, said that are no performance criteria to prove the effec- the hazardous substances “will not be perma- tiveness of the technology. Moreover, one of nently destroyed” and “hazardous chemicals the objectives for the SITE demonstration is to still remain in that [treated] mass. ” And the FS determine the “integrity of the solidified soil for the site said: “Considerable research data over a period of five years.” (U.S. Environ- exists demonstrating the effectiveness of this mental Protection Agency, “Superfund Inno- technology in immobilizing a wide range of vative Technology Evaluation (SITE) Program,” contaminants, primarily inorganic. A substan- HWERL Symposium, May 9-11, 1988.) tial amount of data does not exist, however, to accurately judge the long-term reliability of the But when will the technology be completely process.” Of particular significance to the use evaluated? Some innovative technologies have of stabilization at Chemical Control, the Liq- been demonstrated several times and still have uid Disposal ROD also selected a slurry wall uncertainties for broad use. The information and impermeable cap around and over the available on the SITE demonstration is incon- treated material, in part because it is necessary sistent with the schedule for implementation to “protect the solidified soil/waste from degra- of the selected remedy given in the Chemical dation by upgradient ground water that is Control ROD and seasonal constraints may also slightly contaminated with chemicals” (empha- delay cleanup. sis added). Moreover, the SITE demonstration will not A recent EPA study found “large losses of be performed on a comparable contaminated organics during the [stabilization] mixing proc- material. The presence of volatile organics and ess.” (L. Weitzman et al., “Evaluation of Solidifi- metals, for example, makes the Chemical Con- cation/Stabilization As A Best Demonstrated trol project significantly different. A positive Available Technology,” paper presented at SITE demonstration result will not, therefore, EPA’s Fourteenth Annual Research Sympo- substitute for a treatability study on Chemical sium, May 1988.) Another EPA study showed Control material nor should it justify perform- that stabilization was not competitive with ther- ing a pilot study at Chemical Control. mal and chemical treatment technologies and 21 soil washing for organic contamination. (R.C. ducted in the field. The level of treatment Thurnau and M.P. Esposito, “TCLP As A Meas- achievable is variable, depending on the waste ure of Treatment Effectiveness: Results of and soil conditions. The potential for long-term TCLP Work Completed on Different Treatment immobilization is unknown at this time. The Technologies for CERCLA Soils,” paper pre- reliability of the treatment is unknown since sented at EPA’s Fourteenth Annual Research there is no information on its long-term effec- Symposium, May 1988.) A demonstration of a tiveness.” (R. Sims et al., “Contaminated Sur- stabilization technology under EPA auspices face Soils In-Place Treatment Techniques,” concluded that “for the organics, the leachate Noyes Publishers, 1986.) concentrations were approximately equal for the treated and untreated soils.” (P.R. de Per- The Feasibility Study for the Crystal City site cin and S. Sawyer, “SITE Demonstration of rejected in situ chemical stabilization: “Im- Hazcon Solidification/Stabilization Process,” mobilization, chemical treatment, and physi- paper presented at EPA’s Fourteenth Annual cal treatments have not been shown to be fea- Research Symposium, May 1988.) sible for in situ treatment of these contaminants as it is not possible to get a good, uniform, well Senior EPA people have made an important distributed treatment.” The focused FS for the observation about in situ stabilization:” . . . the Conservation Chemical Co. site in Missouri re- decision to use a stabilization technique should jected in situ stabilization: “Technology [was] be made only after the chemical and physical attempted and [was] found not feasible at other properties of the solidified waste have been ex- similar sites. Technology is not sufficiently de- tensively tested to insure that the required prop- veloped.” An addendum study also rejected in erties have been developed. ” (D.E. Sanning and situ immobilization (which cannot be differen- R.F. Lewis, “U.S. EPA Research on In-Situ tiated from stabilization): “Technology would Treatment Technology,” Anatomy of Super- not effect organic and other non-metallic con- fund, proceedings of the 8th National Ground tamination; thus, these substances would con- Water Quality Symposium, September 1986.) tinue to be a source of contamination. Immobili- The ROD for Chemical Control does not meet zation reactions are reversible.” The FS for the these requirements. Pristine site in Ohio rejected in situ chemical detoxification: “Treatability study is required An educational short course and two recent to assure effectiveness. It is difficult to ensure EPA documents on cleanup technologies make proper reactant mixing and verify effectiveness.” no reference to the selected technology. (U.S. Environmental Protection Agency, “Remedial The same contractor that prepared the Chem- Engineering of Hazardous Waste Sites,” The ical Control FS has tested in situ chemical National Hazardous Materials Training Cen- stabilization elsewhere. (“Feasibility Study Salt- ter, October 1987; “A Compendium of Tech- ville Waste Disposal Site, Smyth County, Vir- nologies Used in the Treatment of Hazardous ginia,” August 1986.) Successful laboratory and Waste,” September 1987; “RCRA/CERCLA pilot tests led to the fieldtesting of a particular Treatment Alternatives for Hazardous Wastes,” in situ treatment; however, the fieldtests failed October 1987.) The latter EPA document is used and the approach was dropped. This twist il- to teach people implementing hazardous waste lustrates the uncertainty of a technology, even programs about waste treatment and says: after successful laboratory tests. “Solidification technologies are designed to be used for final waste treatment. This means the The ROD said that the selected remedy “offers technology should be applied only after other a level of long-term protection comparable to treatment techniques have been applied, i.e., or exceeding that of any of the other alterna- incineration, chemical treatment or other.” tives.” However, if the selected remedy has not been shown to destroy the organic contami- Another observer has commented on this ap- nants the way incineration could, is this asser- proach: “Experimental studies have been con- tion of comparable permanence correct? In- 22

cineration followed by chemical fixation of the site disposal (where a baseline of 21,000 cubic residue to immobilize toxic metals offers a yards of soil was used) and $00 per cubic yard higher level of protection. Therefore, the addi- for the combination of onsite incineration and tional estimated cost for the incineration op- onsite disposal (where a baseline of 27,000 cu- tion (three times more than fixation) does not bic yards of soil was used). These differences eliminate its cost-effectiveness. Also, the cost in unit cost and soil volume do not make sense. estimate for the selected remedy is unreliable Other vendors are now quoting less than $3OO because the technology has not been used be- per cubic yard for the volume of work at this fore on such a site. site and, indeed, $3OO per cubic yard was quoted in the FS for the Davis Liquid Waste site in This ROD illustrated the benefit of examining the supporting Feasibility Study. In this Rhode Island (where onsite incineration was case, the FS introduced several new elements: selected) and $186 per cubic yard was quoted in the FS for the Seymour Recycling site. The 1. The ROD did not reveal several facts about Chemical Control FS also stated that “Rotary the site and the selected remedy that were in kiln and fluidized bed incinerators are the only the FS: a) the curing time for the fixation mate- types of mobile units currently available.” This rial is about one month; b) there is a volumet- statement is not true. According to EPA, there ric increase in the waste after treatment that are three other types of full-scale mobile ther- depends on site materials and conditions; c) “It mal technologies available: circulating bed, in- is unlikely that solidification can be effected frared, and wet air. (U.S. Environmental Pro- in contaminated areas at the interface between tection Agency, “RCRA/CERCLA Treatment the river and the site. This residual contami- Alternatives for Hazardous Wastes,” October nation will continue to flush from such areas 1987.) If $300 per cubic yard were used (with surrounding the solidified mass”; d) “This alter- other costs factored in separately) for onsite mo- native will not reduce any potential human bile incineration of 18,000 cubic yards (the same health or environmental impacts associated amount of material as for the selected remedy), with the contamination detected in the gravel then the cost would be about $14 million. This cover atop the site”; e) “ . . . even under noncost compares to the ROD cost for onsite in- flood conditions, the water table is quite close cineration with onsite disposal for 27,000 cu- to the surface of the site”; and f) The estimated bic yards at $22 million. The ROD omitted FS costs for the bench test and pilot test for the low and high cost estimates for the options. in situ fixation alternative are $770,000. Since there is no field experience within situ fixation, its high cost estimate of $14 million Z. The FS analysis of the selected remedy sug- is significant as an estimate. Thus it is possible gested that there is reliable information on which to base conclusions. The text contained that incineration at $14 million might be about phrases such as: “has been demonstrated,” “the the same cost as the selected remedy and not available literature, ” and “it is reported,” sug- three times more. More recently, after the ROD, a news story reported that the cleanup project’s gesting that technical literature and EPA re- estimated cost is $10 million, with $750,000 ah ports were used. But, all the information came located for the one-year design job, an amount from a single vendor (identified in figure 3-6 in the FS). The FS also referred to the GE Hia- which could not account for the cost of the treat- leah site as a Superfund site, which it is not, ability study. (Superfund, Feb. 1, 1988.) and the FS said (in June 1987) that “a field ap- 4. The analytical framework used to evalu- plication of the emerging technology is pres- ate alternative cleanup approaches is inconsist- ently underway, ” which it was not. ent with commonly accepted practice and with 3. The FS analysis of incineration is poor. The EPA’s recommendations. A July 1987 EPA unit cost for incineration only (with other costs directive clearly recommended the use of nine figured separately) was $750 per cubic yard for criteria; an earlier directive was less clear. Only the combination of onsite incineration and off- four criteria were used in this FS: technical fea- 23 sibility, public health and environmental con- A traditional containment wall was exam- cerns, institutional considerations, and cost. ined, but the ROD said that a containment wall However, analysis using additional criteria, might need replacement in the future. But the particularly factors such as reliability and im- ROD noted that a wall “would offer effective plementability which are normally stressed, protection if institutional controls were im- would have worked against the selected rem- posed to prohibit any future digging at the site.” edy. (Indeed, this analytical result happened in Interestingly, in the Chemical Control situa- the ROD for the Conservation Chemical Co. site tion, a containment approach might have made in Missouri, in which the only soil treatment sense—as an interim measure—because there alternative faired poorly on reliability and im- are many other sources of river contamination plementability and thus was rejected.) The around the site that could reduce the effective- Chemical Control ROD evaluated each cleanup ness of a remedy that leaves site material vul- alternative separately. An explicit comparison nerable to recontamination. The ROD noted the of alternatives weighing relative advantages risk of recontaminating clean material back- and disadvantages was not done in the FS or filled into the site. Contaminants from the river the ROD. (At Conservation Chemical, the ROD might also affect the effectiveness of chemical comparative analysis of cleanup alternatives stabilization. was done within individual discussions of evaluation criteria, a very useful approach com- RIFS contractor. —State-led; NUS Corp. under pared to discussing each alternative for all subcontract to Ebasco Services, Inc.; about $1 criteria.) million obligated. SCAP indicated a RIFS from 9/28/83 - 2/15/85 at a cost of $208,000, a subse- To sum up, an analysis of the FS shows that quent ROD on 2/15/85, a Remedial Design at there was little basis to select the in situ chemi- $504,000, and a Remedial Action at $485,000. cal fixation alternative and that the FS analy- There was no information in the 1987 ROD’s sis was biased in favor of the selected remedy site history on the earlier RIFS, ROD, and re- and against using onsite incineration. Onsite medial action, and SCAP listed separately the incineration is a proven, more cost-effective, interim remedial measures and removal actions and more reliable cleanup alternative than the with their costs. To confuse things still more, FS and ROD indicated. a master EPA list of all Superfund RODS showed an earlier ROD on 9/19/83. 2) Accurate assessment of land disposal and State concurrence.—New Jersey agreed with the containment alternatives. —To a significant ex- selected remedy. tent, the selected remedy is a land disposal/containment approach because, unless shown Community acceptance.—The ROD said: “The otherwise with positive test results, chemical main concern of local officials is that a thor- fixation cannot be assumed to detoxify all con- ough, permanent remedy be expeditiously im- taminants. Leaving the contaminated soil on- plemented.” The responsiveness summary also site and beneath the water table raises ques- indicated some public concern about the effec- tions about future failure. No technical criteria tiveness of the selected remedy, particularly be- were established to determine failure of the cause of its inattention to toxic metals. selected remedy. Failure is a real possibility, Special comments. —The ROD did not address since the same generic treatment failed at the the problem of highly contaminated river resi- Conservation Chemical Co. site. A volume in- dues. The chief reason is the other sources of crease in the treated waste with the addition contamination: “ . . . remediation of the river of fixation chemicals raises questions about the sediments is premature. ” If the river residues integrity of the resulting solidified mass within are considered part of the Chemical control the site with in situ use. Costs were not esti- site, then this ROD does not offer a final per- mated to repair a failure of the selected remedy. manent cleanup of the entire site. 24 ●

Why did the State leave the contaminated soil that a case could have been made to evaluate in place in a previous site action and cover it it as the ROD did for in situ stabilization. up with permeable gravel? This action compli- cates a permanent remedy and will contribute The ROD had no summary of the adminis- to continued leaching of contaminants into the trative record which, because of the inconsist- river and its sediments for about 10 years. Even ent information on dates, would have been very though EPA said that the contaminants adhere useful. to the soil, not all the contaminants could be- General collusions.–EPA’s high confidence and have so ideally nor, as discussed in the FS, will certainty about the selected remedy is unsup- all contaminants be treated to reduce their mo- ported by analysis. In this case as in several bility. The ROD gave no data to support the con- other case studies, the ROD did not follow tention that all the contaminants are tightly EPA’s guidance that ROD analysis “must be bound to the soil. The ROD noted: “The NJDEP based on a specific process within [a] technol- has also indicated that contaminant concentra- ogy category . . . to ground the analysis in hard tions in the soil at Chemical Control exceed data.” (U.S. Environmental Protection Agency, State guidelines.” It is not clear why an interim “Interim Guidance on Superfund Selection of containment action at the site, such as a slurry Remedy,” Dec. 24, 1986.) The ROD package wall and cap, was not implemented years ago. contained a letter of August 31, 1987, in which Other innovative in situ treatment technol- EPA told the City of Elizabeth’s Director of ogies—including biological treatment and Health, Welfare and Housing:” . . . we feel that vitrification—could have been considered in in-situ fixation will protect public health and treatability studies as viable candidates, but the environment from any hazards posed by were not. The FS rejected in situ vitrification the site.” As with the statements in the ROD (ISV) on grounds that also could have been used about the selection being permanent, the state- to reject the selected remedy. Yet ISV was ment in the letter was inconsistent with the selected for the Pristine site, and the Crystal need to prove the effectiveness of the remedy, City FS evaluated ISV favorably, although it did through a treatability study, after it was selected not select ISV or any treatment alternative: but before it is fully applied. In the same letter “[ISV] has been successfully demonstrated in to the city official, EPA said: “In the event that laboratory and bench testing. IISV] was deter- these tests show that in-situ fixation would fail mined to be feasible given the existing infor- to offer protection of public health and the mation available and is retained for further environment, the ROD would be amended as evaluation. ” A recent NJDEP report’s discus- necessary. ” Examining alternatives, selecting sion on innovative/alternative technologies another remedy, amending the ROD, and im- said: “various technologies presently exist plementing another remedy would, of course, which can adequately address contaminated take considerable time. Indeed, this scenario soil and other contaminated media. For exam- has happened at other Superfund sites, includ- ple, waste vitrification (imbedding waste in ing the Re-Solve site in Massachusetts, the Con- glass) can immobilize organic or inorganic conservation Chemical Co. site in Missouri, and taminants while generating residuals that are at Love Canal in New York. delistable and environmentally safe.” More- A systematic bias against incineration was over, the NJDEP report also noted that treata- suggested in the site FS evaluation, particularly bility studies are done during the RIFS to “fill for cost. Use of mobile incineration might not data gaps . . . and supply information needed cost significantly more than the selected remedy, to select a design alternative.” (New Jersey De- but it would offer more certain effectiveness. partment of Environmental Protection, “Com- prehensive Management PIan 1988-1992,” Oc- The Chemical Control site illustrates the prob- tober 1987.) The point here is not whether ISV lem of delaying a treatability study until the de- is the best cleanup technology for the site but sign phase. While chemical fixation is consid- 25 ered a treatment, it cannot be assumed to constructing something on it, despite the un- detoxify all contaminants. A treatability study certainty of the selected cleanup, despite the during the design phase should be limited to contaminated materials to remain onsite, and obtaining data necessary for the detailed, engi- despite the other nearby sources of contami- neering design of the selected technology and nation. The FS indicated that the State of New also to develop technical criteria to guide po- Jersey owns the land and that, with the selected . tential bidders on the project. If a treatability remedy, New Jersey’s own law regarding real study is necessary to show effectiveness, as in estate transfer would be violated if the site was this case, then it should be done, as it some- put into commercial reuse “Since some resid- times is, during the RIFS process. If it is de- ual contamination will exist in the gravel cover layed, then “Such a deferral may result in a under [the selected remedy], and since the sub- premature (and administratively ‘irreversible’) surface contamination will still be present (al- commitment to a technology that may not be though immobile), it is unlikely that this alter- appropriate for a given site.” (D. Truitt and J. native ,will comply with the concentration Caldwell, “Evaluation of Innovative Waste requirements of ECRA [New Jersey’s Environ- Treatment Technologies,” Waste Management mental Cleanup Responsibility Act]. It is felt Conference-Focus on the West, Colorado State that this alternative will be consistent with the University, June 1987.) intent of ECW, however.” The FS noted a cancer risk above 1 in 1 million for contact with The selected remedy for Chemical Control the contaminated gravel, a risk that has impli- can be considered a land disposal/containment cations for future use of the site and onsite approach. OTA does not mean to challenge the workers. merits of the in situ chemical fixation technology but does question the decisionmaking process used at this site. Making the remedy selec- Case Study 2 tion before treatability test results are available Compass Industries, Tulsa County, Oklahoma; may mean that EPA was in a hurry to promote innovative treatment technology and to issue EPA Region 6 the ROD. Capsule OTAfindings.-Capping (containment) of The ROD also did not assure a permanent waste was chosen over incineration, capping remedy for the site because it ignored the was called a cost-effective, permanent cleanup cleanup of the highly contaminated river sedi- even though it does not provide permanent pro- ments, ignored the contamination in the gravel, tection co~arable to incineration. Treatment and ignored the untreated material at the river’s of contaminated groundwater is not yet planned. edge. The ROD over estimated the cost of onsite incineration, which could achieve more Kay datas: permanent, more complete, and more certain ● Entered Superfund system: 10/1/80 cleanup at a cost of about $14 million, instead ● preliminary Assessment: 4/1/80 of the ROD’s $22 million estimate. ● Site Inspection: 7/1/82 The serious complication of other nearby ● National Priorities List sources of contamination shows that Superfund –proposed date: 9/83 sites cannot be seen in isolation. The ROD noted a –final date: 9/84 that recontamination of the site is a p~tential —site rank: #483 out of 770 problem. Therefore, a case could have been ● RIFS start and completion: 6/29/84- 7/13/87 made for coordinating this cleanup with others ● Public comment period before Record of to assure an overall, permanently effective so- Decision: 7/22/87 - 8/31/87 lution for all of them. ● Signing of ROD: 9/29/87 ● Estimated complete remediation: 9/90 There seems to be an unusual interest in the RIFS and ROD process in reusing the site and Totai time.—lO years 26

Brief description of site.-The site is” . . . an aban- the contaminated groundwater in the upper, doned landfill located west of Tulsa, Oklahoma. perched water bearing zone; and s) installing The site occupies an abandoned limestone fences and signs along the perimeter of the cap. quarry. From 1972 to 1976 the site was per- “This alternative consists of site grading, cap mitted and operated as a solid and industrial placement, diversion of surface water, and air waste landfill. physically, the site is situated emissions monitoring. The site cap will be re- on a bluff approximately one-quarter mile south quired to meet RCRA specifications. Ground- and zoo feet above the Arkansas River. An water will be treated at a later date if found to elementary school lies within one-half mile and be necessary. The site will be monitored for a a major regional park is immediately adjacent period of at least so years . . . to ensure that no to the site.” significant contaminant concentrations migrate from the site” (emphasis added). Major contamination/environmental threat.—” . . . a large number of organic and inorganic priority Estimated cost: $12 million. pollutants were detected. They include a total Satisfaction of SARA statutory requirements: of 12 inorganic priority pollutants and at least 33 organic priority pollutants . . . pathways of 1) Selection of permanent cleanup.—The ROD said that “ . . . [the selected] remedy uti- possible off-site contaminant migration are surface water, groundwater, and air. The possi- lizes permanent solutions and alternative treat- bility also exists for direct contact at the site ment technologies to the maximum extent prac- with contaminated source materials, such as ticable.” The ROD acknowledged that capping sludge, soil, or sediments. The majority of the only reduces the mobility of contaminants; groundwater treatment would also “reduce the contamination in the groundwater is confined volume and toxicity of wastes on site to some to the upper aquifer. Samples of groundwater from monitoring wells on the site are highly degree.” Hazardous residues from water treat- contaminated. This indicates a degradation of ment would be shipped offsite to a landfill. groundwater quality due to waste disposal in Full onsite thermal destruction was examined both the perched and deep aquifers. The vol- as an alternative and was given highest ratings ume of waste was determined to be approxi- “because this process would destroy the or- mately 620,000 cubic yards. The average ground- ganic compounds in the waste.” Partial onsite water flow rate of both aquifers is 720 gallons thermal destruction of 2 percent of the waste, per day or an estimated 263,000 gallons of water coupled with capping and groundwater treat- per year [into Arkansas river]. The most recent ment, would have been an improvement over fire burned for several years before it appar- capping and groundwater treatment alone. Re- ently burned out in 1984. . . . there exists a po- garding full onsite thermal destruction: “ . . . tential for future fires. . . . [During fires] ele- this remedy is not cost-effective ($339 million vated levels of air contaminants may present vs. $12 million). ” However, the ROD acknowl- a health hazard.” edged that full incineration would give the best overall environmental protection. Regarding HRS scores .—groundwater 11.05; surface water partial thermal destruction: “ . . . the increase 18.46; air 59.49; total 36.57 in cost does not justify the negligible increase Removai actions.—None indicated. in protection to human health and the environment.” Also, regarding implementability: “On- Cieanup remedy selected.—Two major alternatives site incineration remedies . . . will require rela- were considered: 1) leaving waste in the ground, tively more attention during design than other capping the site, and treating groundwater; and remedies . . . and were therefore rated lower 2) incineration of excavated wastes. There are than the other alternatives.” That is, treatment three parts to the selected remedy: 1) capping requires more work than waste containment. the site; 2) if deemed necessary through compliance, monitoring after installation of the 2) Accurate assessment of land disposal and cover material, collecting and treating onsite containment alternatives. —Regarding short- 27 term health effects for capping, there are” . . . the French Limited site (in the same EPA re- essentially no risks to workers or residents.” gion as Compass), use of a slurry wall and cap For long-term effectiveness, the remedy “ . . . to contain hazardous waste was described as will successfully reduce long term risks to hu- a “temporary solution” for which the “volume man health and the environment.” Standard and toxicity would not be affected” and “the operation and maintenance for site and cap was potential would always exist for failure of ei- planned. Regarding future actions: No future ther the cap or the slurry wall allowing for the remedial actions are anticipated. The selected movement of unstabilized wastes contained remedial action is considered permanent” (em- onsite. ” phasis added). RIFS contractor. –State led, $624,000; John The selected capping remedy, was given two Mathes & Associates. higher levels of ranking for reduction of toxicity and volume of waste as compared to the no State concurrence.—“The State . . . has concurred action alternative, even though capping does with the capping portion of this remedy. . . . the no more to waste than no action. Ratings for State did not support any of the other proposed the reduction of mobility for the selected remedies.” remedy are probably too high, especially be- Community acceptance.—” . . . the public was in cause it is not certain whether or not ground- favor of [capping] over thermal treatment of the water movement would be affected. waste. . . . the public concern was that the ther- Several other areas of uncertainty remain: mal treatment unit would create hazardous

● emissions and increase the potential for ex- “Future land use considerations will be posure.” evaluated in the upcoming design phase based on the needs of protection of the Special comments: cap. ” ● No treatment technology other than ther- ● That some water infiltration through the mal destruction was considered in the fi- cap, which would cause migration of con- nal analysis, although other possibilities taminants into groundwater, might happen existed. is indicated by the possible use of a syn- ● No commitment to using a liner was made thetic liner in the cap: “The long term even though “ . . . Subtitle C of the Re- advantage to the liner is that less water source Conservation and Recovery Act, would be generated from the seeps” (em- which requires a cap with liner, is relevant phasis added). and appropriate. ” ● With regard to long-term impacts: “The po- ● No specific technical criteria were used for tential for future fires and continued off- deciding what types and levels of contami- site migration of contaminants pose ad- nation found via groundwater monitoring verse human health and environmental im- would trigger actual groundwater treat- pacts. Other impacts which the site may ment. There was no comment on level of pose cannot be effectively predicted, A certainty that groundwater monitoring RCRA cap and groundwater treatment would in fact detect plumes of contami- would mitigate these problems as well as nation. most of the unseen, long term problems” ● No consideration was given to the effect (emphasis added). of leaving wastes in the ground and to the In the FS for the Pristine site in Ohio, land- effect of contaminants that have already filling the contaminated soil was rejected “Be- migrated into the subsurface. These sub- cause there is no treatment of soils to reduce surface contaminants can cause future the mobility, toxicity or volume [it] is not a per- contamination of groundwater that moves manent remedy. . . and [it] is the least preferred into and through the site area and eventu- under SARA” (emphasis added). In the FS for ally into the Arkansas river, even though 28

capping reduces water infiltration through at the Seymour Recycling site in Indiana, it was the site surface. six locations per acre; and at the Tacoma Tar ● Although the ROD acknowledged the SARA Pits site in Washington, it was one-and-a-half requirement to review the chosen remedy, locations per acre.) Hence, there was insuffi- which leaves waste onsite, every five years, cient data to consider how partial excavation there is no explicit commitment to doing and incineration for the most contaminated so. areas might be cost-effective. General conclusions.-The remedy selected (cap- Doing enough soil sampling to assess a site ping) and its supporting analysis do not satisfy accurately enough to detect hot spots has been statutory requirements on remedy selection. studied by EPA. Soil sampling is a major ef- The selected remedy is not, as the ROD as- fort: “Systematic sample site selection is nor- serted, a permanent remedy. A number of state- mally used when attempting to determine areal ments in the ROD contradict the claim of per- extent of contamination or when evaluating manency. For example, the possibility clearly spatial variations. Sampling locations are de- exists for future remedial action because wastes fined by a grid or coordinate system and sam- are left untreated in the ground: “If however, ples are collected at preselected locations in future migration does occur appropriate reme- a uniform pattern. ” (R.J. Bruner, “A Review dial actions will be taken.” The long-term un- of Quality Control Considerations in Soil Sam- certainties, the potential environmental risks, pling,” Quality Control in Remedial Site Inves- and future cleanup costs for capping have not tigation, American Society for Testing and Ma- been examined. Moreover, the perspective on terials, 1986, pp. 35-42.) The critical tradeoff land disposal and capping in this ROD is in- between the cost of taking more or less cleanup consistent with work at other Superfund sites. action has been summed up by EPA: “If the EPA said that the selected remedy is less envi- cost of a false positive (incremental cleanup of ronmentally effective than thermal destruction; additional area) is less than the cost of a false therefore, the chosen remedy is less cost-effec- negative (health risk due to not cleaning an tive. Despite the extremely high cost of total area), then the larger probability of false posi- incineration, the issue of the environmental ef- tive is acceptable. If the [contaminant concen- fectiveness of capping remains. If capping is tration action level] were raised, the probabil- not effective, then its lower, more attractive cost ity of false positives (unnecessary cleanup] does not make it cost-effective and does not would be lessened, but with an increase in the make it a permanent remedy. It is not an either- probability of a false negative (leaving a ‘dirty’ or situation. area).” (G.T. Flatman, “Design of Soil Sampling Programs: Statistical Considerations,” Quality To reduce cost, the partial incineration op- Control in Remedial Site Investigation, Amer- tion of hot spots of contamination could have ican Society for Testing and Materials, 1986, been a compromise option. Perhaps spending pp. 43-56.) The latter happens when average two to three times more money than capping, site concentrations are used to decide what instead of 20 to 30 times more for complete in- cleanup to perform, because the average value cineration, could have provided a permanent, is below the action level. If hot spots are found, cost-effective remedy. The ROD suggested that their concentrations will be above the action the site area is some 100 acres, but a statement level and false negative (dirty) areas, as well in the responsiveness summary refers to 32 as false positive (clean) areas, are avoided; that acres for the cap. In either case, the amount is, dirty areas are cleaned, but clean areas are of soil sampling at the site—28 locations—was not. insufficient to accurately characterize contaminant distribution. (Assuming there are 32 acres, At the Compass site, another strategy could sampling is about one location per acre. For have been to delay cleanup or to see capping comparison, at the Renora site in New Jersey, as an interim remedy until more work could sampling was done in 12 locations per acre; be done to fully examine alternative treatment 29 technologies. The large amount of waste at the Case Study 3 site poses a difficult problem for which some Conservation Chemical Company, Kansas innovative cleanup technologies, including in City, Missouri, EPA Region 7 situ techniques to avoid excavation, could have been considered and examined in treatability Capsule OTA findings. —A hydraulic containment studies. system to pump and treat some contaminated groundwater and capping of the site were cho- Why the State and the community chose cap- sen over the alternative of excavating and treat- ping, an impermanent and incomplete remedy, ing contaminated soil and buried wastes, which over incineration and groundwater treatment, was recommended in an EPA study and by the which are permanent and complete, is not en- State. Water treatment cannot remove all the tirely clear. But avoidance of higher costs and diverse contaminants at the site. No estimate pessimism about the safety of incineration seem was said to be possible for the duration of the to be the critical factors. Concerns about in- cleanup. cineration can be addressed through effective communication of state-of-the-art incineration Key ~atm: technology including effective pollution con- Entered Superfund system: 1/1/79 trol technology. The concern about enormous Preliminary Assessment: 3/1/79 cost for incineration could have been addressed Site Inspection: 3/1/79 - 11/1/80 through either the hot spot or search for alter- National Priorities List native treatment strategies. –proposed date: 4/1/85 —final date: none Actually, the groundwater contamination —site rank: none problem did not get enough attention, as cap- RIFS start and completion: Complex his- ping cannot completely stop further contami- tory of studies by PRPs and EPA nation. (The ROD had a tacit acknowledgment Public comment period before Record of of a carcinogenic risk factor of 1 in 100,000 for Decision: 3/26/87 - 5/8/87 contaminant migration through groundwater Signing of ROD: 9/30/87 seeps entering the Arkansas River. It is not cer- Estimated complete remediation: none tain that the cap alone, or even in combination possible with groundwater cleanup, would permanently reduce this to EPA’s typical goal of limiting risk Total time.—Unpredicted but probably a very to 1 in 1 million.) long time–decades. This site also illustrates a subtle and largely Brlaf description of site.–” The site is approxi- ignored issue in Superfund cleanups—cumula- mately 6 acres in size and is situated on the tive risk. To what extent is a cleanup at one floodplain of the Missouri River near the con- site planned relative to neighboring cleanup fluence of the Missouri and Blue Rivers, on the sites that can contribute environmental risk to river side of the levee. [The aquifer under the the same population? Assessment of environ- site] is used as a source of drinking water by mental risk at only one Superfund site seldom both private residents and public water sup- acknowledges human exposures from another ply companies.” site and, therefore, what seems for one site to “Waste disposal operations began [in 1960] be a safe level of contamination, exposure, and and continued until approximately 1980. CCC risk may not be so cumulatively. The Compass employed a variety of waste handling practices, Industries site and the Sand Springs petrochem- including but not limited to solvent incinera- ical Complex Superfund site face each other tion, solvent resale, pickle liquor neutralization, across the Arkansas River. The migration of cyanide complexation, chromic acid reduction, contaminated groundwater into the river from and ferric chloridelferric sulfate recovery. Re- both sites would increase the danger to the same sidual materials from the various treatment downstream users. processes were generally disposed of on site 30 in the basins. Drums, bulk liquids, sludges, and The 1987 ROD indicated a previous ROD in solids were buried at the site. Some wastes, such mid-1985 that adopted a circumferential con- as drummed cyanide wastes and arsenic and tainment approach with interior pumping. But phosphorus containing wastes, were disposed its implementation was stopped in 1986 when of on site without treatment. . . . approximately geotechnical investigations found that the depth 93,000 cubic yards of materials are buried on to bedrock ranged so high (to 160 feet) that “the site. ” construction of a circumferential impermeable barrier could be more difficult than origi- Major contamination/environmmtal threat.—There are nally believed, ” (Neither SCAP or EPA’s mas- 21 substances “substantially in excess of appli- ter list of all RODS indicates an earlier ROD cable criteria or standards for water quality. for this site.) These include six metals, cyanide, four pheno- lic compounds, and 10 volatile organic com- Many cleanup alternatives for the Conserva- pounds (VOCS).” Other substances “cause con- tion Chemical site have been examined, and cern for aquatic life.” Also, dioxin was detected most have been eliminated. Because the enact- at levels up to 29 parts per billion (ppb). (A level ment of SARA came after the initial studies, of 1 ppb has been EPA’s guideline for soil EPA performed two more studies in 1987. How- cleanup.) ever, the current ROD evaluated only three main cleanup alternatives in what is a well- The greatest risk comes from the use of con- structured and well-presented analysis: 1) the taminated groundwater. Next is the risk from 1985 remedy, 2) onsite containment of contami- contaminated soils, which may be “transported nants by onsite pumping and groundwater by precipitation runoff into surface water bod- treatment, and 3) excavation followed by soil ies or the groundwater. Contaminated soils also treatment. present hazards from direct contact and wind dispersion of particulate.” The 1987 ROD chose a remedy that includes: 1) the use of a permeable cap to allow water T~he] groundwater is considered to be a cur- intrusion to assist groundwater cleanup; 2) a rent drinking water source since groundwater withdrawal well system to achieve an inward is used for drinking water within a two mile groundwater gradient; 3) a groundwater treat- radius of the site.” ment system based on several unit operations, HRS scores .—groundwater 51.02; surface water including “at a minimum, such treatment proc- 9,45; air 0.00; total 29.99. esses as metals precipitation (utilizing both hydroxide and sulfide precipitation), filtration, Removal actions.–None indicated in ROD; SCAP biological treatment, and carbon absorption”; indicated over $2 million spent on removal, and 4) offsite groundwater monitoring. Some starting in 1985. descriptions of the chosen remedy from the Cleanup remedy selected. —This remedial cleanup ROD on the selected remedy are: ● 66 is the third selected for the site. The first clean- . . . relies on hydraulic, rather than struc- up, done by the original owner, was discovered tural, containment to prevent migration of in 1985 to have failed. The State had approved contaminants from the site. ” a closure “which called for the addition of ab- ● “Although designed primarily for contain- sorbents and cementing materials to the waste ing the on-site contaminants, [it] would also in the uppermost 5 feet of each basin. Waste clean up a portion of the off site contami- acids, predominantly pickle liquor, and fly ash nation. ” were mixed with the upper layer of waste ma- ● “while the treatment technologies that will terials in the basins. Tests conducted in 1985 be employed provide high levels of treat- indicated that the desired pozzolonic cement- ment, they do not remove 100 percent of like properties have not formed. Also there are the contaminants. ” indications that this material has deteriorated ● “This cleanup process could take a sub- and will continue to deteriorate. ” stantial time period. ” 31

● “There is no methodology available to esti- water cleanup. Such estimates have been made mate the length of time required for at other Superfund sites. If there is insufficient cleanup.” data to make such an estimate, then there is ● “This cleanup would include the discharge a remarkable degree of uncertainty about the of acceptable levels of contaminants re- functioning of the groundwater pump and treat maining in the groundwater after treat- approach. ment to surface waters and the need to dis- pose of solid wastes resulting from the The important feature of the selected remedy groundwater treatment processes.” is that it does not directly deal with the con- <6 ● .0. because an active pumping system taminated soil and buried wastes on site. Draw- relies upon the use of currently available ing water through the site, or flushing, is not technology, which can be constructed in likely to remove all contaminants. Depending the shortest time frame, this alternative on soil conditions and what chemicals are would provide expeditious implementation present, some contaminants are difficult to re- of the remedial action with substantial cer- move by flushing. It is difficult to conceive of tainty as to its effectiveness in protecting water drawn through the site’s hazardous ma- public health and the environment.” terials being able to dissolve or otherwise re- ● 66 . . . may prove to be the least costly move all the diverse contaminants at the Con- remedy that would meet the environmental servation Chemical site. And what will happen goals and requirements of CERCLA.” when the pumping is stopped? Indeed, the ROD did not claim complete removal of the site’s con- Satisfaction of SARA statutory requirements: tamination. 1) Selection of permanent cleanup.—The The ineffectiveness of flushing was shown ROD said the selected remedy “is cost-effective, at tests at the Volk Field Air Force site in Wis- consistent with a permanent remedy [and] consin. EPA laboratory research on the use of applies permanent solutions and alternative surfactants to remove organic contaminants treatment technologies to the maximum extent from soil had been successful. However, the practicable.” On the issue of permanency: “Con- field study found that in situ soil washing with taminants present at the site will be contained aqueous surfactants “was not measurably effec- at the site, thereby eliminating further uncon- tive,” (U.S. Environmental Protection Agency, trolled releases into the environment.” Also: “Project Summary—Field Studies of In Situ Soil “Treatment of hazardous substances to reduce Washing,” February 1988.) their volume, toxicity and mobility by treating the extracted groundwater is the principal ele- The selected remedy leaves a very large ment of [the selected remedy] . . . the uncertain- amount of untreated hazardous material on the ties of [the soil treatment alternative’s] techni- site. The ROD said that protection against the cal feasibility at this site raise substantial risk posed by contaminated soil will be ad- question as to its practicability. Extensive re- dressed first by the permeable cap on the site search would be necessary prior to its imple- during the pump and treat stage and second mentation to resolve this question. For these by the placement of a RCRA cap upon comple- reasons, [the selected remedy] offers treatment tion of the groundwater cleanup. Limiting site to the maximum extent practicable.” access is also offered as a means of minimizing risk. Regarding the dioxin contamination: The selected groundwater treatment part of “Since all the [dioxin] containing samples were the remedy, however, does not involve destruc- obtained from sludge and surface soil samples, tion technology to a large extent. Most of the the waste containment strategy and surface cap unit processes are separation technologies. The will minimize possible contact with TCDD. ” result is the generation of hazardous sludges requiring management and the discharge of A major issue for the Conservation Chemi- contaminants to the air. Moreover, no attempt cal site is the rejection of the alternative of ex- was made to estimate the duration of the ground- cavation followed by soil treatment, which an 32

EPA contractor recommended in a special On safety: “The potential safety risks for study for Superfund’s enforcement program [it] appear to be greater.” (Jacobs Engineering, “Analysis of Alternative On public acceptance: “the alternatives are Remedial Action For The Conservation Chem- generally equivalent based on anticipated ical Co. Site, Kansas City, MO,” March 1987.). public acceptance.” (Avery different state- The study concluded: “The only treatment ment than one in the Jacobs Engineering method which would meet the environmental report.) protection goal (of permanent removal or de- On cost: The ROD’s estimated cost of the toxification of contaminants), and therefore the selected remedy is $21 million and for soil only method likely to gain public acceptance, treatment $24 million. “While there are a is excavation followed by soil treatment. There- number of uncertainties for each alterna- fore, we recommend that excavation followed tive . . . [they are] the greatest for [soil by soil treatment be considered as an alterna- treatment].” tive treatment technology, under the require- On operation and maintenance: “[Soil ments of SARA, for implementation at the CCC treatment], if feasible, should require a’sub- site.” Admittedly, the study dealt with soil treat- stantially shorter period of operation and ment or washing in general terms. Moreover, maintenance than [the others].” even this treatment is a separation technology which would, like the selected groundwater Z) Accurate assessment of land disposal and treatment, produce concentrated residues which containment alternatives.—Although there is would have to be managed. Overall, the study a groundwater treatment component to the was an excellent analysis of cleanup alterna- selected remedy, the cleanup rests on contain- tives and carefully considered the pros and cons ment of the hazardous materials that are the of a number of options including in situ bio- source of the groundwater contamination. reclamation, in situ soil flushing, and excava- There was no significant analysis of the long- tion followed by landfilling. The ROD included term uncertainties and possible failures of the the recommended alternative but consistently containment and capping aspects of the clean- evaluated it more negatively than the selected up. Considering the proximity of the site to both remedy. Some relevant ROD comments sup- surface and groundwater, this lack of analysis porting rejection of soil treatment are: is a major shortcoming of the selected remedy. ● On reliability: “while [it] could be imple- In the FS for the French Limited site in Texas, mented, extensive testing and studies would use of a slurry wall and cap to contain hazard- be necessary to verify this prior to imple- ous waste was described as a “temporary solu- mentation. ” tion” for which the “volume and toxicity would ● On implementability: “[it] applies a new not be affected . . . [and] . . . the potential would technology and, as a result, there are sub- always exist for failure of either the cap or the stantial uncertainties associated with im- slurry wall allowing for the movement of un- plementation of this alternative which may stabilized wastes contained onsite. ” take considerable time to resolve before RIFS contractor. —Information on the complex this alternative could be implemented” ● RIFS history is missing, However, the SCAP On technical effectiveness: “There are still notes that on 5/21/87 there was an EPA takeover unresolved uncertainties associated with the of the RIFS, but the takeover came after the technical effectiveness of [soil treatment].” RIFS reports were completed. ● On environmental concerns: The ROD noted that there would be short-term impacts be- State concurrence.— The front of the ROD said: cause of excavation which could be mini- “The State of Missouri has been consulted on mized but not eliminated. “The option also the selected remedy.” In the responsiveness involves the discharge of low levels of con- summary at the end of the ROD, reference was taminants and the generation of treatment made to a written comment by the Director, plant sludges requiring disposal.” Division of Environmental Quality, Missouri 33

Department of Natural Resources: “The Mis- as the EPA contractor report also concluded. souri Department of Natural Resources recom- The cleanup selected for the Seymour Recy- mends the alternative incorporating excavation cling site in Indiana sets a better example be- of the wastes and soil washing with downgra- cause it includes two components to treat site dient groundwater pumping and treatment be contaminants in addition to the pump and treat utilized for the remedial action at the CCC site. component for the groundwater. The commentor stated that this recommenda- The rejection of the soil excavation and treat- tion is consistent with the final recommenda- ment alternative seems to be based on uncer- tion contained in an EPA contractor’s report tainty about its effectiveness. This uncertainty on alternative remedial action technologies at exists because no treatability study was con- the CCC site.” The ROD contained a copy of ducted as part of the RIFS. But this uncertainty an internal EPA memo stating that EPA head- must be balanced against the uncertainties of quarters and the Department of Justice support the selected pump and treat remedy: How long the selected remedy. will water be pumped and treated? What con- Cominunity acceptance. —The ROD said that there taminants in what amounts will be removed was a “low level of community concern. No and what will remain on site? How protective major public concerns have been received at is the capon the site? What is the ability of the this time.” One comment noted in the respon- hydraulic containment system to prevent con- siveness summary is from the Coalition for the taminants from moving off site in the ground- Environment, Kansas City. The questions touched water? Also, although the pump and treat can upon the length of time for the groundwater be started sooner, the ROD acknowledged that cleanup. the soil treatment remedy could be completed in a much shorter time. Special comments. —The ROD says: “Total risk from all carcinogens should be between one Many uncertainties weaken the claim that the in ten thousand to one in ten million. ” This is selected remedy is cost-effective. Even if a treat- a very broad range which could mean that ac- ability study for soil treatment was successful, tual risks associated with cleanup goals might the selected remedy—with its comparable un- be 100 times as great as the 1 in 1 million 1evel certainties—would not offer the same overall used most frequently by EPA. level of long-term environmental protection. Therefore, regardless of cost, it would not be The site’s HRS groundwater score seems low cost-effective. in the context of the information in the ROD on what contaminants are present and the use The rejected soil treatment alternative ($24 of the aquifer for drinking water. million) was estimated to cost about the same as the selected remedy ($21 million). However, General conclusions.–Contrary to what the ROD because soil treatment was only discussed in concludes, the selected remedy does not offer general terms, its cost is highly uncertain, espe- a permanent remedy which effectively reduces cially when compared to the selected remedy the toxicity of the site’s contaminants. Contami- which uses off-the-shelf equipment, The esti- nant volume would decrease somewhat through mated cost of the soil treatment alternative was groundwater treatment, because groundwater probably underestimated. Indeed, for a simi- moving through the site would flush some con- lar cleanup at the Chemical Control site the cost taminants from the soil. Groundwater treat- of soil treatment was about the same as for Con- ment, for the most part, removes some un- servation Chemical, even though the amount known amount of contaminants which maybe of material treated at Chemical Control (some landfilled somewhere else or may be discharged 20,000 cubic yards) was a small fraction of that into the air. at Conservation Chemical (the ROD indicated Although the soil excavation and treatment about 100,000 cubic yards). alternative is not a true destruction approach, EPA’s contractor report on alternative reme- it is more consistent with the intent of SARA, dial action at Conservation Chemical used a 34 figure of 298,000 cubic yards for contaminated The attractiveness of the selected remedy, soil to be treated and included about $5 mil- therefore, rested in part on its certain cost of lion (out of about $20 million) for groundwater $21 million relative to the underestimated cost pumping and treatment. Thus, the average cost of $24 million for the soil treatment option. The for cleaning the excavated soil was about $0o comparable ROD costs appear to remove low per cubic yard, a very low price for any kind cost as a deciding factor, Would the selected of contaminated soil treatment. (The ROD did remedy seem less attractive from the SARA per- not contain any theoretical or experimental spective of preferring a permanent remedy if data to show that a very simple form of soil the soil treatment option was significantly washing could be effective in removing a di- higher in cost? True, there is a legitimate issue verse set of contaminants.) for excavating materials and the risks associated with it. But such excavation has been The cost of the rejected alternative can be selected at other sites because there are estab- recalculated in two ways. First, the unit cost lished techniques to mitigate such risks (e.g., of the treatment can be altered and a range con- wetting materials to avoid dust). A technical sidered for the amount of material to be treated. case for not excavating materials, given for the A figure of $5OO per cubic yard is comparable Seymour Recycling site in Indiana because of to an estimate for a similar cleanup in the FS large amounts of volatile chemicals, was not for the Chemical Control site and for the Re- made for Conservation Chemical. Solve site. The FS for the Crystal City site had a cost of over $1,000 per cubic yard for a soil Finally, this case may illustrate the lack of washing alternative for both organic contami- management oversight of RIFSS and RODS in nants and arsenic. And a company with a mo- the Superfund program (even if the case does bile soil washing technology, not applicable to not indicate a high level of interest in reaching metal contaminants, has indicated a cost of a settlement with the PRPs). Even a cursory ex- $450 to $1200 per cubic yard. (Tufts Univer- amination of the data for and of the uncertain- sity, “Transportable Treatment Unit Technol- ties about the volume of treated material and ogies,” July 1986.] cost would probably have spotted the cost under estimate for the soil treatment alternative. The ROD acknowledges the uncertainty about onsite contaminants. If it is assumed that about 100,000 cubic yards of material would be ex- Case Study 4 cavated and treated (a figure consistent with information in the ROD), then, at a unit cost Crystal City Airport, Crystal City, Texas, of $5OO per cubic yard, the total cost would be EPA Region 6 about $5o million, about twice what was estimated. For the figure of 298,000 cubic yards Capsule OTA findings. —Excavation of contami- from the 1985 and 1987 studies and the cost nated soils and wastes (which were buried in a previous removal action) and their disposal of $5OO per cubic yard, the ● otal cost is $150 million. Thus, the range is $5o to $150 million. in an unlined landfill with a cap over it were selected over incineration. No treatability study A second way to recalculate is to ask whether supported the conclusion that the selected an estimate closer to the ROD’s can be ob- remedy is permanent on the basis of the adsorp- tained? If a volume of contaminated material tion of diverse contaminants to site soil. Major halfway between the two estimates in the re- failure modes for the landfill were not ex- ports is assumed (200,000 cubic yards) and the amined. calculation is based on a un”it cost of soil treat- Key dates: ment at $2OO per cubic yard (between the optimistic value of $5o per cubic yard and the . Entered Superfund system: 8/1/83 above $5OO per cubic yard), the total is $4o mil- ● pre~iminary Assessment: 3/1/87 lion. In this conservative scenario the cost is ● Site Inspection: 9/1/84 still about twice that used in the ROD. ● National Priorities List 35

–proposed date: 10/84 Cleanup remedy selected. —In addition to a num- –final date: 6/86 ber of containment alternatives, incineration —site rank: #639 out of 770 and critical pressure fluid extraction were ● RIFS start and completion: 9/28/85 to evaluated. The ROD described the remedy as: 7/13/87 “Onsite consolidation of alI material which ex- ● Public comment period before Record of ceeds the health-based criteria of 100 milligrams Decision: 8/10/87 to 9/14/87 per kilogram (mglkg) total pesticides. Placement ● Signing of ROD: 9/29/87 of a RCRA cap over the consolidation cell. Mon- ● Estimated complete remediation: 8/89 itor site for a minimum of 30 years following construction bf selected remedy. Deep-well in- Total time.—6 years jection of decontamination liquids. Five year Brief description of site.—” The site is comprised review of selected remedy. ” The argument was of approximately 120 acres of land. Surround- made that “By consolidating the contaminated ing the airport property . . . is land used for soil away from the runway and taxiways, land grazing animals . . . a municipal landfill . . . an use could be maintained.” Estimated cost: $1.6 elementary and high school as well as a resi- miIlion. dential area . . . Since 1949 the city has oper- Satisfaction of SARA statutory requirements: ated the facility as a municipal airport. Several private companies conducted aerial pesticide 1) Selection of permanent cleanup.—’’The applicating businesses at the airport untiI selected remedial action is considered perma- 1982.” nent. consolidating this ‘naturally treated’ waste under a hazardous waste cap is . . . con- Major contamination/environmental threat.—’’The esti- sidered permanent. ” However, while the ROD mated volume of contaminated soil exceeding did not say that an alternative treatment tech- 100 parts per million (ppm) total pesticide is nology was selected, which it was not, the ROD 12,000 cubic yards.” Although a large number did suggest that alternative treatment would be of contaminants have been detected, “The con- “inappropriate,” No reduction in toxicity or vol- taminants of greatest concern at the site (tox- ume was claimed and the ROD correctly noted aphene, DDT, and arsenic) were chosen from that these reductions are “not a requirement the compounds detected based on their wide- of the [SARA] provision. ” The ROD said that spread distribution over the entire site as well incineration ’did not conform with the Super- as the relative toxicity and concentration. ” fund statute as well as the consolidation/cap- There are also buried materials from an earlier ping remedy” and that health and environ- removal action and contaminated buildings. Di- mental protection is equal for incineration and rect contact, surface water, and air emissions consolidationl/capping alternatives. are major routes of exposure. The ,worst case exposure scenario is for residents of a.nearby The ROD’s case for permanency for consoli- housing project. dation/capping rested on these facts: ● " . . . soils [are] characterized by high clay HRS scores.—ground~ater 33.01; surface water 12.92; air 43.08; total 32.26 content and extremely low permeabilities.” ● " . . . [the] aquifer is located 750 feet below Removal actions.— Immediate removal: 10/31/83 the surface of the site and is isolated from to 11/8/83 for $33,000; 40 cubic yards of waste the contaiminated surface soils of the site and between 50 and 70 drums of material were by thick clay layers.” ● " buried in two onsite landfill cells. Second re- . . . contaminants are already highly im- moval: 4/24/84 to 4/25/84 for $25,000; 19 drums mobilized and fixed within a solid soil matrix." (FS says 21 drums) were buried in an offsite " landfill, the site was fenced, warning signs were ● . . . arsenic and organic pesticides [are] posted, and according to the FS: “eroded areas locked into [the] top foot of the alkaline soils of the clay caps were repaired. ” at the site. ” 36

Ž “The degree that contaminents are bound This uncertainty about adsorption is why a up is of the same degree that would have treatability study on actual site contaminants been achieved if the pure contaminants had and site materials would be necessary to ver- been’ processed by a solidification tech- ify that some form of effective natural stabili- nology.” zation would take’ place at’ the site. (’ The permanence of consolidation/capping is A complication at the Crystal City site is the uncertain for two reasons. First, no sound tech- presence of solvents that can affect be adsorp- nical case was made that all of the diverse range tion of other contaminants. EPA research con- of contaminants would be adsorbed tightly to cluded that” . . . the effects of solvents in haz- the site soil. A treatability study could have been ardous waste contaminated soils may include conducted to demonstrate whether significant two fadtors: 1) decrease in total sorption’to soils, leaching of contaminants is likely. No liner will nd 2) increase in leaching potentialal through be used to separate the waste from underlying changes in soil struture’U.S: Environmental soil. Data on the contaminants and the soil are Protection Agency, Review Of In- Place Treat- pertinent. For example: “The primary indica- ment Techniques for (Contaminated Surface tor used to determine the degree to which an Soils, vol. 2, Novembern1984] Also, the FS did organic contaminant binds to soil particles is not say whether any work was done to ideni- the organic carbon partition coefficient (Koc),. tify.the forms) of arsenic at the site. The abil- A higher Koc for an organic compound indi- ity of arsenic to remain immobilized because cates a greater tendency to adsorb to organic of adsorption to soil is not’straightforward. Of particles in soils, although migration may still the two Chemical forms of arsenic, the more occur throughout the site.” [Feasibility Study toxic arsenite is more mobile,than arsenate and for the Renora site in New Jersey.) K.OC ‘data adsorption,is affected by the presence of cer- are in the FS for Crystal City but were not fully tain metals in the soil and by the pH which can discussed; site contaminants vary remarkably change overtime. (U.S. Environmental Protec- (by a factor of a million) and some contaminants tion Agency, Review of In-Place Treatment have low Koc values which suggest poor ab- Techniques for Contaminated Surface Soils, sorption and the ability to migrate. vol. 2, November 1984.) Mineral surfaces can also greatly affect the ‘Second, not all the site’s contaminants are mobility of an organic contaminant. A research planned to be consolidated. The FS indicated paper that found no adsorption of phenol to that only half the site’s contaminated soil might a mineral has noted the problem of uncertainty be capped. The choice of the cleanup criterion about adsorption of organic compounds to of 100 ppm of total pesticide in combination clays: “Hemphill and Swanson found phenol with the decision to continue to allow the site adsorption on untreated kaolinite, montmoril- to be used as a municipal airport is question- lonite, and illite. Others did not find any phe- able. The ROD said that the cancer risk ap- nol adsorption on untreated clays. Luh and proaches 1 in 100,000 for onsite exposure of Baker, however, did discover significant ad- 22o days a year, which is possible for onsite sorption by clays for substituted phenols . . . “ workers and which is a higher risk than is the (E.C. Yost and M.A. Anderson, Environ. Sci. 1 in 1 million usually sought by EPA. More- Technol., vol. 18, pp. 101-106, 1984.) Another over, the FS indicated that a significant health research paper that examined the interactions threat would persist if site use was not “limited between organic compounds and clay minerals to 10 to 15 days per year.” This issue is impor- concluded that in relation to ideal, laboratory tant because of the absence of future land use conditions “rates and selectivity maybe differ- restrictions. ent and difficult to predict under environmental conditions.” (E.A. Voudrias and M. Reinhard, Moreover, no cleanup criterion was estab- Geochemical Processes, at Mineral Surfaces, lished for arsenic, which is significant because ACS Symposium Series 323, September 1985.) in correspondence to government officials EPA ——

37 said: “Of the two types of contaminants, arsen- Transportation, Tourism; and Hazardous Ma- ic compounds predominate, are more toxic, terials, Dec. 7, 1987.) Arsenic is a contaminant and more persistent in the environment [than at the Tacoma Tar Pits site in Washington, organic pesticides].” EPA also described arse- where stabilization was, selected. Moreover, nic as “the most toxic contaminant” at the.site. within the same EPA region as Crystal City, Also, risks for inhalable dust particles maybe the ROD for the French Limited site in Texas incorrect because such small particles were not said: “The PCBs and arsenic can be controlled tested to determine actual level of contami- by stabilization of the treatment residues.” nation. Biological treatment for arsenic is another On the subject of incineration, the ROD said: alternative that could have been examined in “Treatment ‘will not significantly reduce ‘the a treatability study. Then it would be unneces- mobility of the contaminants due to both the sary to design a “custom [innovative) system, ” characteristics of the contaminants as well as A recent report said: “Arsenic compounds tend the impermeable nature of the soils." In fact, to be converted by bacteria into volatile forms incineration would have offered more certain that disperse to harmlessly-low concentrations. ” permanence. The case against incineration is (R.U. Ayres, et. al., Toxic Chemicals, Health, flawed for several reasons. and the Environment, The Johns Hopkins Uni- versity Press, 1987, pp. 38-70.) However, not The ROD said: “A secondary treatment tech- all bacteria could treat arsenic and some de- nology (soil washing) would be necessary tore- velopment would probably be necessary. move the arsenic compounds from the ‘treated’ soil.” However, no consideration was given to A third alternative is a thermal treatment and the proven feasibility of using chemical fixa- recycling facility in Louisiana which has been tion or stabilization for arsenic in incinerator used to treat a number of cleanup waste soils ash followed by landfilling. There is consider- and sludges. The treatment facility can handle able information to support this approach. In very large volumes of hazardous waste. Its August 1987, EPA published extensive infor- unique process would result in a residual ma- mation in the Federal Register on effective treat- terial which appears to safely contain residual ment of arsenic in the context of the Resource metals, such as stabilization does. Moreover, Conservation and Recovery Act regulatory pro- the cost is reported to be relatively low; trans- gram. EPA said that “all the available data show portation costs must be added, but even then that the [Extraction procedure] regulatory level the total costs might be competitive to mobile of 5.0 [milligrams per liter] for arsenic can be incineration with the added advantage that the achieved.” (52 Federal Register 29992; Aug. 12, cleanup at Crystal City might be done quickly. 1987.) Moreover, OTA has examined a removal ac- Treatability tests for solvent &traction and tion at Southern Crop Services in Delray Beach, chemical fixation of arsenic contaminated soil Florida, (not an NPL site) in which mobile in- and sediment from the Vineland Chemical Co. cineration was selected for cleanup of the same Superfund site in New Jersey have been suc- type of pesticide and arsenic contamination. cessful. (These tests were done for the same The Florida kite has the same history as Crys- RIFS contractor as at Crystal City. The reports tal City Airport. EPA noted that a “naturally were filed in December 1987, although the tests rich organic layer near the surface of the soil” were probably planned and executed ‘much explains why the pesticides are concentrated earlier.) For some time, a commercial chemi- and localized and why downward migration cal fixation company has made available ex- of the pesticides into groundwater has been tensive data on the effectiveness of its treatment slowed; (U.S. Environmental Protection Agency, on relatively high levels of arsenic in incinerator Region 4, “Action Memorandum” for South- ash. Treatment costs were said to be between ern Crop Services site, Sept. 8, 1987.) As much $30 and $55 per ton. (Chemfix Technologies, as 5,000 cubic yards of contaminated soil over Inc., testimony before House Subcommittee on 2.5 acres will be incinerated at a maximum cost 38 of $2.5 million. The EPA document made no Curiously, the alternative using chemical mention of any problem with arsenic in the in- stabilization was analyzed for cost assuming cinerator ash. The cost of the incineration is that a lined landfill would be used for the treated estimated at $300 to $500 per cubic yard. Setup material. The selected remedy uses an unlined of the mobile incinerator was to begin in No- landfill for untreated material. Use of a lined vember 1987 with completion in August 1988. landfill adds an extra $700,000,j a cost that was An EPA Headquarters review on October 9, included in the stabilization option but ex- 1987, of the memorandum by the Regional of- cluded in the selected remedy of onsite con- fice added the following: “It was determined solidation. that on-site incineration was the most suitable method to cleanup the site because it destroys 2) Accurate assessment of land disposal and the hazardous waste and eliminates the need containment alternatives.—“Failure of this to transport the waste off-site.” This memoran- remedy is unlikely as long as proper mainte- dum also added that, relative to offsite land dis- nance of the cap is conducted. ” Nevertheless, posal at half the cost of incineration, the “ad- many failure modes are possible but were not ditional cost is considered reasonable because examined, including the gross, disruption of incineration provides a permanent solution.” landfilled material (disposal cell will be about Moreover, EPA said: “The disposal of the waste 190 by 190 feet) and its dispersal due to an air- in a landfill represents a less permanent solu- plane crash, perhaps with fire and explosion; tion to the problem than incineration, and is a natural disaster such as an earthquake, flood, therefore less desirable in light of SARA em- drought, and cracking of the soil that are appli- phasis on more permanent solutions.” cable to the site; the uptake of contaminants by biota, bioaccumulation, and ingestion by ani- At Crystal City Airport, the estimated cost mals in the food chain; the undetected or un- of the incineration option of $10.8 million may corrected erosion of the cap; greatly increased have been over estimated; $2.9 million was for downward transport through highly permea- the arsenic treatment (about $200 per cubic ble soils due to large-scale pathways such as yard). Without the special arsenic treatment, cracks and root holes; and the perforation of the cost of the alternative would be $7.9 mil- cap liners by animals and bugs and the subse- lion; thus the incineration option would cost quent intrusion by water. Without considering about $575 per cubic yard (total, not unit cost), these possibilities, the ROD overrates the tech- which is more than the cost range used for the nical feasibility of the selected remedy, Florida site but which agrees with the FS data for the Davis Liquid Waste site, where inciner- This ROD and its FS also illustrate another ation was also selected at a cost of about $600 common problem in technology selection—the per cubic yard. If the expensive arsenic treat- technical literature, including EPA’s own, is ment could be substantially cut, the cost for the rarely researched and cited to support conclu- incineration alternative at Crystal City might sions. For example, in a recent’ report, EPA decrease by more than $1.5 million. As noted summed up good practice with caps over land- above, use of the facility in Louisiana would fills: “Major storm events must also be consid- probably reduce costs significantly more. ered, since even an arid region can be subjected Moreover, there are instances where the ROD to infrequent but major storms that cause anom- was biased against incineration and its com- alous ground saturation and percolation to a monly accepted benefits. For example: “organic depth ordinarily not reached. Accordingly, a contaminants may be reduced through an in- rather complete review of expectable storm tegrated incineration system,” and “incinera- events and their frequencies should be required tion would remove the organic contaminants in preparing the background on the hydrologi- from the solid” (emphasis added). The truth is cal system.” (U.S. Environmental Protection that incineration would definitely destroy the Agency, “Project Summary—Design, Construc- organic contaminants. tion, and Maintenance of Cover Systems for 39

Hazardous Waste: An Engineering Guidance impermanent remedy and contributed to the Document,” November 1987.) need for remedial cleanup today. The ROD was ambiguous about the perma- Although the ROD said that “the organic com- nence of the selected remedy: “If however, fu- pounds will continue to degrade under the cap ture migration does occur, appropriate actions into less toxic compounds,” no actual data or will be taken. ” The term “significant unfore- analysis was given to support natural biodegra- seen offsite contamination” was also used. dation under the conditions expected at the site. In contrast, the ROD for the Pristine site was The FS gave data that “suggests that the con- realistic: “The lifetime of a RCRA multilayer taminants are migrating offsite through water/ cap is finite, and the contaminated soils will sediment transport.” This observation merits be left in place to contribute to groundwater more attention and an explanation of the ex- contamination at some future time should the act mechanism of transport. cap fail. . . . there are no data available on the long term effectiveness and permanence of The case for concluding that RCRA is not applicable was that “the contaminated mate- RCRA caps.” In the. FS for the French Limited rial will be consolidated in the unit or area of site (in the same EPA region as Crystal City), contamination from which they originated. ” use of a slurry wall and cap to contain hazard- This conclusion is inconsistent with decisions ous waste was described as a “temporary solu- at other Superfund sites and means that cer- tion” for which the “volume and toxicity would tain relevant aspects of RCRA on regulatory not be affected . . . [and] . . . the potential would requirements for hazardous waste landfills, always exist for failure of either the cap or the such as liners and leachate collection, were not slurry wall allowing for the movement of un- applied as required by SARA. stabilized wastes contained onsite. ” In comparison with Crystal City, these are prime exam- The data on contaminant detection frequency ples of inconsistency across Superfund sites. in table 2 of the ROD were different than the data given in the FS. RIFS contractor.—State led; $218,000 ($726,000 obligated); Ebasco Services Inc. The ROD said no groundwater was encountered, yet the HRS groundwater subscore is not State concurrence. —The ROD said the State “has zero. remained silent. ” General conclusions. —No sound technical case Community acceptance.—The ROD indicated that supported the conclusion that containing the the community favored incineration. There is wastes onsite constitutes a permanent remedy also a lot of other evidence, because of a Con- according to the intent of SARA, All of the congressional hearing in Crystal City on this issue taminants may not bind tightly to the site soil, (Apr. 11, 1988), that the community and others relevant regulatory requirements will not be strongly opposed and continue to oppose the met, health risks may be greater than normally remedy chosen by EPA. A large number of lo- acceptable levels, and a number of major fail- cal, State, and national government officials ure modes of the containment system were not and organizations have requested EPA to change examined;, its decision. The cost of the incineration’ alternative was Special comments. —The Preliminary Assessment over estimated because of the residual arsenic was completed several years after the Site In- contamination in the ash. In fact, stabilization spection, according to CERCLIS; but SCAP in- of such a contaminant has been successfully dicates that the preliminary assessment started demonstrated and is relatively low cost; bio- on 9/26/84, slightly after the site inspection, logical treatment is also known to be feasible. The first removal action (in 1983) that buried The advantages of incineration over the se- hazardous materials set the precedent for an lected remedy for the organic contaminants 40

were discounted. Moreover, in comparison to technology selection. The effectiveness of some the decision to use mobile incineration at the cleanup technologies depends on specific phys- Southern Crop Services site with a nearly iden- ical properties which can vary substantially tical type of pesticide and arsenic contamina- among contaminants. There were a number of tion, the negative view of incineration at Crys- contaminants identified-at Crystal City that are tal City Airport seems inconsistent and even not likely to adsorb tightly to the soil. For ex- contrived. ample, toxaphene is far less likely to bind tightly to soil than DDT. Both DDT and toxaphene pose Since incineration is proven for the organic a problem’ for safeguarding water quality be- contaminants at the Crystal City site and “provides better overall protection than consolida- cause both have laboratory detection limits which are above their water quality limits. (R.H. tion/capping—contrary to the ROD’s claim th`at the two choices are equal—a cost-effective Plumb and J.R. Parolini, “Organic Contamina- remedy was not chosen. The justification used tion of Ground, Water Near Hazardous Waste by EPA for picking incineration at Southern Disposal Sites: A Synoptic Overview,” paper presented at a Geological Society of America Crop Services, in terms of its greater benefits over land disposal, particularly regarding per- conference, Phoenix, October 1987.) Moreover, manency of remedy, undercuts the evaluation some chemicals—particularly solvents-may affect the adsorption of others present. Adsorp- by EPA at Crystal City Airport. tion of contaminants to soil was asserted to The FS analysis of treatment technologies for make the case that containment was similar to the Crystal City Airport illustrates a nationwide solidification treatment, but no analysis or treat- problem–current technology evaluations and ability tests were made to confirm the hypothe- the decisions based on them are not explaina- sis. Without such efforts, it is not reasonable ble by site-specific conditions. Several technol- to assume adsorption of all the contaminants ogies rejected for Crystal City could have been under all future conditions. justified as well as they were at other sites where they were chosen (and are discussed in other Keeping the cost of remedial cleanup low case studies in this report). For example, at seems to have been an important goal. The ROD Crystal City, in situ chemical stabilization was indicated that no responsible party is available rejected: “Immobilization, chemical treatment, to pay for, cleanup. The cost for the selected and physical treatments have not been shown remedy was estimated at $1.6 million, while in- to be feasible for in situ treatment of these con- cineration was estimated at $11.4 million, The taminants as it is not possible to get a good, FS contained an unusual statement: “The cost uniform, well distributed treatment.” (This of a cleanup technology is also a factor of con- technology was selected for the Chemical Con- cern in the primary screening step.” Alterna- trol site in New Jersey and elsewhere.) Biologi- tives were kept in the analysis “if their esti- cal treatment was rejected: “[it] is generally in- mated costs are not more than an order of effective for destroying these wastes as the magnitude higher than an alternative technol- treatment is not performed in a controlled envi- ogy which performs to the same approximate ronment. Several processes are being developed extent.” Generally, FSS do not cut cleanup alter- which show potential However, none of these natives from preliminary screening on the ba- processes have been developed past the labora- sis of cost. tory stage. Therefore, biological treatment has Invoking cost is done in the ROD when a been ruled out.” (This technology was selected sound case can be made for equivalent envi- for the Renora site in New Jersey.) In situ vitrifi- ronmental protection among different alterna- cation was rejected; it was selected for the Pris- tives. Then, the issue of cost and when to esti- tine site in Ohio. mate it takes on new importance because of The Crystal City site illustrates the problem SARA’s requirements on technology selection. of using a small number of indicator contami- “It is difficult enough to estimate costs at this nants not just for risk assessment but also for early [screening] stage of the feasibility study 41 when ‘old’ technologies are involved; it is hardly as scheduled, the complete remedy would end in late 1993, but this estimate maybe optimistic.) prudent to try to estimate the costs of innova- . tive technologies before a much more detailed Total time.—13 years analysis (not to mention extensive pilot testing) is performed.” (D. Truitt and J. Caldwell; “Eval- Brief description of site. —This ROD addresses an uation of Innovative Waste Treatment Tech- operable unit or interim remedial action of the nologies,” Waste Management Conference-Fo- overall remedy. “The Industrial Excess Land- cus on the West, Colorado State University, fill is a closed sanitary landfill . . . From 1968 June 1987.) At Crystal City, the big difference to 1980 the site was operated ., . for the dis- between the RIFS obligation ($726,000) and the posal of a variety of solid waste materials. Dur- actual money spent ($218,000), if the data are ing this time, the landfill accepted municipal, correct, may also indicate that less work, such commercial, industrial, and chemical wastes as treatability studies, was done than could have of substantially undetermined and unknown been done and should have been done to bet- composition, primarily from the rubber indus- ter evaluate cleanup alternatives. try in Akron, Ohio. Large quantities of chemical and liquid waste were dumped onto the ground either from 55-gal@ drums or from Case Study 5 tanker trucks. Although much of the liquid Industrial Excess Landfill, Uniontown, Ohio, wastes were listed as latex and oil at the time EPA Region 5 of disposal; witnesses have described the disposal of solvents and volatile industrial chemicals with foul odors. ” The county ordered a —Providing alternate water Capsule OTA findings. stop to the dumping of chemical wastes in Jan- to houses that have or are likely to have con- uary 1972. It was not until 1980 that a court taminated wells was a satisfactory interim re- ordered closure and a closure plan was engi- medial action. However, actions to address the neered and implemented; the site was covered source of contamination and to stop and treat and seeded. contaminated groundwater are taking a very long time. “The IEL [Industrial Excess Landfill] site is located on a tract of approximately 30 acres Key dates: which had previously been the site of mining Entered Superfund system: 12/1/80 operations (sand and gravel and possibly coal). Preliminary Assessment: 12/1/83 (from The landfill has a relatively pervious soil cover.” CERCLIS); 12/9/83 (ROD) Major contsfnination/environmental threat. -’’About Site Inspection: 8/1/84 (CERCLIS); 3/5/84 80 percent of the site is believed to be under- (ROD) lain by buried solid waste materials. There are National Priorities List over 400 residential homes located within a 0.5 –proposed date: 10/1/84 mile radius of the landfill. . . . over long periods –final date: 6/1/86 of time, the sand and gravel and immediately —site rank: #164 out of 770 underlying bedrock at IEL will act as a single RIFS start and completion: 12/28/84 to 8/87 aquifer.” The landfill “is located in permeable (final focused FS) soils without an impermeable liner.” Public comment period before Record of Decision: 8/12/87 to 9/10/87 Citizen complaints prompted testing in 1983 Signing of ROD: 9/30/87 that verified contaminated drinking water. ” . . . EPA discovered contamination of several pri- Estimated complete remediation: 12/89 for vate drinking water wells near the site. The this interim remedial action. (According to the Agency determined that the cause of the con- ROD, it will take 5 years to design and imple- tamination was the migration of hazardous sub- ment an aquifer restoration remedy. If a ROD stances from the Industrial Excess Landfill. . . . for the final cleanup is issued by October 1988 contaminants have migrated approximately 600 42

feet from the western edge of the landfill, im- 100 homes . . “ The cost was estimated to be pacting the groundwater of 10 homes. Some around $2 million. of the residential wells sampled contained organic contaminants (vinyl chloride and chlo- “The primary objective . . . is to protect hu- roethane) which are attributable to the landfill man health by providing a reliable supply of and inorganic contaminants (barium, copper, safe, potable water to residents whose ground- cadmium, and nickel) above background levels, water is currently contaminated or has the po- also attributable to the landfill. In March 1987, tential for being contaminated by IEL before U.S. EPA found levels of vinyl chloride and bar- the site itself is remediated. If unchecked, con- ium exceeding federal drinking water stand- tamination will continue to migrate westward, ards in approximately ten residential wells near affecting the groundwater of approximately 100 the landfill.” homes in a 15 year time period. U.S. EPA expects to implement a remedy for the IEL site Contamination was also found in samples before contaminants can migrate beyond this from shallow monitoring wells onsite near the projected area.” site borders at “levels which exceed standards Satisfaction of SARA statutory requirements: the observed levels of vinyl chloride [2 to 7 parts per billion (ppb)] in 3 of the 51 wells I) Selection of permanent cleanup.—Because sampled are equal to or exceed the Safe Drink- of its goal to provide alternative water as an ing Water Act Maximum Contaminant Level interim measure, this ROD did not examine or (MCL) of 2 ppb ..." select treatment technologies. “A permanent remedy at IEL will almost certainly involve A risk assessment found risks greater than some sort of groundwater treatment to reduce 1 in 1 million excess lifetime cancer risk. the level of contamination.” HRS scores.—groundwater 88.46; surface water To support the plan to provide new water for 0.00; air 0.00; total 51.13. homes not yet contaminated, the ROD correctly stated that " . . . groundwater flow and contami- Removal actions.-EPA performed interim emer- nant migration predictions are not exact sci- gency actions to protect residents in the short ences, and that predictions concerning the tim- term. The Superfund Comprehensive Accom- ing and effectiveness of remedial action are not plishments Plan (SCAP) showed that work always fulfilled ., . “ started on 12/2/85 at a cost of $973,000. The ROD contained no summary description of ex- But why wasn’t anything done to stop the actly what was done, but there were indications movement of the contaminated groundwater? of several actions, including air stripper treat- Beyond actual source control or treatment and ment for contaminated groundwater, methane groundwater treatment, several interim meas- venting, and some evacuation of houses. “While ures would have been consistent with a final, the air strippers effectively deal with vinyl chlo- permanent remedy. Examples include: 1) vacuum ride contamination, they will not remove other extraction and destruction of volatile organic hazardous substances, such as heavy metals chemicals from the site, 2) testing for and ex- and semi-volatile organics, which threaten to cavation of hot spots of contamination, 3) in- migrate from the IEL site. . . . the Agency de- stallation of a containment wall or barrier, 4) termined to go forward with a permanent alter- plume stabilization pumping, and 5) placement native water supply, rather than continuing to of a more impermeable cap or cover. Land use proceed on a piecemeal basis with air strippers, restrictions could also have been considered. whose long-term liability to protect public Such actions might well have been taken earlier; health cannot be guaranteed.” the site has contaminated local water supplies since it was closed by court order in 1980. The Cleanup remedy selected.–’’Provide alternate site’s HRS groundwater score, determined in water to an area comprised of approximately 1984, is exceptionally high. 43

The ROD granted that “The statutory prefer- accordingly. It contained an incomplete sum- ence for treatment is not satisfied because this mary of the administrative record, and it did action constitutes an operable unit for the over- not decide on the actual alternate water source, all site remedy. Treatment alternatives for the a contentious issue locally. EPA was respond- overall site will be addressed in the compre- ing to community concerns by deferring the hensive RI/FS documents.” water decision.

Z) Accurate assessment of land disposal and In the responsiveness summary, EPA made containment alternatives.—The ROD did not a very interesting statement on cost-effective- consider these types of alternatives, except that ness: “U.S. EPA is required to select the less a de facto no action on the buried wastes meant expensive alternative, given that effectiveness that the original land disposal remained in and implementability are equal” (emphasis effect. added). Adding some factor other than effectiveness is new. It could, however, act against RIFS contractor. —Camp Dresser and McKee; selecting newer cleanup technologies, which about $1 million obligated. usually have not been used on a large scale.

State concurrence. —The State of Ohio concurred The argument for not relying on air strippers with the selected remedy. was technically sound in view of the chemical complexity of the site contaminants. Community acceptance. —It was the community that called on government officials to take seri- General conclusions. —It probably will be closer ous action in the first place. For this ROD, in- to 20 years than 13 between the time the IEL tense community concerns focused on getting site was closed and the time some form of new water for more houses and on identifying source control and groundwater treatment is the exact source of the water. applied to the site. Is EPA correct that the selected interim remedy “is fully consistent with A local newspaper, the Akron Beacon Jour- a permanent remedy?” Although the ROD re- nal, November 15, 1987, reported: “Recently ferred to aquifer restoration, it does not men- the EPA agreed to connect 110 homes in tion removal of the source of contamination. Uniontown to a new water supply when 1,500 While providing alternate water does not stand were in possible danger. Fortunately, Gov. in the way of a permanent remedy, neither does Celeste saw no reason to threaten the rest and it do anything to make the permanent cleanup agreed to spend state money to build a new easier, Indeed, this ROD acknowledged that the water system for all the homes. Gov. Celeste groundwater contamination will get worse. rightly saw that any error should be on the side Thus, because additional interim measures of safety. The federal EPA’s approach seems would stop or slow down the migration and to be the opposite. A telling example was the help to alleviate the source of the problem, this release of documents showing the air tests had interim measure is not fully consistent with an been bungled. The documents became public ultimate permanent remedy. only after citizens invoked the federal freedom of information law to obtain them. Information The assertion that it will take 5 years to de- about mistakes in the air tests had been know sign and implement an aquifer restoration rem- to the EPA since Oct. 21. A responsible EPA edy was probably too optimistic and is not would have informed the public immediately. ” likely to build community confidence. Experi- ence at similar sites suggests that groundwater Special comments.–Any ROD signed on 9/30/87, cleanup can take considerably longer. For ex- the end of the Federal fiscal year, may have been ample, at the Seymour Recycling site in Indi- a rush job; the ROD for IEL may have suffered ana the groundwater treatment was estimated 44 to take from 28 to 42 years; at the Davis Liquid Case Study 6 Waste site in Rhode Island, it was estimated Pristine, Inc., Reading, Ohio, EPA Region 5 to take from 5 to 10 years; and at the Re-Solve site in Massachusetts, it was estimated to take Capsule OTA findings.—In situ vitrification was 10 years. There is, however, no obvious alter- developed originally for radioactive soils, but native for groundwater cleanup at IEL, EPA its use for chemically contaminated sites is still has said: “The actual performance of a ground unproven. Without treatability test results, in water remedial action is difficult to predict until situ vitrification was selected for this site chiefly the remedy has been implemented and opera- because its estimated cost was about half that tional data have been assessed.” (U.S. Environ- of onsite incineration. But the estimated cost mental Protection Agency, “Guidance on Reme- for incineration is probably high by a factor of dial Actions for Contaminated Ground Water two. Incineration offers more certainty and At Superfund Sites,” draft, October 1986.) probably costs no more than the selected remedy. Before a permanent remedy is implemented, Groundwater will be pumped and treated by more wells may become contaminated. Also, air stripping and carbon adsorption. contaminants can migrate at substantially dif- Key dates: ferent rates, and, therefore, the nature of the spreading contamination can change over time. ● Entered Superfund system: 4/1/79 ● Such changes are well verified by research. Preliminary Assessment: 1/1/83 (See, for example, R.L. Johnson et al., Ground ● Site Inspection: 9/1/82 Water, September/October 1985.) Wells that are ● National Priorities List contaminated early can get worse as new, more –proposed date: 12/1/82 slowly moving contaminants reach them; wells –final date: 9/1/83 not yet contaminated eventually see the effects —site rank: #531 out of 770 of the most rapidly moving contaminants, On ● RIFS start and completion: 9/5/84 to 11/87 this point the ROD noted: “Since the publica- ● Public comment period before Record of tion of the FFS [Focussed Feasibility Study], re- Decision: 11/13/87 to 12/11/87 cent data revealed that levels of nickel exceeded ● Signing of ROD: 12/31/87 Ambient Water Quality Standards.” Also, “vi- Estimated complete remediation: 8/91; 2 years nyl chloride has migrated off-site quickly . . . for source control, 5 to 10 years for ground- while its parent compounds pose a threatened water cleanup (August 1991 is given in ROD, release from the site because they are migrat- but this seems optimistic and inconsistent with ing at a slower rate. “ The ROD also noted: “the other groundwater cleanups; 10 years for ground- shallow and deep aquifers are continuous and water cleanup is more realistic) linked to one another.” Therefore, the complex- it y of groundwater contamination and its clean- Total time.—2O years up could worsen significantly. The possibility Brief description of site. —The site is in a suburb of upgradient chemical migration should not of Cincinnati. The site is 2.2 acres and “is bor- be ruled out. (See R.H. Plumb, Jr., Proceedings dered by residential and industrial areas. There Second Canadian/American Conference on are two aquifers under the site.” In the late Hydrogeology, 1985, pp. 69-77.) 1970s, a liquid waste incinerator was operated The long history of the site, the extensive at the site. “In April 1979, as many as 8,000 groundwater contamination offsite, and the de- to 10,000 drums and several hundred thousand lay in addressing the source of the problem is gallons of bulk liquids were on site, consisting feeding community lack of confidence in gov- of acids, solvents, pesticides, PCBs and other ernment efforts and demands for new water chemicals. ” A consent order shut down the fa- to more houses. cility in September 1981. 45

Major contamination/environmental threat.—” . . . over der a consent decree. Some responsible parties 90 compounds were detected in the ground- removed waste and soil from March 1984 to water, soil, sediment, and surface water. ” July 1984 under an administrative order. The ROD did not say how much material was re- “Groundwater in the upper aquifer is con- moved, nor its disposition, but it was probably taminated primarily with volatile organic com- landfilled. pounds (VOCs) such as benzene, vinyl chloride, tetrachloroethene (TCE), and 1,2-dichloroethane. Cleanup remedy selected. —This ROD was a final Semi-volatile compounds (semi-VOCs) and pes- source control remedial action but also in- ticide compounds occurred in relatively lower cluded groundwater cleanup. Another ROD concentrations. The lower aquifer is contami- might be issued for additional groundwater nated with benzene and 1,2-dichloroethane. cleanup. There are also elevated levels of lead and The key component of the selected remedy fluoride.” is in situ vitrification (ISV) for 37,700 cubic " . . . the presence of VOCs in the [Reading yards of contaminated soil and sediment. ISV municipal wells] indicates that the groundwater was chosen over onsite incineration. ISV is an quality in the vicinity is compromised and con- innovative technique that uses electrodes in the tinued monitoring is recommended . . . The ground to pass electricity through soil, melt it, lower aquifer is the source for the regional vaporize and at least partially destroy organic water supply [13,000 people]. ” This route of ex- chemicals, and leave in place a chemically in- posure poses the largest risk. ert, stable, glass-crystalline mixture. Temper- and soil in the upper two feet atures in the range of 2,000 to 3,6000 F are pos- “Sediment . . . sible. Different cells of soil are melted in order of the site are contaminated with a variety of to cover a site. The melt grows downward and VOCs semi-VOCs, and pesticides. Principal outward as power is applied. As the vitrified contaminants in surface soils are benzene, di- zone grows, it incorporates nonvolatile elements eldrin, and DDT.” Low levels of dioxins and furans were also found. and destroys organic components by pyrolysis. The pyrolysis byproducts migrate to the sur- “Subsurface soil contained VOCs . . . There face of the vitrified zone, where they combust were also elevated levels of cadmium, lead, mer- in the presence of oxygen. “The estimated time cury and zinc. ” required to complete the vitrification process is two years assuming the use of one vitrifica- “Surface water was contaminated with VOCs, tion unit.” semi-VOCs and pesticides . . . There were also elevated levels of inorganic compounds (cad- Groundwater will be pumped and treated mium, chromium, and mercury).” with an air stripper and carbon adsorption. That is, separation, not destruction, technol- A good risk assessment established cleanup ogy was selected. Groundwater monitoring was goals at 1 in 1 million risk level. The RIFS calcu- set up. The possibility of deed restrictions was lated how much soil would have to be removed raised. "It is estimated that it will take five to to “eliminate both the risk associated with ad- ten years to extract and treat the contaminated sorption and ingestion of soils and ingestion of groundwater,” groundwater contaminated through leaching from the soil.” Estimated cost: $22 million. HRS scores.—groundwater 60.00; surface water Satisfaction of SARA statutory requirements: 10.91; air 0.00; total 35.25 1) Selection of permanent cleanup.—In its ini- Removal actions. —The site operator removed tial two screening stages, the FS examined a waste from June 1980 to November 1983 un- large number of treatment technologies. How- 46 ever, after more detailed screening, many of Document: Treatment Technologies for Halo- the treatment technologies said to be applica- genated Organic Containing Wastes, vol. 1, Jan- ble were dropped without much justification. uary 1988.) This EPA report describes ISV as: For example, three in situ treatment technol- “Not commercial, further work planned. No ogies passed the initial screening, but solution [performance] data available, but DREs [de- mining and soil vapor extraction were dropped struction-removal efficiencies] of over six nines and only vitrification was retained for more reported. ” detailed analysis. Onsite treatment technologies that passed the initial screening included fixa- ISV’s developer, Battelle Pacific Northwest tion/solidification, soil washing, and dechlori- Laboratories, which supports ISV’s inclusion nation, but only incineration was analyzed in SITE, has commented on treatability test- further. ing: “While the results are promising, feasibility testing to confirm applicability is strongly The ROD said that the selected ISV remedy recommended prior to any commitment to de- “will significantly reduce the mobility, toxic- ploy the process on a site that contains signifi- ity, and volume of hazardous substances in the cant quantities of organics that are unconfined soil through treatment. The mobility of the con- in the soil column. . . . feasibility testing is rela- taminants will be reduced significantly, such tively inexpensive [a few thousand dollars]. The that no leachate is expected to be produced focus of the feasibility testing is the perform- from the vitrified material. This is a permanent ance requirements for the off-gas treatment sys- technology, the results of which are expected tem and the type and quantity of secondary to last for a million years. The toxicity of or- waste generated. Experience with low boiling ganic components will be decreased because point organics that are uncontained in the soil the organics are destroyed or changed to other column is very limited, and feasibility testing forms by pyrolysis or vaporization. The volume with actual site samples prior to application is of the soil will be reduced by 25 to 30 percent strongly recommended” (emphasis added). because the vitrification causes the soil mass (V.F. Fitzpatrick, “In Situ Vitrification–A Can- to consolidate.” didate Process for In Situ Destruction of Haz- A chief issue is whether or not ISV is a proven ardous Waste,” Proceedings of the 7th Confer- technology. ISV is an alternative treatment and ence on the Management of Uncontrolled an innovative technology that was developed Hazardous Waste Sites, December 1986, pp. originally for treating radioactive contaminated 325-332.) soils, but its use for chemical contamination ISV depends on the effectiveness of the col- raises new questions. How should ISV be clas- lection and treatment system for released gases sified? Is it thermal destruction or stabilization? to keep undestroyed organic contaminants (or EPA’s SITE technology demonstration pro- products of incomplete combustion) from en- gram categorizes it as stabilization, as do others. tering the environment, This off-gas system is (N. Nelson et al., Toxic Chemicals, Health, and like a separation technology; hazardous residues the Environment, The Johns Hopkins Univer- can be either destroyed or landfilled after car- sity Press, 1987, pp. 205-279.) Stabilization is bon adsorption. The greater the volatility of con- a reasonable label because metal contaminants taminants, the greater their release into the off- remain in the final glass-like material and be- gas collection system. At Pristine, many of the cause the leaching of metals and the complete contaminants are highly volatile at relatively destruction and removal of organic contami- low temperatures. By the time the soil is melted, nants are uncertain. Although very high tem- therefore, many contaminants have moved. peratures are reached, not all organic contaminants will either be destroyed or be able to What happens to organic contaminants in escape and be captured. However, EPA also ISV is crucial to understanding its cleanup ef- calls ISV thermal destruction. (U.S. Environ- fectiveness relative to other technologies, such mental Protection Agency, Technical Resource as incineration. A published paper reported on 47 a test of ISV on PCB contaminated soil: “Small building up concentration in the isotherm quantities of PCBs, furans, and dioxins were layers corresponding to each chemical’s boil- detected in the untreated off-gas, but none were ing point. This writer expects the DRE to be detected in the vitrified mass. A few samples only 25-50 percent.” (Larry Penberthy, letter directly adjacent to the block contained meas- to Laura A. Ringenbach, attorney for respon- urable concentrations up to 0.7 ppm PCBs. ” sible parties, Mar. 28, 1988; Pyro 32A and 32 (R.R. Battey and J.T. Harrsen, “In Situ Vitrifi- newsletter of Penberthy Electromelt Interna- cation for Decontamination of Soils Contain- tional, Inc., Apr. 7 and 13, 1988.) This vapor ing PCBs,” Proceedings of the Oak Ridge Model retreat phenomenon could lead to increased Conference, February 1987, pp. 233-245.) In contamination of groundwater. Moreover, in another report on the same experiment, the order to test for this effect it would be neces- “process destruction was slightly greater than sary to test a rather large volume of soil so that 99.9 percent. The small amount of material re- temperatures away from the molten zone are leased to the off-gas system was effectively re- low enough to have condensation of vaporized moved, yielding an overall system DRE of contaminants, However, most testing is done >99.9999 percent.” (V.F. Fitzpatrick, “In Situ on too small a volume of contained material Vitrification—A Candidate Process for In Situ to see this effect. Destruction of Hazardous Waste,” Proceedings of the 7th Conference on the Management of Penberthy has a number of other criticisms Uncontrolled Hazardous Waste Sites, Decem- of the tests performed by Battelle which, even ber 1986, pp. 325-332.) after examination of Battelle’s comments on Penberthy’s analysis, seem important enough The ROD responsiveness summary said: to require additional study and testing. More- “with worse case conditions, 97 percent of all over, Penberthy has raised important safety organics are destroyed. Most tests indicate that questions, such as effects from soil heating and 99 to 99.99 percent destruction is achieved.” subsidence, about using ISV at such a heavily Another ROD statement is more optimistic industrialized area as the one around Pristine. about destruction versus removal: “The test re- No significant examination of the risks posed sults [on PCBs] indicate that the organics are by ISV has been made. destroyed and not merely collected in the off- gas system.” All this information shows that The ROD did not focus on the depth of ISV. ISV might be very effective but that the issue The plan is to go down to 8 feet for half the of total destruction of organics through both site and 12 feet for the other half. The Battey thermal treatment and off-gas collection, re- and Harrsen article (see above) noted that the moval and possible treatment needs clarifica- greatest efficiencies for ISV occur when it is tion. Lateral migration of vaporized organics used to depths of 10 to 20 feet. The technology into adjacent soil or perhaps downward into does not work well when contaminants are on groundwater is also important and needs de- the surface and therefore, soil covers are some- tailed resolution for application of ISV to any times used, The ROD also noted: “The equip- large, uncontained site. ment must be specially designed and produced, ” The depth and equipment issues are possible This last issue has received major attention causes of underestimated costs. by Larry Penberthy, who calls it vapor retreat. While Penberthy is a competitor of ISV, he The other big issue for the Pristine site is the makes a good technical arguement: “Instead rejection of onsite incineration. The ROD ac- of being destroyed, the vaporizable chemical knowledged: “incineration is a proven technol- contaminants simply move away from the hot ogy. “ “Incineration ., . is fully protective of hu- core melt by Vapor Retreat, unaltered. They man health and the environment since the move downwardly below the melt core as well ingestion and leachability threats are elimi- as horizontally away from the melt core. This nated.” The ROD did not acknowledge that vaporizing/condensing action is progressive, some stabilization of the incinerator residues 48

might be necessary because of toxic metal con- $300 to $500 per cubic yard for less than 5,000 taminants. But overall, the ROD did not note cubic yards of soil. any disadvantages of incineration: “The use of mobile incinerators is common and the per- Actual (bid) costs for incineration can vary formance of these systems has been demon- but do not explain the high estimate used for strated. It is relatively easy to operate the system Pristine. A recent report showed a cost of about although a trained operator will be needed.” $750 per cubic yard (comparable to the Pris- tine ROD estimate) for a cleanup, but the soil Nevertheless, incineration was not selected. quantity was under 10,000 tons and there was The ROD said: “vitrification is the lower cost more than just soil to clean. (J.F. Frank et al., alternative. Therefore, incineration is not rec- “Use of Mobile Incineration to Remediate the ommended for implementation at the Pristine, Lenz Oil Site,” paper presented at Superfund Inc. site.” However, the costing for incinera- ’87, 1987, pp. 459-464.) At another site, a ven- tion seems too high. The ROD’s total cost esti- dor (a subsidiary of the Pristine FS subcontrac- mate of $51 million for incineration was based tor) got $250 per ton under a turnkey arrange- on a unit direct cost of $730 per cubic yard (for ment; total costs probably were about $450 per 37,700 cubic yards). Meanwhile, the FS said that cubic yard for cleaning between 7,500 and the unit cost ranges from $350 to $500 per cu- 10,000 tons of PCB contaminated soil. (J.W. bic yard. Several other recent FSS (for the Davis Noland, remarks at Weston Environmental Fo- Liquid Waste and Re-Solve sites) by the same rum, Washington, DC, February 1988.) Mobile RIFS contractor provided detailed vendor costs incineration was used to incinerate materials and analysis for onsite incineration. From those at the Nyanza Superfund site in Massachusetts. two FSS, OTA used the cost data for three differ- The vendor charged about $600 per cubic yard ent technologies for a range of contaminated for a very small quantity, about 200 cubic yards. soil to be treated (4,300 to 57,000 cubic yards) At the Prentiss Creosote Superfund site in Mis- and obtained a (conservative) estimate of a unit sissippi, a vendor charged about $200 per ton cost of about $300 per cubic yard for the level ($300 per cubic yard) for mobile incineration of effort at Pristine. This range is consistent of 7,500 tons; at the Southern Crop Services information that other vendors gave to OTA. Superfund site in Florida, the vendor charged EPA said recently that mobile infrared in- $360 per ton for 3,000 tons, cineration of contaminated soils costs “from The Pristine ROD cites no technical factors $120 to $225 per ton [which could be as high to explain a $730 per cubic yard cost the site. as $180 to $340 per cubic yard], depending on For example, no mention has been made of the number of tons incinerated per day.” (U.S. buried drums. Even if the ash were to be chem- Environmental Protection Agency, memoran- ically stabilized because of toxic metal content, dum from John H. Skinner, Office of Research the additional cost would not account for the & Development, Dec. 10, 1987.) The FS for Sey- cost discrepancies noted above. Moreover, in mour Recycling gave costs for onsite incinera- an internal inconsistency the FS calculation for tion over a very broad range of amounts of con- a cleanup of only 8,100 cubic yards used a unit taminated soil: for 35,000 cubic yards the cost cost of $658 per cubic yard; instead of the ex- was $186 per cubic yard; at the smallest scale pected higher unit cost for a smaller volume, (18,000 cubic yards) the unit cost was $349 per a lower figure was used. cubic yard. In the Crystal City FS the unit cost for onsite incineration was $240 per cubic yard In the groundwater cleanup, the ultimate dis- for about half the amount of material at Pris- position of the collected hazardous substances tine. In a recent decision for a Superfund re- is uncertain because it is not clear how the car- moval action at the Southern Crop Services site bon that becomes contaminated by removing in Florida, where mobile incineration was se- organic substances will be managed. The ROD lected, EPA said that it expected bids at from said: “Bench scale studies will be done to de- 49 termine the need for metals treatment.” Air ● “[There is] limited demonstrated perform- stripping only removes volatile organics and ance. 99.9999 percent DRE [destruction carbon adsorption is not likely to be effective removal efficiency] are expected for dioxin for the metals. Maximum reported values of and PCBs.” lead in groundwater are 178 ppb and 148 ppb ● “Because this is not a proven technology, in the upper and lower aquifers; the drinking prior to implementation of this remedial water Maximum Contaminant Level (MCL) for action, bench and/or engineering pilot lead is 50 ppb and the proposed MCL Goal is scale studies will be required to confirm 20 ppb. For cadmium the corresponding ground- the effectiveness and applicability of this water levels are 39 ppb and 9.4 ppb and the technology to site conditions.” standards are 10 ppb and 5 ppb. The complex- ● “If this treatment method is found to be ity and intensity of groundwater contamination ineffective, this Record of Decision may are great enough to warrant more detailed anal- need to be reopened.” ysis of groundwater treatment, as was done for ● “Monitoring will be conducted during the the Operating Industries site in California. treatment process to determine if contamination is migrating through the soil as a Any cleanup of contaminated surface water result of the treatment.” (Such monitoring at Pristine is left uncertain by the ROD. is not routine with in situ techniques.) 2) Accurate assessment of land disposal and containment alternatives.-The ROD contained ISV is not easy to implement—at least, the an excellent rationale for not leaving contami- technology costs more—when water content is nated material onsite: “It may leach into the high. Thus the ROD noted: “Because of con- groundwater at levels that will exceed ARAR’s cern over the effectiveness of vitrifying the up- [regulatory standards] at some future time and per outwash lens, consideration will be given, thus increase the groundwater treatment time during these bench and/or pilot studies, to or require additional future remedial action. whether the lens should be drained prior to The lifetime of a RCRA multilayer cap is finite, vitrification. ” Also, the responsiveness sum- and the contaminated soils will be left in place mary said twice that the site’s soil has “high to contribute to groundwater contamination at moisture content” when it defended why vacuum some future time should the cap fail.” The extraction of VOCs was not feasible. However, responsiveness summary said: “there are no when the selected ISV remedy was defended, data available on the long term effectiveness the responsiveness summary—four pages later and permanence of RCRA caps.” —said “the moisture content . . . is not high. ” Either high is high, or actual measured values By its very nature, ISV leaves treated contami- could be used to show it is high for one tech- nated material onsite. The ROD acknowledged nology but not too high for the other technol- uncertainties about ISV: ogy, if that was the case; however, no actual data were used. (This may illustrate a lack of ● “Vitrification is expected to be effective on ROD quality control and ROD rushing at the the soil type present. . . “(emphasis added). end of a fiscal year quarter.) ● “Air monitoring will be conducted to ensure that the hood is collecting and treat- To its credit, the Pristine ROD specified: “The ing the gases.” (Such monitoring is not rou- Toxicity Characteristic Leaching Procedure tine when mobile incinerators are used.) (TCLP) is the testing mechanism that should ● “Some limited monitoring of the vitrified be used to verify the complete treatment. If this mass will be required to assure that it is treatment method is found to be ineffective, this a reliable and permanent remedy.” (Such Record of Decision may need to be reopened. ” monitoring is not routine when stabiliza- However, a recent technical report said that tion is used.) vitrified contaminated soil performed poorly: 50

“The vitrified product as evaluated with these nated site.” In August 1987 it was reported that standard leaching tests did not perform well. EPA Region 5 had “conditionally accepted” The reason for this is not known. It maybe that ISV for an emergency response action for PCB the nature of the tests maybe inappropriate for contaminated materials at the Greiner’s Lagoon monolithic, vitrified masses, or vitrification site in Fremont, Ohio. (Hazardous Waste News, might not be as effective as chemical stabiliza- Aug. 24, 1987.) As of May 1988, OTA was in- tion for simple metal systems.” (J.J. Barich et formed by EPA that no date had been set for al., “Soil Stabilization Treatability Study at the the test—the actual removal action at the site— Western Processing Superfund Site,” paper pre- and EPA confirmed that this site action would sented at Superfund ’87 conference, 1987, pp. constitute the ISV test for EPA’s SITE program. 198-203.) According to the report, for example, Only if new test data confirm the presence of leaching of zinc–over a short period in a stand- high concentrations of PCBs will ISV be used ard EPA leach test—was about 10 times greater at the Greiner’s Lagoon site and, even then, than for conventional stabilization. The respon- probably not before Spring 1989. If the PCBs siveness summary portrayed much more cer- are low, then another site would probably be tainty: “the metals are encapsulated and bound selected, delaying the SITE demonstration still up in the ISV process.” The responsiveness more. The Pristine responsiveness summary summary also said: “ISV has been tested on haz- said that the demonstration will be performed ardous waste and has been successful.” These prior to use of ISV at Pristine. two statements contradict the Barich findings. Failure of the vacuum off-gas collection sys- Battelle has said that the TCLP is technically tem is possibile and of concern because of the inappropriate for monolithic waste forms. high population density near the site. The re- Their principal concern is that reducing treated sponsiveness summary said: “Should this oc- material to fine particles—and exposing un- cur, the organics will be rapidly dispersed in bonded contaminants–is based on the false as- the air, allowing for a very low probability of sumption that treated material may not main- any adverse impacts through inhalation down- tain monolithic properties. Does this mean that, wind of the site. ” if Battelle is successful in making its point about the TCLP, the Pristine ROD will be reopened ISV received very detailed examination in the because the TCLP will not be used to test ISV’s FS for the BF Goodrich and AIRCO site in Ken- effectiveness? tucky, which was completed several months after the Pristine ROD. ISV was not selected at A recent EPA study that examined eight the BF Goodrich/AIRCO site, primarily because emerging treatment processes for decontami- its high cost made it not cost-effective. Several nation of PCB contaminated sediments ranked of the comments about ISV in that FS (from a ISV last by using two sophisticated methodol- different EPA region and contractor) are im- ogies. The report said: “all the processes ex- portant relative to the decision to use ISV at cept In Situ Vitrification appear to merit fur- Pristine: ther development for this application.” (U.S. Environmental Protection Agency, “Report on ● “The effectiveness of off-gas collection and Decontamination of PCB-Bearing Sediments,” treatment is not known. ” October 1987.) While PCB contamination is not ● “The complexities of this repetitive proc- the dominant problem at Pristine and sediments ess [incremental movement across a site] pose a special problem for ISV unless they are are not known since it has not been fully dewatered, this report is important because demonstrated on a large site.” almost all previous information on ISV has ● “There is very little data available as to come from its developer, including a lot of em- whether vitrification is a reliable technol- phasis on tests on PCB material. The Pristine ogy.” (Compare this to the Pristine ROD: ROD said: “An additional application [of ISV] “results indicate vitrification to be a relia- is being planned by EPA for a PCB contami- ble technology.”)

1------51

“The mass could take several years to cool, pleted “several gaps were identified” and ad- depending on the size; therefore, a tem- ditional work was done which took another porary fence should be constructed to pre- year. vent physical contact with the cooling ma- State concurrence.—The ROD said that Ohio’s terial.” letter of concurrence is forthcoming. This sug- “Implementation of the vitrification proc- gests that the ROD was rushed to get it out by ess has yet to be demonstrated on a com- the end of the fiscal year quarter. mercial scale . . . It is probable that the frequency of operation and maintenance Community acceptance.–“The community and problems would be higher than for a proven PRPs are generally in agreement with the ground- technology. As a result, reliability would water extraction and treatment component of be lower than for proven technologies. the alternative. Some members of the commu- Emissions during implementation would nity have fully supported U.S. EPA’s recom- require extensive control for both VOCs mended alternative, while the PRPs rejected and dust. ” These concerns are especially vitrification and have proposed installation of relevant to Pristine because of its location a RCRA cap with soil gas venting. The City of in a highly industrialized area. Reading prefers that U.S. EPA fund a less expensive remedial action and give it the remain- Similarly, ISV was rejected at several Super- ing funds to build a new municipal treatment fund sites in Colorado in 1987 (Denver Radium plant.” Operable Units: Open Space, Card Property, 1000 West Louisiana Properties, and 12th and Special comments.—The FS had an initial discus- Quivas Properties). What is especially interest- sion of the two main treatment alternatives ing is that contamination by radioactive mate- where unit costs and other data were presented. rials would seem to be an ideal application of For in situ vitrification the sole technology de- ISV because it was originally developed as a veloper, Battelle Pacific Northwest Labora- cleanup technology for radioactive materials. tories, was acknowledged and was the source However, all three RODS say the same thing: of the data. For onsite incineration, the source “[ISV] was eliminated during the initial screen- of information is described as “Firm A.” Only ing because its implementability for this par- rotary kiln incineration is used in the detailed ticular application is unproven. [ISV] has not analysis. Because a wholly-owned subsidiary been demonstrated on a large scale of utilized of the FS subcontractor, Roy F. Weston, Inc., in a highly-populated urban area like that of does cleanups with a transportable rotary kiln the Card property.” Moreover, in these cases incinerator it seems likely that subsidiary was ISV was also rejected because of the possible Firm A. “escape of radon gas and associated radon de- The ROD acknowledged some uncertainty cay products.” This would seem to also be appli- about the groundwater cleanup: “The extent cable to escape of trapped gaseous organic con- of contamination from Pristine, Inc., will be taminants at a site like Pristine, and the concern determined by additional studies during the re- undermines the belief that monolithic, solidi- medial design.” However, this ROD addressed fied ISV material offers secure, tight, and per- the final source control remedy and should not manent encapsulation. be expected to be definitive about the groundwater cleanup. The ROD also noted that con- RIFS contractor. –Camp Dresser and McKee tamination in the lower aquifer may be the (CDM); the cost was at least $500,000. The ROD result of “a multi source groundwater contami- did not indicate who did the FS, but a copy of nation problem in the area. ” According to the the FS shows it was CDM with Roy F. Weston ROD, a variety of other types of actions might performing the FS as a subcontractor. The ROD be used, including RCRA corrective action. said that several figures in the FS were wrong and recalculated figures were used in the ROD. The responsiveness summary had an inter- The ROD also said that after the RI was com- esting interpretation of the provision in SARA 52

Section 121(b)(2) that says that a technology can preferential handling in the FS and ROD: “Be- be selected even if it has not “been achieved cause of its obvious bias in favor of ISV . . . the in practice” elsewhere. The interpretation was FS does not properly evaluate all existing rele- that the provision allows a technology selec- vant technologies . . . “ EPA was accused of be- tion without directly applicable data on its ef- ing “arbitrary and capricious. ” EPA defended fectiveness before the ROD. Legally, this posi- its selection at great length, including a discus- tion seems correct because congressional intent sion of why vacuum extraction of VOCs, appar- was not to prevent full-scale application just ently considered a viable alternative by the re- because there has been no prior full-scale ap- sponsible parties, is not applicable to the site. plication of a new technology. But traditional (The technology was not analyzed in the FS.) engineering practice does not condone choos- However, EPA’s discussion did not resolve the ing or using a technology without supportive questions raised here about ISV, nor did EPA site-specific test data. The issue is that ISV has go into any technical depth in discussing vacuum not been tested sufficiently at Pristine nor on extraction. Extensive work done for the respon- large, unconfined contaminated soil not the ab- sible parties indicates that vacuum extraction, sence of successful full-scale use of it at a simi- which has been selected for cleanups else- lar site. where, may be feasible and cost-effective at Pristine (see below). To the extent that ISV re- The responsiveness summary would have duces risk mainly through removal of volatile been more useful had it provided the sources organic contaminants, it performs functionally of specific comments, which is normally done. like vacuum extraction removal of volatile or- General conclusions.—There area number of out- ganics. But vacuum extraction is intrinsically standing aspects to the Pristine ROD. The com- a lower cost technology that uses less capital mitment to meeting SARA’s preference for per- intensive equipment and energy than does ISV. manent treatment technology is excellent. The The problem is the decisionmaking process commitment to an innovative technology is and the accuracy of crucial data upon which commendable. The risk assessment is excep- it is based, As with so many sites, treatability tionally good, detailed, and well presented. For tests for Pristine were postponed to the post- example, using a 1 in 1 million risk for inges- ROD Design Phase even though test data are tion and direct contact, the cleanup targets for necessary to fully support the selection of soil are 3,182 ppb and 15,041 ppb for benzene remedy. This criticism is not directed at the ISV and trichlorethylene; but, when removing the technology itself, which might eventually work threat to groundwater from leaching of the con- at the site and which is an important new taminated soil was used, the targets for the two cleanup technology. But the wisdom of choos- chemicals decreased to 116 ppb and 175 ppb ing ISV for the Pristine site remains question- to meet drinking water standards. able for several reasons. Unlike most FSS, the Pristine FS presented Consider the following initial criterion to be a preferred alternative that was recommended complied with for a technology to pass the ini- for implementation. It is OTA’s understanding, tial screening in a feasibility study: “There must from speaking to RIFS contractors, that EPA be a demonstrated history of successful use of usually directs them not to give a recommended the technology in environments similar to the cleanup. Why was ISV so strongly supported . site. All technologies of a research and de- for Pristine, especially before the use of it in velopment nature, and which cannot be rea- the removal action planned by Region 5? sonably said to be in common use, are rejected.” As reported in Pristine’s responsiveness sum- This criterion is from the 1987 ROD for the mary, EPA’s selection of ISV got a poor recep- Northern Engraving Corp. site in Wisconsin tion, apparently from responsible parties, be- in the same EPA Region as Pristine. ISV tech- cause of its unproven state, high cost, and nology could not meet that criterion, nor is the —

53 criterion consistent with SARA, but the point given in the FS is $250 to $350 per cubic yard. here is inconsistency within the Superfund The FS used a mid-range value of $290 per cu- program. bic yard. If a unit cost of $350 is used, it leads Either EPA’s SITE program needs to prove to a total cost of $26 million instead of the ROD in situ vitrification or the technology should estimate of $22 million. (Total costs include bur- be released from the demonstration program dens and groundwater cleanup.) and accepted as proven technology in ROD The FS contained initial screening costs for selections. Without treatability study results, the two treatment options: ISV and incinera- the uncertainty about ISV’s effectiveness, when tion. At the screening stage, the two costs coupled with no cost advantage over proven looked comparable: ISV at $13 million and in- incineration (see below), weakens Pristine’s cineration at $18 million. These figures seem ROD and the government’s attempts to get re- to come from using the upper range of the unit sponsible parties to take over the cleanup. Al- costs given in the FS ($500 for incineration and though SARA allows the use of new innovative $350 for ISV, with no burden added). But the cleanup technologies, there is little engineer- ISV cost increased by 70 percent in the final ing or public support for selecting an unproven cost calculations in the FS, while the incinera- technology without supportive data. Data are tion cost increased by 183 percent, with no ex- necessary to substantiate a technology’s abil- planation provided. There is more uncertainty ity to meet specific cleanup goals, especially about the cost of the ISV option, however, than when other proven and cost-effective perma- about incineration and just the opposite change nent treatment technologies are available. More- in cost would have seemed more plausible. over, in the Pristine FS, treatment technologies As discussed earlier, the cost of onsite in- other than incineration were eliminated from detailed evaluation with no technical basis. Al- cineration was seriously over estimated and beyond EPA’s standard allowable range of +50/ though the FS and ROD were clearly commit- ted to using a treatment technology, the alter- –30 percent. (This allowable range, itself, is natives became very restricted. Aside from ISV large enough to invalidate a technology decision based on cost.) The FS did not use the $425 and conventional incineration technology, no per cubic yard mid-range value from its own consideration was given, for example, to above incineration data ($350 to $500); data which ap- ground vitrification of excavated soil in a fur- pears higher than reliable cost estimates. OTA nace. This technology was said, by its developer (Penberthy Electromelt International, Inc.), to recalculated the cost of the onsite incineration option. Instead of using $730 per cubic yard cost $180 per ton ($280 per cubic yard) in di- (which the ROD used to calculate incineration rect unit costs; a transportable furnace would costs), OTA used $300 per cubic yard (see dis- take about 500 days to perform the work at Pris- cussion above). Including all the other costs for tine. These figures are competitive with ISV. the incineration alternative, such as ground- Is EPA prepared to alter its decision in the water cleanup, as done in the FS, the total cost Design Phase if test results are negative? The for the incineration alternative then becomes ROD said that manufacture of the specially de- $23 million (close to ISV’s cost). If the cost of signed equipment “will occur concurrently the ISV option is also recalculated to reflect the with the remedial design. ” In other words, a high end of the cost range supplied by the ven- big investment will be made before test results dor, then its cost is $26 million. The conclu- can support the selection. sion is: onsite incineration is not likely to be The cost of the ISV choice may have been more expensive than ISV at Pristine. under estimated for several reasons discussed In these recalculations the indirect or bur- earlier (depth of treatment and building spe- den costs (83 percent) are those used in the Pris- cial equipment). The range of unit cost for ISV tine FS; however, these are much higher than 54

those in the FSS for the Davis and Re-Solve sites. site incineration is likely to cost more than At Pristine, various contingencies, construc- vacuum extraction of volatile organic contami- tion services, and design costs amounted to a nants, which the responsible parties favor. But burden of 83 percent, while in the Davis and vacuum extraction is a separation technology, Re-Solve FSs the burden is 35 percent; the lat- and an important issue (as it is for ISV) is what ter explicitly included costs for pilot study is done with the extracted contaminants. If they work, while the figures for Pristine did not. For are destroyed rather than landfilled after car- Crystal City, the burden was 29 percent; for bon adsoption, the costs increase. Moreover, Chemical Control, 56 percent; and for Seymour the diversity of contaminants at Pristine re- Recycling, 60 percent. For the BF Goodrich/ quires careful analysis of vacuum extraction’s AIRCO site in Kentucky where ISV received ability to remove them; it might be able to do so. extensive examination, the total overhead on the ISV direct costs was 60 percent (and the Although at Pristine ISV was rated compara- unit cost was $275 per cubic yard). The markup ble in effectiveness to incineration and better at Pristine was substantially higher than the for cost, its implementability is lower than that markups in these other recent FSs. That is, both for incineration because ISV has not been rou- treatment estimates at Pristine would have been tinely used for chemical waste cleanups. ISV lower if a lower burden were used; with 35 per- has a higher level of uncertainty with regard cent, used by the same RIFS contractor in other to site conditions, and there is a need for site- FSs, the cost for ISV becomes $16 million in- specific design. Indeed, the FS said, “There is stead of $22 million and the incineration op- more data to support incineration . . . Of the tion (at $730 per cubic yard) becomes $38 mil- soil contaminant destruction alternatives, only lion instead of $51 million. [incineration] has a demonstrated performance and reliability, ” Incineration offers consider- Indeed, mobile incineration might be less ably more certainty as to effectiveness, relia- costly; if the low $186 per cubic yard figure from bility, and cost. the Seymour site (which is in agreement with other data given above for recent actual con- Cost aside, incineration is a less risky selec- tracted costs) is correct and is used for Pris- tion at this time in the absence of treatability tine, with the same high burden rate as the other study data that could remove uncertainties alternatives, the total cost of the rejected in- about ISV for the Pristine site, especially with cineration option is about $15 million [instead regard to off-gas collection and treatment, the of the estimated $51 million) versus $22 mil- migration of contaminants into surrounding lion for the selected remedy of in situ vitrifica- soil, the degree of destruction of all organic site tion. With the lower, more typical burden rate contaminants, and safety uncertainties for the of 35 percent, the incineration option comes surrounding community. to about $11 million (versus $16 million for ISV There is another uncertainty about the im- at the same, lower burden rate). The Seymour plementation of ISV. Battelle has exclusive site is in the same EPA Region as Pristine. These rights from the Department of Energy to mar- two sites illustrate that more regional oversight ket ISV for nonradioactive sites. The ROD ac- is necessary to catch inconsistencies in data knowledged that “ISV is a patented process critical to technology selection. which requires a license.” Other companies The lower figures for incineration are impor- who do actual cleanup work are not familiar tant in the context of the government obtain- or experienced with the technology. However, ing a settlement with the responsible parties; the Pristine responsiveness summary said: if this estimate proves correct on closer scru- “The selection of ISV is not patently unfair tiny and true costs of incineration at Pristine since the developer will be licensing firms to are indeed much lower than the cost of ISV, carry out the process and the bid process will then incineration becomes the more attractive be competitive.” Subsequently, Battelle changed cost-effective permanent remedy. However, on- the way it offers the technology such that its 55

availability and, hence, competition may be re- ● Preliminary Assessment: 8/1/82 stricted. Rather than supplying the technology ● Site Inspection: 9/1/81 to 8/1/82 to cleanup companies, which has not proven ● National Priorities List a successful strategy, Battelle has helped form –proposed date: 12/1/82 a new company with startup capital. The new –final date: 9/1/83 company, the GeoSafe Corp., will seek addi- —site rank: #378 out of 770 tional venture capital and will enter the haz- ● RIFS start and completion: 5/85 to 8/87 ardous waste cleanup business directly with ● Public comment period before Record of ISV; it has the exclusive worldwide rights for Decision: 8/18/87 to 9/10/87 this market. No competitive bid process now ● Signing of ROD: 9/29/87 ‘ appears possible. ● Estimated complete remediation: 1 to 2 The ROD omitted any commitment to SARA’s years after signing requirement for 5-year reviews when hazard- Total time.—8 years ous material remains onsite, a requirement which applies in this case because of the stabili- Brief description of site.–The site “is an approxi- zation aspect of the technology. For example, mately one acre parcel of land in an area zoned the 5-year review was called for in the Chemi- for light industrial use. The surrounding area cal Control ROD, which selected in situ stabili- is residential with three sensitive uses (a nurs- zation, and at the Tacoma Tar Pits, which se- ery school, senior citizens center, and an apart- lected stabilization. ment complex) within two thousand feet of the site. . . . two residential developments [were] The Pristine case illustrates how different built in close proximity to the site during the offices of the same EPA contractor and how period of time the RI/FS was conducted. From different Superfund contractors can use sub- 1978 to 1982 Renora Inc., transported and ac- stantially different data. One contractor’s of- cepted materials containing hazardous sub- fice used a cost for incineration for Pristine of stances for transfer, storage, blending and ulti- about twice what another of the contractor’s mately, disposal through abandonment at the offices used for the Davis Liquid Waste and Re- site [in 1982].” Solve sites. A close examination of the calculations for estimated costs at Pristine reveals Major contamination/environmental threat.—Evidence that a very high indirect or cost burden was of contamination problems started in 1978. In used, compared to indirect costs in FSs for sev- 1985, the Remedial Investigation (RI) said: “Sur- eral sites. Such cost variations have no techni- ficial soils (0 to 2 feet) are primarily contami- cal basis. nated with polychlorinated biphenyls (PCBs) and polynuclear aromatic hydrocarbons (PAHs) and to a lesser extent with volatile organic com- Case Study 7 pounds (VOCs), acid extractable compounds Renora, Inc., Edison Township, New Jersey, (AECs), other base/neutral organic compounds EPA Region 2 (BNCs) and heavy metals. Shallow groundwater beneath the site is contaminated with low levels Capsule OTA findings:-The selected remedy makes of chloroethane, (a volatile organic compound) use of offsite landfilling for soils contaminated and heavy metals. Surface water and sediment with PCBs. Biological treatment was selected samples show levels of heavy metals, tetra- for soils contaminated with diverse organic chloroethene, phenols and pesticides. No evi- compounds and toxic metals and for contami- dence of air contamination was found at the nated groundwater, but no treatability study site, No buried drums were found at the site. ” supported its selection. The RI concluded that the significant path- Key dates: ways of exposure are direct contact and sub- ● Entered Superfund system: 5/1/81 sequent incidental ingestion by children tres- 56 passing the site, future onsite workers, and A number of cleanup alternatives were ex- future site residents. amined, including containment approaches, treatment of less material, use of incineration According to the ROD:” . . . there are no off- instead of landfilling, and conventional ground- site impacts directly attributable to site operations. Therefore, no management of migration water treatment. measures were selected as part of the overall Satisfaction of SARA statutory requirements: remedy for any environmental media. Although groundwater does not pose a public health risk, 1) Selection of permanent cleanup.—The ROD said: “Overall, [the selected remedy] is protec- achievement of target treatment/residual levels tive of public health and the environment. An will result in restoration of groundwater quality to potable water standards. ” innovative treatment technology would be utilized as a major portion of the remedy. There HRS scores.-groundwater 69.32; surface water is complete reduction of the toxicity, mobility, 9.4; air 0.00; total 40.44 and volume of the contamination. The remedy is permanent and would not require long-term Removal actions.-The ROD said: “A removal ac- management” (emphasis added). More cau- tion was initiated in October 1984 and continued through April 1985. During the cleanup, tiously, the ROD said the remedy “significantly reduces the toxicity, mobility and volume of approximately 33,000 gallons of liquid waste contaminants” (emphasis added). Also: “Upon and 28,000 gallons of PCB contaminated waste oil along with approximately 500 cubic yards completion of the remedy future site uses will be unrestricted.” of non-PCB contaminated soils and 560 cubic yards of PCB-contaminated soils were shipped There is no specific technical information in off-site for proper disposal [presumably in a the ROD or Feasibility Study to support the landfill].” The SCAP shows a Federal removal selection of biological treatment for the Renora action 10/23/84 to 10/31/84 at a cost of $27,000 site. There are no test data, no citations to the and that the responsible parties performed one, technical literature, nor reference to previous 9/28/84 to 4/16/85. Data from the New Jersey use at specific sites. The ROD stated: “A prereq- Department of Environmental Protection indi- uisite to implementation of the bioremediation cates that the responsible parties spent $4 mil- portion of the alternative is a pre-design treat- lion for their removal action. ability study to refine parameters of the operation.” There are a large number of contami- Cleanup remedy selected.—The remedy has four key components: nants, and many of the organic contaminants and heavy metals are considered difficult to bi-

1. “ . . ● excavation of all PCB-contaminated odegrade. The biological approach is not off- soils containing concentrations above 5 the-shelf cleanup technology, except for a few ppm [parts per million] (approximately simpler types of cleanups. 1,100 [cubic yards]) and off-site land fill disposal . . . “ A key issue is the extent of destruction by biotreatment. While it can be easy to get some 2. “. . . biodegradation of all PAH-contaminated soils containing concentrations destruction, it can be very difficult to get complete destruction or as much, for example, as above 10 ppm (approximately 4400 [cubic required for incineration (99.99 percent de- yards]) . . .“ struction). Finding ways to enhance biodegra- 3. “ . . . use of groundwater as an irrigation medium for the bioremediation system dation for a complex set of chemicals and for recalcitrant contaminants can be difficult. For . . . “ [and] example, a recent research paper discussed the 4. “ . . . backfilling, grading and revegetation.” “degradation of Benzo[a]pyrene and other The cost of the selected remedy was estimated recalcitrant PAHs” and explained its failed at- at $1.4 million. tempt to foster biodegradation by noting that

I 57

“organic amendments which are readily uti- operations before full clean-up is effective. lized for carbon and energy are often ineffec- Other technologies, e.g., incineration, would tive in stimulating degradation of recalcitrant provide equal destruction efficiencies in a organic compounds.” (M.P. Coover and R.C. shorter time frame. ” Sims, Hazardous Waste & Hazardous Materi- The Renora ROD said: als, vol. 4, No. 2, 1987, pp. 151-158.) ● " ... bioremediation of soils is considered The current state of technical knowledge and an innovative treatment technology in the experience does not support the cleanup selec- field of hazardous waste management.” tion in the absence of site-specific data to prove ● “Although available scientific literature ineffectiveness in meeting the cleanup goals. dicates implementation of the bioremedi- There are substantially different forms of bio- ation portion of the alternative is feasible; logical treatments, ranging from simple land a pre-design treatability study would be re- treatment to the sophisticated use of bioreac- quired to confirm the operational reliabil- tors using a variety of additions to promote and ity of the alternative. ” sustain biological destruction to desired resid- ● " . the bulk of the contaminated soils (ap- ual levels of contaminants, but the ROD dealt proximately 60 percent of the total) which with the technology only in its simpler, generic remain are amenable to onsite bioremedi- terms. ation. Available scientific literature and its Biodegradation was selected for the French use in the oil refining industry indicate that Limited site in Texas, but the ROD emphasized the bioremediation aspect of the selected that “biodegradation of PCBs to the criterion alternative will achieve the target treat- (23 ppm) has not been demonstrated.” EPA re- ment/residual levels.” quired, therefore, that a secondary stabilization “The prospect for long-term reliability of treatment be used on the residue from the bio- the alternative would be established by the treatment. pre-design treatability work and subsequent verification sampling. However, as The FS for the Liquid Disposal site in Michi- this remedy is permanent and substantially gan examined biological treatment in more de- reduces the toxicity, mobility and volume tail than most studies and did not select it. A of contamination the likelihood of remedy chief reason was: “The level of effectiveness replacement is low” (emphasis added; com- of the biodegradation technologies on a non- pare to different statement above). homogeneous waste stream is unknown.” The study noted that extensive testing would be nec- The choice of offsite landfilling over inciner- essary to prove the technology effective for the ation was not discussed in detail in a direct way site. for the alternative selected. Other alternatives which would include more use of incineration, Biological treatment was rejected in the Fea- because no bioremediation would be used, were sibility Study for Crystal City because “[it] is said to “not result in providing any greater pro- generally ineffective for destroying these wastes tection of public health or the environment that as the treatment is not performed in a controlled would justify the incremental cost increase. ” environment. Several processes are being developed which show potential. However, none Z) Accurate assessment of land disposal and of these processes have been developed past containment alternatives.—The use of offsite the laboratory stage. Therefore, biological treat- landfilling for the PCB contaminated soil in- ment has been ruled out. ” Biodegradation was stead of treatment is contrary to the intent of also rejected in the ROD for the Tower Chemi- SARA, even though the amount is relatively cal Superfund site in Florida: “Biodegradation small. A recent EPA study on PCB cleanup con- does not address the metals contamination cluded: “Landfilling of such materials, where found at the site and would require long term legal, is a potential source of groundwater con- 58

lamination, and only a temporary measure at condition and since the HRS groundwater best.” (U.S. Environmental Protection Agency, score was quite high. However, the responsive- Office of Research & Development, “Bengart ness summary did include a monitoring step & Memel PCB Site Soil Decontamination in its description of the selected remedy. Project,” undated but apparently 1987.) The FS General conclusions. —A key issue is the choice for the Liquid Disposal site in Michigan said: of offsite landfilling over offsite incineration “However, moving wastes from one site to for the PCB contaminated soil. A major driv- another does not constitute a permanent re- ing force behind SARA’s requirements for per- medial action” (emphasis added). For Renora, manently effective treatment technologies was there was no discussion of the negative aspects the long-term ineffectiveness of moving buried of using offsite landfilling. The Renora ROD hazardous waste from Superfund cleanup sites looked strictly from the perspective of this site: to other land disposal sites. This ROD, consid- “Excavation of PCB contaminated soils and off- ering only this site, implied that offsite disposal site landfilling will physically remove hazard- results in maximum protection. It did not con- ous substances, pollutants and contaminants sider the long-term consequences at another site from the site.” of landfilling materials transported from this RIFS contractor.—The RIFS was paid for by a site. In addition to the two EPA sources already group of potentially responsible parties and noted, the FS for the Pristine site in Ohio re- conducted under contract by BCM Eastern Inc.; jected the option of sending contaminated soil $250,000; an endangerment assessment was to an offsite landfill because “ . . . there is po- done by Camp Dresser and McKee under con- tential for the contaminated soil to cause a prob- tract to EPA. lem at the off-site facility. . . . the alternative is not permanent and is the least preferred under State concurrence.—The State of New Jersey con- SARA.” The reasons for ruling out offsite in- curred with the selected remedy. cineration of the PCB contaminated soil at Community acceptance.—The responsiveness sum- Renora were not given. mary gives little information on what the com- The Renora remedy also perpetuates a trend munity felt about the selected remedy. started by the major removal action completed Special comments. –The analysis of cleanup al- at the site in 1985. A viable alternative, as ex- ternatives was somewhat confusing because amined in the ROD, was to incinerate the waste some options, including the selected one, re- instead of landfilling it. Indeed, the ROD con- ferred to offsite disposal consisting of either tained an important statement on this point in landfilling or incineration. Except for cost, the discussion of the alternative that was se- there were no distinctions made within an alter- lected: “If the excavated PCB contaminated native for the use of landfilling versus inciner- soils are incinerated instead of landfilled, there ation. Therefore, the selected alternative might would be a permanent reduction in the toxic- have received overly high evaluations because ity, mobility and volume of contaminants in incineration was included as an option but ulti- soils.” The clear implication is that landfilling mately not selected. is not comparable in meeting SARA’s requirements. The chief reason for not selecting in- The ROD contained a good statement on cap- cineration of the PCB contaminated soils ap- ping: “[it] would not be considered permanent pears to be its greater cost, an additional $4.6 since the toxicity and volume of contaminants million. The ROD noted: “it is likely that the in the soil would remain essentially unchanged.” [potentially responsible parties] will implement The ROD did not commit to groundwater the selected remedy.” In other words, approval monitoring after the selected remedy is imple- of offsite landfilling by EPA may have facili- mented, which seems relevant, since it says that tated getting an agreement from the responsi- the groundwater will be restored to a potable ble parties to clean up the site. 59

The interest in using innovative treatment self and selected solidification for most of the technology is commendable, and some biologi- cleanup, which the responsible party had found cal remedy may, in fact, be found effective. But effective in its treatability study. Incineration the ROD decision was made without support- is to be used if solidification technology is not ing technical information. The ROD and the successfully demonstrated or fails after solidi- FS contained no details about the technology. fied material is landfilled on the floodplain site, There are many forms of biological treatment, but the criteria for failure are unspecified. and it is still too new a cleanup technology to make assumptions on effectiveness at a site Key dates: with so many organic contaminants, some at ● Entered Superfund system: 8/1/80 relatively high concentration. Technology ● Preliminary Assessment: 6/1/80 specificity is high for biological treatment, ● Site Inspection: 11/1/80 which means that it is difficult to extrapolate ● National Priorities List success from one waste to another. Delay of –proposed date: 10/84 treatability testing until after the ROD creates –final date: 6/86 considerable uncertainty and the potential for —site rank: #761 out of 770 actions which are not fully protective of pub- • RIFS start and completion: 6/29/84 to 5/4/87 lic health and environment because of either ● Public comment period before Record of substantial loss of time or a compromise of Decision: 7/29/87 to 9/1/87 cleanup goals if the testing shows problems in ● Signing of ROD: 9/29/87 the biotreatment. Indeed, because of the need ● Estimated complete remediation: 11/91 for extensive treatability testing, the estimated time for complete implementation seems overly Total time.—11 years optimistic. Such biological treatment (both aer- Brief description of site.–”The site operated as a obic and anaerobic processes) of contaminated refinery from the turn of the century through soil was rejected at an early screening stage in the 1940s. The property has since been devel- the Pristine FS because “Mixed wastes and low oped as an industrial area and consists of an concentrations (less than 100 ppm) are difficult abandoned solvent and waste oil recycler, an to treat. ” The same condition exists at Renora. active transformer salvage/recycler, active Because of the use of landfilling and the selec- chemical manufacturers and various other in- tion of an unproven treatment technology, the dustries. . . . the site is located on the northern selected remedy cannot be assured to be per- bank of the Arkansas River, immediately west manent. Moreover, compliance with SARA’s of Tulsa, Oklahoma. The site encompasses ap- requirement on the reduction in toxicity, mo- proximately 235 acres [and] includes unlined bility, or volume was described in three differ- acid sludge pits, a surface impoundment, sur- ent ways: complete, substantial, and significant. ficial sludge contamination, solvent and waste This puzzling situation may indicate end-of- oil lagoons and contaminated sediments. The fiscal-year ROD rushing or confusion over the [site] is located in the alluvial floodplain of the capabilities of the selected remedy. Arkansas River.” Major contamination/environmental threat.—’’Total Case Study 8 known waste volume is approximately 130,000 Sand Springs Petrochemical Complex, cubic yards. During the period of operation haz- Tulsa County, Oklahoma, EPA Region 6 ardous substances were stored or disposed of in drums, tanks, unlined pits and lagoons or Capsule OTA findings.–EPA originally said solid- buried on-site. These substances include vari- ification technology was ineffective for the high ous volatile and non-volatile organics, chlori- organic content site wastes and incineration nated solvents, and sludges containing heavy was effective. Nevertheless, EPA reversed it- metals. Waste pits have contaminated local 60

groundwater and caused migration of surface solidification, accepting a five-part proposal by contaminants. ” Atlantic Richfield Company (ARCO), a responsible party, that included: EPA concluded that there are four major sources of risks: direct contact with organic car- 1. excavation and offsite thermal destruction cinogens and highly acidic wastes and surface of some unspecified volume of surficial waters; air emissions of acid fumes and vola- sludges; tile organic compounds; surface waters pol- Z. solidification and/or stabilization of all re- luted by runoff during heavy rains; and ground- maining sludges and containment of the water being contaminated directly by lagoons resulting material in an onsite RCRA haz- and indirectly from site runoff: “in heavy rains ardous waste landfill; the site is submerged.” 3. demonstration that solidification technology meets EPA approved criteria and, According to the FS: “It is believed that con- should it not do so, use thermal destruc- taminants from the pits, ponds, and lagoons are tion [apparently onsite]; leaching into the alluvial aquifer, therefore, a 4 no liability release for the site or from fu- major pathway for migration is probably ground- ture maintenance and monitoring; and water. However, by definition this pathway has 5 repair or restoration of the landfill to en- been excluded from consideration during the sure no migration or destruction or treat- FS for the Operable Unit.” The same contami- ment of all or a portion of its contents, as nation problem exists for surface water migra- EPA deems appropriate, should monitor- tion offsite. The FS also said: “Several on-site ing show that the solidification/stabiliza- ponds and lagoons have a history of breaching tion remedy fails. their containment structures: there have been incidents of dike walls breaching for one of the It appears that ARCO is anticipated by EPA Glen Wynn lagoons, as well as flow of materi- to sign a consent decree, agreeing to pay for als from the river acid sludge pits into the Ar- the cleanup. kansas River, which have occurred in the past Cost data on the selected remedy is absent . . . . the contents of the large and small acid because the combination of solidification and sludge pits had breached their dike walls on incineration was not evaluated in the FS. It is several occasions.” unclear how much material will be incinerated HRS scores.-groundwater,44 .90; surface water offsite initially. But if solidification is used it 21.82; air 0.00; total 28.86 will cost less than incinerating all the waste. Incineration was estimated in the ROD to cost Removal actions.—A private party performed a $67 million and complete solidification was removal action in 1984; there are no details in estimated at $38 million (the comparable figures the ROD. in the FS are $54 million and $31 million). OTA Cleanup remedy selected.—EPA designated this estimates that the probable comparable cost of cleanup as a source control operable unit that the five-part remedy is $45 million, but this fig- covered surface liquids, sludges, and heavily ure is highly uncertain because there are many contaminated solids but not minimally contami- different forms of solidification. nated soil or groundwater. The latter is to be Satisfaction of SARA statutory requirements: addressed in a subsequent ROD. Originally, before the ROD was officially signed, EPA I) Selection of permanent cleanup.—EPA said selected onsite incineration of wastes and in the ROD that the selected remedy, based solicited public comment on it as part of the mostly on solidification, fulfills the statutory RIFS public comment period; the agency had preference. However, as EPA stated: “on-site already evaluated solidification and onsite land- thermal destruction of wastes . . . appears to fill and solvent extraction, all of which were meet more statutory selection criteria than the rejected. EPA changed its mind and selected other remedies evaluated. ” With solidification, 61

“unlike on-site thermal destruction, the toxic- The FS summed up its evaluation of onsite ity of wastes would not be reduced and the vol- solidification: “Not a proven technology for ume of wastes would be increased.” high organic waste. Contaminant source isolated, may not be rendered nonhazardous. May EPA views in the ROD on the chosen solidifi- not meet ARAR.” It remains unclear whether cation option at this site included the following: the test results supplied by ARCO removed all ● " . . . [there was a] lack of demonstrated of the above concerns for EPA; but the inclu- permanence.” sion of the third provision in the five-part ● " . . . the capability of solidification or remedy which requires demonstration of solid- stabilization techniques to permanently ification technology suggests that EPA was not bind with high organic wastes, such as fully convinced by the ARCO test data or that those found at Sand Springs, has not been it did not have enough time to fully evaluate demonstrated in the pilot studies conducted it prior to signing the ROD. on-site. ” ● " . . . without further treatment free liquid EPA has tied the environmental acceptabil- contaminant concentrations were not re- ity of the solidification remedy to two condi- duced to meet RCRA land ban restrictions.” tions: “if the effectiveness of this concept is ade- ● " . . . the unconfined compressive strength quately assured or if ARCO undertakes the of the stabilized material . . . does not meet corrective actions deemed appropriate by EPA the recommended disposal criteria. ” should the remedy fail.” No such conditions ● " leaching tests conducted by EPAs would have been attached to the originally cho- Cincinnati laboratory show that the solid- sen thermal destruction remedy, which ARCO ified material leaches contaminants.” also examined in its treatability study and ● " . . . leaching of contaminants, and incom- which was found to work effectively. Cause for plete encapsulation [small globules of waste EPA rejecting the incineration option in the were seen] raises questions about the long ROD was said to be its “serious implementa- term effectiveness and permanence of the tion problems,” but EPA’s FS analysis also said process.” that all the processes that would treat waste on- ● the waste” . . . contains 50 percent organic site “are judged to each have the same degree compounds raising doubts about the abil- of implementability.” Moreover, the ROD stated: it y of stabilized or solidified waste to meet “Actual implementation time for solidification RCRA requirements in the long term.” and thermal destruction is comparable . . . “ ● samples “ . . . show obvious degradation EPA said: “The proposed remedy is consid- of the solidifying matrix following analy- ered permanent.” And that it “is cost-effective sis for total organic content. ” " compared to equally environmentally protec- ● the net assessment is that solidifica- tive alternatives.” But EPA also said that the tion or stabilization processes present dif- thermal destruction alternative offered more ficult problems with respect to meeting overall protection than solidification. The FS ARARs [standards].” " summed up its case for onsite incineration: ● . . . possible air emissions.” " “Proven technology destroys hazardous mate- ● . . . volumetric increase of 50 [to 200 rial, Containment source worker health and percent].” safety addressed in remediation. Meets ARAR.” ● " . . . the potential for failure was determined to be greatest for the on-site solidifi- 2) Accurate assessment of land disposal and cation remedy.” containment alternatives.—With regard to fu- ● " . . . the source of the contamination will ture operation and maintenance, the ROD said not be destroyed.” that these “will be minimized since the source ● additional “ . . . studies will need to be of the contamination will be removed.” But, performed on the subsurface petroleum as the ROD also stated, solidification does not wastes. ” destroy the source of contamination that will 62 be left onsite. The ROD contained a replace- meets the design specifications outlined in ment cost of $100 million should failure occur RCRA.” for the onsite solidification and landfilling option but no cost for the incineration alterna- The issue of effectiveness of solidification tive. For these two pure (single technology) op- technology for organics is critical for Sand tions, the reduction in cost of about $30 million Springs. The FS for the Re-Solve site in Massa- for solidification is offset by a possible future chusetts rejected stabilization because “there re-remediation cost of $100 million. And, that has been limited success in chemically fixing tradeoff still exists. organic contaminants such as solvents and PCBs.” The ROD for the Liquid Disposal site The onsite landfill is supposed “to reduce in Michigan, which also selected stabilization groundwater infiltration and the chances of any for soil contaminated with organic chemicals, contaminants migrating off-site. ” But if “ . . . said that the hazardous substances “will not significant, unforeseen, off-site migration or be permanently destroyed” and “hazardous contamination occurs as a result of the site, chemicals still remain in that [treated] mass. ” appropriate remedial measures will be taken, ” And the FS for the site said: “Considerable re- No detailed analysis of future failures was search data exists demonstrating the effective- given. The ROD did not express concern about ness of this technology in immobilizing a wide having the onsite landfill in a location that is range of contaminants, primarily inorganic. submerged in heavy rain and that is in a flood- A substantial amount of data does not exist, plain adjacent to the river. The ROD stated fur- however, to accurately judge the long-term ther: “While a hazardous waste landfill of so- reliability of the process. Long-term leaching lidified waste would protect health in the short and volatilization can be expected for soluble term, the long term stability of this material is and volatile organic wastes.” Although, stabili- not proven. ” zation was selected for Liquid Disposal, so was For Sand Springs, the importance of the the use of a slurry wall and impermeable cap water level to remedy selection is striking. The around and over the treated material, as a sec- ROD for the Tower Chemical Superfund site ond level of control. The ROD for the French in Florida, for instance, commented on the use Limited site in Texas (same EPA region as of chemical stabilization followed by onsite Sands Spring) said: “Fixation is questionable land disposal, a remedy the ROD rejected. The due to high organic content of untreated soils.” comments apply directly to the Sand Springs An EPA report’s observations on halogenated site: “Although this process is effective in ad- organic wastes also apply to the selection of dressing inorganic contamination, the volume chemical stabilization for Sand Springs: “the of materials would increase, thus causing in- area of solidification/encapsulation is one re- creased disposal facility requirements. In addi- quiring additional study before it can be con- tion, the soils being solidified contain signifi- sidered viable technology. ” (U.S. Environmental cant amounts of organic compounds which Protection Agency, Technical Resource Docu- could affect the integrity of the cement mono- ment: Treatment Technologies for Halogenated lith. The presence of organics will require con- Organic Containing Wastes, vol. 1, January tainment of the monolith within an on-site land- 1988.) fill built above the land surface due to the locally high water table. This technology would also Another EPA document, used to teach cleanup require long-term (30 years) monitoring which workers about waste treatment says: “Solidifi- is less favorable than technologies which pro- cation technologies are designed to be used for vide permanent destruction of wastes. . . . a final waste treatment. This means the technol- high water table at the site makes it infeasible ogy should be applied only after other treatment to solidify or build an on-site landfill which techniques have been applied, i.e., incineration, 63 chemical treatment or other. ” (U.S. Environ- of the area’s air pollution, and harm to the lo- mental Protection Agency, “RCRA/CERCLA cal economy. (Building an incinerator for in- Treatment Alternatives for Hazardous Wastes,” dustrial waste is frequently sold by industry to October 1987.) communities as a local economic advantage; several efforts to site a hazardous waste inciner- A specific type of solidification tested for ator in Oklahoma are underway.) Although Sand Springs was mentioned in the FS for Crys- EPA tried to allay the community’s concerns tal City, but the solidification/landfill alterna- about incineration, ultimately the community tive was not selected at Crystal City. Data pro- preferred the uncertainties of the solidification vided by the vendor, on waste from some other technology and accepted the assurances that site, and reported in the Crystal City FS on two incineration would be used if solidification was contaminants also present at Sand Springs (2- less effective. methylnaphthalene and phenanthrene] showed high levels in the leachate. A demonstration of On this issue of the safety of mobile inciner- the same stabilization technology under EPA ation, Sand Springs can be compared to the auspices concluded: “for the organics, the Davis Liquid Waste site in Rhode Island where leachate concentrations were approximately there also was substantial, documented com- equal for the treated and untreated soils.” (P.R. munity concern about onsite incineration, con- de Percin and S. Sawyer, “SITE Demonstra- cern to which EPA responded with good tech- tion of Hazcon Solidification/Stabilization Proc- nical points but, unlike Sand Springs, did not ess,” paper presented at EPA’s Fourteenth An- alter its choice of incineration. Also, in the ROD nual Research Symposium, May 1988.) for the French Limited site (in the same EPA A recent EPA study found “large losses of region as Sand Springs), EPA defended mobile organics during the mixing process.” (L. Weitz- incineration: “Performance standards for air man et al., “Evaluation of Solidification/Stabili- emissions from incinerators would be met, min- zation As A Best Demonstrated Available Tech- imizing the risk from these emissions. EPA con- nology,” paper presented at EPA’s Fourteenth siders the implementation of an incinerator to Annual Research Symposium, May 1988.) be relatively simple in comparison to the other Another EPA study showed that stabilization alternatives evaluated in the summary. ” was not competitive with thermal and chemi- Moreover, it is not clear that the community cal treatment technologies and soil washing for was totally aware of air pollution problems with organic contamination. (R.C. Thurnau and M.P. solidification. EPA’s responsiveness summary Esposito, “TCLP As A Measure of Treatment said: “Pilot studies have shown that some vola- Effectiveness: Results of TCLP Work Com- tile compounds are driven off during excava- pleted on Different Treatment Technologies for tion and mixing of the waste with the solidify- CERCLA Soils,” paper presented at EPA’s ing agent. Mass emission rates have not been Fourteenth Annual Research Symposium, May quantified.” 1988.) Special comments.—The ROD’s analysis of clean- RIFS contractor. —State led; $1.1 million; John up alternatives said that for any alternative it Mathes & Assoc. will be necessary to pump and treat surface im- State concurrence. —The State of Oklahoma fa- poundment liquids and to discharge them into vored solidification over incineration, the Arkansas” River; no details were given. Moreover, the accepted ARCO proposal made Community acceptance. –EPA judged that the no mention of these needs. community was more in favor of solidification than incineration. The community was very General conclusions.—Sand Springs has some concerned about the future use of an incinera- good points: 1) pilot treatability studies were tor for waste from other sites, the worsening used to evaluate treatment technologies; 2) alter- 64

native treatment remedies were analyzed; 3) leach out of the solidified mass, because the EPA responded to the concerns and interests treated material will be placed into a landfill of responsible parties, the community, and the on a floodplain adjacent to the Arkansas River. State; and 4) some preferred treatment technol- Landfill failure was not considered, nor is its ogy was selected by EPA. location compatible with regulated use of land But Sand Springs has many problems too, in- disposal. Moreover, cause for concern about cluding a lot of confusion about what test data independent testing and verification of solidifi- were used by EPA and when. The ROD selected cation’s effectiveness is driven by ARCO’s position that any form of waste treatment is un- not a reliable permanent remedy but a plan with necessary: “Improved site security and a clay several contingencies, with no assurance of a permanent remedy. Over a short period of time cap would mitigate this [accidental direct con- —about one month—EPA reversed a well-sup- tact] potential risk.” (ARCO, letter to Carl Ed- lund, EPA Region 6, Aug. 31, 1987.) ARCO’s ported, technically sound decision to use incineration rather than solidification/stabilization critique of EPA’s FS of August 31, 1987, said that fencing and a cap “could be a sufficient technology. In the ROD, EPA said: “Solidifi- remedy.” Moreover, in this document ARCO cation was considered in detail during the Fea- sibility Study and actual pilot studies. Adequate also said: “Long term effectiveness of inciner- information is available on which to base a de- ation, stabilization and solidification are comparable.” These views of ARCO suggest that cision.” Was EPA talking about the information available when it signed the ROD, including the selection of solidification, with costs much lower than incineration, was a compromise the ARCO test results, or information obtained made by both EPA and ARCO and that future, by EPA prior to the RIFS public comment post-ROD actions require close EPA scrutiny. period? The ROD suggested that EPA had conducted its own tests on solidification of site ma- ARCO’s effort to get EPA to retrench from terials. But it may have used a very different its original decision to use incineration prob- type of solidification technology. In the ROD ably was helped by its apparently successful responsiveness summary EPA said: “The [ARCO] criticism of the quality of the RIFS. Indeed, a pilot studies had a major influence on the rem- number of ARCO’s comments are consistent edy selected. ” If that was so, then why did the with OTA’s observations in this report for ROD still contain so many negative comments RIFSsin general. For example, ARCO said: 1) about solidification? Part of the answer may “significant gaps exist in the data presented and be revealed in another statement by EPA in the considered in the FS”; 2) “The analysis reflected ROD’s responsiveness summary: “Solidifica- by the FS is cursory and of limited detail”; 3) tion pilot tests [presumably ARCO’s] were only “The lack of back-up, the limited detail and the conducted on the surficial acid sludge waste. lack of references suggested that the analysis Additional waste characterization and pretreat- may not have involved the development of any ment studies will need to be performed on the additional information beyond that provided subsurface petroleum wastes.” Another part of by the authors’ experience”; and 4) “The FS the answer is that EPA had conflicting test data is characterized by an over-reliance on assump- from two different sources on several solidifi- tions rather than actual performance data.” cation technologies. Therefore the question per- sists: Was there enough test data to justify the A big question still remains. Who bears the ROD’s selection of remedy? burden of proof that a treatment technology works before EPA officially endorses its use The ROD contained no details on how EPA at an actual cleanup? There is no basis in the will assure that independent, detailed, and technical literature for concluding that solidifi- timely testing will track progress on the selected cation/stabilization technology is likely to be remedy and detect ineffective performance in effective for wastes with so much and so many the long term, if that occurs. If the treatment different kinds of organic contamination. The technology is ineffective, contaminants will presence of negative laboratory results, which 65

EPA suggested it had prior to the ARCO test After the ROD was issued, a news publica- data, would normally prevent application of a tion reported: “Solidification poses ‘very little cleanup technology at the site in question. It risk whatsoever,’ says an EPA headquarters is true that unusual conditions might justify an source, who is encouraged that the agency is unproven technology; for example, in an emer- willing to allow its use at a Superfund site. The gency situation and where all other treatment technology has not been proven to ‘truly bind technologies are less applicable. But this site organics,’ but any release of organic substances is not generally considered to require emer- would be slight, the source explains. ” (Inside gency attention, and the pilot study on inciner- EPA, Nov. 27, 1987, p. 13.) ation was successful. (Note that stabilization of incinerator residue contaminated with me- Moreover, the solidification technology eval- tals only is proven technology.) It may well be uated by ARCO appears to be one which is in that, as EPA says, the technology proposed by EPA’s SITE technology demonstration pro- ARCO is “a promising innovative technology,” gram to “resolve issues standing in the way of but sanctioning its full-scale application actual full-scale application, ” The demonstra- through a ROD on the basis of limited data ob- tion was conducted October 13-16, 1987, at the tained by the responsible party that conflicts Douglassville Disposal Superfund site in Penn- with data obtained by the government, is a big sylvania, but results of the test were not made step, especially because the data are inconsist- available before the ROD and EPA made no ent with what is generally understood about reference to them for Sand Springs. Results re- the capabilities of the technology. In fact, there cently made available were negative for organic are a number of very different proprietary contaminants. forms of solidification and stabilization (ARCO actually tested two and got similar results), and The way the remedy selection was made il- it is not clear that either ARCO or EPA has con- lustrates what can happen when there is much sidered or evaluated enough of them and their pressure to issue a ROD by the end of the fiscal performance relative to the technologies used year. ARCO submitted its treatability study re- in the ARCO treatability study. sults to EPA on July 15, 1987. As late as Au- gust 21, 1987, results from ARCO’s pilot tests The degree to which the ARCO data support were still being obtained and disseminated EPA’s decision also raises the issue of how ac- among EPA staff. Several formal ARCO reports curately current EPA tests—in this case for haz- are dated August 31, 1987, including one criti- ardous waste treated by stabilization—predict cizing EPA’s RIFS. This site, like a very high long-term environmental effectiveness. Much percentage of all Superfund sites, had its ROD more rigorous testing appears necessary to issued in the last days of the fiscal year (Sep- make the case that stabilization of hazardous tember). In this case, EPA might have stayed organic material, such as at Sand Springs, as- with its original, technically supported decision sures insignificant leaching of organic contami- and kept on schedule or delayed issuing the nants under long-term conditions at the site. ROD while it: 1) designed more tests for ARCO Therefore, although the ARCO test data do look to carry out to convincingly demonstrate, be- good, they are limited by the test procedures fore actual use, the long-term effectiveness of themselves. Nor were the ARCO test data ob- solidification/stabilization technology for the tained by using standard test protocols and diverse wastes at the site; and 2) developed quality assurance procedures to assure the pub- detailed protocols for future testing and moni- lic and the government that the data are relia- toring as well as technical criteria which would ble. Doing this is a major effort and an impor- trigger the switch to incineration, if solidifica- tant characteristic of EPA SITE program and tion failed. In this case, the responsiveness of most treatability testing done by or for the gov- EPA to local pressures seems to be related to ernment. selecting a lower cost technology and facilitat- ing the government’s attempts to have a pri- nebago, in close proximity to the site.” Vari- vate party fund the cleanup. ous types of waste were disposed there from As with the Compass Industries Superfund 1968 to 1979. site directly across the Arkansas River from Major contamination/environmental threat.—Accord- Sand Springs, little attention was given to posing to the ROD, there was “Initial on-site sam- sible groundwater contamination from the re- pling by the State of Wisconsin and the U.S. medial action for the present hazardous mate- Army Corps of Engineers in early 1979 . . . Lead rial and to the risks to downstream water users and chromium were also detected in relatively from cumulative discharges into the river. high concentrations at several sampling sta- tions. Lead and trivalent chromium [the least Case Study 9 toxic form of chromium] were found through- Schmalz Dump Site, Harrison, Wisconsin, out [soils at] the site at concentrations ranging from detection limits to 1940 milligrams per EPA Region 5 kilogram (mg/kg) and 964 mg/kg respectively Capsule OTA findings.-A simple compacted earth . . . . lead and [trivalent] chromium in soils cover over the soil contaminated with lead and . with concentrations greater than 14 mg/kg chromium was selected. Solidification/stabili- and 100 mg/kg respectively, pose an unaccept- zation treatment was rejected, although this able lifetime risk from direct contact.” The ROD was a textbook example of appropriate use of indicated a volume of contaminated material the technology. Voluntary well abandonment of 8,000 cubic yards. and monitoring was chosen over pumping and The second environmental threat was con- treating contaminated groundwater. taminated groundwater. “ . . . the silty sand Key dates: aquifer beneath the site appears to be separated from the lower aquifer by a fairly thick, con- ● Entered Superfund system: 5/1/83 tinuous clay layer. It is therefore unlikely that ● Preliminary Assessment: 5/23/83; (5/1/83 in contaminants from the site would enter the CERCLIS) lower aquifer and reach residential wells. The ● Site Inspection: 5/23/83; (5/1/83 in CERCLIS) shallow aquifer beneath the site contains levels ● National Priorities List of trivalent chromium above background but –proposed date: 9/1/83 not above the MCL [Maximum Contaminant –final date: 9/1/84 Level]. ., . the water table is three to five feet —site rank: #190 out of 770 below the land surface and direction of flow ● RIFS start and completion: 12/28/84 to is , . . towards Lake Winnebago.” Regarding the 8/14/87 ability of lead and chromium to leach from the ● Public comment period before Record of soil, the ROD said: “Results of the tests show Decision: 8/17/87 to 9/8/87 that very low levels of both lead and chromium ● Signing of ROD: 9/30/87 are leachable. . . . trivalent chromium has an ● Estimated complete remediation: 20 months affinity to fine grained, silty soils like those (around 6/89) found in the site area.” A groundwater model- Total time.—6 years ing study “indicates that in fifty years, groundwater containing chromium would have mi- Brief description of site. –The ROD said: “The grated just beyond the site boundary.” Schmalz Dump Site is located on the north shore of Lake Winnebago in the Town of Har- A complex technical issue for this site is that rison, The site occupies approximately 7 acres there are two different chemical forms of chro- in the Waverly Beach Wetlands area. The neigh- mium: the more toxic hexavalent form and the boring city of Appleton, with a population of less toxic trivalent form. The analysis for soil 60,000, has its drinking water intake approxi- contamination recognized that even trivalent mately 1200 feet from the shore of Lake Win- chromium is toxic. However, the analysis of 67 the groundwater threat discounted the hazard tion, the ROD said the selected remedy “would from trivalent chromium or from it turning into provide adequate protection from contami- hexavalent chromium. nated soils on site.” Treatment was rejected for both contaminated groundwater and soil. The HRS scores.-groundwater 27.62; surface water ROD noted for both treatment approaches: 80.00; air 0.00; total 48.92 “Treatability or compatibility testing is re- Removal actions.— Earlier operable unit cleanup quired . . . prior to design and construction.” involved removal of PCB contaminated mate- EPA has said, however: “Toxic metals rep- rials in 1987; Remedial Design, $81,000 and Re- resent a long term threat in the soil environ- medial Action, $2.6 million. ment. This threat can be reduced considera- Cleanup remedy selected. —This ROD is the sec- bly if the heavy metals can be permanently ond one for the site. The first ROD in August immobilized by either chemical or physical 1985 addressed PCB contamination at the site methods.” (U.S. Environmental Protection and selected “removal of construction debris Agency, Review of In-Place Treatment Tech- and sediments containing elevated concentra- niques for Contaminated Surface Soils, vol. 2, tions of PCBs. ” November 1984.) Alternatives examined included: 1) ground- The ROD acknowledged that solidification/ water pumping with treatment, slurry wall and stabilization of excavated soil could be “a per- cap; 2) a RCRA cap; and 3) solidification/stabili- manent remedial action to limit the off-site mo- zation treatment of soil and onsite disposal of bility, volubility [an unusual term instead of vol- residues. The ROD said: “The preferred alter- ume] and toxicity of the heavy metals.” The native involves the installation of a low permea- selected remedy “is expected to significantly bility, compacted earth material cap over ap- reduce the mobility of lead and chromium by proximately seven acres of lead and chromium containment in the site soils, but do nothing contaminated soils, and implementation of to reduce toxicity or volume of contaminants.” groundwater monitoring for lead and chro- The chief reasons for rejecting solidifica- mium. A voluntary well abandonment program tion/stabilization were that: for nearby wells is also proposed.” ● Excavation is risky “due to potential air- Estimated costs for the groundwater treat- borne migration of dusts from the site.” ment alternative were $3.4 million; for solidifi- ● “The reliability ., , is unknown principally cation and/or stabilization, $2.8 million; for a due to the lack of data documenting long- RCRA cap, $2.4 million; and for the selected term success or failure of similar projects.” remedy of a soil cover $800,000. (In the body ● It “ . . . is not expected to significantly min- of the ROD the selected remedy is called a soil imize risks associated with ingestion of cover, but in the beginning of the ROD it is soils without additional restrictions on use called a cap. Soil cover is more accurate be- ✌ of the site (e.g., additional fencing).” " cause a cap implies a more complex, engineered ● . . . there is considerable research data approach to containing hazardous waste.) to suggest that silicates used together with a cement setting agent can stabilize a wide Satisfaction of SARA statutory requirements: range of materials including metals. How- 1) Selection of permanent cleanup.—The ever, the feasibility of using silicates for Schmalz ROD was straightforward: “The stat- any application must be determined on a utory preference for treatment set forth in . . . site specific basis, particularly in view of SARA is not satisfied because treatment was the large number of additives and differ- found to be impracticable due to questionable ent sources of silicates which maybe used, ” technical feasibility, inadequate short-term pro- (Interestingly, the paragraph is verbatim tection, and inappropriate site conditions.” from the book Hazardous Waste Treatment With regard to overall environmental protec- Technologies, by Gerald Rich and Kenneth 68

Cherry, 1987; the statement is a good ex- precautionary measure to ensure that no po- planation of the need for a treatability study tential for exposure exists should contaminant for solidification.) levels in groundwater increase in the future.” ● “Based on the content of soils on the site The ROD’s case is supported by the results . [the alternative] maybe difficult to im- of leachability tests which found low levels of plement. Contaminated soils consist of both lead and chromium; however, the test em- solid waste, wood, brick, and car bodies, ployed does not necessarily describe long-term which would make implementation dif- effects under actual site conditions. The ROD ficult.” also noted that the background groundwater ● its “ . . . reliability. . . is unknown. ” level for chromium is 5 µg/l; therefore, there ● it” . . . is not conducive to a wetlands envi- is contamination from the site. However, the ronment. Capping and vegetation of the ROD also said: “groundwater contamination site is. ” is not above MCLs and there is no leachate With the groundwater problem, there are two release.” subtle issues. First, the ROD emphasized the The ROD for the Tower Chemical Superfund sampling data for chromium concentration in site in Florida, where there is chromium con- the shallow aquifer, which ranged from 14 to tamination in the groundwater, had a target 48 micrograms per liter (µg/l) and the fact that cleanup level for the treatment of 100 million the values are below the MCL of 50 µg/l. How- gallons of water of 0.05 µg/l even though the ever, the highest value is close enough to the groundwater standard is 50 µg/l. The public MCL value to worry about the precision of sam- health threat there is considered minimal. pling and the possibility of future increases in concentration. Indeed, EPA work indicates that The second groundwater issue is a second plume of contamination which the ROD de- chromium can be underreported by more than enough to make the 48 µg/l observation above scribed as “an isolated off-site anomaly west the standard. (K.A. Aleckson et al., “Inorganic of the Schmalz Site.” Two concentrations re- Analytical Methods Performance and Quality ported for the location are 185 µg/l and 1140 Control Considerations,” Quality Control in Re- µg/l of dissolved chromium. The contamination medial Site Investigation, American Society for “appears to emanate from a localized point Testing and Materials, 1986, pp. 112-123.) The source. Based on the history of dumping in the ROD said: “Groundwater was determined not area, this phenomenon is not unusual.” It seems to be a public threat because chromium con- that a narrow definition of the site boundary centrations are below [Safe Drinking Water Act] kept these higher concentrations from being drinking water standards. However, leaching considered a significant factor in selecting the of chromium and/or lead to groundwater could cleanup remedy. Information in the ROD clearly potentially cause drinking water standards to indicates that the second plume would also flow be exceeded. Based on the above discussions into Lake Winnebago. [of small amount of leaching according to stand- Overall, the critical ROD conclusion was: ard tests and affinity to soil], onsite soils are “Based on the rate of groundwater movement, not likely to ever increase chromium and lead and taking into consideration the dilution that concentrations in the ground water to greater would occur once ground water discharges to than the drinking water MCLs of 50 µg/l. How- the Lake, the levels of chromium in the ground- ever, because there is a remote possibility that water should never pose a threat to Appleton’s this pathway could later become a concern, it water supply. ” Groundwater treatment was was determined that groundwater should be also rejected because “several problems can oc- monitored over time. In addition, residents in cur at each component stage. This could result the vicinity of the site will be asked to volun- in delays or inability to implement the alter- tarily abandon any existing wells. This is a native. ” 69

In summing up its comparison of alternatives, The decision not to use a RCRA cap was in- EPA said that groundwater treatment “does not consistent with statutory requirements about protect against direct contact” and that the satisfying current regulations, and it raises sig- solidification/stabilization treatment of the soil nificant uncertainties about future failure, Offi- “would be protective upon implementation, cial EPA guidance notes: “A key task of cover however, there are several problems associated design is the selection of suitable materials. The with implementation of this alternative that cover usually will include a synthetic mem- make it undesirable.” The use of a slurry wall brane and a large volume of soil or soil-like ma- and cap or a RCRA cap are “not cost effective terial . . . “ (U.S. Environmental Protection because they provide excess protection for Agency, “Project Summary-Design, Construc- groundwater.” tion, and Maintenance of Cover Systems for Hazardous Waste: An Engineering Guidance 2) Accurate assessment of land disposal and Document,” November 1987.) The selected containment alternatives.—The word perma- cover is very simple and poses substantial un- nent was not used to describe the overall se- certainty about future risks. In the ROD for the lected remedy. The ROD said that the cap is Pristine site EPA defends its rejection of cap- protective and that the groundwater monitor- ping by saying” ., . there are no data available ing “will provide protection from potential fu- on the long term effectiveness and permanence ture releases. ” of RCRA caps. ” Clearly the situation is worse Regarding permanency of the selected remedy, for a simple soil cover. the ROD said that” . . . the only potential need With regard to groundwater monitoring: for replacement is seen to be that of the cap “Any increase in existing levels of chromium or soil cover. This need could occur if the origi- or lead will be evaluated as to whether correc- nal cap was washed out by some storm event, tive action is necessary based on levels found.” if heavy equipment were to abrade the cover, The ROD did not give any specific technical or if unforeseen subsidence were to occur.” No criteria for deciding when other remedial ac- restrictions on future land use were set, and tions will be necessary. the ROD noted that EPA has legal authority “to issue an order for corrective action, should the There was no discussion of the possible fu- owner make an attempt to damage the cap.” ture oxidation of trivalent chromium to the A letter from the State said, “the cap could be more toxic and mobile hexavalent chromium. damaged by the landowner, who has indicated EPA research has noted: “under conditions a desire to build on the site.” The ROD for the prevalent in many soils, Cr(III) can be oxi- Tower Chemical site in Florida eliminated the dized.” (U.S. Environmental Protection Agency, alternative of using a soil mixture cap because: Review of In-Place Treatment Techniques for “This technology is unproven and has exten- Contaminated Surface Soils, vol. Z, November sive monitoring requirements. Development of 1984,) Such oxidation constitutes a potential dessication cracks could cause failure. High mode of failure for the selected remedy, espe- failure potential” (emphasis added). cially in the context of future land use. Although OTA does not know whether the site soil poses In the FS for the French Limited site (in EPA this problem, it is an important enough issue Region 6), use of a slurry wall and cap to con- to have been examined by EPA. Another po- tain hazardous waste was described as a “tem- tential effect is that trivalent chromium in water porary solution” for which the “volume and can be oxidized to the hexavalent form in cer- toxicity would not be affected . . . [and] . . . the tain types of chlorination treatments; therefore, potential would always exist for failure of ei- leaching of trivalent chromium into ground- ther the cap or the slurry wall allowing for the water that eventually enters a drinking water movement of unstabilized wastes contained supply can be a problem, especially if it is un- onsite. ” expected. 70

RIFS contractor.—Camp Dresser and McKee; The body of the ROD includes a part of the over $600,000 for all of the RIFS work. selected remedy that is not in the remedy’s “The State of Wisconsin description at the beginning of the document. State concurrence.— That is the recommendation that “adjacent supports our preferred alternative, however it property be evaluated under the pre-remedial has several concerns related to implementation program.” Presumably this action refers to the of the remedy. . . . due to the excess cost in- hot spot of contamination just outside the site volved, they do not feel that a groundwater boundaries used by EPA. Such an examination treatment alternative is warranted. The State might, therefore, open up the possibility of has concerns over whether adequate cap pro- another site cleanup, but the ROD only recom- tection is available for alternatives involving mended the evaluation instead of requiring it. capping the site. . . . the State has agreed to attempt to obtain a voluntary agreement from the General conclusions.-The Schmalz site is defi- landowner.” nitely not one of the worst Superfund sites. But Community acceptance. —The ROD said the “com- even though the environmental threat from the munity does not perceive the site as an imme- site is not severe, the handling of the remedial diate danger. . . . at least some residents feel cleanup raises important questions. The degree that a full RCRA Subtitle C Cap should be in- of certainty expressed by EPA for the long-term stalled.” effectiveness of its selected remedy is inconsistent with the technical limits of the remedy. Special comments.—Although the ROD said the A good example of EPA’s over optimism is in selected soil cover had a low permeability, un- the responsiveness summary: “Following im- like some other RODS no specific permeability plementation of the selected remedy, exposure value was given. to contamination from the Schmalz Site will Documents obtained from EPA on the use be eliminated” (emphasis added). This state- of treatability studies show that a study was ment is inconsistent with the technical limita- completed for the Schmalz site in April and tions of a soil cover and with uncertain failures, May 1987; the study focused on “surface water responses to monitoring results, and land use. contaminated with PCBs, Cr [chromium], and Moreover, ROD statements that the “ground- Pb [lead].” This attempt to verify physical and water is not contaminated” are incorrect. The chemical treatment may have been done as part issue is whether the risks estimated by EPA are of the earlier operable unit cleanup for PCBs, correct and stable or whether groundwater but the results are relevant to groundwater cleanup is warranted now or whether it may cleanup for the Schmalz ROD, which does not become necessary. EPA seemed to place heavy mention the study. emphasis on a technicality, namely that con- In the RIFS, the detection limit for hexava- tamination within the bureaucratic boundaries lent chromium was 20 times higher than for of the site was slightly below the current regu- trivalent chromium, which might explain why latory standard for allowable chromium con- hexavalent chromium was not found, if it was tamination in drinking water. What if the meas- present. ured chromium contamination goes up the 5 percent necessary to bring it over the stand- Some of the estimated costs for the rejected ard? Moreover, the ROD’s selected remedy cleanup alternatives look high. For example, omits the statutory requirement of reexamin- the estimate of $2.8 million for stabilization is ing the site at least every 5 years because un- higher than estimates at other Superfund sites. treated hazardous waste will be left onsite. With a figure of $200 per cubic yard for total stabilization costs, consistent with data at the The volume of contaminated soil (8,000 cu- Sand Springs and Liquid Disposal sites where bic yards) at Schmalz is relatively small for a stabilization was selected, the correct value for cleanup site. Nevertheless, various ROD state- Schmalz is probably about $1.6 million. ments indicate that minimizing cleanup costs

I 71 was a significant motivation in the remedy tion, resulting from likely continued leaching selection. Selecting both soil and groundwater and migration of contaminants into Lake Win- treatment was estimated to cost eight times as nebago, even if water intrusion is curtailed much as the soil cover. The ROD said alterna- through the surface of the site. The problem tive treatment remedies for both contaminated is that groundwater movement will still occur soil and groundwater were rejected because beneath the surface where the contaminated “they are more costly while achieving the same soils reside. At the Liquid Disposal Superfund desired results” as the selected remedy. Only site in Michigan, a RCRA hazardous waste cap by using a narrow, short-term objective of limit- and a containment wall will be built around ing exposure to the hazardous waste are there chemically stabilized material, landfilled onsite, comparable environmental results of soil treat- to prevent just this type of leaching and movement and a soil cover. ment of contaminated groundwater. For the soils problem, the rejection of the Moreover, in the Schmalz ROD, no specific solidification/stabilization alternative is incon- criteria were given for groundwater monitor- sistent with generally understood capabilities ing or for triggering a decision that a ground- of the technology. Indeed, having soil contami- water remedy and a better soils remedy are nated with only two toxic metals, lead and chro- needed. This site highlights a problem found mium, offers a textbook example of when chem- by recent EPA research: “many [Superfund] in- ical fixation works best. None of the arguments vestigations are not producing sufficient data given in the ROD against using this soil treat- to adequately characterize ground water con- ment describe especially unique or difficult ditions near these sites.” (R.H. Plumb, Jr., “A problems. For example, the presence of large Comparison of Ground Water Monitoring Data buried objects is faced routinely; they can be From CERCLA and RCRA Sites,” Ground Water washed and reburied. Regarding implementa- Monitoring Research, fall 1987, pp. 94-100.) tion risks, wetting excavated materials for dust suppression is routine and, in exceptional Case Study 10 cases, inflatable domes have been used. The technology has been selected for Superfund Tacoma Tar Pits, Tacoma, Washington, sites posing much more challenging kinds and EPA Region 10 levels of contamination for which solidification Capsule OTA findings:–No treatability study re- is unproven. sults supported the selection of chemical stabili- The argument that stabilization technology zation. Significant amounts of untreated con- needs to be verified for a site is an argument taminants as well as the treated materials will for conducting a treatability study, preferably be left onsite. The effectiveness of the treatment during the RIFS, not for rejecting the alterna- is uncertain. Incineration was said to offer no tive. If, as the ROD acknowledged to be the case, better protection and was rejected because of the soil is contaminated enough to pose a true its higher cost. environmental risk, then the selected remedy Key dates: of a soil cover is not permanently effective. The absence of land use restrictions is particularly ● Entered Superfund system: 2/1/82 worrisome. ● Preliminary Assessment: 4/1/82 ● Site Inspection: 3/1/83 The rejection of groundwater treatment does ● National Priorities List not consider several factors: 1) values for chro- –proposed date: 10/81 mium contamination very close to the MCL; –final date: 9/83 2) the hot spot of high chromium contamina- —site rank: #347 out of 770 tion apparently just outside the site boundary; ● RIFS start and completion: 11/84 to 9/87 and 3) the exact environmental and health ef- ● public comment period before Record of fects, which might not be eliminated by dilu- Decision: 11/6/87 to 12/6/87 72

● Signing of ROD: 12/30/87 of tar is 5,000 cubic yards; it is mostly in three ● Estimated complete remediation: Assume areas at depths of several feet and more. PCBs 2 years after ROD are found up to 204 parts per million (ppm), Total time.—7 years and lead in soil in the 2,000 to 8,000 ppm range. Three shallow aquifers have varying degrees Brief description of site.-The site is approximately of contamination, and the ROD noted the “cur- 30 acres and “within a heavily industrialized rent lack of understanding of local groundwater area . , . [the] site is part of the Commencement hydrology.” ● Bay-Nearshore/Tideflats Superfund site located within the Tacoma Tideflats industrial area Four indicator compounds were used to esti- near Commencement Bay.” (The ROD does not mate risks and establish cleanup goals: ben- describe this action as an operable unit, but that zo(a)pyrene, PCBs, benzene, and lead. For the is what it appears to be. However, Tacoma Tar most part, the cleanup levels are consistent with Pits is listed as a separate site in CERCLIS, three a 1 in 1 million cancer risk level, except for lead years after the large Commencement Bay site in soil which seems high relative to the MCL entered the system.) “A coal gasification plant value. The major emphasis was correctly put on was in operation on site from 1924 through exposure of onsite workers over short periods. 1956. The study area currently contains a metal HRS scores.—groundwater 6.12; surface water recycling facility . . . a natural gas transfer sta- 10.91; air 71.92; total 42.20 tion . . . a rail freight loading yard . . . a meat Removal actions.—None indicated. packing plant . . . and a railroad switching yard . . . “ Cleanup remedy selected. –This ROD apparently Major contamination/environmental threat.—"The site is for an operable unit of the larger Superfund currently contains two ponds, a small tar pit, site of which it is a part, even though it did not and various surface-water drainage ditches. use the term. Besides the selected remedy, The study area is located near several major cleanup alternatives examined included containment and landfilling, incineration, in situ surface water bodies [waterways]. . . the Puyal- lup River, and Commencement Bay. Although vitrification, and groundwater treatment. none of these water bodies are used for water “The preferred remedial alternative . . . is a supply, the bay and river do support extensive combination of source control measures, meas- fish and shellfish populations. Several portions ures to control contaminant release, and also of Commencement Bay have been identified measures to reduce human exposure to con- as being severely contaminated, resulting in ad- taminants. This alternative consists of the ex- verse biological effects . . . contamination of the cavation of the most severely contaminated local groundwater resource is also of concern. soils, stabilization of these soils using a tech- Many local industries use groundwater from nique which immobilizes contaminants, capon-site wells . . . “ ping of the stabilized material [with asphalt], With regard to site contaminants, first found treatment of the surface water, continued groundwater monitoring, regulatory controls on water in 1981: “Many of these organic compounds are toxic and several are considered to be car- usage for both surface and groundwater, and restrictions on site access. ” Thoroughly mixed cinogenic. These compounds include aromatic hydrocarbons (i.e., benzene, toluene), poly- excavated materials will be “fed to a mixing vessel where silicate polymers, cement, and nuclear aromatic hydrocarbons collectively known as PAHs (i.e., napthalene, benzo(a)py- water from the site ponds is added.” rene), as well as numerous other classes of However, in site areas that are not severely hydrocarbons and cyanide. Heavy metals . . . contaminated with PAHs, soils and sediments include arsenic, mercury, and lead.” The au- “will be excavated to a depth not to exceed 3 tomobile recycling facility has also caused lead feet.” This requirement means that significant and PCB contamination. The estimated volume amounts of contaminants may not be excavated —

73 and treated. Exactly what amount will not be essary to evaluate and implement an alterna- excavated cannot be judged from the informa- tive remedial action that includes groundwater tion given in the ROD, but the site’s history and extraction. A commitment to groundwater complex contamination suggests that this may monitoring was made, but given the acknowl- bean important limitation to the selected rem- edged complexity of local hydrogeology and edy. Many of the contaminants have been given EPA’s interest in minimizing cost, it ne- present long enough to have migrated down- cessitates a major, carefully planned effort. ward a significant distance. “The total esti- With regard to the selected stabilization treat- mated volume of material to be excavated is ment technology: “No bench or pilot studies 45,000 cubic yards.” The total cost for the have been performed to date, these being left selected remedy is $3.4 million. This cost im- until the Remedial Design is commenced . . . “ plies a rather low cost for the stabilization part The diverse and highly concentrated contami- of the cleanup of about $50 per cubic yard. nants pose a major challenge for a chemical Satisfaction of SARA statutory requirements: stabilization technology. The current state of knowledge and experience does not support an 1) Selection of permanent cleanup.—The assumption of effectiveness for the Tacoma Tar ROD said: “This remedy satisfies the prefer- Pits site. The ROD noted: “Laboratory experience expressed in SARA for treatment that re- ments will be performed to ensure that the duces toxicity, mobility, and volume. . . . it is stabilization process effectively immobilizes determined that this remedy utilizes permanent contaminants. Following this activity, a larger solutions and alternative treatment technol- scale ‘pilot study’ will stabilize a larger volume ogies to the maximum extent practicable.” of contaminated material from the site. This The uncertainty about contaminants left un- pilot study will determine the effectiveness of treated onsite means the selected remedy may the stabilization process.” The uncertainty not be permanent. The intent to comply with about effectiveness means that the selected SARA’s requirement for 5-year review confirms remedy did not merit the high (maximum pos- that the selected remedy is not assuredly per- sible) rankings it received for effectiveness in manent. This situation raises questions about the ROD’s analysis of cleanup alternatives. future land use. Further uncertainty about per- Moreover, there is no basis for saying: “The manence is indicated by the ROD’s comment: chemical stabilization process should signifi- “If as a result of this frequent reassessment, cantly reduce the toxicity and leachability of the remedial action is shown to have decreased site soils,” The ability of any chemical stabili- performance, the nature and extent of addi- zation technology to reduce toxicity of a wide tional actions will be considered.” range of organic and inorganic contaminants has not been proven nor is it generally accepted An important issue for Tacoma Tar Pits con- in the technical community, But it is reason- cerns the cleanup of contaminated ground- able to claim that the mobility of the contami- water. Although the ROD contained cleanup nants might be reduced through stabilization; standards for groundwater, “the remedial ac- to claim more than that would require actual tion does not currently provide for groundwater test data on site materials, Elsewhere the ROD extraction and treatment.” The basis for this said: “Permanent treatment can be provided lack of cleanup was that the selected remedy through the immobilization of contaminants.” will reduce surface water intrusion and con- This statement is, however, overly emphatic at taminant flow into the water, that existing con- the current stage of knowledge about the site. tamination will be swept away and into its ultimate discharge, and that “Action levels of An important statement on the selected contaminants in groundwater have not been stabilization technology was in the responsive- consistently exceeded at off-site locations. ” A ness summary, given in response to a concern claim of permanence at this time is premature about its effectiveness: “Although the cement/ because the ROD said that it may become nec- polymer stabilization process is a proven tech- 74 nique for immobilization of heavy metals, this chemical treatment or other. ” (U.S. Environ- technique has not been conclusively proven to mental Protection Agency, “RCRA/CERCLA be effective in immobilizing organic contami- Treatment Alternatives for Hazardous Wastes,” nants in coal tars, Therefore, both laboratory October 1987.) and bench scale treatability studies will be performed during the design phase of the re- A recent EPA study found “large losses of medial action to ensure the process will be ef- organics during the mixing process.” (L. Weitz- fective and permanent. . . . the soils/tars con- man et al., “Evaluation of Solidification/Stabili- taining the highest tar content . . . may be zation As A Best Demonstrated Available Tech- considered for an alternate type of treatment/ nology,” paper presented at EPA’s Fourteenth disposal (i.e., incineration) if the stabilization Annual Research Symposium, May 1988.] process is found to be ineffective for the waste Another EPA study showed that stabilization matrix” (emphasis added). Also: “Criteria to be was not competitive with thermal and chemi- used to evaluate the effectiveness of the stabili- cal treatment technologies and soil washing for zation process during laboratory and bench organic contamination. (R.C. Thurnau and M.P. scale studies . . . “ will be addressed in the de- Esposito, “TCLP As A Measure of Treatment sign phase. Effectiveness: Results of TCLP Work Com- pleted on Different Treatment Technologies for The FS for the Re-Solve site in Massachusetts CERCLA Soils,” paper presented at EPA’s rejected stabilization for organic contamination Fourteenth Annual Research Symposium, May because “there has been limited success in 1988.) A demonstration of a stabilization tech- chemically fixing organic contaminants such nology under EPA auspices concluded that “for as solvents and PCBs,” The ROD for the Liq- the organics, the leachate concentrations were uid Disposal site in Michigan, which also approximately equal for the treated and un- selected stabilization for soil contaminated with treated soils.” (P.R. de Percin and S. Sawyer, organic chemicals, said that the hazardous sub- “SITE Demonstration of Hazcon Solidifica- stances “will not be permanently destroyed.” tion/Stabilization Process,” paper presented at And the FS for Liquid Disposal said: “Consid- EPA’s Fourteenth Annual Research Sympo- erable research data exists demonstrating the sium, May 1988.) effectiveness of this technology in immobiliz- The rejection of other treatment technologies ing a wide range of contaminants, primarily for Tacoma Tar Pits did not have much techni- inorganic. A substantial amount of data does cal analysis behind it. The analysis in the ROD not exist, however, to accurately judge the long- rests mainly on a very simple rating system. term reliability of the process.” For example, all 10 alternatives, including cap- The ROD for the Tower Chemical Superfund ping the waste and incinerating it, received the site in Florida said the following about chemi- same rating of high for technical feasibility (in- cal stabilization processes, which it rejected: cluding effectiveness, useful life, operation and “This technology would also require long-term maintenance requirements, possible failure (30 years) monitoring which is less favorable modes, constructability, implementation time, than technologies which provide permanent de- worker safety, and neighborhood safety). All struction of wastes.” The inference is that the but the two no action or nearly no action op- technology does not provide permanent de- tions received the same high rating for public struction of wastes. health impacts (including minimization of chemical releases, exposures during remedial Another EPA document, used to teach peo- action, and exposures after remedial action). ple about waste treatment said: “Solidification But sound technical bases exist for finer dis- technologies are designed to be used for final tinctions among such abroad range of alterna- waste treatment. This means the technology tives. For example, the incineration options should be applied only after other treatment offered substantially greater effectiveness, relia- techniques have been applied, i.e., incineration, bility, permanency, and certainty of destruction of toxic substances than capping the waste tions resulted from the two earlier RIFSs and or the selected remedy. no completion dates for them are indicated; $2.5 million was spent by the government on the The estimated costs of the alternatives prob- first one in 1982 and 1983. ably weighed heavily: two incineration options (including stabilization of residue) had total State concurrence.—” The State of Washington costs of $17 million (only surface soils) and $243 has been consulted and has verbally concurred million (all soil with contamination with a risk with the selected remedy.” Verbal concurrence greater than 1 in 1 million cancer risk); both may indicate a rushed ROD at the end of the options included groundwater pumping and fiscal year quarter. treatment. In comparison, the selected remedy (stabilization) would cost $8 million if ground- Community acceptance.—The ROD noted that water treatment is included in the calculation. community interest “has not been actively dem- The cost will be only $3.4 million, however, be- onstrated.” The reasons given for the lack of cause groundwater treatment was excluded. community interest are the site’s location within the larger Commencement Bay Superfund site, 2) Accurate assessment of land disposal and the lack of private residences nearby, and a containment alternatives.—There is a statement number of cleanup actions already taken in the on future land use: “Land use restrictions will area. be imposed to prevent or require stringent control of future excavation on the site, to prevent Special comments.—Although there are state- future use of surface water and shallow ground- ments about restricting future land use, there water, and to prevent site access by personnel are also statements that suggest that those re- other than site workers. ” However, there is no strictions may be applied only in the short term, detailed analysis of possible future failures of For example: “The [stabilization] reagent com- the landfill in which the contaminated materi- position is formulated to provide a high-strength als will be re-buried after stabilization. surface capable of supporting trucks and other vehicles. ” The ROD for the Tower Chemical site made a good point about concrete or asphalt caps (as Cleanup goals were set for indicator contami- selected for Tacoma Tar Pits): “The risk of fail- nants. While these goals make risk assessment ure . . . is high due to the potential for fracture more manageable, there can be problems with formation.” The FS for the Pristine site says: using them for analysis of the effectiveness of “Asphalt is photosensitive, and subject to crack- a cleanup technology that is chemical specific, ing due to settling, chemical action, and vege- such as chemical stabilization. Moreover, the tation. Frequent inspections are required to en- cleanup very much depends on data that re- sure cap integrity. ” The asphalt cap option was veal areas of high tar concentration, where rejected at both of these sites. there is no excavation depth limit. However, there may be other areas that have high con- RIFS contractor. –The studies were paid for by centrations of other contaminants and that may responsible parties. The ROD noted that, al- be either overlooked or fall under the provision though EPA and the State found the documents of the excavation depth limit. Compounding the acceptable, “EPA has prepared an addendum problem is the relatively small amount of soil for each document addressing issues that the sampling that has been reported, averaging only studies have inadequately or incompletely ad- about 1.5 locations per acre. dressed.” Geotechnology, Inc., performed the RI; Envirosphere Company (Ebasco) performed The administrative record indicated a large the FS. For the entire Commencement Bay/ number of contractors have performed studies Nearshor/Tideflats site, the SCAP indicates on the site. It is not clear whether EPA had in- three different RIFSs with the last one labeled dependent work done to verify work done for Tar Pits started on 9/23/83 and then taken over the responsible parties. There is also some con- by the PRPs on 11/1/84. It is not clear what ac- fusion about the relationship between Tacoma 76

Tar Pits and the larger Commencement Bay stabilization and incineration; therefore, the site because the former is listed separately in claim that the selected remedy is cost-effective CERCLIS and the ROD does not use the term is unsupported. operable unit. However, incineration for a comparable vol- General conclusions.—The technical information ume of contaminated soil definitely would be obtained prior to and used in the ROD did not much more expensive ($242 million) and, there- support claims that the selected remedy is fore, this site, like many others, shows how permanent nor even that it will be effective. De- important it is to examine the issue of compara- laying testing of the chosen stabilization tech- ble environmental protection for a cost-effec- nology and setting of criteria for its effective- tiveness decision. Moreover, Tacoma Tar Pits . ness until after the ROD undercuts the claim illustrates the need to consider a broader range that a permanent remedy has already been of treatment technologies to reduce cleanup selected. The scope and depth of analysis of costs. Biological treatment for such a site alternative cleanup technologies was less than deserves attention. The case for its considera- seen in any other ROD examined in this report. tion at Tacoma Tar Pits, for example, was as This shortcoming directly affected technology good, if not better, than for the Renora site selection; it maybe related to the strong involve- where it was chosen without treatability test ment of the responsible parties, particularly in data to support the decision. A research pro- conducting the RIFS. (An experienced attor- gram on developing biotechnology for clean- ney advises responsible parties: “Participation ing up old manufactured gas plant sites such in the IRIFS] study provides an opportunity to as Tacoma Tar Pits is underway at the Univer- generate information that can sway EPA deci- sity of Tennessee with support from the Gas sion makers on important issues. We all know Research Institute, The choice at Tacoma Tar that one can interpret the same data a number Pits could have included postponing the reme- of ways. Differing conclusions can be made and dial action or conducting treatability studies supported from the same data. A company that for biological treatment. uses its experts to argue convincingly in favor Moreover, the cost of the selected remedy of one conclusion often can influence the ulti- may have been significantly under estimated. mate decision.” [P. H. Hailer, Hazardous Ma- Data from a vendor of the stabilization tech- terials, January-February 1988.]) nology most likely to be effective on this site While the ROD’s interest in alternative treat- suggests a cost of about $150 per cubic yard ment technology is commendable, the chief instead of the $50 indicated in the ROD. If major costs for treatability studies are added in, driving force for selecting the remedy appears the cost of the selected remedy could be about to be cost: “The final selected remedy meets $5 million more than the $3.4 million estimated the requirement of cost-effectiveness as this in the ROD for a total of $8.4 million. The higher alternative provides for permanent treatment, and contaminant release minimization for a cost matches the low range of the ROD’s inciner- cost significantly less than other alternatives ation options without groundwater treatment. exhibiting a similar level of protection. Addi- The lack of a commitment to groundwater tional cost of these [other alternatives] is the cleanup may be linked to the location of the result of the use of more costly technologies site in a highly contaminated area, including such as incineration . . , or the excavation of contaminated major water bodies. This is a sit- larger volumes of soils coupled with off-site uation where analyzing the cleanup in isola- landfilling.” But no data support the conten- tion may be misleading and ultimately in- tion of similar or equal levels of protection for efficient.