Placing Superfund Liability Determinations in the Hands of Scientists

James A. Rogers

Editors' Summary: Relying on federal courts to allocate a major contaminated site's Superfund liability can take years, and "quick and dirty" efforts to allocate liability are rarely successful. Faced with these harsh realities, over 30 parties potentially responsible for contamination underlying the eastern San Fernando Valley near Los Angeles, California, decided to rely on scientists to allocate liability for the cleanup of the Glendale operable units. In this Dialogue, the author describes this novel approach of placing Superfund liability determinations in the hands of scientists. It begins with an overview of the San Fernando cases and the parties' early efforts to obtain a bargain-basement allocation. After these efforts proved unsuccessful, the parties agreed to a binding arbitration in which a panel of three scientists would be the arbiter. Therefore, much of the Dialogue describes the arbitration agreement, the alternative dispute resolution process, and the panel's decision. The arbitration was successful in that it enabled the parties to avoid contentious litigation. The panel's decision was largely based on an independent, advanced understanding of contaminant movement in groundwater, geochemistry, mathematical computer modeling, and environmental engineering that is not readily available to federal judges. However, the extent to which the panel's decision represents an improvement over traditional judicial decisionmaking depends on the party and how that party fared in the panel's decision.

James A. Rogers is a partner and the head of the environmental practice group at Wilmer, Cutler & Pickering in Washington, D.C. He served as the Chairman of the Board of Directors of the Environmental Law Institute from 1987 to 1990, and he was the Associate General Counsel for Water and Solid Waste at the U.S. Environmental Protection Agency from 1977 to 1981.

[29 ELR 10259]

After living for almost two decades with what has become commonly referred to as the Superfund law, environmental lawyers have learned that, even under the best of circumstances, the process of allocating liability for a major contaminated site can take years and often consumes more money than the cost of the underlying cleanup. There also seems to be widespread agreement among weary Superfund lawyers that attempting to divvy up responsibility through traditional litigation is among the least desirable of alternatives: the federal dockets are crowded and most judges do not want to devote months to hearing truck drivers speculate on the nature of wastes they may have transported 30 years ago to a dump. In addition, the rigid rules of traditional discovery and evidence presentation are out of sync with the whole Superfund world — where the U.S. Environmental Protection Agency (EPA) may designate responsible parties on the basis of sketchy information that would not meet a preponderance of the evidence test in an actual trial.

East of the Mississippi River, the typical Superfund site involves a defunct and often sham solvent recycling facility, a landfill, or the contaminated grounds of a factory. The area affected is usually described in terms of acres, and liability often hinges on which entities sent wastes to the site and whether the facility's owners and operators can be found and have financial resources. Rarely is it difficult to find alternative sources of drinking water for the people living near the site because there are plentiful supplies of surface water and groundwater.

In the Far West, however, the situations are quite different at the large "regional" groundwater sites. There, the alleged contributors to the contamination often occupy facilities physically located over large plumes of contaminated groundwater, and the chemicals in that water got there for the most part through routine releases from decades of normal industrial production. The scope of the problem is measured in many square miles rather than acres, and, because water is scarce in these areas, often the remedies selected by the government include cleansing the groundwater and making it available for use — even human consumption — in nearby communities. In these areawide cases, liability hinges more on forensic hydrogeology than on review of shipping invoices. Hordes of contributors to the [29 ELR 10260] contamination problem may escape the government's attention altogether because their spills occurred decades ago, and because the high concentration "slugs" have moved well downgradient under someone else's property. Often, there are no groundwater monitoring wells on or near many possible sources, and there are changes in property use through transfer of ownerships, etc. Faced with these harsh realities, 30-odd companies named as potentially responsible parties (PRPs) by EPA in connection with the plumes of contamination underlying the eastern San Fernando Valley near Los Angeles decided to place the key liability determinations in the hands of three scientists, rather than with the federal judicial system.

The San Fernando Valley Cases

In the early 1980s, health officials detected volatile organic compounds (VOCs) — primarily trichloroethylene (TCE) and perchloroethylene — in the groundwater of the eastern San Fernando Valley, a densely populated area between the Santa Monica and San Gabriel Mountains; they promptly closed many of the community wells that had been used for drinking water purposes. In 1986, EPA placed four sites within the Valley on the national priorities list, and subsequently divided one of these sites — the Crystal Springs Site — into the Glendale North and South operable units. Not surprisingly, these two operable units are located near Glendale, California. According to EPA, the plume of contamination underlying the Valley covers an area roughly 11 miles long and up to 3 miles wide. (Although the size of the contaminated area is large, it is not unusual; there are several other contaminated groundwater sites of this magnitude in the Los Angeles Basin.)

In June 1993, EPA issued records of decisions for the two Glendale operable units that called for the construction and operation of groundwater extraction and treatment systems designed to "inhibit vertical and horizontal migration of groundwater contamination" and to "begin to remove contaminant mass from the upper zone of the aquifer." The cost of building and operating these interim remedial measures ("interim" because EPA will decide what further cleanup is required after observing the operation of this system for 12 years) will be well in excess of $ 50 million.

To say that the liability picture was murky is an understatement. The eastern San Fernando Valley area of California includes the large communities of Burbank and Glendale, which are now booming commercial and residential areas, perhaps most famous for the movie and TV studios located there. However, earlier in the century this area was the home of the aircraft manufacturing industry and two airports that were hubs of military air operations during World War II (one providing the set for the foggy departure of Ingrid Bergman in Casablanca). Within this Valley over the past six decades there are or have been major electronics producers, vehicle depots, railroad rights of ways, landing gear manufacturers, aircraft instrument makers, dozens of dry cleaners, and a host of other solvent-using operations. Some of the properties have had several owners, and the present occupant may have had nothing to do with groundwater contamination. Other properties have been demolished, paved over, or redeveloped, obliterating signs of former contaminating activities. While state regional water quality officials had ordered soil and groundwater investigations on the premises of current businesses that were possible sources of contamination, and EPA had prepared summaries of available soil, soil gas, and groundwater data for some PRPs, neither regulatory body had attempted to attribute specific liability to any PRP.

The situation was further clouded by the absence of groundwater quality data in parts of the Valley. For a variety of reasons, there are unfortunate gaps in the geographic coverage of the database. The liability of a specific facility can best be determined by the VOC concentrations immediately upgradient and downgradient as well as on the property itself. When that information is not available, one is forced to develop liability shares by extrapolating from more distant data, usually relying heavily on computer groundwater transport models. In addition, the Glendale operable units are downgradient of two other operable units in the Valley, which means that sources in those upgradient areas are possible contributors to the Glendale contamination. In particular, one of the largest companies in the area is located just north of the Glendale operable units, and it had contributed many millions of dollars to the construction of the Burbank operable unit pump-and-treat remedy that was designed in part to capture the contamination before it leaves the Burbank area. EPA designated this large company as a PRP for purposes of the Glendale operable units, and it was perceived by the Glendale parties as a major contributor to the Glendale plumes as a result of its historic operations. Yet, its attitude was that it had "given at the office" and should have only a minor stake in the Glendale cleanup. (For purposes of this Dialogue, this party will be referred to as "Company A.")

The Early Efforts to Obtain a Bargain-Basement Allocation

There were many thorny policy issues that had to be answered before any allocation could be achieved. The PRPs had struggled with these for many months, but had been unable to come close to reaching any agreements. For example, what if a PRP was located miles from the nearest proposed downgradient extraction well, yet within the overall EPA plume? If this PRP could show that little or none of its contamination would reach the extraction well during the 12-year period of the interim remedy, should it be required to pay only a minimal sum? Also, the parties debated the importance of soil and soil gas data, and the significance of "clean areas" of soil between the surface contamination and the groundwater. There were the usual questions of the validity of employing surrogates in lieu of actual data on a facility — for example, the size, years of operation, number of employees, amount of solvents used, etc. — in the absence of reliable evidence as to the quantity of chemicals released into the ground. In addition, the parties debated whether liability should be based on the mass of contamination introduced by a given facility, or the size of the respective plumes created by the individual PRPs — given that most of the remedy costs are driven by the volume of water pumped through the system regardless of the level of contamination. (A small amount of contamination released many years ago can cause a large plume.) In the end, the PRPs agreed that the costs of the remedy was driven 70 percent by volume and 30 percent by mass, and that this breakdown could be considered in the allocation of liability.

[29 ELR 10261]

SAN FERNANDO VALLEY BASIN

[SEE ILLUSTRATION IN ORIGINAL]

[29 ELR 10262]

While they were trying to determine how to allocate liability, the PRPs formed a group, elected officers, and retained a national environmental consulting firm to begin the process of designing the groundwater extraction and treatment system. Eventually, the group also retained a small firm to handle the movement of information among the group: among other tasks, it distributed documents, arranged conference calls, and maintained a master document repository. The PRP group also hired a lawyer to advise it on commercial issues, such as property acquisition (for well sites, pipelines, and the treatment plant), leases and other agreements with property owners and governmental agencies, and taxes. Finally, the PRP group formed a limited liability corporation to carry on many of the construction activities.

It seems that there is a ritual at any multiparty site that occurs early in the process: the cry for a "quick and dirty" allocation that will somehow give birth to a consensus allocation without having to retain outside help, run up legal bills, share sensitive technical studies, or evaluate certain parties' ability to pay. This certainly occurred at the Glendale site and now can be recalled with some amusement. After several months of floundering, the group decided to retain outside expertise to assist in allocation. It hired a very competent and experienced team of professionals: an environmental lawyer, a process chemist, and a hydrogeologist. They were asked to oversee a voluntary, interim allocation that would be nonbinding in any later proceeding to reallocate costs and establish final shares. To save money, this proceeding was compressed into a tight schedule, and the allocation team was not allowed to "dig into" the data, as it yearned to do. The parties responded to a series of interrogatories, and the allocation team employed a matrix of factors to develop an allocation schedule with which all parties professed to be profoundly unhappy. However, in the absence of something better, this interim allocation served for several years as the basis of group operations.

The Decision to Let Scientists Be the Arbiters

Many companies believed that the key to a fair allocation lay in proper appreciation of the available environmental data, and there was a growing clamor for a scientifically based proceeding. Several companies, including Company A, insisted that the next effort be binding, so as to ensure a permanent financial foundation for future group funding of the construction and operation of the interim remedy and to avoid another futile round of allocation debate. It took many more months of negotiation to develop acceptable procedures and schedules for a binding arbitration. Under this new approach, the allocation determination would rest on the judgment of three scientists, who would allocate liability for all sources located north of the Glendale units (the "Burbank sources") on the one hand, and all sources located within those units (the "Glendale sources") on the other. This became known as the "Burbank-Glendale split" allocation proceeding. To further simplify matters, Company A agreed to act as the stand-in for all the Burbank sources (it had arrived at side deals with most of the other sources as part of resolving the Burbank case), and the Glendale sources would be treated as a unified group. The Glendale entities believed that if the question of the role of the Burbank sources could be resolved, their respective "intra-Glendale" shares could be established quickly thereafter in a separate proceeding.

The scientific panel was designated as "sole judge of the facts and arguments presented by the [parties]." It consisted of Dr. David B. McWhorter, professor of chemical and bioresource engineering at Colorado State University; Dr. Steven M. Gorelick, professor of geological and environmental sciences at Stanford University; and Dr. David A. Stephenson, then-president and principal hydrogeologist at South Pass Resources, Inc., in Scottsdale, Arizona, and now a consultant in Jackson, Wyoming. The arbitration agreement also provided for a limited right of appellate review within the framework of the voluntary process. A retired Los Angeles County Superior Court Judge, Harry V. Peetris, was retained to play this limited appellate role. The agreement allowed this "judicial neutral" to overturn:

an action taken by the Panel [that] is (1) in conflict with this Agreement or any written procedural rules adopted by Mutual Agreement among the Allocation Participants, (2) based upon a clear error of fact, or (3) subject to reversal based on the standards set forth in California Code of Civil Procedure sec. 1286.2(a), (b) or (d) [governing procurement of award by fraud or corruption, or when the arbitrator exceeds his powers].

There continues to be a dispute as to the exact meaning of this paragraph and, in particular, the extent to which the internal appeal procedures would supplant any right to seek redress in court.

In preparation for the presentations to the arbitrators, the parties engaged in voluntary discovery, answering interrogatories about such things as chemical usage and on-site groundwater and soil analyses, and submitted all relevant technical studies to a repository where they were available to the other parties and to the arbitration panel. In addition, all the relevant EPA technical documents were sent to the repository. Because the panel was charged with resolving all issues that pertain to the allocation, such as the purpose of the remedy selected by the government, the panel was authorized to and did meet with the EPA project manager and the Agency's outside contractor in charge of the Superfund remedies for this portion of the Valley, as well as with the group's own consultant who had been involved in designing the proposed interim remedy. These "ex parte" meetings were memorialized in memoranda placed in the repository. Included in the substantial database available to the panel were computer models of the contamination in the Valley prepared by the government and by some of the parties. The panel found these models invaluable in deciphering relative causation.

The panel received several rounds of written briefs and convened two "workshop" sessions at which the parties were permitted to elaborate on the written presentations that had been submitted earlier. The briefs ranged from a mix of traditional arguments on legal and policy issues to analysis of the most highly complex groundwater modeling matters. The panel issued more interrogatories to all the participants and responded to certain presentations with highly specific requests for more information from some of the parties' consulting scientists, including computer simulations and model verification data. The workshops were attended by scores of experts and attorneys and were held in hotel meeting rooms in Los Angeles. The panel issued a proposed decision, [29 ELR 10263] reviewed further briefs in response thereto, and then issued a final decision.

The Panel's Decision

The panel's decision certainly was based on science, as the participants had bargained for; the extent to which it represented an improvement over traditional judicial decisionmaking, of course, is an opinion that is heavily influenced by which side was perceived to have done well. However, no one can disagree that the panel reviewed an enormous amount of complex information in a concentrated, intense effort and employed a sophisticated approach to solving the problem that simply would not be readily available to a federal judge because it was premised on an independent, advanced understanding of contaminant movement in groundwater, geochemistry, mathematical computer modeling, and environmental engineering. In a deceptively simple approach, the panel calculated the cost to each of the two PRP groups (Burbank versus Glendale) of capturing the plume of contamination attributable to that group alone, and then compared that cost to the sum of the two stand-alone remedial costs. (E.g., It would cost Party A $ 200 to treat just its share of the plume. It would cost Party B $ 100 to treat just its share of the plume. Thus, A's share is 200/300 or 2/3.) The allocation methodology was not tied to a site-specific remedial design because that would involve reliance on particular well locations and other details that might bring about unfair results to a party that happens to be located near an extraction well (which may have been located there for a variety of reasons entirely unrelated to that PRP's contamination — such as it is an existing well site, or is in a park and doesn't require dismantling a building or tearing up a street). The panel calculated the water volume of each of the parties' largest contaminant plume (TCE) by selecting cross-sections (width and thickness) and multiplying the representative flow through that section by the 12-year period of the interim remedy. As noted, the panel had at its disposal not only computer models generated by EPA's consultant and the group's design consultant, but extensive modeling and calibration studies prepared by certain companies participating in the arbitration. In addition, the panel had on its staff experts who were able to synthesize assumptions from different groundwater transport models. It is difficult to see how a federal court — even with the help of a court-appointed expert — could have duplicated the sophisticated fact-finding effectiveness generated by the free exchange of information between a well-staffed panel of experts and the many hydrogeologists employed by the participants. The scientists involved in the extensive debates that took place over the validity of modeling assumptions are among the leaders in the field of chemical subsurface contamination, and the free exchange of information and challenges to assumptions — presented in briefs and at the workshops — served to give the panel more confidence in the accuracy of its final product.

A Follow-Up Mediation

Having resolved the "split" between sources in Burbank and Glendale, there remained the issue of how the Glendale parties would allocate liability among themselves. While the split proceeding was underway, the "intra-Glendale" parties devised a somewhat different and more compressed nonjudicial resolution process. In a large part based on the assumption that the extensive exchange of information in the split proceeding would educate them and refine the issues, the intra-Glendale parties chose to undertake a nonbinding mediation exercise, but with some of the scientific support employed in the split proceeding.

The intra-Glendale parties executed their own allocation agreement, which relied on two mediators — one technical and one legal — to sort through the many complex "cases" presented by each party and to attempt to cajole a consensus agreement out of the many participants. Randall W. Wulff, a San Francisco attorney and experienced mediator, and Dr. Stavros S. Papadopulos, a nationally known hydrogeologist with his own Washington, D.C.-area firm, were selected to serve in these roles. As with the Burbank-Glendale split proceeding, each party was required to submit relevant information to a common repository, to certify as to the completeness of its document production, and to update the repository with any new data. The mediators were given access to all the fundamental EPA technical information, as well as the massive remedial design-related studies that had been recently prepared by the consultant hired by the PRPs. (Throughout the split proceeding and the intra-Glendale mediation, this common consultant was effectively shielded from any discussions of allocation.)

The mediation agreement established two phases. In the first, participants were allowed to submit papers of no more than 20 pages — on any subject — promoting any theory of liability or allocation. For the most part, these addressed the facility-specific situations, such as the import of soil and groundwater data; some, however, went beyond this to suggest overall allocation results. Phase I also included workshop and mediation sessions that were to take place over three or four days, with the possibility of follow-up group or individual sessions with the mediators if a settlement appeared to be within reach. If Phase I was not successful, the agreement provided that the parties would move into the next phase, which called for more detailed written submittals, another workshop, and, finally, the issuance of a written recommended allocation report prepared by the two mediators. The agreement stated that the recommended allocation could become the nucleus of a binding settlement among some of the parties if at least 85 percent of the shares of liability recommended in that report agreed to that result. After intense negotiations following Phase I, the intra-Glendale parties reached a global settlement, and the procedures contemplated by Phase II were never tested.

A Closing Observation

The most important element in the Glendale case, or in any similar Superfund alternative dispute resolution (ADR) proceeding for that matter, is not the sophistication or open-mindedness of the participants, or even the skill of the arbitrator or mediator. Rather, it is the role that the government is willing to assume, and specifically the approach taken by EPA and the U.S. Department of Justice staff assigned to the site. The federal regulatory representatives must give ADR a fighting chance, and to do that they must communicate this message clearly and often to the PRPs: "The group-sponsored ADR process is the only game in town; we will not entertain a separate deal with you unless we are convinced [29 ELR 10264] that the group ADR effort will be a failure." In the Glendale case, many companies and individuals over the years were convinced that the split arbitration and the intra-Glendale mediation would generate allocation shares that would be unacceptable to them. They hoped to find a "kinder and gentler" EPA that might listen more sympathetically to their portrayal of their relative roles. A number of them made confidential appeals for individual "cashouts" with EPA. Had even one or two of these parties been able to obtain individual settlements with the government, it would probably have doomed the Glendale ADR process and made all-out litigation war a virtual certainty.