Full opinion text
MEMORANDUM GILES, Chief Judge. TABLE OF CONTENTS I. INTRODUCTION.364 II. FINDINGS OF FACT AND CONCLUSIONS OF LAW A. Location and Description of the Site. 1. The Courtyard Area. 2. The Southern Area. 3. The River Sediments Area. B. Ownership and Early Use of the Property. C. Summary of Investigative and Enforcement History D. Unsuccessful Remediation Attempt. E. Nature of the Contamination. 1. PCBs. fc-CO 2. PAHS. to t-CO 3. Other Contaminants . ÍO t~ CO F. Migration of Contaminants. t-t-CO G. Health and Environmental Risks oi t-CO 1. Risks to the Ecosystem. Oi t-CO 2. Risks to Human Health _ w oo CO III. ANALYSIS CERCLA Liability. A. 1. Factual Basis for Finding Union Corporation is a “Covered Person” a. Under Corporate Veil Piercing . a-1. Legal Analysis. b. As Owner/Operator (Judicial Estoppel) . b-1. Legal Analysis.: c. As an “Arranger”. c — 1. Legal Analysis. d. As Lessee/ De Facto Owner. d-1. Legal Analysis. 2. Factual Basis for Finding Metal Bank is a “Covered Person”. a. As a-1. Legal Analysis. b. As an “Arranger”. b-1. Legal Analysis. 3. Factual Basis for John B. and Irvin G. Schorsch, Jr. are “Covered Persons” CD a. As Owners/Operators. ÍD a-1. Legal Analysis. CO Factual/Legal Basis for Finding Site is a “Facility” ^ Actual or Threatened Release. CR a. Factual Basis. a-1. Legal Analysis. 6. Response Costs . a. Factual Basis for finding EPA has Incurred. a-1. Legal Analysis. RCRA Liability. 1. Imminent and Substantial Endangerment. a. Factual Basis. a-1. Legal Analysis. 2. Endangerment Stemming from Handling, Storage, Treatment, Transportation or Disposal of Solid or Hazardous Waste. o o a. Factual Basis. ^ o o a-1. Legal Analysis. o 1 — 1 3. Factual Basis for Liability of the Schorsch Brothers. o CO a. As Owners/ Operators and Corporate Decision-makers who Contributed to the Contamination. 00 o ^ a-1. Legal Analysis. CO o 4. Factual Basis for Liability of Metal Bank. ^ o ^ a. As Owner/ Operator who Contributed . ^ o a-1. Legal Analysis. io o ^ 5. Faetual/Legal Basis for Liability of Union Corporation as Owner/ Operator who Contributed to the Contamination. io o ••ÑF IV. CONCLUSION 406 I. INTRODUCTION The United States (“Plaintiff’ or “the Government”) sues individual defendants Irvin G. Schorsch, Jr. and John B. Schorsch (“the Schorsch brothers”), and corporate defendants Union Corporation ‘(“Union”) and Metal Bank of America, Inc. (“Metal Bank”), seeking reimbursement for past and future response costs to investigate the Cottman Avenue Superfund site in Philadelphia and for enforcement pursuant to section 107 of the Comprehensive Environmental, Response, Compensation and Liability Act (“CERCLA”), 42 U.S.C. § 9607. The Government also seeks in-junctive relief and remediation of the site pursuant to section 7003 of the Resource Conservation and Recovery Act (“RCRA”), 42 U.S.C. § 6973. Defendants have filed a third party complaint against Baltimore Gas & Electric Company, Jersey Central Power & Light, Long Island Lighting Company, Metropolitan Edison Company, Orange and Rock-land Utilities Corporation, PECO Energy Company, Potomac Electric Power Company, PP & L Electric Utilities Corporation Virginia Power Company, Consolidated Edison of New York, Public Service Electric & Gas Company of New Jersey, (collectively “the utilities”), The Monsanto Company (“Monsanto”) and against the City of Philadelphia (“the City”). Filed in 1980, this matter was placed in suspense in 1983 pending attempted remediation of the site but was restored to the court’s active trial docket in 1998, upon the Government’s claim that remediation had failed or had not addressed all contamination concerns. Following extensive discovery and pre-trial proceedings, trial was phased as follows: Phase One would determine whether defendants were liable and whether response costs were incurred by the Government; Phase Two would determine whether the Government’s response costs, if any, were reasonable and recoverable, as well as the scope of any further remedial action; and Phase Three would determine the liability, if any, of the third-party defendants. Trial of Phase One issues commenced August 19, 2002. This memorandum explains the court’s disposition of Phase One issues in favor of the Government. This court has jurisdiction pursuant to section 113(b) of CERCLA, 42 U.S.C. § 9613(b) and section 7003 of RCRA, 42 U.S.C. § 6973, as well as under 28 U.S.C. § 1331 as a civil action arising under the laws of the United States and under 28 U.S.C. § 1345 as a civil action commenced by the United States. II. FINDINGS OF FACT AND CONCLUSIONS OF LAW A. Location and Description of the Site The Metal Bank Superfund site (“the Site” or “the Property”) is a former industrial property located on the Delaware River at 7301 Milnor Street in Philadelphia, Pennsylvania. The property is bordered by Cottman Avenue on the west, Milnor Street on the north, Safe Disposal Systems (an appliance recycling company) and Morris Iron & Steel Company (a metal salvage yard) on the east, and the Delaware River on the south and southwestern edge. The Site also includes the mudflat in the shallow embayment to the west. (Gov. Ex. 488 at 4, Gov. Ex. 481 at 2-1). A six-foot high fence surrounds part of the property, but the fence is not sufficient to exclude trespassers. There is also a gate for vehicular access from Cottman Avenue and the property is also accessible from the river and mudflats to the south and southwest. Graffiti evidences that trespassers have recently entered the property. (T. Tr. at IV — 111, V — 236-238; Gov. Ex. 481 at 2-1). The Site is generally divided into three areas of concern: the Courtyard Area, the Southern Area, and the River Sediments area. (See Gov. Ex. 488 at 14; Gov. Ex. 481 at 2-1 to 2-2). 1. The Courtyard Area The Courtyard Area consists of an open area at the northern section of the property near Milnor Street. The buildings near the Courtyard Area were demolished in 1998. (Gov. Ex. 481 at 2-2). 2. The Southern Area The Southern Area is an open area bordering the Delaware River where transformer processing operations once took place. Historical aerial photos show that most of the Southern Area is located in what was once part of the Delaware River and was gradually filled in beginning in approximately 1950. Heterogeneous fill material, most of which was placed before 1968, is about 15 feet thick. Its origin is unknown, but it contains construction debris, including chunks of concrete, brick, lumber, cloth and metal. The Southern Area sits approximately 10 feet above Mean Sea Level, and most of it is located within the 100-year floodplain. The outer edge of the Southern Area is steeply sloped, with large concrete blocks of material apparently placed for erosion control along approximately 550 feet of shoreline of the present-day Delaware River. (Gov. Ex. 488 at 11-14; Gov. Ex. 481 at 2-1 through 2-2; see also T. Tr. at 11-248; Gov. Ex. 494 at 1-3). This is the area known as the “upper rip rap.” There is also a “lower rip rap” area sometimes referred to as the “beach.” It consists of small heterogeneous fill material placed there as part of a clean up response ordered by the United States Coast Guard following an “oil spill” in that area that occurred in 1972. 3.The River Sediments Area The River Sediments Area, located adjacent to the southern and western boundaries of the Property, includes both mudflat and river bottom. The Delaware River is tidal in the vicinity of the Cott-man Avenue Property, with six to seven foot tides that reach maximum and minimum water levels every twelve hours. To the immediate west of the Property, the River forms a shallow embayment, which is completely submerged at high tide and which forms an exposed mudflat five to seven acres in size at low tide. The mudflat consists primarily of fine silts and clays, with' some occasional gravel in the subsurface, with the amount of gravel increasing closer to the border with the Property. The river bottom is composed of gravelly and sandy material, and it slopes gradually away from the Property. (T. Tr. at 11-247-249; Gov. Ex. 488 at 14; Gov. Ex. 481 at 2-2). The embayment is bordered to the north by St. Vincent’s School' — -a former orphanage currently serving as a day care center and an emergency shelter for at-risk children from the City of Philadelphia — and to the west by the Quaker City Yacht Club, which serves as a boat launch for recreational boaters. (Gov. Ex. 488 at 4). A municipal combined stormwater/sewer outfall (“CSSO”) owned by the City of Philadelphia is located at the foot of Cottman Avenue and empties into the northeastern corner of the embayment during and following periods of heavy rain. (See Gov. Ex. 488 at 4; Gov. Ex. 481 at 2-1). Groundwater underneath the Site flows into the Delaware River, generally from north to south. Depth of groundwater varies from seven to sixteen feet. (See Gov. Ex. 488 at 13). Recreational and subsistence fishing takes place in the Delaware River approximately 200 yards south of the Site from a public access ramp that was formerly used for boat access but is now open only for fishing. (T. Tr. at III — 195, 210). On April 11, 2001, the Commonwealth of Pennsylvania issued a general statewide fish advisory for recreationally caught sport fish which advises the public to eat no more than one meal (approximately one half pound) of sport fish caught in the state’s waterways. In addition, the Commonwealth has issued a more protective advisory for the Delaware River south of Yardley advising the public to limit or avoid consumption of white perch, striped bass, carp, channel catfish, and ameriean eel due to PCB contamination and to limit consumption of smallmouth bass due to mercury contamination. (Gov. Ex. 467A at 37-38; see also T. Tr. at III — 188). The fish advisory is posted on the Fish and Boat Commission’s website and is contained in a summary of fishing regulations and laws provided to holders of a Pennsylvania fishing license. Persons under the age of sixteen are not required to have fishing licenses and would not likely become aware of the warnings through the license application’s procedures and documents. The brochure is published only in English, although many of the people who regularly fish the Delaware River in that area are of Asian descent who profess not to speak English when approached by fishing authorities. (T. Tr. at III — 188, 192-93). An enforcement officer with the Pennsylvania Fish and Boat Commission testified at trial that she has seen people fishing in the Delaware River and then cooking and eating their catch along the shore. (T. Tr. at III — 196). The Commonwealth of Pennsylvania has cited merchants in some Asian markets in Philadelphia for selling illegally Corbicula clams that may have come from the Delaware River. (T. Tr. at III — 203). B. Ownership and Early Use of the Property In the early 1960’s, the Property was owned by L. Goldstein Sons, Inc. (“LGS”), a Pennsylvania Corporation. Irvin G. Schorsch and John B. Schorsch (collectively “the Schorsch brothers”) were LGS’s executive officers and sole shareholders. At some point in the 1960’s, LGS began doing business as the Metal Bank of America, Inc. (T. Tr. at XII — 23-24; see also Gov. Ex. 62 at 1-2). On December 4, 1968, LGS agreed to sell its business assets, except for its real estate at 6801 State Road (the “State Road Property”) and the Cottman Avenue property, to Defendant Union Corporation (“Union”), a New Jersey corporation. Union paid all the consideration for the LGS assets. As a result, the Schorsch brothers received 170,612 shares of Union common stock valued at an average of $12.24 per share (or $2,087,547), and $4,891,028 in cash and notes for an aggregate value of $6,978,575. Union agreed to (1) provide the Schorsch brothers with employment contracts to continue as Metal Bank’s executive officers; (2) provide them with incentive compensation in the form of more Union stock to be paid in the future; and (3) assume all of LGS’s liabilities. (T. Tr. at XII — 25-28; see also Gov. Ex. 25, Gov. Ex. 144, Gov. Ex. 758). On December 4, 1968, Union formed a wholly-owned subsidiary, MBA, Inc. (“Metal Bank”), a Pennsylvania corporation, to complete the LGS purchase. Union has been Metal Bank’s sole shareholder at all times since its creation. On or about December 9, 1968, Metal Bank changed its formal corporate name from MBA, Inc. to the Metal Bank of America, Inc., and then to U.C.O.-M.B.A., Inc. in 1985. (See Gov. Ex. 62 at 1). Union assigned its interest in the LGS purchase contract to Metal Bank, which closed on the transaction on or about December 27, 1968. Metal Bank continued the same metals recycling business it had conducted before the sale to Union. (T. Tr. at XII— 28-29; see also Gov. Ex. 62 at 1). Irvin Schorsch was President of Metal Bank from 1968 until at least 1980, when his title changed to Metals Investment Supervisor. He was a member of Metal Bank’s board of directors from 1969 until 1978. John Schorsch was Executive Vice President and a director of Metal Bank from 1968 until 1974, when he left Metal Bank to work in another of Union’s subsidiaries. (T. Tr. at XII — 87; see also Gov. Ex. 144, Gov. Ex. 758 at 34-35). LGS leased the Cottman Property to Metal Bank from December, 1968, until 1970. (See Gov. Ex. 198). In about March 1969, LGS liquidated and distributed its remaining assets, including the Cottman Avenue Property and the State Road Property, to its shareholders, the Schorsch brothers. They leased both properties to Union Corporation from 1970-72 and from June, 1972 through September, 1980. (See Gov. Ex. 199). Under the June 1972 lease, Union agreed to lease the Cottman Avenue Property “for refining, processing, conversion, separation of ferrous and non-ferrous metals and alloys, sales and administrative offices and uses and occupation necessary or desirable” .thereto. (Gov. Ex. 199 at ¶ 2). Through the lease, Union assumed responsibility for keeping the Property clean and in good repair (Id, at ¶ 8(b)) and for any damage to “any ... person” caused by “breakage or leakage” from “the drains, pipes, or plumbing works” at the properties (Id. at ¶ 11(a)). A rider attached to the lease expressly committed the lessors, Schorsch brothers, to provide “oil storage tanks” to lessee, Union. (Id. at ¶ 30). In late 1968 or early 1969, Metal Bank began recycling electrical transformers at the Property. It reclaimed the iron castings and copper cores from the transformers. Most of the transformers contained transformer oil which had to be drained from them before the metal reclamation could occur. An underground storage tank (“UST”), with a capacity of approximately 4,000 gallons, was installed underneath a bermed concrete pad in the Southern Area of the Property, and was located approximately 25 to 30 feet from the bank of the Delaware River. (Gov. Exs. 16, 722; T. Tr. at XII — 151). Metal Bank operated the Property as a scrap metal and transformer reclamation facility from some time shortly after the sale to Union until at least the summer of 1973. While the Property was in operation, all of Metal Bank’s transformer operations were conducted there. (T. Tr. at XII — 147-49; see also Gov. Ex. 439A at ¶ 5). Irvin Schorsch established business relationships with several major electric utility companies along the East Coast and arranged to purchase from them their scrap electrical transformers for recycling. (T. Tr. at XII — 32). Metal Bank processed many thousands of transformers at the Site. (See, e.g., Gov. Exs. 520, 521A, 757, 763). An electrical transformer generally has an iron or steel casing, within which a core of wound copper wire is immersed in a dielectric fluid or oil that is used for insulation and cooling purposes. Dielectric fluids used included mineral oil, as well as fluids consisting primarily of polychlorinat-ed biphenyls (“PCBs”). From 1968 or 1969 until 1973, the scrap electrical transformers processed by Metal Bank frequently contained PCBs, polycyclic aromatic hyrdrocarbons (“PAHs”) and other semi-volatile organic compounds (“SVOCs”), volatile organic compounds (“VOCs”), metals and other hazardous substances. (T. Tr. at 117-68; Gov. Ex. 718 at v-vi; Gov. Ex. 494 at 1-3). Metal Bank’s procedure was to leave its traders at the utility facilities. When the trailers were filled with scrap transformers, Metal Bank would send its trucks to pick up the trailers and bring them to the State Road site for weighing before driving them to the Site for processing. (T. Tr. at XII — 147-149). The trailers were then driven to an area near the concrete disassembling pad near the river. A crew of employees would take the tops off the transformers using hand-held power tools and would tip the transformers over on their sides. Oil from the transformers drained onto the concrete pad and flowed down the drain into the UST. Periodically, the oil was pumped out of the UST by a private contractor and was hauled from the Site and sold for industrial fuel purposes. When the transformers had finished draining, they were stacked to the side, away from the pad. After their copper coils were pulled out and recovered, and the steel canisters were loaded by crane onto trucks adjacent to the pad. (T. Tr. at XII — 147-149,188-190; Gov. Ex. 16). The Cottman Avenue reclamation operations were very sloppy. Photographs of the Site amply demonstrate this. (See Def. Ex. 786-1). Moreover, the Site operators did not regard the oil as posing any health risks and were not careful about the transformer draining and storage procedures. (T. Tr. at XII — 153-54). Oil dripped or spilled from leaking transformers as they were first unloaded at the Property. “Clam buckets” used to unload the transformers from the trucks, several at a time, sometimes ruptured transformers filled with oil that splashed onto the ground. Oil dripped from the stacks of transformers piled next to the pad as they awaited further processing. Transformers stacked on the pad for draining toppled off of the pad and the oil spilled into the ground rather than onto the pad. Rainwater sometimes got into the tank, forcing oil from the tank onto the ground. The ground near the UST became saturated with oil and was stained black. Sometimes, rivulets of oil ran across the surface of the property and down the riverbank into the Delaware River. (T. Tr. at I — 46, V — 253, XII — 188-190; Gov. Ex. 425 at 4-5; Gov. Ex. 16). At some time prior to August 3, 1972, a significant leak or rupture developed in the UST at the Site resulting in PCB-laden oil being released into the groundwater and thereby into the Delaware River. (Gov. Ex. 439A at ¶ 5; see also Gov. Exs. 16, 670). Then Metal Bank employee Roosevelt Thornton observed oil floating in the river near the tank. He noticed that the level of oil in the tank would change at times when oil had neither been added nor pumped out. For example, if employees ceased work on Friday and capped the tank, they would return on Monday to find the tank overflowing if it had rained over the weekend. When they then lowered a pump hose deep into the tank, they would pump water out of the bottom. The levels in the tank also rose when the tide was in, even if the tank was capped. (See T. Tr. at IX — 175-76). The volume of oil observed entering the river at the Property evidenced to the United States Coast Guard (“USCG”) and to Pennsylvania environmental inspectors that oil was flowing from the Site into the river, principally as a result of a leaking UST. In subsequent litigation with its insurance companies over clean-up costs, Metal Bank asserted and admitted that the UST was ruptured and leaking. Witnesses who were present at the Site at the time corroborated that the tank leaked. (Gov. Ex. 16; Gov. Ex. 425; Gov. Ex. 439A; Gov. Ex. 475; T. Tr. at XII — 149-151). Following the Coast Guard’s identification of the UST as the source of the oil spill into the Delaware River in August of 1972, Metal Bank had the tank inspected. In this record, there is no documentation of a leak being found or repaired. However, Metal Bank would not have admitted in litigation with its insurance company that the tank was ruptured and leaking if it had not been convinced by its own investigation that this was so. By inference, the court finds by a preponderance of the evidence that the tank did leak and was repaired during the tank inspection process. Defendants claim that the tank shows no evidence of leaking at this time. Metal Bank frequently received scrap capacitors mixed in with the bulk shipments of transformers from the various utility companies. The scrap capacitors contained extremely concentrated mixtures of PCBs. Metal Bank did not process these capacitors because they had no salvage value. However, some of the capacitors arrived broken or crushed, or were broken and crushed after delivery, such that dielectric fluids containing high concentrations of PCBs were released onto and into the Property. (T. Tr. at IV — 46; T. Tr. at V — 162-169, 177-190). Capacitors were strewn about the Property, in whole or in parts. Dr. Edward W. Klep-pinger, Defendants’ environmental consultant, observed scrap capacitors on the mudflat adjacent to the Cottman Avenue Property in 1980-81. Dr. John D. Schell, Defendants’ expert on environmental risk assessment, observed capacitors or parts of capacitors in the lower rip-rap and on the mudflat adjacent to the Property during his work at the Site in 2000 and after-wards. (T. Tr. at V — 177-190; T. Tr. at VII — 108,112-113). Along with the transformer processing business, Metal Bank conducted other scrap metal operations at the Site. These operations included receiving cans of cobalt residue, which was stored on Site prior to re-sale. Marvin Dayno, a former Metal Bank officer, testified that Metal Bank received shipments of cobalt residue from DuPont in five-gallon containers. He described the cobalt residue as a thick, black, tar-like, coal-type material. Metal Bank stored the cobalt residue outside on the grounds of the Property and resold the cobalt residue to its customers. Metal Bank shipped approximately two to three truckloads of cobalt residue per year. (T. Tr. at XII — 155-156; see generally Gov. Ex. 494 at 1-7 through 1-10). At the Site, Metal Bank also recovered copper from insulated copper wiring. Dr. Kleppinger testified that Metal Bank burned insulation off of copper wires in small furnaces known as “sputniks.” (T. Tr. at V — 153-54). Ash residue from the furnaces was spread upon the ground. (T. Tr. at IV — 37-38). In addition to the UST, there were four other underground storage tanks located on the Site. A storage tank registration form prepared by Metal Bank for submission to Pennsylvania’s Department of Environmental Resources identified four underground storage tanks at the Site: the UST (identified as a 6,000 gallon capacity tank), two 10,000 gallon tanks, and one tank with an estimated capacity of 15,000 gallons. Dr. Kleppinger, who assisted in the preparation of this form, testified that one of these tanks was used for the storage of heating oil. The contents and use of the other two tanks are not known at this time. Dr. Kleppinger also testified that there is another 35,000 gallon tank that is bisected by the Metal Bank property line. It has tested positive for benzene, a volatile organic compound, although its use and contents are also presently unknown. (Gov. Ex. 772; T. Tr. at V — 193-200). Whether tanks other than the UST ever leaked has not been determined. C. Summary of Investigative and Enforcement History On the morning of August 3, 1972, the USCG and Pennsylvania Environmental Inspectors responded to an oil spill in the Delaware River in the vicinity of the Quaker City Yacht Club and traced its origin to the Property. Numerous dead fish were reported. An inspection four days later revealed that the entire area near the bank of the Delaware River was saturated with oil, and areas were visible where oil had seeped through and poured over the bank of the Site into the river. (Gov. Exs. 670, 263; see also Gov. Ex. 439A at ¶ 5, Gov. Ex. 16 at ¶ 5). The court finds that this oil spill related solely to the Site and cannot be attributed to Bunker “C” oil from vessels passing the site in the Delaware River, as defendants contend. Various technical consultants estimated that the amount of oil released into the Site range from 11,700 gallons to 46,750 gallons. (Gov. Ex. 151 at 7; see also T. Tr. at 11-41; V — 8-9, 131). Although these estimates assume that all of the oil came from Metal Bank’s operations, even if a substantial portion came from upgra-dient sources as defendants contend, the defendants’ spills became commingled with all of the other oil, making any responsible defendant responsible for the entire clean up. During 1972 and 1973, the USCG collected samples from the Site, including samples of the soil and of the oil spill. Initially, the USCG did not detect PCBs in the oil samples. In late 1972 and 1973, in response to recommendations made by the USCG, Metal Bank took limited actions to clean up its property. It performed some surficial clean up of the southern portion of the property where the concrete pad was located, placed “booms” out to contain and collect oil in the river and along the shoreline, installed cylindrical caissons to capture the oil as requested by the USCG, and covered the ground with clean soil. However, Metal Bank did not undertake any efforts at that time to clean up subsurface contamination or to prevent subsurface oil from migrating into the river. (See Gov. Ex. 62 at 4). In the mid-1970s, in response to a USCG request that Metal Bank armor the shore, Metal Bank installed the lower rip rap, consisting of smaller pieces of rubble, broken bricks and concrete. (T. Tr. at IX — 126). The lower rip rap was designed to stabilize the lower portion of the upper rip rap and to act as a buffer both for oil spills coming from the land toward the water and for oil that might come from the River toward the land. Thus, the lower rip rgp likely was placed on top of oil that had not been recovered by the methodologies utilized by Metal Bank. At trial, although defense expert Dr. Kirk Brown did not agree that oil or oil residue is currently seeping from the Property through the lower rip rap, he opined that some oil could have been trapped by the lower rip rap installation and that the perceived oil sheens could be extractions of the oil deposit that are coming to the surface rather than seepage of oil trapped in subsurface areas of the Property bordered by the upper rip rap. (T. Tr. at IX — 141). In September 1977, in response to continuing concerns, the USCG and other government agencies re-analyzed the 1972 samples using more sophisticated analytical technology and detected the presence of PCBs in concentrations over 800 parts per million (“ppm”). (See Def. Ex. 53 at 1-4; Gov. Ex. 62 at 5; see also T. Tr. at I — 67). Analyses of soils and liquids samples at the Site detected the presence of PCBs at levels up to 1579 ppm. (See Def. Ex. 53 at 1-1 — 1-4). Based on these findings, EPA hired Roy F. Weston, Inc. (‘Weston”), to define more fully the nature and extent of PCB contamination at the Site. Weston conducted an investigation and documented its findings in two reports dated October 1978 and March 1980. (Def. Exs. 58 and 68). The 1978 Report showed that as many as 21,000 gallons of PCB-contaminated oil had pooled in the subsurface of the Cottman Avenue Property. The report concluded that this oil was releasing PCBs to the underlying groundwater and that PCBs from the Property were contaminating the Delaware River through oil and groundwater discharges. {See Def. Ex. 53 at 4-1). At Metal Bank’s request, Energy and Environmental Analysis, Inc. (“EEA”) also investigated conditions at the Property. In its May 8, 1979 report, EEA estimated that the extent of the oil spill from the ruptured UST affected an area of approximately 75,000 square feet and contained about 11,700 to 46,750 gallons of oil, and involved 115 to 460 pounds of PCBs. The EEA report also estimated that groundwater transporting oil to the Delaware River moved at the rate of 17,053 gallons per day. (See Gov. Ex. 151 at 6-7,11, 23). On April 23, 1980, the United States filed suit seeking injunctive relief and costs. Based on a Hazard Ranking System score of 33.23, EPA listed the Site on the Superfund National Priorities List (“NPL”) in 1983. See 48 Fed.Reg. 40669, 40673 (Sept. 8,1983). D. Unsuccessful Remediation Attempt On December 13, 1983, the court approved a Stipulation between the United States and Defendants that required Metal Bank to install and operate an oil recovery system until all recoverable oil was removed from the subsurface of the Site. Defendants hired Dr. Kleppinger to implement the system. The system consisted of three recovery wells, several oil separation units, and several 55-gallon drums containing activated carbon to treat groundwater. (T. Tr. at IV — 40-48). The system removed most, but not all, of the subsurface oil at the Site. (T. Tr. at V — 8, 260-261). In the late 1980s, defendants shut down the system and dismantled it, placing approximately one to two feet of clean fill material over the surface of the Southern Area. (T. Tr. at I — 91; IV — 109). After eight years of remedial efforts, defendants took the position that all feasible remediation had occurred and that the clean-up had been successful in that all remaining oil was permanently trapped and posed no risk of migration to the Delaware River or to the area of the embayment. However, the Government did not agree that the Site no longer posed a substantial hazard to human health or the environment. EPA monitoring in 1989 showed that despite eight years of groundwater pump and treat operations at the Site, a layer of PCB-contaminated oil at least three inches thick was still floating on the groundwater at some portions of the Site. (Gov. Ex. 644 at 9; Gov. Ex. 494 at 1-13— 1-14; Gov. Ex. 488 at 9). PCB concentrations measured in the oil layer were 1,539 ppm in 1977 prior to the oil recovery operation and almost the same, 1,540 ppm, in 1989 when the oil recovery operation was being terminated. (Gov. Ex. 488 at 17). On May 1991, EPA signed an Administrative Order by Consent with ten utilities, which had sent transformers to the Site and organized themselves as the Cottman Avenue PRP Group (“PRP Group”). Pursuant to the Administrative Order, the PRP Group conducted a Remedial Investigation/Feasibility Study (“RI”) to define the nature, extent, and sources of contamination at the Site and to estimate the health and environmental risks associated with the contaminants at the Site. The RI Final Report, dated October 14, 1994, documented widespread contamination by PCBs, TPH, PAHs and other hazardous substances at the Site. (Gov. Ex. 494). Pockets or layers of oil beneath the ground surface were found to contain PCBs at concentrations in the range of 520 ppm to 1,090 ppm. Courtyard Area soils at the ground surface and to a depth of approximately two feet were found during the RI to contain PCBs at concentrations up to 140 ppm. Southern Area surface soils were found during the RI to contain hazardous substances such as: • arsenic at concentrations up to 6.8 ppm; • copper at concentrations up to 149 ppm; • lead at concentrations up to 220 ppm; • PCBs at concentrations up to 4.7 ppm; • total SVOCs at concentrations up to 11.8 ppm. (Gov. Ex. 494 at Tables 4-3 to 4-5). Southern area subsurface soils were found during the RI to contain the following hazardous substances: • PCBs at concentrations up to 42 ppm; • the pesticide 4,4’-DDD at concentrations up to 11 ppm; • total VOCs at concentrations up to 907 ppm; • total SVOCs at concentrations up to 2,008 ppm; • arsenic at concentrations up to 21.1 ppm; • lead at concentrations up to 227,000 ppm (or 22.7% of the sample); • Mercury at concentrations up to 10.5 ppm. (Gov. Ex. 494 at 4-17 to 4-48 & Tables 4-6 to 4-10). Southern Area subsurface soils were found during the Pre-Design Investigation (“PDI”) to contain PCBs at concentrations up to 680 ppm. Twenty-one sample locations contained PCBs in excess of ppm; fifteen sample locations contained PCBs in excess of 50 ppm; and seven locations contained PCBs in excess of 100 ppm. A majority of the samples in excess of 100 ppm were located near the UST in the southwest corner of the Site. (Gov. Ex. 481 at 5-3 & Tables 5-3 to 5-5). Groundwater at the Site was found during the RI to contain the following hazardous substances: • PCBs at concentrations in water up to 25.6 parts per billion (“ppb”), and at a concentration in a floating oil sample of 1,090 ppm; • total VOCs at concentrations up to 5.6 ppm; • total SVOCs at concentrations up to 22.7 ppm; • total pesticides at concentrations up to 61.3 ppb; • arsenic at concentrations up to 369 ppb in unfiltered samples, and up to 67 ppb in filtered samples; • chromium at concentrations up to 288 ppb in unfiltered samples, and up to 102 ppb in filtered samples; • lead at concentrations up to 1,382 ppb in unfiltered samples, and up to 7.6 ppb in filtered samples; • mercury at concentrations up to 22.2 ppb in unfiltered samples and up to 0.9 ppb in filtered samples. (Gov. Ex. 494 at 4-53 to 4-70 & Tables 4-13, 4-15 to 4-17). Sediments in the Delaware River adjacent to the Site were found during the RI to contain the following hazardous substances: • PCBs at concentrations up to 6.8 ppm; • total SVOCs at concentrations up to 244 ppm; • lead at concentrations up to 2,030 ppm. (Gov. Ex. 494 at 4-70 to 4-79 & Tables 4-20 to 4-23). Sediments in the mudflat and Delaware River were found during the PDI to contain PCBs at concentrations up to 6.1 ppm. Out of a total of 45 samples, 11 contained PCBs at concentrations exceeding 1 ppm. (Gov. Ex. 481 at 5-5 & Table 5-10). Sediments in the mudflat area were found during a January 2002 sampling event to contain PCBs at concentrations up to 22.1 ppm. (Gov. Ex. 499 at 9-10 & Table 3). Sheens of oil originating from the sediments adjacent to the'Site were witnessed as recently as December 2001. (T. Tr. at TV — 17-18, 19-20). Defendants’ trial experts opined that sheens present in the waters adjacent to the Site are biological in nature. The court credits the witnesses who concluded that the sheens are oil. They testified that the sheens come from pockets of oil that can be produced upon digging shallow depths in the beach or lower rip rap., and that what looked and smelled like oil, is oil. Manufacturers of PCB-containing transformers used PCB Aroclor 1260 as the principal component of the transformers’ dielectric fluid. (Defs. Ex. 534 at 5). Ar-oclor 1260 has been found on-site, in the rip rap area, in the beach area and in the Delaware River sediments adjacent to the Site. (Gov. Ex. 494 at Ch. 4; Gov. Ex. 481 at 5-3 to 5-5). Similarly, before 1971, manufacturers of PCB-containing capacitors used Aroclor 1242 as a principal component of capacitor dielectric fluid. (Defs. Ex. 534 at 5). Aroclor 1242 has also been found on-site, in the rip rap area, in the beach area and in the Delaware River sediments adjacent to the Site. (Gov. Ex. 494 at Ch. 4; Gov. Ex. 481 at 5-3 to 5-5). Transformers manufactured during the period from approximately 1950 to 1970 contained polycyclic aromatic hydrocarbons (“PAHs”), including naphthalene, phenanthrene, and 2-methylnaphthalene. (Gov. Ex. 718 at Table 3-7, pp. 3-8 to 3-9). Naphthalene, phenanthrene and 2-methyl-naphthalene are among the PAHs that have been found in the soils and sediments at the Site. (Gov. Ex. 494 at Ch. 4). Based on the results obtained in the RI, EPA prepared a Proposed Plan for remediation at the Site, which it circulated for public comment in 1995. (Gov. Ex. 488 at 10). Following review of the comments received, EPA issued a Record of Decision (“ROD”) in December, 1997. The ROD established PCB cleanup action levels at the Site — 10 ppm in surface soils, 25 ppm in subsurface soils, and 1 ppm in mudflat sediments — and documented selection of the remedial action to be implemented. (Id. at 1-3). On June 26, 1998, the EPA issued an Administrative Order for Remedial Design and Remedial Action to the Defendants and the members of the PRP Group. In accordance with the Administrative Order, the PRP Group conducted a Pre-Design Investigation (“PDI”) to collect engineering data in support of the design for the remedy and to further determine the scope of contamination. (Gov. Ex. 481). The PDI final report, issued on January 21, 2000, confirmed the existence of a layer of oil floating on the groundwater table beneath the Southern Area and further delineated the extent of PCB contamination at the Site. (Id. at 1-1 to 1-2). A measurable layer of oil ranging in thickness from 0.125 inches to 5.75 inches was detected in the southwest corner of the property. (Id. at 4-7). In the Courtyard Area, PCBs were detected in two of eleven samples, one of them at a concentration of 8.2 ppm and the other at 190 ppm. (Id. at 5-1). In the three areas of concern delineated within the Southern Area, PCBs at concentrations above the EPA cleanup action level of 25 ppm were detected in 38 out of 231 subsurface borings. (Id. at 5-2 to 5-3). The highest concentration (680 ppm) was found in a soil sample taken from near the UST. (Id.). Sediment samples taken from 45 locations in the River Sediments Area, including the mudflats, had total PCB concentrations ranging from non-detect to 6.1 ppm. (Id. ■ at 5-5). As.part of the PDI, the combined sewer system located underneath Cottman Avenue was inspected and seven sediment samples were collected from inlet pipes leading into the sewer. (Id. at 3-12 to 3-14). Only one of them had a PCB concentration higher than the EPA cleanup action level of 1 ppm for sediments, and that concentration was 1.3 ppm. (Id. at 5-5). During the summer of 2000, Defendants conducted a field study to look for the presence of Light Non-Aqueous Phase Liquids (“LNAPL”) at the Site, issuing a report in September, 2000, entitled “Data Report, Cottman Site Investigation, July 2000” (“2000 Trench Study”). (Gov. Ex. 103; see also T. Tr. at I — 88-109). As part of that study, Defendants opened up seventeen previously installed monitoring wells and piezometers to check for the presence of oil over the course of about four weeks. (Gov. Ex. 103 at 3). During this time, they found measurable oil in 10 of the wells and piezometers, with the thickest measurement being 15 inches of oil in one of the piezometers. (Id. at Table 1). Defendants also excavated five trenches using a trackhoe and observed oil in all of the trenches except for Trench # 3, which was not fully excavated due to the presence of a gas line. (Id. at 36^13). A layer of oil approximately four inches thick was found floating on the groundwater in one of the trenches. (Id. at 38). In addition to measuring the oil thickness, Defendants tested for PCBs in the groundwater and oil underneath the property, as well as in a groundwater “seep” discharging at the base of the upper rip-rap into the mudflat area. (Id. at 6, 34-35). PCBs in groundwater were measured at levels up to 7.2 ppm; PCBs in oil were measured at levels up to 530 ppm; and PCBs in the groundwater seep were measured at a concentration of 0.22 ppb. (Id. at Tables 2 and 3). In January 2002, EPA’s consultants returned to the Site to sample mudflat sediments in the area on the western edge of the Site. Sediment samples yielded PCB concentrations up to 22.1 ppm, and samples of oily liquid that pooled in small holes excavated in the mudflat yielded PCB concentrations up to 360 ppb. (Gov. Ex. 499 at 10 (Table 3), 12 (Table 5)). Both the sediment and groundwater samples also showed significant levels of PAHs. (Id. at 11 (Table 4) and at 13 (Table 6); see also T. Tr. at I — 102-04). All Government (EPA) witnesses present during the January 2002 sampling event described a strong, oily, petroleum smell coming from the shallow holes dug to gather sediment samples. (Gov. Ex. 499 at 6; T. Tr. at I— 105,106-07). EPA’s consultants again sampled mudflat sediments and groundwater on June 19, 2002. Analysis of these samples indicated the presence of dioxin-like PCB con-geners, PAHs, volatile organic compounds (“VOCs”), semivolatile organic compounds (“SVOCs”), dioxins, and furans. (Gov. Exs. 688 and 642A; T. Tr. at I — 116-24). E. Nature of the Contamination The Metal Bank Site is contaminated with PCBs, PAHs, SVOCs, VOCs, petroleum hydrocarbons (“TPH”), metals, dioxins, furans, pesticides and other hazardous substances. (Gov. Ex. 494 at Table 6-1; Gov. Ex. 488 at 14-27; T. Tr. at 163-65). The levels and scope of contamination have been documented in the RI/FS, the ROD, the PDI, the 2000 Trench Study and the mudflat data collected during 2002. (Gov. Ex. 717; T. Tr. at 1-52-66). 1. PCBs PCBs are a group of synthetic organic chemicals which were widely used in the United States because of their chemical stability and low reactivity. (Gov. Ex. 725 at 1; Gov. Ex. 644 at 19). Each PCB molecule consists of a chlorinated biphe-nyl — two hexagonal rings of carbon atoms connected by a carbon-carbon bond — containing from one to ten chlorine atoms attached in various locations. (Gov. Ex. 725 at 1). There are 209 different types— or “congeners” — of PCBs, based upon the number of chlorine atoms and their positions on the carbon rings. (T. Tr. at 1123, no). PCBs were manufactured in the United States between 1929 and 1977 as complex mixtures of individual congeners known as “Aroclors.” (Gov. Ex. 725 at 1; T. Tr. at II-50). Each Aroclor mixture has a product number, the last two digits of which generally refer to the average percent of chlorine by weight. “Aroclor 1260,” for example, contains 60% chlorine by weight. (T. Tr. at 1150). The individual congeners that make up Aroclors have different physical properties, including different rates of solubility, volatilization and degradation. (T. Tr. at 11-51). These differences can affect each congener’s persistence in the environment and its rate of migration, with higher chlorine congeners tending to be more stable in the environment. (Id. at III-52). As PCB mixtures move through the environment, the absolute and relative concentrations of individual congeners change over time due to differences in their physical properties as well as differences in the rate of bioaccumulation by living organisms. (Gov. Ex. 725 at 1). The low reactivity and high chemical stability of PCBs made them useful for a number of industrial purposes, including as a constituent of insulating fluid used in electrical transformers and capacitors. (Gov. Ex. 725 at 1). “Askarel transformers,” used a fluid containing 60 percent or more PCBs by weight. (T. Tr. at II — 26-27, 67; T. Tr. at V — 170). “Mineral oil transformers” contained a grade of mineral insulating oil that was frequently contaminated with PCBs through recycling operations or reuse of contaminated mineral oil. (Gov Ex. 527 at 1-2; Gov. Ex. 113 at ¶ 15; T. Tr. at 1-67-71,11-27). PCBs can cause a variety of adverse health effects. They are classified as suspected human carcinogens and may damage the immune system, may cause developmental problems in children and impair reproductive systems. (T. Tr. at III — 113; Gov. Ex. 488 at 15). PCBs have been shown to cause severe effects on exposed aquatic organisms and wildlife, including suppression of immune responses, impairment of reproduction and development, disruption of endocrine function, cancer, and organ enlargement and malfunction. (Gov. Ex. 647 at 31). Most PCBs do not degrade very quickly in the environment, and they can persist for many years in sediments, where they can cause adverse effects, not only to individual organisms, but also to entire aquatic populations or ecosystems. (Gov. Ex. 725 at 18). The most common way to test for PCBs in the environment is to perform Aroclor testing by using a gas chromatograph instrument to detect the pattern of PCBs that is currently present in a sample. (T. Tr. at II — 53, 57-58). Where chromato-grams, which are print-outs from the chro-motograph, match the pattern of a known formula, the presence of a particular Aroc-lor can be confirmed. (T. Tr. at II — 53-57, 243-244). Aroclor testing has its shortcomings because as PCB mixtures are exposed to the environment, the concentration of individual congeners can change over time. (T. Tr. at II — 53-55, III — 115). In addition, the presence of a contaminant in a sample can interfere with analysis, causing inaccuracies in interpreting chromatograms. (T. Tr. at II — 61, 243-244). Although Aroclor testing is helpful for screening a- site, it may be insufficient to characterize risk fully. (T. Tr. at III — 117). The more accurate, but more expensive and less frequently used, means of measuring PCBs in the environment is congener testing. (T. Tr. at II— 64, III — 117; see also Gov. Ex. 647 at 81). Dr. Allen Medine, an expert witness for the United States in the field of analytical chemistry and environmental engineering, reviewed chromatograms from various Ar-oclor testing at the Site, including chroma-tograms from the Remedial and Pre-De-sign Investigations, and determined that PCB contamination at the Site has been consistently under-reported due to problems inherent in Aroclor analysis. (T. Tr. at II — 57-63, 241-246). PCBs are known to be present in surface soils in the Courtyard Area at concentrations up to 190 ppm (Gov. Ex. 481 at 5-1), in subsurface soils in the Southern Area at concentrations up to 680 ppm (Id. at 5-3), in oil underneath the Southern Area at concentrations up to 530 ppm (Gov. Ex. 103 at Table 2), and in the mudflat sediments at concentrations up to 22.1 ppm. (Gov. Ex. 499 at 10 (Table 3)). 2. PAHs Polycyclic aromatic hydrocarbons (“PAHs”) are a group of semi-volatile organic compounds formed during the incomplete combustion of organic substances. (Gov. Ex. 644 at 21). PAHs are components of the raw material that is used to make various grades of oil, including mineral insulating oil. (T. Tr. at II— 67). Research has shown that mineral insulating oil contains a high percentage of PAHs. (T. Tr. at II — 26; see generally Gov. Ex. 718). PAHs tend to persist in the environment for a long time, where they pose a risk to both humans and aquatic life. (T. Tr. at II — 240; T. Tr. at III1 — 52; see also Gov. Ex. 647 at 36-38, Gov. Ex. 488 at 15). Exposure to high doses of PAHs can be acutely toxic to aquatic life. The harmful effects from lower doses may include immune system dysfunction, organ failure and physiological impairment. (T. Tr. at III — 51—52; Gov. Ex. 647 at 36). In addition, some types of PAHs, including benzo(a)pyrene, are known to be carcinogenic. (T. Tr. at II— 240; Gov. Ex. 488 at 15). The National Oceanic and Atmospheric Administration has determined the toxicity threshold level for PAHs (that is, the level at or above which an organism exposed to a hazardous substance would be expected to suffer a toxic effect) to be 44.8 ppm. (T. Tr. at III — 53-54; see also T. Tr. at II — - 234-235). The toxicity threshold for ben-zo(a)pyrene is 7 ppm. (T. Tr. at III — 56-58). Total PAHs and benzo(a)pyrene have been measured in the mudflats adjacent to the Property at levels well above the toxicity thresholds. In January 2002, total PAHs were detected at concentrations as high as 869 ppm, and benzo(a)pyrene was measured at concentrations of up to 46.6 ppm. (Gov. Ex. 499 at Table 4). In the early 1990’s, the Delaware River Basin Commission (“DRBC”) studied PAH concentrations in the Delaware River at locations both upstream and downstream from the Metal Bank Site. (T. Tr. at III — 33-34). Samples taken from the mudflats at the Site had PAH concentrations ranging from ten to sixteen times greater than the concentrations reported by DRBC at other locations in the Delaware River. (T. Tr. at III — 33-34). 3. Other Contaminants Various metals — including arsenic, cadmium, copper, lead, and mercury — have been detected in the mudflats at the Site in concentrations above the toxicity threshold level. (Gov. Ex. 488 at Table 9). Arsenic is of particular concern at the Site because it was detected in the mudflats at a concentration up to 290 ppm when the screening level (i.e., toxicity threshold) is only 8.2 ppm. (T. Tr. at II — 237-238). Arsenic exposure at this level poses a threat to the aquatic ecosystem. (T. Tr. at 11-238). Dioxins and furans were also found at the Site. Dioxins and furans are halogenated aromatic hydrdocarbons whose molecules, like PCBs, consist of two benzene rings with variable amounts of chlorine attached. (Gov. Ex. 644 at 20; T. Tr. at II — 69). Dioxins are naturally produced through the incomplete combustion of organic materials and are extremely toxic to human health. (T. Tr. at II — 71, III — 112; Gov. Ex. 644 at 20). Furans are inadvertent contaminants present in PCB mixtures. (T. Tr. at 11 — 71). F. Migration of Contaminants The evidence shows that contaminants from the Property have migrated, and are continuing to migrate, into the Delaware River and mudflats adjacent to the Property. (T. Tr. at II — 12; see also T. Tr. at III — 25, 28). Thousands of gallons of transformer oil are still present at the Site, and some of that oil is continuing to migrate from the Property into the River and mudflats. (T. Tr. at II — 22). Oil has been observed in the subsurface of the Southern Area within the past two years. (Gov. Ex. 103 at Table 1). Because of the heterogeneous nature of the fill and uncertainty as to the amount of oil that was originally released at the Site, it is difficult to quantify precisely how much oil remains. (T. Tr. at II — 41). Oil has been visually observed in the mudflats. Linda Dietz, EPA Senior Remedial Project Manager, and Drs. Jerome Diamond and Allen Medine, two expert witnesses for the United States, all testified that they have seen oil sheens coming from the mudflats in the area near the rip-rap at the southwest corner of the Property. (T. Tr. at 1 — 120-121, 11 — 17, III — 6). All three testified that when they dug down a few inches into this area of the mudflats, they saw a dark, oily-looking and oil-smelling liquid percolate into the holes. (T. Tr. at 1 — 113, 11 — 18, III — 6). The experts testified that they did not see or smell this oily material in areas of the mudflat further away from the southwest corner of the property. (T. Tr. at II — 19, III — 6). Another way to measure oil is to look for Total Petroleum Hydrocarbons (“TPH”), which are derived from oil. (T. Tr. at II— 34-35, III — 29-30). Concentrations of TPH in the range of 10,000 ppm and greater have been measured at the Property. (T. Tr. at II — 35). The highest concentrations have been measured in the southwest corner of the Property, in the vicinity of the underground storage tank. (T. Tr. at II — 35). TPH has also been measured in the mudflats near the southwest corner in the range of 1,000 to 6,000 ppm. (T. Tr. at II — 36). Contaminant concentrations decrease in a gradient from the southwest corner of the Property westward across the mudflats. (T. Tr. at 11-12-13, 32-36). While TPH concentrations in the mudflats are highest adjacent to the southwest corner of the Property, (T. Tr. at II — 36), TPH concentrations further away from the southwest corner, both toward the CSSO and towards the Yacht Club, are much lower. (T. Tr. at II — 36-39). Similarly, in the mudflats, PCBs are found in the highest concentration in the area adjacent to the southwest corner of the Property, and the concentrations decrease with distance away from the southwest corner. (T. Tr. at II — 65, III — 23-24, 34; Gov. Ex. 644 at Map 2). Most PAH hotspots in the mudflats — those with total PAH concentrations of 40 ppm or greater — are located in the area of the mudflats adjacent to the southwest corner of the Property. (T. Tr. at III — 26-28; Gov. Ex. 137.06 — 137.09). During the June 2002 mudflat sampling, the concentrations of PCBs, PAHs, volatile organic compounds, dioxins and furans detected in the samples from near the yacht club were from ten to one hundred times lower than the samples from near the Property. (T. Tr. at 11 — 32-33; see generally Gov. Ex. 688). In addition, the same group or suite of contaminants can be found on the Property and in the adjacent mudflats and river sediments. (T. Tr. at II — 49). This suite includes PCBs, PAHs, dioxins, furans, and volatile organic chemicals. (T. Tr. at II— 49). The similarities between the contamination found in the two areas demonstrates by a preponderance of the evidence that the contamination in the mudflats came from the Site. (See T. Tr. at II — 40, 49). The subsurface oil at the Property, groundwater flow and tidal actions account for transformer oil containing PCBs and other contaminants leaching into the mudflat. (T. Tr. at II — -12). The lower rip rap installation was over oil residue that now has found pathways through the fill to the surface as sheens and to areas beneath the beach that are subject to disturbance. In addition, the historical evidence shows that there was an oil pathway to the embayment through or under the upper rip rap in 1972 when the underground oil leaked from the UST. The lower rip rap installation did not close off that pathway entirely, inasmuch as the contamination in the mudflats at the southwestern portion of the Property is consistent with the suite of contaminants found on, and under, the Site in the vicinity of the UST. The court credits Dr. Allen Medine’s opinion that “... oil can find ways to migrate, regardless of whether or not you have a thick layer floating on the groundwater environment.” Dr. Medine explained that where there are pockets of oil, “we just need to move little bits of it and carry those contaminants with it.” Thus, “[tjhere would be opportunities, that may be affected by tides, ... precipitation ... or drought conditions, that can cause some of that [oil thought to be trapped] to move.” (T. Tr. at II — 43-45). Groundwater at the Property generally flows in a southwesterly direction from the Southern Area to the mudflats. (T. Tr. at II — 28). Because the bulk of the Southern Area is constructed on artificial fill, the specific flow pathways for the groundwater and oil flow are difficult to characterize. (T. Tr. at II — 29-30). The presence of chunks of concrete, bricks, wire, pipe and other material in the heterogeneous fill may offer preferential pathways for the migration of contaminants. (T. Tr. at II — 30). The court finds that the oil has found, and continues to find, pathways from the ground and groundwater through the rip rap, upper and lower and into the beach and mudflat areas. Defendants have failed to show that the pollution in the mudflats was caused exclusively by a source other than the Cottman Avenue Site. A lampblack factory was in operation on a neighboring property from 1849 until 1970. Lampblack is a fine powdery material produced by burning low grade oils, creosote, coal tar, anthracene oil and crude oil. During operations, coal tar and crude oils were stored in tanks on the property. Due to the absence of waste product controls, discharges of lampblack, and raw materials used to manufacture it, possibly may have occurred into the area surrounding the factory, although there are no records of spills. However, the court finds that this possibility does not account for the PCBs, PAHs, and petroleum hydrocarbons at the Site and in the adjacent mudflat and river sediments. The topography of the area makes it improbable that contaminants from the lampblack factory migrated to the Site. An historic channel running southeast past the lampblack property would have provided a natural pathway to the Delaware River for any discharges from the factory and would not account for contamination of the southwest portion of the Site or for groundwater contamination thereunder. (T. Tr. at II — 86-88, 93-94, 99-100; Gov. Exs. 688, 688-A, 470.03, 470.04; Defs. Ex. 1102). Similarly, the court finds that the CSSO is not a significant source of contamination to the mudflats. (T. Tr. at II — 74-75; see also Gov. Ex. 647 at 15). If it were, the PCBs and other contaminants would be distributed across the mudflats and would not be concentrated in the area adjacent to the southwest corner of the property. (Id. at 11-76). Given the overwhelming evidence tracing the contamination of the Metal Bank Site to its own operations, the court is unpersuaded by Defendants’ argument that background chemicals (naturally-occurring or other non-site related chemicals) account for the contamination at the Site. G. Health and Environmental Risks 1. Risks to the Ecosystem The mudflat in the embayment adjacent to the Property is located in a relatively undisturbed and environmentally sensitive area, with the mudflat providing attractive habitat for a wide diversity of aquatic organisms, including plant life, invertebrates, fish and birds. (T. Tr. at II — 248-54). A number of contaminants have been detected in the River Sediment Area at levels well above EPA’s screening toxicity levels (that is, at levels likely to cause toxic effects to aquatic organisms). (T. Tr. at II — 234-35). These contaminants include PCBs, PAHs (such as benzo[a]pyrene, fluoranthene, fluorene, and pyrene), heavy metals (such as arsenic, cadmium, lead and mercury), and pesticides (such as DDT and its derivatives). (T. Tr. at II — 234-235; Gov. Ex. 488 at Table 9). The upper mudflat area furthest from the Property supports a community of aquatic emergent plants. The lower mudflat area furthest from the Property supports filamentous algae. These plants offer potential habitats and food sources for various invertebrate, fish and bird species in the Delaware River. (T. Tr. at II — 251-254; Gov. Ex. 647 at 23). The combination of rocky substrate and fine sediments at the mudflat provide habitat for a number of invertebrate species, including the Asiatic clam Corbicula, am-phipods, midge larvae, snails and various species of freshwater worms. (Gov. Ex. 494 at 3-24; Gov. Ex. 647 at 24). These organisms fill an important niche by feeding on sediments and recycling nutrients and energy in the mudflat ecosystem. (T. Tr. at III — 13-14). They also serve as valuable food sources for fish and birds species. (T. Tr. at III — 14). During the RI, Corbicula sampled at and near the Site were found to have total PCB concentrations ranging from 0.23 to 1.03 ppm. After normalizing for lipids (accounting for the increased solubility of PCBs in fatty tissue), the PCB concentration in Corbicula on a lipid-basis ranged from 17.4 ppm to 75.8 ppm. (Gov. Ex. 494 at 4-108). In the early 1990s, the DRBC conducted a study of PCB concentrations in mollusks at four different locations in the Delaware River downstream from the Site. (T. Tr. at III — 36-38; Gov. Ex. 647 at 27; Gov. Ex. 137.11). PCB concentrations in Corbi-cula collected from the mudflats near the southwest corner of the Property were approximately 2 to 6 times higher than PCB concentrations in mollusks from the DRBC study. (Gov. Ex. 647 at 27). Cor-bicula are filter feeders, so their primary exposure to PCBs is through the water column as opposed to through sediment particles. (T. Tr. at III — 36; Gov. Ex. 647 at 26). Because PCBs favor sediments over water, worms and other organisms that feed in the sediments will likely show higher PCB concentrations than filter feeders. (T. Tr. at III — 36-37, Gov. Ex. 647 at 26). PCB concentrations observed in Corbicula are environmentally conservative estimates of bioaccumulation by aquatic species near the Metal Bank Site. (Gov. Ex. 647 at 26). Because PCBs bioaccumulate in the environment by building up in the tissues of living organisms over time, low concentrations o