Chemical Weapons Working Group, Inc. v. United States Department of the Army

MEMORANDUM DECISION AND ORDER CAMPBELL, District Judge. This matter is before the court on plaintiffs’ second motion for preliminary injunctive relief to enjoin defendants from incinerating chemical warfare agent at the Tooele Chemical Agent Disposal Facility (TOCDF). Plaintiffs claim that “new” evidence, that is, evidence discovered after the conclusion of hearings on plaintiffs’ first motion for preliminary injunctive relief, demonstrates that continued incineration of agent at TOCDF poses a threat of irreparable harm. Plaintiffs also contend that the new evidence requires that defendants prepare a supplemental environmental impact statement (SEIS). The new evidence presented by plaintiffs falls into two general categories: (1) operation of TOCDF and (2) stack emissions. Procedural Background Plaintiffs initiated this suit on May 10, 1996. The amended complaint alleges that defendants have violated the National Environmental Policy Act (NEPA), the Resource Conservation and Recovery Act (RCRA), the Toxic Substances Control Act (TSCA), the Defense Authorization Act, and the Clean Water Act (CWA) and that defendants’ operation of TOCDF constitutes a nuisance under Utah law. The court granted defendants’ motions to dismiss the RCRA CWA and nuisance counts. Subsequently, plaintiffs initiated several proceedings before the Utah Solid and Hazardous Waste Control Board (“Utah Board”) to challenge various issues and decisions regarding TOCDF’s hazardous waste permits. These parallel proceedings are, to date, ongoing. Plaintiffs’ first motion for preliminary injunctive relief (“first motion”), filed on June 12, 1996, sought to enjoin defendants from commencing trial bums of chemical warfare agent at TOCDF. After a nine-day evidentiary hearing, this motion was denied on August 13, 1996. See Chemical Weapons Working Group, Inc. v. Department of the Army, 935 F.Supp. 1206 (D.Utah 1996) (“CWWG I”). On October 11, 1996, plaintiffs filed a notice of appeal of the court’s denial of their first motion and dismissal of various claims alleged in the first amended complaint. Seven days later, on October 18, 1996, plaintiffs moved the United States Court of Appeals for the Tenth Circuit to stay TOCDF operations pending resolution of their appeal. Because plaintiffs had failed to first seek a stay in the district court, the Tenth Circuit denied plaintiffs’ motion on December 6, 1996. Chemical Weapons Working Group (CWWG) v. Department of the Army, 101 F.3d 1360 (10th Cir.1996). The Tenth Circuit did not address the merits of plaintiffs’ appeal. On January 11, 1997, plaintiffs filed a consolidated motion for a stay and second motion for preliminary injunction. A hearing on plaintiffs’ consolidated motion was held over six days from March 3, 1997 through March 10, 1996. Having considered the evidence presented at that hearing, the memoranda filed by the parties, and the arguments presented by counsel, the court denies plaintiffs’ second motion for a preliminary injunction 1 and enters the following findings of fact and conclusions of law. FINDINGS OF FACT Background 1. In its previous Memorandum Decision and Order, the court made detailed factual findings concerning the physical facility at TOCDF, the nature of the chemical warfare agent stockpile stored at Deseret Chemical *1086 Depot 2 (“Depot”), the Army’s nearly twenty-years of experience with large-scale incineration of agent materials, and the compliance process dictated by NEPA. CWWG I, 935 F.Supp. at 1209-14. The court will not repeat its prior findings here except as necessary to explain the pending consolidated motion. TOCDF Operations 2. On August 22, 1996, TOCDF began destroying chemical agent pursuant to Trial Burn Plans approved by the Utah Division of Solid and Hazardous Waste. 3 Two of TOCDF’s five furnaces became operational— the Deactivation Furnace System (DFS) and the first of two Liquid Incinerators (LIC-1). The DFS is used to incinerate munitions which, after being drained of agent, remain contaminated. Agent drained from munitions and ton containers is destroyed in the LIC. As of February 4, 1997, the DFS had functioned for more than 569 hours and the LIC for over 736 hours, resulting in the destruction of 11,472 rockets and 122,750 pounds of the nerve agent GB. This amount represents thirty-eight percent of the stockpile of GB-filled rockets stored at the Depot. 3. On January 17, 1997, TOCDF began processing ton containers 4 stored at the Depot in the Metal Parts Furnace (MPF) and GB nerve agent drained from those containers in the second LIC (LIC-2). As of February 4, 1997, fifty-one such ton containers and 76,500 pounds of GB had been destroyed. 4. All present activities at TOCDF are governed by the Trial Burn Plans. TOCDF is currently operating in the “shakedown” phase, a period designed to identify possible mechanical difficulties, ensure that the facility has reached operational readiness, and achieve steady-state operating conditions pri- or to conducting the trial burns. 5. A munitions processing schedule governs the order in which the various munitions stored at the Depot are to be destroyed at TOCDF. Each portion of the schedule, termed a “campaign,” is devoted to the disposal of a specific item in the stockpile inventory. Defendants submitted evidence in the form of declaration testimony by Gary J. Boyd, author of the quantitative risk assessment for TOCDF, that since the court’s decision in CWWG I, the munitions processing campaigns have been reordered to provide for the destruction of higher risk munitions earlier in the processing schedule. Mr. Boyd testified that accidents involving nerve agent GB represent the majority of the risk from potential stockpile accidents; accordingly, the first campaign in the reordered munitions processing schedule provides for the disposal of GB-filled rockets and ton containers. The reorganization of the munitions processing campaigns will reduce the stockpile risks much more rapidly than would have occurred under the original processing schedule. 6. During the shakedown period, three events have occurred which have caused defendants to halt operation of TOCDF: (1) detection of low levels of agent in two filter containment vestibules; (2) leakage of a small quantity of decontamination fluid through hairline cracks in a second level cement floor to a first floor electrical room; and (3) migration of agent into an observation corridor. In addition, TOCDF has experienced other operational and personnel difficulties. Citing these events and difficulties, plaintiffs maintain that TOCDF’s present-day operation is substantially different from that contemplated during the NEPA compliance process and presents a risk of irreparable harm. Specifically, plaintiffs contend that while the “lessons learned” program and the operation of prototype facilities at the Depot (CAMDAS) and at Johnson Atoll (JA-CADS) were designed to identify, analyze, and correct problems of this type prior to the commencement of agent operations at *1087 TOCDF, the facility is being operated in a reactive, trial-and-error manner. Agent Migration Into Filter Vestibules 7. The primary means of preventing an airborne agent release to the environment or the spread of agent vapor within TOCDF is the Heating, Ventilation, and Air Conditioning (HVAC) system. The HVAC maintains negative pressure throughout the facility so that air from areas least likely to be contaminated with agent flows to areas where contamination is more likely. Air from areas in which contamination is most likely is steadily removed and directed through a bank of filters designed to extract agent. Of the nine filter units in the system, seven are typically in use at any given time with the remaining two units serving as reserves. Air passing through the filters is funneled into a common exhaust stack and monitored for the presence of agent. Additional agent monitors are located between the charcoal banks that comprise the filter units. 8. The HVAC filters are enclosed within a metal structure containing sealed access doors. In March 1995, it was discovered that the gaskets surrounding similar access doors at JACADS allowed small amounts of agent to escape to the environment. It was subsequently determined that different door gaskets and clamping mechanisms would prevent future agent releases. It was also determined that secondary containment structures, not part of the original design plans for either JACADS or TOCDF, should be constructed over the access doors at both facilities. These structures, called “vestibules,” were constructed at both facilities in 1996. 9. The filter vestibules are pre-fabrieated wooden structures. The interiors of the vestibules are modified so that each is lined with 22-gauge stainless steel sheeting and caulked with silicon; however, the structures are not designed to be airtight. In August 1996, the vestibules were monitored for the presence of agent through the use of Depot Area Air Monitoring System (DAAMS) units. DAAMS units collect air contents onto an absorbent material inside a testing tube over an extended period of time, typieally eight hours. The absorbent material is then analyzed for the presence of agent. The other monitoring system used at TOCDF, Automatic Continuous Air Monitoring System (ACAMS), is a self-contained chromatograph used to determine agent concentrations on a near real-time basis. 10. On August 23, 1996, GB nerve agent migrated into the filter vestibules attached to filter units 107 and 108. The release was detected and confirmed the next day during laboratory analysis of absorption tubes removed from DAAMS units in both vestibules. The testing revealed that between 8:00 a.m. and 4:00 p.m. on August 23, 1996, agent was present in vestibule 108 at a level of 3.62 Time Weighted Average (TWA) 5 and in vestibule 107 at a level of .25 TWA. During this eight hour period, filter unit 107 was in operation for approximately nine minutes; filter unit 108 was off-line for the entire day. On the afternoon of August 23, 1996, three TOCDF employees were present in the vestibule 107 for approximately twenty minutes. Filter 107 was off-line at the time. Medical tests of the personnel indicated that they suffered no agent exposure. 11. The plant shift manager was notified of the confirmed agent levels in the filter containment vestibules at approximately 2:32 p.m. on August 24, 1996. Processing of M55 rockets was immediately suspended. An announcement made over the public address system declared the area in and around the filter units off-limits to all but essential personnel, a 200-foot clear zone was established and cordoned off with barricade tape, and ACAMS agent detection systems were placed outside the filter vestibules and at the plant boundaries. None of these ACAMS units registered the presence of chemical agent. In addition, any personnel entering the clear zone were required to wear protective masks. Richard Holmes, Associate Project Manager for TOCDF, testified that he could not recall whether a masking alarm sounded to warn TOCDF personnel in the area to don their *1088 masks; however, the evidence suggests that the alarm was not sounded. Finally, the ambient air present in the filter vestibules was drained back into TOCDF. 12. An investigation of the incident revealed that the most probable cause of the agent migration into the filter containment vestibules was the procedure used to place the filter units off-line, a phenomenon not encountered at JACADS. After consultation with the Executive Secretary of the Utah Board, defendants altered the procedure to ensure that negative pressure was maintained within the filter units at all times. Defendants also installed additional filtered valves in all of the filter vestibules. Finally, defendants permanently installed ACAMS agent detection systems in the vestibules. All of these actions were taken before agent operations resumed on August 30,1996. Cracks in the Concrete 13. On September 18, 1996, TOCDF experienced a leak of approximately eight ounces of decontamination liquid through the ceiling of an electrical room located on the first floor of the Munitions Demilitarization Building from a decontamination area located above. Upon learning of the leak, the plant manager halted agent processing activities. Tests detected no chemical agent present in the liquid, and no electrical equipment was affected. An inspection of the decontamination room, used by TOCDF personnel to decontaminate their protective equipment after exiting Explosive Containment Rooms (ECRs), revealed several hairline cracks in the concrete floor which allowed fluid to leak into the electrical room. Mr. Holmes testified that the cracks were sealed by an injection of a low viscous grout. 14. Cracks in concrete have occurred in other areas at TOCDF. Donald Smith, EG & G’s Senior Quality Assurance Specialist at TOCDF, testified that he noticed cracks as early as 1990 during construction of the facility. However, Timothy Thomas, the Army’s TOCDF Project Manager, testified that cracks in the concrete identified during construction of the facility were repaired. John Russell Hall, an Engineering Technician for EG & G at TOCDF from February 6,1994 to January 4, 1996, testified that he noticed cracks in the concrete floors of the Toxic Maintenance Area (TMA) and the Residue Handling Area (RHA) in 1995. Mr. Hall testified that he completed a work order for the cracks in the TMA on November 22, 1995, but he was unable to recall whether these cracks or those located in the RHA were repaired by the time he left EG & G’s employ. In fact, Mr. Hall testified in his deposition that he had no knowledge of any corrective action taken by defendants to address cracks in the concrete at TOCDF since his employment was terminated. Further, the court notes that Mr. Hall’s tenure at TOCDF was during systemization, a phase of operations designed to identify and correct problems in TOCDF’s physical plant and equipment. 15. Defendants have established corrective measures to identify, map, evaluate, and repair cracks in the concrete at TOCDF. Cracks are identified through routine and scheduled inspections of TOCDF’s structural integrity by EG & G personnel. When a crack is identified, a work order is prepared and the crack is evaluated to determine its nature and the appropriate method of repair. Minor and hairline cracks in the concrete or the floor coating are repaired by applying a filling compound resistant to decontamination fluid and re-coating the area with chemical-resistant epoxy paint. Major cracks are repaired in a similar manner after being injected with a bonding compound. When cracks are identified in a toxic area, agent operations in that area are suspended until the crack is fixed. Thomas A Kurkjy, EG & G’s Risk Management Division Director, testified that hundreds of minor cracks and three major cracks — cracks having a discernible void — have been identified in the concrete floor and floor coating at TOCDF and have been repaired. Agent Migration Into Observation Corridors 16. On January 26,1997, GB nerve agent vapors migrated into an unoccupied observation corridor adjacent to the first floor buffer storage area. An ACAMS alarm in the corridor sounded, and TOCDF personnel donned their protective masks and evacuated the building. In addition, TOCDF operations were halted pending an investigation *1089 into the alarm. The ACAMS monitoring system indicated that agent had been present in the observation corridor at a level of 1.04 TWA Agent was not released to the environment and no TOCDF employees were exposed to agent. The Army notified the appropriate regulatory authorities of the event, and agent operations resumed only after state approval was received. 17. An investigation revealed that the event was triggered when, at approximately 10:30 p.m. on January 25, 1997, an interior door between an air lock and the DFS room was opened during maintenance operations. The opening of the door caused the air pressure in nearby Toxic Cubicle 6 to rise slightly, which, in turn, activated an alarm in TOCDF’s control room. A control room operator responded by opening a toxic cubicle bypass damper to lower the air pressure in the Cubicle, an action that violated TOCDF’s standard operating procedures. Opening the damper had a secondary effect of causing a pressure imbalance between the observation corridor and the buffer storage area. As a result, when a drained GB ton container was moved through the storage area en route to the MPF, GB vapor was permitted to migrate into the corridor. 18. Defendants have taken corrective measures to prevent such an event from recurring: operator procedures and system changes have been implemented, control room operators have received additional training on the secondary effects of opening bypass dampers, TOCDF’s standard operating procedures have been modified to emphasize the appropriate use of the toxic cubicle bypass damper, and additional air pressure alarms have been installed in TOCDF’s control room. Other Operational Events at TOCDF 19. TOCDF has experienced additional operational events during the shakedown period. These include: the failure of heating elements in the slag removal system in LIC-1; incidents during rocket processing initialization, loss of electrical power, temporary HVAC imbalance during a test of the fire suppression system, malfunction of the agent quantification system, and use of a “hot cutout” procedure to remove TOCDF personnel from their protective clothing. Slag Removal System Operation 20. The incineration of liquid agent produces acidic by-products which condense on the walls of the secondary combustion chamber of the LICs to form a molten slag. This slag slowly flows down the walls and collects in a pool at the bottom of the chamber. The slag removal system employs sixteen heating elements to maintain the slag in a molten state so that it may be drained from the LICS. During agent operations in LIC-1, several of the elements failed, requiring defendants to halt processing in LIC-1 temporarily so that the faulty elements could be replaced. Defendants intend to modify the slag removal system by fitting protective sleeves around the heating elements to increase their useful life expectancy. Incidents During Rocket Processing Initialization 21. On October 14,1996, an end cap from the rear of an M55 rocket shipping and firing tube was inadvertently removed by a feed gate designed to allow rockets to pass into an ECR for disassembly. Operation of the affected rocket processing line was halted temporarily to allow TOCDF personnel to examine the tube and the rocket. Once it was determined that the rocket was intact and stable, the employees secured the end cap to the tube and processing resumed. Rocket processing procedures have since been modified to minimize the likelihood that such an event will occur in the future. 22. On two occasions — once in November 1996 and once in December 1996 — rocket parts have jammed in the chute feeding into the DFS. After both incidents, operations at TOCDF were halted to allow the jams to be cleared and an evaluation to be conducted. An investigation revealed that a build-up of heated materials on the feed chute prevented sheared rockets from being fed properly into the DFS. Feed chute jams experienced at JACADS had a different root cause, one not observed at TOCDF. To reduce the likelihood of future jams at TOCDF, defendants *1090 have heightened inspection of the chute area and have modified the chute to allow for ready access, should a jam occur. Loss of Electrical Power 23. In mid-September 1996, TOCDF experienced a loss of commercial electrical power for 38 minutes. TOCDF’s emergency power supply activated and provided power to essential equipment in the plant, including the HVAC system. During the restart of the HVAC system, only one of the two normally operating air supply handlers came on line, causing an imbalance in the HVAC system air pressure. The masking alarm sounded and notification was given to the Deseret Chemical Depot Emergency Operations Center. Within minutes of the power loss, the second air supply handler was started manually and the masking signal was withdrawn. No chemical agent migrated from the facility and no TOCDF personnel were exposed to agent. 24. Power failures are not uncommon at TOCDF. However, there has never been an occasion when the backup power system failed to properly activate upon loss of power. Because ACAMS systems operate on an independent and uninterrupted power supply, their operation is not affected by a loss of commercial power. Temporary HVAC Imbalance During Fire Suppression System Test 25. On September 2, 1996, a temporary imbalance in the HVAC system occurred in the Unpack Area (UPA) during a test of the fire suppression system. In conducting the test, maintenance personnel shut off the water supply to the sprinkler system. This action triggered an abnormal water pressure alarm and caused the internal fire dampers in the UPA ventilation system to close automatically. This closure caused a temporary pressure imbalance in the UPA HVAC system. Negative pressure was maintained throughout TOCDF during the event and the UPA HVAC system was stabilized quickly. No agent migrated from primary containment areas and employee safety was maintained. TOCDF maintenance personnel have since received corrective training to ensure that such an incident does not recur. Agent Quantification System Operation 26. The Agent Quantification System (AQS) is designed to measure the amount of agent drained from munitions. The measurement is made not for security purposes but to determine the amount of residual agent being fed into the DFS. This calculation is required by TOCDF’s RCRA permits. During initial agent operations it was discovered that the AQS was improperly indicating the presence of agent in rockets that had been completely drained. An investigation revealed that the AQS allowed a small quantity of agent to flow into the agent holding tank before being measured. Proper operation of the AQS was restored by the installation of a metal plate vertically from the top of the AQS downward into the tank, and similar problems have not recurred. Hot Cut-Outs 27. TOCDF personnel working in contaminated areas are required to wear demilitarization protective ensemble (DPE) suits. These plastic suits are completely sealed from the outside environment and must be physically cut to be removed from employees. When the cut-out procedure is performed in an area where agent concentration exceeds 1 TWA, the exit is termed a “hot” cut-out. Plaintiffs argue that the hot cut-out procedure exposes TOCDF employees to chemical agent. In support of their claim, plaintiffs presented documentary evidence that the number of hot cut-outs performed at TOCDF increased steadily from ten in September 1996 to a high of fifty in November 1996. The increased frequency of hot cut-outs was due to two factors. First, practices in place at the time attempted to minimize the amount of decontamination liquid utilized outside the ECRs. Second, the nature of the work performed during those months required more frequent handling of equipment bearing liquid agent. While the hot cut-out procedure requires employees to exit the DPE suit in the presence of agent, GB nerve agent is primarily an inhalation hazard and every worker is equipped with an independent breathing apparatus. Because TOCDF will process only GB nerve agent for at least one year, the court finds that the hot cut-out *1091 procedure does not present a threat to employee safety prior to trial. 28. None of the events cited by plaintiffs or other operational difficulties experienced at TOCDF have resulted in loss of life, injury to TOCDF personnel, or harm to the environment. James J. Cudahy, an expert in the evaluation, design, operation, and permitting of hazardous waste incineration facilities, testified that the number of safety related incidents at TOCDF is not unusually high when compared to typical start-ups of modem complex systems for hazardous waste incineration. TOCDF Management 29. Plaintiffs argue that defendants’ management of TOCDF does not ensure protection of public health and the environment. In support of their assertions, plaintiffs presented documentary evidence and testimony from former TOCDF employees Gary Millar, John Hall, and James DeHaven, and current employee Donald Smith. 30. Plaintiffs rely heavily on a November 9, 1996 letter from Gary Millar, former General Manager at TOCDF, to Fred Parks, President of EG & G, written shortly after Mr. Millar’s employment with EG & G was terminated. In the letter, Mr. Millar raised several issues about TOCDF operations and management. Mr. Millar indicated that on the date agent operations began at TOCDF, August 22,1996, the facility was at a marginally acceptable state of safety readiness. The letter also speaks of numerous safety, quality, environmental, and operational deficiencies which, in Mr. Millar’s opinion, are excessive in a “high risk business like TOCDF.” In addition, Mr. Millar was highly critical of management actions at TOCDF which he analogized to the those preceding the nuclear accident at Three Mile Island and the Challenger Disaster and of a corporate “mindset” which, according to Mr. Millar, presents a high risk to TOCDF employees, the public, and the environment. 31. Mr. Millar’s testimony under oath belies many of the concerns raised in the November 9, 1996 letter. On December 12, 1996, Mr. Millar testified to the Utah Board that TOCDF was being operated safely and that state regulatory agencies charged with overseeing the facility were doing a “good job” keeping TOCDF operations and the public safe. Mr. Millar further testified that he never intended his letter to become public and that he considered the issues raised therein to concern EG & G’s internal management, not plant safety. Mr. Millar’s deposition testimony echos his testimony before the Utah Board. Mr. Millar acknowledged that when he wrote the letter, he was upset about his recent firing and that he did not intend the letter to be disseminated to the Army, the State of Utah, or the public. Further, Mr. Millar testified that he believed TOCDF to be “inherently safe” and that he did not consider the risks described in the November 9,1996 letter to be so serious that he was required to disclose them to state regulatory - authorities. The court finds Mr. Millar’s testimony more credible and more probative than the contents of his November 9,1996 letter. 32. Plaintiffs also presented evidence in the form of deposition and live testimony from John Hall. Mr. Hall testified that he had noticed and reported various problems at the facility, including cracks in concrete flooring and leaks of sulfuric acid from batteries powering the emergency power system. However, Mr. Hall also testified that his employment with EG & G ended on January 4, 1996 and that he has no knowledge of conditions or operations at TOCDF since that date. Accordingly, insofar as Mr. Hall’s testimony relates to events occurring since he left EG & G’s employ or the potential for future problems at the facility, the court finds his testimony to be of little probative value. 33. Plaintiffs also presented evidence in the form of testimony by James DeHaven, an emergency medical technician employed at the Tooele Health Clinic from October 1, 1996 to February 14, 1997. Mr. DeHaven testified that electrocardiograms (EKGs) he administered'to several Depot personnel revealed a higher than normal incidence of bradycardia (slow heart rate) and “blocks” (interruptions of the electrical pathways to the heart), both symptoms of nerve agent exposure. Mr. DeHaven further testified that his medical supervisors, Army person *1092 nel, and representatives from the Utah Department of Health ignored his concerns. On cross-examination, Mr. DeHaven acknowledged that he did not know the prior medical histories of those employees exhibiting bradycardia or blocking, how long they had exhibited either condition, possible other causes of the symptoms, how long the workers had been employed at the Depot, or whether they worked at TOCDF or some other area of the Depot. 7 Mr. DeHaven also testified that all EKGs administered at the Tooele Health Clinic were transmitted to an Army hospital in Texas to be read by a cardiac specialist. Mr. DeHaven was not aware of any instance in which an EKG administered at the Depot had been deemed abnormal by a cardiac specialist. Accordingly, the court finds that Mr. DeHaven’s testimony did not constitute evidence that TOCDF personnel had suffered nerve agent exposure. 34. Plaintiffs also relied on the testimony of Donald Smith and on entries contained in private journals in which Mr. Smith recorded concerns and frustrations he experienced as EG & G’s Senior Quality Assurance Program Development Coordinator at TOCDF. Throughout his • testimony, Mr. Smith made clear that his journals did not represent his professional work product, were intended to be private, and were often used to- “[vent] emotional feelings at the time.” Hearing Tr. at 76 (March 6, 1997). Mr. Smith also testified that the latter portions of his journal were written while he was being treated with heavy medication. Finally, Mr. Smith testified that many, if not most, of his journal entries relating to TOCDF were based not on his personal knowledge, but on hearsay evidence. Accordingly, the court finds Mr. Smith’s testimony to be of little probative value. 35. Finally, plaintiffs attempted to establish that Mr. Millar and Mr. Hall were terminated from their employment with EG & G in retaliation for raising concerns regarding safety at TOCDF. Defendants presented evidence in the form of Timothy Thomas’ supplemental declaration that Mr. Millar’s firing was predicated on his inability to effectively communicate his management priorities and instructions and on his management approach, which caused stress in other EG & G personnel. In his deposition testimony, Mr. Millar admitted being told that he was causing morale problems among the workforce at TOCDF and that several EG & G managers had filed written complaints about him. Defendants also presented Mr. Hall’s deposition testimony in which he conceded that he had no direct evidence that his firing was retaliatory in nature. The court finds that plaintiffs have produced insufficient evidence to establish that either Mr. Millar or Mr. Hall were terminated from their employment with EG & G for raising concerns regarding safety at TOCDF. Quantitative Risk Assessment 36. The Army’s Program Manager for Chemical Demilitarization, Major General Robert D. Orton, has directed that a quantitative risk assessment (QRA) and risk management program be developed for each of the eight planned chemical demilitarization facilities in the continental United States, including TOCDF. The TOCDF QRA estimates the probabilities and public health consequences of potential accidental releases of chemical agent during chemical storage and disposal activities. Releases resulting from internal initiating events (those originating inside the facility or directly from the activity being performed) and from external events (e.g. earthquakes, aircraft accidents, and tornadoes) were included. In addition, the TOCDF QRA assesses the public risk associated with storage of the chemical munitions at the Depot absent demilitarization operations. The TOCDF QRA is based on the “as-built TOCDF design” and incorporates data derived irom JACADS. 37. At the time of the court’s decision in CWWGI, the QRA for TOCDF was in draft form. Since that time, the final QRA for TOCDF has been issued. The final QRA reflects several changes from the draft version. First, as discussed above, the muni *1093 tions processing campaigns have been reordered so that higher risk munitions are destroyed earlier in the processing schedule. Second, the expected duration of processing operations at TOCDF has increased from 6.2 to 7.1 years. This change reflects the final QRA’s reliance on the now de-classified actual processing schedule for TOCDF rather than on an estimated schedule, as had the draft QRA Third, the final QRA incorporates a more realistic model for measuring risks associated with agent spills in storage area igloos. The draft QRA assumed that all igloo spills would completely exit the igloos and be subject to outside evaporation rates. The final QRA assumes that smaller spills not having sufficient volume to reach the doorway will be confined to the interior of the igloo. Fourth, the final QRA more realistically predicts the number of M55 rocket igloos that would explode during an earthquake. The draft QRA assumed that the explosion of one igloo would trigger all of the remaining igloos to explode. The current model predicts the number of igloos that would explode in earthquakes of varying sizes. 38. The final QRA concludes that, on average, 34 days of continued storage of the stockpile incurs a public risk equal to that associated with the entire 7.1 years of TOCDF agent operations. If rare events such as earthquakes and aircraft accidents are removed from the assessment, the finding is stronger — the risk to the public from the entirety of TOCDF’s operations is equaled by the risk of only 2.3 days of continued storage. 8 The final QRA also concludes that a one year delay in processing will approximately double the risk to the population surrounding the stockpile. Stack Emissions 39. Plaintiffs allege that “the health risk from the on-going daily stack emissions of toxic chemicals [at TOCDF] including both dioxin and nerve agent is both more real and quantitatively greater than previously disclosed.” Plaintiffs’ Consolidated Memorandum in Support of Plaintiffs’ Motion for an Injunction Pending Appeal and Plaintiffs’ Second Motion for Preliminary Injunction, at 3. Plaintiffs base this claim on “new” evidence that: (1) nerve agent GB is being emitted from the stacks at TOCDF; and (2) the Utah Department of Environmental Quality (DEQ) improperly manipulated the screening health risk asséssment (SRA) performed for TOCDF by reducing estimates of mustard gas emissions, deleting risk calculations associated with open burning and open detonation, and omitting risk scenarios based on dioxin exposure to breast-fed infants and subsistence farmers. GB Emissions 40. Stack effluent gasses at TOCDF are regularly monitored for a number of analytes, including GB, by ACAMS and DAAMS monitoring systems installed in the stacks. Stack samples are analyzed, and the results are forwarded to state regulatory authorities. 41. Plaintiffs argue that analyses of stack particulate emissions at TOCDF indicate the presence of nerve agent GB in the stack effluent. Plaintiffs also allege that defendants’ testing and analysis methodologies underestimate the amount of nerve agent actually escaping the stack and overestimate the agent destruction and removal efficiencies calculated for the DFS. In support of their claims, plaintiffs presented evidence in the form of declaration and live testimony by Pat Costner and a document prepared by EG & G’s subcontractor, Battelle, which reports non-zero readings for nerve agent GB in stack emissions at TOCDF. The court finds that the evidence presented by plaintiffs is insufficient to demonstrate conclusively that nerve agent GB is being emitted from the stacks at TOCDF. All of the positive results cited in the Battelle document were below the level of quantification (LOQ), that is, the sensitivity or calibration range, of the monitoring equipment, a level approximately equivalent to a GB stack concentration over 5,500 times less than the maximum allowable regulatory-based GB stack concentration. Values below the LOQ have a lower confidence that the quantity of GB detected is *1094 accurate and could reflect machine “noise,” an interferant, or a false positive. The court finds that Ms. Costner’s testimony is only marginally probative in that much of the scientific data underlying her opinions was not directly applicable to the TOCDF facility. 42. Plaintiffs attempt to analogize the potential health effects of alleged GB emissions with illnesses suffered by veterans of the Persian Gulf War. In lieu of written or live testimony, the parties introduced documentary evidence, mainly in the form of articles from periodicals. The court finds that because the etiology of Gulf War illness is not known with any degree of reasonable certainty, the evidence submitted on this subject is not probative of the issues raised by plaintiffs’ consolidated motion. Screening Health Risk Assessment 43. Prior to approving trial burns of chemical agent at TOCDF, DEQ performed an SRA which analyzed the impacts of expected TOCDF emissions on human health and the environment. The SRA followed Environmental Protection Agency (EPA) guidance in adopting conservative assumptions. 44. In CWWG I, the court discussed in detail the risk assessment performed for TOCDF. 935 F.Supp. at 1213-14. After carefully considering the health risks associated with dioxin exposure and DEQ’s decision to eliminate from its February 1996 SRA risk scenarios regarding a subsistence farmer and a breast-feeding infant, the court found that: [although plaintiffs have shown that the assumptions applied in the State’s health risk assessment may indicate a higher level of risk for some hypothetical persons, this does not constitute a showing that there is an actual risk to some person or persons posed by the emissions levels predicted for the facility. Id. at 1214. None of the new evidence presented by plaintiffs undermines the court’s prior finding. 45. As they did with their first motion, plaintiffs rely heavily on a draft chapter of the EPA’s “Health Assessment Document for 2, 3, 7, 8, Tetrachlorodibenzo-p-Dioxin (TCDD) and Related Compounds” (Dioxin Reassessment). However, this document, which by its terms is not to be cited or quoted, remains subject to review by EPA’s Science Advisory Board and possible public comment and does not represent the EPA’s final position. The court finds that scientific knowledge regarding health risks associated with dioxin exposure and the methods to assess the health impacts of dioxin emissions at TOCDF remains unchanged since the previous hearing. 46. In CWWG I, the court addressed the omission of the subsistence farmer and breast-feeding infant scenarios from the SRA. 935 F.Supp. at 1213-14. The court noted that the final version of the SRA “considered three farmer scenarios based on a survey of actual farming practices in the area, and simply deleted the breast-feeding infant scenario.” Id. at 1214. Plaintiffs argue that the final SRA erroneously omitted consumption of local dairy products from the risk calculus. In support of this contention, plaintiffs presented evidence in the form of a written summary of farming activity surrounding TOCDF drafted by Rachel Shilton, an engineer in DEQ’s Division of Solid and Hazardous Waste, and an unofficial table prepared by former EG & G Permitting Manager Gary Harris which, according to plaintiffs, demonstrate that such dairy consumption is, in fact, occurring. The court finds this evidence to be unreliable. While the table appears to identify a family of loeal dairy consumers, no names or other identifying information is provided. The summary indicates, at most, that local dairy production may have occurred in the past. The court finds more probative Ms. Shilton’s testimony that she was unable to locate any persons in the area of TOCDF who presently consume locally produced dairy products. 47. Plaintiffs allege that risk calculations for open burning and open detonation (OB/OD) of chemical weapons at the Depot were improperly omitted from the SRA However, plaintiffs presented no evidence that OB/OD operations were occurring or would occur within the next year. Ms. Shilton testified in her declaration that DEQ will allow open burning to occur at the Depot *1095 only if agent operations at TOCDF have ceased or if emissions from OB/OD could be modeled with the other emissions considered in the SRA to produce a cancer risk no higher than 10 per million. Plaintiffs presented no evidence that either condition has occurred. 48. The SRA models stack emissions for four stacks: the combined stack for the LICs, DFS, and MPF incinerators, the HVACS (ventilation) stack, the Dunnage Incinerator stack, and the CAMDAS stack. Early screening calculations for risks associated with mustard agent (HD) stack emissions assumed that all four stacks would emit both [¶] and GB continuously at the minimum concentration that would trigger a waste-feed cutoff under TOCDF’s operating permits. When these early calculations showed a high risk level for HD, the model was revised to assume that [¶] emissions from the HVACS stack would be at the detection level, 20% of the waste-feed cutoff level. The court finds that this change reflects a more realistic approximation of expected HVACS stack emissions. The court also finds that there is no evidence that [¶] will be processed at TOCDF within the next year. Indeed, according to the revised munitions processing schedule, TOCDF will process only GB-filled M55 rockets and ton containers for at least one year. CONCLUSION OF LAW 1. Plaintiffs bear the burden of establishing that they are entitled to injunctive relief. To meet their burden, plaintiffs must establish: (1) that they will be irreparably injured unless an injunction issues; (2) that the threatened injury outweighs any damage defendants might suffer; (3) that the injunction, if issued, is not adverse to public interest; and (4) that they have shown a substantial probability of success on the merits. Walmer v. U.S. Department of Defense, 52 F.3d 851, 854 (10th Cir.), cert. denied, — U.S. —, 116 S.Ct. 474, 133 L.Ed.2d 403 (1995). If plaintiffs establish the first three requirements for a preliminary injunction to issue, they may establish likelihood of success by showing “questions going to the merits so serious, substantial, difficult and doubtful, as to make the issues ripe for litigation and deserving of more deliberate investigation.” Id. (citing City of Chanute v. Kansas Gas & Elec. Co., 754 F.2d 310, 314 (10th Cir.1985)). Irreparable Harm 2. To constitute irreparable harm, an injury must be certain, great, and actual. Wisconsin Gas Co. v. FERC, 758 F.2d 669, 674 (D.C.Cir.1985). Irreparable harm cannot be speculative; “the injury complained of [must be] of such imminence that there is a ‘clear and present’ need for equitable relief to prevent irreparable harm.” Id. (emphasis in original) (citations and internal quotations omitted); see also Regan v. Vinick & Young, 862 F.2d 896, 902 (1st Cir.1988) (“[speculation or unsubstantiated fears about what may happen in the future cannot provide the basis for a preliminary injunction”). Plaintiffs’ claims of irreparable injury relate to: (1) the alleged public health consequences of the operation of, and emissions from, TOCDF; and (2) defendants’ failure to prepare a supplemental environmental impact statement (SEIS), which, according to plaintiffs, constitutes a violation of NEPA. Operational Risks 3. Plaintiffs assert that the occurrence of several incidents at TOCDF since agent operations began demonstrates that continued operation of the facility poses immediate risks to TOCDF employees, the public, and the environment. The overall record of operations at TOCDF does not support plaintiffs’ claim. Although there have been problems at the facility, some of which required the suspension of operations, none of the events caused harm to TOCDF personnel, the public, or the environment. There is no evidence that human injury or environmental harm is inevitable or likely. In fact, the record suggests that TOCDF’s safety equipment and procedures are effective in preventing such harms. 4. The evidence presented through John Hall, Donald Smith, and James DeHaven does not undermine this conclusion as it lacks sufficient probative value to be of merit. Likewise, in light of Gary Millar’s testimony *1096 in his deposition and before the Utah Board that TOCDF was being operated in a safe manner, his November 9, 1996 letter cannot serve as a basis for finding that TOCDF operations pose a risk of irreparable harm. 5. The problems experienced at TOCDF do not demonstrate that the “lessons learned” program is a failure. To the contrary, the evidence indicates that events experienced at JACADS have not recurred at TOCDF and the root causes of incidents at TOCDF were not observed at JACADS. 6. The fact remains that all of the events at TOCDF occurred during the shakedown period, a phase designed to identify and correct operational difficulties prior to full-scale operations. As testified to by defendants’ expert, James Cudahy, such events are to be expected during shakedown operations for any large-scale hazardous waste incinerator. The court finds that the occurrences cited by plaintiffs are too speculative to support a finding of irreparable harm. Risks from Stack Emissions 7. Plaintiffs argue that the presence of nerve agent GB in TOCDF’s stack effluent constitutes a direct and present threat to public safety and the environment. However, there has never been a confirmed detection of agent in the stack emissions from TOCDF since agent operations began. Nonzero values for GB reported in the stack particulate analysis relied upon by plaintiffs were well-below the level of quantification of the monitoring equipment, a level that is itself more than five thousand times less than the maximum GB stack concentration permitted by the regulatory scheme. No significant degree of scientific confidence can be placed in the results of the particulate analysis; indeed, the evidence indicates that the positive readings for GB could have benign origins such as machine noise or false positives. The court finds that the asserted risks from emissions of GB from the stacks at TOCDF are too speculative to qualify as irreparable harm to plaintiffs. 8. Plaintiffs’ allegations regarding mustard agent emissions cannot support a finding of irreparable harm. The evidence reflects that mustard agent will not be processed at TOCDF before trial on the merits. Thus, plaintiffs will suffer no injury justifying preliminary injunctive relief. Further, while the final SRA prepared by DEQ utilized a less conservative model for mustard agent stack emissions than did earlier assessments, the court finds that the revision more accurately reflects the actual operating conditions at TOCDF. 9. Plaintiffs’ allegation that the SRA underestimates the risks associated with dioxin exposure is not tantamount to irreparable harm. The elimination of the breast-feeding infant and subsistence farmer scenarios is consistent with EPA guidance for facilities like TOCDF. Further, there is simply no reliable evidence that either scenario applies to the areas surrounding the facility. At most, plaintiffs have shown that the assumptions applied in the SRA may indicate a higher level of risk for some hypothetical persons, not that there exists an actual risk to actual persons from projected emissions levels. 10. The SRA’s omission of risk calculations for open burning/open detonation at TOCDF does not support a finding of irreparable harm. Neither activity currently occurs at TOCDF, and plaintiffs produced no evidence that OB/OD operations would commence before trial. The court finds that the asserted risks of harm due to dioxin exposure are too speculative to constitute irreparable harm to plaintiffs. 11. Having carefully considered all of these factors, the court concludes that neither the plaintiffs nor the public will suffer irreparable harm from TOCDF emissions. NEPA Harm 12. The purpose of NEPA is to focus “government and public attention on the environmental effects of proposed agency action.” Marsh v. Oregon Natural Resources Council, 490 U.S. 360, 371, 109 S.Ct. 1851, 1858, 104 L.Ed.2d 377 (1989). To this end, NEPA requires federal agencies to consider all information prior to taking an action which might have significant environmental effects. Id. Plaintiffs contend that defendants’ failure to supplement the 1989 environmental impact statement (EIS) violates NEPA and constitutes irreparable harm. As *1097 discussed more folly in the court’s, examination of plaintiffs’ likelihood of success on the merits, plaintiffs’ “new” information did not require the preparation of a supplemental environmental impact statement. However, even if defendants’ decision not to generate a SEIS did violate NEPA, “pending final resolution of this case, such injury will occur during only a small portion of the expected operating lifetime of TOCDF, and is therefore relatively minimal.” CWWGI, 935 F.Supp. at 1216. Balancing of Harms — Public Interest 13. In this case, the interests of both plaintiffs and defendants coincide with different elements of public interest. Plaintiffs assert that the public interest requires a suspension of operations at TOCDF pending trial; defendants assert that the public interest lies in disposing of the stockpile of lethal chemical agent and munitions stored at the Depot. It is true that halting agent operations at TOCDF could have negative consequences for defendants, including a loss of proficiency in operations and a risk to employees during decontamination of “hot” portions of the facility. However, these harms are best considered in the analysis of where the public interest lies. Indeed, the public has an interest in the safe and efficient operation of TOCDF, and TOCDF personnel are, of course, members of the public. 14. In CWWGI, the court found that the risks of continued storage outweigh the risks of operation of TOCDF during the period before trial. 935 F.Supp. at 1216-17. This conclusion has been strengthened by changes made by defendants in the munitions processing schedule. During the approximately one-year period before trial, the Army will continue to process the volatile GB nerve agent, the source of the majority of the risk from potential stockpile accidents. Reorganizing the munitions processing campaigns to destroy higher risk munitions earlier in the schedule, starting with GB, will reduce the overall stockpile risks faster than would have occurred under the previous schedule. Further, the risks from potential accidental releases of chemical agent are minimized by allowing agent disposal activities at TOCDF to continue. Gary Boyd, author of the QRA for TOCDF, concluded that as of February 1997, when the bulk of GB in the stockpile remained unprocessed, a one-year delay in agent operations would approximately double the risk to the population surrounding TOCDF. 15. There is no general presumption that a NEPA violation will in all cases outweigh other public interests. See Fund for Animals, Inc. v. Lujan, 962 F.2d 1391, 1400 (9th Cir.1992); Concerned Citizens v. Secretan/ of Transportation, 641 F.2d 1, 7-8 (1st Cir.1981); Southern Utah Wilderness Alliance v. Thompson, 811 F.Supp. 635, 641 (D.Utah 1993). Here, even if defendants’ failure to prepare a SEIS violated NEPA, this harm is outweighed by the harm to the public by allowing an injunction to issue. 16. Congress has mandated that the nation’s stockpile of lethal chemical weapons be destroyed and has designated the U.S. Department of the Army to carry out this directive. Pub.L. 99-145 (codified as amended at 50 Ú.S.C. § 1521 (1996)). The deadline for destruction of the stockpile, originally set for 1994, has been extended to December 31, 2004, less than eight years from now. Pub.L. 102-484 (1993). This Congressional mandate is further evidence of the public’s interest in the prompt disposal of the stockpile. 17. Having carefully considered all of these factors, the court concludes that the harms balance in favor of defendants and that the public interest is best served by the continued destruction of chemical agent at TOCDF. Likelihood of Success on the Merits 18. A federal agency’s duties under NEPA do not end when an initial decision is made or when an EIS is prepared. Marsh, 490 U.S. at 371-72, 109 S.Ct. at 1858-59. Rather, there are circumstances which require an agency to supplement an EIS. According to regulations promulgated by the Council on Environmental Quality, an EIS must be supplemented if an “agency makes substantial changes in a [project] that are relevant to environmental concerns; or ... [if] there are significant new circumstances or information relevant to environmental con *1098 ceras and bearing on the [project] or its impacts.” 40 C.F.R. § 1502.9(c)(l)(i) & (ii) (1996); see also Marsh, 490 U.S. at 374, 109 S.Ct. at 1859 (“[i]f there remains major Federal action to occur, and if the new information is sufficient to show that the remaining action will affec[t] the quality of the human environment in a significant manner or to a significant extent not already considered, a supplemental EIS must be prepared” (citations and internal quotations omitted)). Not all new information requires a SEIS; information must be “significant” to trigger the need for supplementation. See Wisconsin v. Weinberger, 745 F.2d 412, 420 (7th Cir.1984) (duty to supplement EIS not triggered simply because information is “worthy of further inquiry or may be considered important research”). 19. The court’s review of the Army’s decision not to prepare a SEIS is narrow. Marsh, 490 U.S. at 378, 109 S.Ct. at 1861-62. “[S]o long as the [Army’s] decision not to supplement the [EIS] was not ‘arbitrary and capricious,’ it should not be set aside.” Id. at 377, 109 S.Ct. at 1861. Because the question is a factual one that implicates the agency’s technical expertise, the court must defer to the agency’s informed discretion. Id. Such deference is not automatic. In this, as in every case involving an agency’s decision to not prepare a SEIS, the court must review the record and satisfy itself “that the agency has made a reasoned decision based on its evaluation of the significance — or lack of significance — of the new information.” Id. at 378, 109 S.Ct. at 1861. 20. The new evidence which plaintiffs claim mandates a SEIS is not significant information. The operational problems encountered at TOCDF caused no injury to workers or the environment, were quickly remedied, and were the type of events common to the start up of a complex industrial facility. The evidence from the EG & G employees, when closely examined, has little probative value. The evidence regarding the dangers from emissions from the stacks is either not “new,” having been previously considered by the court, or is speculative and of little merit. In short, the evidence presented by plaintiffs does not present “a seriously different picture of the likely environmental consequences of [TOCDF].” Weinberger, 745 F.2d at 420. Accordingly, the court finds that the Army’s decision not to prepare a SEIS was not arbitrary and capricious. Conclusion 21. In light of the above analysis, the court finds that plaintiffs have failed to show that they or the public will be irreparably harmed during the pendency of this action. The court also finds that the public interest favors continued operation of TOCDF and that plaintiffs have failed to show a sufficient likelihood of success on the merits to support a preliminary injunction. Plaintiffs’ second motion for a preliminary injunction is therefore DENIED. ' Footnotes [1] The court entered judgment on plaintiffs’ motion for a stay pending appeal by a separate Order. [2] Formerly known as the Tooele Army Depot. [3] TOCDF’s RCRA Part "B” permit specifies distinct sets of conditions for the "long term” and “short term” operation of the facility. The short term period is comprised of three phases— "shakedown,” "trial bum,” and "post-trial bum." The trial bum plans cover the shakedown and trial bum phases only. [4] Ton containers are large steel bulk storage containers. There were approximately 5,709 GB-filled ton containers in the original stockpile. [5] TWA is an agent concentration exposure measurement established by the Office of the Surgeon General. An average individual can be exposed to one TWA for eight hours per day, forty hours per week, for his or her lifetime without suffering adverse health effects. [6] The Toxic Cubicle houses the liquid agent storage tank. [7] Later testimony established that the medical clinic at TOCDF is separate from the Tooele Health Clinic. [8] This is so because earthquakes dominate the risks from disposal more than those associated with storage.