D.C. No.4:08-cv-01372-SBA Appeal from the United States District Court for the Northern District of California Saundra B. Armstrong, District Judge, Presiding
The opinion of the court was delivered by: M. Smith, Circuit Judge:
Argued and Submitted January 11, 2012-San Francisco, California
Before: John T. Noonan, Jr., and Milan D. Smith, Jr., Circuit Judges, and Jed S. Rakoff, Senior District Judge.*fn1
Opinion by Judge Milan D. Smith, Jr.
This case arises out of Plaintiffs-Appellants Tri-Valley CAREs', Marylia Turner's, and Janis Kate Turner's (collectively, Tri-Valley CAREs) second challenge to the sufficiency of the United States Department of Energy's (DOE) Environmental Assessment (EA) of a prospective "biosafety level-3" (BSL-3) facility at the Lawrence Livermore National Laboratory (LLNL). In an earlier round of litigation, we upheld all aspects of the DOE's original EA, except for its failure to consider the impact of a possible terrorist attack. Following our remand, on September 30, 2009, the district court entered summary judgment in the DOE's favor on the grounds that it had sufficiently revised its Final Revised Environmental Assessment (FREA) to adequately consider the environmental impact of an intentional terrorist attack on the BSL-3 facility at LLNL. On November 18, 2010, Tri-Valley CAREs timely appealed the district court's decision, petitioning us to require the DOE to prepare an Environmental Impact Statement (EIS), or, in the alternative, to revise its EA, in light of the allegations set forth in its original complaint, to determine whether an EIS is required.
We hold that the DOE took the requisite "hard look" at the environmental impact of an intentional terrorist attack in the manner required by the National Environmental Policy Act (NEPA) and San Luis Obispo Mothers for Peace v. Nuclear Regulatory Commission, 635 F.3d 1109 (9th Cir. 2011). We further hold that the district court did not abuse its discretion in denying Tri-Valley CAREs' motion to supplement the record. Accordingly, we affirm.
I. FACTUAL AND PROCEDURAL BACKGROUND
A. The LLNL Biosafety Level-Three Facility
On December 16, 2002, the National Nuclear Security Administration (NNSA), an agency within the DOE, authorized the construction of a BSL-3 laboratory at LLNL. Center for Disease Control (CDC) and National Institutes of Health (NIH) guidelines divide biosafety lab operations into four levels: BSL-1 (least hazardous) through BSL-4 (most hazardous). BSL-3 laboratories work with agents that may cause diseases in humans with serious or lethal consequences if untreated, and which have the potential of airborne transmission. Common agents found in BSL-3 facilities include West Nile virus, yellow fever virus, Mycobaterium tuberculosis, and SARS virus. There are more than 1,350 BSL-3 laboratories in the United States. Common examples of BSL-3 facilities include hospital surgical suites, laboratories associated with medical schools, and university research laboratories. At the time of construction, the LLNL BSL-3 facility was the only BSL-3 facility operating in the same facility as a nuclear laboratory.
The DOE decided to undertake the construction of an on-site BSL-3 facility at LLNL because limitations in its BSL-1 and 2 laboratories forced LLNL to conduct its BSL-3 research off-site. This off-site research was difficult and costly because LLNL lacked physical control, and shipping and handling increased the risk of cross-contamination and degradation.
For the new BSL-3 facility, LLNL selected a 1,500 square-foot, prefabricated building to be constructed next to existing BSL-2 facilities. The air-handling system comprised a double High Efficiency Particulate Air-Purifying (HEPA) air filtration system, as is consistent with CDC guidelines. Each HEPA filter removes at least 99.97 percent of bioagents. The facility also has additional safeguards against pathogenic breach, including a ventilation system that would draw the contaminated air back into the facility, a backup power system that would enable employees to shut down portions of the contaminated part of the facility, and a "zone-tight" system that would prevent any air flow in the facility in the event of a total power loss.
B. The 2002 Original Environmental Assessment
The DOE performed an EA for the proposed LLNL BSL-3 laboratory, pursuant to NEPA. The EA considered the environmental impacts of the BSL-3 laboratory on a wide range of issues, including human health, ecological resources, transportation, waste management, geology, soils and seismology, noise, and air quality. The EA also discussed how CDC and NIH guidelines govern the facility's operations and mitigate the risk of infection and accidental release.
In evaluating the public risk potentially caused by the BSL-3 facility, the DOE relied upon three major sources of data: (1) statistics from hundreds of other CDC-registered BSL-3 laboratories; (2) the U.S. Army's Biological Defense Research Program (BDRP) laboratories; and (3) LLNL's BSL-1 and -2 laboratories. In addition to examining the normal operations of the aforementioned sources, the EA analyzed potential abnormal impacts on those sources, using a "catastrophic release" scenario, modeled upon a "Maximum Credible Event" (MCE), simulating the outer bounds of impact caused by a pathogen's accidental release.
The DOE considered numerous possible methods of assessing the threat of release, but it chose a catastrophic release simulation (a centrifuge analysis), that the Army used to perform a NEPA analysis of its own biological research labs. The catastrophic release model used by the Army also was an MCE type of analysis, which simulated a reasonably foresee- able event with a low likelihood of occurrence, but with high risk. In the Army's simulated catastrophic release model, a liter of coxiella burnetii (C. burnetii)*fn2 was hypothetically divided into six centrifuge tubes with loose caps and loose O-rings. When the centrifuge was activated, some of the tubes' contents would be aerosolized, resulting in the production of almost 10 billion airborne pathogens.
The Army then modeled a plume of the airborne pathogens as it moved through the lab and outside via the ventilation system. In order to produce conservative results, the Army simulated only one HEPA filter, operating at only 95 percent effectiveness. The Army concluded the chance of public exposure to an airborne pathogen, at a 50 percent rate of contracting the disease, was extremely remote.
Using the Army centrifuge model, the DOE concluded that the chances of exposure at the LLNL BSL-3 lab were even more remote than those modeled by the Army because the Army scenario assumed one HEPA filter that was 95 percent effective. The LLNL BSL-3 lab, the DOE reasoned, filters all room air through two HEPA filter banks, each of which is at least 99.97 percent effective. The Army scenario also assumed a lab in close physical proximity to the public, whereas the LLNL BSL-3 lab is one-half mile from the nearest public area. Finally, the Army assumed lower wind speeds than are prevalent at LLNL; higher wind speeds would decrease airborne concentrations more quickly. Based on this analysis, the DOE concluded that even if a catastrophic release were to occur, there would be no significant impact on public health or safety. This conclusion thus led the DOE to issue a Finding of No Significant Impact (FONSI).
C. The First Round of Litigation
In August 2003, Tri-Valley CAREs brought suit in the United States District Court for the Northern District of California, challenging the EA on numerous grounds under NEPA. Tri-Valley Cares v. U.S. Department of Energy, No. C 03-3926-SBA, 2004 WL 2043034 (N.D. Cal. Sep 10, 2004). On September 10, 2004, the district court granted summary judgment for the DOE on the ground that the DOE had satisfied the requirements of NEPA in preparing the original EA. Id., at 2004 WL 2043034 at *1. On appeal, we affirmed all aspects of the EA, except for the DOE's failure to consider the environmental impact of a terrorist attack. Tri-Valley Cares v. Dep't of Energy, 203 F. App'x 105, 107 (9th Cir. 2006). We thus affirmed in part and reversed in part, remanding for the sole purpose of the DOE's considering whether the threat of a terrorist act required the preparation of an EIS. Id. at 106-07.
D. The Draft Revised Environmental Assessment (DREA)
On remand, the DOE prepared a DREA in March 2007, addressing the impacts associated with terrorist attacks to determine whether the threat of terrorist attack necessitated preparation of an EIS. The DREA was circulated for public comment from April 11, 2007 through May 11, 2007.
To analyze the threat that terrorist activity posed to the LLNL BSL-3 laboratory, the DOE was required to take a different approach than it did when analyzing the threat posed by accidents. Because there are an infinite number of possible modes of attack, the DOE considered three general types of terrorist threats: (1) a direct terrorist attack at the LLNL BSL-3 facility, resulting in loss of containment; (2) the theft and release of pathogenic material by an LLNL terrorist outsider; and (3) the theft and release of pathogenic material by an LLNL terrorist insider.
1. Direct Attack Resulting in Loss of Containment
The DOE considered various possible modes of direct terrorist attack on the LLNL facility, including a suicidal plane crash or an explosive device delivered by vehicle or on foot. In considering the impact of this type of threat, the DOE used the centrifuge scenario to determine the bounded or outer limits of any release, and then analyzed factors which could mitigate such a release.
The DOE concluded that the outer bounds of dispersion in a terrorist attack would be the same as those of the catastrophic release scenario used in the original EA. The centrifuge model analysis supported a finding of no significance for ...