Notice Concerning the Intent to Grant an Exemption to Parke-Davis Division, Warner-Lambert Company for the Continued Injection of Hazardous Waste Subject to the Land Disposal Restrictions of the Hazardous and Solid Waste Amendments of 1984 (HSWA)

Agency: Environmental Protection Agency (EPA)

Action: Notice of Intent to Grant an Exemption to the Parke-Davis Division, Warner-Lambert Company of Holland, Michigan for the Continued Injection of Certain Hazardous Wastes

Summary: The United States Environmental Protection Agency (USEPA or Agency) is proposing to grant an exemption from the ban on disposal of hazardous wastes through injection wells to the Parke-Davis Division, Warner-Lambert Company (Parke-Davis), Holland, Michigan. Parke-Davis may therefore continue to inject Resource Conservation and Recovery Act (RCRA) regulated hazardous wastes, if the exemption is granted. Parke-Davis submitted a petition to the USEPA under Title 40 of the Code of Federal Regulations (40 CFR) Part 148, which allows any person to petition the Administrator to determine whether its continued injection of certain hazardous wastes is protective of human health and the environment. After a comprehensive review of all material submitted, the USEPA has determined, with a reasonable degree of certainty, that Parke-Davis’s injected wastes will not migrate out of the injection zone over the next 10,000 years.

Date: The USEPA is requesting public comments on today’s proposed decision. Comments will be accepted until March 27, 1998. Comments post-marked after the close of the comment period will be stamped "Late". A public hearing will be scheduled for this proposed action if warranted by public interest and notice will be given in a local paper and to all people on a mailing list developed by the Underground Injection Control (UIC) Program. If you wish to be notified of the date and location of a public hearing (if one is held), please contact the person listed below.

Addresses: Submit written comments by mail to:
United States Environmental Protection Agency, Region 5
Underground Injection Control Branch (WU-16J)
77 W. Jackson Boulevard
Chicago, Illinois 60604-3590
Attn: Rebecca L. Harvey

For Further Information: Contact Stephen Roy, Lead Petition Reviewer, UIC Branch, telephone (312) 886-6556, electronic mail roy.stephen@epamail.epa.gov.

Supplementary Information:

  1. I. Background
    1. Authority The Hazardous and Solid Waste Amendments of 1984 (HSWA), enacted on November 8, 1984, impose substational responsibilities on those who handle hazardous waste. The amendments prohibit the continued land disposal of untreated hazardous waste beyond specified dates, unless the Administrator determines that the prohibition is not required to protect human health and the environment for as long as the waste remains hazardous (RCRA Section 3004(d)(1), (e)(1), (f)(2), (g)(5)). The statute specifically defined land disposal to include any placement of hazardous waste in an injection well (RCRA Section 3004(k)). After the effective date of prohibition, hazardous waste can only be injected under two circumstances:
      1. When the waste has been treated in accordance with the requirements of 40 CFR Part 268 pursuant to Section 3004(m) of RCRA (the USEPA has adopted the same treatment standards for injected wastes in 40 CFR Part 148, Subpart B); or
      2. When the owner/operator has demonstrated that there will be "No Migration" of hazardous constituents from the injection zone for as long as the waste remains hazardous. Applicants seeking an exemption from the ban must demonstrate either:
        1. That the waste undergoes a chemical transformation so as to no longer pose a threat to human health and the environment; or
        2. That fluid flow is such that injected fluids would not migrate vertically upward out of the injection zone or to a point of discharge in a period of 10,000 years by use of mathematical models (40 CFR 148.20(a).
        The USEPA promulgated final regulations on July 26, 1988, (53 F.R. 28118) which govern the submission of petitions for exemption from the injection prohibition (40 CFR Part 148). A time frame of 10,000 years was specified for the demonstration not because migration after that time is of no concern but because a demonstration which can meet a 10,000 year time frame will likely provide containment for a substantially longer time period and also allow time for geochemical transformations which would render the waste nonhazardous or immobile. The Agency’s standard thus does not imply that leakage will occur at some time after 10,000 years; rather it is a showing that leakage will not occur within that time frame.
    2. Facility Operation and Process The Parke-Davis facility in Holland, Michigan (see location map, Figure 1) produces bulk pharmaceutical chemicals, both finished products and intermediates for further pharmaceuticals to be processed and/or packaged elsewhere. The products are manufactured by reaction or blending of pure chemical components in batch operations, which result in a wide variation in waste stream composition. The chemical wastes produced at this facility are treated in various ways: a major portion is recovered and reused or collected and sent to an offsite hazardous waste disposal facility. The liquid wastes generated as a result of equipment and production area washdowns and from chemical processing are collected and pumped to the chemical wastewater treatment facilities. This system consists of large tanks used for equalization and neutralization, a rotary vacuum filter and a flat and frame polishing filter for removal of suspended solids. The treated filtrate is injected down the three deep injection wells. (The filter cakes are sent to an off-site permitted hazardous waste disposal facility.) Total annual flow in recent years has been approximately 41 million gallons. Percent organic material varies between 1.0 and 2.0 per cent by weight.
    3. Waste Minimization RCRA emphasizes the preeminence of source reduction and recycling as a strategy for managing solid waste. There are four major components of waste minimization: 1) inventory management and improved operations, 2) modification of equipment, 3) production process changes, and 4) recycling and reuse. The Parke-Davis facility has implemented a waste minimization program which includes specific actions toward source reduction, recovery and recycling, and waste treatment. As a result of these efforts, for example, the amount of methanol injected in the deep wells during 1995 was 20% less than in 1994, despite increased in production levels and was reduced a further 17% in 1996, again with increased production levels. In addition, waste minimization efforts are now part of the product process development to reduce waste even before products are manufactured.
    4. Submission Parke-Davis submitted a petition for exemption from the land disposal restrictions on hazardous waste injection under the HSWA Amendments to RCRA pursuant to the regulations set for at 40 CFR Part 148 on June 10, 1991. This submission was reviewed and revised documentation was submitted March 1992, April 1996, December 1997 and February 1998. The total submission was reviewed by staff at the USEPA and by consultants hired by the Agency to assist in its determination.
  2. II. Basis for Determination
    1. Waste Description for Analysis (148.22) The wastes to be injected are generated from the production of pharmaceuticals, both finished products and intermediates produced at the Parke-Davis facility in Holland, Michigan. These wastes are also characteristically ignitable and corrosive. In addition, the exemption includes all hazardous materials currently managed at this facility as a protection against inadvertent injection of a banned substance. The specific waste codes from Part 261 included in this exemption are listed below.
      D001 D002 D004 D005 D006 D007 D008 D009 D010 D011 D018 D035 D038 F005 P095 U002 U003 U019 U080 U112 U151 U154 U159 U196 U213 U220 U239
    2. Well Construction and Operation (148.22) The construction of all three Parke-Davis wells is very similar and consists of three casing strings within surface conductor pipe. All casing is cemented over the entire length from the casing shoe to the ground surface to preclude potential avenues for injected fluid to escape the injection zone. (See Figure 2.) Injection takes place through tubing set on a packer and is isolated from the innermost casing by a fluid-filled annulus. The pressure in this annulus is monitored continuously. The monitoring system is designed to trigger an alarm and shut down the injection pumps if the injection pressure or annulus pressure exceeds the maximum permitted levels or if the annulus pressure falls below the minimum permitted level. Injection pressure is limited by permit to 1000 pounds per square inch gauge (psig), which is below the value calculated using the equation in 40 CFR 147.1153. Average flow is approximately 3.4 million gallons per month.
    3. Mechanical Integrity Test Information To assure that the waste does not leak out of the casing prior to reaching the injection zone, mechanical integrity tests (MITs) of the wells are required. 40 CFR Section 148.20(a)(2)(iv) requires submission of satisfactory MIT results within one year of petition submission. The test consisted of a radioactive tracer survey (RTS) and an annulus pressure test on each well. MITs have been run annually on the Parke-Davis wells, most recently in October 1997. Results of these tests demonstrated that the wells have mechanical integrity and confirmed the positive results recorded on continuous monitoring equipment. From both construction and operation standpoints the Parke-Davis injection wells ensure transmission of the injected fluid to the injection zone without leakage with a reasonable degree of certainty.
    4. Site Description As part of the "no migration" demonstration under Part 148, Subpart C, any Class I hazardous waste injection well petitioner must identify the strata within the injection zone which will confine fluid movement within the injection interval and the strata which act as a confining zone. In evaluating the confinement properties of these strata and the geologic suitability of the site for hazardous waste injection, the USEPA used the standards set forth in 40 CFR Part 146. All three Parke-Davis wells have approximately 4800 feet of separation between the lowermost underground source of drinking water and the top of the injection zone. This separation zone is composed of shales, limestones, dolomites and evaporites.
      1. Regional Geology The Michigan Basin, the large scale structure into which the Parke-Davis wells inject, is an almost circular sedimentary basin which extends to a depth of 14,000 feet in the center of the basin. The geologic strata generally dip less than 1 degree toward the center of the basin. The sedimentary formations can be classified into four general sequences: 1) the Cambrian sandstone sequence, 2) the Ordovician to middle Devonian carbonate-evaporite sequence, 3) the late Devonian to Mississippian shale-sandstone sequence, and 4) the Pennsylvanian coal bearing sequence. The Cambrian sandstone sequence (the Munising Formation), the lowest sedimentary unit in the basin, has the most favorable properties for the disposal of liquid waste because of its high permeability and porosity. There are very few historical seismic occurrences in the Michigan Basin; the nearest to the Parke-Davis site was a magnitude 3.0 event which occurred in 1906 and was epicentered 22 miles from the site. The most severe event within 100 miles occurred in 1883 with an estimated magnitude of 5.0, epicentered 43 miles from the site. Experience with Alaskan earthquakes has shown that damage to injection wells is generally limited to surface facilities.
      2. Injection Zone Description The injection zone must have sufficient permeability, porosity, thickness and areal extent to prevent migration of fluids out of the injection zone. The injection zone is divided into two parts: the lower part is the injection interval, into which the fluid is directly emplaced, sometimes called the emplacement interval, and the upper part is the arrestment interval, in which upward movement of the injected fluid (due to diffusion, for example) is contained. The injection interval at Parke-Davis consists of the Mt. Simon Sandstone, which occurs at a depth of approximately 5080 feet and is approximately 740 feet thick at the Parke-Davis site. The average permeability is 198 millidarcies (md), the average porosity is 14 percent. The arrestment interval is composed of the EauClaire, Dresbach, Franconia Members of the Munising Formation and the Trempeleau Formation, of which the Eau Claire Member is most important because it immediately overlies the injection interval. The Eau Claire Member is composed of shale, sandstone and dolomite. Based on core measurements, its porosity averages 13.2 percent and its horizontal and vertical permeabilities average 0.4 md and 0.008 md, respectively, due to the presence of shales and clays. The entire arrestment interval is approximately 630 ft thick. All the formations occuring at the site are generally laterally continuous across the entire Michigan Basin.
      3. Confining Zone Description The confining zone must be (1) laterally continuous, (2) free of transecting, transmissive faults and fractures over an area sufficient to prevent fluid movement and (3) of sufficient thickness and lithologic and stress characteristics to prevent vertical propagation of fractures. The confining zone is made up of the Prairie du Chien, Trenton and Black River Groups and the Utica Shale. As noted above, all formations occuring at the site are generally laterally continuous across the entire Michigan Basin. No complex geological features such as faults or extensive fractures are known to exist within the area of review and no indication of boundary effects has been seen in any well test results. The entire thickness of the confining zone at the site is over 1100 ft. The thin shale beds and argillaceous (clayey) nature of the dolomite beds should be adequate to prevent the vertical propagation of fractures. This is particularly true of the Utica Shale, which is approximately 180 ft thick in this area.
      4. Geochemical Conditions The characteristics of the injection and confining zone fluids and lithologies must be adequately described in order to determine the wastestream’s compatibility with the zones. The injection zone is composed mainly of sandstone, with minor amounts of dolomite, siltstone and shale. These rocks are generally very resistant to chemical degradation and therefore little, if any, compatibility problem is expected and none was shown in core tests. Tests run to determine compatibility between the injected fluid and the formation fluids in the injection interval showed no signs of precipitation. Perhaps of most significance, no compatibility problems have shown up in the years these wells have operated. Parke-Davis has injected approximately 675 million gallons of wastewater during the period from 1975 to 1997, and during this time, no adverse pressure response has occurred to indicate that the wells have suffered from a compatibility problem. The confining zone is composed of a variety of rock types including sandstone, dolomite, shale and limestone and should have little compatibility problems with the injected fluid. No problems were revealed during testing with core samples.
      5. Area of Review The area of review (AOR) is the area within which the petitional must identify all wells which penetrate the confining zone and demonstrate whether they have been properly completed or plugged and abandoned. For the Parke-Davis facility, the USEPA has designated an AOR consisting of a circle of 2.0 mile radius centered between the three Parke-Davis wells with a 3.8 mile extension to the southwest (S30.5?W) based on a flow model which shows that the waste plume will move this distance in 10,000 years. There are nine Class I injection wells in the AOR: BASF has 3 Class I wells,- Well #1 was properly plugged in 1980. BASF Well #2 and Well #3 were properly plugged in 1995. Parke-Davis and Heinz each have three properly constructed Class I wells. In addition to the nine existing Class I injection wells, there are 13 other wells within the 2-mile radius, none of which penetrates the injection zone. Therefore, no corrective action is required at this facility.
    5. Model Demonstration of No Migration Over the past quarter century, mathematical modeling has emerged as the preeminent tool for the predictive analysis of hydrogeologic systems. It is appropriate then that the demonstration of no migration of hazardous constituents from the injection zone involve the use of predictive mathematical models. "Injection Forecast", the model used in this demonstration of no migration, was developed by REC. It is a collection of analytical models which is capable of simulating a variety of complex hydrogeological conditions, such as multiple wells with individual rate histories, multi-layer aquifers, leaky shale formations, plume migration over 10,000 years and molecular diffusion into overlying formations (the confining zone). Because Injection Forecast is a collection of analytical models, validation consists of verifying that the computer model correctly evaluates the equations comprising the model. This has been done for many subcomponents of the program. In addition, comparison of predicted and observed values for a variety of different situations has provided support for this model. Although a simulation model cannot prove anything, it is the best tool available to test hypotheses against all available data in a much more rigorous way than can be done by any other method.
      1. Model Development and Calibration Assignment of values for parameters used in the modeling is a critical task. The input to Injection Forecast was developed by incorporating hydrogeological data of the site and surrounding area into a conceptual model. These values were derived from well logs, cores, published literature and well tests. The model includes values for the Mt. Simon Sandstone (the Parke-Davis injection interval), the Eau Claire Member (the Parke-Davis arrestment interval) and the Dresbach Sandstone (used by other injection wells in the AOR but not by Parke-Davis). Initial calibration was performed by matching simulated results to results of an interference test conducted in 1989. This shows that the important parameters, taken as a group, adequately simulate the pressure behavior of the injection zone. The parameters for the Mt. Simon Sandstone included a permeability of 198 md and a porosity of 14%. Values for the Eau Claire Member were estimated from logs and conservative values chosen: permeability of 0.1 and 0.01 md (for sensitivity analysis) and porosity of 12% were used.
      2. Model Predictions Two simulation periods were investigated: the operational life of the well (estimated to end in 2009) and a post-operational period of 10,000 years. For the operational period, pressure buildup in the injection interval, waste plume configuration and vertical penetration were calculated. For the post-operational period, pressure decrease, waste plume lateral movement due to natural hydraulic gradient, gravity drift and dispersion and vertical penetration due to diffusion were calculated. Modeling results and the parameter choices which ensure that these results represent reasonably conservative conditions are presented below.

        For the operational period, actual injection rate histories were used through January 1991 (when the calculations were performed). Constant projected injection rates were used from that time through February 2009: an average monthly injection rate of 100 gpm for each Parke-Davis well (more than twice the average rates of the Parke-Davis wells), 75 gpm for each of the two BASF wells (which were plugged in 1995) and 90 gpm for each of the three Heinz wells (more than 150% of recent rates). The use of rates higher than actual provides a conservative cushion to the demonstration by causing an over-prediction of modeled pressure build-up and waste migration.

        1. Operational Lifetime Pressure Buildup Analysis: Projected pressure increase in the injection zone is critical to the issue of confinement because this increase drives permeation of fluid out of the injection zone into the immediately overlying portion of the arrestment interval and could, in principle, cause fluid movement up an improperly abandoned wellbore. The maximum projected pressure rise will occur at the end of the projected operational life of the wells in 2009 at the wellbore of Well No. 4. This rise is approximately 83 psi, a 3.5% increase over the pre-injection pressure in the injection zone, estimated to have been 2386 psi. The pressure rise decreases radially away from the cluster of wells.
        2. Operational Lifetime Waste Plume Configuration: Using an effective thickness of 325 ft for the Mt. Simon Sandstone, and the injection histories described above and ignoring dispersion, the maximum radial distance of waste from propagation is approximately 1750 ft. Dispersion will cause the waste front to be irregularly "smeared" from a zone of 100% waste to 0% waste. Isopropanol was used to calculate maximum increase in waste plume size because it is the constituent which undergoes the greatest reduction from concentration in the wastestream to the health-based concentration. For isopropanol, the maximum radial distance to the dispersed waste front is 3180 ft from Well No. 3.
        3. Operational Lifetime Vertical Migration: The vertical penetration out of the injection interval into the immediately overlying arrestment interval is most sensitive to the pressure rise history of the site since it is this pressure which provides the driving force for this process. Since this pressure is greatest near the injection wells and decreases radially away from the wells, the depth of penetration will also follow this pattern. Using a value of 0.1 md for the vertical permeability of the arrestment interval, the maximum vertical penetration is calculated to be 9.3 ft. The thickness of the arrestment interval is over 600 ft. Using 0.1 md for the vertical permeability is very conservative, since measured vertical permeability at the site ranged from 1.7 x 10-5 to 4.0 x 10-3 md. Dispersion will also affect vertical penetration and is estimated to be 17 ft. Thus the maximum vertical penetration is 26.3 ft.
        4. Ten Thousand Year Pressure Decrease: After the projected shut-in of the Parke-Davis wells in 2009, pressure above original will decline logarithmically to approximately 10% of its value at shut-in within 10 years. In 10,000 years, there will be no remaining pressure build-up in the Mt. Simon due to injection into the Parke-Davis disposal wells.
        5. Ten Thousand Year Vertical Fluid Penetration: Because vertical fluid penetration into the arrestment interval is driven by pressure increases in the injection zone, it actually decreases after injection is terminated. During the 10,000 year period essentially all of the fluid which penetrates the arrestment interval will flow back into the injection interval.
        6. Ten Thousand Year Waste Plume Movement: During the injection period, waste plume movement is dominated by the injection process and dispersion as the waste front moves through the rock. Once injection has ceased, different processes become dominant. They include the natural hydraulic gradient within the injection interval, gravity-induced movement due to differences in density between the injected fluid and the in-situ fluids, and additional dispersion as the plume continues to move.

          Determining the natural hydraulic gradient in deep aquifers is difficult because of the lack of data. The natural flow in the Mt. Simon Sandstone appears to be on the order of 0.5 to 1.4 ft/yr to the southwest. Thus in 10,000 years, the fluids injected by Parke- Davis will move 5,000 ft to 14,000 ft.

          Calculation of the flow driven by density differences used 1.16 for the specific gravity of the in-situ Mt. Simon fluid and 1.0 for the injection fluid. Since the actual specific gravity of the Parke-Davis injection fluid has varied between 1.0 and 1.2; using 1.0 for the calculation maximizes this effect and therefore is conservative. The plume will flow an additional 8949 ft in an up-dip direction (approximately southwest) during the 10,000 years following the cessation of injection due to density differences.

          As mentioned above, dispersion will cause further displacement of the waste material. Using a value of 20 ft for the coefficients of lateral and transverse dispersion, the calculated effect of dispersion will be an additional movement of 5394 ft.

          The total distance from the center of the injection system during 10,000 years will be the sum of these effects: 1750 ft (undispersed lifetime distance) + 14000 ft (hydraulic gradient) + 8949 ft (density differences) + 5394 ft (cumulative diffusion) = 30093 ft = 5.7 miles. The plume will be elongated in the direction of hydraulic drift and density- difference driven flow; the width will be approximately 3180 ft as calculated above.

          Note that a great deal of conservatism has been included in these calculations: as the plume moves and spreads, the concentration will continually decrease but the calculation has assumed that the concentration remains constant. In addition, mixing with waste from Heinz or BASF injection will also lower concentrations of hazardous constituents.

        7. Ten Thousand Year Vertical Diffusion of Injected Materials: Calculation of this value includes the following factors: concentration of each particular constituent of the waste (taken as approximately twice the observed maximum concentrations to be conservative), the health-based limit for each constituent, the diffusion coefficient for each constituent; the porosity of the arrestment interval (12% as worst case) and the injection interval (14%); the relative tortuosity of the arrestment interval and the injection interval (taken as 1.0 as worst case), and the time period for diffusion (10,000 years). Vertical diffusion was calculated using two different methods which had different initial assumptions. The results agreed within a few percent, indicating that vertical diffusion is not sensitive to these assumptions. The greatest distance of vertical diffusion was calculated for methanol and was 158 ft. Neglecting the effect described in paragraph E above to be conservative, the total vertical penetration will be 167 ft into the arrestment interval. Since the arrestment interval is over 600 ft thick, this distance is more than 433 ft below the top of the arrestment interval.
    6. Quality Assurance and Quality Control Parke-Davis and its consultants have demonstrated that adequate quality assurance and quality control plans were followed in preparing the petition. Parke-Davis has followed appropriate protocol for locating records for penetrations in the area of review, for collection and analyses of geologic and hydrogeologic data, for waste characterization and for all tasks associated with the modeling demonstration.
  3. III. Conditions of Petition Approval As a condition of granting this proposed exemption from the ban on injection of hazardous wastes, the USEPA requires that the following conditions be met by Parke-Davis:
    1. (1)The monthly average injection rate is limited to 100 gallons per minute per well.
    2. (2)Injection shall occur only into the Mt. Simon Sandstone Member of the Munising Formation between the depths of 5080 and 6027 feet.
    Date: ___________________________________

    ________________________________________
    Jo Lynn Traub
    Director, Water Division
    Region 5, U.S. Environmental Protection Agency

    Figure 1: Site Map General Location Map of Lower Peninsula of Michigan Figure 2: Well Diagram All three wells are very similar in construction, though exact depths vary slightly. The numbers shown here refer to well #3.