You are on page 1of 14

Federal Register / Vol. 72, No.

187 / Thursday, September 27, 2007 / Proposed Rules 54875

This proposal also does not have minority or low income populations and for the District of Columbia Circuit has
Tribal implications because it would not determined it would not. issued several opinions construing
have a substantial direct effect on one or In reviewing State Plan submissions, section 112 (d) of the Clean Air Act, and
more Indian tribes, on the relationship EPA’s role is to approve State choices, one of those opinions has called into
between the Federal Government and provided that they meet the criteria of question the legality of some of the
Indian tribes, or on the distribution of the CAA. In this context, in the absence standards for hazardous waste
power and responsibilities between the of a prior existing requirement for the combusters. This notice discusses the
Federal Government and Indian tribes, State to use voluntary consensus standards that EPA promulgated in
as specified by Executive Order 13175 standards (VCS), EPA has no authority October 2005, and specifically identifies
(65 FR 67249, November 9, 2000). to disapprove a State Plan for failure to which standards EPA believes are
This proposed action also does not use VCS. It would thus be inconsistent consistent with the Act and caselaw,
have Federalism implications because it with applicable law for EPA, when it and which standards are not and need
would not have substantial direct effects reviews a State Plan submission, to use to be reexamined through a subsequent
on the States, on the relationship VCS in place of a State Plan submission rulemaking. With respect to those
between the national government and that otherwise satisfies the provisions of standards EPA intends to retain, this
the States, or on the distribution of the CAA. Thus, the requirements of notice indicates the portions of the
power and responsibilities among the section 12(d) of the National rationale upon which EPA intends to
various levels of government, as Technology Transfer and Advancement rely, and which portions EPA would no
specified in Executive Order 13132 (64 Act of 1995 (15 U.S.C. 272 note) do not longer rely upon as a justification for the
FR 43255, August 10, 1999). This action apply. This proposed rule would not October 2005 standards. EPA is seeking
merely proposes to approve a State rule impose an information collection public comment on this analysis. EPA
implementing a Federal standard. It burden under the provisions of the has also placed edited versions of
does not alter the relationship or the Paperwork Reduction Act of 1995 (44 various support documents in the
distribution of power and U.S.C. 3501 et seq.). public docket, edited to remove portions
responsibilities established in the CAA. of the rationale on which EPA no longer
This proposed rule also is not subject to List of Subjects in Part 62
plans to rely, and seeks public comment
Executive Order 13045 ‘‘Protection of Environmental protection, Air on these edits.
Children from Environmental Health pollution control, Electric utilities, DATES: Comments must be received on
Risks and Safety Risks’’ (62 FR 19885, Intergovernmental relations, Mercury, or before October 18, 2007.
April 23, 1997), because it proposes to Reporting and recordkeeping.
approve a State rule implementing a ADDRESSES: Submit your comments,
Dated: September 19, 2007. identified by Docket ID No. EPA–HQ–
Federal standard.
Executive Order 12898, ‘‘Federal John B. Askew, OAR–2004–0022, by one of the
Actions to Address Environmental Regional Administrator, Region 7. following methods:
Justice in Minority Populations and [FR Doc. E7–19120 Filed 9–26–07; 8:45 am] • www.regulations.gov: Follow the
Low-Income Populations,’’ requires BILLING CODE 6560–50–P on-line instructions for submitting
Federal agencies to consider the impact comments.
of programs, policies, and activities on • E-mail: a-and-r-docket@epa.gov.
minority populations and low-income ENVIRONMENTAL PROTECTION • Fax: 202–566–1741.
populations. EPA guidance 1 states that AGENCY • Mail: U.S. Postal Service, send
EPA is to assess whether minority or comments to: Air and Radiation Docket
low-income populations face risk or a 40 CFR Part 63 (2822T), Docket ID No. EPA–HQ–OAR–
rate of exposure to hazards that is [EPA–HQ–OAR–2004–0022; FRL–8474–2] 2004–0022, U.S. Environmental
significant and that ‘‘appreciably Protection Agency, 1200 Pennsylvania
RIN 2050–AG29 Avenue, NW., Washington, DC 20460.
exceed[s] or is likely to appreciably
exceed the risk or rate to the general Please include a total of two copies.
NESHAP: National Emission Standards • Hand Delivery: In person or by
population or to the appropriate for Hazardous Air Pollutants:
comparison group.’’ (EPA, 1998) courier, deliver comments to: HQ EPA
Standards for Hazardous Waste Docket Center, Public Reading Room,
Because this rule merely proposes to Combustors
approve a state rule implementing the EPA West, Room 3334, 1301
Federal standard established by CAMR, AGENCY: Environmental Protection Constitution Avenue, NW., Washington,
EPA lacks the discretionary authority to Agency (EPA). DC 20004. Such deliveries are only
modify today’s regulatory decision on ACTION: Solicitation of comment on legal accepted during the Docket’s normal
the basis of environmental justice analysis. hours of operation, and special
considerations. However, EPA has arrangements should be made for
already considered the impact of CAMR, SUMMARY: On October 12, 2005, deliveries of boxed information. Please
including this Federal standard, on pursuant to section 112(d) of the Clean include a total of two copies.
minority and low-income populations. Air Act, EPA issued national emission Instructions: Direct your comments to
In the context of EPA’s CAMR standards for hazardous air pollutants Docket ID No. EPA–HQ–OAR–2004–
published in the Federal Register on (NESHAP) emitted by various types of 0022. The EPA’s policy is that all
May 18, 2005, in accordance with hazardous waste combusters. EPA comments received will be included in
Executive Order 12898, the Agency has subsequently granted reconsideration the public docket without change and
mstockstill on PROD1PC66 with PROPOSALS

considered whether CAMR may have petitions relating to certain issues may be made available online at
disproportionate negative impacts on presented by the rules. 71 FR 14665, www.regulations.gov, including any
52564, but has not yet issued a final personal information provided, unless
1 U.S. Environmental Protection Agency, 1998.
determination on reconsideration. the comment includes information
Guidance for Incorporating Environmental Justice Following the close of the comment claimed to be Confidential Business
Concerns in EPA’s NEPA Compliance Analyses.
Office of Federal Activities, Washington, DC, April, period on the proposed reconsideration Information (CBI) or other information
1998. rule, the United States Court of Appeals the disclosure of which is restricted by

VerDate Aug<31>2005 17:46 Sep 26, 2007 Jkt 211001 PO 00000 Frm 00016 Fmt 4702 Sfmt 4702 E:\FR\FM\27SEP1.SGM 27SEP1
54876 Federal Register / Vol. 72, No. 187 / Thursday, September 27, 2007 / Proposed Rules

statute. Do not submit information that and cannot contact you for clarification, Reading Room, EPA West, Room 3334,
you consider to be CBI or otherwise EPA may not be able to consider your 1301 Constitution Avenue, NW.,
protected through www.regulations.gov comment. Electronic files should avoid Washington, DC 20004. The Public
or e-mail. The www.regulations.gov Web the use of special characters, any form Reading Room is open from 8:30 a.m. to
site is an ‘‘anonymous access’’ system, of encryption, and be free of any defects 4:30 p.m., Monday through Friday,
which means EPA will not know your or viruses. For additional information excluding legal holidays. The telephone
identity or contact information unless about EPA’s public docket visit the EPA number for the Public Reading Room is
you provide it in the body of your Docket Center homepage at http:// (202) 566–1744, and the telephone
comment. If you send an e-mail www.epa.gov/epahome/dockets.htm. number for the HQ EPA Docket Center
comment directly to EPA without going Docket: All documents in the docket is (202) 566–1742. A reasonable fee may
through www.regulations.gov, your e- are listed in the www.regulations.gov be charged for copying docket materials.
mail address will be automatically index. Although listed in the index, FOR FURTHER INFORMATION CONTACT: For
captured and included as part of the some information is not publicly more information on this notice, contact
comment that is placed in the public available, e.g., CBI or other information Frank Behan at (703) 308–8476, or
docket and made available on the the disclosure of which is restricted by behan.frank@epa.gov, Office of Solid
Internet. If you submit an electronic statute. Certain other material, such as Waste (5302P), U.S. Environmental
comment, EPA recommends that you copyrighted material, will be publicly Protection Agency, 1200 Pennsylvania
include your name and other contact available only in hard copy. Publicly Ave., NW., Washington, DC 20460.
information in the body of your available docket materials are available SUPPLEMENTARY INFORMATION: Entities
comment and with any disk or CD–ROM either electronically in Potentially Affected by this Action.
you submit. If EPA cannot read your www.regulations.gov or in hard copy at Categories and entities potentially
comment due to technical difficulties the HQ EPA Docket Center, Public affected by this action include:

NAICS
Category Potentially affected entities
codea

Petroleum and coal products manufacturing ..................................... 324 Any entity that combusts hazardous waste as defined in the final
rule.
Chemical manufacturing .................................................................... 325
Cement and concrete product manufacturing ................................... 3273
Other nonmetallic mineral product manufacturing ............................ 3279
Waste treatment and disposal ........................................................... 5622
Remediation and other waste management services ....................... 5629
a North American Industry Classification System.

This table is not intended to be issued on October 12, 2005 (70 FR Tips for Preparing Your Comments.
exhaustive, but rather provides a guide 59402) and the two petition for When submitting comments, remember
for readers regarding entities likely to be reconsideration notices published on to:
impacted by this action. This table lists March 23, 2006 (71 FR 14665) and • Identify the rulemaking by docket
examples of the types of entities EPA is September 6, 2006 (71 FR 52624). number and other identifying
now aware could potentially be Preparation of Comments. Do not information (subject heading, Federal
regulated by this action. Other types of submit this information to EPA through Register date and page number).
entities not listed could also be affected. www.regulations.gov or e-mail. Clearly • Follow directions—The agency may
To determine whether your facility, mark the part or all of the information ask you to respond to specific questions
company, business, organization, etc., is that you claim to be CBI. For CBI or organize comments by referencing a
affected by this action, you should information in a disk or CD–ROM that Code of Federal Regulations (CFR) part
examine the applicability criteria in 40 you mail to EPA, mark the outside of the or section number.
CFR 63.1200.1 If you have any questions disk or CD–ROM as CBI and then • Explain why you agree or disagree;
regarding the applicability of this action identify electronically within the disk or suggest alternatives and substitute
to a particular entity, consult the person CD–ROM the specific information that language for your requested changes.
listed in the preceding FOR FURTHER is claimed as CBI. In addition to one • Describe any assumptions and
INFORMATION CONTACT section. complete version of the comment that provide any technical information and/
How Do I Obtain a Copy of this includes information claimed as CBI, a or data that you used.
Document and Other Related copy of the comment that does not • If you estimate potential costs or
Information? In addition to being contain the information claimed as CBI burdens, explain how you arrived at
available in the docket, an electronic must be submitted for inclusion in the your estimate in sufficient detail to
copy of today’s proposed rule will also public docket. Information so marked allow it to be reproduced.
be available on the World Wide Web will not be disclosed except in • Provide specific examples to
(WWW). Following the Administrator’s accordance with procedures set forth in illustrate your concerns, and suggest
signature, a copy of this document may 40 CFR part 2. Send or deliver alternatives.
be posted on the WWW at http:// information identified as CBI to only the • Explain your views as clearly as
mstockstill on PROD1PC66 with PROPOSALS

www.epa.gov/hwcmact. This Web site following address: Ms. LaShan Haynes, possible.
also provides other information related RCRA Document Control Officer, EPA • Make sure to submit your
to the NESHAP for hazardous waste (Mail Code 5305P), Attention Docket ID comments by the comment period
combustors including the NESHAP No. EPA–HQ–OAR–2004–0022, 1200 deadline identified.
Pennsylvania Avenue, Washington DC, Organization of this Document. The
1 Unless otherwise noted, all regulatory references 20460. Clearly mark the part or all of the information presented in this notice is
in this notice are to 40 CFR. information that you claim to be CBI. organized as follows:

VerDate Aug<31>2005 17:46 Sep 26, 2007 Jkt 211001 PO 00000 Frm 00017 Fmt 4702 Sfmt 4702 E:\FR\FM\27SEP1.SGM 27SEP1
Federal Register / Vol. 72, No. 187 / Thursday, September 27, 2007 / Proposed Rules 54877

I. Background metals in all inputs to the units which performer; over time it emits less
II. Consideration of Variability in are not hazardous waste); hydrogen cadmium than source A. Indeed, given
Establishing MACT Floors chloride/chlorine (measured as total that the chief health risks of most HAP
III. Discussion of Individual Standards chlorine) (‘‘TCl’’); carbon monoxide/ emitted by Hazardous Waste
A. Standards for Particulate Matter
1. Standards for Incinerator, Cement Kilns, total hydrocarbons (‘‘CO/HC’’) (as Combustors results from chronic rather
Lightweight Aggregate Kilns, and Solid surrogates for non-dioxin organic HAP than acute exposure (i.e., amount of
Fuel Boilers (and in a few cases, dioxin as well); and repeated exposure over time as opposed
2. Standards for Liquid Fuel Boilers destruction removal efficiency (‘‘DRE’’) to single exposure incidents), floor
B. Standards for Semivolatile Metals and (an aspect of control of non-dioxin standards based on evaluation of
Low Volatility Metals organic HAP, and in a few cases, sources’ performance over time (i.e.,
1. Methodology to Establish Floor Levels dioxin). standards which account for sources’
2. Alternatives to the Particulate Matter On March 13, 2007, the United States variability) best address the sources’
Standard for Incinerators, Liquid Fuel Court of Appeals for the District of ultimate impacts on human health. See
Boilers, and Solid Fuel Boilers
Columbia Circuit (D.C. Circuit) issued 70 FR at 59533–35 where EPA discusses
3. Alternative Mercury, Semivolatile
Metals, Low Volatile Metals, and Total its decision in Sierra Club v. EPA, 479 human health benefits of the standards
Chlorine Standards for Cement Kilns and F.3d 875 (2007) (‘‘Brick MACT’’). EPA considering reductions in chronic
Lightweight Aggregate Kilns has reexamined the rule to determine if exposure to HAP.
4. Alternative Mercury Standards for it is compliant with the statute with
respect to the issues discussed in the II. Consideration of Variability in
Cement Kilns and Lightweight Aggregate
Kilns Under the Interim Standards Court’s opinion, and specifically Establishing MACT Floors
C. Standards for Total Chlorine whether the MACT floors for each EPA may consider variability in
1. Incinerators standard are compliant. For the most identifying best performers and their
2. Cement Kilns part, EPA believes that they are. The level of performance. See 70 FR at
3. Lightweight Aggregate Kilns 59436. See also Brick MACT, 479 F.3d
basic reason, for those standards EPA
4. Liquid Fuel Boilers
plans to retain, is that the rule identifies at 881–82 (variability of best performing
5. Solid Fuel Boilers
6. Hydrochloric Acid Production Furnaces as best performers—the best performing sources may be taken into account in
D. Standards for Dioxins/Furans 12 per cent or best performing five establishing MACT floors).
1. Incinerators sources in smaller source categories for EPA in this rule identified two types
2. Cement Kilns existing sources, and the best controlled of variability, run-to-run variability and
3. Lightweight Aggregate Kilns single source for new sources—those test-to-test variability. Run-to-run
4. Liquid Fuel Boilers sources which are likely to emit the variability ‘‘encompasses variability in
E. Non-Dioxin Organic HAP least HAP over time, and reasonably individual runs comprising the
F. Mercury estimates these sources’ level of compliance tests, and includes
1. Incinerators uncertainties in correlation of
performance. Put another way, the rule
2. Cement Kilns monitoring parameters and emissions,
3. Lightweight Aggregate Kilns identifies as best performers those
emitting the least HAP considering and imprecision of stack test methods
4. Liquid Fuel Boilers
5. Solid Fuel Boilers variability (i.e., their performance over and laboratory analyses.’’ 70 FR at
G. Normalization time), and accounts for that variability 59437. A shorthand description is that
H. Potential Implications to the as much as possible in estimating these this is within-test variability. EPA
Compliance Date Provisions If Standards sources’ level of performance. See 70 FR quantified run-to-run variability using
Are Remanded to EPA at 59346 (‘‘best performers are those that the statistical methodology set forth in
perform best over time (i.e., day-in, day- Technical Support Document (‘‘TSD’’)
I. Background
out)’’). Vol. III section 7.2; 2 see also 70 FR at
The Hazardous Waste Combustor The statute does not address the 59437/1–2, 59438, and 59439 explaining
(HWC) Maximum Achievable Control question of whether, in assessing which the reasonableness of this statistical
Technology (MACT) rule, 70 FR 59402 sources perform best or are best approach. The chief element of this
(October 12, 2005), adopts separate controlled, emission levels should be quantification is simply the standard
standards for six source categories, the evaluated over time, or in a single test deviation in the performance test data
common link being that sources in each result. Nor does Brick MACT, which (standard deviation being the usual
category burn hazardous waste. These states at 479 F.3d 880 that ‘‘section statistical measure for assessing
sources are incinerators, cement kilns, [112(d)(3)] requires floors based on the variation within a data set by comparing
lightweight aggregate kilns, solid fuel emission level actually achieved by the a single result with the average of the
boilers, liquid fuel boilers, and best performers (those with the lowest data comprising the data set). The result
hydrochloric acid production furnaces. emission levels)’’, but does not refer to is an estimate of the value which the
Liquid fuel boilers are further a time period for measurement. The source would achieve in 99 of 100
subcategorized into those burning following example shows why it is future tests if it replicated the operating
higher heating value hazardous wastes reasonable to determine which sources conditions of the compliance test. 70 FR
and lower heating value hazardous are the best performers by accounting in at 59437; 3 see also 69 FR at 21232 and
wastes. The following hazardous air the first instance for what their n. 69 (April 20, 2004).
pollutants (‘‘HAP’’) are regulated for emissions are over time. Assume that
each of these source categories: dioxins source A in a single test emitted 10 2 USEPA, ‘‘Technical Support Document for HWC

and furans (‘‘D/F’’); semivolatile metals units of cadmium, and source B emitted MACT Standards, Volume III: Selection of MACT
Standards,’’ (TSD Vol. III) September 2005. Unless
mstockstill on PROD1PC66 with PROPOSALS

(lead and cadmium) (‘‘SVM’’); low 15 units. However, assume further that otherwise specified, all TSD references in this
volatile metals (arsenic, beryllium and over time source A emits cadmium at a notice are to this document, which is available in
chromium) (‘‘LVM’’); mercury, rate of 40 units and source B emits the docket to the rule. See docket items EPA–HQ–
particulate matter (‘‘PM’’) (as a surrogate cadmium at a rate of 25 (the difference OAR–2004–0022–0453, 0457, 0459, and 0460.
3 More precisely, this is a modified prediction
for the remaining HAP metals being that source B’s performance is less limit that ensures at the 95% confidence level that
(antimony, cobalt, manganese, nickel, variable). It is at the very least the average of the best performing sources could
and selenium), and also to control HAP reasonable to view source B as the better Continued

VerDate Aug<31>2005 17:46 Sep 26, 2007 Jkt 211001 PO 00000 Frm 00018 Fmt 4702 Sfmt 4702 E:\FR\FM\27SEP1.SGM 27SEP1
54878 Federal Register / Vol. 72, No. 187 / Thursday, September 27, 2007 / Proposed Rules

Existence of run-to-run variability is single emission tests may not be the by the standards for semi-volatile and
confirmed most evidently by the wide lowest emitters over time due to their low-volatile HAP metals (referred to as
variations within different runs of the test-to-test variability. 70 FR at 59437– ‘nonenumerated metals’ in this
best performers’ performance tests. 438 and TSD Vol. III chapters 16 and 17; rulemaking). See section III.B. below. In
Moreover, simply averaging these see also 70 FR at 59439 explaining why addition, as explained in section III.B.,
different run results would lead to total variability is not accounted for by the PM standard also controls all non-
standards which not even the best of the compliance test conditions. mercury HAP metals (i.e., semi-volatile,
best performers would achieve over EPA was able to provide a low volatility, and nonenumerated HAP
time. TSD Vol. III section 16.4. quantitative estimate of test-to-test metals) in all nonhazardous waste
Comparative test results of best variability in only one instance—where inputs to HWCs. 70 FR at 59459. Since
performing sources (i.e., tests of the fabric filters are used to capture the PM standards are measured by total
same source at a different time) strongly particulate matter. See discussion of PM end-of-stack output, these standards
suggest that run-to-run variability can be standards in section III.A. below. In account for all HAP metal input to
appreciable (although not the only other instances, EPA accounted for test- hazardous waste combustion devices
measure of variability), since these to-test (i.e., long-term) variability in one (other than mercury). Id.
sources have been shown consistently to of two ways: (a) Selecting as best EPA used the Air Pollution Control
emit more than the averaged emissions performers those which minimized their Device methodology to establish floors
from the performance test identifying long-term (i.e., test-to-test) variability by for PM. Under this methodology, EPA
the source as best performing. See TSD best controlling the aspects of determined as a matter of engineering
Vol. III Tables 16–4, 16–5, 17–1, 17–3. performance (notably removal efficiency judgment which devices best control
Failure to consider run-to-run evaluated systemwide and hazardous PM emissions, ranked these means of
variability could seriously waste HAP feedrate) within their control, and selected as the best
underestimate a source’s emissions over control, or (b) using a surrogate for the performers those with the lowest PM
time. See TSD Vol. III section 17.3.3, HAP where EPA could assess the long- emissions using the best control device.
showing that even the lowest emitting term variability associated with See TSD Vol. III section 7.4; see also id.
Straight Emission sources could have emissions of that surrogate, but could at 16–2 ranking PM control devices from
emissions higher than floor levels under not otherwise assess long-term best to worst for each source category.
a methodology that considers run-to-run variability. The floor for each source category was
variability. EPA has comparative data EPA also carefully assessed a floor then established based on the average of
from a number of lowest emitting methodology which simply assumed these lowest emitting sources’ PM
incinerators for PM in single test results. that the lowest emitters in individual emissions (or the lowest emitter of these
In other tests, these same sources were performance tests were the best sources for the new source floor).
performers. The major problem with In most instances, the lowest emitters
typically unable to achieve the same
such a methodology is that it ignores the in the performance test used for
level of performance, sometimes
sources’ performance over time, leading determining best performers were
emitting up to seven times more PM. 69
to situations where the sources’ level of equipped with the best control device—
FR at 21232 and n. 69 (April 20, 2004).
performance may be assessed some type of fabric filter (‘‘FF’’).
Test-to-test variability results from
improperly. See TSD Vol. III chapters 16 Occasionally, a lower PM emitter in a
variability in pollution device control
and 17; 70 FR at 59442–446 (explaining single test was equipped with some
efficiencies over time (depending on
why lowest emitters in individual other type of control device, or, in the
multitudinous factors, including for
performance tests 4 are not always the case of three incinerators, no control
fabric filters the point in the
best performers). EPA consequently device, but EPA ranked these sources as
maintenance cycle at which the source lower (i.e., worse) performing than FF-
is tested, and for electrostatic used this methodology to identify best
performers and their level of equipped sources. EPA reevaluated
precipitators variations in combustion carefully whether the lower ranking of
gas moisture and particle resistivity), as performance when it was not possible to
assess sources’ waste feedrate and these sources, in some instances
well as measurement variability resulting in their omission from the pool
resulting from different sampling crews systemwide removal efficiency.
of best performers, is consistent with the
under different meteorological III. Discussion of Individual Standards holding of Brick MACT, 479 F.3d at
conditions and different analytical 882–83, as well as Cement Kiln
laboratories. Id. and n. 63. A shorthand A. Standards for Particulate Matter
Recycling Coalition v. EPA, 255 F.3d
description is that this is long-term 1. Standards for Incinerator, Cement 855, 863–65 (D.C. Cir. 2001), that floors
variability. EPA demonstrated generally Kilns, Lightweight Aggregate Kilns, and are not to be set only on performance of
that: (a) Test-to-test variability exists; (b) Solid Fuel Boilers sources equipped with certain
it is not encompassed in EPA’s EPA adopted standards for particulate technology unless that is the only factor
statistical quantification of run-to-run matter (‘‘PM’’) for all of the hazardous affecting emissions, and that EPA must
variability; (c) the amount of test-to-test waste combuster source categories consider all means of control when
variability can be significant such that except for hydrochloric acid production selecting best performers.
failing to account for it in some manner furnaces.5 Particulate Matter is a EPA of course accepts these holdings,
means that the sources’ performance surrogate for the HAP metals antimony, and believes its approach here is
over time can be seriously cobalt, manganese, nickel, and consistent with the statute and
underestimated (i.e., since their long- selenium, the HAP metals not covered applicable case law. EPA selected as
term variability would be ignored); and
mstockstill on PROD1PC66 with PROPOSALS

best performers (or as the best


(d) sources which are lowest emitting in 4 The heading to this preamble section should controlled source) those sources it
have explicitly included the words ‘‘in individual estimated to have the lowest PM
achieve the emission level in 99 or 100 future test performance tests’’ in the section title. emissions over time. EPA’s selection
conditions based on a three-run average, assuming 5 The alternative metal standards, in lieu of PM

the best performers could initially replicate the standards, for incinerators, and liquid and solid fuel
process has a reasoned basis. Sources
compliance test conditions. TSD Vol. III at 7–7; 70 boilers are discussed in section III.B discussing equipped with control devices other
FR at 59437. standards using the SRE Feed floor methodology. than FFs are likely to emit more over

VerDate Aug<31>2005 17:46 Sep 26, 2007 Jkt 211001 PO 00000 Frm 00019 Fmt 4702 Sfmt 4702 E:\FR\FM\27SEP1.SGM 27SEP1
Federal Register / Vol. 72, No. 187 / Thursday, September 27, 2007 / Proposed Rules 54879

time than they do in individual test Variability Factor algorithm for fabric approach here is consistent with these
conditions, even after adjusting test filters, which is derived from the requirements.
results to account for run-to-run quantified measure of the total The D.C. Circuit has stressed in both
variability. (Put another way, these variability (i.e., both run-to-run and Brick MACT and CKRC that factors such
sources’ performance in individual test long-term test-to-test variability) of the as low HAP feed that influence
conditions are likely not representative FF-equipped hazardous waste emissions cannot be ignored in
of what they will emit over time.) This combusters identified as best performers assessing performance. 479 F.3d at 882–
is because test-to-test variability, that is, based on the historical test conditions 83; 255 F.3d at 864–65. EPA thus
long-term variability, has not been taken for those sources. See TSD Vol. III carefully reexamined those instances
into account. Since these other control section 5.3. where low PM emitters in single tests
devices are known to be more variable As a result, for HWCs EPA has a were not equipped with any pollution
and less efficient than FFs, TSD Vol. III considerably more reliable idea of what control equipment so that their emission
pp. 16–3 to 4 and 11, failure to consider fabric filter-equipped sources’ actual levels necessarily reflected low ash
long-term variability (i.e., looking performance for PM is over time than inputs. There are three incinerators that
exclusively at results of single for any other type of control device- had lower PM emissions in single tests
performance tests) results in these equipped source (or for sources without that were lower than the worst of the
sources’ performance not being fully air pollution control devices). Second, lowest-emitting FF-equipped
characterized. Long-term variability as just noted above, the record incinerators on whose performance the
exists due to, among other things, demonstrates that the performance data floor standard is based. TSD Vol. III
variation over time in control device from sources that emitted less PM in App. F at APCD–INC–PM. EPA
performance and varying ash feed individual performance tests but are not continues to believe that it properly
rates.6 EPA confirmed in a series of equipped with FFs significantly chose not to include these sources
analyses of HWCs that this test-to-test underestimates the amount of PM these among the pool of best performers. First,
variability for non-FF equipped devices sources emit over time (i.e., fails to even in single test conditions, these
both exists and is appreciable. See TSD account for their long-term variability). sources’ emissions were not
Vol. III section 16.5 showing among Third, over time, these emissions in significantly lower (0.0018 to 0.0009 gr/
other things that ostensibly lowest some instances exceed (i.e., are higher dscf lower, that is, roughly a 7–14%
emitting, non-FF equipped sources in than) the lowest emitting FF-equipped difference) than the average of the best
other tests (i.e., other occasions when sources, even though emitting less in an performing 12% of sources EPA
the same source was tested) were unable individual performance test. 70 FR at identified as best performing using the
to duplicate (i.e., achieve): (a) Their own 59448; TSD Vol. III section 16.5. Putting Air Pollution Control methodology. Id.
level of performance (i.e., their all this together, EPA selected the These sources also emit more PM than
performance in the other test), (b) their lowest emitting FF-equipped sources as all but one of the best performing
own performance adjusted to account the best performing. 70 FR at 59448. incinerators in EPA’s pool of best
for run-to-run variability, (c) floors This approach is consistent with the performers, and the difference in
based on the average of the lowest single statute and applicable case law. EPA performance between these
test emitters’ performance, (d) design selected as best performers (or best uncontrolled sources and the last of the
level of the floor actually adopted in the controlled sources) those sources it EPA pool is small, roughly a factor of 2.
rule (i.e., the level sources would design reasonably estimated to have the lowest Id. Since these devices lack any
to in order to comply with the rule), PM emissions over time. Performance of pollution control equipment, their
and, in one case, (e) the floor level units equipped with fabric filters can be performance over time will be highly
established in the rule (i.e., the floor reliably estimated over time—i.e., all of variable as ash feedrates vary and their
reflecting application of the Air the variability can be quantified. emissions could 7 well exceed the
Pollution Control Device methodology). Performance of other units over time emissions of the sources comprising
EPA further examined whether this cannot be estimated as reliably (the EPA’s pool of best performing
difference in performance resulted from long-term variability cannot be incinerators. Second, and of at least
legitimate operating variability, rather quantified at all), but is known to be less equal importance, low ash feedrates are
than from differing ash feed rates, and efficient and more variable. Short-term not a guarantee of low HAP metal
in the instance where direct comparison performance tests thus demonstrably emissions. Low PM emissions from
was possible, determined that it did not. and dramatically understate the amount uncontrolled sources could still reflect
TSD Volume III pp. 16–15 through 17. of PM (and HAP metal) these sources high metal HAP emissions since, if the
In contrast, EPA was able to quantify emit, so that these units could (and ash has high metal content, all of it
the long-term performance (i.e., demonstrably do in some instances) would be emitted. See 70 FR at 59449
performance accounting for both run-to- emit more PM (and therefore more HAP (‘‘ash feedrates are not reliable
run and test-to-test variability) of HWC metal) than the lowest emitting FF- indicators of nonmercury metal HAP
sources equipped with FFs. This is the equipped sources notwithstanding feed control levels and are therefore
only type of air pollution control device lower PM emissions in individual tests. inappropriate parameters to assess in
for HWCs, and the only pollutant, for The D.C. Circuit has held repeatedly the MACT evaluation process. For
which such a calculation is possible. that EPA may use reasonable means to example, a source could reduce its ash
The reason this quantification is estimate the performance of best feed input by reducing the amount of
possible is that FFs are less variable performing sources, and may account silica in its feedstreams. This would not
than other control devices, and perform for sources’ variability in doing so. result in * * * emission reductions of
mstockstill on PROD1PC66 with PROPOSALS

relatively constantly regardless of input CKRC, 255 F.3d at 865–66; Mossville, metal HAP’’). In contrast, ‘‘particulate
loadings. 70 FR at 59449. EPA thus 370 F.3d at 1240, 1242; National Lime matter emissions from baghouses [e.g.,
developed a so-called Universal Ass’n v. EPA, 627 F. 2d 416, 431 n. 46, FF-equipped units] are not significantly
443 (D.C. Cir. 1980); see also Brick affected by inlet particulate matter
6 Ash content is an indicator of the MACT, 479 F.3d at 881–82 (estimates of
noncombustible matter (i.e., inorganic content, variability are to be for the variability of 7 There are no comparative test data in the record

including metals) in the feed to the source. the best performing sources). EPA’s for these sources.

VerDate Aug<31>2005 17:46 Sep 26, 2007 Jkt 211001 PO 00000 Frm 00020 Fmt 4702 Sfmt 4702 E:\FR\FM\27SEP1.SGM 27SEP1
54880 Federal Register / Vol. 72, No. 187 / Thursday, September 27, 2007 / Proposed Rules

loadings’’, id., so that PM (and hence EPA is less certain that these LFBs emitters as the sources with the best
HAP metal emissions) from these units could emit more PM over time than the combination of hazardous feed rate
will remain best controlled regardless of FF-equipped sources EPA selected as control and back end control (removal
relative amounts fed to the device. See best performers and therefore will efficiency across the entire system).
also TSD Vol. III section 17.7 reexamine the standard with a view to Some of these sources were also the
documenting that PM emissions from amending it. However, EPA notes lowest emitters in single test results, but
FF-equipped sources are not affected further that this difference in emission were not in all cases. EPA selected this
appreciably by inlet loadings. EPA is levels between controlled and methodology, rather than the so-called
thus giving preference as best uncontrolled sources suggests that Straight Emissions approach of simply
performers to those incinerators we subcategorization may be appropriate. identifying best performers as those
know are effectively controlling non- EPA intends to investigate that with the lowest emissions after
mercury metal HAP because they are the possibility in subsequent rulemaking. accounting for run-to-run variability,
lowest emitting of the most efficiently because the SRE Feed methodology
controlled sources. Moreover, although B. Standards for Semivolatile Metals better identifies who the lowest emitters
a severable part of the rationale, EPA and Low Volatility Metals will be over time, and better assesses
believes it reasonable that most 1. Methodology To Establish Floor their performance (i.e., how much SVM
efficiently controlled sources can be Levels or LVM they will emit as they operate).
viewed as ‘‘best performing’’ and ‘‘best 70 FR at 59441–442; TSD Vol. III at 17–
EPA used the so-called system
controlled’’ under appropriate 1. SRE Feed best performers are likely
removal efficiency/hazardous waste
circumstances. See discussion in section to emit less of these metals over time
feed control (‘‘SRE Feed’’) methodology
B.1 below. than sources identified as best under the
EPA does, however, believe that to establish floor levels for semivolatile
Straight Emissions methodology—
certain parts of the justification for the metal HAP (‘‘SVM’’—lead and
averaged performance of lowest emitting
PM standards in the final rule are not cadmium) and low volatile metal HAP
sources in the most recent performance
proper after Brick MACT, and EPA is no (‘‘LVM’’—arsenic, beryllium, and
test accounting for run-to-run variability
longer relying on them and will revise chromium) for all source categories
(see TSD Vol. III at section 7.2)—
the record accordingly. The principal except hydrochloric acid production
because the Straight Emissions
revisions are to discussions relating to furnaces. Under this methodology, best
methodology (even after accounting for
how EPA considered raw material performers are ranked by hazardous
run-to-run variability) ignores sources’
inputs in assessing which sources are waste feed rate of metal HAP, and by
long-term (test-to-test) variability, and
best performers. See Brick MACT, 479 F. system removal efficiency (the degree to
so underestimates (indeed, ignores)
3d at 882–83. The specific alterations which HAP are removed from stack
their performance over time. The SRE
EPA is contemplating (generally emissions across the entire system, be it
Feed methodology accounts for test-to-
excising existing language) are found in by an air pollution control device or by
test variability, albeit qualitatively. Id.8
red line/strike out versions of the any other means). 70 FR at 59441. Best
For the same reason, the SRE Feed
Preamble, Technical Support performers are those with the best
methodology better estimates sources’
Documents, and Response to Comment combination of hazardous waste feed
performance over time since it accounts
Document which EPA has placed in the rate for the HAP at issue and system
in some measure for their long-term
docket for this rule. removal efficiency (i.e., lowest
variability instead of ignoring it. As
hazardous waste feed rate and best
2. Standards for Liquid Fuel Boilers discussed earlier, elements of long-term
removal efficiency). EPA assessed SVM
variability include such things as
EPA’s initial decision is not to defend and LVM separately, so that there are
chlorine feed rates (since metals are
the PM standard for liquid fuel boilers separate pools of best performing
more volatile in the chlorinated form),
(LFBs), and we thus contemplate sources for each of these HAP metal
back-end control devices’ controllable
requesting the Court to remand the groups for each of the source categories.
Once best sources are identified by operating parameters (e.g. ESP power
standard so that EPA can reexamine it.
this methodology, EPA calculated the levels, pressure drop across baghouses,
Most of the liquid fuel boilers with
lowest PM emissions are uncontrolled floor (accounting for run-to-run and other such operating parameters),
units with extremely low ash feeds. TSD variability) based on the averaged the matrix in which the metal is fed
Vol. III App. F at APCD–LFB–PM. emission levels of SVM or LVM from (solid, liquid, pumpable) and the
Unlike the situation with incinerators, these best performing sources (or for hazardous waste feedrate. TSD Vol. III at
the difference in PM emissions between new sources, the SVM or LVM emission p. 17–5. SRE Feed best performers are
these sources and those lowest-emitting level of the single best performer). For those that best control these and other
LFBs equipped with FFs is great, source categories where SVM and LVM controllable parameters and therefore
ranging from a factor of 6 (comparing standards are normalized by hazardous are less variable (i.e., are more efficient
lowest emitting FF-equipped LFB with waste heat input (cement kilns, at controlling SVM and LVM
lowest emitting uncontrolled LFB) to lightweight aggregate kilns, and the emissions), and therefore likely to emit
over three orders magnitude (comparing higher heating value hazardous wastes less SVM and LVM over time. Id. at p.
worst of the lowest emitting FF- subcategory for liquid fuel boiler), see 17–11. Put more broadly, the
equipped LFB to lowest emitting 70 FR at 59451–53, the standard is methodology best evaluates the two
uncontrolled LFB). Id. These expressed exclusively in terms of SVM things sources can do to control SVM
uncontrolled sources’ emissions are also or LVM attributable to hazardous waste and LVM emissions: limit the feed rate
of these HAP in hazardous waste (since
mstockstill on PROD1PC66 with PROPOSALS

roughly an order of magnitude lower inputs. For all source categories, total
than the promulgated floor based on SVM and LVM emissions are addressed hazardous waste feed rate is controlled
performance of FF-equipped sources. Id. and controlled by the PM standard. under RCRA rules), and manage
There are also ten uncontrolled LFBs in The SRE Feed methodology does not 8 See TSD Vol. III at 17–1 to 4 explaining why
the data base with lower PM emissions always identify the lowest emitters of long-term variability for SVM and LVM cannot be
than the lowest emitting FF-equipped SVM or LVM in single tests as the best determined quantitatively, even for sources
LFB. Id. Under these circumstances, performers; it identifies the lowest equipped with baghouses (FFs).

VerDate Aug<31>2005 17:46 Sep 26, 2007 Jkt 211001 PO 00000 Frm 00021 Fmt 4702 Sfmt 4702 E:\FR\FM\27SEP1.SGM 27SEP1
Federal Register / Vol. 72, No. 187 / Thursday, September 27, 2007 / Proposed Rules 54881

controllable parameters to limit emitted more SVM and LVM in entity which does not even presently
emissions across the entire system (both previous performance tests than they exist. See 71 FR 14665 (March 23, 2006).
through emission control device control did in the single performance test As noted above, the SVM and LVM
and by any other means), the result identifying them as a straight emission standards which are normalized by
being that these sources are likely to best performer (or were projected to do hazardous waste thermal input apply
emit less SVM and LVM over time. 70 so under the same reasonable only to SVM and LVM contributed by
FR at 59441. assumptions), and often exceeded their the hazardous waste. MACT standards
Data confirm that lowest emitters in earlier performance by wide margins must address all HAP emitted by a
single tests (i.e., performers identified as (failing routinely, for example, to source, not just some portion of the
best under the Straight Emissions achieve their own performance test HAP. Brick MACT, 479 F.3d at 882–83
methodology) can and do emit more results adjusted upward to account for (raw material input contributions to
SVM and LVM over time than the run-to-run variability, the Straight HAP emissions must be addressed by
sources EPA identified as best Emissions approach floor level (which MACT floor). Although most SVM and
performers using the SRE Feed also accounts for run-to-run variability), LVM emitted by these sources comes
methodology. See TSD Vol. III sections and the design level of the SRE Feed from the hazardous waste,11 hazardous
17.2 and 17.3.1 and 17.3.2. Looking at floor level). See TSD Vol. III sections waste is not the sole input of these
all the data in the record where there 17.2 and 17.3.1 and 17.3.2. metals. However, all SVM and LVM
were multiple test results (i.e., tests EPA’s approach is consistent with the emissions from these sources is
conducted at different times) from statute and with applicable caselaw. controlled by virtue of the PM standard.
sources with the lowest SVM or LVM EPA may consider variability in In addition, although the SVM and LVM
emissions in single tests, EPA found assessing sources’ performance, and it floor standards for cement kilns and
that a) three of four of these sources did so here for the evident reason that lightweight aggregate kilns are
emitted more SVM or LVM in historical variability is an aspect of a source’s normalized by hazardous waste thermal
tests than allowed under the Straight performance. CKRC, 255 F.3d at 865–66; input, EPA also capped these standards
Emissions floor (i.e., average emissions Mossville, 370 F.3d at 1242. Here, short- by the interim standards for SVM and
(not considering run-to-run variability) term and long-term variability (i.e., run- LVM, which are standards that control
of SVM or LVM were higher than the to-run and test-to-test) in SVM and LVM all SVM and LVM emissions emitted
average of the best performers using the performance demonstrably exists. The from the combustor, not just emissions
Straight Emissions methodology (which SRE Feed methodology accounts for of SVM and LVM from hazardous
considers run-to-run variability)) (id. both types of variability. The Straight waste.12 Moreover, there is strong direct
Table 17–1); 9 (b) 5 of 15 of these Emissions methodology demonstrably correlation between the control of total
sources were projected to emit more does not. The Straight Emissions PM and control of metal HAP (including
SVM or LVM than allowed under the methodology thus not only consistently SVM and LVM), so that emission limits
SRE Feed floor using the reasonable underestimates sources’ performance, reflecting best PM control will also
assumption that these sources fed the but identifies as best performers those similarly control the total SVM and
same amount of LVM and SVM in which may emit more SVM and LVM LVM. Sierra Club v. EPA (‘‘Primary
hazardous waste as they did in the over time. For these reasons we believe Copper MACT’’), 353 F.3d 976, 984–85
performance test identifying them as a the record of this rulemaking (D.C. Cir. 2004) (PM proper surrogate for
demonstrates that the SRE Feed HAP metals ‘‘even in light of the
best performer (lowest emitter) under
methodology better accounts for potential variability of impurities in
the straight emission approach, but had
variability, and hence performance, than copper ore’’). Furthermore, as a cross-
the system removal efficiency
does the Straight Emissions approach check, EPA determined that total SVM
demonstrated in their other tests. Id. at
(even with consideration of run-to-run and LVM emissions from the sources
Tables 17–2 and 17–3; 10 and (c) 8 of 13
variability), and consequently, the SRE EPA identified as the PM best
straight emission best performers would
Feed methodology more accurately performers from these source categories
exceed the SRE Feed floor if their
identifies the best performing sources are generally comparable to (and often
system removal efficiency from all tests
and their level of performance. lower than) total SVM and LVM
(i.e., whether the system removal It is also no answer to say that the emissions from the sources identified as
efficiency was higher or lower than that Straight Emissions best performing best performers under EPA’s SRE Feed
demonstrated in the single performance sources could simply retrofit their methodology.13 Thus, on the facts here,
test identifying it as a best performer devices to achieve over time what they the thermally normalized floors for SVM
under the straight emissions were able to achieve in a single and LVM (i.e., the SVM and LVM
methodology) were pooled and applied performance test. Section 112(d)(3) standards for cement kilns, lightweight
to the hazardous waste federate for LVM requires EPA to determine the best aggregate kilns, and the higher heating
or SVM used in the single performance performers and their level of value hazardous wastes subcategory of
test identifying it as a best performer performance based on sources as they liquid fuel boilers), in combination with
under the straight emissions now exist, not how they might be the PM standards, provide control of
methodology. Id. at 17.3.2 and Tables retrofitted. Requiring even the pool of
17–6 and 7. In addition, most of the best performers (i.e., those whose 11 See Source Data for Hazardous Waste
straight emissions best performers performance was measured at below the Combustors, Source Category Summary Sheets, at
average of the best performers) to retrofit http://www.epa.gov/epaoswer/hazwaste/combust/
9 It should be noted that source 3016 was feeding finalmact/source.htm.
more LVM in this test than in its most recent
to meet a floor level is a de facto beyond 12 See 70 FR at 59457–458, § 63.1220(a)(3)(ii),
mstockstill on PROD1PC66 with PROPOSALS

performance test, although the source was operating the floor standard and therefore (a)(4)(ii), (b)(3)(ii), and (b)(4)(ii), and
within its permit limits, and so far as can be impermissible unless costs and other § 63.1221(a)(3)(ii), (a)(4)(ii), (b)(3)(ii), and (b)(4)(ii).
determined was also otherwise properly designed factors under section (d)(2) factors are 13 See note from Bob Holloway, USEPA, to Docket
and operated in this test. considered. 70 FR at 59445. Moreover, ID No. EPA–HQ–OAR–2004–0022 entitled ‘‘SVM/
10 EPA also showed that these sources were LVM Emissions from PM Best Performers Are
operating properly in the tests where they removed
a source so retrofitted would not be an Generally Comparable to SVM/LVM Emissions from
SVM and LVM less efficiently. TSD Vol. III at 17– existing source as required by section SVM/LVM Best Performers,’’ dated August 23,
14 to 15 and Tables 17–4 and 5. 112(d)(3), but rather some hypothetical 2007.

VerDate Aug<31>2005 17:46 Sep 26, 2007 Jkt 211001 PO 00000 Frm 00022 Fmt 4702 Sfmt 4702 E:\FR\FM\27SEP1.SGM 27SEP1
54882 Federal Register / Vol. 72, No. 187 / Thursday, September 27, 2007 / Proposed Rules

SVM and LVM reflecting the average performance, for SVM and LVM burning higher heating value hazardous
SVM and LVM emissions of the best emissions. wastes and those burning lower heating
performing sources. Second, as a legal matter, section value hazardous wastes. The alternative
EPA further justified its use of the 112(d)(3) does not specifically address to the PM standard allows sources to
the question of whether ‘‘best comply with standards limiting
SRE Feed methodology on two
performing’’ sources are those with the emissions of all SVM and LVM metals,
additional bases, both of which are
lowest net emissions, or those which including the five nonenumerated metal
severable from the analysis just
control HAP emissions the most HAP not covered by the standards for
presented. First, EPA appropriately
efficiently. 70 FR at 59443. EPA posited SVM and LVM, in lieu of complying
utilized the SRE Feed methodology
the example of whether a source with the PM standard. Under these
because the Straight Emissions emitting 100 units of HAP and feeding alternatives, the numerical emission
approach would force some best- 100 units of the HAP must be limits for SVM and LVM HAP are
controlled commercial hazardous waste considered better performing than a identical to the promulgated standards.
treatment units to stop burning source emitting 101 units of the HAP However, for SVM, the alternative
hazardous waste (or to burn less waste), but feeding 10,000 units. Id. Indeed, standard applies not only to the
even though hazardous waste must be floors for new sources are to be based combined emissions of lead and
treated before it can be land disposed on the performance of the ‘‘best cadmium, but also includes selenium, a
under sections 3004(d), (e), (g), and (m) controlled’’ similar source. Section semivolatile nonenumerated metal HAP;
of RCRA and combustion is the only 112(d)(3). In the example just given, a for LVM, the standard applies to the
means of successfully treating the source with control efficiency of 99.9 combined emissions of arsenic,
hazardous waste. 70 FR at 59442; TSD per cent can naturally be viewed as beryllium, chromium, antimony, cobalt,
Vol. III section 17.4. EPA noted further better controlled than one with 0 per manganese, and nickel, the latter four
that the Clean Air Act requires that EPA cent control efficiency. EPA’s decision being low volatile nonenumerated metal
take into account RCRA requirements to incorporate control efficiency (i.e., HAP.
when issuing MACT standards for system removal efficiency) into the SRE As noted above, some SVM and LVM
hazardous waste combustion units.14 Feed methodology as one of the two standards are normalized by hazardous
CAA section 112(n)(7). Although a factors used to identify best performing/ waste thermal input and apply only to
severable part of EPA’s rationale, 70 FR best controlled sources reasonably SVM and LVM contributed by the
at 59447/3, EPA continues to believe reflects that the statute allows hazardous waste. For these standards,
that use of the Straight Emissions performance to be evaluated in terms of SVM and LVM emissions from
methodology is unreasonable here control efficiency. See further nonhazardous waste inputs is controlled
because it could have significant discussion of this issue in the analysis by the PM standard. However, if a
adverse cross-media environmental of the total chlorine emission standard source were to elect to comply with the
impacts by reducing the amount of for hydrochloric acid production alternative to the PM standard, then the
needed, and statutorily mandated furnaces. nonhazardous waste inputs would not
hazardous waste treatment capacity. See EPA does, however, realize that be controlled because, under the
id. at 59442 (‘‘EPA doubts that a certain parts of the justification for the alternative, the source would not be
standard which precludes effective SVM and LVM standards in the final required to comply with a PM standard.
treatment mandated by a sister rule may not be consistent with Brick In such instances, the alternative to the
environmental statute must be viewed MACT, and EPA is no longer relying on PM standard would not address all HAP
as a type of best performance under them. These relate principally to how emitted by a source. This does not
section 112(d)’’). EPA’s concern here is MACT standards reflect HAP metal appear to be consistent with the holding
not that certain sources are unable to inputs from variable raw materials. The of Brick MACT that the standard must
achieve a floor standard. See Brick specific alterations EPA is apply to all HAP emitted. 479 F.3d at
MACT, 479 F.3d at 881–82. Rather, the contemplating (generally excising 882–83. Of the source categories for
concern is the adverse cross-media existing language) are found in red line/ which EPA promulgated alternatives to
environmental impact resulting from strike out versions of the Preamble, the PM standard, the higher heating
undermining ‘‘the heart of RCRA’s Technical Support Documents, and value hazardous wastes subcategory for
hazardous waste management program’’, Response to Comment Document which liquid fuel boilers is the only category
the restrictions on land disposal of EPA has placed in the docket for this for which SVM and LVM standards
untreated hazardous waste. Chemical rule. normalized by hazardous waste thermal
Waste Management v. EPA, 976 F.2d 2, 2. Alternatives to the Particulate Matter input were established. Therefore, EPA
23 (D.C. Cir. 1992). Section 112(n)(7) of Standard for Incinerators, Liquid Fuel believes (subject to comment) that it
the Clean Air Act requires EPA to Boilers, and Solid Fuel Boilers must reassess the alternative to the PM
consider RCRA standards when standard for this subcategory (and
adopting section 112(d) standards for EPA promulgated alternatives to the intends to seek remand of this
RCRA sources, and EPA’s consideration PM standard for incinerators, liquid fuel standard). See § 63.1217(e)(2)(ii) and
of the issue here reinforces the boilers, and solid fuel boilers.15 In the (e)(3)(ii).
conclusion that the SRE Feed case of liquid fuel boilers, separate
alternatives to the PM standard were 3. Alternative Mercury, Semivolatile
methodology is reasonable, and the
finalized for each subcategory: those Metals, Low Volatile Metals, and Total
proper means here of assessing which
Chlorine Standards for Cement Kilns
mstockstill on PROD1PC66 with PROPOSALS

sources are best, and their level of 15 For incinerators, the alternative to the PM and Lightweight Aggregate Kilns
standard are promulgated §§ 63.1206(b)(14) and
14 EPA investigated the possibility of 63.1219(e). For the higher and lower heating value EPA promulgated provisions that
subcategorizing by commercial/non-commercial hazardous wastes subcategories for the liquid fuel allow cement kilns and lightweight
sources but found this undesirable because it would boiler category, the alternatives are promulgated aggregate kilns to petition the
lead to anomalously high floors for some under § 63.1217(e)(2) and (e)(3). The alternative to
subcategories due to sparse available data. 70 FR at the PM standard is under § 63.1216(e) for solid fuel
Administrator for alternative mercury,
59442 and n. 78. boilers. semivolatile metals, low volatile metals,

VerDate Aug<31>2005 17:46 Sep 26, 2007 Jkt 211001 PO 00000 Frm 00023 Fmt 4702 Sfmt 4702 E:\FR\FM\27SEP1.SGM 27SEP1
Federal Register / Vol. 72, No. 187 / Thursday, September 27, 2007 / Proposed Rules 54883

and total chlorine standards.16 64 FR at C. Standards for Total Chlorine it in a subsequent rulemaking. EPA
52962–967 and 70 FR at 59503–504. EPA established standards for total notes further that the health-based
Under these provisions, the alternative chlorine (TCl, which controls emissions compliance alternatives for total
standard was not prescribed, and could of both hydrochloric acid and chlorine) chlorine under § 63.1215 would not be
take the form of an operating for all of the source categories. For all affected by this reexamination and thus
requirement, such as a hazardous waste would provide an alternative means of
of the source categories except HCl
feedrate limitation of metals and demonstrating compliance.
production furnaces, EPA established
chlorine or an emission limitation, floors using the SRE Feed methodology 3. Lightweight Aggregate Kilns
subject to approval by the described in the previous section. For Choice of a floor methodology for TCl
Administrator. The rule discusses two HCl production furnaces, EPA selected is essentially academic for existing
sets of circumstances under which a sources with the best removal efficiency lightweight aggregate kilns, since both
source could petition for such an as the best performers. EPA believes that the SRE/Feed and Straight Thermal
alternative standard. One reason is that most of these standards are consistent Emission (and Straight Mass Emission)
the source cannot achieve the standard with the statute and applicable caselaw, methodologies yield floor levels higher
due to contributions of metals and although certain of the standards than the interim standard for these
chlorine HAP in the raw materials. The probably are not. devices, in which case the floor level is
second reason is limited to mercury, 1. Incinerators capped by the level of the interim
and applies in situations where a source standard. 70 FR at 59457; see TSD Vol.
cannot comply with the mercury For hazardous waste incinerators, all III appendices C, D, and E for data and
standard when mercury is not present in of the best performers using the SRE calculations. The reason for this
the raw materials at detectable levels Feed methodology were also the lowest seeming anomaly in all the
(e.g., the mercury emission standard emitters using the Straight Emissions methodologies is that EPA has little data
could be exceeded by a source if it methodology. Thus, choice of floor from this source category (and there are
assumed mercury is present in the raw methodology is not at issue here. only a few sources to begin with), so
materials at the detection limit). These However, EPA found that the analytic that differences in individual
circumstances appear to be method used to gather these data is performance runs are magnified when
biased below 20 ppmv. 70 FR at 59427– the standard is calculated. In addition,
inappropriate bases for an alternative
428. EPA’s determination of how to all of the data in the record came from
standard after Brick MACT.
estimate these best performers’ level of tests conducted before EPA adopted the
Accordingly, EPA currently intends to
performance is explained in detail in 71 interim standards. This is especially
seek a remand of these alternative FR at 52628–30 (Sept. 6, 2006). As there
metals and total chlorine standards and relevant for this standard because the
stated, this determination is consistent interim standard is a beyond-the-floor
remove these provisions in a subsequent with Brick MACT and all other
rulemaking. standard. See generally TSD Vol. III
applicable statutory and caselaw. chapter 19. The interim standard thus
4. Alternative Mercury Standards for 2. Cement Kilns remains the best measure of evaluating
Cement Kilns and Lightweight best performing sources.
Aggregate Kilns Under the Interim EPA used the SRE Feed methodology However, for new sources, EPA noted
Standards to establish floors for new and existing only that the new source floor
sources, but believed that the data did calculated using the SRE Feed
EPA promulgated an alternative to the not fully reflect variability that best methodology would be less stringent
interim standards for mercury for performing kilns experience due to than the interim standard but did not
cement and lightweight aggregate kilns fluctuating alkalinity levels within the closely examine whether the
in 2002. Section 63.1206(b)(15) and 67 kiln. Rather, the TCl emissions data methodology clearly identified the best
FR 6792 (February 13, 2002). Under this reflect the alkalinity of the limestone controlled source. TSD Vol. III section
alternative, sources are allowed to raw material used at the time of 12.6.3. EPA therefore intends to
comply with a hazardous waste performance tests. 70 FR at 59469–70, reexamine this standard in a subsequent
maximum theoretical emissions TSD Vol. III section 13.7.1. To account rulemaking, subject to consideration of
concentration of mercury.17 This for this variability, EPA assumed a 90 comment (and to seek remand of the
alternative mercury standard does not per cent system removal efficiency for standard).
address all mercury emitted by a source, all cement kiln sources. The best
performing sources then effectively 4. Liquid Fuel Boilers
and, therefore, is not permissible in
become the lowest chlorine feeders. a. Higher Heating Value Hazardous
light of the holding of Brick MACT that
Although this assumed system removal Wastes Subcategory. EPA believes
the standard must apply to all HAP
efficiency has some factual basis, see (subject to comment) that it must
emitted. 479 F.3d at 882–83.
Table 1 at 70 FR 59470 showing that the reassess this standard (for both new and
Accordingly, EPA currently intends to
median of the best performing sources existing sources) since the standard
seek a remand of these alternative (Ash Grove) demonstrated removal applies only to TCl attributable to
standard provisions and remove them in efficiencies ranging from 85.1 to 98.8%, hazardous waste inputs, and currently
a subsequent rulemaking. the standard reflects concerns relating to intends to seek remand of the standard.
raw material variability, and also may See § 63.1217(a)(6)(ii). This is not
16 The alternative standard provisions are
reflect a level that is achievable (albeit permissible in light of the holding of
promulgated under § 63.1206(b)(9) for lightweight
mstockstill on PROD1PC66 with PROPOSALS

aggregate kilns and § 63.1206(b)(10) for cement by best performers) rather than actually Brick MACT that the standard must
kilns. achieved. Neither of these rationales is apply to all HAP emitted,
17 Maximum theoretical emissions concentration permissible after Brick MACT, 479 F.3d notwithstanding variable HAP levels in
(MTEC) is a term to compare metals (and chlorine) at 880–81, 882–83. Accordingly, subject raw materials. 479 F.3d at 881–82.
feedrates across sources of different sizes. MTEC is
defined as the metals (or chlorine) feedrate divided
to consideration of comments on this b. Lower Heating Value Hazardous
by the gas flow rate and is expressed in units of ug/ issue, EPA currently intends to seek a Wastes Subcategory. The SRE Feed and
dscm. remand on this standard and reexamine Straight Emissions methodologies give

VerDate Aug<31>2005 17:46 Sep 26, 2007 Jkt 211001 PO 00000 Frm 00024 Fmt 4702 Sfmt 4702 E:\FR\FM\27SEP1.SGM 27SEP1
54884 Federal Register / Vol. 72, No. 187 / Thursday, September 27, 2007 / Proposed Rules

the same floor value for this express performance solely in terms of The opinion does hold, however, that
subcategory, and the standard applies to numerical limits, rather than different HAP levels in raw materials
all TCl emissions from the boiler, not performance efficiency, it said so could not justify a conclusion that floor
just those attributable to hazardous explicitly. See CAA section 129(a)(4). standards were unachievable, so that
waste. See § 63.1217(a)(6)(i). The issue The policy reason for EPA’s emissions attributable to raw material
is how to account for analytical bias at interpretation here is that a standard HAP had to be accounted for in the
levels below 20 ppmv, and EPA’s limiting volumetric TCl emissions standard. 479 F.3d at 882–883. The TCl
resolution of the issue is explained at 71 means that less product is produced, standard at issue here accounts for
FR at 52628–630. EPA does not believe since these sources recover hydrogen emissions from all HAP inputs, 70 FR at
this approach raises issues under the chloride to produce hydrochloric acid. 59450, and so does not present this
statue, or under Brick MACT or other TSD Vol. III at 15–6; 70 FR at 59450. deficiency. Nor are the floor standards
applicable caselaw. EPA does not believe that the MACT designed to be achieved by all sources
floor provisions should compel an with a specific emission control
5. Solid Fuel Boilers
otherwise best performing source to technology. 479 F.3d at 880–81. The
The SRE Feed and Straight Emission limit the amount of product it produces. removal efficiency standard is not based
methodologies give the same floor level See 2 Legislative History at 3352 (House on performance of any particular
for both existing and new solid fuel Report) (‘‘MACT is not intended to technology, and simply is the averaged
boilers, so the issue of appropriate floor * * * drive sources to the brink of (or single best) efficiencies of the best
methodology is academic. TSD Vol. III shutdown’’). performing sources (after accounting for
at App. E and C. run-to-run variability).
Moreover, all that is at issue here is
6. Hydrochloric Acid Production how to express the performance of EPA, however, does not (subject to
Furnaces sources ranked as best performing under comment) believe that the alternative
both EPA’s methodology and under the standard of 150 ppmv by volume for
The TCl standard for this source existing sources (section 63.1218 (a) (6)
category controls TCl emissions and Straight Emissions methodology. This is
(i)) should be retained and EPA
also serves as a surrogate for all metal because, with one exception, the best
currently intends to seek remand of this
HAP. TSD Vol. III sections 15.2 and performing sources are the same under
alternative standard. The standard
15.3. EPA selected as best performers EPA’s methodology as those identified
appears inconsistent with the SRE
sources with the best TCl system as best performing under the Straight
MACT standard, since it allows sources
removal efficiency (or, for new sources, Emissions methodology. TSD Vol. III
to operate with less efficient system
the single source with the best TCl App. C at E–HCLPF–CL and App. E at
removals.
system removal efficiency). The SO–HCLPF–CL. The one exception is EPA also recognizes that certain parts
standard is then expressed as a required where EPA chose a parallel test of the rationale for the standard,
degree of control: 99.923 percent for condition which exhibits more generally related to whether standards
existing sources (the average efficiency variability to characterize the source’s are to reflect varying raw material HAP
of the five best controlled sources), performance (source 855 condition 11 inputs, do not appear to be consistent
99.987 percent for new sources (the rather than condition 13), and with Brick MACT. EPA is making
control efficiency of the single best consequently resulted in this source not appropriate revisions to the key record
controlled source). Id. section 15.3. being selected as a best performer. documents, which are available in red
EPA continues to believe that this Given this documented variability, this line strike out versions in the
standard is consistent with the statute is a reasonable choice. Thus, EPA is administrative record.
and applicable caselaw. First, the selecting as best performers those with
statutory language requiring floors to be the lowest measured emissions of D. Standards for Dioxins/Furans
based on ‘‘best controlled’’ (new) /‘‘best chlorine, but chose to express their Polychlorinated dioxins and furans
performing’’ (existing) does not specify performance in terms of system removal (D/F, or ‘dioxins’) are typically not
whether ‘‘best’’ is to be measured on efficiency to avoid impacts on amount present in any of the inputs to
grounds of control efficiency or of product these best performing sources hazardous waste combustion devices.
emission level. See Sierra Club v. EPA, produce. EPA continues to regard this Rather, they are formed post-
167 F.3d 658, 661 (‘‘ ‘average emissions choice as reasonable. combustion (often from some type of
limitation achieved by the best EPA has carefully reexamined this chlorinated precursor, which precursor
performing 12 percent of units’ * * * standard in light of Brick MACT. The is itself typically a product of
on its own says nothing about how the opinion does not address the issue incomplete combustion). 70 FR at
performance of the best units is to be directly, since no standard there was 59461. As combustion efficiency
calculated’’). The requirement that the determined or expressed in terms of increases, complex organic molecules
new source floor reflect ‘‘emission control efficiency. Moreover, as noted which can be D/F precursors are
control’’ achieved in practice reinforces above, unlike section 129, section 112 oxidized to form carbon dioxide or
that the standard can be determined and contains no directive to express carbon monoxide, helping to minimize
expressed in terms of control efficiency. standards as numerical limits (see D/F formation and emission. Id.
Existing floors determined and section 129(a)(4)), further supporting Different levels of chlorine in waste or
expressed in terms of control efficiency EPA’s view that it could reasonably other inputs do not appreciably
are likewise consistent with the choose to express this standard in per influence D/F emission rates. TSD Vol.
requirement that the floor for existing cent reduction terms. See also section IV 18 section 3.3 (documenting that D/F
mstockstill on PROD1PC66 with PROPOSALS

sources reflect ‘‘average emission 112(i)(5)(A), which allows sources that formation and emission is ordinarily not
limitation achieved’’, since ‘‘emission achieve early reductions based on dependent on feed levels of chlorinated
limitation’’ includes standards which measured rates of removal efficiency a
18 USEPA, ‘‘Technical Support Document for
limit the ‘‘rate’’ of emissions on a reprieve from MACT, a provision
HWC MACT Standards, Volume IV: Compliance
continuous basis—exactly what the reasonably read to allow section 112(d) with the HWC MACT Standards’’ (TSD Vol. IV),
standards do here. CAA section 302(k). performance to be expressed in terms of September 2005. See docket item EPA–HQ–OAR–
Moreover, where Congress wanted to rate of removal efficiency. 2004–0022–0435.

VerDate Aug<31>2005 17:46 Sep 26, 2007 Jkt 211001 PO 00000 Frm 00025 Fmt 4702 Sfmt 4702 E:\FR\FM\27SEP1.SGM 27SEP1
Federal Register / Vol. 72, No. 187 / Thursday, September 27, 2007 / Proposed Rules 54885

materials); TSD Vol. III at 10–6. Nor device is generally the dominant factor lowest emitters in single tests, but these
does burning hazardous waste generally affecting D/F emissions),22 there is no sources had strikingly varied results in
have an appreciable impact on CDD generic, dominant factor affecting other tests, with one ‘best’ performer
formation and emissions, so that it is emissions. In these instances, EPA (source 3016) having emissions over
technically appropriate in some consequently selected as best 1000 times greater than its previous test,
instances to consider D/F emission performers those sources which best and well in excess of the floor level
levels from sources which do not burn minimized the formation of dioxin established by EPA. TSD Vol. III at 10–
hazardous waste in evaluating emission precursors by maintaining the most 6.24 Under these circumstances, EPA
potential from those that do. TSD Vol. efficient combustion conditions, as was unable to conclude that single test
III at 11–4 and n. 72. measured by carbon monoxide (CO) or results adequately represented the
Precise formation and control total hydrocarbon emissions (HC), as sources’ performance over time (i.e.,
mechanisms of D/Fs are thought to be well as by destruction/removal of their long term variability). TSD Vol. III
fairly well understood for systems with hardest-to-burn hazardous waste at 10–6 (lowest emitters in single tests
dry air pollution control devices (or constituents at an efficiency of 99.99 would prove unable to duplicate their
extensive ductwork containing percent. The floor standards for these performance in other tests due to their
particulates on surfaces, such as for sources consequently is either meeting a variability). Without a means to assess
certain lightweight aggregate kilns). For CO standard of 100 ppmv or an HC long-term performance, EPA used the
these systems, D/Fs are formed on standard of 10 ppmv, plus interim standard as the measure of best
particles entrained in the control device demonstrating a destruction/removal performers’ performance over time. Id.
by surface-catalyzed reactions where efficiency (DRE) of 99.99 percent on the EPA continues to believe that this is a
entrained particulate matter provides hardest-to-combust hazardous reasonable estimate of best performance,
the reaction surfaces.19 D/F formation constituents present in the hazardous and that the standards are consistent
can increase exponentially as gas waste. In instances where the interim with the statute and applicable caselaw.
temperatures increase from 400 °F to standard applied to such sources, EPA
750 °F.20 Formation mechanisms, or used that standard as the measure of 2. Cement Kilns
their degree, are less well understood best performers’ good combustion The calculated floor for existing
for systems with wet air pollution instead of quantified CO/HC and cement kilns using the straight
control or no air pollution control destruction/removal efficiency. emissions approach was slightly higher
systems, making it less certain how Our assessment of these standards, (less stringent) than the low end of the
much D/F these sources may emit over subject to comment, is: interim standard (0.28 as opposed to
time. TSD Vol. III pp. 10–5 to 6. 0.20 ng TEQ/dscm). However, available
EPA used the Straight Emissions 1. Incinerators
historical D/F emissions data for cement
methodology rather than the SRE Feed a. Dry Air Pollution Control Device
kiln best performers (other test
methodology as the starting point for Subcategory.23 EPA used the Straight
conditions conducted at different times
calculating floors for D/F because Emissions approach to establish floor
from cement kiln sources identified as
dioxins/furans do not come from inputs levels for existing and new sources for
best performing, which test conditions
(but rather are formed post-combustion), this subcategory. The existing source
reflect temperature optimization) show
so that it is not possible to calculate floor, calculated in this manner, was
that these sources performance
system removal efficiencies (which is slightly higher than the interim
considering run-to-run variability
calculated from inputs and outputs). standard, so the floor is capped at the
exceeded both the floor level calculated
However, for a number of the source level of the interim standard. TSD Vol.
using the Straight Emissions
categories where best performers do not III p. 10–4. The standard for new
methodology and the interim
have dry air pollution control devices, sources is based on the performance of
standard.25 In light of this documented
EPA’s professional judgment was that the single lowest emitting source. Id. at
variability, EPA considered the interim
this methodology did not give an 10–11. EPA believes this standard to be
standard the more stringent and
accurate assessment of the best consistent with the statute and all
consequently used the interim standard
performing sources’ performance over applicable caselaw.
b. Incinerators with Wet Air Pollution (0.20 ng TEQ/dscm or 0.40 ng TEQ/
time (i.e., the best performers’
Control Systems or No Air Pollution dscm and a temperature of 400 °F or less
variability). This is because there are
Control Systems. For both new and at the inlet of the dry air pollution
myriad factors that can affect D/F
existing sources, EPA selected the control device) as the floor. The
emissions for these sources 21 and,
calculated floor for new cement kilns
unlike sources equipped with a dry interim standard as the floor standard.
emission control device where gas Id. at 10–6 and 10–11. EPA considered using the straight emissions approach
temperature at the inlet to the control basing the floor on the performance of was slightly higher (less stringent) than
one part of the interim standard for new
19 USEPA, ‘‘Draft Technical Support Document 22 For sources with dry emission control devices, cement kilns (0.21 ng TEQ/dscm as
for HWC MACT Standards, Volume IV: Compliance D/F emissions during the compliance tests EPA opposed to 0.20 ng TEQ/dscm), and in
with the HWC MACT Standards,’’ March 2004, used to characterize emissions would generally be addition, the lowest emitter in a single
Section 3.0. at the upper end of the range of normal operations. test condition (source 323B3) exhibited
20 To be clear, the dry air pollution control device Because an operating limit is established on gas
does not control D/F emission (except insofar as temperature at the inlet to the control device based
enormous variability in other
some of the formed dioxins/furans adsorb to on levels achieved during the compliance test,
particulate which is collected). Rather, the inlet to operators had the incentive to maximize gas 24 See also Note from Bob Holloway, USEPA, to

these devices serves as an agent for the actual temperatures while still complying with the D/F Docket ID No. EPA–HQ–OAR–2004–0022 entitled
mstockstill on PROD1PC66 with PROPOSALS

formation of the chemical, to the availability of a emission standard under part 266, subpart H ‘‘Incinerators: Comparison of D/F Emissions
surface catalyzed reaction which occurs under these (§ 266.104(e)). Variability for Best Performers and Other Sources
conditions. 23 EPA explained a number of times that it did not with Wet or No APCD,’’ dated April 5, 2007.
21 Factors that can affect D/F emissions from subcategorize incinerators by control device. 25 See data for test conditions 228C4, 403C4, and

sources with a wet control device or no control Rather, the presence or absence of a dry air 404C3 in Note from Frank Behan, USEPA, to Docket
device include: Soot buildup on boiler tubes and pollution device relates to differences in dioxin ID No. EPA–HQ–OAR–2004–0022 entitled
presence of metals in the feed that can catalyze formation mechanisms and consequent dioxin ‘‘Comparative D/F Data for the Cement Kiln Best
D/F formation reactions. 70 FR at 59502. emission levels. See e.g. 70 FR at 59467. Performers,’’ dated August 23, 2007.

VerDate Aug<31>2005 17:46 Sep 26, 2007 Jkt 211001 PO 00000 Frm 00026 Fmt 4702 Sfmt 4702 E:\FR\FM\27SEP1.SGM 27SEP1
54886 Federal Register / Vol. 72, No. 187 / Thursday, September 27, 2007 / Proposed Rules

performance tests (see test condition EPA also recognizes that not all of the invariably based on performance of
323C1; the other lowest emitters rationale adopted for these standards is combustion technology.) EPA adopted
likewise showed significant variability consistent with Brick MACT, standards quantifying good combustion
in other tests (id.)) so EPA adopted the particularly discussions relating to conditions for non-dioxin organic HAP
level of the interim standard as the whether sources other than those in the emitted by liquid fuel boilers, solid fuel
MACT floor for new sources. TSD Vol. best performing half of the MACT pool boilers, and hydrochloric acid
III p. 11–7. EPA believes that these of best performers could replicate best production furnaces. The floor
standards are consistent with the statute performers’ level of performance. EPA standards for these sources is either
and applicable caselaw. has made appropriate edits to the key meeting a CO standard of 100 ppmv or
support documents which are available an HC standard of 10 ppmv, plus
3. Lightweight Aggregate Kilns for comment in red line strikeout form demonstrating a destruction/removal
The calculated floors for existing and in the administrative record. efficiency (DRE) of 99.99 percent on the
new lightweight aggregate kilns using b. Sources with Wet or Without Air hardest-to-combust hazardous
the Straight Emissions approach were Pollution Control Equipment. EPA has constituents present in the hazardous
higher (less stringent) than the interim decided (subject to comment) not to waste. In the event a source chooses to
standard, so EPA adopted the level of defend most of the dioxin standards for comply with the 100 ppmv CO
the interim standard as the MACT floor sources with wet air pollution control standard, it must also demonstrate that
for both existing and new sources. TSD equipment or without air pollution it is achieving 10 ppmv HC standard in
Vol. III pp. 12–4 and 12–6. EPA control equipment.27 These include the a single performance test, and establish
continues to believe that this approach standards for liquid fuel boilers with continuously monitored parameters
uses the best measure of evaluating the wet or no air pollution control systems reflecting the conditions of that
best sources and their level of and standards for hydrochloric acid performance test (including operating
performance, and that these standards production furnaces. EPA continues to temperature, maximum feed rates,
are consistent with the statute and adhere to its analysis that these sources minimum combustion zone residence
applicable caselaw. experience enormous operating time, and operating requirements on the
variability based on dioxin formation hazardous waste firing system that
4. Liquid Fuel Boilers and control mechanisms which are optimize liquid waste atomization
uncertain and presently not efficiency). Sections 63.1216(a)(5),
a. Sources with Dry Air Pollution
quantifiable. However, based on the 63.1217(a)(5), and 63.1218(a)(5).
Control Devices. EPA used the Straight
discussion at 70 FR 59202/2, EPA does The basis for these standards is that
Emissions approach to establish a floor
not believe that it is certain that the good combustion, as measured by 100
for existing liquid fuel boilers equipped
promulgated standard based on ppmv CO or 10 ppmv HC, plus meeting
with a dry air pollution control device,
quantified good combustion addresses 99.99 percent DRE, is the best measure
which yielded an extremely high
all of the potential formation and of the performance over time of best
standard of 3.3 ng TEQ/dscm.26 TSD performers. However, in contrast to
control mechanisms for dioxins as
Vol. III p. 13–7. The floor standard also dioxin, EPA has more knowledge of
required. See Brick MACT, 479 F.3d at
includes an alternative based on formation mechanisms and means of
882–83; CKRC, 255 F.3d at 862–63.
meeting temperature control of 400° F at control over time. Non-dioxin organics
Moreover, the cited preamble discussion
the inlet to the dry air pollution control (of which there are over 100 on the list
suggests that additional dioxin
device. Id. EPA also adopted a beyond- of HAP) can be present in hazardous
formation and control mechanisms can
the-floor standard for these sources waste (or other inputs) or can be formed
be quantified directionally, if not with
which is (necessarily) more stringent as products of incomplete combustion.
exactitude. This again may not be
than the level of the floor. Id. at 13–8. Organics are destroyed when wastes are
consistent with Brick MACT, 479 F.3d at
This beyond the floor standard would combusted, and best performers are
883 (lack of data resulting in inability to
be ascertained identically whether or those which destroy organics through
quantify variability related to non-
not the existing source floor included the most efficient combustion. 70 FR at
technology factors does not by itself
the temperature control alternative. EPA 59463; see also Horsehead Resource
justify by itself a less stringent floor
believes that this standard is consistent Development v. Browner, 16 F.3d 1246,
standard). EPA intends to seek a remand
with section 112 (d) (2) of the statute, 1265 (D.C. Cir. 1994) (‘‘A kiln’s utility
(subject to consideration of public
and that the floor is also consistent with as a means of destroying hazardous
comment) and to investigate these
the statute, but not of direct relevance wastes turns on its ability to fully
issues further in subsequent rulemaking.
given that the actual standard is beyond- destroy them. In practice, destruction of
the-floor. E. Non-Dioxin Organic HAP hazardous wastes in the fuel is a
For new sources, EPA adopted a floor Hazardous wastes contain non-dioxin function of the combustion efficiency of
standard of the lowest emitters’ organic HAP which are destroyed by the kiln: Under poor conditions of
performance, or meeting temperature effective combustion. Treatment of efficiency, the principal organic
control of 400° F or less at the inlet to hazardous waste by destruction of hazardous constituents * * * of the
the dry air pollution control device. organics is indeed the chief reason that toxic organic compounds contained in
Subject to comment, EPA does not there is a hazardous waste combustion the hazardous waste fuel will be only
believe that this standard accounts for industry. See 40 CFR 268.42. (RCRA partially broken down, thereby
all the factors that could influence treatment standards for organic increasing the production of [products
dioxin emissions from new sources, see hazardous wastes, reflecting application of incomplete combustion]’’).
mstockstill on PROD1PC66 with PROPOSALS

Brick MACT, 479 F.3d at 881–82, and of Best Demonstrated Available Furthermore, 100 ppmv CO or 10
therefore intends to seek a remand of Technology (see Hazardous Waste ppmv HC are long-recognized levels
the standard and further examine it in Treatment Council v. EPA, 886 F.2d representing good combustion
a subsequent rulemaking. 355, 363–64 (D.C. Cir. 1989)), are conditions. 70 FR 59463–464
(explaining further that lower levels are
26 The basis for subcategorizing in this way is the 27 For the same reasons, we will not defend the unlikely to be associated with good
same as for incinerators. dioxin standards for solid fuel boilers. combustion and so no longer serve as a

VerDate Aug<31>2005 17:46 Sep 26, 2007 Jkt 211001 PO 00000 Frm 00027 Fmt 4702 Sfmt 4702 E:\FR\FM\27SEP1.SGM 27SEP1
Federal Register / Vol. 72, No. 187 / Thursday, September 27, 2007 / Proposed Rules 54887

measure of organic destruction). EPA emission levels. 370 F.3d at 1242. The molecule, and levels lower than 100
adopted these levels here as the best difference here is that EPA is using a ppmv are no longer reliably associated
measure of the sources’ long-term quantified surrogate to do so, but EPA with levels of organic HAP). Finally, the
performance (and reiterates that finding believes this is a difference without factors mentioned by the commenter
here). Id. and TSD Vol. III at 13–35, 14– legal significance given the which can influence organic destruction
26, and 15–9. In addition to good reasonableness of the surrogate on the are in fact encompassed within the CO
combustion being the long-recognized facts presented here. Indeed, EPA and HC standards because, as EPA
metric for organic destruction and selected here an existing regulatory explained, sources must conduct a
performance, EPA lacked any data on standard as a measure of best performance test for HC and DRE, and
individual organic HAP emissions from performers’ performance over time continuously monitored parameters,
these devices, so had no choice but to (RCRA standards for CO/HC and DRE), including minimum operating
use some type of surrogate to evaluate just as in Mossville EPA selected the temperature, maximum feed rates,
sources’ performance. existing uniform vinyl chloride minimum combustion zone residence
EPA views these standards as regulatory standard as that measure. 370 time, and operating requirements on
consistent with the statute and F.3d at 1240.28 hazardous waste firing systems (i.e., all
applicable caselaw. Regarding use of the One commenter maintained that CO/ of the factors mentioned by the
quantified good combustion surrogate, HC standards should be numerically commenter), are established based on
the D.C. Circuit has held repeatedly that lower to reflect lowest CO/HC the conditions established in that
EPA may select a surrogate for control emissions, and further maintained that performance test. 70 FR at 59464/1. EPA
of HAP in adopting section 112(d) CO and HC are not the sole measures of consequently views all of these
standards. See, e.g. National Lime Ass’n organic combustion efficiency, which, standards as consistent with Brick
v. EPA, 233 F.3d 625, 639 (D.C. Cir. as EPA noted, can be influenced by such MACT and the statute.
2000); Sierra Club v. EPA (‘‘Primary factors as inadequate time, temperature Edited versions of the key support
Copper MACT’’), 353 F.3d 976, 984–85 and turbulence within individual documents for this standard, edited to
(D.C. Cir. 2004). EPA has shown here a combustion zones, and, the argument reflect changes necessary in light of
valid basis for choosing good goes, are therefore improper or Brick MACT, are available in red line
combustion as a surrogate: There is a inadequate surrogates. 70 FR at 59463/ strike out format for comment in the
strong correlation between optimized 2; cf. National Lime, 233 F.3d at 639. administrative record.
combustion conditions and minimized EPA addressed these issues in the
organic emissions in that oxidation of record. 70 FR at 59462–63. With respect F. Mercury
heavier, more complex organic to the level for CO/HC, extremely low 1. Incinerators
molecules will be maximized when CO floors are unlikely to be met at all
combustion conditions are optimized, times by best performers due to all the For existing incinerators, both the
thus minimizing emission of organics. potential minor sources of variability. SRE/Feed methodologies and straight
70 FR at 59463; see also id. at 59461– So the 100 ppmv standard—which must emissions methodologies (even without
62; see also National Lime, 233 F.3d at be met continuously (and is measured calculation of run-to-run variability)
639 (upholding EPA’s selection of PM as by a continuous emission monitor), is produced floors which were higher than
a surrogate for HAP metals where EPA the best measure of best performers’ the interim standard. TSD Vol. III
demonstrated a correlation between variability and hence performance over appendices C and E, tables E–INC–HG
removal of PM and metal HAP, and time. TSD Vol. III at 13–35, 14–26 and CT and SF–INC–HG, respectively. EPA’s
further holding both that EPA need not 15–9 (best sources’ inability to duplicate decision to use the interim standard as
quantify the precise amount of metal a lower level of performance at all times the level of the floor consequently does
HAP removed, and that the amount of for these reasons); see also Mossville, not raise issues vis-á-vis Brick MACT.
HAP metal removed may vary); Primary 370 F.3d at 1242 (if floor standard must See also Mossville, 370 F.3d at 1241–42
Copper MACT, 353 F.3d at 984. EPA has be met continuously, then the best (selection of regulatory standard as floor
further demonstrated the reasonableness performers’ maximum variability must is a legitimate means of assessing best
of 100 ppmv CO or 10 ppmv HC as be reflected in that standard). Of equal performers’ variability when these
measures of good combustion. importance, lower levels of either CO or performers demonstrably emit at a level
National Lime further indicates (in HC are no longer likely to be associated close to that regulatory level).
dicta) that choice of a surrogate may not with increased organic destruction For new incinerators, EPA selected
be valid if emissions of the HAP could efficiency. 70 FR at 59462–64 (CO itself the emission level of the lowest emitting
increase by some mechanism for which is a conservative indicator of source since the same source was the
the surrogate fails to account, combustion efficiency because it is a lowest emitter under both the SRE
specifically noting that if HAP metal thermally stable, refractory compound methodology and the Straight Emissions
feedrates decreased and PM emissions which is the final stage of the methodology, TSD Vol. III appendices C
did not decrease proportionately, PM combustion process of an organic and E, tables E–INC–HG CT and SF–
might not be a valid surrogate. 233 F.3d INC–HG, respectively, again raising no
at 639. This discussion has no direct 28 Brick MACT holds that EPA may not select
issues vis-á-vis Brick MACT.
factual applicability here since organic floor standards to assure that all sources in the
category will be able to meet the standards. 479 2. Cement Kilns
emissions are not input dependent. See F.3d at 880–81. EPA did not do so here. The CO/
also Primary Copper MACT, 353 F.3d at HC and DRE standards are EPA’s best estimate of For both new and existing cement
985 (rejecting argument that input best performers’ performance over time. As in kilns, the mercury floor standard
mstockstill on PROD1PC66 with PROPOSALS

variability made PM an arbitrary Mossville, EPA selected an existing regulatory limit appears inconsistent with the Brick
not because all sources were (by definition) meeting
surrogate for metals). The situation here that regulatory limit, but because no other means MACT opinion and the statute because
is similar to that in Mossville, where the of accurately assessing variability were available. it is based in whole or in substantial
court held that EPA could account for 370 F.3d at 1240. Moreover, sources will establish part on emissions attributable
best performers’ performance over time, parametric monitoring conditions, which will vary exclusively to hazardous waste control.
by source, as part of the process of meeting the 10
and could estimate performance over ppmv HC standard, so the standards in fact are not The standard thus does not result in
time by some means other than uniform across the source category. control of all mercury which could be

VerDate Aug<31>2005 17:46 Sep 26, 2007 Jkt 211001 PO 00000 Frm 00028 Fmt 4702 Sfmt 4702 E:\FR\FM\27SEP1.SGM 27SEP1
54888 Federal Register / Vol. 72, No. 187 / Thursday, September 27, 2007 / Proposed Rules

emitted by cement kilns (mercury in currently expects to seek remand of this baseball pitcher is the better performer,
raw materials being the notable standard. Brick MACT, 479 F.3d at 882– the one who has given up 6 earned runs
example), and so appears to require 83. or the one who has given up 20. Unless
revision. 479 F.3d at 882–83. Subject to b. Lower Heating Value Hazardous and until the figure is normalized over
comment, it is thus EPA’s intent to Wastes Subcategory. The mercury floor 9 innings pitched, the question is
amend this standard and to seek remand standard for this subcategory for both meaningless. 70 FR at 59451 n. 101.
of the standard. existing and new sources is based on the EPA sees nothing in the statute which
3. Lightweight Aggregate Kilns Straight Emissions methodology. TSD precludes use of normalization in
Vol. III at 13–16 and 13–18; see also 69 determining who best performers are for
The methodology for developing floor FR 21286–87 (because so many of the purposes of MACT floor determinations.
standards for mercury for lightweight data measurements were non-detects, Section 112(d)(3) does not specifically
aggregate kilns is essentially a Straight EPA was unable to calculate removal address the issue (the terms ‘‘best
Emissions approach for mercury efficiencies, and so did not use the SRE performing’’ and ‘‘best controlled’’ being
contributed by hazardous waste.29 The Feed methodology). The standard also amenable to an interpretation allowing
floor calculated thereby produced applies to all mercury emitted by the comparisons of normalized emissions to
existing and new source floors higher source, not just that attributable to assess which source is ‘‘better’’ or
than the interim standard of 120 µg/ hazardous waste. Section ‘‘best’’). The issue of normalization was
dscm total mercury emissions (110 µg/ 63.1217(a)(2)(i). EPA does not believe not presented in Brick MACT, so that
dscm for new sources), which EPA that this approach creates any issues EPA likewise does not view the opinion
therefore adopted as the floor standard. vis-á-vis Brick MACT. as precluding the approach.
TSD Vol. III at 12–8 to 9, 12–12 and
section 7.2.3.5. EPA continues to believe 5. Solid Fuel Boilers H. Potential Implications to the
that the interim standard remains the EPA used the SRE Feed methodology Compliance Date Provisions if
best measure of best sources’ to identify best sources and their level Standards Are Remanded to EPA
performance given the available data. of performance for both new and
However, the interim standard contains The compliance date of the final rule
existing solid fuel boilers. TSD Vol. III is October 14, 2008. As discussed above,
a compliance option based solely upon at 14–7, 14–9. The floor standards are
mercury emissions attributable to we are contemplating requesting the
identical to those using the Straight Court to remand several standards so
hazardous waste. Section Emissions methodology because the best
63.1206(b)(15). Subject to comment, this that we can reexamine them in a future
performing sources (and single best rulemaking, a process that likely would
alternative compliance mechanism performing source) are the same under
appears to be inconsistent with Brick be concluded well after the compliance
either methodology. TSD Vol. III at App. date of the rule. It is not our intent to
MACT since it would not control all C (E–SFB–HG–CT) and E (SF–SFB–HG).
mercury emitted by the kiln. 479 F.3d ask the Court to vacate any standards,
EPA does not believe that these including those standards that may have
at 882–83; see also section III.B.3 above. standards pose issues vis-á-vis Brick
Subject to consideration of public to be revised in a future rulemaking. As
MACT. a result, sources would need to comply
comment, EPA intends to seek a remand
of this alternative standard and to G. Normalization with the standards promulgated in
consider this issue further in subsequent October 2005 according to the
A number of the standards are compliance date provisions codified
rulemaking. ‘‘normalized,’’ that is expressed as a under § 63.1206(a). See NRDC v. EPA,
4. Liquid Fuel Boilers given amount of pollutant per amount of 489 F.3d 1364, 1373–74 (D.C. Cir. 2007).
some production related parameter such
a. Higher Heating Value Hazardous List of Subjects in 40 CFR Part 63
as air flow or thermal inputs. See
Wastes Subcategory. The mercury floor
generally 70 FR at 59451. Most Environmental protection, Air
standard for this subcategory for both
technology-based standards are pollution control, Hazardous
existing and new sources accounts only
expressed in terms of some type of substances, Reporting and
for mercury emissions from hazardous
normalizing parameter in order to allow recordkeeping requirements.
waste. TSD Vol. III pp. 13–14 and 13–
meaningful comparison between
16. These standards thus appear to Dated: September 21, 2007.
performance of different sources.
require revision, and EPA accordingly Stephen L. Johnson,
Weyerhaeuser v. Costle, 590 F.2d 1011,
1059 (D.C. Cir. 1978). As EPA pointed Administrator.
29 EPA used the Straight Emissions approach here

for data-specific reasons explained at section 7.5.3.2 out, comparing unnormalized [FR Doc. E7–19097 Filed 9–26–07; 8:45 am]
of Volume III of the TSD. performance is like asking which BILLING CODE 6560–50–P
mstockstill on PROD1PC66 with PROPOSALS

VerDate Aug<31>2005 17:46 Sep 26, 2007 Jkt 211001 PO 00000 Frm 00029 Fmt 4702 Sfmt 4702 E:\FR\FM\27SEP1.SGM 27SEP1