Re: EPA Action Memorandum for Removal Action at Illinois Central Springs

November 10, 1998

Moreover, as discussed in greater detail throughout this letter, EPA's decision to construct and operate a water treatment system at ICS is based on erroneous and incomplete facts, is arbitrary and capricious, and is not in accordance with law. Specifically, EPA's conclusion that the PCBs in the water at ICS create an imminent and substantial endangerment to human health, welfare or the environment is not supported by the facts. Although EPA claims serious risks to humans eating fish contaminated with PCBs from the ICS, it is unable to document this route of exposure. EPA only presents limited evidence of fishing in Clear Creek and no evidence of human consumption of fish. Moreover, although EPA bases much of its concern about human consumption of fish on the fact that the State of Indiana has imposed a Level 5 Fish Advisory on all species caught in Clear Creek, the most recent fish data, including 1997 data obtained by EPA, show that the Level 5 Fish Advisories are not warranted for game fish species.

Also, EPA's decision is arbitrary and capricious in rejecting CBS's proposed treatment system in favor of a more costly alternative which will take much longer to implement. Despite EPA's assertion that it is deciding to undertake a time-critical removal action, the Agency has chosen a more expensive and time-consuming approach over a more practical one that could be implemented much more quickly.

Further, in reaching its decision, EPA failed to follow the procedural requirements of the Administrative Procedure Act ("APA"), the Comprehensive Environmental Response Compensation and Liability Act ("CERCLA"), and the National Contingency Plan ("NCP"), 40 C.F.R. Part 300 and denied CBS due process of law. In particular, EPA failed to include relevant materials in its administrative record, failed to consider information which undermines its decision, and failed to solicit public comment, not even seeking comment from CBS, the party whom it has threatened to sue for cost recovery.

Finally, the Action Memorandum provides an incomplete and erroneous record of the negotiations among the Consent Decree parties over an interim system. EPA states that negotiations stopped when an impasse was reached at the time of the August 14, 1998 status conference before Special Master Foster, and that CBS failed to propose an acceptable system. Actually, at the close of that status conference, CBS representatives informed the government that CBS was prepared to continue negotiations. The EPA representatives indicated they did not have authority at that time to continue negotiations, and scheduled a meeting for the following week when they expected to have authority to continue the discussions. Subsequently, EPA cancelled the meeting and announced that it would build its own treatment system. Thus, it was EPA's unilateral action that brought discussions to a close.

Accordingly, CBS submits this letter to protest EPA's unlawful, ill-considered and ultra hires action, to identify the substantive and procedural flaws in EPA's decision- making, and preserve its defenses against the Agency's threatened cost recovery action.

1. EPA's Cost Recovery Claim is Barred by the Bloomington Consent Decree

At the outset, CBS notes that the United States has already agreed in the Consent Decree not to sue CBS for response costs related to water treatment at Illinois Central Springs. The statement made by EPA on page 3 of the Action Memorandum that "Illinois Central Springs was not included in the settlement" contained in the 1985 Consent Decree between the United States and CBS Corporation (under its former name of Westinghouse Electric Corporation) is simply wrong. Although the parties to the Consent Decree agreed not to treat PCB-contaminated Groundwater emerging from ICS, the Covenant Not to Sue in the Consent Decree clearly bars the United States from pursuing environmental claims relating to ICS against CBS.

At the time, the Consent Decree was entered by the Court, the parties were well aware that groundwater contaminated with PCBs from Lemon Lane Landfill was flowing through karst conduits to emerge at Illinois Central Springs. The parties agreed in the Consent Decree, however, not to treat this contaminated ground water, and the governmental parties agreed not to sue CBS for further relief with respect to ICS. Indeed, in one of the public comments submitted to the Department of Justice with respect to the Consent Decree, an individual named Ron Smith criticized the settlement for not addressing PCB contamination at ICS:

A contaminated spring about 1500 feet southeast of Lemon Lane is gushing high levels of PCB. . . This indicates that a very large amount of material has already traveled from the upper portion of the site through sinkholes and dissolution cavities present in the fractured karst topography surrounding this site. If this aquifer is not treated immediately, the contamination will remain, and spread . . .

Letter of Ron Smith to Assistant Attorney General, June 24, 1985.2 In the "Response of the United States to Public Comments on the Consent Decree and Request to Enter Consent Decree," filed with the Court, the United States acknowledged the presence of PCB contamination in groundwater near Lemon Lane and other sites, but responded to criticisms by Mr. Smith and others by stating that "[g]iven the infeasibility oftraditional groundwater remedial measures, the proposed settlement adequately addresses potential groundwater problems at the sites." Response at p. 30. The groundwater measures in the Consent Decree consisted of a groundwater monitoring program within a 5000-foot radius around the sites, including Lemon Lane Landfill, a residential well survey, and a commitment by Westinghouse (now CBS) to provide alternative drinking water supplies, if drinking water wells within 5000 feet of a site become contaminated with PCBs above detection. See Consent Decree pp69-82

The Covenant Not to Sue in Paragraph 111(a) of the Consent Decree expressly protected CBS from any environmental claims by the United States relating to PCB-contaminated Groundwater from Lemon Lane Landfill The Covenant Not to Sue states that it bars any claim under an environmental law "resulting from or relating to:" (1) "[t]he past disposal or discharge of PCBs or materials contaminated with PCBs at . . . the sites and areas specified" in certain paragraphs of the decree; and (2) "[t]he release or threatened release of PCBs or material contaminated with PCBs from" those areas.

Under subparagraph 111(a)(l), the Covenant Not to Sue bars any environmental claim that "results from or relates to" the disposal or discharge of PCBs in the designated areas, including Lemon Lane Landfill. In the Action Memorandum, at pp. 2-3, EPA asserts that "Lemon Lane Landfill is the source of PCBs" at ICS, and that PCB- contaminated water flows from ICS to Clear Creek. Thus, EPA has alleged that the contamination at ICS and Clear Creek "results from" and "relates to" the disposal of PCBs at Lemon Lane Landfill.³ Accordingly, claims relating to addressing this contamination are barred by the Covenant Not to Sue.

Also, under subparagraph 11 I(a)(2), the Covenant Not to Sue precludes any claim under an environmental law that "results from or relates to: . . . (2) [t]he release or threatened release of PCBs or materials contaminated with PCBs from" the specified sites and streambeds, including Lemon Lane Landfill. This language expressly covers PCBs that have been released into the Groundwater and migrated from the sites where they had been originally disposed and released.⁴ Accordingly, the Covenant Not to Sue bars claims relating to PCBs and PCB-materials that have been released into the Groundwater at Lemon Lane Landfill and have migrated to ICS and Clear Creek.

Finally, paragraphs 78 and 82 of the Consent Decree further demonstrate that claims relating to the ICS were resolved. In paragraph 78, Westinghouse (now CBS) agreed to provide an alternative drinking water supply if PCBs were detected in drinking wells within 5000 feet of Lemon Lane Landfill. In paragraph 82, the United States and other governmental parties agreed that "provision by Westinghouse of a permanent alternative potable water supply source in compliance with [paragraph 78] shall be the exclusive civil or administrative remedy available to the United States . . . against Westinghouse to remedy the problem of PCB-contaminated Groundwater . . . within the 5000-foot radius of the boundaries of . . . Lemon Lane Landfill." (Emphasis supplied). As EPA's Action Memorandum acknowledges, ICS is located about 2000 feet from Lemon Lane Landfill, well within the 5000-foot radius.⁵

Collectively, these provisions demonstrate that, at the time the Consent Decree was entered, the United States was aware that PCB-contaminated Groundwater emerged from ICS, but agreed not to bring future environmental claims against CBS with respect to that situation.

II. Procedural Errors

The Action Memorandum purports to document an EPA decision to undertake a removal action. It fails to meet the procedural requirements of the APA, CERCLA, and the NCP for such decisions in three significant respects: (1) EPA failed to solicit public comments on its action, particularly comments from CBS; (2) EPA reached its decision before receiving much of the information which it now cites to support its decision; and (3) the administrative record excludes substantial relevant information available to EPA which conflicts with the decision reached.

A. EPA Failed to Solicit Public Comment. Including Comment from CBS

In United States v. Seymour Recycling Corp., 679 F. Supp. 859 (S.D. Ind. 1987), the U.S. District Court for the Southern District of Indiana held that the Due Process Clause ofthe U.S. Constitution required EPA to allow a PRP to submit written comments on a proposed response action "at a meaningful time, in a meaningful manner." Id. at 864, quoting Armstrong v. Manzo, 380 U.S. 545, 552 (1965). The court further held that "[t]he most meaningful time to comment on selection of a remedy is before the decision is made." 679 F.Supp. at 864. Similarly, the court in United States v. Rohm & Haas Co., 669 F. Supp. 672, 683 (D.N.J. 1987) held that where EPA only allowed PRPs five days to submit comments on a proposed action, it Wiled to provide a meaningful opportunity to comment in violation of their rights to due process.

In issuing its Action Memorandum, EPA provided no opportunity for CBS to submit comments at all. Indeed, EPA reached a decision summarily on August 18, 1998, and then issued a supporting Action Memorandum 43 days later on September 30, 1998. But at no time - either before its summary decision or its Action Memorandum - did EPA allow for public comments. This is an egregious procedural error depriving CBS of the opportunity to persuade EPA to take a different course of action, and the opportunity to create a record of its position to defend against a future cost recovery action.

EPA apparently is trying to justify its failure to solicit public comment by describing its removal action as "time critical." But EPA includes no explanation in the Action Memorandum why this action should be considered so time critical that public comment is not allowed. Therefore, EPA's decision to treat this removal action as time- critical is itself arbitrary and capricious, and not justified by the administrative record.

Even if EPA had attempted to justify its characterization of this action as "time- critical," the facts do not support that position. EPA has known about PCBs in the water at ICS for more than 13 years, and similarly, has known about the Fish Advisories in Clear Creek since at least the 1980's. Indeed, as discussed in greater detail below, the August 19, 1997 fish data upon which EPA principally relies in its risk analysis actually shows decreases over the PCB levels found in samples taken in 1996. Moreover, these 1997 levels are well below the 1900 ppb standard that the State of Indiana uses for setting Level 5 Fish Advisories. In other words, had EPA submitted this data to the State of Indiana (which it did not), this data may have provided support for removing the Level 5 Fish Advisories which so alarm EPA.

Moreover, EPA does not plan to have its ICS water treatment system operational until the summer of 1999 at the earliest.6 Thus, EPA is not acting at a time critical pace. Indeed, had EPA intended to act as quickly as possible, it would have approved CBS's proposal for a system that could be installed within a couple of months and capture about 55% of the estimated annual PCB mass emerging through the ICS, rather than waiting almost a year to install its chosen system.

Finally, more than 30 days elapsed between EPA's August 21, 1998 announcement to the U.S. District Court in Indianapolis that it intended to construct its own treatment system and its September 30, 1998 Action Memorandum. Therefore, EPA had time to collect public comments on its proposal. Accordingly, there is nothing time- critical about EPA's decision. Under the circumstances, EPA cannot rely on a characterization that the selection of a treatment system at ICS is "time critical" to deprive CBS of its right to submit comments.

B. EPA Improperly Relied on Post-Hoc Rationalizations to Support a Previously Made Decision.

By its own admission, EPA had made its decision to build the interim treatment plant by at least as early as August 18, 1998. In the United States' Status Report of August 21, 1998 to the U.S. District Court, at p. 5, the federal government specifically announced that "[o]n August 18, 1998, the United States informed the other parties to the lawsuit that the United States would resolve the issue of interim water treatment by constructing and funding the initial operation of the interim water treatment system." Significantly, EPA did not say that it was proposing or considering the construction of such a system, but that it had already decided to build one.

Nonetheless, the September 30, 1998 Action Memorandum included in the administrative record several documents which were only created after EPA had made its decision on August 18, 1998. These documents include Administrative Record items nos. 10- 15, many of which were relied upon for the decision in the Action Memorandum. For example, items nos. 10 and 13 were relied upon to support EPA's (erroneous) conclusion that fishing for food occurs in Clear Creek.' As discussed below, under traditional rules of administrative law, an agency is not allowed to use post hoc rationalizations to support a previously made decision. Federal Power Comm. v. Texaco, 417 U.S. 380, 397 (1974), Burlington Truck Lines. Inc. v. United States, 371 U.S. 156, 168-69 (1962); SEC v. Chenerv Corp., 332 U.S. 194, 196 (1947). In its September 30, 1998 Action Memorandum, EPA has attempted to justify its August 18, 1998 decision, by using information that it did not have at the time the decision was made. This action contravenes the accepted norms of administrative decision-making. In Corrosion ProofFittings v. EPA 947 F.2d 1201, 1212 (5th Cir. 1991), the Fifth Circuit overturned an EPA decision because it relied on a methodology and information that had not been disclosed to the public, and therefore had not been subject to comment. See Aqua Slide 'N' Dive v. CPSC, 569 F.2d 831, 842-43 (5th Cir. 1978). Here, EPA has not only deprived CBS of the opportunity to comment on this data, but more significantly, EPA has relied on this data to justify a decision that it actually had made weeks before it knew this information.

C. EPA has failed to Include Relevant Documents and Information in the Administrative Record

Pursuant to 40 C.F.R. .. 300.810, 300.820, EPA is required to compile a complete administrative record for a decision selecting a removal action. In connection with its decision to install a water treatment system at ICS, EPA has failed to include in the administrative record many documents in its possession that are substantially related to the decision. Many of the documents omitted from the record provide data and analysis which undermine the rationale for EPA's decision.

EPA's chief omission is its Failure to include the Agency for Toxic Substances and Disease Registry's ("ATSDR") health assessment of the Bloomington sites (including Lemon Lane Landfill and Illinois Central Springs). Under 40 C.F.R. . 300.810(a)(1), EPA is expressly required to include in an administrative record supporting a removal action all "ATSDR health assessments." In this case, EPA's omission is particularly noteworthy because, as discussed in greater detail below, the ATSDR health assessment for this site expressly contradicts the basis for EPA's finding of an imminent and substantial endangerment. Thus, in making its decision,, EPA deliberately ignored the findings of its sister agency, which is directly tasked under CERCLA with conducting assessments of "the potential risk to human health posed by individual sites and facilities." See 42 U.S.C. . 9604(i).

Moreover, EPA omitted from the administrative record, numerous documents and other information about the PCB-contaminated water at ICS provided to it by CBS and others. These documents include correspondence between CBS and EPA throughout 1998 concerning water treatment, overheads and other demonstrative documents used by CBS at meetings with EPA (including sets of overheads containing annotations which EPA requested CBS add), and CBS's own Status Report to the U.S. District Court, submitted on September 10, 1998.

The omission of these documents is critical because they describe the technical and legal issues ignored by EPA in reaching its decision. These documents also describe CBS' position about the need for treatment at ICS and the selection made by EPIC By both omitting these documents from the record, and not providing CBS with an opportunity for public comment, EPA has attempted unfairly to exclude all information supporting CBS's position from the administrative record and has selectively included only documentation that supports its own conclusion.

III. EPA's Finding of an Imminent and Substantial Endangerment is Arbitrary and Capricious. and based on Incomplete and Erroneous Evidence

EPA's conclusion that PCB-contaminated water at ICS presents "an imminent and substantial endangerment" is arbitrary and capricious, and is based on incomplete and erroneous evidence. Indeed, in some instances, EPA's conclusions are based on no data at all.

In the Action Memorandum, EPA states that "[e]stimates using data submitted by CBS show that in 1996, approximately 41.23 pounds of PCBs were released at ICS" and that discharges during a storm period from April 15 to April 19, 1998 resulted in "approximately 6.40 pounds of PCBs being released into Clear Creek." Not only are both estimates wrong, they are more than double the true values for these periods. Using the same data, CBS has calculated that the mass of PCBs released in 1996 - a comparatively wet year- was 15.4 pounds, less than half the amount estimated by EPA With respect to the April 15 to 19, 1998 period, CBS estimated the mass of PCBs emerging from the spring to be 1.8 pounds, less than a third of the amount estimated by EPA. CBS does not know how EPA made its errors. CBS has reviewed administrative record item no. 11, containing the estimates of PCB mass by EPA's contractor, Tetra Tech. This document does not provide sufficient information about how Tetra Tech made its calculations to allow CBS to determine how Tetra Tech made its errors. CBS would be willing to meet with EPA and Tetra Tech to review its calculations.

Although EPA concludes that PCBs at ICS "pose a threat via inhalation, ingestion, and direct contact," the Action Memorandum contains no information demonstrating human exposure through inhalation, ingestion of water, or direct contact with PCBs. Indeed, EPA's Dr. Milton Clark has stated in meetings with CBS and the City of Bloomington that ICS presents no risk to humans through these routes of exposure. Water from the ICS is not suitable for drinking for reasons having nothing to do with PCBs, and EPA has no information that any one actually drinks it. Therefore, this route of exposure creates no risk. No discussion is given in the Action Memorandum about how PCBs from ICS may be inhaled, and in what quantities. Thus, this route of exposure is totally unjustified. As for the possibility of direct contact, there is no evidence of swimming or other contact with the water and EPA admits that the area around the ICS has been fenced off. The only information that EPA provides related to a direct contact scenario is information from the Fish and Wildlife Service that its bird nest boxes nearby have been vandalized. This is not substantial evidence of a significant direct exposure pathway.

The ATSDR health assessment, which EPA excluded from the administrative record, considered these potential routes of exposure in great detail and concluded "neither children nor adults are likely to engage in activities in the . . springs and streams that would lead to significant exposures to site-related contaminants." ATSDR Report at 4. EPA has provided no evidence which contradicts ATSDR's conclusion, but simply Ignores it.

C. EPA's Theory of Human Exposure through Eating Fish Is Arbitrary and Capricious and Not Supported by the Evidence

Although EPA pays lip service to other routes of exposure, EPA's discussion of human risks is based almost entirely on its theory that people will eat fish contaminated with PCBs from ICS. EPA's theory has three analytical supports: (1) its reliance on the Indiana State Fish~Advisories as an indicator of endangerment to human health; (2) its use of incomplete and erroneous data to conclude that people fish in Clear Creek and eat the fish caught; and (3) its erroneous estimate of cancer risk. As discussed below, upon close analysis, each of these supports collapses of its own weight.

The cornerstones of EPA's conclusions that eating fish from Clear Creek represents a serious human health risk are its 1997 PCB data from the fish caught in the Creek and the State of Indiana Level 5 Fish Advisory, warning against eating fish from the Creek. In particular, EPA emphasizes that Clear Creek is one of ten streams in Indiana where there is a Level 5 Fish Advisory against eating all species.

EPA describes the August 19, 1997 data as showing "elevated" levels of PCBs; however, this data actually shows decreases as compared to fish sampling in 1996. More significantly, EPA's 1997 results would not justify the State of Indiana Level 5 Fish Advisories. EPA indicates that its 1997 results show a mean concentration of PCBs in edible fish tissue of: 446 ppb in rock bass, 1620 ppb in spotted sucker, and 872 ppb in largemouth bass. All of these figures, however, are substantially below the level of 1900 ppb that the State of Indiana uses as a bench mark for establishing a Level 5 Fish Advisory for PCBs;⁹ indeed, the PCB concentration in largemouth bass is less than half the 1900 ppb trigger for a Level 5 Fish Advisory, and the PCB concentration in rock bass is less than a fourth ofthe 1900 ppb trigger. Similarly, IDEM's own 1997 data do not show PCB concentrations which would justify Level 5 Fish Advisories for game fish species.

Thus, EPA's "elevated" 1997 data would have actually provided a basis for removing the Level 5 Fish Advisories at Clear Creek on at least some species. In which case, Clear Creek would be no different from all of the other streams in Indiana, since the State of Indiana has placed a Level 5 Fish Advisory on every stream in Indiana for PCBs (and mercury) in at least one species, and most other waterways in Indiana have lower level Fish Advisories for PCBs in other species. The fact that significant PCB levels have been found in at least one species in every stream in Indiana indicates that PCBs are ubiquitous in Indiana waterways, and implies other sources may account for the PCB levels in fish in Clear Creek. More significantly, this information shows that the PCB levels in the fish in Clear Creek do not present an unusual risk to humans that is not encountered elsewhere in Indiana.

Also, in order to show that fish in Clear Creek represent a serious health risk to humans, EPA must produce evidence that Clear Creek is actually used on a frequent basis for fishing, and that people actually eat the fish caught there. EPA has gathered limited anecdotal data on fishing in Clear Creek, but has provided no data to indicate that people actually eat fish from Clear Creek at all, let alone on a frequent basis.

EPA relies on three pieces of data to support its conclusion that fishing occurs frequently in Clear Creek, at least two pieces of this data came into existence after EPA announced its decision on August 18, 1998. The first piece of data is a five-line memorandum from Thomas Alcamo dated September 24, 1998 (administrative record item no. 13), stating that persons attending two Bloomington Citizens Information Committee meetings informed him that "people do fish frequently in Clear Creek." This memorandum fails to identify who are Mr. Alcamo's informants, who fishes in Clear Creek, how many people do so. how often is meant by "frequently," whether fish are caught of the appropriate species and size to be eaten, or whether people eat them.

The second and third pieces of data come from the September I 1, 1998 letter (and its attachments) from Scott E. Pruitt of the U.S. Fish and Wildlife Service to Mr. Alcamo (administrative record item no. 10) Mr. Pruitt's letter includes the anecdotal statement that "[e]vidence of fishing can be seen at every bridge crossing in lower Clear Creek." But no description of this evidence is offered, other than a statement that on one occasion a father was seen with two children going to fish in Clear Creek. Even with respect to this observation, there is no information that these people were fishing for food, indeed, the section ofthe stream at Gordon Pike where they were seen is too small to support a population of edible fish.

Mr. Pruitt's letter also refers to the 1991 and 1994 State of Indiana studies of fishing in the Lake Monroe area. In these studies, data is presented that in 1991 there were 14,492 angler/days per year in Salt Creek below the Monroe Reservoir darn, and in 1994, there were 9,635 angling days per year in the same area. Mr. Pruitt fails to point out that this data concerned fishing in the portion of Salt Creek before it reaches the confluence with Clear Creek. In other words, this data concerns fishing in a portion of Salt Creek, that is not impacted by waters from Clear Creek. There is no data in these reports concerning fishing in Clear Creek itself. Thus, the data referred to by Mr. Pruitt does not support a conclusion that Clear Creek waters are heavily fished. Indeed, the 1991 and 1994 studies demonstrate that Lake Monroe and other areas nearby provide better locations for fishing than Clear Creek, and that these locations are fished more frequently. In a 1997 newspaper article, Indiana state of ficials were quoted as saying that the State tested fish in the area just below the dam, and determined that "they came back clean," i.e., not at concentrations which would support Level 5 Fish Advisories. S. Hinnefeld, "Health Advisories Cover Many Area Fishing Spots," Bloomington Herald- Times (Aug. 10, 1997) (copy attached). Moreover, the article quoted State officials as indicating that there was little fishing in Salt Creek below the confluence with Clear Creek because of limited public access.

Although EPA includes attempts to include anecdotal evidence of fishing, it offers no data to indicate that fish caught in Clear Creek is actually consumed as food (as opposed to recreational sport fishing, in which fish are simply thrown back). The same 1997 newspaper article reported that there was much catch and release fishing around Lake Monroe, particularly for bass It is important to note that the Indiana Fish Advisories provide warnings about eating fish, but they do not restrict catch and release fishing for sport.

Furthermore, EPA ignores the conclusions of the ATSDR in its health assessment that Clear Creek did not provide a significant source of food for human consumption. ATSDR specifically concluded that the PCB levels in fish in Clear Creek were not a health concern because the stream was too small to support fishing for food, and that most of the species close to Lemon Lane Landfill are not considered a human food source. ATSDR Report at p. 5. EPA does not try to contradict this conclusion, but again simply ignores it.

The ATSDR findings are corroborated by the statements of long time Monroe County fisherman Dan Combs, who was quoted in the 1997 newspaper article. The 44- year old Mr. Combs said that "[n]ot since I was a kid have I seen anyone fishing for food in [Clear Creek]." Mr. Combs further explained:

It just got such a horrid reputation. And what gave it the horrid reputation was the Winston-Thomas raw sewage instead of the PCBs.

S. Hinnefeld, "Health Advisories Cover Many Area Fishing Spots," Bloomington Herald-Times (Aug. 10, 1997). Although the Winston-Thomas plant has been closed, the Dilman Road sewage plant is operating and discharging into the same area.

3. EPA's Calculation Overestimates Cancer Risks

EPA calculates a supposed cancer risk associated with eating fish from Clear Creek as follows: "[for the average freshwater sport fish consumer, ingesting an average of 15 g/day of largemouth bass, cancer risks would exceed I x 10E-4 (one in 10,000)." This conclusion is based on two unreliable assumptions. First, EPA uses an ingestion rate of 15 g/day. Except in the Great Lakes area, the State of Indiana uses a standard fish ingestion rate of 6.5 g/day to calculate cancer risk levels. 12 327 IAC 2-1-8.6. The standard of 6.5 g/day also is used generally by EPA as a national average for fish consumption in calculating cancer risks. EPA does not explain why in this calculation, it has used an ingestion rate that is more than double the accepted rate that serves as its own national standard and the standard in Indiana.

Second, EPA assumes that its hypothetical consumer eats fish from Clear Creek on an ongoing weekly basis. If someone were fishing for food on a weekly basis, that person would be more likely to fish in Lake Monroe or the area of Salt Creek just below the dam and before the confluence with Clear Creek, where the fish are both larger and more plentiful, than to try to catch them in Clear Creek where they are smaller and less abundant. [3 The more regularly someone would fish for food in the Bloomington waterways, the more likely that person would find the better fishing spots. Therefore, the possibility that someone would continue to eat fish from Clear Creek week in and week out, while ignoring better fishing in Lake Monroe, is absurd. In other words, EPA has created an unrealistic hypothetical to justify its endangerment determination.

IV. EPA's Selection of an Interim Treatment System is Arbitrary and Capricious. and Inconsistent with the NCP

Even assuming that there was a need for action and that need was time-critical, EPA has made an arbitrary and capricious selection of an interim treatment system. In settlement discussions, CBS had proposed a 300 8pm system. This system could catch all the water flowmg through the ICS during normal weather (when the estimated flow is in the range of 100 8pm) and during mild wet weather events; it would also capture a significant portion of flow during more severe wet weather events. On an annual basis, it could capture an estimated 55% of the estimated PCB mass. 14 The principal virtue of CBS's proposal was that it could be installed within a couple of months at an estimated capital cost of $75,000. By comparison, it will take EPA almost a year to install its selected system at a capital cost of $1 .3 million. ' For almost a year of delay, and over $1 million additional expenditure, EPA will be able to achieve a capture efficiency of 80% of the PCB mass.

If EPA is truly concerned about taking an interim response to a time-critical problem, its decision is highly questionable. By delaying almost a year in implementing its system, EPA would be allowing the PCB mass to continue to flow through ICS unreduced throughout that time, while CBS's proposal could be capturing 55% of PCB mass during that year. Over a three year period, EPA's proposal would capture an annual average of 53% of the PCB mass - 0 % in year I and 80 % in each of years 2 and 3 - about the same average capture percentage that CBS's system would achieve over the same period. But CBS's approach would allow capture to start about a year earlier and save over a $1 million in capital costs alone. In selecting an interim system that is only likely to be in use for a period between one and five years, there is no rational basis for choosing EPA's system rather than CBS's proposal.

Curiously, nowhere in EPA's Action Memorandum does the Agency address the question of how soon the various systems - either EPA's selected system or the one proposed by CBS - could be put into operation. Although the Agency calls its decision "time-critical," it completely omits any discussion of the timing of its action. EPA's failure to address the question of timing is significant, because the Agency has chosen, arbitrarily and capriciously, to reject the most timely approach in favor of one that will take almost a year to implement.

Moreover, although EPA describes its approach as an "interim" system, the Agency has ignored the importance of avoiding unnecessary duplication of effort and expense in constructing both an interim and a final system. One of the reasons why CBS had proposed a simple and inexpensive interim system was to allow time to complete on- going activities at Lemon Lane Landfill that may enable a more effective and efficient final system to be constructed. In particular, CBS has been undertaking an investigation of geologic conduits under the Landfill. This investigation may provide information about how to divert clean ground water and/or storm water before it can be contaminated with PCBs. Moreover, this study may help identify a better location for a final treatment system than the location selected by EPA. In addition, CBS is investigating different water management practices that may reduce the flow of water through ICS or the level of PCB contamination in that water.

EPA's Action Memorandum completely ignores these efforts. Instead, EPA calls for a year long effort to design and construct a massive system in a particular location, which may or may not be the location of a final system. EPA's decision is likely to result in the wasteful construction of an expensive facility that could be obsolete within a few years. ¹⁶

In the Action Memorandum, EPA has obligated $2,015,223 to be spent over a period of more than one year. 17 In doing so, EPA has violated section 104(c)(l) of CERCLA, 42 U.S.C. . 9604(c)(1), which prohibits EPA from obligating more than S2 million or undertaking a removal action for more than one year unless the Agency makes certain specific findings. ax The Agency has not made the necessary findings, nor could it do so based on the factual record.

The statute allows EPA to exceed the $2 million and one year limits if: (1) "response actions are immediately required to prevent limit or mitigate and emergency," (2) "there is an immediate risk to human health, welfare or the environment," and (3) "such assistance would not otherwise be provided on a timely basis." 42 U.S.C. . 9604(c)(1); see also the NCP at 40 C.F.R . 300.415(b)(5).~9 Thus, the statutory standard which must be met for EPA to commit more than S2 million for a removal action or to undertake removal action beyond one year is more stringent than the imminent and substantial endangerment standard. To avoid these limits on its removal action power, EPA must show that an actual emergency exists and that the risks are immediate. This standard has clearly not been met. Although EPA has asserted (without adequate support) that there is an imminent and substantial endangerment, EPA has not asserted or presented evidence to support a finding that there is an immediate risk to human health or that this system is needed to prevent an emergency. Here EPA admits in the Action Memorandum itself that it has known about the situation at ICS for over 13 years. Moreover, the 1997 fish data upon which EPA relies so heavily, shows decreases over 1996 data and concentrations that do not support Level 5 Fish Advisories. Most significantly, EPA has not decided to take immediate action. Rather, it has chosen to spend almost a year designing a treatment system, while rejecting CBS's proposal of a system that could be implemented in a couple of months.

Finally, EPA cannot support a conclusion that assistance would not be otherwise provided on a timely basis. CBS had offered to install its proposed system that would have been operational in a couple of months, long before EPA's system will be ready. But EPA rejected this proposal. CBS also offered to continue discussions of compromise approaches, but EPA cancelled a meeting scheduled for further discussions with CBS and announced its unilateral decision to take action and try to make CBS pay for it.

EPA's decision not only violates the express requirements of CERCLA, but conflicts with the Congressional policy behind the limitations on EPA's removal power. These limits are imposed to prevent EPA from skirting the substantive and procedural requirements for remedial decision making by calling long term actions "removals." This is precisely what EPA has done here: by calling its decision a time-critical removal, EPA has decided on a remedial action without going through the process, or meeting the substantive requirements, for selecting a remedy.

¹Nonetheless, in the course of settlement discussions with the government relating to alternatives to the original Consent Decree remedies, CBS offered to implement an appropriate interim treatment system at ICS, and to consider implementing a final system.

² Mr. Smith also stated that "[t]he majority of groundwater beneath Lemon Lane moves in conduit patterns responding to rain events." So the general relationship between rain events and PCB-contaminated ground water at ICS was also known at the time.

³Moreover, the geographic scope of the Covenant Not to Sue in paragraph 1 1 I(a) of the Consent Decree was not limited to six disposal sites, but extended to various "sites and areas" identified in specific paragraphs of the Consent Decree, including certain streambeds listed in paragraph 51. One of those streambeds is Clear Creek, which EPA now contends has been impacted by PCB-contaminated water from Illinois Central Springs. Therefore, the Covenant Not to Sue expressly protects CBS from environmental claims relating to contamination at Clear Creek.

⁴ See Rumpke of Indiana. Inc. v. Cummings Engine Co., 107 F.3d 1235 (7th Cir. 1997), in which the Seventh Circuit construed language in a CERCLA consent decree that barred claims relating to contamination transported from the site to preclude claims relating to contamination which leaches from the site into the Groundwater, reasoning that "the word 'from' is understood to relate to . . . phenomena, such as leaching and other similar leakage from the Seymour site itself." Id. at 1243 (Emphasis supplied).

⁵CBS also notes that under the heading "Site Conditions and Background," EPA listed CERCLIS ID # IND 9807-4341 as the relevant CERCLIS number for the affected site. This CERCLIS number refers to Lemon Lane Landfill. There is no separate CERCLIS number for Illinois Central Springs. This demonstrates that EPA has always regarded Illinois Central Springs as part of the Lemon Lane site.

⁶ Under 40 C.F.R. .. 300.415(n), 300.820, a public comment period is required whether or not on-site removal will occur within six months; however, if there is a planning period of at least six months, EPA must also perform an Engineering Evaluation/Cost Analysis ("EE/CA"), 40 C.F.R. . 300.415(a)(4)(i), and submit the EE/CA for public comment. 40 C.F.R. .. 300.415(n), 300.820. Even though the treatment system will not be operational for more than six months, EPA has not prepared an EE/CA

⁷ CBS also objects that administrative record item no. 10, the September 10, 1998 letter from Scott E. Pruit of the U.S. Fish and Wildlife Service to Mr. Alcamo contains references to preliminary data concerning songbirds downstream from ICS, but does not include this data. For almost a year now, Fish and Wildlife has referred to this data, but has failed to produce it despite numerous requests from CBS and frequent assurances by lawyers from the Department of Justice and the Department of the Interior that the data will be provided. It violates not only due process, but also every principle of good government, for the United States to invoke against CBS data that it refuses to provide. These references should be struck from the administrative record.

⁸An additional factual error in the Action Memorandum is worth noting. EPA states that "as many as 60,000 capacitors were disposed of at Lemon Lane Landfill." Although many estimates have been given of the number of capacitors disposed of at Lemon Lane Landfill, we have never seen an estimate anywhere near that high. The highest estimate we have seen is 40,000 capacitors.

⁹ According to James Stahl of IDEM, EPA never submitted its 1997 data to IDEM for consideration in establishing fish advisories.

¹⁰ Even though IDEM had collected its own 1997 fish data, that information was not considered by IDEM in setting the 1998 Fish Advisories; rather, IDEM simply continued to apply the pre-existing Level 5 Advisories to Clear Creek in 1998 without examining new data. Except in the case of the spotted sucker, which is not generally considered a game fish, IDEM's 1997 data would not support Level 5 Fish Advisories.

¹¹The letter itself is clearly dated after EPA announced its decision on August 18, 1998. The attachments, 1991 and 1994 studies of fishing in Lake Monroe and 1993 sample data, are from an earlier date, but there is no information that EPA was aware of that information until the date of Mr. Pruitt's letter.

¹² Indiana uses an ingestion rate of 15 gJday to calculate the risks of eating Great Lakes fish, because the fish caught in the Great Lakes are generally larger than those caught in smaller waterways, and because this rate is considered appropriate for a specific subpopulation that consumes Great Lakes fish.

¹³ The 1991 and 1994 fishing studies attached to Mr. Pruitt's letter substantiate that fishing occurs very frequently at Lake Monroe, which provides a good location for catching edible fish.

^|4 In the Action Memorandum, EPA incorrectly makes reference to a 200 8pm system which would capture 55% of the PCB mass at an estimated cost of S75,000. This appears to be an error, and CBS's proposal for a 300 8pm system is intended.

¹⁵ In Thomas Alcamo's letter of July 20, 1998 to Dorothy Alke (administrative record item no. 7), EPA raised concerns that CBS's proposed 300 8pm system would not work as intended. CBS disputed these concerns and prepared responses to each of EPA's criticisms. When CBS offered to present those responses at a face to face meeting with EPA and representatives of other Consent Decree parties on August 5, 1998, however, the governmental parties, including EPA, told CBS it was not necessary to respond because even if each ofthese criticisms were adequately addressed, CBS's proposal would still be rejected because they wanted a bigger system.

¹⁶ EPA does not identify the ARARs (applicable or relevant or appropriate requirements) in its Action Memorandum, but simply states that ARARs will be identified later. CBS, of course, would reserve all rights to challenge EPA's determination of ARARs in any cost recovery action.

¹⁷ As part of this cost figure, EPA includes a 40% extramural contingency of $465,051; this amount is calculated on top of a 20% contingency already built into the construction cost. EPA does not explain what this 40% contingency is for.

¹⁸ EPA's proposed selection of an interim system includes almost a year of design and construction work, followed by a year of operation by EPA before turning the system over to the State.

¹⁹ An additional exception to these limits exists where the removal action is consistent with a selected remedial action. But the Action Memorandum indicates that EPA has not determined to select any remedial action for the ICS.