US Magnesium, LLC v. Environmental Protection Agency

United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT

Argued November 5, 2010 Decided January 14, 2011

No. 09-1269

US MAGNESIUM, LLC,
PETITIONER

v.

ENVIRONMENTAL PROTECTION AGENCY,
RESPONDENT

On Petition for Review of a Final Action
of the Environmental Protection Agency

Laurence S. Kirsch argued the cause for petitioner. With
him on the briefs were Valerie E. Ross, David W.
Tundermann, and M. Lindsay Ford.

T. Monique Jones, Attorney, U.S. Department of Justice,
argued the cause and filed the brief for respondent.

Before: SENTELLE, Chief Judge, GARLAND, Circuit
Judge, and WILLIAMS, Senior Circuit Judge.

Opinion for the Court filed by Senior Circuit Judge
WILLIAMS.
 2

WILLIAMS, Senior Circuit Judge: The National Priorities
List (“NPL”) is a list of places, commonly known as
“superfund sites,” considered national priorities for
environmental remediation because of known or threatened
releases of hazardous substances. The Comprehensive
Environmental Response, Compensation, and Liability Act
(“CERCLA”), 42 U.S.C. §§ 9601-75, requires the President to
establish “criteria for determining priorities among releases or
threatened releases [of hazardous substances] throughout the
United States for the purpose of taking remedial action . . . .”
42 U.S.C. § 9605(a)(8)(A). The Environmental Protection
Agency developed the Hazard Ranking System (“HRS”) to
fulfill that mandate. 40 C.F.R. § 300.425(c)(1); see generally
Eagle-Picher Industries Inc. v. EPA, 759 F.2d 905, 909-11
(D.C. Cir. 1985). The HRS is the principal guide used by the
EPA to place sites on the NPL. 40 C.F.R. Pt. 300, App. A,
§ 1.1.

This case concerns the NPL listing of a magnesium plant
located in Tooele County, Utah, approximately 40 miles west
of Salt Lake City and adjacent to the Great Salt Lake. The
plant, which is now owned by petitioner US Magnesium LLC
(“USM”), has produced molten magnesium since 1972,
creating chlorine gas and hydrochloric acid as by-products. A
network of ditches carries waste from the plant to an active
waste pool. Just beyond that pool is an inactive waste pool,
which was previously a recipient of waste.

The EPA completed an HRS evaluation for the US
Magnesium site in 2008. The HRS requires the agency to
analyze four “pathways”: ground water migration, surface
water migration, soil exposure, and air migration, and to plug
the resulting individual pathway scores into a formula to
obtain the site score. 40 C.F.R. Pt. 300, App. A, § 2.1.1. The
EPA calculated scores for two out of these four possible
“pathways”—air migration and soil exposure. Based on
 3

these, the EPA computed a total HRS score of 59.18 for the
US Magnesium site. Because this score is above the threshold
for inclusion on the NPL, the EPA published a Proposed Rule
proposing to list USM on the NPL. National Priorities List,
Proposed Rule No. 49, 73 Fed. Reg. 51,393, 51,393 (Sept. 3,
2008). After receiving and responding to comments on the
proposed listing, the EPA added the site to the NPL. National
Priorities List, Final Rule No. 48, 74 Fed. Reg. 57,085, 57,087
(Nov. 4, 2009).

USM challenges the NPL listing as “arbitrary, capricious,
an abuse of discretion, or otherwise not in accordance with the
law.” 5 U.S.C. § 706(2)(A); Board of Regents of University
of Wash. v. EPA, 86 F.3d 1214, 1217 (D.C. Cir. 1996). It
claims that the EPA made four errors in calculating an HRS
score and that if these errors were corrected, the USM site’s
HRS score would fall below the 28.5 threshold for listing on
the NPL. Because the score assigned to the USM site is far
above the 28.5 cutoff for inclusion on the NPL, USM would
have to prevail on its first alleged error and some combination
of the other three alleged errors in order for the NPL listing to
be arbitrary or capricious. (If we were to remand based on all
the other three alleged errors, the site would still receive a
total score of at least 50—21.5 more than the minimum score
for inclusion on the NPL.) Although placement on the NPL
does not require any action or determine any party’s liability
for cleanup costs, it may inflict damage on business reputation
and cause a loss in property values. Kent County v. EPA, 963
F.2d 391, 394 (D.C. Cir. 1992).

We are not persuaded that the EPA in fact erred in the
first decision element claimed by USM to have been
erroneous—the scoring of multiple “sources” for the air
pathway.
 4

* * *

The essence of USM’s objection to the EPA’s scoring of
the air pathway is that the EPA multiplied the plant’s rather
high “release” score by the site’s total “waste characteristics
factor,” a factor that here was driven overwhelmingly by the
ponds’ relatively high waste quantity scores.1 Obviously such
a procedure has the potential to make a site’s score artificially
higher than that of a factually far more dangerous site in
which release and quantity were, say, middling at a single
source. It is a bit like choosing the winner of the “best team”
award at a track meet by multiplying each team’s highest
score in any single event by the team’s total number of
competitors (no matter how well or badly all other team
members may have performed). As we shall see, however,
the HRS directs precisely this procedure, and the EPA Hazard
Ranking System Guidance Manual (“HRS Manual”)2 invoked
by USM neither contradicts it nor suggests a different
treatment.

In scoring the air pathway, the EPA evaluated four
sources: the plant, the active waste pond, the inactive waste
pond, and three anode dust boxes located in the manufacturing
area of the plant. HRS Documentation Record, U.S.
Magnesium, EPA ID No. UTN000802704, (Nov. 2009)
(“HRS Documentation Record”) at 10, 13, 16, 19, Joint
Appendix (“J.A.”) 378, 381, 384, 387. USM appears to

1
In accordance with the HRS, the EPA did not calculate a
potential to release score for the ponds. HRS § 6.1.2. USM
assumes, as will we, that the potential to release score for the ponds
would be negligible.
2
United State Environmental Protection Agency, Hazard
Ranking System Guidance Manual, Interim Final, Publication No.
9345.1-07, EPA 540-R-92-026 (Nov. 1992).
 5

suggest that instead of using the system summarized above the
EPA ought to have scored each of the four sources for the air
migration pathway separately and used the highest of the four
individual source scores as the value for the air pathway.
Appellant’s Br. at 35-37.

But contrary to USM’s contention, the HRS clearly
contemplates that a pathway score for a site be computed by
the system of multiplication across sources to which USM
objects. A pathway is defined as a “[s]et of HRS factor
categories combined to produce a score to measure relative
risks posed by a site in one of four environmental
pathways . . .” 40 C.F.R. Pt. 300, App. A, § 1.1 (emphasis
added here and elsewhere in quotations from the HRS). A site
“may include multiple sources and may include the area
between sources.” Id. A source is “any area where a
hazardous substance has been deposited, stored, disposed, or
placed . . . .” Id. The air pathway score is calculated by
multiplying three factor categories: (1) likelihood of release,
(2) waste characteristics, and (3) “targets.” Id. § 6.0. The
targets score, which represents threats to nearby residents,
natural resources, or ecosystems, is not implicated in the error
alleged here.

The likelihood of release score is based either on an
“observed release” or on a score for “potential to release.”
Id. § 6.1. The waste characteristics score is obtained by
multiplying a score for toxicity/mobility and a score for
hazardous waste quantity. Id. § 6.2.3. USM’s complaint, in
essence, is that the EPA multiplied a likelihood of release
score based on an observed release from the plant by a waste
characteristics score based on waste quantity values from the
waste pools.

A likelihood of release score of 550 is assigned for the air
pathway if an “observed release” is documented for the site.
 6

Id. § 6.1.1. Otherwise, a “potential to release” score must be
calculated and used as the likelihood of release value. Id.
§ 6.1.3. An “observed release” may be established by
“demonstrating that the site has released a hazardous
substance to the atmosphere.” Id. § 6.1.1. The HRS plainly
requires the EPA to assign a likelihood of release of 550 for
any observed release into the atmosphere at the site. This is
exactly what the EPA did—based on direct observations of
release of chlorine gas from the plant, it assigned a score of
550. HRS Documentation Record at 32-33, J.A. 400-01. This
step is uncontested.

The waste characteristics score is the product of two
separate values: waste toxicity/mobility and waste quantity.
The HRS instructs the agency to “[e]valuate only those
hazardous substances available to migrate from the sources at
the site to the atmosphere” and “assign a toxicity factor value,
a mobility factor value and a combined toxicity/mobility
factor value” for “each hazardous substance.” 40 C.F.R. Pt.
300, App. A, §§ 6.2, 6.2.1. Once all sources have been
scored, the agency is to “[u]se the hazardous substance with
the highest toxicity/mobility factor value to assign the value to
the toxicity/mobility factor for the air migration pathway.” Id.
§ 6.2.1.3. The EPA evaluated both PCBs and
Hexachlorobenzene, determined that the latter had a higher
toxicity/mobility score and used its score of 1,000 for the air
pathway calculation. HRS Documentation Record at 34, J.A.
402. This scoring element is also uncontested.

The HRS instructs the EPA to

Evaluate the hazardous waste quantity factor by first
assigning each source (or area of observed
contamination) a source hazardous waste quantity value
as specified below. Sum these values to obtain the
 7

hazardous waste quantity factor value for the pathway
being evaluated.

40 C.F.R. Pt. 300, App. A, § 2.4.2. For three pathways,
including the air pathway, the agency is to “assign a source
hazardous waste quantity value to each source” including
only sources “having a containment factor value greater than 0
for the pathway being evaluated,” an exclusion in essence for
sources that effectively wall hazardous substances off from
escape. Id. §§ 2.4.2.1, 6.1.2.1.1, 6.1.2.2.1. In order to
calculate the final hazardous waste quantity factor value for
the site, the agency must “[s]um the source hazardous waste
quantity values assigned to all sources . . . for the pathway
being evaluated . . .” Id. § 2.4.2.2.

Here the EPA found that each of the four sources had
containment values of greater than zero for the air pathway.
HRS Documentation Record at 11, 14, 17, 19, J.A. 379, 382,
385, 387. It assigned quantity values for both waste pools.
Id. at 34, J.A. 402. It also found that the hazardous waste
quantity for the anode dust boxes and for the plant stack were
greater than zero but that the total amount was unknown. Id.
Therefore, as HRS § 2.4.2 explicitly directs, it added the
quantity values for the two waste ponds to obtain the waste
quantity factor value for the pathway. Id. It then multiplied
the toxicity/mobility factor value by the waste quantity factor
value to obtain a waste characteristics score as provided by
HRS § 6.2.3. Id.

The remainder of the calculation was purely mechanical.
The EPA entered the likelihood of release score and the waste
characteristics score in the air migration pathway scoresheet,
HRS Table 6-1. Id. It calculated the final air pathway raw
score by multiplying the likelihood of release score, the waste
characteristics score and the targets score and dividing the
 8

product by 82,500. 40 C.F.R. Pt. 300, App. A, § 6.4, Table 6-
1.

The HRS thus requires the EPA to score likelihood of
release on the basis of observed releases from any source on
the site, and to score waste quantity on the basis of the sum of
scores from all sources at the site. These two scores are
eventually multiplied, even in cases where the observed
release is from a source with a trivial waste quantity. The
same system prevails when an “observed release” cannot be
shown and the agency relies on “potential to release.” See 40
C.F.R. Pt. 300, App. A, § 6.1.2. The multiplication is not a
product of agency discretion, but an artifact of the scoring
methodology mandated by the HRS. So to the extent that
USM’s claim asserts arbitrary application of the HRS, it must
fail.

* * *

USM’s argument that the EPA violated the letter or spirit
of the source aggregation provisions in the HRS Manual is
similarly unavailing. The HRS Manual provides detailed
guidance on how to apply the HRS. The section invoked by
USM, § 4.2, discusses how to score sites with multiple
sources. It defines source aggregation as “[t]he treatment of
two or more areas that could be considered individual sources
as one discrete source.” HRS Manual § 4.2 at 49. And it goes
on to list various criteria appropriate to consider when
deciding whether to aggregate sources. Id. at 51. In general,
the EPA may aggregate sources that are very similar when
doing so would have little impact on the final score. Id. The
HRS Manual suggests that source aggregation is desirable
when this would have no impact on the overall HRS score
“because this should limit the number of separate sources
evaluated.” Id.
 9

In this case, the EPA aggregated three anode dust boxes
by treating them as one source for the purpose of scoring.
HRS Documentation Record at 16-18, J.A. 384-86. In the end
this aggregation had no material effect on the USM plant’s
score. It did, however produce a modest economy of effort:
the EPA took only two samples, whereas if each dust box had
been treated as a single source it would have had to take at
least one for each dust box. HRS § 2.2.2. Similarly, because
of the way scoring for “containment” proceeds, if the EPA
aggregated a source with zero containment with one with a
positive containment factor value, the positive finding would,
in most cases, trump the zero for the whole aggregated source.
See HRS § 6.1.2.1.1, Table 6-3. Aggregation in such a case
would risk violating the precept against any aggregation that
is likely to affect the score. HRS Manual § 4.2 at 51.

Here the EPA, though aggregating the dust boxes, did not
aggregate that “source” with the other three—the plant stack
and the two waste ponds. Rather, it analyzed each of these
four sources individually and then applied the HRS scoring
methodology to score the site as a whole. USM’s quarrel is
not with failure to apply the Manual, or failure to apply it
correctly, but with the process explicitly mandated in the HRS
and not contradicted by any passage in the Manual.

USM appears to believe that HRS Manual § 4.2 counsels
a policy inconsistent with USM’s scoring of the USM site.
But § 4.2 recognizes that some components of the HRS
pathway score are created by adding values for all sources at a
site (such as the waste quantity score under HRS § 2.4.2),3
while other components take the highest value for any source

3
In a typo, the EPA uses “hazardous waste quality” when its
citation to the HRS makes clear that it intends “hazardous waste
quantity.” See Manual § 4.2 at 50.
 10

at the site (such as the potential to release score for the air
pathway under HRS § 6.1.2). Id. We can find nothing in the
Manual that contradicts the clear language of the HRS.

* * *

USM presents its claim as one of arbitrary application of
the rules to its site. Certainly if the HRS gave the EPA a
choice of air migration pathway formulas and the EPA had
exercised discretion to use this formula rather than another
formula under these circumstances, that decision might well
have been arbitrary and capricious. But the HRS gives the
EPA no discretion to alter the air migration pathway score
formula when it produces peculiar results (or, indeed, on any
other ground).

In a sense, then, the real thrust of USM’s argument seems
more properly directed at the HRS regulations themselves.
Here it has not posed a challenge to the rationality of the HRS
regulation. See Oral Arg. Recording at 5:55-6:15, 15:44-
16:24. Nor does it argue that the EPA should have declined to
place the USM site on the NPL notwithstanding a HRS score
of over 28.5. Since the EPA’s scoring was consistent with the
HRS, we are left with no theory on which we may overturn
the EPA’s decision.

We note that CERCLA imposes exceptional limits on
efforts to attack the EPA’s regulations in this field:

Review of any regulation promulgated under this
chapter may be had upon application by any interested
person only in the Circuit Court of Appeals of the United
States for the District of Columbia. Any such application
shall be made within ninety days from the date of
promulgation of such regulations. Any matter with
respect to which review could have been obtained under
 11

this subsection shall not be subject to judicial review in
any civil or criminal proceeding for enforcement or to
obtain damages or recovery of response costs.

42 U.S.C. § 9613(a). We have said that this leaves in place a
party’s usual ability to petition for a rulemaking to revise such
regulations, see RSR Corp. v. EPA, 102 F.3d 1266, 1270 (D.C.
Cir. 1997), for the denial of which it could obtain judicial
review, see Massachusetts v. EPA, 549 U.S. 497, 527-28
(2007). In addition, even under § 9613 there may be some
room to challenge a regulation when litigating its application.
See RSR Corp., 102 F.3d at 1269-70; see also NLRB Union v.
FLRA, 834 F.2d 191, 195-97 (D.C. Cir. 1987); cf. National
Air Transportation Ass’n v. McArtor, 866 F.2d 483, 486-87
(D.C. Cir. 1989). Regardless of whether a challenge
paralleling USM’s “arbitrariness” contention here but framed
as a statutory attack on the HRS would be permissible under
such cases, USM failed to meet the prerequisite of raising the
issue before the EPA. See Linemaster Switch Corp. v. EPA,
938 F.2d 1299, 1308-09 (D.C. Cir. 1991); Letter from M.
Lindsay Ford, et al., Counsel to US Magnesium LLC to
United States Environmental Protection Agency (Nov. 24,
2008), J.A. 129-272.

* * *

Because the EPA followed the HRS precisely in scoring
the air migration pathway and affirmation of the EPA on that
first issue results in a score above the cut off for inclusion on
the NPL, listing of the USM site on the NPL was not arbitrary
or capricious. The petition for review is therefore

Denied.