Emhart Industries, Inc. v. New England Container Company, Inc et al
Filing
745
PHASE II FINDINGS OF FACT AND CONCLUSIONS OF LAW. So Ordered by Chief Judge William E. Smith on 8/17/2017. Associated Cases: 1:11-cv-00023-S-LDA, 1:06-cv-00218-S-LDA(Jackson, Ryan)
UNITED STATES DISTRICT COURT
FOR THE DISTRICT OF RHODE ISLAND
___________________________________
)
)
)
Plaintiff and Counterclaim )
Defendant,
)
)
v.
)
)
NEW ENGLAND CONTAINER COMPANY,
)
INC; et al.,
)
)
Defendants and Counterclaim )
Plaintiffs.
)
___________________________________)
)
EMHART INDUSTRIES, INC.,
)
)
Plaintiff and Counterclaim )
Defendant,
)
)
v.
)
)
UNITED STATES DEPARTMENT OF THE
)
AIR FORCE; et al.,
)
)
Defendants, Counterclaim
)
Plaintiffs, and Third-Party )
Plaintiffs,
)
)
v.
)
)
BLACK & DECKER, INC.; et al.,
)
)
Third-Party Defendants.
)
___________________________________)
EMHART INDUSTRIES, INC.,
C.A. No. 06-218 S
C.A. No. 11-023 S
PHASE II FINDINGS OF FACT AND CONCLUSIONS OF LAW
I. Introduction
Dioxin and other toxic chemical pollution at the Centredale
Manor Restoration Project Superfund Site (“Centredale Site” or
“Site”) in North Providence, Rhode Island, has led to going on ten
years of litigation over which parties are responsible and what is
the appropriate remedy. The Court divided the litigation into three
phases. (See Eighth Revised Case Management Order, ECF No. 295.)1
At the close of trial in Phase I, the Court found Emhart jointly
and severally liable for the release of dioxin at the Site. The
Court has now concluded Phase II of the trial 2 and must provide
findings of fact and conclusions of law addressing the following
two issues: (1) whether the Environmental Protection Agency’s
(“EPA”) remedy-selection process was arbitrary, capricious, or
otherwise not in accordance with law; and (2) whether Emhart had
sufficient cause to refuse to comply with EPA’s June 10, 2014
Administrative Order. The necessary contributions, if any, of
third-party defendants will be addressed in Phase III of the trial.
The Court provided a comprehensive background discussion and
procedural history of this case in its Phase I opinion and need
1
The present case was also preceded by other insurance
coverage litigation. The Centredale Site is truly a litigation
gift that keeps on giving.
2
The Court received evidence over the course of thirteen
days, concluding on January 19, 2017. (See Trial Trs. vols. 1-13,
ECF Nos. 448-57, 494-96.) This included thousands of pages of
exhibits from both the United States (Bates numbered “US__”) and
Emhart (Bates numbered “Emhart__”). The parties supplied posttrial briefs thereafter. (See Emhart Post-Trial Brief, ECF No.
461; Gov’t Post-Trial Brief, ECF No. 466; Emhart Post-Trial Reply
Brief, ECF No. 497.) Lastly, the Court heard oral argument on April
4, 2017. (See Trial Tr. vol. 14, ECF No. 534.)
2
not repeat it here. (See Phase I Findings of Fact and Conclusions
of Law (“Phase I Findings”), ECF No. 405.) However, since issuing
the Phase I Findings, there have been three new and important
developments in the case relating to the Phase II litigation.
First, the parties have come to an agreement regarding the United
States’ past response costs in light of the Court’s findings in
Phase I. (See Stipulation Regarding United States’ Past Response
Costs, ECF No. 444.) As such, the Court need not determine the
amount of past response costs.
Second, the Government moved to limit the scope of judicial
review during Phase II. (Gov’t Mot. to Limit Disc. and Scope of
Review, ECF No. 415.) The Government argued that, because judicial
review under CERCLA is strictly limited to the administrative
record, Emhart should be prohibited from presenting any evidence
or arguments not found in the administrative record. Emhart opposed
the Government’s motion, arguing that review under CERCLA is not
so limited. (Emhart Obj., ECF No. 416.) The Court denied the
Government’s motion without prejudice, thereby allowing Emhart to
conduct
discovery
and
present
the
objected
to
evidence
and
arguments at trial. (Order Den. Mot. to Limit Disc. and Scope of
Review 6, ECF No. 421.) However, the Court has reserved its
judgment on the ultimate admissibility of the challenged evidence
and arguments to this point. (Id. at 6.)
3
Third, and lastly, the Government moved to exclude certain
portions of testimony provided by Emhart’s expert, Mr. Jeffrey
Loureiro. (Gov’t Mot. to Exclude Certain Test. of Jeffrey Loureiro,
ECF No. 447.) The Government argued that significant portions of
Mr. Loureiro’s opinions had not been disclosed in his expert report
in violation of Rule 26(a)(2) of the Federal Rules of Civil
Procedure. According to the Government, the admission of Mr.
Loureiro’s testimony would be “highly prejudicial . . . given the
United
States’
discovery
or
inability
to
counter
to
explore
them
such
effectively
opinions
through
through
rebuttal
testimony.” (Id. at 2.) In light of the Government’s concerns, the
Court decided to allow the testimony of Mr. Lourerio but provided
the Government an opportunity to conduct additional discovery and
to designate and utilize rebuttal experts to address the concerns
highlighted in the Government’s motion. The Court therefore denied
that motion as moot. (Text Order of 10/07/2016.)
In this decision, the Court will set forth its findings of
fact and conclusions of law with regards to the question of remedy,
and address along the way the various issues reserved to this
point. See Fed. R. Civ. P. 52(a)(1). The decision begins with an
overview of the CERCLA process by which EPA may choose a response
action, as well as the standards of judicial review for challenges
to a response action. Next the Court provides specific findings of
fact, starting with a history of EPA action at the Site, and then
4
moving
into
the
topics
which
provide
the
bases
for
Emhart’s
challenge. After each topic the Court provides conclusions of law.
II. Remedy Selection
Congress enacted the Environmental Response, Compensation,
and Liability Act (“CERCLA”), 42 U.S.C. § 9601 et seq., “in
response to the serious environmental and health risks posed by
industrial pollution.” United States v. Bestfoods, 524 U.S. 51, 55
(1998). To address those risks, CERCLA “grants the President broad
power to command government agencies and private parties to clean
up hazardous waste sites.” Id. (quoting Key Tronic Corp. v. United
States, 511 U.S. 809, 814 (1994)). These cleanup efforts are called
“response actions,” and may require both “the cleanup or removal
of released hazardous substances” as well as other “remedial action
. . . to prevent or minimize the release of hazardous substances.”
42 U.S.C. § 9601(23)-(25).
EPA acts on the President’s behalf in choosing the appropriate
response action. See Exec. Order No. 12580(b)(1), (g). In doing
so, EPA must work within the framework provided by CERCLA and the
National
Contingency
Plan
(“NCP”).
See
42
U.S.C.
9604(a)(1)
(requiring EPA to act “consistent with the national contingency
plan”); National Contingency Plan, 40 C.F.R. § 300, et seq. In
addition, EPA has established various guidance documents to assist
in the CERCLA process. While these guidance documents are nonbinding on EPA, they do represent EPA’s collective wisdom as to
5
best practices. 3 The steps required by CERCLA and the NCP, and
implemented with the assistance of EPA guidance documents, are
outlined below.
A. National Priorities List
The first step in the CERCLA process is placing a site on the
National Priorities List (“NPL”). See 42 U.S.C. § 9605(a)(8)(b);
40 C.F.R. § 300.425. A site is appropriately included on the NPL
if, for example, EPA determines that a hazardous substance poses
“a significant threat to public health.” Id. § 300.425(c)(3)(ii).
A notice and comment period is required before a site is officially
placed on the NPL. Id. § 300.425(d)(5). Once the notice and comment
process is complete, and if EPA determines that NPL listing is
still appropriate, EPA may begin the process of developing a
response action for the site.
B. Remedial Investigation and Feasibility Study
EPA is required to conduct a remedial investigation (“RI”)
and feasibility study (“FS”) before choosing a response action.
The end goal of the RI/FS process is “to assess site conditions
and evaluate alternatives to the extent necessary to select a
remedy.” Id. § 300.430(a)(2).
The first step is the RI, which
seeks to “collect data necessary to adequately characterize the
3
(See, e.g., Ms. Taylor Test., Trial Tr. vol. 2, 5:3–9, ECF
No. 449 (acknowledging that EPA guidance represents “the
collective information and practices developed by EPA through its
experience in implementing remedies”).)
6
site
for
remedial
the
purpose
of
alternatives.”
developing
Id.
§
and
evaluating
300.430(d)(1).
EPA
effective
is
given
significant leeway to develop a RI process specific to the site.
But, at a minimum, EPA must conduct a field investigation (i.e.,
collect site-specific data) and create a baseline risk assessment.
Id.
The data collected during the field investigation includes
the physical characteristics of the site and the hazardous material
present
as
well
as
the
exposure
pathways
through
which
the
hazardous material may affect human health and the environment.
Id.
§
300.430(d)(2)(i)-(vii).
Part
of
this
process
involves
estimating the reasonable maximum exposure that is likely to occur
for both current and potential future land use at the site. (EPA,
Interim Final Risk Assessment Guidance for Superfund (“IFRAGS”),
vol. I, Emhart579-24.) EPA then uses this data in the baseline
risk
assessment
to
understand
the
extent
to
which
hazardous
materials pose a threat to human health and the environment, as
well as what would be “acceptable exposure levels” for the site
going forward. 40 C.F.R. § 300.430(d)(4).
At
this
point,
EPA
transitions
from
the
RI
process
of
collecting data to the FS process of finding a remedy. 4 “The
4
While the RI and FS are labeled as separate steps in the
CERCLA process, the RI does not end when the FS begins. Instead,
“the RI and FS are interactive processes that are conducted
concurrently,” such that “field investigation activities will be
7
national goal of the remedy-selection process is to select remedies
that are protective of human health and the environment, that
maintain protection over time, and that minimize untreated waste.”
Id. § 300.430(a)(1)(i). As an aid to developing potential remedies
that accomplish this overarching goal, EPA first establishes what
are
labeled
Preliminary
Remediation
Goals
(“PRGs”)
targeting
“acceptable exposure levels that are protective of human health
and the environment.” Id. § 300.430(e)(2)(i). These PRGs must,
among other things, comply with federal and state environmental
laws 5
and
limit
the
lifetime
cancer
risk
from
carcinogenic
exposure. Id. § 300.430(e)(2)(i)(A)(1)-(2). EPA then develops a
range of response alternatives that may achieve those PRGs.
In order to choose among the response options for the site,
the
various
alternatives
are
initially
screened
using
three
criteria: (1) effectiveness; (2) implementability; and (3) cost.
Id. § 300.430(e)(7)(i)-(iii). At this stage, alternatives that
provide “significantly less effectiveness” or are “technically or
ongoing during the development and screening of remedial action
alternatives.”
EPA,
The
Remedial
Investigation:
Site
Characterization and Treatability Studies, Emhart509-1; see also,
40 C.F.R. § 300.430(e)(1) (“Development of alternatives shall be
fully integrated with the site characterization activities of the
remedial investigation.”); id. § 300.430(e)(1) (“Preliminary
remediation goals should be modified, as necessary, as more
information becomes available during the RI/FS.”)
5
These are referred to in CERCLA as Applicable and Relevant
and
Appropriate
Requirements
(“ARARs”).
40
C.F.R.
§ 300.505(d)(2)(iii).
8
administratively
infeasible”
may
be
eliminated.
Id.
§ 300.430(e)(7)(i)-(ii). Additionally, “[c]osts that are grossly
excessive compared to the overall effectiveness of alternatives
may be considered as one of several factors used to eliminate
alternatives.” Id. § 300.430(e)(7)(iii).
Alternatives that survive the initial culling proceed to the
“detailed analysis of alternatives.” Id. § 300.430(e)(9). This
requires “an assessment of individual alternatives against each of
nine evaluation criteria and a comparative analysis that focuses
upon the relative performance of each alternative against those
criteria.” Id. § 300.430(e)(9)(ii). The nine criteria used by EPA
to compare alternatives are: (1) overall protection of human health
and
the
environment;
(2)
compliance
with
federal
and
state
environmental laws (i.e., “ARARs” 6); (3) long-term effectiveness
and permanence; (4) reduction of toxicity, mobility, or volume
through
treatment;
implementability;
(7)
(5)
cost;
short-term
(8)
state
effectiveness;
acceptance;
and
(6)
(9)
community acceptance. See id. § 300.430(e)(9)(iii)(A)-(I). Once
sufficient information has been gathered such that EPA can compare
the alternatives based on the nine evaluation criteria, EPA may
proceed with remedy selection.
6
See id.
9
C. Remedy Selection
During
criteria
remedy
into
selection
three
EPA
categories.
places
The
the
first
nine
evaluation
category,
labeled
“threshold criteria,” are the criteria “that each alternative must
meet
in
order
to
§ 300.430(f)(1)(i)(A).
be
The
eligible
two
for
threshold
selection.”
criteria
are
Id.
overall
protection of human health and the environment and compliance with
ARARs (criteria (1) and (2), above). 7
Once EPA has screened out alternatives that do not meet the
threshold criteria, EPA then compares the remaining alternatives
based
on
the
balancing
second
criteria.”
category
Id.
§
of
criteria,
labeled
300.430(f)(1)(i)(B).
“primary
These
include
long-term effectiveness and permanence, reduction of toxicity,
mobility, or volume through treatment, short-term effectiveness,
implementability, and cost (criteria (3)-(7), above). Id. As the
label suggests, these “primary balancing criteria” are balanced
against
one-another.
In
doing
so,
the
NCP
provides
certain
preferences. For instance, a remedy must be “cost-effective” in
that “its costs are proportional to its overall effectiveness.”
Id.
§
300.430(f)(1)(ii)(D).
emphasize
long-term
In
effectiveness
7
addition,
and
“balancing
reduction
of
shall
toxicity,
While compliance with ARARs is one of the “threshold
criteria,” there are limited circumstances where a chosen remedy
need not meet those standards. See id. § 300.430(f)(1)(ii)(C).
10
mobility,
or
volume
through
treatment,”
thereby
focusing
on
“permanent solutions . . . to the maximum extent practicable.” Id.
§ 300.430(f)(1)(ii)(E). However, in the end, the NCP does not
dictate exactly how the primary balancing criteria should be
weighed.
Instead,
the
NCP
provides
EPA
with
considerable
discretion to select a remedy that “reflect[s] the scope and
purpose of the actions being undertaken and how the action relates
to
long-term,
comprehensive
response
at
the
site.”
Id.
§ 300.430(f)(a).
The
last
criteria.”
Id.
category
§
EPA
must
consider
300.430(f)(1)(i)(C).
The
is
the
“modifying
modifying
criteria
include state and community acceptance (criteria (8) and (9),
above). Because the state and community provide formal input during
the notice and comment period for the proposed plan, state and
community acceptance is typically only considered, at this stage,
“to the extent that information is available during the FS.” EPA,
A Guide to Preparing Superfund Proposed Plans, Records of Decision,
and
Other
Remedy
Selection
Decision
Documents,
Emhart518-33.
However, “after public comment is received on the Proposed Plan,”
the modifying criteria will be “fully considered.” Id. This means
that, “[i]n the final balancing of trade-offs between alternatives
upon which the final remedy selection is based, modifying criteria
are of equal importance to the balancing criteria.” Id. This does
not provide the state or community the power to veto a selected
11
remedy 8; their input is simply considered along with the other
balancing criteria.
After
completing
the
“detailed
analysis
of
alternatives”
using the nine criteria, EPA chooses a remedy through a “two-step
process.” 40 C.F.R. § 300.430(f)(1)(ii). First, EPA “identifies a
preferred alternative and presents it to the public in a proposed
plan, for review and comment.” Id. The proposed plan functions not
only to “supplement the RI/FS,” but also to “provide the public
with
a
reasonable
opportunity
to
comment
on
the
preferred
alternative for remedial action, as well as alternative plans under
consideration, and to participate in the selection of remedial
action at a site.” Id. § 300.430(f)(2). To accomplish this, the
proposed plan must, among other things, “[p]rovide a brief summary
description of the remedial alternatives evaluated in the detailed
analysis” as well as “[i]dentify and provide a discussion of the
rationale that supports the preferred alternative.” Id. The public
is then given “a reasonable opportunity, not less than 30 calendar
days, for submission of written and oral comments” as well as an
“opportunity for a public meeting.” Id. § 300.430(f)(3)(i)(C)-(D).
After receiving public input on the proposed remedy, the
second step for remedy selection requires EPA to “review the public
comments and consult with the state (or support agency) in order
8
See, e.g., 42 U.S.C. § 9604(j); id. § 9621(e)(1); 40 C.F.R.
§ 300.400(d).
12
to determine if the [proposed plan] remains the most appropriate
remedial
action
for
the
site
or
site
problem.”
Id.
§ 300.430(f)(1)(ii). The NCP anticipates that public comments may
provide “new information or points of view” that prompt EPA “to
modify aspects of the preferred alternative or decide that another
alternative
provides
a
more
appropriate
balance.”
Id.
§ 300.430(f)(4). To the extent EPA makes any significant changes
to the remedy, those changes must be documented. 9 EPA must then
make the final remedy selection. Id. § 300.430(f)(4)(i).
Just as EPA is not required to remove all uncertainty at the
RI/FS stage regarding the conditions at the site, EPA is also not
required to provide complete details of the final remedy at the
selection stage. The NCP envisions that EPA will fill in the
details of the final remedy during the implementation (or “remedial
design”) phase and that the final remedy may require modifications.
See generally id. § 300.435. The NCP also allows EPA to reserve
decisions regarding how to handle certain portions of the remedy
9
EPA must provide documentation in the record of decision
where the final remedy “significantly differs from the original
proposal.” Id. § 300.430(f)(3)(ii). If the change is so drastic
that it could not have been “reasonably anticipated by the public”
based on the information in the proposed plan and administrative
record, then EPA must allow for an additional notice and comment
period on the revised plan. Id. § 300.430(f)(3)(ii)(B). If the
changes to the plan could have been “reasonably anticipated,” then
EPA need only “[i]nclude a discussion in the record of decision of
the significant changes and reasons for such changes.” Id.
§ 300.430(f)(3)(ii)(A).
13
until remedial design. See, e.g., id. § 300.825(a)(1)-(2). This
process is discussed below in the “Remedial Design and Remedial
Action” section.
D. Documentation of the Remedy-Selection Process
After selecting the final remedy, EPA must “establish an
administrative record that contains the documents that form the
basis for the selection of a response action.” 40 C.F.R. § 300.800.
An important piece of the administrative record is the Record of
Decision (“ROD”), which includes “all facts, analyses of facts,
and site-specific policy determinations considered” by EPA in
selecting the final remedy. Id. § 300.430(f)(5)(i). The ROD is
essentially EPA’s justification for its decision, explaining, for
example, “[h]ow the selected remedy is protective of human health
and
the
environment”
and
proportional to its costs.”
part
of
this
“responsiveness
“provides
effectiveness
Id. § 300.430(f)(5)(ii)(A), (D). As
justification,
summary,”
overall
the
which
ROD
is
“a
also
must
written
include
summary
a
of
significant comments, criticisms, and new relevant information
submitted during the public comment period and the lead agency
response to each issue.” Id. § 300.430(f)(3)(i)(F). Lastly, if EPA
chooses to reserve certain decisions for a later date, it may,
“[w]here appropriate, provide a commitment for further analysis.”
Id. § 300.430(f)(5)(iii)(D).
14
The administrative record is not a static document. Even after
the final remedy is selected, EPA is responsible for updating the
administrative record where necessary. For example, if the ROD did
not address a portion of the response action or reserved certain
decisions until the implementation phase, EPA must document those
later decisions in the administrative record. Id. § 300.825(a)(1).
Additionally, if EPA decides to modify the final remedy during
remedial design, it must document those changes either through an
“explanation of significant differences” or a ROD amendment, as
appropriate. Id. § 300.825. 10
E. Remedial Design and Remedial Action
The final step is “the development of the actual design of
the selected remedy and implementation of the remedy through
construction.”
Id.
§
300.435(a).
The
NCP
labels
this
as
the
“remedial design/remedial action (RD/RA) stage.” Id. While the ROD
establishes the final remedy, the NCP leaves it to the RD/RA stage
for EPA to determine the remedy’s final design. The “initial
building block in developing” the final design is the information
contained in the RI/FS, but EPA guidance also envisions that
additional
“data
necessary
during
acquisition”
RD/RA.
and
EPA,
“sample
Scoping
the
analysis”
Remedial
will
be
Design,
Emhart516-2. EPA then recommends going through multiple design
10
See infra note 11.
15
phases
before
coming
to
a
“final
design”
and
beginning
construction. Id. at 0001. Once a final design is complete, EPA
must – as a final notice to the public - “issue a fact sheet and
provide, as appropriate, a public briefing prior to the initiation
of remedial action.” 40 C.F.R. § 300.435(c)(3). EPA may then
implement
the
remedy,
and
as
long
as
the
“remedial
action
objectives and remediation goals in the ROD” are accomplished, the
CERCLA
response
action
process
is
largely
complete.
Id.
§ 300.435(f)(1); 42 U.S.C. § 9621(d) (stating that the response
action must “attain a degree of cleanup . . . which assures
protection of human health and the environment”).
If strict adherence to the final design proves unworkable at
any point, EPA guidance provides for significant “flexibility” to
account for “any constraining factors of the particular site.”
EPA,
Scoping
the
Remedial
Design,
Emhart516-1.
The
NCP
also
foresees that “[a]dditional work” may be “needed as a result of
such unforeseen situations as newly discovered sources, types, or
quantities of hazardous substances.” 40 C.F.R. § 300.435(e)(1)(i).
Because the “chief task” of RD/RA is “to achieve the goals of the
Record of Decision . . . in a timely manner,” EPA, Scoping the
Remedial Design, Emhart516-1, as opposed to blind adherence to any
particular design, EPA is permitted to change the remedial design
16
at any point. The NCP simply requires that sufficient notice and
opportunity to comment is provided to the public. 11
F. Unilateral Administrative Order
The
parties
responsible
for
the
release
of
the
hazard
materials at the site are liable for the costs associated with the
response action. See 42 U.S.C. § 9607(a). EPA can either complete
the response action and seek reimbursement from the responsible
parties
or
require
the
responsible
parties
to
implement
the
response action themselves. Where EPA determines that “there may
be an imminent and substantial endangerment to the public health
or welfare or the environment,” EPA is authorized to issue a
unilateral administrative order (“UAO”) requiring the responsible
parties to implement the response action “as may be necessary to
protect
public
health
and
welfare
and
the
environment.”
Id.
§ 9606(a).
If a responsible party “willfully violates, or fails or
refuses to comply with” the UAO, EPA may seek an order “in the
11
If,
after
additional
data
collection
or
during
construction, EPA determines that the final design will “differ[]
significantly from the remedy selected in the ROD with respect to
scope, performance, or cost,” EPA must provide the public notice
of this change. 40 C.F.R. § 300.435(c)(2). Where the differences
“significantly change but do not fundamentally alter the remedy
selected in the ROD,” the lead agency will publish an “explanation
of
significant
differences.”
Id.
§
300.435(c)(2)(i).
Alternatively, if EPA does “fundamentally alter” the final remedy
in its final design, EPA must amend the ROD and provide for another
period of notice and comment. Id. § 300.435(c)(2)(ii).
17
appropriate United States district court to enforce” the UAO. Id.
§ 9606(b)(1). Additionally, if the district court finds that the
responsible
party
refused
to
comply
with
the
UAO
without
“sufficient cause,” the responsible party is subject to daily fines
during the period of non-compliance as well as treble damages for
any work EPA performed at the site. Id. § 9606(b)(1), 9607(c)(3).
By regulation, the daily fine amount is $37,500 per day for every
day of non-compliance between December 6, 2013 and November 2,
2015, and $54,789 per day thereafter. 40 C.F.R. §§ 19.2, 19.4.
III. Scope of Review
Having described the remedy-selection process as outlined by
CERCLA and the NCP, the Court will now provide its findings of
fact
and
conclusions
of
law
regarding
EPA’s
remedy-selection
process at the Site. Before doing so, however, the Court must first
determine what evidence and arguments it will consider, as well as
the standard of review. As discussed previously, the Government
moved, pre-trial, to limit the scope of discovery and judicial
review. According to the Government, the Court should not consider
evidence and arguments not contained in the administrative record
when making its determination about the appropriateness of EPA’s
chosen remedy. The Court denied the Government’s motion without
prejudice,
permitting
Emhart
to
take
discovery
and
present
evidence and arguments at trial not contained in the administrative
record. The Court “reserve[d] ruling on the admissibility of any
18
particular extra-record evidence until the time of trial.” (Order
Den. Mot. to Limit Disc. and Scope of Review 6.) The Government
has reiterated its arguments at trial and in its post-trial briefs,
and Emhart has again responded. (See Gov’t Post-Trial Brief 8-13;
Emhart Post-Trial Reply Brief 1-22.)
The two questions the Court must answer are: (1) whether the
Court should consider evidence not contained in the administrative
record;
and
(2)
whether
the
Court
should
consider
Emhart’s
arguments that were not made during the notice and comment period.
A. What Evidence Should the Court Consider?
CERCLA provides the following limitation to judicial review:
“In any judicial action under this chapter, judicial review of any
issues concerning the adequacy of any response action taken or
ordered by the President shall be limited to the administrative
record.”
42
U.S.C.
§
9613(j)(1)
(emphasis
added).
CERCLA
reiterates this limitation when describing the applicable standard
of review: “In considering objections raised in any judicial action
under this chapter, the court shall uphold the President’s decision
in selecting the response action unless the objecting party can
demonstrate, on the administrative record, that the decision was
arbitrary and capricious or otherwise not in accordance with law.”
19
Id. § 9613 (j)(2) (emphasis added). 12 As these provisions make
clear, “[u]nder CERCLA, judicial review normally is limited to the
administrative record as it existed at the time of the challenged
agency action.” United States v. JG-24, Inc., 478 F.3d 28, 33-34
(1st Cir. 2007); see also Murphy v. Comm’r of Internal Revenue,
469
F.3d
27,
31
(1st
Cir.
2006)
(“[T]he
Supreme
Court
has
consistently stated that review of administrative decisions is
ordinarily limited to consideration of the decision of the agency
. . . and of the evidence on which it was based.”) (internal
quotations omitted). This ensures that the Court “take[s] into
account ‘neither more nor less information than did the agency
when it made its decision.’” Linemaster Switch Corp. v. EPA, 938
12
The administrative record is typically closed once the
decision document has been signed, except in the following narrow
circumstances:
(a) The lead agency may add documents to the
administrative record file after the decision document
selecting the response action has been signed if: (1)
The documents concern a portion of a response action
decision that the decision document does not address or
reserves to be decided at a later date; or (2) An
explanation of significant differences required by §
300.435(c), or an amended decision document is issued,
in
which
case,
the
explanation
of
significant
differences or amended decision document and all
documents that form the basis for the decision to modify
the response action shall be added to the administrative
record file.
40 C.F.R. § 300.825.
20
F.2d 1299, 1305 (D.C. Cir. 1991) (quoting Walter O. Boswell Mem’l
Hosp. v. Heckler, 749 F.2d 788, 792 (D.C. Cir. 1984)).
However, while generally confining the scope of judicial
review to the administrative record, CERCLA also provides that
“[o]therwise applicable principles of administrative law shall
govern whether any supplemental materials may be considered by the
court.” 42 U.S.C. § 9613(j)(1). Typically, “[c]ourts require a
strong showing of bad faith or improper behavior before ordering
the supplementation of the administrative record.” Town of Norfolk
v. U.S. Army Corps of Eng’rs, 968 F.2d 1438, 1458–59 (1st Cir.
1992); see also JG-24, Inc., 478 F.3d at 34 (“Normally, we do not
allow supplementation of the administrative record unless the
proponent points to specific evidence that the agency acted in bad
faith.”). This exception provides no assistance to Emhart, as there
is no evidence that EPA crafted a remedy for the Site in bad faith
or based on an improper motive, and EPA’s “designation of the
Administrative
Record,
like
procedure,
entitled
to
is
any
a
established
presumption
of
administrative
administrative
regularity.” Nw. Bypass Grp. v. U.S. Army Corps of Eng’rs, No. CIV
06-CV-00258-JAW, 2007 WL 1498912, at *2 (D.N.H. May 14, 2007)
(quoting Bar MK Ranches v. Yuetter, 994 F.2d 735, 740 (10th Cir.
1993)).
Emhart must therefore find some other legal avenue if it is
to successfully inject evidence outside the administrative record
21
into this proceeding. Outside of a showing of bad faith, the First
Circuit recognizes two other “exceptions to the rule against
supplementation.” Nw. Bypass Grp., 2007 WL 1498912, at *2.
First, “supplementation may be proper when . . . there is a
record so inadequate that it prevents judicial review.” Id. at *23 (citing Murphy, 469 F.3d at 31). This applies in very limited
circumstances. For instance, the Supreme Court has recognized
that, “where there are [no formal] administrative findings that
were made at the same time as the decision, . . . it may be that
the only way there can be effective judicial review is by examining
the decisionmakers themselves.” Citizens to Pres. Overton Park,
Inc. v. Volpe, 401 U.S. 402, 420 (1971) (citing Shaughnessy v.
Accardi,
349
U.S.
280
(1955)).
This
applies
where
the
administrative agency’s decision cannot be discerned or justified
on the record. See Camp v. Pitts, 411 U.S. 138, 142–43 (1973)
(reviewing
whether
“there
was
such
failure
to
explain
administrative action as to frustrate effective judicial review”).
Therefore, supplemental materials are typically unnecessary when
the administrative agency’s judgment is based on a substantial
record. See Nw. Bypass Grp., 2007 WL 1498912, at *3 (finding an
administrative record “more than sufficient to allow for judicial
review” because the administrative record was “hefty, 3,233 pages
over seven volumes, with documents spanning from 1989 to 2006”).
Where a substantial record is available, even if the administrative
22
agency’s
explanation
unnecessary
as
long
is
as
“curt,”
the
supplemental
explanation
materials
are
“indicate[s]
the
determinative reason for the final action taken.” Pitts, 411 U.S.
at 143.
The second exception to the “rule against supplementation”
applies where “additional testimony by experts” will “aid to
understanding highly technical, environmental matters.” Nw. Bypass
Grp., 2007 WL 1498912, at *2 (quoting Valley Citizens for a Safe
Env’t v. Aldridge, 886 F.2d 458, 460 (1st Cir. 1989)). The Court
may
even
allow
“additional
factual
evidence
as
an
aid
to
understanding.” Valley Citizens for a Safe Env’t, 886 F.2d at 460.
The necessity of such evidence “is discretionary with the reviewing
court.” Id. However, it is important to note that the purpose of
such evidence is “simply to help [the Court] understand matters in
the agency record.” Id. Therefore, the Court still “looks first
and foremost at the record before the agency.” Id.; see also Olsen
v. United States, 414 F.3d 144, 155 (1st Cir. 2005) (“The focal
point for judicial review should be the administrative record
already in existence, not some new record made initially in the
reviewing court.”) (quoting Pitts, 411 U.S. at 142); United States
v. Dravo Corp., No. 8:01CV500, 2003 WL 21434761, at *3 (D. Neb.
June 20, 2003) (reviewing an EPA response action and refusing to
“permit supplementation of the record absent a showing that it is
23
explanatory and not intended to invoke new material into the
case”).
In this case, given the extensive record compiled by EPA, as
well as EPA’s documentation of its decision-making process in the
ROD, the Court does not find the record “so inadequate” as to
require
supplemental
materials.
However,
the
subject
matter
involved certainly falls under the umbrella of “highly technical,
environmental matters” where the Court has discretion to consider
“additional
testimony
by
experts”
and
“additional
factual
evidence” as an aid to understanding the administrative record.
Nw. Bypass Grp., 2007 WL 1498912, at *2 (quoting Valley Citizens
for a Safe Env’t, 886 F.2d at 460). The Court will therefore
consider the expert testimony presented by both parties. With that
said, in considering the expert testimony, the Court acknowledges
that
it
must
remain
focused
“first
and
foremost”
on
the
administrative record, Valley Citizens for a Safe Env’t, 886 F.2d
at 460, as “the focal point for judicial review should be the
administrative record already in existence, not some new record
made initially in the reviewing court.” Olsen, 414 F.3d at 155
(quoting Pitts, 411 U.S. at 142).
B. What Arguments Should the Court Consider?
Having established what evidence is admissible in this case,
the Court must determine which of Emhart’s arguments the Court
will consider. According to EPA, the Court should not consider any
24
of Emhart’s arguments that were not previously submitted during
the official notice and comment period on the proposed remedy.
Emhart, on the other hand, urges the Court to consider all of its
arguments, including those made for the first time at trial.
There is limited precedent directly addressing this issue in
the context of CERCLA. Generally speaking, arguments not made
before an administrative agency are waived when a court reviews
the administrative agency’s decision:
[w]e have recognized in more than a few decisions, and
Congress has recognized in more than a few statutes,
that orderly procedure and good administration require
that objections to the proceedings of an administrative
agency be made while it has opportunity for correction
in order to raise issues reviewable by the courts.
United States v. L.A. Tucker Truck Lines, Inc., 344 U.S. 33, 36–
37 (1952). This requirement is commonly referred to as “issue
exhaustion.” Sims v. Apfel, 530 U.S. 103, 107 (2000). As the First
Circuit has explained, “this rule preserves judicial economy,
agency
autonomy,
and
accuracy
of
result
by
requiring
full
development of issues in the administrative setting to obtain
judicial review.” Pepperell Assocs. v. EPA, 246 F.3d 15, 27 (1st
Cir. 2001) (citing Northern Wind, Inc. v. Daley, 200 F.3d 13, 18
(1st Cir. 1999)).
The doctrine of issue exhaustion was adhered to relatively
recently by the First Circuit in Upper Blackstone Water Pollution
Abatement Dist. v. EPA, 690 F.3d 9, 18 (1st Cir. 2012). That case
25
involved a sewage company (“petitioner”) appealing a limitation
imposed
on
it
by
EPA
under
the
National
Pollutant
Discharge
Elimination System. After notice and comment on the limitation,
petitioner
challenged
EPA’s
decision
in
the
First
Circuit.
Petitioner sought review of, among other things, “the limit placed
on aluminum discharge, arguing that the EPA assembled and then
relied upon an erroneous data set in deriving the limit.” Id. at
33. The First Circuit refused to consider that argument, however,
because it had not been raised during the notice and comment
period.
The First Circuit explained that, by regulation, petitioner
was required to “raise all reasonably ascertainable issues and
submit
all
reasonably
available
arguments
supporting
their
position by the close of the public comment period . . . .” Upper
Blackstone, 690 F.3d at 30 (quoting 40 C.F.R. § 124.13). In light
of this statutory requirement, the court found that petitioner had
“waived the argument by failing to raise it during the public
comment period of the permitting process. . . . By failing to give
the
EPA
an
opportunity
to
address
the
argument
during
the
permitting process, [Upper Blackstone] has waived its claim.” Id.
(citing several cases, including L.A. Tucker Truck Lines, 344 U.S.
33, and Pepperell Assocs., 246 F.3d 15).
Emhart argues that Upper Blackstone is distinguishable from
this
case
because,
unlike
the
26
regulation
at
issue
in
Upper
Blackstone, CERCLA and its implementing regulations do not contain
a clear issue exhaustion provision. The Court disagrees.
A determination of whether issue exhaustion applies to an
administrative
characteristics
process
of
“requires
the
careful
particular
examination
administrative
of
the
procedure
provided.” Sims, 530 U.S. at 112–13 (O’Connor, J., concurring)
(quoting McCarthy v. Madigan, 503 U.S. 140, 146 (1992)). And while
the regulation reviewed by the First Circuit in Upper Blackstone
may have been slightly more definitive than the language in CERCLA,
the Court finds that CERCLA and its implementing regulations, when
taken as a whole, clearly require interested parties to present
arguments to EPA before bringing those issues before a federal
court.
After identifying the “preferred” remedy, EPA is required to
“present[] it to the public in a proposed plan, for review and
comment.” 40 C.F.R. § 300.430(f)(1)(ii). This ensures that the
public has “a reasonable opportunity to comment on the preferred
alternative for remedial action, as well as alternative plans under
consideration, and to participate in the selection of remedial
action at a site.” Id. § 300.430(f)(2). At the completion of the
notice
and
comment
“responsiveness
period,
summary,”
EPA
which
is
is
required
“a
written
to
create
summary
a
of
significant comments, criticisms, and new relevant information
submitted during the public comment period and the lead agency
27
response
to
responsiveness
each
issue.”
summary
is
Id.
then
§
300.430(f)(3)(i)(F).
placed
in
the
The
administrative
record. Id.
Importantly
though,
the
responsiveness
summary
does
not
address comments made outside the public comment period. See id.
§ 300.825. As is mandated by CERCLA’s implementing regulations,
the responsiveness summary will not address comments made “after
the close of the public comment period” unless the comments: (1)
“contain significant information not contained elsewhere in the
administrative record file”; (2) the information “could not have
been submitted during the public comment period”; and (3) the
information “substantially support[s] the need to significantly
alter the response action.” Id. § 300.825. Therefore, unless an
issue raised after the public comment period falls under this
exception,
that
issue
will
not
be
responded
to
in
the
responsiveness summary or included in the administrative record.
And it is in this context that CERCLA explicitly limits judicial
review to the information contained in the administrative record.
See 42 U.S.C. § 9613(j)(1)-(2).
This statutory and regulatory scheme, when viewed as a whole,
requires parties to make all of their known and available arguments
regarding the merits of a remedy to EPA during the notice and
comment period in the first instance. Only then, after EPA has had
the opportunity to provide its response in the administrative
28
record, may a federal court review EPA’s decision. See, e.g., JG24,
Inc.,
478
F.3d
at
33-34
(“Under
CERCLA,
judicial
review
normally is limited to the administrative record as it existed at
the time of the challenged agency action.”); Arco v. Travelers
Ins. Co., 730 F. Supp. 59, 69 (W.D. Mich. 1989) (“[R]efusal to
participate
in
this
administrative
process
[under
CERCLA]
essentially allows the EPA a free-reign in dictating response
methods since judicial review is limited to the administrative
record.”). To allow Emhart to make arguments it could have made
directly to EPA for the first time on judicial review would
frustrate this administrative scheme by depriving EPA of the
opportunity to address Emhart’s arguments in the first instance on
administrative review. Upper Blackstone, 690 F.3d at 30.
However,
while
judicial
review
is
generally
limited
to
arguments presented during the notice and comment period, the Court
also recognizes that there are some narrow exceptions to this rule.
For instance, the Court of Appeals for the District of Columbia
Circuit has recognized that EPA must justify “key assumptions” in
its analysis, regardless of whether a party specifically objects
during the notice and comment process:
EPA has a preexisting duty to examine key assumptions as
part of its affirmative burden of promulgating and
explaining a non-arbitrary, non-capricious rule and
therefore . . . must justify that assumption even if no
one objects to it during the comment period.
29
Oklahoma Dep’t of Envtl. Quality v. EPA, 740 F.3d 185, 192 (D.C.
Cir. 2014) (quoting Appalachian Power Co. v. EPA, 135 F.3d 791,
818 (D.C. Cir. 1998)). 13
While the First Circuit has not directly addressed this narrow
exception, the Court agrees with the District of Columbia Circuit
that issue waiver cannot absolve EPA of its responsibility to
explain the key assumptions that underpin its remedy and that
contain
obvious
mistakes.
Therefore,
while
the
Court
will
generally not consider Emhart’s arguments that were not presented
to EPA during the notice and comment period, the Court will
consider several obvious issues relating to key assumptions that
formed the basis of EPA’s selected remedy.
C. Standard of Review
With
the
bounds
of
admissible
evidence
and
argument
established in the preceding sections, the Court turns to the
applicable standard of review. The Court will uphold EPA’s decision
“unless the objecting party can demonstrate, on the administrative
record, that the decision was arbitrary and capricious or otherwise
13
Other Circuit Courts of Appeals have recognized similar
exceptions for cases in which the allegedly waived issue should
have been obvious to the reviewing agency. See, e.g., Sierra Club,
Inc. v. Bostick, 787 F.3d 1043, 1048 (10th Cir. 2015) (holding
that an issue not presented during the notice and comment period
is not waived where that issue was “obvious”); Portland Gen. Elec.
Co. v. Bonneville Power Admin., 501 F.3d 1009, 1024 (9th Cir. 2007)
(“In general, we will not invoke the waiver rule in our review of
a notice-and-comment proceeding if an agency has had an opportunity
to consider the issue.”).
30
not in accordance with law.” 42 U.S.C. § 9613(j)(2). The “law”
with which EPA must comply in selecting a remedial action is
primarily found in CERCLA and the NCP. EPA’s decisions made within
that legal framework will qualify as “arbitrary and capricious” if
EPA
fails
to
“examine
the
relevant
data
and
articulate
a
satisfactory explanation for its action.” FCC v. Fox Television
Stations, Inc., 556 U.S. 502, 513–14 (2009) (quoting Motor Vehicle
Mfrs. Ass’n of United States, Inc. v. State Farm Mut. Auto. Ins.
Co., 463 U.S. 29, 43 (1983)).
This is a “narrow” standard of review. Id. EPA’s explanation
need only be “plausible in light of the record as a whole” and
“supported
by
substantial
evidence
in
the
record.”
Leahy
v.
Raytheon Co., 315 F.3d 11, 17 (1st Cir. 2002) (citations omitted).
Furthermore, in reviewing the evidence, the Court is not permitted
“to substitute its judgment for that of the agency.” Fox Television
Stations, 556 U.S. at 513–14 (quoting Bowman Transp., Inc. v.
Arkansas–Best Freight Sys., Inc., 419 U.S. 281, 286 (1974)).
Particularly when the Court reviews “a purely factual question
within the area of competence of an administrative agency . . .
and when resolution of that question depends on ‘engineering and
scientific’
considerations,”
the
Court
must
“recognize
the
relevant agency’s technical expertise and experience, and defer to
its analysis unless it is without substantial basis in fact.”
Browning-Ferris Indus. of S. Jersey, Inc. v. Muszynski, 899 F.2d
31
151, 160 (2d Cir. 1990) (quoting Federal Power Commission v.
Florida Power & Light Co., 404 U.S. 453, 463 (1972)).
While certainly deferential, the arbitrary and capricious
standard is not a shibboleth by which EPA may completely avoid
judicial scrutiny. As the Supreme Court has explained, an action
may qualify as arbitrary and capricious where the administrative
agency: (1) “relied on factors which Congress has not intended it
to consider,” (2) “entirely failed to consider an important aspect
of the problem,” (3) “offered an explanation for its decision that
runs
counter
to
the
evidence
before
the
agency,
or
is
so
implausible that it could not be ascribed to a difference in view
or the product of agency expertise,” or (4) seeks to have its
action upheld based on “post hoc rationalizations.” Motor Vehicle
Mfrs. Ass’n of U.S., Inc., 463 U.S. at 43. Courts also review
whether an administrative agency has treated similar cases with
“apparent irrational discrimination.” Puerto Rico Sun Oil Co. v.
EPA,
8
F.3d
73,
78
(1st
Cir.
1993)
(citing
Green
Country
Mobilephone, Inc. v. FCC, 765 F.2d 235 (D.C. Cir. 1985)). Put
simply, the Court must review EPA’s analysis to ensure that it is
“rational” and “makes sense.” Penobscot Air Servs., Ltd. v. FAA,
164 F.3d 713, 720 (1999) (quotations, citations, and brackets
omitted).
Emhart’s
arguments
as
to
why
EPA’s
actions
qualify
as
arbitrary and capricious or otherwise not in accordance with law
32
address several aspects of EPA’s processes as well as the merits
of the selected remedy itself. Each of Emhart’s arguments are
addressed below.
IV. Case Specific Findings of Fact and Conclusions of Law
A. Background
1. Findings of Fact
In 1996 dioxin was discovered in fish collected from the
Woonasquatucket
River.
(See
Phase
I
Findings
12.) 14
An
EPA
investigation of the surrounding area – later labeled as the
Centredale Manor Restoration Project Superfund Site - ensued. The
Site covers a three-mile stretch of the Woonasquatucket River,
which includes a nine-acre peninsula that has been identified as
the “Source Area” of the Site’s hazardous substances. (Id. at 1011.) The Source Area contains two elderly housing facilities (Brook
Village and Centredale Manor) and is bounded to the north by Smith
Street, to the south by Allendale Pond, to the west by the
Woonasquatucket River, and to the east by the “tailrace,” a remnant
of a narrow body of water used for water power by the mills that
once occupied the peninsula. (Id.)
After passing by the Source Area, the Woonasquatucket River
leads to the remainder of the Site. The river first runs into
Allendale Pond, a .65-mile-long dammed pond that spans fifteen
14
This and various other findings of fact are taken directly
from the Court’s Phase I Findings.
33
acres and has depths ranging from .5 to ten feet. 15 (ROD, US144412, 30.) The river next runs into Lyman Mill Pond, a .85-mile-long
dammed pond spanning twenty-four acres with depths similar to
Allendale Pond. (Id.) In between Allendale and Lyman Mill Ponds is
the Oxbow Area, a forty-acre forested wetland habitat adjacent to
the river channel. (U.S. Army Corps of Eng’rs, Oxbow Area Report,
US1227-0005.) There are also various abutting residential and
commercial properties throughout the Site. (Phase I Findings 1011.) 16
Ultimately, EPA determined that the entire Site, not just the
Source Area, was polluted by a variety of contaminants, including
dioxins
(2,
compounds,
3,
7,
8–TCDD,
polychlorinated
in
particular),
biphenyls,
volatile
semi-volatile
organic
organic
compounds, polycyclic aromatic hydrocarbons, and various metals.
(Phase I Findings 11-12.) This determination led EPA to list the
Site on the National Priorities List of Superfund sites in 2000.
(Id.)
The Site, in short, is complicated. It contains several types
of toxic hazardous waste, and the Site’s size and diversity require
EPA to consider various types of flora, fauna, physical features,
and human uses. As part of the cleanup effort EPA has required
15
Unless specified otherwise, the lengths, sizes, and depths
discussed by the Court are approximated.
16
The ROD contains a map of the Site. (ROD, US1444-0012.)
34
several
removal
and
remedial
actions.
These
involved
the
construction of three interim protective caps as well as one RCRA17
cap over portions of the Source Area 18; reconstruction of the
Allendale Dam and restoration of Allendale Pond to prevent further
downstream migration of contaminants; excavation and removal of
one hundred cubic yards of soil from eleven areas along Allendale
and Lyman Mill Ponds; and erection of fences along the residential
properties adjacent to the Site in order to prevent access to the
contamination. (ROD, US1444-15-16.) To varying degrees, Emhart has
participated in each of these removal actions. (Phase I Findings
n.110.)
Additionally,
EPA
has
pursued
a
“comprehensive”
remedial
action at the Site that will address “all current and potential
future risks caused by soil, sediment, groundwater and surface
water contamination.” (ROD, US1444-6.) The first step in this
process (after listing the Site on the NPL) was the remedial
17
RCRA stands for the “Resource Conservation & Recovery Act.”
See 42 U.S.C. §§ 6901 et seq. “Congress enacted RCRA, a
comprehensive environmental statute that governs the treatment,
storage, and disposal of solid and hazardous waste, based . . . on
its finding that waste disposal had become a national problem
requiring federal involvement.” AES Puerto Rico, LP v. TrujilloPanisse, 857 F.3d 101, 103 (1st Cir. 2017) (quotations and
citations omitted). RCRA governs the disposal of “hazardous waste”
under subtitle C. Id. (citing 42 U.S.C. §§ 6921-6939g).
18
The ROD contains a map of the work done on the Source Area.
(ROD, US1444-13.)
35
investigation (“RI”). EPA’s RI 19 utilized Site-specific data and
modeling to characterize the nature and extent of contamination at
the Site. (ROD, US 1444-16-17.) The culmination of this process
was the RI report released in 2005. (ROD, US 1444-16-17; see also
RI, US1098.) The RI included an assessment of the risks to human
health (the Baseline Human Health Risk Assessment, or “BHHRA”) 20
and ecology (the Baseline Ecological Risk Assessment, or “BERA”) 21
posed by the Site.
Using the information collected during the RI, EPA developed
remediation goals that, if achieved, would likely mitigate the
risks to human health and the environment identified in EPA’s risk
assessment. (Feasibility Study, US1254-73-84.) EPA conducted an
extensive feasibility study (“FS”) in order to determine which
remedial alternative would best achieve those targets. 22 The FS
19
The RI actually included a series of individual
investigations conducted between 1999 and 2004. (See ROD, US 144416-17.)
20
The BHHRA is comprised of several individual assessments,
including: (1) 2005 BHHRA (US1101-1103); (2) 2011 Supplemental
BHHRA & BERA (US1287); and (3) May 2012 Technical Memorandum on
Impact of Dioxin Reassessment (“2012 Technical Memorandum”)
(US1392).
21
The BERA is comprised of the 2004 BERA (US1040-1044) and
the 2011 Supplemental BHHRA & BERA (US1287).
22
In total, the FS consisted of 329 pages of analysis
exclusive of references and appendices. Additionally, EPA issued
an FS Addendum in September 2011 in order to incorporate
information brought to EPA’s attention after publication of the
FS. (See FS Addendum, US1311.)
36
divides the Site into five “action areas,” and provides potential
remedial actions for each. (FS, US1254-10-14.) As is required by
the NCP, the FS includes both an initial screening as well as a
detailed analysis of the remedial alternatives for each action
area. (FS, US1254-147-321.)
Based on the analysis in the FS, EPA drafted a Proposed
Remedial Action Plan (“PRAP”) in the fall of 2011. (See PRAP, US
1328.) This version of the remedy did not last long though. Soon
after its publication, EPA released a nationwide change to its
non-cancer toxicity value for dioxin. Since dioxin is present at
the Site, EPA was forced to issue a “Technical Memorandum” updating
the BHHRA, cleanup levels, and FS for the Site. (See 2012 Technical
Mem., US1392.) These findings required several changes to the PRAP
in the form of a PRAP Amendment. (See PRAP Amendment, US1393.)
While much of the PRAP went unchanged, the PRAP Amendment did
require, among other things, an expanded cleanup area at the Site.
These changes were published in July, 2012.
Both the PRAP and PRAP Amendment were subject to notice and
comment after their publication. The notice and comment period on
the PRAP and PRAP Amendment went from November 14, 2011 to March
2, 2012, and July 19, 2012 to September 17, 2012, respectively.
(ROD, US1444-24-25.) During that time EPA participated in public
hearings and also accepted comments from a variety of sources,
37
including Emhart. (See Emhart Comments on PRAP, US1383; Emhart
Comments on PRAP Amendment, US1418.)
On
September
28,
2012,
with
the
public
comment
period
complete, EPA issued its Record of Decision (“ROD”) explaining the
remedial action plan. (See ROD, US1444.) The ROD provides EPA’s
justification
for
“Responsiveness
the
chosen
Summary”
remedial
that
action
addresses
as
well
significant
as
a
public
comments submitted to EPA on the PRAP and PRAP Amendment. In the
end, the plan outlined in the ROD was substantially similar to the
plan provided for in the PRAP and PRAP Amendment.
The remedy, as described in the ROD, contains the following
basic characteristics. In the Source Area the ROD requires removal
and off-Site treatment or disposal of waste material, installation
of
a
RCRA
underground
C cap 23 over
utilities
existing
into
clean
surfaces,
corridors.
and relocation
(ROD,
of
US1444-7.)
Sediment and floodplain soil in Allendale and Lyman Mill ponds
will also be excavated. However, unlike in the Source Area, the
majority of the excavated material in and around the ponds will be
placed in a confined disposal facility (“CDF”) near the Site,
leaving only a small portion of the excavated material (estimated
at approximately ten percent of the total) to be shipped off-Site 24
23
See supra note 17.
24
EPA requires off-Site disposal or treatment only for
particularly
toxic
materials
that
exceed
EPA’s
treatment
38
for disposal or treatment. (Id. at 7.) A thin-layer cover will be
placed over the remaining contaminated areas in the Oxbow wetland
area and, if necessary, over remaining contaminated sediment in
the Woonasquatucket River. (Id.)
requires continuous
institutional
activities
Site,
Going
at
groundwater.
the
(Id.)
such
forward,
controls
as
Additionally,
that
construction
long-term
the
limit
and
ROD
certain
use
maintenance
of
and
monitoring is required, including of the CDF, damns, sediment,
water, and biota located at the Site. (Id.) In total, EPA estimates
that the remedial action will cost approximately $104,600,000.
(Id. at 213.)
Lastly, the ROD commits that EPA will collect additional
information during the remedial design phase. For instance, EPA
will have to determine the location of the CDF. (ROD, US1444-6.)
EPA must also collect additional soil samples to determine the
precise amount of excavation required. (ROD, US1444-304.) It is
possible that the remedy will change based on this information. 25
However, EPA maintains that it has the capacity to adequately
standards. Materials that do not exceed EPA’s treatment standards
may be disposed in a CDF.
25
One example of this has already occurred since the ROD’s
publication. Several third party defendants collected additional
information on the Source Area that led EPA to conclude that it
“no longer expects that the removal of waste material and off-site
treatment or disposal will be necessary in the Source Area.” (Gov’t
Post-Trial
Brief
n.137
(citing
Revised
Draft
Pre-Design
Investigation Final Report, US1459).)
39
address such changes through, for example, an amendment to the
remedial
action
plan
or
an
“explanation
of
significant
differences.” 26
EPA
has
compiled
an
extensive
administrative
record
to
document its remedy-selection process. The administrative record
includes essentially all documents related to the development,
creation, and implementation of the remedial action. The next step
in EPA’s process is to create a more in-depth design of the
remedial action plan and implement it. 27 However, before the final
design
and
implementation
of
the
remedy
occurred,
Emhart
challenged EPA’s proposed remedy on several grounds.
2. Conclusions of Law
As a general matter, the Court finds that EPA followed the
basic steps mandated by CERCLA and the NCP in developing its
remedial
action
for
the
Site.
These
legal
requirements
were
previously outlined by the Court. See supra Section II (“Remedy
Selection”).
individual
However,
actions
Emhart
and
argues
decisions
that
along
the
several
way
of
were
EPA’s
either
arbitrary, capricious, or not in accordance with CERCLA or the NCP
such that the remedy cannot withstand judicial scrutiny. The Court
addresses each of Emhart’s arguments below.
26
See supra note 9.
27
The NCP labels this as the “remedial design/remedial action
(RD/RA) stage.” 40 C.F.R § 300.435(a).
40
B. Excavation and Treatment of Soil and Sediment
1. Findings of Fact
a. Excavation of Allendale and Lyman Mill Ponds
EPA’s remedy calls for significant excavation of the sediment
in Allendale and Lyman Mill Ponds. The goal of excavation is to
achieve dioxin levels of approximately fifteen parts per trillion.
(FS, US1254-75.) In order to determine how much excavation will be
required to meet this target, EPA took sediment samples from both
Allendale
and
Lyman
Mill
Ponds.
Based
on
these
samples,
EPA
estimates the average excavation depth in Allendale Pond required
to achieve the target dioxin level is 2.2 feet. (ROD, US 1444170.) For Lyman Mill Pond, the average excavation depth required
to achieve EPA’s target is 2.7 feet. (Id. at 170.) Both of these
estimates assume .25 feet of over-excavation will occur. (Id.)
The data set used by EPA to come up with these estimates
included 250 data records from Allendale Pond (ranging from 0.5 to
twelve feet in depth) and 160 data records from Lyman Mill Pond
(ranging from 0.5 feet to four feet in depth). (ROD, US1444-303;
see also RI, US1098-29; FS, Tables G-3 and G-4, US1254-1458-1477.)
The data most heavily relied upon by EPA were core samples taken
in 2003 and 2005, which included ten sediment cores taken from
Allendale Pond and sixteen sediment cores taken from Lyman Mill
Pond.
(FS,
US1254-363-64.)
Each
core
includes
multiple
soil
samples all of which were collected with the specific objective of
41
detecting the vertical extent of contamination in the ponds. (Id.
at 42, 679.) To that end, EPA conducted laboratory analysis using
high resolution mass spectroscopy, a highly accurate method that
can detect dioxin in parts per trillion. (Dr. Medine Test., Trial
Tr. vol. 12, 13:16-14:18, ECF No. 495.)
Emhart’s expert, Mr. Loureiro, testified that these samples
were inadequate because only a small portion was taken at depths
greater than one foot. 28 Furthermore, Mr. Loureiro pointed out
that, in certain instances, dioxin levels were detected at levels
above fifteen parts per trillion at depths greater than EPA’s
anticipated excavation depths. 29 As such, Mr. Loureiro opined that
the available data was insufficient to accurately estimate the
vertical extent of the contamination and, in turn, the amount of
excavation that will be necessary to achieve EPA targets.
EPA
agrees
that
additional
sampling
is
needed.
As
EPA
explained in the Feasibility Study, “[t]he proposed cleanup areas
or remedial footprints are conceptual and more precise cleanup
28
Mr. Loureiro testified that he evaluated the data and found
only forty samples from Allendale Pond and eighteen samples from
Lyman Mill Pond at depths below one foot. (Mr. Loureiro Test.,
Trial Tr. vol. 6, 105:7-9, 113:23-117:1, ECF No. 453.)
29
Because EPA’s anticipated excavation depth is an average,
Emhart is able to point to samples where dioxin is still above
fifteen parts per trillion past the average excavation depth, and
EPA is able to point to samples where dioxin is less than fifteen
parts per trillion even before reaching the average excavation
depth.
42
footprints will be developed during the remedial design. For
example, additional coring will need to be performed at Allendale
and
Lyman
Mill
contamination.”
Ponds
to
confirm
the
(FS,
US1254-327-28.)
vertical
For
this
extent
of
reason,
the
EPA
committed itself in the ROD to “perform[ing] additional sampling
and analysis closer to the time of remediation to confirm the
sediment cleanup depth and volume.” (ROD, US1444-304.)
Given
the
uncertainty
as
to
the
vertical
extent
of
the
contamination, the exact amount of excavation required will likely
differ from the amount estimated in the ROD. Precision in this
area is likely impossible, however. As Mr. Loureiro pointed out,
“certainly in my experience with excavation of all types, even a
robust data set aren’t adequate to actually describe the conditions
you run into in the field.” (Mr. Loureiro Test., Trial Tr. vol. 6,
19:5-8.) This means that this component of the cost of the remedy
is
to
some
extent
uncertain
and
could
potentially
be
more
expensive.
Emhart did raise this general issue during the notice and
comment period on the PRAP. (Emhart PRAP Comments US1383-8 (“EPA
fails
to
adequately
define
the
volume
of
soil
and
sediment
requiring excavation.”); see also id. at 51.) EPA responded in the
ROD’s “Responsiveness Summary” by explaining that, in its opinion,
sufficient
data
had
been
collected
to
provide
excavation
estimates, particularly in light of the 0.25-foot over-excavation
43
allowance built into the estimate. (ROD, US1444-303-04.) EPA also
noted that additional sampling will be done during remedial design
in order to refine those estimates. (Id.) Lastly, to the extent
that the proposed level of excavation does not achieve target
dioxin levels, EPA suggests that “a 6-inch soil cover on the
sediment bottom” could be used where “additional excavation is not
feasible.” (Id.)
b. Excavation of the Oxbow Area and the Floodplain Soil
of Allendale and Lyman Mill Ponds
In
February
2012,
after
publication
of
the
PRAP,
EPA
established a nation-wide non-cancer toxicity value for dioxin.
This new information forced EPA to reevaluate its remedial design
for the Site. EPA did so by issuing a Technical Memorandum on the
Impact of Dioxin Reassessment that updated EPA’s human health risk
assessment
and
feasibility
study.
(2012
Technical
Memorandum,
US1392.) The analysis of each remedial action alternative did not
change significantly. (See ROD, US1444-347 (“Impacts resulting
from these changed conditions are presented in EPA’s May 2012
Technical Memorandum and are generally consistent with evaluations
presented in . . . the FS.”).)
However, EPA did determine that certain areas, not previously
identified in the PRAP, would require excavation. These areas were
primarily located in the Oxbow Area and floodplain soil around
Allendale and Lyman Mill Ponds. (See PRAP Amendment, US1393-002.)
44
EPA’s expanded excavation plan was documented in the PRAP Amendment
and later adopted in the ROD. (PRAP Amendment, US1393-7-9; ROD,
US1444-176-192.) Emhart submitted comments on the PRAP Amendment
in which it argued that EPA had not collected sufficient Sitespecific data in order to adequately characterize contamination at
the Site. (See Emhart Comments on PRAP Amendment, US1418-7-9.) 30
With regards to the floodplain soils, EPA determined which
areas would require excavation using two types of Site-specific
data: soil samples and FEMA floodplain maps. The soil samples were
collected during the remedial investigation on the eastern shore
floodplains 31 of the Allendale and Lyman Mill Pond reaches. EPA
collected 226 samples in total, 212 of which detected some level
of dioxin. (ROD, US1444-345.) Of those 212 samples detecting
dioxin, “there are approximately 100 sampling locations where
floodplain
residential-use
soil
samples
have
[dioxin]
concentrations greater than cleanup levels.” (Id.; see also 2012
Technical Mem., US1392-24.) Based on this data EPA expanded the
30
A review of those comments reveals that Emhart provided no
specifics for what level of sampling it believed was necessary
going forward or where that sampling needed to occur. Emhart simply
asserted that additional sampling is required.
31
EPA defined “floodplain areas” as “the area of water and
land inundated during the highest point of the base, or 100-year,
flood using maps prepared by the Federal Insurance Administration
of the Federal Emergency Management Agency . . . .” (ROD, US1444306.)
45
area requiring excavation and estimated an average excavation
depth of one foot. (See 2012 Technical Mem., US1392-62.) 32
Having established that unsafe levels of dioxin had migrated
into floodplain soils, EPA then used Federal Emergency Management
Agency (“FEMA”) floodplain maps to determine other areas that
likely contain similar dioxin levels. FEMA floodplain maps cannot
substitute for field samples in determining the precise nature or
extent of dioxin contamination in a given area. (Mr. Loureiro
Test., Trial Tr. vol. 5, 183:23-184:11, ECF No. 452.) However, as
Dr. Medine explained, such maps are commonly used “to aid in
designing sampling programs to characterize the nature and extent
of floodplain contamination” because they can act as “a guide to
where to look for contamination and serve as indicator of where
contamination may have come to rest.” (Dr. Medine Test., Trial Tr.
vol. 12, 81:13-15, 21-22, ECF No. 495.)
As for the Oxbow Area, EPA relied primarily on two Sitespecific data sets in order to estimate the amount of required
excavation. The first was a data set collected by EPA during the
remedial investigation. (FS, US1254-188.) The second was a data
set
collected
by
Emhart
as
part
32
of
its
2010
Oxbow
Area
EPA’s limited subsurface data suggests that excavation will
be limited to one foot in most areas, though some areas will also
likely require deeper excavation. (2012 Technical Mem., US139224.) The ROD explains that, generally speaking, excavation below
surface soils is unlikely because “dioxins do not migrate easily
through the soil column.” (ROD, US1444-346.)
46
Investigation. (ROD, US1444-318.) That investigation collected
forty-four
floodplain
soil
samples
and
twenty-eight
sediment
samples at depths ranging from zero to thirty-six inches. (ROD,
US1444-36.) After EPA came out with its toxicity value for dioxin,
EPA utilized both data sets in the FS Addendum. The FS Addendum
determined that 5,600 cubic yards of material needs to be excavated
in
addition
to
the
areas
identified
in
the
PRAP.
(See
PRAP
Amendment, US 1393-7, 10.)
For both the Oxbow Area and the floodplain soils the ROD
requires
that
EPA
collect
additional
information
during
the
remedial design phase to delineate the exact counters and depths
of excavation. (See, e.g., ROD, US1444-176, 346.) This information
will consist primarily of additional soil samples. (Id. at 176,
181-83.) Once excavation is complete, “confirmation sampling will
be conducted to verify that the cleanup levels are achieved . . .
.” (Id. at 177; see also id. at 183.)
c. Amount of Soil Requiring Incineration
Once contaminated soil and sediment are excavated, the ROD
calls for those materials to be either stored in a CDF or shipped
off-Site for treatment (i.e., incineration). (ROD, US1444-172-73.)
Whether treatment is required depends on the toxicity of the
material. Material that contains dioxin at or above ten parts per
billion must be treated; material that contains dioxin below ten
parts per billion can be stored in a CDF. (Id. at 172.)
47
At
the
Site,
EPA
estimates
that
ninety
percent
of
the
excavated soils can be disposed of in a CDF while ten percent will
need to be shipped off-Site for treatment. (ROD, US1444-172.) As
with EPA’s excavation estimates, certainty in this area is likely
impossible. (See, e.g., Mr. Loureiro Test., Trial Tr. vol. 6, 19:58.)
Emhart’s
2010
Oxbow
Area
Investigation
exemplifies
this
uncertainty. EPA initially estimated that no Oxbow Area floodplain
soil would require incineration. (ROD, US1444-318; FS Addendum,
US1311-27.) However, after more extensive sampling was conducted
in the 2010 Oxbow Area Investigation, EPA changed its opinion and
now estimates that up to ten percent of the Oxbow Area floodplain
soil may require treatment. (ROD, US1444-318; FS Addendum, US131127.) As this example demonstrates, additional sampling can reveal
that more material requires treatment than initially estimated
thereby increasing the cost of the overall remedy.
However,
uncertainty
does
not
necessarily
mean
that
the
ultimate remedy will involve additional treatment or increased
cost compared to the estimates in the ROD. The Source Area provides
an
example
where
the
amount
of
material
requiring
treatment
decreased as a result of additional sampling. At the time the ROD
was published, EPA “assumed that all of the excavated potential
buried waste material [in the Source Area] would be taken to an
off-site incinerator for treatment.” (ROD, US1444-164; see also
id. at 214.) However, the ROD also called for additional sampling
48
“to determine off-site treatment requirements.” (Id. at 164.)
After
publication
of
the
ROD,
several
third-party
defendants
collected additional information on the Source Area that led EPA
to conclude that it no longer anticipates that the Source Area
material will require off-Site treatment. (See Revised Draft PreDesign Investigation Final Report, US1459.)
The lack of certainty in the ROD as to the amount of material
requiring treatment was addressed by Emhart during the notice and
comment period on the PRAP. As Emhart explained, “EPA’s estimate
that only 10% of the soil and sediment will exceed the alternative
treatment standards for soil is based on limited data. . . . [I]t
appears possible that a much higher percentage of soil and sediment
will
contain
dioxin
at
concentrations
above
the
standard,
therefore requiring off-Site disposal.” (Emhart PRAP Comments,
US1383-71.)
EPA responded to Emhart’s comment in the ROD’s Responsiveness
Summary.
EPA
explained
that
its
estimates
were
based
on
approximately 400 sediment samples from Allendale and Lyman Mill
Ponds and 250 floodplain soil samples (not including Source Area
soil). (ROD, US1444-318.) Dr. Medine also described how those
samples were used in combination with a contouring analysis in
order to estimate the volume of material requiring treatment. (See
Dr. Medine Test., Trial Tr. vol. 12, 28:23-29:13; FS, US1254-147879.) As discussed above, EPA has committed to refining these
49
estimates through further sampling and analysis during remedial
design.
2. Conclusions of Law
Emhart argues that EPA failed to collect sufficient Sitespecific data to conduct a feasibility study in accordance with
the NCP. Specifically, Emhart argues that EPA did not collect
enough Site-specific data to accurately estimate the amount of
material that will have to be excavated or shipped off-Site for
treatment. Emhart’s comments on the PRAP and PRAP Amendment did
not provide any specifics as to the number of samples it believes
need
to
be
collected
from
the
Site
in
order
to
conduct
a
feasibility study under the NCP. However, Emhart did raise this
issue as a general matter by discussing the need “to adequately
define the volume of soil and sediment requiring excavation” or
“off-Site disposal” (Emhart PRAP Comments, US1383-8, 71), and EPA
had the opportunity to respond. (See, e.g. ROD, US1444-303-04,
318.)
Under
these
circumstances
the
Court
finds
that
Emhart
adequately preserved this issue.
As to the merits of Emhart’s arguments, both Emhart and EPA
agree that additional sampling is needed in order to characterize
the lateral and vertical extent of contamination at the Site and
obtain
a
more
accurate
excavation
and
incineration
volume
estimate. The disagreement is about the timing of that additional
data collection. Emhart argues that such data should have been
50
collected for use in the feasibility study. EPA contends that it
collected sufficient data to conduct a feasibility study and that
the necessary additional sampling can be done during remedial
design.
While
the
NCP
requires
that
EPA
conduct
a
remedial
investigation and feasibility study that includes the collection
of site-specific data, it does not require a specific number of
samples or data points. Instead, EPA must collect sufficient data
“to adequately characterize the site for the purpose of developing
and
evaluating
effective
remedial
alternatives.”
40
C.F.R.
§ 300.430(d)(1); see also id. § 300.430(a)(2) (requiring EPA to
“assess site conditions . . . to the extent necessary to select a
remedy”). One of the factors that EPA must be able to “adequately
characterize” is the cost of the remedy. Id. § 300.430(e)(7)(ii)
(cost
is
an
initial
screening
factor);
see
also
id.
§ 300.430(e)(9)(iii)(G) (cost as one of nine factors to be balanced
during remedy selection).
Beyond these general guidelines, EPA is given significant
leeway to develop a remedial investigation and feasibility study
process specific to a site. This leeway includes the option of
reserving a certain amount of data collection for remedial design.
The NCP specifically allows the ROD to reserve certain decisions
for a later date and, “[w]hen appropriate, provide a commitment
for
further
analysis
.
.
.
.”
51
Id.
§
300.430(f)(5)(iii)(D).
Therefore, while the information contained in the feasibility
study provides the “initial building block in developing” the final
design,
EPA
guidance
also
envisions
that
additional
“data
acquisition” and “sample analysis” may be necessary during the
remedial
design
phase.
EPA,
Scoping
the
Remedial
Design,
Emhart516-1-2.
There is no doubt that the amount of soil that must be
excavated or shipped off-Site for incineration will affect the
cost of the remedy. The question before the Court is whether EPA
collected sufficient data to “to adequately characterize” the
remedial alternatives, including their estimated costs, such that
those
alternatives
could
be
effectively
compared
during
the
feasibility study. The Court finds that EPA has met that standard.
As described in the findings of fact, EPA analyzed multiple
data sets that include hundreds of soil and sediment samples from
across
the
Site.
In
examining
these
samples
EPA
used
high
resolution mass spectroscopy to detect dioxin levels down to the
parts per trillion. The results of this sample analysis were
reviewed in conjunction with other types of information, such as
FEMA floodplain maps and contouring analysis, to further refine
EPA’s estimates. It is true that in none of the Site areas
discussed in this section – Allendale and Lyman Mill Ponds, the
floodplain soils, or the Oxbow Area – has EPA collected sufficient
data to fully implement the remedy. As EPA recognizes, additional
52
sampling and analysis will be required during remedial design. But
this sort of commitment to further analysis is permitted by the
NCP and EPA guidance. At this point in the remediation process EPA
is only required to have collected sufficient data to “adequately
characterize” the estimated excavation and treatment volumes such
that remedial alternatives can be compared in a feasibility study,
which EPA has done.
While
the
Court
finds
that
the
current
state
of
EPA’s
excavation and treatment estimates does not violate the NCP, this
finding
is
additional
based,
sampling
in
part,
and
on
analysis
EPA’s
commitment
during
remedial
to
conduct
design.
If
information uncovered during remedial design reveals that the cost
will differ significantly from the cost outlined in the ROD, EPA
has the responsibility to update the administrative record as
necessary.
This
could
potentially
require
an
“explanation
of
significant differences,” a ROD amendment, or even an updated
feasibility study. EPA has shown a willingness to take these sorts
of actions in the past. (See, e.g., FS Addendum, US1311; PRAP
Amendment, US1328.) And while CERCLA and the NCP leave it up to
EPA to determine whether such actions are necessary in the first
instance, the Court retains jurisdiction over this matter in the
event that Emhart, in light of new evidence, seeks to challenge
EPA’s decision.
53
C. Location of the Confined Disposal Facility
1. Findings of Fact
The ROD calls for the majority of excavated soil to be placed
in
a
CDF.
(ROD,
US1444-167.)
EPA
initially
identified
three
locations at the Site that could potentially accommodate a CDF.
(PRAP Amendment, US1328-21; see also FS, US1254-389.) However, as
Emhart noted in its comments on the PRAP, EPA did not establish
the final location of the CDF. (Emhart Comments on PRAP, US138365.) In addition, the Town of Johnston has made it clear that it
does not want a CDF placed at or near the Site. 33
Despite these concerns, the ROD states that a CDF will be
part of the final remedy. The ROD explains that EPA “continues to
believe the upland CDF disposal option is the best approach to
address contaminated sediment/soil.” (ROD, US144-172.) As for the
exact location, the CDF could still be located on-Site. However,
given the lack of community acceptance for this option, EPA has
also “expanded the area where an upland CDF could be located to
locations outside the Town of Johnston and beyond what is in very
close proximity to the Site.” (Id. at 172.) The ROD further commits
33
The Town was initially supportive of the CDF being located
in Johnston. (See, e.g., PRAP Hearing at Centredale Manor, US133330.) The Town has since sent a letter to EPA explaining that it
“is no longer supportive of consideration of any upland disposal
of contaminated material associated with the Centredale Manor
Superfund Site.” (Letter from Town of Johnston to EPA dated
12/29/2011, US1344-1.)
54
EPA “to identify additional locations where an upland CDF could be
located” as part of remedial design. (Id. at 172.)
EPA has already begun this process. (See Technical Mem. for
Assessment
of
CDF,
US1475.)
However,
to
date,
EPA
has
not
identified a location for the CDF. Whether EPA will be able to
identify
a
location
in
close
proximity
to
the
Site
that
is
acceptable to the local community is unclear. If a CDF is not
utilized as part of the remedy the cost of the remedy will likely
rise significantly. Given the local community’s resistance to a
CDF and the likely cost increase that would result if a CDF is not
built, Emhart’s expert, Mr. Loureiro, provided his opinion that
“there should be a defined location for the CDF, whether it’s
purchased or under contract or part of the site itself” before the
remedy is selected. (Mr. Loureiro Test., Trial Tr. vol. 10, 10:810.) Having a “defined location,” according to Mr. Loureiro, is
the only way to ensure that the proposed remedy is “feasible” and
therefore a “viable alternative.” (Id. at 10:11-12.) EPA disagrees
and “believes a location can be identified [during remedial design]
that addresses most or all of the concerns raised by the public.”
(ROD, US1444-172.)
2. Conclusions of Law
Emhart contends that the CDF discussed in the ROD “does not,
and will never, exist.” (Emhart Post-Trial Brief 175.) It argues
that EPA’s failure to identify a “defined location” for the CDF in
55
the ROD violates the NCP. (Id. at 175; see also Mr. Loureiro Test.,
Trial Tr. vol. 10, 10:8-13.) EPA disagrees and argues that the
location of a CDF can be determined during remedial design. Emhart
preserved this issue in its comments on the PRAP. (Emhart Comments
on PRAP, US1383-65.)
The Court finds that EPA’s decision to determine the location
of the CDF during remedial design does not constitute a violation
of the NCP. There are two problems with Emhart’s argument. First,
Emhart’s claim that the CDF can “never” exist is premised on the
idea that the Town of Johnston has the authority to prohibit EPA
from constructing a CDF at the Site. But this is not the case.
See, e.g., 42 U.S.C. § 9604(j); id. § 9621(e)(1); 40 C.F.R.
§ 300.400(d). And while the Town of Johnston has made clear that
it does not wish to provide a space for a CDF, community acceptance
is only one of nine factors that EPA must consider. See 40 C.F.R.
§ 300.430(e)(9)(iii)(A)-(I). Therefore, it is possible that EPA
could determine, in its final analysis, that other factors outweigh
a lack of community acceptance and override the Town of Johnston.
Second, there is no provision of the NCP requiring the ROD to
provide
a
“defined
location”
for
a
CDF.
Nor
is
there
any
requirement that the land on which a CDF will be located must be
either purchased or under contract before publication of the ROD.
Instead,
EPA
need
only
collect
sufficient
information
to
effectively compare remedial alternatives. EPA has expressed its
56
judgment that it will be able to find a suitable location for a
CDF during remedial design and compared remedial alternatives
accordingly. (ROD, US1444-172.) This does not violate the NCP.
Of course, as discussed in the preceding section, EPA is
required to document all significant changes made during remedial
design. Therefore, if EPA ultimately does not select a location
for the CDF in accordance with the cost estimate and description
in the ROD, EPA must document that fact in accordance with the
NCP. This could potentially require additional analysis by EPA of
remedial
alternatives
and
possible
challenges.
At
this
point
though, before EPA has completed its remedial design, the lack of
a defined location for the CDF does not constitute a violation of
the NCP.
D. Dewatering Sediment, Controlling Ground and Surface Water,
and Constructing Haul Roads
1. Findings of Fact
The ROD calls for significant excavation of contaminated
sediment at the Site. This will require some level of ground and
surface
water
controls
in
order
to
access
the
contaminated
sediment, the construction of haul roads to move the sediment after
excavation, and the creation of a dewatering facility to remove
the
water
from
the
sediment.
At
trial
Emhart’s
expert,
Mr.
Loureiro, opined that these activities will require funding that
is not provided for in EPA’s cost estimate.
57
For instance, Mr. Loureiro testified that, while the remedy
will require a dewatering facility, the ROD does not account for
the cost of creating and removing that facility. (Id. at 147:412.) Mr. Loureiro also testified that controlling ground and
surface water will require a complex system of bypass pumping, and
sheet piling will be necessary to separate excavation areas from
river flow. (See, e.g., Mr. Loureiro Test., Trial Tr. vol. 6, 7:118:8; Mr. Loureiro Test., Trial Tr. vol. 10, 23:2-19.) Lastly, Mr.
Loureiro described how that the ROD does not allocate sufficient
funds for the construction of haul roads that will be necessary to
transport contaminated sediment and soil. (Id. at 145:9-147:4.)
Emhart never mentioned any of these costs in its extensive
comments on the PRAP or PRAP Amendment. Additionally, while the
Court found Mr. Loureiro’s overall presentation persuasive, there
is
reason
to
believe
that
his
cost
estimates
are
somewhat
exaggerated. EPA’s expert, Dr. Medine, noted several instances
where
Mr.
Loureiro
may
have
overlooked
less
expensive
alternatives. 34 Mr. Loureiro also did not address all of the cost
allocations already provided for by EPA’s estimate, including
34
For example, Dr. Medine called into question whether Mr.
Loureiro’s proposed bypass pumping system was necessary in light
of cheaper alternatives such as installing culverts. (See, e.g.,
Dr. Medine Test., Trial Tr. vol. 12, 71:24-72:10.) Dr. Medine also
called into question whether the cost of Mr. Loureiro’s proposed
dewatering facility was necessary in light of a smaller (i.e.,
cheaper) alternative. (Id. at 51:24-52:2.)
58
approximately $10,800,000 for dewatering contaminated sediment 35,
$3,400,000 for acquisition, installation and removal of sheet
piling 36, and $1,800,000 for construction of haul roads. 37 Lastly,
Mr. Loureiro’s cost estimate for water management assumes that the
river channel adjacent to the Source Area must be excavated despite
the fact that the ROD calls for dredging, not excavation, of that
area. (ROD, US1444-170; see also FS, US1254-228.)
2. Conclusions of Law
Emhart argues that EPA failed to take into account several
costs associated with the remedy, including costs of dewatering
sediment, controlling surface and ground water, and construction
of haul roads. None of these arguments were provided to EPA during
the notice and comment period on the PRAP or PRAP Amendment. (See
Emhart Comments on PRAP, US1383; Emhart Comments on PRAP Amendment,
US1418.) As such, Emhart’s arguments related to these topics are
waived. See Upper Blackstone, 690 F.3d at 30. In addition, Emhart
35
(See FS, US1254-1647:48-51 (describing approximately
$8,613,000 in costs); see also id. at 1635 (providing for a 25%
contingency).)
36
(See FS Addendum, US1311-868:20-21 (estimating $804,600 to
purchase the sheet piling and $1,933,742 to install and remove
it); see also FS, US1254-1635 (providing for a 25% contingency).)
Of note, EPA also allocated a “unit cost” for excavation, which
could potentially provide funding for other water management
requirements. (See FS Addendum, US1311-889.)
37
(See ROD, US1444-198 (estimating cost for “mobilization and
temporary roads”); see also FS Addendum, US1311-868 (explaining
costs associated with haul roads).)
59
has failed to demonstrate that EPA miscalculated costs for the
above-mentioned activities and that those miscalculations were
“key assumptions” on which EPA based its selected remedy. See
Oklahoma
Dep’t
of
Envtl.
Quality,
740
F.3d
at
192
(quoting
Appalachian Power Co., 135 F.3d at 818).
EPA
estimates
that
the
remedy
will
cost
approximately
$104,600,000. (ROD, US1444-213.) This estimate includes funds for
dewatering contaminated sediment, controlling ground and surface
water, and constructing haul roads. While Mr. Loureiro believes
EPA’s estimates in these areas are low, it is unclear to what
extent. But regardless of the exact number, any miscalculation by
EPA in these areas likely constitutes a relatively small percentage
of the total cost of the remedy, and any necessary recalculation
is unlikely to fundamentally alter EPA’s remedy selection. Under
these circumstances, the Court finds that EPA did not fail to
justify
“key
assumptions”
such
that
an
exception
to
issue
exhaustion is warranted.
E. Baseline Ecological Risk Assessment
1. Findings of Fact
EPA conducted a remedial investigation of the Site that
included an assessment of ecological risks. The results of that
investigation can be found in EPA’s Baseline Ecological Risk
60
Assessment,
demersal
or
“BERA.” 38
fish,
The
pelagic
BERA
fish,
characterized
piscivorous
the
risks
wildlife,
to
and
insectivorous wildlife, and the FS provided proposed cleanup goals
based on those risks. (See FS, US1254-930-33.) However, EPA’s
remedial goals at the Site were primarily based on risks to human
health, not risks to ecological receptors. (See, e.g., id. at 73
(“Biota PRGs are not presented in this FS because the sediment
PRGs (based on fish consumption and direct contact/incidental
ingestion) were used to determine proposed cleanup areas . . .
.”); see also id. at 930-33; ROD, US144-120-22.) Furthermore,
Emhart’s objections to EPA’s ecological risk calculations were not
raised in its comments on the PRAP or PRAP Amendment. (See Emhart
Comments on PRAP, US1383; Emhart Comments on PRAP Amendment,
US1418.)
2. Conclusions of Law
At
no
point
in
Emhart’s
comments
on
the
PRAP
and
PRAP
Amendment did Emhart address potential flaws in EPA’s ecological
risk assessment, including EPA’s findings on risks to demersal
fish,
pelagic
fish,
piscivorous
38
wildlife,
and
insectivorous
See supra note 21. The focus of the BERA is the “actual or
potential impacts of site contaminants on plants and animals. . .
. And the objectives of an ecological risk assessment . . . is to
identify and characterize the current and potential threats to the
environment” and “to identify cleanup levels that would protect
those natural resources from risk.” (Dr. Keenan, 138:24-139:8, ECF
No. 451.)
61
wildlife. (See Emhart Comments on PRAP, US1383; Emhart Comments on
PRAP
Amendment,
US1418.)
Emhart’s
arguments
related
to
these
topics are therefore waived. Upper Blackstone, 690 F.3d at 30. In
addition,
Emhart
determinations
has
related
failed
to
to
demonstrate
ecological
risks
that
constitute
EPA’s
“key
assumptions” on which EPA’s selected remedy is based such that an
exception to the doctrine of issue exhaustion is appropriate. See
Oklahoma
Dep’t
of
Envtl.
Quality,
740
F.3d
at
192
(quoting
Appalachian Power Co., 135 F.3d at 818).
A review of both the FS and the ROD reveals that EPA’s
remedial goals at the Site were primarily based on risks to human
health, not risks to ecological receptors. (See, e.g., FS, US125473, 930-33; ROD, US144-120-22.) Even if the ecological risks were
somehow
mitigated,
EPA
has
made
clear
that
it
would
pursue
essentially the same remedial design based on the risks to human
health posed by the Site. (See, e.g., Gov’t Post-Trial Brief 87
(explaining that “the results of the ecological risk assessment
had little impact on the selection of the cleanup remedy” because
“site cleanup was not driven by unacceptable ecological risk.”).)
As such, Emhart has failed to show how a change to EPA’s ecological
risk
determinations
would
fundamentally
62
change
EPA’s
selected
remedy such that it constitutes a “key assumption” on which that
remedy is based. 39
F. Vernal Pool Habitats Located in the Oxbow Area
1. Findings of Fact
A vernal pool is a “seasonal pool of water that can provide
[a] habitat for some plants and animals.” (Dr. Keenan Test., Trial
Tr. vol. 5, 92:9-10, ECF No. 452.) The 2004 BERA noted that
“[s]everal likely vernal pools were observed in the forested
floodplain area downstream from the Allendale damn.” (2004 BERA,
US1040-49.) The ROD also noted the potential presence of vernal
pools in the Oxbow Area. (See ROD, US1444-181-82; 2011 Supplemental
BHHRA & BERA, US1287-218.)
EPA’s proposed remedy for the Oxbow Area includes excavation
and the installation of a thin-layer cover. (ROD, US1444-7.) If
vernal pools are found in the Oxbow Area, EPA recognizes that the
current remedy “will face additional implementation issues.” (FS,
US1254-282.)
The
“implementation
issues”
consist
primarily
of
taking “special care” not to disturb the vernal pool habitats
during design and construction of the thin-layer cover. (ROD,
39
The Court notes that, just as Emhart cannot establish that
the remedy is invalid based on EPA’s ecological risk calculations,
EPA also cannot justify the proposed remedy based on those same
ecological risk calculations. As EPA has argued, and the Court has
agreed, EPA’s selected remedy was the result of a process that
focused on the risks to human health at the Site. EPA’s selected
remedy must therefore stand or fall based on those human health
risk determinations.
63
US1444-156.) In light of these findings, EPA has committed to
conducting
“pre-design
and
design
investigations
[that]
will
include physical and ecological surveys to further . . . identify
any potential vernal pools.” (Id. at 181-82.)
Emhart did not raise any issues related to vernal pools in
either
its
comments
on
the
PRAP
or
the
PRAP
Amendment.
Furthermore, at trial, Emhart’s expert (Dr. Keenan) could not
testify that EPA’s remedy would cause damage to any potential
vernal pools. (Dr. Keenan Test., Trial Tr. vol. 5, 92:12-16.)
Instead, Dr. Keenan merely suggested that “it’s something that
needs to be investigated before it’s implemented.” (Id.)
2. Conclusions of Law
Emhart failed to address the potential presence of vernal
pool habitats in the Oxbow Area in its comments on the PRAP and
PRAP Amendment. (See Emhart Comments on PRAP, US1383; Emhart
Comments on PRAP Amendment, US1418.) As such, Emhart’s arguments
on this topic are waived. See Upper Blackstone, 690 F.3d at 30.
Moreover,
Emhart
confirming
the
has
failed
presence
of
to
demonstrate
vernal
pools,
that
EPA,
disregarded
by
a
not
“key
assumption” on which the remedy is based. See Oklahoma Dep’t of
Envtl. Quality, 740 F.3d at 192 (quoting Appalachian Power Co.,
135
F.3d
at
818).
The
presence
of
vernal
pools
will
not
fundamentally change the remedy as a whole, but instead simply
require that “special care” be taken not to disturb the vernal
64
pools in one of the five action areas at the Site. (ROD, US1444156.) EPA has committed to further investigation of this issue
during remedial design and can adjust the remedy as needed.
G. Classifying Site Soils and Sediments as Principal Threat
Waste and F020 Listed Waste
1. Findings of Fact
EPA classified certain Site soils and sediments as Principal
Threat Waste (“PTW”) 40 and F020 listed waste. 41 Emhart challenged
these classifications in their comments on the proposed remedy.
(See Emhart Comments on Proposed Plan, US1383-83-90.) However, the
ROD makes clear that the PTW and F020 classifications had no
practical impact on remedy selection. (See, e.g., ROD, US1444-332
(explaining the PTW classification “had no practical effect on the
selected
remedy”);
classification
[conceptual
Loureiro,
“is
site
id.
at
not
important
model]”).)
essentially
288
At
conceded
(explaining
for
trial
this
the
that
purposes
Emhart’s
point.
the
F020
of
the
expert,
Mr.
(See,
e.g.,
Mr.
Loureiro Test., Trial Tr. vol. 6, 161:12-15 (“And ultimately, at
the end of the day it doesn't appear that [the PTW classification
40
PTW is defined as material that is highly mobile and
contains high concentrations of toxic compounds. 40 C.F.R.
§ 300.430(a)(1)(iii)(A).
41
F020 listed waste is defined as “[w]astes . . . from the
production or manufacturing use . . . of tri- or tetrachlorophenol,
or of intermediates used to produce their pesticide derivatives.”
Id. § 261.31.
65
is] being used for any purpose in defining the scope of the work
that needs to be done.”); id. at 170:11-12 (“I don’t think [the
F020 classification has] any practical implication with regard to
the costs of the remedy.”).)
2. Conclusions of Law
Emhart argues that EPA’s classification of certain soils and
sediments
as
PTW
and
F020
listed
waste
was
arbitrary
and
capricious. The Court finds that this question is moot. The ROD
makes clear that these classifications had no practical impact on
remedy selection. (See, e.g., ROD, US1444-288, 332.) Emhart agrees
that the PTW classification “was not used for any purpose in
defining the scope of work to be done.” (Emhart Post-Trial Brief
228; see also Mr. Loureiro Test., Trial Tr. vol. 6, 161:12-15.)
And Emhart’s expert essentially conceded this same point with
respect to the F020 classification. (See Mr. Loureiro Test., Trial
Tr.
vol.
6,
170:11-12.)
The
Court
need
not
address
the
appropriateness of classifications that, if changed, would have no
practical impact on the remedial action itself.
H. Classifying Source Area Ground Water as Drinking Water
1. Findings of Fact
For
years
the
Rhode
Island
Department
of
Environmental
Management (“RIDEM”) has considered the Source Area groundwater to
be so contaminated that it is unsuitable for potential use as
drinking water. (FS, US1254-43.) The contamination is caused by
66
the Source Area, but also possibly by waste sites located close to
the Woonasquatucket River upgradient from the Source Area. (ROD,
US1444-58.)
EPA’s
remedial
investigation
emphasized
RIDEM’s
classification of the Source Area groundwater as non-potable and
discussed several important uncertainties in its groundwater data,
including the “vertical extent” of the contamination. (RI, US109888.)
Based
on
this
information,
the
FS
determined
that
the
groundwater was not a potential source of drinking water, which
EPA labels as “Class III” groundwater. (FS, US1254-43.)
EPA
later
determined
that
its
classification
of
the
groundwater as Class III was not appropriate. As explained in the
ROD, RIDEM’s groundwater classification system had not “obtained
EPA approval of a Comprehensive State Ground Water Protection
Program.” (ROD, US1444-57.) 42 EPA therefore could not base its
classification decision on RIDEM’s findings, and had to instead
focus on the standards set by EPA guidance. (Id. at 343.) Citing
these standards, the ROD explains that the Class III designation
does not apply because the groundwater was not “so contaminated by
naturally occurring conditions or the effects of broad-scale human
activity (unrelated to a specific activity) that [it] cannot be
42
This fact was brought to EPA’s attention by the National
Remedy Review Board, which “is a committee of EPA regional
personnel that provide technical assistance and review large,
complex, and costly Superfund sites.” (Mr. Maccarone Test., Trial
Tr. vol. 2, 169:25–170:9.)
67
cleaned up using treatment methods reasonably employed in public
water supply systems.” (Id. at 344.) This led EPA to label the
Source Area groundwater as “Class II,” which applies to “current
and potential sources of drinking water.” (Id.)
EPA made this determination despite the fact that EPA has
recognized data gaps for Source Area groundwater. (See, e.g., RI,
US1098-88.) For example, it is uncertain to what extent dense nonaqueous phase liquid (“DNAPL”) is present in the Source Area and
at what depths. (Compare Mr. Loureiro Test., Trial Tr. vol. 6,
172:9-173:6 with Dr. Medine Test., Trial Tr. vol. 12, 88:1-89:4.) 43
The
vertical
extent
of
that
(and
other)
contamination
is
particularly important because the remedy – which partially relies
on a RCRA C cap to prevent further contamination – will not be
effective in stopping contamination sources located below the cap.
(Mr. Loureiro Test., Trial Tr. vol. 6, 177:24-178:12; id. at
171:17-25; see also Mr. Loureiro Test., Trial Tr. vol. 5, 143:24144:17.)
43
EPA has collected some data that suggests DNAPL is not
present in the Source Area groundwater. (See, e.g., RI, US1098205-06; ROD, US1444-197.) However, EPA has also found some
groundwater contamination in intermediate and deep samples in the
southern end of the Source Area, the origin of which is not clear.
(RI, US1098-54.) This issue was likely not more thoroughly
investigated during the remedial investigation because, at that
time, EPA did not consider the groundwater to be a potential source
of drinking water.
68
In addition to possible DNAPL contamination, EPA recognizes
that
“there
are
numerous
non-Superfund
sources
.
.
.
that
contribute or have the potential in the future to contribute to
exceedances of drinking water standards away from the Source Area.”
(ROD, US1444-58.) Under these conditions, the ROD explains that
“[f]uture
groundwater
uses
are
not
expected
to
change
significantly” and that water will continue to be supplied to the
Source Area by outside sources. (Id.; id. at 62.) Nevertheless,
EPA persists in classifying the groundwater as a potential source
of drinking water.
The change in classification from Class III to Class II led
EPA
to
adopt
stricter
cleanup
goals
for
the
Source
Area
groundwater. (FS Addendum, 1311-6, 79.) EPA also used the new
classification as part of its justification for requiring a RCRA
C cap over the Source Area. (ROD, US1444-339.) As explained in the
ROD, the RCRA C cap will be used to prevent the migration of
contamination from the soil to the groundwater. (ROD, US1444-7.)
However, it is unclear whether this will be sufficient to prevent
further
Source
Area
groundwater
contamination.
(Id.
at
342.)
Therefore, the ROD calls for groundwater monitoring wells to be
installed so that EPA can determine whether federal drinking water
standards are achievable. (Id. at 7.)
69
2. Conclusions of Law
Emhart argues that EPA inappropriately classified Source Area
groundwater as Class II as opposed to Class III. Emhart challenged
EPA’s classification of Source Area groundwater in its comments on
the PRAP. (Emhart PRAP Comments, US1383-102-05; see also ROD,
US1444-338.)
Emhart
has
therefore
preserved
this
issue
for
judicial review.
The NCP directs EPA “to return usable ground waters to their
beneficial uses wherever practicable, within a timeframe that is
reasonable given the particular circumstances of the site.” 40
C.F.R.
§
300.430(a)(1)(iii)(F).
Where
restoration
is
“not
practicable” EPA need not restore the groundwater, but must instead
take steps to “prevent further migration of the plume, prevent
exposure to the contaminated ground water, and evaluate further
risk reduction.” Id. To that end, EPA has established guidelines
that place groundwater into different “classes.” (ROD, US1444343.) Class II groundwater is considered a “current and potential
sources of drinking water.” (Id. at 344.) “Class III” groundwater,
on the other hand, is “not considered potential sources of drinking
water.” (Id.) However, even where EPA guidelines would classify
groundwater as Class II (i.e., a “potential” source of drinking
water),
the
application
NCP
still
requires
of
drinking
water
EPA
to
determine
standards
is
whether
the
“technically
impracticable.” 40 C.F.R. § 300.430(f)(1)(ii)(C)(3); see also 42
70
U.S.C. § 9621(d)(4)(C). If restoration of the groundwater is
impracticable, then a waiver is appropriate. Id.
In this case, EPA classified the groundwater as Class III in
the vast majority of the Site. However, for the Source Area, EPA
classified the groundwater as Class II. (ROD, US1444-342-43.) EPA
explains that this finding is based on EPA guidelines, which permit
the Class III label only where groundwater is “so contaminated by
naturally occurring conditions or the effects of broad-scale human
activity (unrelated to a specific activity) that [it] cannot be
cleaned up using treatment methods reasonably employed in public
water supply systems.” (Id. at 344.)
As EPA has emphasized at several points in this litigation,
EPA guidance is non-mandatory. (See, e.g., Gov’t Post-Trial Brief
56-57.) Moreover, agency guidelines, while informative, cannot
alter fundamental regulatory requirements. See United States v. S.
Union Co., 643 F. Supp. 2d 201, 211-12 (D.R.I. 2009), aff’d, 630
F.3d 17 (1st Cir. 2010), rev’d and remanded on other grounds, 567
U.S. 343 (2012). Therefore, to justify the remedy, EPA cannot
merely show that it followed EPA guidelines in classifying the
groundwater as Class II. Instead, EPA must demonstrate that it
complied with the NCP.
Unlike EPA guidelines, the NCP neither calls for nor permits
a bright line rule that groundwater must be considered a potential
source
of
drinking
water
anytime
71
contamination
is
neither
“naturally
occurring”
nor
the
result
of
“broad-scale
human
activity.” Instead, the NCP requires that EPA make a case-specific
determination
as
groundwater
to
to
whether
its
it
is
“practicable”
“beneficial
use.”
to
40
restore
C.F.R.
§ 300.430(a)(1)(iii)(F). In addition, EPA is required to collect
sufficient information to determine whether the remedy is likely
to be “effective” in restoring the groundwater to any such use.
Id. § 300.430(d)(1); see also id. § 300.430(e)(7)(i).
The evidence makes overwhelmingly clear that the Source Area
groundwater is currently far too contaminated to provide a source
of drinking water and that “[f]uture groundwater uses are not
expected
to
change
significantly.”
(ROD,
US1444-58.)
The
ROD
recognizes that there are likely several off-Site sources of
contamination
contributing
to
the
contamination.
(Id.)
In
addition, EPA has been on notice as far back as the remedial
investigation that there are several important uncertainties in
its data for the Source Area groundwater, including the “vertical
extent” of the contamination. (RI, US1098-88.)
In light of these issues, the Court finds that EPA has not
collected sufficient information or conducted sufficient analysis
on which to base its finding that the Source Area groundwater is
a potential source of drinking water or that the remedy is likely
to
effectuate
that
outcome.
Specifically,
there
is
currently
insufficient analysis regarding (1) the vertical extent of the
72
contamination at the Site; and (2) the extent to which off-Site
sources
contribute
to
contamination
in
the
Source
Area
groundwater. Going forward, EPA has broad discretion to determine
how
best
to
characterize
the
Site
and
make
any
necessary
adjustments to the remedy. But, at a minimum, if EPA continues to
classify the Source Area groundwater as a potential source of
drinking
water,
EPA
must
present
sufficient
information
and
analysis to justify what are key findings under the NCP: that
restoration of the groundwater is “practicable” and that the remedy
will be “effective” in bringing about that restoration.
I. Requiring a RCRA C Cap in the Source Area
1. Findings of Fact
The Source Area, as its name suggests, contains a significant
amount of contamination. Three interim protective caps as well as
one RCRA cap have already been placed over several portions of the
Source
Area.
(ROD,
US1444-13.)
Significant
portions
of
what
remains of the Source Area are covered by paved roads and parking
lots that service two elderly housing facilities (Brook Village
and Centredale Manor). (Id.)
EPA
previously
identified
the
existing
caps
and
paved
surfaces as a potential long-term solution. (See, e.g., EPA Action
Mem. dated May 4, 1999, US605-14; 2004 Technical Mem. on Long-Term
Remedy for Source Area Soils, US1048-25; 2005 BHHRA, US1101-57.)
However, after EPA classified Source Area groundwater as potential
73
drinking water, EPA identified the RCRA C cap as the preferred
remedy. (ROD, US1444-6.) Separate from remediation of the Source
Area groundwater, EPA has presented evidence that the RCRA C cap
is more durable and more likely to prevent direct human contact
with contaminated soil. (Id. at 333.) However, it is unclear from
the
record
whether,
absent
the
change
to
groundwater
classification, the ROD would have called for the installation of
a RCRA C cap.
There are significant challenges to implementing a RCRA C cap
on the Source Area. For example, installation of the cap will
require removing paved surfaces and excavating substantial amounts
of soil. This will inevitably cause inconvenience for the residents
of Brook Village and Centredale Manor, and EPA will have to take
special care to ensure that those residents are not exposed to the
toxic soil that is being excavated in close proximity to their
homes. (See, e.g., Mr. Loureiro Test., Trial Tr. vol. 6, 45:1-8
(discussing
“life
safety
issues”
for
Site
residents).)
EPA
considered many of these concerns in the FS. (See, e.g., FS,
US1254-208, 313.)
2. Conclusions of Law
EPA is required to address whether a RCRA C cap is “relevant
and appropriate” for the Source Area. See 42 U.S.C. § 9621(d). EPA
determined that a RCRA C cap is “relevant and appropriate.” (ROD,
US1444-296.) Emhart concedes that the RCRA C cap is “relevant,”
74
but argues that it is not “appropriate.” (Emhart Post-Trial Brief
190-91.) Emhart preserved this issue by bringing it to EPA’s
attention in its comments on the PRAP. (See generally Emhart
Comments on PRAP, US1383-90-109.)
The Court already determined that EPA has not conducted
sufficient analysis on which to base its finding that restoration
of Source Area groundwater is practicable. On the current record,
the Court also finds that EPA’s decision to include a RCRA C cap
as
part
of
the
remedy
for
the
Source
Area
is
inextricably
intertwined with EPA’s current remediation goals for Source Area
groundwater. Therefore, EPA cannot justify the RCRA C cap absent
further analysis with regards to the Source Area groundwater or a
finding that the RCRA C cap is necessary regardless of EPA’s
groundwater
remediation
goals.
This
analysis
will
necessarily
include a determination as to whether the benefits of the RCRA C
cap outweigh the inconvenience and potential health and safety
risks
to
Site
residents.
The
Court
reiterates
that
EPA
has
significant leeway in characterizing the Site going forward and
making adjustments to the remedy as necessary. However, on the
current record, the portion of the ROD requiring a RCRA C cap
cannot survive judicial scrutiny.
75
J. Baseline Human Health Risk Assessment (“BHHRA”)
1. Findings of Fact
a. Sampling of Residential Soils
EPA conducted a residential soil exposure risk assessment as
part of the BHHRA. The goal was to evaluate the risks to human
health caused by contact with the soils along Allendale and Lyman
Mill Ponds. (See generally 2012 Technical Mem., US1392; PRAP
Amendment,
US1422.)
Emhart
submitted
comments
on
the
PRAP
Amendment arguing that EPA had failed “to collect sufficient data
in the area of the residential floodplain to derive appropriate
exposure point concentrations for use in the [BHHRA].” (Emhart
Comments on PRAP Amendment, US1418-14-15.)
During
the
remedial
investigation
conducted
in
1999
EPA
collected several samples from each of sixty-two residential lots
located along the Woonasquatucket River. (RI, US1098-25.) EPA also
collected a small number of additional samples in 2001. (See NTCRA,
US1099-37-42.) In total, EPA evaluated 226 data records for dioxin
in
determining
residential
the
Site
potential
soils.
concentrations
“above
residential-use
soil”
(ROD,
the
was
human
risks
US1444-348-49.)
cleanup
found
health
“in
level
posed
Dioxin
by
at
for
floodplain
approximately
[forty-five
percent] of the locations sampled along the eastern shore of
Allendale and Lyman Mill Ponds.” (Id. at 348-49.) While the ROD
commits EPA to conducting additional sampling during remedial
76
design to define the exact contours of excavation, EPA determined
that the current level of sampling was sufficient to assess the
risk to human health posed by the Site. (Id. at 349.)
b. Estimates Used to Determine Risks from Recreational
and Residential Pond and Soil Exposure
i. Relative Bioavailability of Dioxin
The BHHRA 44 considered the risks from dioxin exposure to
recreational users and residents of the Site. This required that
EPA estimate the relative bioavailabilty (“RBA”) of dioxin in the
soil and sediment. RBA is the availability of dioxin in the medium
to which humans are likely being exposed (i.e., the soil and
sediment) relative to the medium that was the source of the
exposure. (Dr. Vorhees Test., Trial Tr. vol. 8, 65:14-25.) This
number helps EPA estimate the rate at which humans are likely to
absorb toxins at the Site. (Id.) EPA guidance suggests that RBA
for dioxin “can be expected to be less than [one] hundred percent.”
(Id. at 66:21-24 (discussing EPA, Dioxin RBA Report, US1542-10).)
Emhart therefore requested that EPA assume a RBA value of less
than one hundred percent for the Site. (Emhart Comments on PRAP
Amendment, US14418-12-14.)
EPA responded in the ROD that “there is not . . . a consensus
protocol for determining a site-specific RBA for dioxin in soil,
nor are such assessments a common practice.” (ROD, US1444-357-58;
44
See supra note 20.
77
see also Dr. Vorhees Test., Trial Tr. vol. 8, 66:25-67:5.) In the
absence of a “consensus protocol” EPA declined to conduct any Sitespecific RBA testing during the BHHRA. Instead, EPA used the
“protective assumption[]” that that the Site soil and sediment had
a RBA of one hundred percent. (Dr. Vorhees Test., Trial Tr. vol.
8, 68:7-11; see also ROD, US1444-357-58.) This is common practice
for EPA at other CERCLA sites. (Id. at 67:10-22.) It is also
consistent with EPA guidance. (See EPA, Dioxin RBA Report, US154210.)
ii. How Often People Use the Ponds
Having made a dioxin RBA assumption of one hundred percent,
EPA then went on to consider the risks to persons swimming and
wading in the ponds. In doing so EPA had to estimate how often
people use the ponds. The BHHRA assumed that the Site is used for
recreational activities from May through October, but specifically
for wading and swimming only from June through August. (2005 BHHRA,
US1101-81.) During these months the BHHRA used the Risk Assessment
Guidance for Superfund to estimate the maximum exposure that would
occur, which represents the upper percentile (95th-99.9th) of
exposure for persons using the Site. (See EPA, Risk Assessment
Guidance
for
Superfund
(“RAGS”),
vol.
III,
US1612-207.)
The
estimated maximum exposure frequency for wading was fifty-two days
per year for young children (ages six and below) and thirty-nine
days per year for adults and older children. (2005 BHHRA, US110178
82.) Based on the “recommended exposure time for recreational
swimming,”
EPA
estimated
that
each
exposure
would
last
approximately one hour. (Id.)
Emhart submitted comments to EPA on the BHHRA. While Emhart
conceded that the ponds were used recreationally, Emhart suggested
that a lower frequency of use should be assumed because “[i]t is
unlikely that most individuals will spend substantial amounts of
time wading in the river or ponds that do not have beach areas.”
(Emhart Comments on BHHRA, US1151-47.) There is anecdotal evidence
in the record suggesting that Allendale and Lyman Mill Ponds have
not been frequently used for swimming and wading in the past. (See,
e.g.,
12/7/2011
Public
Hearing,
US1334-37-39;
8/15/2011
Email
Correspondence, US1309.) This sentiment was echoed during trial by
Emhart’s expert. (Mr. Loureiro Test., Trial Tr. vol. 6, 196:14197:1.) That visitors and residents would refrain from swimming
and wading is understandable, as the ponds have been subject to
historic industrial pollution (RI, US10998-57) and have portions
that some consider unattractive. (See, e.g., Dr. Keenan Test.,
Trial Tr. vol. 4, 98:25-99:6; Mr. Loureiro Test., Trial Tr. vol.
6, 196:20-197:7.)
However, EPA presented evidence at trial demonstrating that
the
ponds
are
undoubtedly
used
for
various
recreational
activities. Residents at the Site have yards that lead directly
into the ponds, and various residents have placed docks and boats
79
in the ponds. (See Dr. Medine Test., Trial Tr. vol. 12, 58:2559:6.) EPA’s expert, Dr. Vorhees, also testified that she observed
fisherman as well as evidence of other recreational activities at
the Site. (Dr. Vorhees Test., Trial Tr. vol. 8, 11:21-12:19, 47:2349:10.) Additionally, as a counter to Emhart’s experts, Dr. Vorhees
testified that she personally found various portions of the ponds
to be “attractive.” (Id. at 48:7-49:20.) This evidence suggests
not only that the ponds are currently being used recreationally,
but also that increased use could be expected after cleanup is
complete.
iii. Incidental Ingestion Rates
In addition to estimating the frequency of pond use at the
Site,
EPA
had
to
estimate
the
rate
at
which
residents
and
recreational pond users incidentally ingest soils and sediments. 45
EPA estimated the upper range of incidental ingestion was a rate
of approximately two hundred milligrams per day for young children
and one hundred milligrams per day for older children and adults.
(2012 Technical Mem., US1392-58.) These ingestion rates are based
on EPA guidance. (See, e.g., IFRAGS, vol. I, US1620-13 (discussing
“Standard Default Exposure Factors”).) Dr. Vorhees testified at
trial that the use of these ingestion rates is standard practice
45
These ingestion rates were also used to estimate the risks
from residential soil exposure.
80
at other CERCLA sites. (Dr. Vorhees Test., Trial Tr. vol. 8, 64:713.)
Emhart’s expert, Dr. Keenan, testified at trial that these
soil ingestion rates were unnecessarily high. Dr. Keenan explained
that EPA’s estimate is based primarily on research conducted by
Dr. Edward Stanek and Dr. Edward Calabrese. (Dr. Keenan Test.,
Trial Tr. vol. 4, 103:16-21.) Those researchers have since cut
their recommendations in half, suggesting a soil ingestion rate of
approximately 100 milligrams per day for young children and fifty
milligrams per day for older children and adults. (See Emhart
Comments on BHHRA, US1151-61-64.) However, EPA has specifically
considered
this
information
and
rejected
the
researchers’
recommendations. (See EPA Response to Emhart Comments on the BHHRA,
US1450-5.) As EPA’s 2011 Exposure Factors Handbook explains, EPA
has
considered
the
available
information
(including
the
new
research done by Dr. Stanek and Dr. Calabrese) and still recommends
an upper percentile for soil ingestion of two hundred milligrams
per
day
for
young
children.
(See
EPA,
2011
Exposure
Factors
Handbook, Emhart578-268-318.)
iv. Number of Days Per Year Residents are Exposed to
Site Soils
As part of the residential soil risk assessment, EPA estimated
the
number
surrounding
of
days
their
Site
homes.
residents
EPA
are
estimated
81
exposed
that
to
the
exposure
soil
occurs
approximately
350
days
per
year.
(PRAP,
US1393-5.)
Emhart
challenged this estimate during the notice and comment period on
the PRAP, arguing that it was unrealistically high given seasonal
weather changes in Rhode Island. (Emhart Comments on PRAP, US141811.) Emhart has also presented evidence that other CERCLA sites in
New England have used the lower exposure frequency estimate of
approximately 150 days per year. (ROD, US1444-356-357.)
There is no consensus way to determine the specific number of
days EPA should use as the exposure frequency rate. (Dr. Vorhees
Test., Trial Tr. vol. 8, 59:16-60:3.) EPA guidance provides for a
national default exposure frequency rate of 350 days per year.
However, EPA guidance also states that this default rate “may not
be appropriate for all regions.” (EPA, RAGS, vol. I, Emhart58432.)
For
CERCLA
sites
in
New
England,
EPA
has
previously
recommended a lower exposure frequency of approximately 150 days
per year. (Dr. Vorhees Test., Trial Tr. vol. 9, 4:21-6:15, 63:2022.)
However, EPA released updated regional guidance in 2002 and
since that time has typically used the national default rate of
350 days per year. (ROD, US1444-356-57 (citing EPA, Supplemental
Soil Screening Guidance, US1637).) This is particularly true in
Rhode Island where EPA’s local counterpart (RIDEM) has also adopted
a default exposure frequency rate of 350 days. (Dr. Vorhees Test.,
Trial Tr. vol. 8, 59:16-60:3.) For instance, the “Peterson Puritan”
82
site, located in Rhode Island, used an exposure frequency rate of
350 days. (Id. at 60:16-21.)
Furthermore, as noted in the ROD, there are reasons why EPA
would assume that a 150-day estimate is too low for the Site.
Unlike
other
cleanup
locations,
the
Site
contains
a
large
residential population such that “access to this site is not
restricted or otherwise limited.” (EPA, IFRAGS, vol. I, Emhart579116; see also id. (“[R]esidential land use is most often associated
with the greatest exposure . . . .”); ROD, US1444-356.) As such,
the residents (unlike recreational users) may have little choice
but to venture outside and come into contact with the Site even in
inclement weather. Additionally, because residents’ houses are
located on the Site, staying indoors does not prevent contact with
contamination because Site soils and dust are easily tracked into
the home. (Dr. Vorhees Test., Trial Tr. vol. 8, 60:5-15.)
c. Risks from Fish Consumption
EPA conducted a fish consumption risk assessment as part of
the BHHRA. The goal was to determine the risks to human health
posed by fish consumption at the Site and the cleanup levels of
pond sediments necessary to mitigate those risks. To accomplish
this goal, EPA had to estimate the frequency with which people
consumed fish from the Site, the species of fish consumed, and the
contamination levels present in the fish. EPA then determined the
biota-to-sediment accumulation factor (“BSAF”), which is the ratio
83
between contamination levels in fish and contamination levels in
the sediment. (Dr. Keenan Test., Trial Tr. vol. 4, 42:11-43:4.) By
analyzing the frequency and toxicity of fish consumed in the ponds,
in
combination
with
the
BSAF,
EPA
was
able
to
make
risk
calculations and establish pond sediment remediation goals. Emhart
challenged
several
aspects
of
EPA’s
fish
consumption
risk
assessment in its comments on the BHHRA. (See Emhart Comments on
BHHRA, US1151; see also EPA Response to Emhart Comments on BHHRA,
US1450.)
i. Fish Sampling
Allendale Pond is sustained, in part, by a dam that prevents
the majority of Allendale Pond waters from flowing downstream into
Lyman Mill Pond. The dam has breached several times, including
once in 1991 and twice in April and May of 2001. (See FS, US125443; Dr. Keenan Test., Trial Tr. vol. 4, 14:22-15:7.) The dam was
restored after the 2001 breaches and by early 2002 Allendale Pond
was restored to pre-breach water levels. (FS, US1254-43.)
Unfortunately, EPA’s collection of fish samples in Allendale
and Lyman Mill Ponds was conducted in July 2001, after the May
2001 breach but before Allendale pond had been restored. (ROD,
US1444-71.) This led to two sampling issues. First, some of the
fish collected from Lyman Mill Pond “may actually have been washed
into Lyman Mill Pond from Allendale pond at the time of the
84
Allendale Dam breach.” (MACTEC Mem., US1100-34.) 46 This means that
the contamination levels of fish collected from Lyman Mill Pond
“may not bear a meaningful relationship” to the sediment in Lyman
Mill pond, such that EPA could not calculate an accurate BSAF.
(CSTAG
Recommendations,
US1080-5.)
The
Contaminated
Sediments
Technical Advisory Group (“CSTAG”) 47 highlighted this issue and
“strongly recommend[ed] that new, co-located sediment and fish
tissues samples be collected to develop a BSAF.” (Id.)
EPA recognized this problem and employed MACTEC to find an
adequate solution. MACTEC compared EPA’s data collected after the
Allendale dam breach to data collected before the breach and
determined that, if certain steps were taken, “the available data
appear to be sufficient for PRG development and additional sampling
and analysis is not recommended at this time.” (MACTEC 4/29/2005
Letter, US 1100.) The steps MACTEC recommended were for EPA to use
fish tissue data and BSAFs from other parts of the Woonasquatucket
River to determine PRGs. EPA complied with this recommendation. As
EPA explained in the ROD, “[f]or [] Lyman Mill Pond, arithmetic
mean BSAF for each fish species from five other exposure areas
46
MACTEC is an EPA contractor that was used during the risk
assessment to analyze data collected at the Site.
47
CSTAG is EPA’s internal technical advisory group that
monitors the progress and provides advice for certain large and
complex Superfund sites involving contaminated sediment. (See
Principles for Managing Contaminated Sediment Risks at Hazardous
Waste Sites, US909.)
85
. . . for each contaminant was used in the derivation of speciesspecific
fish-consumption-based
sediment
PRGs
for
Lyman
Mill
Pond.” (ROD, US1444-71 n.6.) This allowed EPA to estimate a BSAF
for Lyman Mill Pond without using the data points objected to by
Emhart.
The second sampling issue caused by the Allendale dam breach
involved the species of fish collected. EPA collected equal numbers
of eel, white sucker, and largemouth bass from Lyman Mill Pond.
(2005 BHHRA, US1101-60.) However, when EPA sampled Allendale Pond,
EPA collected only eel and white sucker, but no largemouth bass.
(Id.) Based on that data EPA’s risk calculations assumed that
largemouth bass are not typically present in Allendale Pond. (See
id. at 60, 78-79.)
This assumption is likely erroneous. As MACTEC explained to
EPA after reviewing the BHHRA, “[t]he absence of largemouth bass
in Allendale Pond in July 2001 suggests that largemouth bass may
[] have been washed out of Allendale Pond into Lyman Mill Pond.”
(MACTEC Mem., US1100-34.) EPA recognized that assuming the absence
of largemouth bass in Allendale Pond “may have resulted in an overor under-estimation” in its risk assumptions. (ROD, US1444-71.)
However, unlike with its BSAF calculations, EPA did not adjust its
species assumptions to take into account the likely effects of the
Allendale dam breach.
86
ii. Species of Fish Consumed
In 1999 the Rhode Island Department of Health established a
fish-consumption advisory at the Site in order to deter people
from consuming contaminated fish. (Dr. Vorhees Test., Trial Tr.
vol.
8,
26:18-24.)
As
discussed
below,
despite
this
fish-
consumption advisory, people still fish and consume fish from the
Site. In order to estimate the risk to human health from fish
consumption at the Site, EPA came up with protective assumptions
about the species of fish people currently consume from the Site
as well as the species of fish people would consume under possible
future conditions. (See, e.g., id. at 32:14-16.) This reasonable
maximum exposure is essentially the highest exposure that could be
reasonably expected to occur at the Site. (Id. at 37:16-19; see
also id. at 40:18-19 (describing “reasonable maximum exposure” as
“the high end of possible exposure”).)
One common way of accomplishing this is through a survey of
local anglers. However, an angler survey would be of limited value
at the Site because of the fish-consumption advisory that has been
in place since 1999. (Id. at 26:18-27:12, 33:7-14.) The effects
were evident, for example, in EPA’s 2001 survey of local anglers.
In
that
survey
various
respondents
admitted
to
catching
and
consuming fish from the Site. (2005 BHHRA, US1101-80.) However,
“most respondents were aware of the current advisory against
consumption of fish and other biota from the river” and “indicated
87
they no longer consume fish from the river.” (Id.) Given the effect
of the fish-consumption advisory, EPA reasonably concluded that a
Site-specific angler survey would be of limited value in coming up
with reasonable estimates of fish consumption, particularly in
estimating
the
reasonable
maximum
exposure
of
likely
future
consumption.
This led EPA to use several other sources of information to
estimate current and future fish-consumption rates at the Site. As
discussed above, EPA sampled fish from Allendale and Lyman Mill
Ponds to determine the types of fish typically caught at the Site.
These included bass, white sucker, and eel. (Id. at 60.) EPA also
used surveys of people living near the site, academic literature,
and input from local stakeholders to confirm that people were
consuming fish at the Site and were likely to do so in the future.
(Id. at 80.) Based on this information, EPA determined that the
group most likely to be the subject of maximum exposure was certain
Asian populations living near the Site.
The academic literature suggested that various parts of New
England (including Rhode Island) are home to Asian populations
that fish both recreationally and as a form of subsistence. (See,
e.g., id. at 576, 583-89.) These populations often either ignore
or misunderstand fishing advisories and regularly consume their
catches
regardless
of
size.
(See,
e.g.,
id.
at
583-4.)
Additionally, they catch and consume a wide variety of species,
88
including species identified at the Site. (See, e.g., id. at 583;
U.S. Army Corps of Eng’rs, Identification of Preferred Target
Species, US847-3-4.) EPA, in consultation with a wide variety of
local
subject
matters
experts 48,
considered
these
findings
persuasive.
The information garnered from the academic literature and
local experts also corresponded with some Site-specific data. For
instance, two surveys conducted at or near the Site suggested that
certain Asian populations fish in that area. The first is the 1998
Urban River Use Survey of the Woonasquatucket River, which observed
a diverse set of people – including persons of Asian descent –
fishing in the Woonasquatucket River. (Id. at 80.) The second was
the Tool Kit for Urban Rivers, which focused on the demographics
of nearby Providence and found the “Southeast Asian population
(Hmong, Camobodian, Laotian, & Vietnamese) and other ethnic groups
were . . . high consumers of fish, eel, and turtles from local
waterways, including the Woonasquatucket River.” (Id.; see also
EPA, Tool Kit for Urban Rivers, US1029-3; Dr. Vorhees Test., Trial
Tr. vol. 8, 31:6-21.)
48
EPA consulted with private groups (such as universities and
non-profits) as well as RIDEM, the Rhode Island Department of
Health, and the Rhode Island Division of Fish and Wildlife. (See,
e.g., BHHRA, US1102-577, 583; U.S. Army Corps of Eng’rs,
Identification of Preferred Target Species, US847-3.)
89
EPA recognized that the lack of an angler survey made EPA’s
analysis of the data more difficult. This meant that “[t]here was
no
quantitative
information
available
that
could
be
used
to
quantify the potential combined fish diet in a more detailed
manner.” (EPA Response to Emhart Comments on BHHRA, US1450-3.) EPA
was
therefore
forced
to
use
its
“professional
judgment
in
identifying the composition of the hypothetical future combined
fish diet.” (Id.) Taking into account the fact that certain Asian
populations consumed a wide variety of fish, EPA decided to make
the
conservative
estimate
that
the
species
of
fish
found
in
Allendale and Lyman Mill Ponds were consumed in equal portions.
In
its
comments
on
the
BHHRA
Emhart
challenged
EPA’s
assumption that certain populations may consume white sucker and
eel at the same rate as largemouth bass. (Emhart Comments on BHHRA,
US1151-41.) In the absence of a Site-specific angler survey, Emhart
suggested that EPA consult the Maine Angler Survey (“MAS”). The
MAS is an assessment of fish consumption practices in Maine from
1992. EPA’s Exposure Factors Handbook includes some MAS data when
providing recommended consumption rates. (Dr. Keenan Test., Trial
Tr. vol. 4, 49:17-50:21; EPA, 1997 Exposure Factors Handbook,
Emhart598-458-461.) Specifically addressing the consumable species
identified at the Site, the MAS recommends a consumption ratio of
approximately 83:11:3 (largemouth bass - white sucker – eel).
90
EPA, however, determined that the MAS should not be relied
upon at the Site. Instead, EPA opted to use its “professional
judgment”
to
determine
the
reasonable
maximum
exposure
of
a
“hypothetical future combined fish diet.” (EPA Response to Emhart
Comments on BHHRA, US1450-3.) As discussed above, EPA determined
its
consumption
ratios
using
academic
literature,
input
from
subject matter experts, and surveys of people living near the Site.
iii. Parts of Fish Consumed
In the BHHRA, EPA had to consider which parts of the fish
people
might
consume.
For
the
purposes
of
determining
the
reasonable maximum exposure, EPA assumed that people consume the
fillets of largemouth bass and the whole body of eels and white
sucker. Based on this assumption, EPA used the contamination
concentrations in the fillets of largemouth bass and the wholebody contamination concentrations for eel and white sucker in the
BHHRA. (Dr. Vorhees Test., Trial Tr. vol. 8, 33:23-34:5.) This
increased the estimated risk to human health because whole-body
contamination concentrations are higher than the contamination
concentrations of fillets due to the fact that contamination
concentrates to a higher degree in fish organs than in the flesh.
(Id. at 34:9-16, 41:3-6.)
Emhart objected to EPA’s use of whole-body contamination
concentrations for white sucker in its comments on the BHHRA. While
recognizing that “it may be reasonable to assume that the entire
91
eel is consumed in certain instances,” Emhart argued that EPA
lacked
sufficient
Site-specific
data
on
which
to
base
its
assumption “that individuals eat the entire sucker, which is an
extremely bony fish.” (Emhart Comments on BHHRA, US1151-42.) At
trial, Emhart also cited EPA guidance which suggests that “[m]ost
fishers in the United States consume fish fillets.” (EPA, Assessing
Chemical Contaminant Data for Use in Fish Advisories, vol. II,
Emhart590-95.) Absent “specific data on fish preparation methods,”
this guidance “recommends using fillets as the standard sample
type for analyzing chemical contaminants.” (Id. at 312.)
EPA responded to this criticism by referring again to its
research on certain Asian populations that consume fish at or near
the Site. Academic literature suggests not only that those groups
consume a wide variety of fish, but also that they consume the
whole body of those fish. For example, one study found that certain
Asian populations fishing in Rhode Island often do not remove the
skin, fat or organs of the fish before cooking. (2005 BHHRA,
US1102-584.) That study also found that “fish cooking methods”
include “boiling” such that “the broth may be used for soup.” (Id.)
Another study provided a similar observation of Asian fishers on
the Housatonic River in Connecticut, noting that “[i]n addition to
fillets,” those populations eat “eyes, skin, and organs.” (Id. at
589.)
92
These conclusions were also supported by EPA guidance. The
same EPA guidance cited by Emhart also notes that “[c]ertain
populations, including some Asian-Americans . . . eat parts of the
fish other than the fillet and may consume the whole fish. Recipes
from many cultures employ whole fish for making soups or stews. As
a result, more of the fish contaminants are consumed.” (EPA,
Assessing Chemical Contaminant Data for Use in Fish Advisories,
vol. II, Emhart590-309.) Based on this information EPA “assumed
that white sucker might be most likely to be consumed if it were
a component of a fish stew or similar meal (as a whole body).”
(EPA Response to Emhart Comments on BHHRA, US1450-3.)
iv. Consumption Amounts from the Site
In
considering
consumption,
EPA
the
had
risks
to
to
human
estimate
health
the
posed
reasonable
by
fish
maximum
consumption rate of fish at the Site. EPA did this, in part, by
considering data from the MAS. (2005 BHHRA, US1101-79.) According
to that data, the upper end (i.e., 90th percentile) of adult
fishers who do not share the fish they catch can be expected to
consume approximately “23 half-pound fish meals in a year or . . .
about one fish meal every two to three weeks” from the Site. (Dr.
Vorhees Test., Trial Tr. vol. 8, 38:5-7; see also MAS, US1541-7;
2005 BHHRA, US1101-79.) This equates to approximately fourteen
grams of fish per day from the Site. (Id.)
93
In
its
comments
on
the
BHHRA
Emhart
challenged
these
consumption rates for two reasons. (Emhart Comments on BHHRA,
US1151-44.) First, Emhart argued that EPA was overestimating the
amount of fish consumed by adults. (Id. at 45.) As Emhart noted,
the MAS found that most anglers share their catch with other
members of the household and therefore typically consume less than
fourteen grams per day. (Id.) Under these circumstances, Dr. Keenan
opined at trial that EPA should have assumed household sharing of
fish in calculating reasonable maximum consumption rates. (Dr.
Keenan Test., Trial Tr. vol. 4, 69:6-71:14.)
Dr. Vorhees disagreed and testified that EPA’s assumptions
were reasonable. As she explained, in calculating the reasonable
maximum exposure EPA “obviously had a concern about the angler who
does not share their catch and wanted to be sure to protect them.”
(Dr. Vorhees Test., Trial Tr. vol. 8, 40:13-15.) According to Dr.
Vorhees, EPA’s decision was therefore reasonable “because [EPA]
wanted to protect consumers who, in fact, catch fish and don’t
necessarily share their catch.” (Id. at 39:3-5.)
Emhart’s second challenge to EPA’s consumption rates was
based on EPA’s assumption that all fourteen grams consumed by the
adult fisher came from the Site. (Emhart Comments on BHHRA, US115145.) EPA has explained that its “fish consumption rates include
only fish caught from the water bodies at the Site, and do not
include fish caught at other locations nor fresh or preserved fish
94
purchased for consumption.” (2005 BHHRA, US1101-79.) This suggests
that EPA is assuming that the adult fisher consumes fourteen grams
per day from the Site and also consumes significant additional
fish from other sources.
While this assumption may be reasonable, it does not conform
to the MAS data on which EPA basis its analysis. (Id.) In the
context of its fourteen-gram estimate, The MAS explains that
the study was designed to collect data on consumption
from all flowing bodies of water, and not just the . . .
contaminated water. Thus, although individuals may fish
in affected river reaches some of the time, it is highly
unlikely that all fishing effort is focused on these
waters,
particularly
because
there
are
numerous
alternative fisheries in close proximity to each river.
. . . Consequently, whereas the estimates for rivers and
streams include all consumed fish from rivers and
streams during the season, it is likely that only a
portion of the consumption can be attributed to a single
water body.
(MAS, US1541-7.) As this makes clear, the MAS’s fourteen-gram
estimate assumes that those fourteen grams come from multiple
sources.
2. Conclusions of Law
Emhart previously challenged the sufficiency of the BHHRA in
comments submitted to EPA. However, several of Emhart’s specific
arguments presented at trial were not submitted during the official
notice and comment period. Such arguments are typically waived
under the doctrine of issue exhaustion. However, because the Court
finds that the portions of the BHHRA challenged by Emhart were so
95
foundational to EPA’s remedy-selection process, and because Emhart
previously brought many of these issues to EPA’s attention, a
limited exception is warranted. EPA has emphasized that the BHHRA
was the primary driver of EPA’s cleanup standards. (See, e.g.,
Gov’t Post-Trial Brief 87.) Under these circumstances, EPA must be
able to justify the BHHRA’s “key assumptions.” See Oklahoma Dep’t
of Envtl. Quality, 740 F.3d at 192 (quoting Appalachian Power Co.,
135 F.3d at 818). These “key assumptions” include EPA’s conclusions
regarding the human health risks posed by soil exposure, pond
exposure, and fish consumption at the Site.
a. Sampling of Residential Soils
Emhart challenges the sufficiency of EPA’s sampling process,
describing it as “sparse.” (Emhart Post-Trial Brief 79.) Emhart
points to the fact that EPA collected fewer than four samples
containing dioxin data for each property at the Site. (Id. at 7980.) However, as the Court explained in previous sections, the NCP
does not require a specific number of samples. Instead, EPA need
only “collect data necessary to adequately characterize the site
for the purpose of developing and evaluating effective remedial
alternatives.” 40 C.F.R. § 300.430(d)(1). In characterizing the
risks from residential soil exposure EPA reviewed approximately
226 data points, which included several samples from each of sixtytwo residential properties at the Site. Emhart has failed to
96
demonstrate how this level of sampling is somehow insufficient
under the NCP.
b. Estimates Used to Determine Risks from Recreational
and Residential Pond and Soil Exposure
Emhart challenges various estimates made by EPA in conducting
the pond and soil risk assessments. Emhart argues that EPA used
inappropriately high estimates in characterizing: (1) the relative
bioavailability (“RBA”) of dioxin; (2) the rates at which people
incidentally ingest contaminants; (3) how often people use the
ponds at the Site; and (4) how often residents are exposed to Site
soils.
In each of these areas, EPA did not rely solely on Sitespecific data, but instead developed “conservative” estimates (in
the sense that EPA erred on the side of protecting people from
exposure) based on available scientific literature, EPA guidance,
and experience at other CERCLA sites. This practice does not
inherently violate the NCP. While the NCP requires that EPA collect
sufficient
site-specific
data
to
adequately
compare
remedial
options, it plainly does not mandate that every piece of data used
by EPA be the result of sampling at the site. And in the absence
of clear direction from either the NCP or CERCLA, the development
of broadly applicable estimates for use on either a regional or
national basis falls squarely within the purview of EPA. See Fox
Television Stations, 556 U.S. at 513–14; Florida Power & Light
97
Co., 404 U.S. at 463. The question is whether EPA’s use of these
estimates for the Site was somehow arbitrary and capricious. The
Court finds that it was not.
EPA’s estimates are not “without substantial basis in fact.”
Muszynski, 899 F.2d at 160 (quoting Florida Power & Light Co., 404
U.S. at 463). EPA has surveyed the available data and provided
reasonably conservative estimates for use at CERCLA sites. To be
sure, there is no universal consensus on how to estimate the RBA
of dioxin, incidental ingestion rates, or pond and soil exposure
frequency. But it is precisely in these areas that involve evolving
scientific
considerations
that
deference
to
EPA’s
“technical
expertise and experience” is appropriate. Id. (quoting Florida
Power & Light Co., 404 U.S. at 463). And while EPA’s assumptions
are certainly conservative, that is by design. The NCP mandates
that EPA’s “goal . . . is to select remedies that are protective
of human health and the environment, that maintain protection over
time,
and
that
minimize
untreated
waste.”
40
C.F.R.
§ 300.430(a)(1)(i). In so doing, EPA estimates the reasonable
maximum exposure that is likely to occur for both current and
potential future land use at the Site. (EPA, IFRAGS, vol. I,
Emhart579-24.) The use of a conservative estimate under these
circumstances – erring on the side of caution when it comes to the
risk of cancer (and other maladies) for those living in and around
98
the Site – is neither arbitrary, capricious, nor a violation of
the NCP.
c. Risks from Fish Consumption
Emhart challenges several aspects of EPA’s fish consumption
estimates in the BHHRA. With regards to EPA’s estimates about the
types and parts of fish consumed from the Site, Emhart’s challenges
focus primarily on EPA’s decision to forego a Site-specific angler
survey and to rely on data from outside the MAS. Neither the NCP
nor EPA guidance mandates that EPA conduct an angler survey or use
data from the MAS. Instead, as discussed above, EPA need only
“collect data necessary to adequately characterize the site for
the
purpose
of
developing
and
evaluating
effective
remedial
alternatives.” 40 C.F.R. § 300.430(d)(1). After collecting that
data, EPA must then draw conclusions that are not arbitrary or
“without substantial basis in fact.” Muszynski, 899 F.2d at 160
(quoting Florida Power & Light Co., 404 U.S. at 463).
In determining species preference at the Site EPA reasonably
concluded that an angler survey would be of little use because of
the fish-consumption advisory in place at the Site. EPA therefore
looked to alternative sources of information, including local
subject matter experts, academic literature on fish-consumption in
Rhode Island and surrounding states, surveys, and fish samples
collected from the Site. This level of data collection provided an
adequate, non-arbitrary, basis on which to determine the likely
99
species preferences of the most vulnerable populations near the
Site for the purposes of developing a reasonable maximum exposure
estimate. EPA certainly had to make judgment calls along the way,
but these decisions deserve deference considering EPA’s “technical
expertise and experience,” id. (quoting Florida Power & Light
Co., 404 U.S. at 463), as well as EPA’s duty to make conservative
estimates
in
order
to
protect
human
health.
See
40
C.F.R.
§ 300.430(a)(1)(i).
With that said, regarding EPA’s estimates for the amount of
fish consumed from the Site, the Court finds two instances where
EPA’s analysis was arbitrary. First, EPA relied on knowingly nonrepresentative sampling to assume that no largemouth bass are
consumed
from
Allendale
Pond.
Despite
being
on
notice
that
largemouth bass are likely present in Allendale Pond, EPA excluded
them
from
its
Allendale
Pond
risk
calculation.
Emhart
has
demonstrated that this misstep arbitrarily increased the risk
calculation for Allendale Pond. See Motor Vehicle Mfrs., 463 U.S.
at 50 (describing agency action as arbitrary where the agency
“entirely failed to consider an important aspect of the problem”).
Second, EPA assumed that certain populations consume fourteen
grams of fish per day from the Site. While the fourteen-grams
estimate is based on the MAS, the MAS provides the caveat that
“only a portion of the consumption can be attributed to a single
water body.” (MAS, US1541-7.) While EPA may ultimately determine
100
that
fourteen
grams
is
the
appropriate
reasonable
maximum
consumption rate at the Site, on the current record (i.e., basing
its estimate on the MAS data alone) EPA’s decision is arbitrary.
See Motor Vehicle Mfrs., 463 U.S. at 50 (describing agency action
as arbitrary where the agency “offered an explanation for its
decision that runs counter to the evidence before the agency.”)
Both of these missteps must be remedied before moving forward.
How best to address these issues and identify any consequent
changes to the remedy is appropriately addressed by EPA in the
first instance. That is EPA’s prerogative, to be reviewed by the
Court only as provided under CERCLA.
K. Notice and Reasonable Opportunity to Comment
1. Findings of Fact
EPA drafted a PRAP in the fall of 2011. (PRAP, US1328.) Based
on subsequent analysis, EPA made several changes to the PRAP in
the form of a PRAP Amendment, which was published the following
summer. (PRAP Amendment, US 1393.) Both the PRAP and PRAP Amendment
were subject to notice and comment after their publication. The
notice and comment period on the PRAP and PRAP Amendment went from
November 14, 2011 to March 2, 2012, and July 19, 2012 to September
17, 2012, respectively.
(ROD, US1444-24-25.) EPA provided the
public notice of the PRAP and PRAP Amendment through newspaper
announcements. (Id.) EPA then participated in public hearings and
considered comments from a variety of sources, including Emhart.
101
(Id.; see also Emhart Comments on PRAP, US1383; Emhart Comments on
PRAP Amendment, US1418.)
2. Conclusions of Law
Emhart argues that EPA failed to provide the public notice
and a reasonable opportunity to comment on the PRAP and PRAP
Amendment. Before publication of the ROD, EPA is required to
present a PRAP to the public “for review and comment.” 40 C.F.R.
§
300.430(f)(1)(ii).
The
PRAP
must
“briefly
describe[]
the
remedial alternatives analyzed by the lead agency, propose[] a
preferred
remedial
action
alternative,
and
summarize[]
the
information relied upon to select the preferred alternative.” Id.
§
300.430.
EPA
must
then
provide
the
public
“a
reasonable
opportunity, not less than 30 calendar days, for submission of
written and oral comments” as well as an “opportunity for a public
meeting.”
Id.
§
300.430(f)(3)(i)(C)-(D).
The
purpose
of
this
process is to provide “the public with a reasonable opportunity to
comment on the preferred alternative for remedial action, as well
as alternative plans under consideration, and to participate in
the selection of remedial action at a site.” Id. § 300.430(f)(2).
EPA complied with these NCP requirements. However, the Court
has found several instances where EPA’s analysis was flawed and
the results of that analysis were included in the PRAP and PRAP
Amendment. EPA has also committed to conducting further analysis
during remedial design and reserved the right to modify the remedy.
102
EPA
is
therefore
responsible
for
updating
the
administrative
record as necessary to correct flaws in its analysis and document
future
analysis
conducted
during
remedial
design.
Id.
§ 300.825(a)(1). To the extent EPA’s corrections or the remedial
design process lead to a significant change to the remedy 49, EPA
must
provide
for
an
additional
notice
and
comment
period
as
required by the NCP.
L. Failure to Comply with the Unilateral Administrative Order
1. Findings of Fact
Emhart
initiated
this
litigation
in
May
2006.
(Phase
I
Findings 3-4.) EPA did not issue the ROD until over six years
later, in September 2012. (See ROD, US1444.) EPA then issued its
Unilateral Administrative Order (“UAO”) on June 10, 2014. (See
UAO, US1490.) The UAO ordered Emhart “to perform the Remedial
Design, Remedial Action, and Operation and Maintenance for the
selected remedy . . . as described in the Record of Decision.”
(Id. at 5.) That UAO currently remains in place.
Since EPA’s issuance of the UAO the Court has completed Phase
I and II of this litigation. The Phase I trial lasted over twenty
days beginning on May 18, 2015 and resulted in Emhart being found
liable under CERCLA. However, the Court also found that “the
49
What constitutes a “significant change” includes, among
other things, a significant increase in the estimated amount of
time it will take to implement the remedy, costs, and potential
disruption to local residents.
103
Government is not yet entitled to judgment in its favor on this
claim because the issues of costs and whether the remedy selected
by the EPA is consistent with CERCLA first need to be litigated in
the second phase of this trial.” (Phase I Findings 185-86.) Emhart
has continued to refuse to comply with the UAO through Phase II of
this litigation based on its objections to the selected remedy.
2. Conclusions of Law
CERCLA
necessary
authorizes
to
protect
EPA
to
public
issue
health
“such
and
orders
welfare
as
may
and
be
the
environment.” 42 U.S.C. § 9606(a). This includes situations where
there “may be an imminent and substantial endangerment to the
public health or welfare or the environment because of an actual
or threatened release of a hazardous substance from a facility.”
Id. Any entity that “willfully violates” such an order “without
sufficient cause” may be liable for fines and punitive damages.
Id. §§ 9606(b)(1); 9607(c)(3). These penalties are not mandatory,
but are instead authorized at the discretion of the Court. See id.
§§ 9606(b)(1), 9607(c)(3) (explaining that violators “may” be
liable for penalties); Gen. Elec. Co. v. Jackson, 610 F.3d 110,
119 (D.C. Cir. 2010) (“[T]he district court has authority to decide
not to impose fines even if it concludes that a recipient ‘without
sufficient cause, willfully violate[d], or fail[ed] or refuse[d]
to comply with’ a UAO.”) (citing 42 U.S.C. §§ 9606(b), 9607(c)(3)).
104
What constitutes “sufficient cause” under CERCLA has not been
addressed by the First Circuit. However, multiple other Circuit
Courts of Appeal have interpreted “sufficient cause” to mean a
“good faith” or “objectively reasonable basis for believing that
the EPA's order was either invalid or inapplicable to it.” Solid
State Circuits, Inc. v. EPA, 812 F.2d 383, 391 (8th Cir. 1987);
see also, e.g., Gen. Elec. Co., 610 F.3d at 119; Emp’rs Ins. of
Wausau v. Browner, 52 F.3d 656, 661 (7th Cir. 1995). A party may
meet
this
standard
by
demonstrating
“that
the
applicable
provisions of CERCLA, EPA regulations and policy statements, and
any formal or informal hearings or guidance the EPA may provide,
give rise to an objectively reasonable belief in the invalidity or
inapplicability of the clean-up order.” Solid State Circuits, 812
F.2d at 392.
In this case, Emhart does not challenge EPA’s finding that
the Site qualifies as a potential “imminent and substantial”
environmental threat sufficient to justify a UAO. Moreover, in the
aftermath of the Phase I trial, Emhart cannot assert that the UAO
is somehow inapplicable to it. Instead, Emhart argues that it has
a “sufficient basis” to not comply with the UAO based on its good
faith challenges to the validity of EPA’s selected remedy.
The Court has determined that several aspects of EPA analysis
qualify
as
arbitrary,
including
the
classification
of
Site
groundwater as a potential source of drinking water and portions
105
of
the
fish
consumption
risk
assessment.
Emhart
made
timely
objections highlighting these issues in its comments on the PRAP.
These sorts of missteps on the part of EPA do not automatically
provide
potentially
responsible
parties
sufficient
cause
to
disregard a UAO. However, given that Emhart has continuously
objected to critical aspects of the remedial design that the Court
has now found arbitrary, and Emhart’s previous participation in
the cleanup process, the Court finds that Emhart’s challenge to
the UAO up to this point has been pursued in objective good faith.
V. Conclusion
EPA has developed a remedial action using the process outlined
in CERCLA and the NCP. That remedial action, if completed, will
mitigate the risks to human health and the environment EPA has
identified at the Site. However, in developing that remedial
action, EPA made several decisions that the Court finds violated
CERCLA because they were arbitrary, capricious, or otherwise not
in accordance with law; these must be addressed before moving
forward with the remedial action.
The Court finds that Emhart has met its burden in establishing
that,
on
decisions
the
were
record
as
currently
arbitrary,
constituted,
capricious,
or
the
otherwise
following
not
in
accordance with law: (1) labelling Source Area groundwater as a
potential source of drinking water; (2) assuming that there are no
largemouth bass in Allendale Pond; and (3) using fourteen grams as
106
the reasonable maximum consumption rate for anglers fishing at the
Site. The UAO is stayed until these matters are resolved, and
Emhart is not required to pay the fines and fees stemming from its
non-compliance with the UAO that have accrued up to this point.
The Court can envision several ways EPA could approach these
deficiencies. EPA could find that the issues identified by the
Court
require
EPA
to
reopen
the
remedial
investigation
and
feasibility study process and publish a PRAP Amendment, as it has
done in the past. The Court takes no view as to the appropriate
course of action at this time and leaves it to EPA to address these
issues
in
the
first
instance.
However,
the
Court
retains
jurisdiction in this matter in order to ensure that the issues are
addressed in a manner consistent with the law and not arbitrary or
capricious.
Lastly, the Court notes that its decision is based, in part,
on EPA’s commitment to further sampling and analysis during the
remedial design phase. EPA concedes that information discovered
during remedial design may require alterations to EPA’s analysis
and the chosen remedy. For instance, EPA’s estimates may change
with regards to the amount of money and time it will take to
implement the remedy or the potential hazards and inconvenience to
local residents caused by implementation. Again, the Court retains
107
jurisdiction over this process to ensure that EPA’s actions are
consistent with the law and not arbitrary or capricious.
IT IS SO ORDERED.
William E. Smith
Chief Judge
Date: August 17, 2017
108
Disclaimer: Justia Dockets & Filings provides public litigation records from the federal appellate and district courts. These filings and docket sheets should not be considered findings of fact or liability, nor do they necessarily reflect the view of Justia.
Why Is My Information Online?