Standards of Performance for New Stationary Sources and Emissions Guidelines for Existing Sources: Hospital/Medical/Infectious Waste Incinerators

Federal RegisterOct 6, 2009

Ask Donna

What actually matters in this document.

Text

ENVIRONMENTAL PROTECTION AGENCY

40 CFR Part 60

[EPA-HQ-OAR-2006-0534; FRL-8959-9]

RIN 2060-A004

Standards of Performance for New Stationary Sources and Emissions Guidelines for Existing Sources: Hospital/Medical/Infectious Waste Incinerators

AGENCY:

Environmental Protection Agency (EPA).

ACTION:

Final rule.

SUMMARY:

On September 15, 1997, EPA adopted new source performance standards (NSPS) and emissions guidelines (EG) for hospital/medical/infectious waste incinerators (HMIWI). The NSPS and EG were established under Sections 111 and 129 of the Clean Air Act (CAA or Act). In a response to a suit filed by the Sierra Club and the Natural Resources Defense Council (Sierra Club), the U.S. Court of Appeals for the District of Columbia Circuit (the Court) remanded the HMIWI regulations on March 2, 1999, for further explanation of EPA's reasoning in determining the minimum regulatory “floors” for new and existing HMIWI. The HMIWI regulations were not vacated and were fully implemented by September 2002. On February 6, 2007, we published our proposed response to the Court's remand. Following recent court decisions and receipt of public comments regarding the proposal, we re-assessed our response to the remand, and on December 1, 2008, we published another proposed response and solicited public comments. This action promulgates our response to the Court's remand and also satisfies the CAA Section 129(a)(5) requirement to conduct a review of the standards every 5 years.

DATES:

The amendments to 40 CFR 60.32e, 60.33e, 60.36e, 60.37e, 60.38e, 60.39e, Table 1A and 1B to subpart Ce, and Tables 2A and 2B to subpart Ce are effective as of December 7, 2009. The amendments to 40 CFR 60.17, 60.50c, 60.51c, 60.52c, 60.55c, 60.56c, 60.57c, 60.58c, and Tables 1A and 1B to subpart Ec are effective as of April 6, 2010. The incorporation by reference of certain publications listed in the regulations is approved by the Director of the Federal Register as of April 6, 2010.

ADDRESSES:

EPA has established a docket for this action under Docket ID No. EPA-HQ-OAR-2006-0534 and Legacy Docket ID No. A-91-61. All documents in the docket are listed on the

http://www.regulations.gov

Web site. Although listed in the index, some information is not publicly available,

e.g.,

confidential business information or other information whose disclosure is restricted by statute. Certain other material, such as copyrighted material, is not placed on the Internet and will be publicly available only in hard copy form. Publicly available docket materials are available either electronically through

http://www.regulations.gov

or in hard copy at the EPA Docket Center, EPA West, Room 3334, 1301 Constitution Ave., NW., Washington, DC. The Public Reading Room is open from 8:30 a.m. to 4:30 p.m., Monday through Friday, excluding legal holidays. The telephone number for the Public Reading Room is (202) 566-1744, and the telephone number for the EPA Docket Center is (202) 566-1742.

FOR FURTHER INFORMATION CONTACT:

Mr. Ketan D. Patel, Natural Resources and Commerce Group, Sector Policies and Programs Division (E143-03), Environmental Protection Agency, Research Triangle Park, North Carolina 27711; telephone number: (919) 541-9736; fax number: (919) 541-3470; e-mail address:

patel.ketan@epa.gov.

SUPPLEMENTARY INFORMATION:

Organization of This Document.

The following outline is provided to aid in locating information in this preamble.

I. General Information

A. Does the Final Action Apply to Me?

B. Where Can I Get a Copy of This Document?

C. Judicial Review

II. Background

III. Summary of the Final Rule and Changes Since Proposal

A. Remand Response

B. Clean Air Act Section 129(a)(5) 5-Year Review Response

C. Other Amendments

D. Implementation Schedule for Existing Hospital/Medical/Infectious Waste Incinerators

E. Changes to the Applicability Date of the 1997 New Source Performance Standards

F. Startup, Shutdown, and Malfunction Exemption

IV. Summary of Major Comments and Responses

A. Applicability

B. Subcategorization

C. MACT Floor Approach

D. Emissions Limits

E. Monitoring

F. Emissions Testing

G. Alternatives to On-Site Incineration

H. Medical Waste Segregation

I. Startup, Shutdown, and Malfunction

J. Economic Impacts

V. Impacts of the Final Action for Existing Units

A. What Are the Primary Air Impacts?

B. What Are the Water and Solid Waste Impacts?

C. What Are the Energy Impacts?

D. What Are the Secondary Air Impacts?

E. What Are the Cost and Economic Impacts?

VI. Impacts of the Final Action for New Units

A. What Are the Primary Air Impacts?

B. What Are the Water and Solid Waste Impacts?

C. What Are the Energy Impacts?

D. What Are the Secondary Air Impacts?

E. What Are the Cost and Economic Impacts?

VII. Relationship of the Final Action to Section 112(c)(6) of the Clean Air Act

VIII. Statutory and Executive Order Reviews

A. Executive Order 12866: Regulatory Planning and Review

B. Paperwork Reduction Act

C. Regulatory Flexibility Act

D. Unfunded Mandates Reform Act

E. Executive Order 13132: Federalism

F. Executive Order 13175: Consultation and Coordination With Indian Tribal Governments

G. Executive Order 13045: Protection of Children From Environmental Health and Safety Risks

H. Executive Order 13211: Actions That Significantly Affect Energy Supply, Distribution or Use

I. National Technology Transfer Advancement Act

J. Executive Order 12898: Federal Actions To Address Environmental Justice in Minority Populations and Low-Income Populations

K. Congressional Review Act

I. General Information

A. Does the Final Action Apply to Me?

Regulated Entities.

Categories and entities potentially affected by the final action are those which operate hospital/medical/infectious waste incinerators (HMIWI). The new source performance standards (NSPS) and emissions guidelines (EG) for HMIWI affect the following categories of sources:

Category

NAICS Code

Examples of potentially regulated entities

Industry

622110, 622310, 325411, 325412, 562213, 611310

Private hospitals, other health care facilities, commercial research laboratories, commercial waste disposal companies, private universities.

Federal Government

622110, 541710, 928110

Federal hospitals, other health care facilities, public health service, armed services.

State/local/Tribal Government

622110, 562213, 611310

State/local hospitals, other health care facilities, State/local waste disposal services, State universities.

This table is not intended to be exhaustive, but rather provides a guide for readers regarding entities likely to be affected by the final action. To determine whether your facility would be affected by the final action, you should examine the applicability criteria in 40 CFR 60.50c of subpart Ec and 40 CFR 60.32e of subpart Ce. If you have any questions regarding the applicability of the final action to a particular entity, contact the person listed in the preceding

FOR FURTHER INFORMATION CONTACT

section.

B. Where Can I Get a Copy of This Document?

In addition to being available in the docket, an electronic copy of this final action will also be available on the Worldwide Web (WWW) through the Technology Transfer Network (TTN). Following signature, a copy of the final action will be posted on the TTN's policy and guidance page for newly proposed or promulgated rules at the following address:

http://www.epa.gov/ttn/oarpg/.

The TTN provides information and technology exchange in various areas of air pollution control.

C. Judicial Review

Under Section 307(b)(1) of the Clean Air Act (CAA or Act), judicial review of this final rule is available only by filing a petition for review in the U.S. Court of Appeals for the District of Columbia Circuit (the Court) by December 7, 2009. Under Section 307(d)(7)(B) of the CAA, only an objection to this final rule that was raised with reasonable specificity during the period for public comment can be raised during judicial review. CAA Section 307(d)(7)(B) also provides a mechanism for EPA to convene a proceeding for reconsideration, “[i]f the person raising an objection can demonstrate to EPA that it was impracticable to raise such objection within [the period for public comment] or if the grounds for such objection arose after the period for public comment (but within the time specified for judicial review) and if such objection is of central relevance to the outcome of the rule.” Any person seeking to make such a demonstration to us should submit a Petition for Reconsideration to the Office of the Administrator, Environmental Protection Agency, Room 3000, Ariel Rios Building, 1200 Pennsylvania Ave., NW., Washington, DC 20460, with a copy to the person listed in the preceding

FOR FURTHER INFORMATION CONTACT

section, and the Associate General Counsel for the Air and Radiation Law Office, Office of General Counsel (Mail Code 2344A), Environmental Protection Agency, 1200 Pennsylvania Ave., NW., Washington, DC 20004. Moreover, under Section 307(b)(2) of the CAA, the requirements established by this final rule may not be challenged separately in any civil or criminal proceedings brought by EPA to enforce these requirements.

II. Background

Section 129 of the CAA, entitled “Solid Waste Combustion,” requires EPA to develop and adopt new source performance standards (NSPS) and emissions guidelines (EG) for solid waste incineration units pursuant to CAA Sections 111 and 129. Sections 111(b) and 129(a) of the CAA (NSPS program) address emissions from new HMIWI, and CAA Sections 111(d) and 129(b) (EG program) address emissions from existing HMIWI. The NSPS are directly enforceable Federal regulations, and under CAA Section 129(f)(1) become effective 6 months after promulgation. Under CAA Section 129(f)(2), the EG become effective and enforceable as expeditiously as practicable after EPA approves a State plan implementing the EG but no later than 3 years after such approval or 5 years after the date the EG are promulgated, whichever is earlier.

A HMIWI is defined as any device used to burn hospital waste or medical/infectious waste. Hospital waste means discards generated at a hospital, and medical/infectious waste means any waste generated in the diagnosis, treatment, or immunization of human beings or animals, in research pertaining thereto, or in the production or testing of biologicals (

e.g.,

vaccines, cultures, blood or blood products, human pathological waste, sharps). As explained in EPA's regulations, hospital/medical/infectious waste does not include household waste, hazardous waste, or human and animal remains not generated as medical waste. A HMIWI typically is a small, dual-chamber incinerator that burns on average about 800 pounds per hour (lb/hr) of waste. Smaller units burn as little as 15 lb/hr while larger units burn as much as 3,700 lb/hr, on average.

Incineration of hospital/medical/infectious waste causes the release of a wide array of air pollutants, some of which exist in the waste feed material and are released unchanged during combustion, and some of which are generated as a result of the combustion process itself. These pollutants include particulate matter (PM); heavy metals, including lead (Pb), cadmium (Cd), and mercury (Hg); toxic organics, including chlorinated dibenzo-p-dioxins/dibenzofurans (CDD/CDF); carbon monoxide (CO); nitrogen oxides (NO

); and acid gases, including hydrogen chloride (HCl) and sulfur dioxide (SO

). In addition to the use of pollution prevention measures (

i.e.,

waste segregation) and good combustion control practices, HMIWI are typically controlled by wet scrubbers or dry sorbent injection fabric filters (dry scrubbers).

Waste segregation is the separation of certain components of the waste stream in order to reduce the amount of air pollution emissions associated with that waste when incinerated. The separated waste may include paper, cardboard, plastics, glass, batteries, aluminum cans, food waste, or metals. Separation of these types of wastes reduces the amount of chlorine- and metal-containing wastes being incinerated, which results in lower potential emissions of HCl, CDD/CDF, Hg, Cd, and Pb.

Combustion control includes the proper design, construction, operation, and maintenance of HMIWI to destroy or prevent the formation of air pollutants prior to their release to the atmosphere. Test data indicate that as secondary chamber residence time and temperature increase, emissions decrease. Combustion control is most effective in reducing CDD/CDF, PM, and CO emissions. The 2-second combustion level, which includes a minimum secondary chamber temperature of 1800 °F and residence time of 2 seconds, is considered to be the best level of combustion control (

i.e.,

good combustion) that is applied to HMIWI. Wet scrubbers and dry scrubbers provide control of PM, CDD/CDF, HCl, and metals, but do not influence CO or NO

and have little impact on SO

at the low concentrations emitted by HMIWI. (

See

Legacy Docket ID No. A-91-61, item II-A-111; 60 FR 10669, 10671-10677; and 61 FR 31742-31743.)

The CAA sets forth a two-stage approach to regulating emissions from

incinerators. EPA has substantial discretion to distinguish among classes, types and sizes of incinerator units within a category while setting standards. In the first stage of setting standards, CAA Section 129(a)(2) requires EPA to establish technology-based emissions standards that reflect the maximum levels of control EPA determines are achievable for new and existing units, after considering costs, non-air quality health and environmental impacts, and energy requirements associated with the implementation of the standards. Section 129(a)(5) then directs EPA to review those standards and revise them as necessary every 5 years. In the second stage, Section 129(h)(3) requires EPA to determine whether further revisions of the standards are necessary in order to provide an ample margin of safety to protect public health or to prevent (taking into consideration costs, energy, safety and other relevant factors) an adverse environmental effect.

See, e.g.,

NRDC and LEAN

EPA,

529 F.3d 1077, 1079-80 (DC Cir. 2008) (addressing the similarly required two-stage approach under CAA Sections 112(d) and (f), and upholding EPA's implementation of same).

In setting forth the methodology EPA must use to establish the first-stage technology-based NSPS and EG, CAA Section 129(a)(2) provides that standards “applicable to solid waste incineration units promulgated under Section 111 and this Section shall reflect the maximum degree of reduction in emissions of [certain listed air pollutants] that the Administrator, taking into consideration the cost of achieving such emissions reduction, and any non-air quality health and environmental impacts and energy requirements, determines is achievable for new and existing units in each category.” This level of control is referred to as a “maximum achievable control technology,” or MACT, standard.

In promulgating a MACT standard, EPA must first calculate the minimum stringency levels for new and existing solid waste incineration units in a category, generally based on levels of emissions control achieved or required to be achieved by the subject units. The minimum level of stringency is called the MACT “floor,” and CAA Section 129(a)(2) sets forth differing levels of minimum stringency that EPA's standards must achieve, based on whether they regulate new and reconstructed sources, or existing sources. For new and reconstructed sources, CAA Section 129(a)(2) provides that the “degree of reduction in emissions that is deemed achievable [* * *] shall not be less stringent than the emissions control that is achieved in practice by the best controlled similar unit, as determined by the Administrator.” Emissions standards for existing units may be less stringent than standards for new units, but “shall not be less stringent than the average emissions limitation achieved by the best performing 12 percent of units in the category (excluding units which first met lowest achievable emissions rates 18 months before the date such standards are proposed or 30 months before the date such standards are promulgated, whichever is later).”

The MACT floors form the least stringent regulatory option EPA may consider in the determination of MACT standards for a source category. EPA must also determine whether to control emissions “beyond-the-floor,” after considering the costs, non-air quality health and environmental impacts, and energy requirements of such more stringent control. EPA made such MACT floor and beyond-the-floor determinations and on September 15, 1997, adopted NSPS (40 CFR part 60, subpart Ec) and EG (40 CFR part 60, subpart Ce) using this approach for entities which operate HMIWI. The NSPS and EG are designed to reduce air pollution emitted from new and existing HMIWI, including HCl, CO, Pb, Cd, Hg, PM, CDD/CDF (total, or 2,3,7,8-tetrachlorinated dibenzo-p-dioxin toxic equivalent (TEQ)), NO

, SO

, and opacity. The 1997 NSPS apply to HMIWI for which construction began after June 20, 1996, or for which modification began after March 16, 1998. The 1997 NSPS became effective on March 16, 1998, and apply as of that date or at start-up of a HMIWI, whichever is later. The 1997 EG apply to HMIWI for which construction began on or before June 20, 1996, and required compliance by September 2002.

On November 14, 1997, the Sierra Club and the Natural Resources Defense Council (Sierra Club) filed suit in the Court. The Sierra Club claimed that EPA violated CAA Section 129 by setting emissions standards for HMIWI that are less stringent than required by Section 129(a)(2); that EPA violated Section 129 by not including pollution prevention or waste minimization requirements; and that EPA had not adequately considered the non-air quality health and environmental impacts of the standards.

On March 2, 1999, the Court issued its opinion in

Sierra Club

EPA,

167 F.3d 658 (DC Cir. 1999). While the Court rejected the Sierra Club's statutory arguments under CAA Section 129, the Court remanded the rule to EPA for further explanation regarding how EPA derived the MACT floors for new and existing HMIWI. Furthermore, the Court did not vacate the regulations, and the regulations have remained in effect during the remand.

On February 6, 2007, EPA proposed a response to the HMIWI remand. The proposed response was based on a reassessment of information and data that were available at the time of promulgation in 1997, in light of the EPA's understanding of the Court's rulings in the

Sierra Club, National Lime Association (NLA) II,

Cement Kiln Recycling Coalition (CKRC)

and other cases discussed in our 2007 proposal notice. The proposed response would have revised some of the emissions limits in both the NSPS and EG. Relative to the NSPS, the emissions limits for CO, Pb, Cd, Hg, PM, and CDD/CDF would have been revised. Relative to the EG, the emissions limits for HCl, Pb, Cd, and CDD/CDF would have been revised. EPA believed that the revised emissions limits proposed in February 2007 as a result of its response to the remand could be achieved with the same emissions control technology currently used by HMIWI to meet the 1997 rule.

On December 1, 2008, EPA re-proposed its response to the Court's remand. EPA's decision to re-propose was based on a number of factors, including further rulings by the U.S. Court of Appeals that were issued after our 2007 proposal was published. In addition, public comments regarding the 2007 proposal raised issues that, upon further consideration, we believed would best be addressed through a re-proposal. One issue regarded the use of emissions limits included in State regulations and State-issued permits as surrogates for estimated actual emissions limitations achieved. Another issue regarded EPA's previous reliance on control technology performance as the sole indicator of HMIWI performance in making MACT floor determinations, which did not necessarily account for other factors that affect emissions (

e.g.,

waste mix, combustion conditions).

As mentioned above, every 5 years after adopting a MACT standard under Section 129, CAA Section 129(a)(5) requires EPA to review and, if appropriate, revise the incinerator standards. In addition to responding to the Court's remand, today's final action constitutes the first 5-year review of the HMIWI standards.

III. Summary of the Final Rule and Changes Since Proposal

A. Remand Response

Today's final response to the remand revises the December 2008 proposed emissions limits for both the NSPS and EG. The emissions limits are being revised in response to a public comment on the December 2008 re-proposal, which requested that EPA adjust the statistical approach used to account for variability in the data and consider the distribution of the emissions data in determining the MACT floor emissions limits. The revised statistical approach results in generally higher limits compared to the December 2008 re-proposal. (

See

section IV.C.6 of this preamble for further information about this revised approach.) We expect most sources should be able to meet the revised limits using control technology already available to the industry (

e.g.,

wet scrubbers, dry scrubbers, or some combination of these controls). (

See

section IV.C.2 of this preamble for further information.) Similar to the 2008 re-proposal, the emissions limits in today's final action do not include percent reduction alternative standards, as discussed further in section IV.D.4 of this preamble.

Table 1 of this preamble summarizes the NSPS emissions limits being promulgated in this action in response to the Court remand for new HMIWI.

Table 1—Summary of Emissions Limits Promulgated in Response to the Remand for New HMIWI

Pollutant (units)

Unit size

Final remand response limit

HCl (ppmv)

5.1

7.7

CO (ppmv)

1.8

Pb (mg/dscm)

0.00069

0.018

0.31

Cd (mg/dscm)

0.00013

0.0098

0.017

Hg (mg/dscm)

0.0013

0.0035

0.014

PM (gr/dscf)

0.0080

0.0095

0.029

CDD/CDF, total (ng/dscm)

9.3

0.47

CDD/CDF, TEQ (ng/dscm)

0.035

0.014

0.013

(ppmv)

130

M, S

(ppmv)

1.6

M, S

1.4

Opacity (%)

L, M, S

6.0

L = Large (>500 lb/hr of waste); M = Medium (>200 to ≤500 lb/hr of waste); S = Small (≤200 lb/hr of waste).

All emissions limits are reported as corrected to 7 percent oxygen.

Table 2 of this preamble summarizes the emissions limits being promulgated in this action in response to the Court remand for existing HMIWI.

Table 2—Summary of EG Emissions Limits Promulgated in Response to the Remand for Existing HMIWI

Pollutant (units)

Unit size

Final remand response limit

HCl (ppmv)

6.6

7.7

810

CO (ppmv)

5.5

S, SR

Pb (mg/dscm)

0.036

0.018

0.31

0.50

Cd (mg/dscm)

0.0092

0.013

0.017

0.11

Hg (mg/dscm)

0.018

0.025

0.014

0.0051

PM (gr/dscf)

0.011

0.020

0.029

0.038

CDD/CDF, total (ng/dscm)

9.3

0.85

240

CDD/CDF, TEQ (ng/dscm)

0.054

0.020

0.013

5.1

(ppmv)

140

M, S

190

130

(ppmv)

9.0

M, S

4.2

Opacity (%)

L, M, S, SR

6.0

L = Large (>500 lb/hr of waste); M = Medium (>200 to ≤500 lb/hr of waste); S = Small (≤200 lb/hr of waste); SR = Small Rural (Small HMIWI >50 miles from boundary of nearest SMSA, burning <2,000 lb/wk of waste).

All emissions limits are reported as corrected to 7 percent oxygen.

B. Clean Air Act Section 129(a)(5) 5-Year Review Response

We are promulgating our response to the remand in

Sierra Club

such that the revised MACT standards, reflecting floor levels determined by actual emissions data, would be more stringent than what we proposed in 2007 for both the remand response and the 5-year review, with the exceptions noted and discussed in sections IV.A. and IV.B of this preamble. Consequently, we believe that our obligation to conduct a 5-year review based on implementation of the 1997 emissions standards will also be fulfilled through this action's final remand response, even as amended compared to the 2008 re-proposed standards. This is supported by the fact that the revised MACT floor determinations and emissions limits associated with the remand response are based on performance data for the 57 currently operating HMIWI that are subject to the 1997 standards, and by the final rule's accounting for non-technology factors that affect HMIWI emissions performance, which the 2007 proposed remand response and 5-year review did not fully consider. Thus, the final remand response more than addresses the technology review's goals of assessing the performance efficiency of the installed equipment and ensuring that the emissions limits reflect the performance of the technologies required by the MACT standards. In addition, the final remand response addresses whether new technologies and processes and improvements in practices have been demonstrated at sources subject to the emissions limits. Accordingly, the remand response in this final action fulfills EPA's obligations regarding the first 5-year review of the HMIWI standards and, therefore, replaces the 2007 proposal's 5-year review proposed revisions.

C. Other Amendments

This final action puts forward the same changes based on information received during implementation of the HMIWI NSPS and EG that were proposed in 2007 and 2008. The changes proposed in 2007 included provisions allowing existing sources to use previous emissions test results to demonstrate compliance with the revised emissions limits; annual inspections of air pollution control devices (APCD); a one-time visible emissions test of ash handling operations; CO continuous emissions monitoring systems (CEMS) and bag leak detection systems for new sources; and several approved monitoring alternatives. The 2008 proposal included changes regarding requirements for NO

and SO

emissions testing for all HMIWI; performance testing requirements for small rural HMIWI; monitoring requirements for HMIWI that install selective non-catalytic reduction (SNCR) technology to reduce NO

emissions; and procedures for test data submittal. The changes included in this final action include revised provisions regarding waste segregation and removal of exemptions regarding startup, shutdown, and malfunction (SSM). The removal of SSM exemptions is discussed in section III.F of this preamble. The performance testing and monitoring amendments, electronic data submittal provisions, waste segregation amendments, and miscellaneous other amendments are summarized in the following sections.

1. Performance Testing and Monitoring Amendments

The amendments require all HMIWI to demonstrate initial compliance with the revised NO

and SO

emissions limits. The 1997 standards did not require testing and demonstration of compliance with the NO

and SO

emissions limits. In addition to demonstrating initial compliance with the NO

and SO

emissions limits, small rural HMIWI are required to demonstrate initial compliance with the other seven regulated pollutants' emissions limits and the opacity standard. Under the 1997 standards, small rural HMIWI were required to demonstrate only initial compliance with the PM, CO, CDD/CDF, Hg, and opacity standards. Small rural HMIWI also are required to determine compliance with the PM, CO, and HCl emissions limits by conducting an annual performance test. On an annual basis, small rural HMIWI are required by the 1997 standards to demonstrate compliance with the opacity limit. The amendments allow sources to use results of their previous emissions tests to demonstrate initial compliance with the revised emissions limits as long as the sources certify that the previous test results are representative of current operations. Only those sources who could not so certify and/or whose previous emissions tests do not demonstrate compliance with one or more revised emissions limits would be required to conduct another emissions test for those pollutants. (Note that most sources were already required under the 1997 standards to test for HCl, CO, and PM on an annual basis, and those annual tests are still required.)

The amendments require, for existing HMIWI, annual inspections of scrubbers, fabric filters, and other air pollution control devices that may be used to meet the emissions limits. The amendments require a visible emissions test of the ash handling operations using Method 22 in appendix A-7 of this part to be conducted during the next performance test. For new HMIWI, the amendments require CO CEMS; bag leak detection systems for fabric-filter controlled units; annual inspections of scrubbers, fabric filters, and other air pollution control devices that may be used to meet the emissions limits; and Method 22 visible emissions testing of the ash handling operations to be conducted during each compliance test. For existing HMIWI, use of CO CEMS is an approved option, and specific language with requirements for CO CEMS is included in the amendments. For new and existing HMIWI, use of PM, HCl, multi-metals, and Hg CEMS, and integrated sorbent trap Hg monitoring and dioxin monitoring (continuous sampling with periodic sample analysis) also are approved options, and specific language for those options is included in the amendments. HMIWI that install SNCR technology to reduce NO

emissions are required to monitor the reagent (

e.g.,

ammonia or urea) injection rate and secondary chamber temperature.

2. Electronic Data Submittal

The EPA must have performance test data to conduct effective 5-year reviews of CAA Section 129 standards, as well as for many other purposes, including compliance determinations, development of emissions factors, and determining annual emissions rates. In conducting 5-year reviews, EPA has found it burdensome and time-consuming to collect emissions test data because of varied locations for data storage and varied data storage methods. One improvement that has occurred in recent years is the availability of stack test reports in electronic format as a replacement for burdensome paper copies.

In this action, we are taking a step to improve data accessibility. HMIWI have the option of submitting to an EPA electronic database an electronic copy of annual stack test reports. Data entry will be through an electronic emissions test report structure used by the staff as part of the emissions testing project. The electronic reporting tool (ERT) was developed with input from stack testing companies who generally collect and compile performance test data electronically. The ERT is currently available, and access to direct data submittal to EPA's electronic emissions database (WebFIRE) will become available December 31, 2011.

See http://cfpub.epa.gov/oarweb/index.cfm?action=fire.main, http://www.epa.gov/ttn/chief/ert/ert_tool.html.

Please note that the option to submit source test data electronically to EPA will not require any additional performance testing. In addition, when a facility elects to submit performance test data to WebFIRE, there will be no additional requirements for data compilation. Instead, we believe industry will benefit from development of improved emissions factors, fewer follow-up information requests, and better regulation development, as discussed below. The information to be reported is already required in the existing test methods and is necessary to evaluate the conformance to the test method. One major advantage of electing to submit source test data through the ERT is to provide a standardized method to compile and store all the documentation required to be reported by this rule. Another important benefit of submitting these data to EPA at the time the source test is conducted is that it will substantially reduce the effort involved in data collection activities in the future. Specifically, because EPA would already have adequate source category data to conduct residual risk assessments or technology reviews, there would be fewer data collection requests (

e.g.,

CAA Section 114 letters). This results in a reduced burden on both affected facilities (in terms of reduced manpower to respond to data collection requests) and EPA (in terms of preparing and distributing data collection requests). Finally, another benefit of electing to submit these data to WebFIRE electronically is that these data will greatly improve the overall quality of the existing and new emissions factors by supplementing the pool of emissions test data upon which the emissions factor is based and by ensuring that data are more representative of current industry operational procedures. A common complaint we hear from industry and regulators is that emissions factors are outdated or not representative of a particular source category. Receiving most performance tests will ensure that emissions factors are updated and more accurate. In summary, receiving test data already collected for other purposes and using them in the emissions factors development program will save industry, State/local/Tribal agencies, and EPA time and money.

The electronic data base that will be used is EPA's WebFIRE, which is a Web site accessible through EPA's TTN. The WebFIRE Web site was constructed to store emissions test data for use in developing emissions factors. A description of the WebFIRE data base can be found at

http://cfpub.epa.gov/oarweb/index.cfm?action=fire.main.

The ERT will be able to transmit the electronic report through EPA's Central Data Exchange (CDX) network for storage in the WebFIRE data base. Although ERT is not the only electronic interface that can be used to submit source test data to the CDX for entry into WebFIRE, it makes submittal of data very straightforward and easy. A description of the ERT can be found at

http://www.epa.gov/ttn/chief/ert/ert_tool.html.

The ERT can be used to document stack tests data for various pollutants including PM (EPA Method 5 of appendix A-3), SO

(EPA Method 6 or 6C of appendix A-4), NO

(EPA Method 7 or 7E of appendix A-4), CO (EPA Method 10 of appendix A-4), Cd (EPA Method 29 of appendix A-8), Pb (Method 29), Hg (Method 29), and HCl (EPA Method 26A of appendix A-8). Presently, the ERT does not handle dioxin/furan stack test data (EPA

Method 23 of appendix A-7), but the tool is being upgraded to handle dioxin/furan stack test data. The ERT does not currently accept opacity data or CEMS data.

3. Waste Segregation

The amendments revise the waste management plan provisions for new and existing HMIWI. Commenters on the 2008 re-proposal recommended that EPA minimize or eliminate from the HMIWI waste stream any plastic wastes, Hg and other hazardous wastes (

e.g.,

Hg-containing dental waste, Hg-containing devices), pharmaceuticals, and confidential documents and other paper products that could be shredded and recycled. One commenter recommended that EPA take action to regulate emissions of polychlorinated biphenyls (PCBs) and polycyclic organic matter (POM) from HMIWI. To address the various commenters' concerns, the waste management plan provisions in §§ 60.35e and 60.55c are revised to promote the segregation of the aforementioned wastes. (

See

section IV.H of this preamble for further information about the change to waste management plan provisions.)

5. Miscellaneous Other Amendments

The amendments revise the definition of “Minimum secondary chamber temperature” to read “

Minimum secondary chamber temperature

means 90 percent of the highest 3-hour average secondary chamber temperature (taken, at a minimum, once every minute) measured during the most recent performance test demonstrating compliance with the PM, CO, and dioxin/furan emissions limits.”

The amendments add definitions for “Bag leak detection system,” “commercial HMIWI,” and “minimum reagent flow rate.” “

Bag leak detection system”

is defined to mean “an instrument that is capable of monitoring PM loadings in the exhaust of a fabric filter in order to detect bag failures,” and examples of such a system are provided. “

Commercial HMIWI”

is defined to mean “a HMIWI which offers incineration services for hospital/medical/infectious waste generated offsite by firms unrelated to the firm that owns the HMIWI.” “

Minimum reagent flow rate”

is defined to mean “90 percent of the highest 3-hour average reagent flow rate at the inlet to the selective noncatalytic reduction technology (taken, at a minimum, once every minute) measured during the most recent performance test demonstrating compliance with the NO

emissions limit.”

The amendments require HMIWI to submit, along with each test report, a description, including sample calculations, of how operating parameters are established during the initial performance test and, if applicable, re-established during subsequent performance tests.

To provide greater clarity, the amendments also include averaging times and EPA reference test methods in the emissions limit tables for existing and new sources. It should be noted that the averaging times and EPA reference test methods added to the emissions limits tables are not new requirements but simply a restating of requirements presented elsewhere in the HMIWI regulations. Also, the inclusion of these additional table columns should not be interpreted as reopening the 1997 standards themselves.

The amendments also incorporate by reference two alternatives to EPA reference test methods (ASME PTC 19.10-1981 and ASTM D6784-02) to provide HMIWI with greater flexibility in demonstrating compliance. These alternative methods are described in greater detail in section VIII.I of this preamble and were first presented in the preamble to the December 1, 2008 re-proposal.

D. Implementation Schedule for Existing Hospital/Medical/Infectious Waste Incinerators

Under the amendments to the EG, and consistent with CAA Section 129, revised State plans containing the revised existing source emissions limits and other requirements in the amendments will be due within 1 year after promulgation of the amendments. That is, revised State plans have to be submitted to EPA on October 6, 2010.

The amendments to the EG then allow existing HMIWI to demonstrate compliance with the amended standards as expeditiously as practicable after approval of a State plan, but no later than 3 years from the date of such approval or 5 years after promulgation of the revised standards, whichever is earlier. Because many HMIWI will find it necessary to retrofit existing emissions control equipment and/or install additional emissions control equipment in order to meet the revised limits, States may wish to consider providing the maximum compliance period allowed by CAA Section 129(f)(2).

In revising the emissions limits in a State plan, a State has two options. First, it could include both the current and the new emissions limits in its revised State plan, which would allow a phased approach in applying the new limits. That is, the State plan would make it clear that the 1997 emissions limits remain in force and apply until the date the revised existing source emissions limits are effective (as defined in the State plan). States whose existing HMIWI do not find it necessary to improve their performance in order to meet the revised emissions limits may want to consider a second approach, where the State would insert the revised emissions limits in place of the 1997 emissions limits, follow procedures in 40 CFR part 60, subpart B, and submit a revised State plan to EPA for approval. If the revised State plan contains only the revised emissions limits (

i.e.,

the 1997 emissions limits are not retained), then the revised emissions limits must become effective immediately, since the 1997 limits would be removed from the State plan.

EPA will revise the existing Federal plan to incorporate the changes to existing source emissions limits and other requirements that EPA is promulgating. The Federal plan applies to HMIWI in any State without an approved State plan. The amendments to the Federal plan for the EG would require existing HMIWI demonstrate compliance with the amended standards not later than 5 years after today's final rule, as required by CAA Section 129(b)(3).

E. Changes to the Applicability Date of the 1997 New Source Performance Standards

HMIWI are treated differently under the amended standards than they were under the 1997 standards in terms of whether they are “existing” or “new” sources, and there are new dates defining what are “new” sources and imposing compliance deadlines regarding the amended standards. All HMIWI that complied with the NSPS as promulgated in 1997 are “existing” sources under the amended standards and are required to meet the emissions limits under the revised EG or the 1997 NSPS, whichever is more stringent, by the applicable compliance date for the revised EG. (Note that the HCl emissions limit for small HMIWI and the PM emissions limit for medium HMIWI are more stringent under the 1997 NSPS than under the revised EG, and HMIWI that complied with those 1997 NSPS are required to continue to do so.) In the interim, those sources will continue to be subject to the NSPS as promulgated in 1997 until the date for compliance with the revised EG. Units for which construction is commenced after the December 1, 2008 proposal, or for which modification is commenced on or after the date 6 months after today's

promulgation of the amended NSPS, are “new” units subject to more stringent revised NSPS emissions limits.

Thus, under these specific amendments, units that commenced construction after June 20, 1996, and on or before December 1, 2008, or that are modified before the date 6 months after the date of promulgation of the revised final NSPS, continue to be or would become subject to the 40 CFR part 60, subpart Ec NSPS emissions limits that were promulgated in 1997 until the applicable compliance date for the revised EG, at which time those units must comply with the amended “existing” source EG or 1997 NSPS, whichever is more stringent for each pollutant. Similarly, HMIWI that met the 1997 EG must meet the revised EG by the applicable compliance date for the revised EG. HMIWI that commence construction after December 1, 2008 or that are modified 6 months or more after the date of promulgation of the revised NSPS must meet the revised NSPS emissions limits being added to the subpart Ec NSPS within 6 months after the promulgation date of the amendments or upon startup, whichever is later.

This approach is justified because most HMIWI will have to install additional emissions controls to comply with the revised standards. CAA Sections 129(g)(2) and (3) define “new solid waste incineration unit” and “modified solid waste incineration unit” based on whether construction of the new unit commences after the date of proposed standards under Section 129 and on whether modification occurs after the effective date of a Section 129 standard, respectively. While these definitions might be read as referring to the dates EPA first proposes standards for the source category as a whole and on which such standards first become effective for the source category, we are interpreting and applying them in this rulemaking to refer to the proposal and effective dates for standards under this new rulemaking record. The evident intent of the definitions plus the substantive new unit and modified unit provisions is that it is technically more challenging and potentially more costly to retrofit a control system to an existing unit than to incorporate controls when a unit is initially designed.

F. Startup, Shutdown, and Malfunction Exemption

The 1997 standards included provisions in 40 CFR 60.56c and 60.37e that exempted HMIWI from the standards during periods of SSM, provided that no hospital waste or medical/infectious waste is charged to the unit during those SSM periods. Neither our 2007 proposal nor our 2008 re-proposal would have changed these provisions. However, soon after the date of our re-proposal, the U.S. Court of Appeals in

Sierra Club

EPA,

551 F.3d 1019 (DC Cir. 2008), vacated provisions in EPA's CAA Section 112 regulations governing emissions of hazardous air pollutants during SSM periods. Specifically, the Court vacated 40 CFR 63.6(f)(1) and 63.6(h)(1), which, when incorporated into CAA Section 112(d) standards for specific source categories, exempt sources from the requirement to comply with otherwise applicable Section 112(d) standards during periods of SSM. While the Court's vacatur did not have a direct impact on source category-specific SSM exemptions such as those contained in the 1997 HMIWI standards, one commenter on the 2008 re-proposal stressed that the legality of SSM exemptions such as those in the 1997 standards is questionable, and urged EPA to remove the exemptions in the final rule. For the reasons set forth later in this notice responding to comments, today's final rule removes the SSM exemption from the HMIWI standards, such that the emissions limits under these subparts apply at all times.

IV. Summary of Major Comments and Responses

A total of 22 separate sets of public comments were received on the December 1, 2008 re-proposal. (One additional comment, received after the deadline for public comments, was an addendum to an earlier comment.

See http://www.regulations.gov

, docket ID no. EPA-HQ-OAR-2006-0534, for the complete public comments.) The comment period ended on February 17, 2009. In addition to the comment letters, speaker comments from a January 15, 2009, public hearing on the re-proposal were recorded, and a transcript of the hearing was placed in the project docket (document no. EPA-HQ-OAR-2006-0534-0361). The following sections summarize the major public comments received on the re-proposal and present EPA's responses to those comments. The major comment topics are applicability; subcategorization; MACT floor approach; emissions limits; monitoring; emissions testing; alternatives to on-site incineration; medical waste segregation; startup, shutdown, and malfunction; and economic impacts.

A. Applicability

Comment:

While this issue was not raised in our re-proposal, one commenter stated that subpart Ec should be amended to exempt units already complying with subpart AAAA—the NSPS for new small municipal waste combustors (MWCs)—or subpart BBBB—the EG for existing small MWCs—consistent with the exemptions provided to MWCs in the 1997 HMIWI rule.

Response:

We are aware of two HMIWI at one facility that are currently subject to rules for both HMIWI and small MWCs. We have considered the appropriateness of exempting the two units from the HMIWI rule or creating a separate HMIWI subcategory for the units, and have concluded that exemptions and creation of a separate subcategory are not warranted. One issue is the technological feasibility for the facility to meet both the HMIWI and small MWC rules if there is the possibility that the facility would have to implement different control strategies to meet the limits in both rules. (Note that we do not currently have any information to suggest that the facility would find it technically impossible to meet both the revised HMIWI standards and the small MWC standards.) For example, if the HMIWI rule were to include stringent CO limits and the small MWC rule were to include stringent NO

limits, it may be challenging for the facility to meet the limits of both rules simultaneously by controlling secondary chamber temperature; increasing the temperature to reduce CO emissions would invariably increase NO

emissions. However, by choosing to burn both types of waste and operate as both a small MWC and a HMIWI, the facility has the responsibility to meet whatever set of rules that applies based on its operating scenario and could avoid this situation by choosing to burn one type of waste or the other exclusively, or at least reducing the other type of waste to co-fired levels. Also, the facility already employs additional control strategies besides combustion control for reducing NO

emissions (urea injection).

The facility typically burns 50 percent hospital/medical/infectious (HMI) waste and 50 percent municipal waste in its two units. If we were to grant an exemption to the HMIWI rule for this facility due to it being subject to the small MWC rule and the facility were to increase the amount of HMI waste burned to 70 percent and reduce the amount of municipal waste burned to 30 percent, we could create a total compliance loophole for the facility, given that the small MWC rule includes a co-fired exemption for units burning 30 percent or less of municipal waste. This would be an unacceptable outcome.

Another option to address this situation would be to create a hybrid waste subcategory to include the two units, based on the rationale that the units are burning a unique mixture of waste. However, we did not provide an opportunity to comment on such an option in the re-proposal, and have not had the opportunity to develop a record to support such a new approach or its possibly unique regulatory framework. Moreover, it is also not clear that such a hybrid subcategory would fit within the statutory divisions of incinerator categories set forth in Section 129(a)(1) of the CAA. Therefore, we decided not to pursue that option for the final rule.

We believe it is reasonable for the facility to be subject to both the HMIWI and small MWC rules when switching back and forth among the types of waste burned, since this ensures that, when the facility operates as either a HMIWI or small MWC, it is regulated as such and does not avoid compliance obligations that all other incinerators operating continuously as either HMIWI or small MWC must meet. We do not expect that continuing to require the facility to comply with both rules will be overly burdensome. The facility should be able to control to the more stringent of the two rules.

B. Subcategorization

Comment:

Four commenters stated that EPA's rationale for subcategorization does not reflect any analysis of how the proposed subcategories will help assure that what has been “achieved” by better performers in a proposed subcategory results in a standard that is “achievable” by other sources in that subcategory. Two of the commenters argued that, without this assessment, the final subcategory decisions will be arbitrary and may result in standards that are unlawfully stringent. The commenters urged EPA to provide the necessary assessment and rationale for its subcategory proposal. Another commenter further urged EPA to reconsider its decision to retain the categories defined by the 1997 HMIWI rule without defining additional subcategories. The commenter suggested that EPA could keep the relation between “achieved” and “achievable” by grouping existing units based on control technology type and that EPA could address variability by establishing subcategories that take into account non-technology factors that affect emissions, as the commenter claimed is required under Section 112(d)(3).

While the commenter cited to CAA Section 112(d)(3), which does not literally apply to NSPS and EG promulgated under Sections 111 and 129, we assume the commenter was referring to factors relevant to MACT floor analyses in general, including those under Section 129(a)(2).

Three commenters stated that EPA must develop a new subcategory for commercial facilities, based on the claimed significant operational differences between commercial and so-called “captive” units that are attached to HMI waste generators. The commenters defined a captive unit as one that is co-owned and co-operated by the generator of the waste, while a commercial operator is in business to receive wastes from third parties. The commenters stated that commercial HMIWI, unlike operators of captive units, cannot use alternative forms of disposal (

e.g.,

landfills), and claimed that EPA views their only alternative to the standards as closure. According to the commenters, EPA not only has the authority under Section 129(a)(2) to further subcategorize HMIWI, but it is also mandated to do so due to an overly stringent standard that is not “achievable” by commercial units. The commenters claimed that wastes sent to a commercial unit are more heterogeneous than for captive units. They also noted that the handling of medical wastes is subject to numerous Federal and State requirements related to worker and public health and safety, which the commenters claimed makes segregation of wastes hazardous and impractical for operators of commercial facilities. Thus, the commenters argued that waste segregation cannot be a control “achieved in practice” that can be used to determine floors for commercial units.

The same three commenters also argued that EPA provides no rationale for its retention of the small rural class in the re-proposed rule, and that its prior rationale regarding the unavailability of alternative means of medical waste treatment beyond 50 miles from the nearest standard metropolitan statistical area (SMSA) is unsupported. According to the commenters, EPA's proposed retention of the small rural subcategory is arbitrary and capricious.

Another commenter recommended that EPA establish new size classifications, claiming that the distribution of HMIWI no longer matches the three size categories EPA identified in 1995 when the rule was first being developed. The commenter also noted that current standards are based on subcategories defined in terms of feed rates with no corresponding heating value. According to the commenter, a reference waste heating characteristic must be established to adjust or rate incinerators, given that there is currently no consistency or basis for determining equivalent charging rate.

The same commenter further recommended that, based on its facility's unique attributes—extremely large processing capacity, customer generated waste material variability, waste mix, waste-to-energy heat recovery technology, CEMS, 2+ second combustion gas retention time, and high British thermal unit (BTU) waste content—EPA should place its facility in a separate subcategory for extra-large HMIWI. The commenter provided a list of suggested standards for such a subcategory, based on upper confidence limits (UCLs) calculated using EPA's methodology, that indicate 7 of the 11 promulgated standards applicable to it could be tightened. The commenter noted that residual risk analyses conducted under Maryland's stringent air toxics regulations (provided in the commenter's public comments) show that the resulting ambient emissions would meet all applicable requirements.

Response:

Regarding the commenters' argument that EPA must show how the proposed subcategories will result in a standard that is “achievable,” we do not believe that the CAA requires such an analysis. In facing a similar claim, the U.S. Court of Appeals for the DC Circuit recently rejected the argument that a facility's claimed differences between itself and other members of a source category in the plywood and composite wood products (PCWP) MACT rule compels EPA to set a unique standard that is achievable for that source. In

NRDC

EPA,

489 F.3d 1364 (DC Cir. 2007), Louisiana-Pacific Corp. (L-P) objected to EPA's refusal to establish a separate subcategory for its wet/wet press process apart from the subcategory of all other press processes, claiming that, at L-P's plant, EPA's identified MACT floor control technology was not feasible and that L-P would experience greater costs in complying with the MACT floor compared to other press operators.

Id.,

at 1375-76. The Court denied L-P's claims, explaining that “cost is not a factor that EPA may permissibly consider in setting a MACT floor. [* * *] To the extent that L-P maintains that it cannot comply with the MACT floor based on complete enclosure and capture of emissions because it cannot enclose its presses, L-P also relies on an incorrect premise that the MACT level of emissions reduction is invalid if it is based on control technology that a source cannot install. The 2004 rule does not require a source to use any particular method to

achieve compliance: If L-P cannot use enclosure and capture, it may utilize other compliance techniques. Hence, L-P fails to show that EPA was arbitrary or capricious in refusing to create a subcategory for it.”

Id.

at 1376. The option provided by one commenter to subcategorize based on control technology type is inappropriate, as it would essentially endorse the type of unique treatment that L-P demanded in the PCWP rule and that the Court rejected. Moreover, we are unaware of any situations in the HMIWI industry where one type of control would be technically applicable, but not another, such that subcategorizing based on the ability to use certain controls would be justified.

We evaluated three different subcategory options to try and address the concerns stated by the commenters. The three options included: (1) Option 1—no change to existing size categories; (2) Option 2—creating a commercial subcategory (as suggested by three commenters) and redistributing the size categories for the captive HMIWI (as suggested by another commenter); and (3) Option 3—redistributing the existing size categories to more evenly distribute the number of HMIWI (also suggested by the other commenter).

Under Option 1, the size distributions would remain the same—large (>500 lb/hr of waste), medium (>200 to ≤500 lb/hr of waste), and small (≤200 lb/hr of waste), with the latter category divided into small rural and non-rural subcategories based on distance from the nearest SMSA.

Under Option 2, commercial HMIWI would be categorized separately from captive HMIWI, and the captive HMIWI further subcategorized as follows—large (>1,000 lb/hr of waste), medium (>500 to ≤1,000 lb/hr of waste), and small (≤500 of waste), with no further subcategorization of the latter category.

Under Option 3, the sizes would be redistributed as follows—large (>1,500 lb/hr of waste), medium (>500 to ≤1,500 lb/hr of waste), and small (≤500 lb/hr of waste), with the latter category divided into small rural and non-rural subcategories as under Option 1.

We conducted MACT floor analyses on all three options, using the following methodology, which is described in more detail later in this notice—(1) Ranking the emissions data from lowest to highest for each pollutant; (2) determining the units in the MACT floor for each pollutant; (3) determining the distribution of test run data for the MACT floor units; and (4) calculating a 99 percent UCL for each pollutant based on that distribution, using Student's t-test statistics. We developed floor-based emissions limits based on these UCL values, rounding up to two significant figures. We compared the emissions limits to average emissions estimates for each HMIWI and determined whether the HMIWI would meet the limits. We estimated the number of HMIWI expected to meet

at least

nine limits, eight limits, seven limits,

etc.

under each option. Based on our analysis, Options 1, 2, and 3 resulted in similar numbers of HMIWI meeting the limits. (For more detailed results,

see

2009 memorandum entitled “Revised MACT Floors, Data Variability Analysis, and Emission Limits for Existing and New HMIWI,” which is included in the docket for today's rulemaking.)

However, since we did not propose any subcategorization option other than the small, medium and large size subcategories identified in the 1997 rule, and did not provide an opportunity to comment on this issue in the re-proposal, we have concluded that it would not be appropriate at this time to promulgate emissions limits based on Options 2 and 3. Moreover, we do not see a compelling need to make the adjustments of Options 2 or 3, given that similar numbers of HMIWI meet the limits under all three options. Simply re-adjusting the size thresholds to reflect an even distribution of units post-MACT compliance among the subcategories is not necessarily reasonable, whereas the size thresholds from the 1997 rule continue to correspond to the basic distinctions between the subcategories of units as currently operated. Therefore, we selected Option 1 (no change to existing size subcategories) as the best subcategory option on which to base the emissions limits for promulgation.

Two other subcategory options were considered and rejected without further analysis. The two options include (1) an extra-large subcategory for one HMIWI facility (as suggested by one commenter), and (2) a mixed waste subcategory for another HMIWI facility (an outgrowth of a comment by another commenter, as discussed in the previous section). In addition to the fact that we did not provide opportunity to comment on this issue, we found no basis for creating a new subcategory for this particular rulemaking to fit a single facility.

We disagree with the argument by three commenters that EPA's retention of the small rural subcategory is unsupported by any rationale. As we explained in the September 15, 1997 notice of final rulemaking (62 FR 48370), alternative means of medical waste treatment may not be available to some facilities that operate small HMIWI in rural or remote locations. Facilities that operate small HMIWI in remote locations could be faced with unique adverse impacts if required to meet the more stringent emissions limits associated with small non-rural HMIWI. Therefore, we continue to support subcategorizing facilities based on the location of the facility and the amount of waste burned, as allowed under Section 129(a)(2). The only remaining small rural units are in Alaska and Hawaii, and the options are very limited for alternative medical waste treatment in those States. There are a very limited number of landfills and MWC facilities in those States, and there are no commercial HMIWI. (The basis for this information is a 2004 Chartwell Information document entitled

Directory & Atlas of Solid Waste Facilities.

)

C. MACT Floor Approach

1. MACT-on-MACT

Comment:

Several commenters argued that EPA's recalculation of the 1997 MACT floors using post-MACT compliance data results in so-called “MACT-on-MACT” standards that cannot be achieved and are contrary to the CAA and the intent of Congress. Three of the commenters stated that the CAA provides for a one-time setting of the MACT floor based on what sources achieved at the time of the initial promulgation, not at the time of subsequent revisions. According to those three commenters, the proposed standards would force the HMIWI industry to shut down and prevent installation of new HMIWI, without any consideration of the costs of additional reductions or whether the emissions posed any risks to human health and the environment. The commenters urged EPA to use the population of pre-1997 HMIWI and their emissions data to establish the revised MACT floors. One commenter stated that new data should only be used for those units that have the same control equipment in place as when EPA undertook the original rulemaking.

Three of the commenters objected to EPA's arguments for using the post-MACT compliance data, namely that EPA is no longer confident in the regulatory limits used in 1997 (based on a comparison of the regulatory limits and emissions test data in the 1997 record) and that the EPA questions their use as surrogates because they do not account for non-technology factors (based on waste segregation data EPA received after the 2007 proposal). Specifically, the three commenters stated that EPA provides no justification for its change in using the post-MACT

compliance data, noting that the Court, in

Sierra Club

EPA,

167 F.3d 658 (DC Cir. 1999), upheld EPA's data-gathering for the 1997 rule, and did not dispute that EPA could make estimates based on the lack of data. The three commenters further stated that EPA provides no support for reassessing its determination in 1997 that emissions controls significantly impact emissions, which the commenters indicated is a finding that EPA continues to assert and that is supported by the data.

Regarding EPA's claim that it reset the floors in response to the remand of the regulation in

Sierra Club

EPA,

167 F.3d 658 (DC Cir. 1999), the same three commenters argued that the Court's remand was limited and did not vacate the 1997 floors. According to the commenters, EPA cites no legal support that subsequent case law invalidates a promulgated regulation not at issue in that case. The commenters stated that, in the past, EPA has declined to account for changes in law after its decision to impose new regulatory obligations, based in part on the general presumption against law having a retroactive effect. According to the commenters, this approach is supported by case law, which holds that agencies are required to apply the law at the time the decision is made.

Aaacon Auto Transport

ICC,

792 F.2d 1156, 1161 (DC Cir. 1986). The commenters also noted that the 2002 data used to set the proposed standards would not have been available had the EPA responded to the 1999 remand in a more timely manner.

The three commenters also argued that new public comments raising issues with the 1997 floors are out of time and insufficient to require EPA to go beyond the Court's remand order. The commenters pointed out that Section 307(b) of the CAA requires any challenges to regulations to be filed within 60 days, which has been held up in the relevant case law. According to the commenters, any required revisions to address the Court's limited remand does not justify reopening the time period for judicial challenge of the floors. The commenters also argued that another exception to the 60-day jurisdictional bar, that there was a substantive violation of the statute, does not apply since the Court did not find the 1997 floors in conflict with the statute.

Response:

First, we disagree with the commenters' assertion that we are employing a MACT-on-MACT approach to set limits that are not achievable by HMIWI. The purpose of this action is not to force units who have complied with a lawfully adopted MACT standard to have to subsequently comply with another round of updated MACT standards, but to respond to the Court's ruling that questioned the basis for the 1997 MACT standards and revise them such that they are clearly compliant with the Court's several pronouncements of how MACT should be set in the first instance. Moreover, the actual emissions data upon which the revised standards rely comes directly from HMIWI that have in fact achieved the resulting levels, which necessarily belies the assertion that no HMIWI can achieve them. Regarding the commenters' argument that our recalculation of the MACT floors was contrary to the CAA and intent of Congress, it is clear from the Court's opinion in

Sierra Club

EPA

that EPA needed to revisit the MACT floors in order to respond to the Court's concerns about the MACT floor approach we used in 1997, as noted in its remand of the HMIWI regulations. The Court explicitly “conclude[d] that there are serious doubts about the reasonableness of EPA's treatment of the floor requirements, and remand[ed] the rule for further explanation.” 167 F.3d at 660. Regarding the existing source floors, the Court even went so far as to suggest that, based on its review of the record for the 1997 rule, “EPA's method looks hopelessly irrational.”

Id.

at 664. Ultimately, the Court ordered the case “remanded to EPA for further explanation of its reasoning in determining the `floors' for new and existing [HMIWI].”

Id.

at 666. This remedy squarely placed the responsibility on EPA to either develop an explanation for the MACT standards derived from the 1997 data set that fully addressed the Court's concerns, or develop a different methodology and/or data set that did so.

In the 2008 re-proposal, we decided to use post-compliance data to recalculate the MACT floors because, based on our analysis, it became impossible to fully address the Court's concerns about the suitability of using regulatory limits and uncontrolled emissions values from the 1997 data set in rationally explaining the MACT floors for the 1997 rule. To respond to those concerns, we conducted an analysis comparing the regulatory limits used in the 1997 data set to actual emissions data for those HMIWI, and we determined that the regulatory limits used to establish the MACT floors were not representative of actual operation and did not account for non-technology factors that affected HMIWI emissions performance. (For further information,

see

2008 memorandum “Comparison of Regulatory Limits with Emissions Test Data,” which is included in the docket.) Since it was no longer possible to obtain actual emissions data from the full set of HMIWI that were operating at the time of the 1997 rule's promulgation, the most available alternative was to use the actual emissions data we received from sources who chose to remain in operation and comply with the 1997 MACT standards. With such data, we could actually identify the emissions levels achieved by use of the MACT technologies and control measures that HMIWI employed in order to meet the 1997 standards—technology and measures which we had at that time assumed would be necessary to comply with the standards. This verifying approach was eminently reasonable, since it relied upon data that HMIWI recorded and reported specifically for purposes of demonstrating compliance with the 1997 HMIWI MACT standards, and it addressed the Court's stated concerns regarding the existing source floors. Those concerns, namely, were that permit levels might not accurately estimate actual emissions performance if sources are over-achieving the permit limits (167 F.3d at 663), and that the assumption that unpermitted HMIWI did not deploy emissions controls of any sort was not substantiated (

Id.

at 664).

While we agree with the commenters that control technology has a major impact on pollutant emissions from HMIWI, we also acknowledge that factors other than control technology (

e.g.,

waste mix, combustion conditions) can affect pollutant emissions and should be accounted for in the MACT floor analysis. These non-control technology factors, however, were not considered or reflected by the permit data and uncontrolled emissions values data used in the 1997 rule. Therefore, we needed to take further steps in order to be able to account for these factors and “provide a reasonable estimate of the performance of the top 12 percent of units.”

. at 662. It is true that the Court in

Sierra Club

did not rule that EPA had impermissibly ignored these factors.

. at 666. However, subsequent case law, specifically

National Lime Ass'n

EPA,

233 F.3d 625 (DC Cir. 2000) (NLA II),

Cement Kiln Recycling Coalition

EPA,

255 F.3d 855 (DC Cir. 2001) (CKRC), and

Sierra Club

EPA,

479 F.3d 875 (DC Cir. 2007) (Brick MACT case), have made it abundantly clear that, in any MACT analysis, EPA is currently expected by the Court to address non-technology factors. Based on the actual emissions data we received, which necessarily reflects both the use of

control technologies and any non-technology measures the best performing sources happen to use, we were able to provide the “reasonable estimate” of the best performers' emissions levels that the Court required in its remand. Therefore, we stand by the reassessment we presented in the re-proposal, although, as discussed later in this notice, we have made some adjustments in our statistical analysis to correct for errors in the 2008 re-proposal.

Regarding the commenters' arguments about the impact of subsequent case law, we do not expect that we could reasonably respond to the Court's 1999 remand of the HMIWI rule in a manner that knowingly disregards other flaws in EPA's prior MACT methodology that the Court has since identified. In a recent MACT ruling in which the Court found that EPA had failed to follow the rulings issued in other MACT cases, the Court admonished the EPA that if “[EPA] disagrees with the Clean Air Act's requirements for setting emissions standards, it should take its concerns to Congress. If EPA disagrees with this court's interpretation of the Clean Air Act, it should seek rehearing en banc or file a petition for a writ of certiorari. In the meantime, it must obey the Clean Air Act as written by Congress and interpreted by this court.” 479 F.3d at 884. EPA takes this directive seriously and acted consistently with the Court decisions in preparing this response to the remand. We do not believe that the Court would view its own post-1999 MACT rulings as having changed “the law” (namely, the MACT requirement of Sections 112 and 129) such that following those rulings' instructions would reflect retroactive application of “new” law. The commenters' reliance on

Aaacon Auto Transport

ICC,

792 F.2d 1156 (DC Cir. 1986) is inapposite, as that case addressed an entirely different situation of retroactive application of a new statutory provision; here, instead, the governing statutory requirements have not changed, EPA is acting in response to a Court's ruling that it had not adequately shown that it had complied with those provisions, and the Agency is acting subsequent to further rulings that interpret those same provisions and purport to set forth general directions for EPA to follow in all cases.

As for the comment that EPA could not have relied upon the 2002 compliance data if it had more swiftly responded to the remand, this only suggests that if EPA had acted earlier the EPA would have been forced to take additional steps to require the HMIWI industry to supply emissions data. In no way would this support EPA disregarding the 2002 data we have in-hand and allow us to continue to rely upon data that does not reasonably estimate emissions levels achieved by the best performing units. Based on our analysis of the record, we determined that the 1997 floors did not in all cases meet the requirements of the CAA as interpreted by the DC Circuit. We attempted to explain one set of revisions to the 1997 floors in a subsequent (February 2007)

Federal Register

notice that relied upon the 1997 data set, and received new public comments on that notice and took account of new case law that convinced us that a new approach was required. Consequently, we have chosen on our own to re-open the issues addressed in the 2008 re-proposal.

Comment:

One commenter stated that EPA's approach to revising HMIWI standards under CAA Section 129(a)(5) is correct. The commenter said that revising the MACT floors to reflect the actual performance of the relevant best units satisfies Section 129(a)(5). However, four other commenters objected to revising the floors under the technology review provisions of Section 129(a)(5). The commenters argued that Section 129(a)(5) does not require resetting the floors, but only requires EPA to consider developments in pollution control at the sources and revise the standards based on our evaluation of the costs and non-air quality impacts. The commenters stated that the use of new emissions data is inconsistent with the reasoning EPA presented in other contexts (

e.g.,

in the coke ovens residual risk/technology review rulemaking) that MACT floors need not be recalculated when the EPA conducts its technology review under CAA Section 112(d)(6). The commenters also argued that this approach is inconsistent with the Court's decision on litigation challenging the Hazardous Organic NESHAP (HON) residual risk/technology review rule that there need not be an “inexorable downward ratcheting effect” for the MACT floors.

See NRDC and LEAN

EPA,

529 F.3d 1077, 1083-84 (DC Cir. 2008). One of the commenters also claimed that EPA's approach sets a precedent for all other sources subject to Section 129 or Section 112 MACT standards that could have dire implications on the future viability of rules covering other sources (

e.g.,

MWCs or waste-to-energy facilities).

Response:

Regarding the comment from the first commenter, as noted in the preamble to the December 2008 re-proposal (73 FR 72971), we do not interpret Section 129(a)(5), together with Section 111, as generally requiring EPA to recalculate MACT floors in connection with this periodic review when such review is not conducted together with any other action requiring EPA to reassess the MACT floor.

See, e.g.,

71 FR 27324, 27327-28 (May 10, 2006) (“Standards of Performance for New Stationary Sources and Emission Guidelines for Existing Sources: Large Municipal Waste Combustors; Final Rule”);

see also, NRDC and LEAN

EPA,

529 F.3d 1077, 1083-84 (DC Cir. 2008) (upholding EPA's interpretation that the periodic review requirement in CAA Section 112(d)(6) by itself does not impose an obligation to recalculate MACT floors). However, in the unique case of HMIWI, MACT floor recalculations for the 2008 re-proposal were conducted in order to respond to the Court's concerns stated in its remand of the 1997 regulations, the public comments received on the February 2007 proposal, and recent court decisions, specifically

Sierra Club

EPA,

479 F.3d (DC Cir. 2007) (Brick MACT). This recalculation would have been necessary even if the periodic review requirement of Section 129(a)(5) did not exist. However, Section 129(a)(5) does exist, and EPA must, in addition to responding to the Court's remand, satisfy its requirements. As we previously explained and continue to believe, in this case, our obligation to conduct a 5-year review based on implementation of the 1997 emissions standards is fulfilled through our current remand response. This is supported by the fact that the revised MACT floor determinations and emissions limits associated with the current remand response are based on performance data for the 57 currently operating HMIWI that are subject to the 1997 standards, and by our accounting for non-technology factors that affect HMIWI emissions performance, which the 2007 proposed remand response and 5-year review did not fully consider. Thus, our current remand response more than adequately addresses the technology review's goals of assessing the performance efficiency of the installed equipment and ensuring that the emissions limits reflect the performance of the technologies required by the MACT standards. In addition, the current remand response addresses whether new technologies and processes and improvements in practices have been demonstrated at sources subject to the emissions limits. Accordingly, our current remand response fulfills EPA's obligations regarding the first 5-year review of the HMIWI standards and, therefore,

replaces the 2007 proposal's 5-year review proposed revisions.

2. Pollutant-by-Pollutant Approach

Comment:

Numerous commenters objected to our continued use of the EPA's longstanding pollutant-by-pollutant approach to choosing the best performing HMIWI. The commenters argued that this approach essentially created a hypothetical “super unit” and resulted in the selection of a set of new and existing MACT floors (and standards) that no one existing source has completely achieved and that cannot be simultaneously achieved by any of the best performing sources. The commenters stated that the “best performing” sources must be real sources, not theoretical or hypothetical, based on the statute and legislative history. S. Rep. No. 228, 101st Cong., 1st Sess. 169 (1989). According to the commenters, the proposed standards do not reflect the performance of actual sources, and as such, these proposed standards are not legal under Section 129.

One commenter argued that Section 129(a)(2) (and the similar Section 112(d)(3)) does not speak in terms of the best performing source for each listed pollutant but the best existing source for all pollutants and what these sources can achieve on an overall basis. The commenter claimed that Congress abandoned Section 112's previous focus on individual pollutant standards in the 1990 CAA Amendments and also adopted the technology-based multi-pollutant approach to regulating toxics in use under the Clean Water Act (CWA).

See

S. Rep. No. 228, 101st Cong., 1st Sess. 133-34 (1989). The commenter concluded that if one source can achieve a tight degree of control for one pollutant but not for another, there may be no justification for including it in the set of sources from which the floor is calculated.

See, e.g.,

Tanners' Council of America

Train,

540 F.2d 1188, 1193 (4th Cir. 1976) (CWA effluent limitations guidelines were deemed not achievable where plants in EPA's data base were “capable of meeting the limitations for some, but not all, of the pollutant parameters”).

Two commenters stated that under CAA Sections 129(a)(2) and 112(d)(2) consideration of a higher level of control than the average aggregate levels achieved by the best sources (

i.e.,

using the pollutant-by-pollutant approach instead of basing floors on levels of the full set of pollutants achieved by particular units) must be done only as a “beyond-the-floor” assessment, required to weigh economics and other factors, and not be “hidden” in the floor evaluation, in which costs may not be considered.

Multiple commenters also questioned the technical feasibility of EPA's pollutant-by-pollutant approach. According to the commenters, establishing MACT standards based on the best achievable emissions limits for each type of pollution control equipment assumes that the equipment can be combined in the same system and that the emissions limits of each system are additive. The commenters stated that, in practice, this outcome is likely not achievable due to the challenge of finding pollution control equipment (

e.g.,

fabric filters for PM removal and wet scrubbers for HCl removal) that can work in concert with each other. The commenters said that EPA should consider how the different emissions controls may interfere with each other if employed simultaneously. As an example, one commenter noted that employing a wet scrubber to control HCl would saturate the gas stream, which would bind the bags in the fabric filter used to control PM, thereby compromising the filter's effectiveness. Some of the commenters also noted that the interrelationships between pollutants must be considered in order to ensure that the emissions control is operating effectively for control of all of the related pollutants, and not just a single pollutant. For example, commenters noted that improving combustion to control CO may affect NO

Multiple commenters suggested EPA should revisit the MACT floors for HMIWI and choose the best performing sources on an overall basis, so that at least one source can meet all of the new source standards and a certain portion of the existing sources can meet the existing source standards. One commenter suggested that EPA combine the individual pollutants into a single analysis to determine which control provides the best overall control or otherwise determine that the MACT floor resulting from the analysis is actually achieved by those sources identified as the “best controlled.” According to various commenters, one possible way for doing this would be to establish rankings for how a HMIWI performs for each of the regulated pollutants and then sum the individual pollutant rankings to determine the overall ranking for the HMIWI.

Response:

We disagree with the commenters who object to setting MACT floors on a pollutant-by-pollutant basis. We continue to interpret Section 129 as supporting the pollutant-by-pollutant approach. Section 129(a)(4) says that the standards promulgated under Section 129 shall specify numerical emissions limitations for each pollutant enumerated in that provision. Section 129(a)(2) requires EPA to establish standards requiring “maximum degree of reduction of emissions.” “Maximum degree of reduction of emissions,” in turn is defined in Section 129(a)(2) as including a minimum level of control (the so-called MACT floor). EPA, therefore, believes—and has long believed—that the combination of Section 129(a)(4), requiring numerical standards for each enumerated pollutant, and Section 129(a)(2), requiring that each such standard be at least as stringent as the MACT floor, supports, if not requires, that floors be derived for each pollutant based on the emissions levels achieved for each pollutant.

We also disagree with the commenters who complain that there may not be any operating unit that currently employs the complete suite of MACT technologies and meets the revised limits. The suite of MACT floor controls identified by the final rule approach (specifically, the combination of dry and wet control systems) is already used by four existing HMIWI that meet most of the MACT floor standards. For example, one HMIWI, equipped with a high-efficiency particulate air (HEPA) filtering system, carbon bed adsorber, and rotary atomizing wet scrubber, is estimated to meet all nine revised emissions limits in the final rule; another HMIWI, equipped with a lime injection system, powdered activated carbon injection system, baghouse, and vertical upflow two-stage multi-microventuri scrubber system, is estimated to meet eight of the nine revised limits. Also, an estimated 42 of the 57 HMIWI are estimated to meet both the CO and NO

revised limits simultaneously with existing combustion controls. (

See

2009 memorandum entitled “Revised Compliance Costs and Economic Inputs for Existing HMIWI,” which is included in the docket for today's rulemaking.) The MACT control techniques for the various pollutants are fully integratable and compatible. There do not appear to be any conflicts where meeting the standard for one pollutant may jeopardize the achievability of meeting another pollutant's limit. This conclusion is supported in part by a review of available data and information. As discussed above, there are currently four units that are achieving most, if not all, of the floor standards (based on actual data for each pollutant) using the complete suite of

MACT floor controls. Thus, we conclude that our approach results in compatible MACT controls. Further, an evaluation of the emissions data from units that have measured data for all pollutants supports our conclusion. Our analysis shows that 12 percent (7 of 57 units) simultaneously meet all of the MACT floor emissions levels. (For further information,

see

2009 memorandum entitled “Revised Compliance Costs and Economic Inputs for Existing HMIWI,” which is included in the docket for today's rulemaking.)

We also disagree with commenters claiming that it is inappropriate to consider a suite of floor control techniques that may not be currently in use by the source category. There is no reason not to consider emissions data and controls in use at sources that may be the best performers from some pollutants but not for other pollutants. The MACT floor controls applicable for one pollutant do not preclude the use of MACT floor controls for another pollutant. Therefore, it is appropriate to consider controls at sources employing MACT controls for some pollutants, but not all. For example, floor controls for existing large HMIWI include wet scrubbers for HCl control, dry scrubbers or combination dry/wet systems for PM and metals control, activated carbon injection for CDD/CDF control, and wet scrubbers or dry scrubbers for SO

control. As noted previously, wet and dry systems are demonstrated to be compatible, and it would be inappropriate to exclude from the MACT floor pool those units equipped with wet or dry systems because some of the control systems do better with some pollutants (

e.g.,

wet scrubbers with HCl) than others (

see

previous memorandum).

EPA disagrees strongly with commenters arguing that Congress has directly addressed the issue of whether the MACT floor can be established on a pollutant-by-pollutant basis. With respect to the MACT floor mandate of Section 112, there appears, rather, to be a substantial ambiguity in the statutory language about whether the MACT floor is to be based on the performance of an entire source or on the performance achieved in controlling particular hazardous air pollutants. The language regarding best performing “sources” (or, for new sources, “source”) could apply either to the sources' (or source's) performance as a whole, or performance as to a particular pollutant or pollutants. The same is true of the definition of “emission limitation” in Section 302(k), which refers to “air pollutants,” but does not address whether the limitation must apply to every pollutant emitted by a source, or just some of them. (The same is true of the reference to “air pollutants” (in the plural) in Section 112(d)(2).) In this regard, we note that commenters in other MACT rulemakings have assumed that Section 129, which governs today's rule and which uses language essentially identical to Section 112 in mandating MACT, requires a pollutant-by-pollutant approach to establishing floors, because EPA is commanded to establish standards for enumerated pollutants under Section 129(a)(4). We further note that the DC Circuit, when reviewing the floor determinations we made in 1997 for HMIWI under Section 129 in

Sierra Club

EPA,

noted that they were set pollutant-by-pollutant and found no error in this approach (

see

167 F.3d at 660) (although this aspect of the rule was not challenged specifically). Indeed, the commenters who object so vehemently to the pollutant-by-pollutant approach in this rule raised no such objection when the opportunity to litigate the same approach in establishing the 1997 HMIWI standards was first presented.

EPA also believes that the commenters' reference to basing MACT floors on the performance of a hypothetical or theoretical unit, so that the limits are not based on those achieved in practice, is not only wrong factually (

see

above), but just re-begs the question of what the language in Sections 112(d)(3) and 129(a)(2) is referring to. We did not base the controls or emissions levels on theoretical sources, but on the performance of actual units in the HMIWI source category. All of the MACT floors are achieved in practice (since they are based on actual performance data). Moreover, the DC Circuit has emphasized that EPA may use any reasonable means to determine what levels of performance are achieved in practice.

Sierra Club

EPA,

167 F.3d at 663, 665. The commenters' reliance on cases that they claim preclude EPA's use of a pollutant-by-pollutant approach does not compel a unit-based approach, and the issue is not critical to EPA's position in any event, since the record shows that some units are meeting all of the floor limits and many are meeting several of them. At the very least, the

CMA

EPA

decision under the CWA supports the proposition that a technology-based standard can be considered achievable even if all limits are not yet met by a single unit. Since the floor standards are demonstrably being achieved in practice by some sources, this issue is largely academic.

In short, EPA is not persuaded that the floors must be established on the basis of a unit's performance for all pollutants overall. We continue to believe, as we explained in the 1997 final rule, that such a reading would lead to results that are at odds with evident congressional intent (and with the Court's rulings in NLA II, CKRC and Brick MACT). To argue that Congress compelled this type of result is at odds with both the language of Sections 112 and 129 and common sense. Indeed, it would necessarily suggest that EPA could continue to adopt floors that reflect “no emissions reduction,” even after the DC Circuit so emphatically forbade that approach in the Brick MACT ruling (

Sierra Club

EPA,

479 F.3d 875 (DC Cir. 2007).

As we stated in the preamble to the 1997 regulation (62 FR 48363), we recognize that the pollutant-by-pollutant approach for determining the MACT floor can, as it does in this case, cause the overall cost of the regulation to increase compared to what would result under a unit-based methodology. For example, the pollutant-by-pollutant approach for the HMIWI regulation results in a stringent MACT floor for HCl based on control using a wet scrubber, and stringent MACT floors for PM and metals based on control using a dry scrubber. We interpret Section 129 of the CAA to require that the MACT floor be determined in this manner, and we believe that Congress did in fact intend that sources subject to regulations developed under Section 129 meet emissions limits that are achieved by the best controlled unit for each pollutant, as long as the control systems are compatible with each other. To our knowledge, there is no technical reason why these two air pollution control systems cannot be combined. (62 FR 48363-4) Combined dry/wet scrubber systems are currently in operation on several HMIWI. In response to commenters' concerns regarding the technical feasibility of combined dry/wet systems, available data on the performance of combined dry/wet scrubber systems indicate that the MACT floor emissions levels are achievable and technically feasible. The performance of dry scrubbers with activated carbon injection and the performance of wet scrubbers are well-documented. The available data for combination dry/wet systems provide no indication of operational or emissions problems that occur as a result of combining dry and wet control systems. Regarding the inverse relationship between CO and NO

with regard to combustion control, it is incumbent upon the HMIWI facility to determine whether combustion

conditions can be adjusted to meet both standards and, if not, install add-on NO

controls as necessary,

e.g.,

SNCR systems.

The MACT floor reflects the least stringent emissions standards that EPA may adopt in accordance with Section 129(a)(2) regardless of costs. Other statutory provisions are relevant, although they also do not decisively address this issue. Section 129(a)(4) requires MACT standards for, at a minimum, PM, opacity, SO

, HCl, NO

, CO, Pb, Cd, Hg, and CDD/CDF emitted by HMIWI. This provision certainly appears to direct maximum reduction of each specified pollutant. Moreover, although the provisions do not state whether there is to be a separate floor for each pollutant, the fact that Congress singled out these pollutants suggests that the floor level of control need not be limited by the performance of devices that only control some of these pollutants well. (62 FR 48364)

Regarding the commenter's suggestion that EPA choose the best performing sources on an overall basis, so that at least one source can meet all of the new source standards and a certain portion of the existing sources can meet the existing source standards, we reviewed this approach and found that the suggested approach does not consistently result in emissions limits that are at least as stringent as would have resulted in 1997 if we had actual emissions data and used the correct methodology. We estimate that four emissions limits for large and small non-rural HMIWI and five emissions limits for medium and small rural HMIWI calculated using the suggested overall unit-based approach would be higher than the 1997 emissions limits. Further, because not all pollutants are required to be tested (

e.g.,

and SO

), a substantial fraction of available emissions data would have to be discarded in order to rank only those HMIWI with a complete set of data for all nine pollutants (PM, SO

, HCl, NO

, CO, Pb, Cd, Hg, and CDD/CDF). Specifically, we would have to discard emissions data for 30 percent of large, 40 percent of medium, 100 percent of small non-rural, and 50 percent of small rural HMIWI in order to calculate MACT floors using the suggested approach. (

See

2009 memorandum entitled “Revised MACT Floors, Data Variability Analysis, and Emission Limits for Existing and New HMIWI,” which is included in the docket for today's rulemaking.)

A unit-based approach would tend to result in least common denominator floors where, as here, multiple pollutants are emitted, whereby floors would no longer be reflecting performance by the best performing sources for those pollutants. For example, if the best performing 12 percent of units for HAP metals did not control acid gases as well as a different 12 percent of units, the floors for acid gases and metals would not reflect best performance. Having separate floors for metals and acid gases in this example certainly promotes the stated purpose of the floor to provide a minimum level of control reflecting what best performing units have demonstrated the ability to do.

Similarly, a unit-based approach that employs ranking of a weighted average of pollutants would require EPA to assume priority for certain pollutants (one unit may have lower NO

emissions but higher CDD/CDF, for example). This approach would similarly tend to require EPA to disregard the factual levels reflecting the best performers for individual performers, but based on value judgments regarding the risks presented by various pollutants. Such considerations are antithetical to strictly performance-based analyses such as MACT floor determinations. Indeed, reviewing EPA's primary copper smelters MACT standard, the DC Circuit rejected the argument that risk-based considerations have any place in the MACT context (

see Sierra Club

EPA,

353 F.3d 976 (DC Cir. 2004).

3. Adequacy of Emissions Test Data

Comment:

Multiple commenters argued that the proposed standards are flawed because EPA has not demonstrated that the actual emissions data on which the proposed rule is based adequately represent the full range of performance of tested facilities. According to various commenters, the emissions data were derived from performance tests conducted under “representative operating conditions,” rather than the “worst reasonably foreseeable circumstances” contemplated by the case law.

See Sierra Club

EPA,

167 F.3d 658, 665 (DC Cir. 1999). Commenters stated that the proposed emissions limits did not adequately account for variability, and said EPA should have sought out more test data and specifically requested continuous monitoring data to properly characterize variability.

Another commenter specifically recommended that EPA gather additional data on emissions of medium HMIWI such as theirs before finalizing the rule to ensure each medium HMIWI has data sufficiently accurate and representative to properly set a MACT standard in accordance with the CAA Amendments. According to the commenter, rigorous quality assurance/quality control (QA/QC) procedures should also be applied to the test data.

One commenter stated that, because the new regulations are solely based on previous stack testing, the actual emissions tests need to be reviewed by EPA for technical accuracy, as well as consistency. Although there may have been insufficient time under the court-ordered schedule, the commenter argued that proposed standards cannot be defended technically in the absence of such an analysis.

The same commenter also stated that revisions to EPA's incinerator test protocol are needed to ensure that the unit is being tested at proper design conditions. At a minimum, the commenter said that incinerator temperature, waste input rate and constituents, auxiliary fuel consumption, quench rates (air and water), and chemical feed rates need to be recorded during an incinerator test to determine whether the operating and testing conditions were representative of the higher emissions rates that can be experienced during normal operations. Given that emissions are determined by waste characteristics, the commenter recommended that a standardized realistic worst-case test waste be used, which includes specific criteria components, as well as moisture content and heating value. Incinerators would be tested with the standard waste and the top 12 percent identified.

Response:

First, in response to industry commenters who claim we should have gathered more data, we note that nothing precluded them from giving us more data to consider in responding to the Court's remand, if they felt that the data submitted to us for purposes of showing compliance with the 1997 standards was not representative of their normal operations. We have reasonably used the data available to us at the time we conducted this rulemaking, in the absence of being provided with any other data. We agree with the commenters that emissions tests might provide information on representative operations only where owners and operators conducting the tests have endeavored to reflect such representative operations at the time of the tests. However, when conducting tests to establish various parameters to be monitored, owners and operators may also endeavor to produce data for a wide range of operating conditions. Moreover, we have taken several steps to try and account for the emissions and operational variability, including (1) obtaining additional emissions test data

from States and EPA Regions representing all available annual test results for each unit, (2) using individual test run data for the best-performing 12 percent of sources to calculate UCL values, (3) using a substantial confidence interval (specifically, a 99 percent UCL value), and (4) closely reviewing how the data are distributed (

e.g.,

normally, lognormally). Also, EPA's own review of emissions factors shows that the variability of emissions between facilities is greater than the variability within facilities.

We believe that the data quality concerns expressed by the commenters have been addressed in a number of ways. First, EPA test methods incorporate data quality assurance and quality control steps and acceptance criteria at several levels. These provisions assure that the data produced are of quality sufficient for decision making, including compliance, when the methods are followed and the acceptance criteria are met. Second, States further assure that testers adhere to the test methods by providing third party oversight and review of compliance tests conducted by industry, such as that being discussed here. The States also implement the source testing audit program when available, further assuring the high quality of emissions testing data. Third, through internal and contractor support efforts for this regulatory project, EPA conducted additional review of the initial emissions test data to check for completeness and appropriate characterization of process operations. Finally, EPA reviewed and accounted for variability inherent in the emissions data used in establishing the applicable emissions limit including applying statistical confidence intervals.

Regarding the comment about revisions to EPA's incinerator test protocol, the factors cited by the commenter could be considered in setting site-specific compliance conditions. Such an approach may be useful at the next technology review. The commenter's suggestion that EPA use a standardized waste for testing is questionable, unless EPA wanted to establish a certification testing program like the residential wood combustion rule. However, such a program would be cumbersome and could potentially eliminate a majority of the industry.

Comment:

Three commenters stated that EPA did not consider the accuracy and precision of the EPA test methods in proposing the emissions limits for new and existing HMIWI. To support their argument, the commenters referenced the findings of the Reference Method Accuracy and Precision (ReMAP) program co-sponsored by the American Society of Mechanical Engineers (ASME). According to the commenters, one of the main objectives of the ReMAP project was to ensure emissions limits would properly consider the inherent accuracy and precision limits of the test methods used to demonstrate compliance, such that a facility would not be in violation of a limit as a result of this inherent variability. The commenters noted that the ReMAP program established Precision Metrics for various reference methods and corresponding pollutants (

e.g.,

±42 percent for CDD/CDF Method 23), and they compared these Precision Metrics to actual stack concentrations and proposed emissions limits for several pollutants. Based on this comparison, the commenters concluded that EPA did not adequately address these Precision Metrics in establishing the proposed limits.

Response:

As noted above, we already took into account variability inherent in the data representing emissions and process operations in establishing the emissions limit. By using UCLs to set our emissions limits, we have inherently accounted for measurement precision. In fact, the adjustments we made to the average stack concentrations for the best-performing 12 percent of units to calculate the final emissions limits more than account for the Precision Metrics cited by the commenters. Thus, any additional adjustments of measurement to account for method precision are unnecessary.

Comment:

One commenter stated that there are significant deficiencies in the emissions data used to establish the standards. Some of the standards are based on data from a limited number of stack tests. According to the commenter, there needs to be a standard for the minimum number of stack tests that must be performed before its data can be used as the basis for determining the top 12 percent performing incinerators. Because of the waste characteristics and variability, the commenter recommended a minimum of four tests. The commenter noted that some of the units included in the top 12 percent are specialty incinerators, which the commenter said are not representative of the subcategory as a whole. The commenter also noted that another unit incinerates municipal waste, which the commenter argued should cause its data to be invalid for the proposed HMIWI standards. According to the commenter, municipal waste would be expected to have a makeup that produces significantly lower emissions for some pollutants (

e.g.,

CDD/CDF, Cd). The commenter recommended developing a testing metric (

e.g.,

heating value, flue gas per pound of feed) and applying it to the data used to indicate possible flaws (

e.g.,

variations and/or abnormalities) which would spur further investigation into the validity of the data. Of the 45 emissions tests used to develop emissions limits for the large subcategory, the commenter concluded that 38 of those tests could be considered invalid because of too little testing or the unrepresentative content of the incinerated waste stream.

Response:

Regarding the commenter's argument about claimed deficiencies in the emissions data used to establish the standards, we do not believe that data from high quality tests should be dismissed simply because there are only a few tests. As noted above, we have reasonably relied upon the data we had available to us, and we have already taken steps to alleviate concerns about the representativeness of the measured data used to establish the emissions limit, including calculating UCL estimates using standard statistical conventions.

Regarding the commenter's concerns about the specialty incinerators and the facility that also incinerates municipal waste, we evaluated creating separate subcategories for captive units (which would include the specialty incinerators) and a separate subcategory for mixed waste units, but as noted above, we ultimately rejected both options because we did not provide an opportunity to comment on the issue of subcategorization in the December 2008 re-proposal or a record that would justify such a significant change in categorization. Another option to address the facility incinerating municipal waste would be to use only the emissions data from those tests conducted with 100 percent medical waste, but that would limit the number of tests for that facility. Also, we have found a significant amount of overlap in emissions (including CDD/CDF and Cd) between the different test conditions at the facility (

e.g.,

100 percent medical waste, 50 percent medical waste, 20 percent medical waste,

etc.

), suggesting that such a distinction in waste type is not very meaningful in this case. (

See

2008 memorandum entitled “Documentation of HMIWI Test Data Database,” which is included in the docket.)

Comment:

Three commenters stated that some emissions test data were improperly excluded from the dataset, including data deemed “non-compliant,” data collected at HMIWI subsequently shut down, and data collected under specific “test

conditions.” The commenters argued that emissions test data from compliance tests that were conducted in accordance with the applicable reference test methods for affected HMIWI should not be arbitrarily excluded from the re-stated MACT dataset, because that undermines the entire data evaluation process. The commenters stated that EPA provides no rationale for arbitrarily including data in some instances, and excluding them in others. Thus, according to the commenters, EPA's proposed standards are arbitrary and capricious. The commenters said that inclusion of all valid test data provides a better representation of the inherent variability of the various test methods and source operation. According to the commenters, EPA's MACT floor dataset was inconsistent, leading EPA to rely on an unrepresentative set of data. The commenters recommended that EPA provide a clear description of “representative HMIWI operation” so that consistent criteria are applied to evaluate whether valid emissions test data were properly included or excluded from the MACT floor dataset.

Response:

Non-compliant emissions data from the initial tests of HMIWI were not included in the emissions database used to establish the emissions limits. At the time of the initial test, operators were still in the process of establishing their operating parameters and tuning their emissions control devices and operating conditions to comply with the regulation. Any non-compliant emissions data from the initial test would be expected to trigger a change in HMIWI operation in order to come back into compliance with the 1997 standards. Consequently, the non-compliant emissions data from these tests would not be representative of the typical operation of these HMIWI.

If non-compliant emissions data from an annual test were substantially higher than the emissions typically seen from the facility or were substantially higher than the emissions limit, this strongly suggested that there was a problem during the test and indicated that the test results would not be representative of the typical operation of the HMIWI. Such data were excluded from the pollutant averages for the particular facility. (It should be noted that the data that were excluded amount to less than 1 percent of the total set of emissions data for the industry.) For example, the emissions data from tests on one unit did not meet the PM or Cd emissions limit during an August 2006 annual test. A subsequent retest of this unit for those same pollutants in November 2006 showed PM emissions results less than 10 percent of those measured earlier, and Cd emissions results about 0.1 percent of the previously measured results. Consequently, we believe that the August 2006 PM and Cd test results were not representative of the typical operation of the HMIWI, and they were not included in the test data database. The PM and Cd retest data from the November 2006 retest were included instead. (

See

previous memorandum.)

We also excluded test data if we found errors in the calculations or the test methods, or some important elements of the data needed to calculate emissions in the form of the standard were missing. For example, we excluded the TEQ emissions estimates provided for a 2005 annual test at a second HMIWI because the reported TEQ estimates were greater than the total CDD/CDF estimates provided, a clearly incorrect result. The total CDD/CDF estimates were believed to be the correct values because they were well within the applicable emissions limit, while the TEQ estimates were a few times higher than the applicable limit. The 2001 annual test results for HCl at a third HMIWI were deemed invalid because the HCl sample train did not meet the method's ±95 percent sample collection efficiency requirement. There was believed to be some contamination in the sample collection and/or recovery during the 2005 Pb test at a fourth HMIWI, so a retest in February 2006 was conducted. The Pb results from the February 2006 retest were included with the results of the 2005 annual test in the test data database, replacing the 2005 Pb results. The first HCl test run during a 2006 test at a fifth HMIWI was below the detection limit, and the laboratory that analyzed the samples did not provide a detection limit for this test run. In this case, we decided to delete the results for this particular test run and calculated the HCl average for the 2006 test using the results from the other two test runs. Similarly, the second Hg test run during the 2003 test at a sixth HMIWI was reported to be below the detection limit, but the data summary did not include the measured Hg detection limit. Attempts to obtain the detection limit for this test run from the facility were unsuccessful. Consequently, we decided to delete the results for this test run and calculated the Hg average for the 2003 test using the results from the other two test runs. (

See

previous memorandum.)

A couple of annual compliance tests were excluded from the unit averages because they were conducted under test conditions (

e.g.,

reduced emissions control) that were not considered representative of the typical operation of the HMIWI. The exclusion of these tests had little impact on most of the pollutant averages for these HMIWI, and it should be noted that these HMIWI are not in the MACT floors of the pollutants of interest. One HMIWI was unable to meet the CDD/CDF emissions limit during the 2003 and 2004 annual compliance tests conducted without activated carbon. Only when activated carbon injection was included as a second test condition during the 2004 annual compliance test was the facility able to meet the CDD/CDF emissions limit. Consequently, we determined that the second test condition was more representative of the typical, current operation of the HMIWI. During a Hg annual compliance test, another HMIWI was unable to meet the Hg percent reduction limit under the test condition with a lower activated carbon injection rate, but was able to meet the limit under the test condition with a higher activated carbon injection rate. The Hg data meeting the limit were considered representative of the typical operation of the HMIWI, and the other Hg data were rejected. (

See

previous memorandum.)

Regarding the argument that EPA improperly excluded data available from HMIWI that subsequently shut down, we believe that it is appropriate in this particular rulemaking to base the MACT floor on emissions data from facilities that are currently operating, since those are the facilities that would be complying with the rule.

Comment:

Three commenters stated that the treatment of individual “non-detect” data points within the MACT floor dataset should be consistent and should represent the actual detection level of the pollutant of concern. The commenters noted that non-detect or zero data provided as part of the latest data request were considered equal to the method detection limit, while CDD/CDF test data already in EPA's project files were calculated at one-half the detection limit. While this approach may be valid for total CDD/CDF, the commenters argued that it could have a profound effect on TEQ.

Response:

In response to the commenters, it should be noted that section 9 of EPA Method 23 specifies that “[a]ny PCDD's or PCDF's that are reported as below the measurement detection level (MDL) shall be counted as zero for the purpose of calculating the total concentration of PCDD's and PCDF's in the sample.” The CDD/CDF results reported in the facilities' initial test reports and provided by States and EPA Regions in the annual test summaries reflect this computation approach. Consequently, by using one-half the detection limit in our review of

CDD/CDF data in full test reports, we were being conservative in our estimation of CDD/CDF emissions. Nonetheless, we looked at those HMIWI in the MACT floor for total CDD/CDF and TEQ to determine whether using the full detection limit would make a substantial difference. When we averaged in the results with all other CDD/CDF results for each facility, we found on average essentially no difference in total CDD/CDF emissions estimates (less than 1 percent) and only a small difference in TEQ emissions estimates (0.1 to 20 percent) for the four HMIWI size categories. (

See

2009 memorandum entitled “Comparison of CDD/CDF Non-Detect Data—Full Detection Limit vs.

1/2

Detection Limit,” which is included in the docket for today's rulemaking.)

4. Non-Technology Factors

Comment:

Numerous commenters stated that the variability in non-technology factors, such as the materials and composition fed to combustion devices, must be adequately addressed in the rulemaking process in order to promulgate a feasible rule,

Sierra Club,

479 F.3d at 883 and

Cement Kiln Recycling Coalition

EPA,

255 F.3d 855, 865 (DC Cir. 2001). According to various commenters, EPA did not identify the non-technology factors in the proposed rule or quantify their effect on actual emissions performance, but instead claimed, without supporting evidence, that using actual emissions levels accurately reflects emissions performance resulting from the use of add-on controls and other emissions reduction measures. Commenters argued that the failure to make these findings renders the proposed standards arbitrary. Another commenter disagreed, stating that EPA's proposed floor approach for new and existing HMIWI is generally correct and that EPA correctly observed that the use of actual emissions levels accounts for all emissions reduction strategies.

Response:

With regard to the commenters' argument, the CAA does not require EPA to quantify the emissions reductions resulting from all non-technology factors, but instead focuses on identifying the emissions levels achieved by best performing sources no matter what means they use to achieve them. This approach is supported by the DC Circuit's decision in the Brick MACT case, which stressed the importance of identifying emissions “levels” achieved by sources. There can be no dispute that both the composition and level of emissions exiting the incinerator reflect both the add-on control technologies used by a unit (

e.g.,

dry scrubber, wet scrubber, activated carbon) that control the emissions and the non-technology factors (

e.g.,

waste material quantity and composition, combustion conditions) that influence the level and composition of emissions. As the

Sierra Club

Court noted in 1999, the less mercury fed into the waste stream, the less mercury emissions will be coming out of the stack. Whatever combination of add-on controls and non-technology measures a unit is employing will, therefore, necessarily affect the resulting emissions levels that are reflected in the actual emissions data upon which the revised floors are set. It would be impossible for those data to not reflect all those measures. This situation is quite the opposite of what was presented in the 1997 rulemaking, in which the floors were primarily derived from permit and regulatory levels that were not necessarily reflective of actual emissions performance but were assumed to reflect levels achievable by add-on control only. At that time, to adjust floors downward to account for non-technology factors, it might indeed have been necessary to be able to quantify additional emissions reductions attributable to such measures. Similarly, as the 2007 proposed remand response still in large part relied upon the permit and regulatory levels, not knowing the quantified reductions achieved by non-technology measures frustrated estimating the emissions levels achieved in practice by HMIWI. But this is simply not an issue under a methodology that depends upon the measured emissions levels that result from whatever mix of add-on or non-technology controls is being used, as under the 2008 re-proposal and today's final rule. The non-technology factors cannot help but affect the actual emissions data, and they are, therefore, necessarily accounted for in the actual emissions data-based floors.

EPA's data gathering effort for this rulemaking included not just initial and annual emissions test data obtained from EPA Regions, State/local governments, and HMIWI facilities, but also a waste segregation practices questionnaire sent to nine representative entities in the HMIWI category (six hospitals, one pharmaceutical facility, one university, and one company that owns 8 of the 14 commercial HMIWI). (

See

2008 memoranda entitled “Documentation of HMIWI Test Data Database” and “Summary of Industry Responses to HMIWI Waste Segregation Information Collection Request,” which are included in the docket.) While our analysis of the emissions test data indicates a strong relationship between add-on control and emissions (

e.g.,

wet scrubbers achieve superior HCl control, while dry scrubbers achieve superior PM and metals control), our review of the questionnaire responses indicates that non-technology factors also play a role in emissions reduction. All of the survey respondents, except for the commercial company, practice onsite waste segregation to reduce the volume of waste being incinerated. Most of the respondents started the practice of waste segregation in the 1980s and 1990s. Five respondents also accept offsite waste and require the offsite waste generators to employ waste segregation practices. The commercial company encourages waste segregation from its waste generator clients through a number of efforts, including a waste management plan, contract requirements and waste acceptance protocols, a dental waste management program, and educational programs and supporting posters. All of the respondents that practice onsite waste segregation separate batteries and fluorescent bulbs (

i.e.,

mercury waste) from the HMI waste stream. Eight respondents separate paper and/or cardboard, four separate glass, and three separate plastics from the HMI waste stream. Other materials that are separated from the HMI waste stream include hazardous waste, waste oil, wood, construction debris, refrigerants, and various metals and metals-containing materials (

e.g.,

aluminum, copper, lead, mercury, steel, and electronics). (For further information,

see

2008 memorandum “Summary of Industry Responses to HMIWI Waste Segregation Information Collection Request,” which is included in the docket.) These waste segregation efforts would certainly have an impact on the emissions of CDD/CDF, mercury, and other pollutants from these HMIWI and would be reflected in the emissions levels measured during their initial and annual emissions tests and used in our test data analysis. As noted previously, the nine entities surveyed were believed to be representative of the HMIWI industry as a whole, so the conclusions reached for the nine entities are also expected to apply to the entire industry as well.

5. Straight Emissions Approach

Comment:

Two commenters argued that the parenthetical language in the Brick MACT decision equating the best performers with “those with the lowest emissions levels” (straight emissions approach) was only a legal dictum to

which EPA is not bound, and which is not cited in either the CKRC decision or the CAA. The commenters cited

Sierra Club

EPA,

479 F.3d 875, 880 (DC Cir. 2007) (Brick MACT), and

Cement Kiln Recycling Coalition

EPA,

255 F.3d 855, 861 (DC Cir. 2001). In citing EPA's justification for the MACT floor approach used in the hazardous waste combustor rulemaking, the commenters stated EPA's position that the CAA does not require the Agency to equate the best performers with the lowest emitters. The commenters specifically cited EPA's statement that, “as a legal matter, CAA Section 112(d)(3) does not specifically address the question of whether `best performing' sources are those with the lowest net emissions, or those which control HAP emissions most efficiently.”

The commenters also noted that, since the Brick MACT decision, EPA has determined that there are other ways to rank the best performing sources and set the MACT floors than a straight emissions approach, such as the approach used in the hazardous waste combustor rulemaking, which combined the hazardous waste fed to the source and the source's system removal efficiency (SRE). According to the commenters, the “SRE Feed” methodology better identifies who the lowest emitters will be over time, better assesses their performance (

i.e.,

how much they will emit as they operate), and better accounts for variability (

e.g.,

non-technology factors).

Response:

It is not necessary to adopt a position regarding whether the Brick MACT Court's references to “emissions levels” is dictum or binding for purposes of this rulemaking. In the 1999 HMIWI case, the Court very clearly stated that EPA's duty here was to use data that allowed the Agency to reasonably estimate the emissions performance of the best performing units. We have discovered that the permit and regulatory data upon which the 1997 rule was based do not reliably serve this purpose. Conversely, the actual emissions data from HMIWI do enable us to estimate the performance of the best performers. We believe that the use of actual emissions data, appropriately adjusted for variability using statistical methods, sufficiently accounts for the performance and variability of HMIWI operation. Regarding the commenters' reference to CAA Section 112(d)(3) to support their argument regarding the definition of “best performing” sources, we assume the commenters also meant Section 129, which governs this rule.

We do not think the SRE Feed methodology can be successfully adapted to determine MACT floors for HMIWI. This is because the SRE Feed approach requires knowledge of the amount of hazardous materials fed into the system and knowledge of the system's removal efficiency for those specific materials, neither of which is known or measured in the HMIWI industry. Such materials are mixed in with other waste and cannot reasonably be measured separately, especially given the occupational safety regulations to which HMIWI operators are subject.

6. Statistical Approach

Comment:

Multiple commenters stated that the statistical methodology EPA used to establish MACT floors did not properly account for underlying non-technology factors such as feed material quantity or composition or for normal operational variability within and across unit operations, which led to unattainable emissions limits.

Three of those commenters supported the conditional use of the 99.9 percent UCL to quantify “emissions limitation achieved” as it applies to variability above average emissions. However, the three commenters had concerns about EPA's methods used to calculate statistical parameters. The commenters stated that EPA should characterize emissions data distributions before calculating statistics, instead of assuming all data are normally distributed. Otherwise, according to the commenters, it is difficult to determine if the statistics are valid. When data are not normally distributed, the commenters recommended that EPA transform the data prior to conducting its statistical calculations. The commenters noted that EPA used the NORMSINV function in Microsoft Excel to calculate the 99.9 percent UCL, which assumes that the actual mean and variance of a data set is known. According to the commenters, when the mean and variance are estimated from random samples or a small subset of the total population, such as stack test runs, the 99.9 percent UCL should be calculated with the Student t-statistic using the TINV function in Excel, not normal statistics.

Two other commenters objected to the use of the 99.9 percent UCL to account for variability in determining emissions limits. One of the commenters argued that EPA provides insufficient explanation or justification of its use of the 99.9 percent UCL. According to the commenter, if the performance of the best performing HMIWI, on average, is estimated to meet the emissions limit 99.9 percent of the time, then it would be expected to exceed the emissions limit 8.76 hours per year, which does not comply with the requirement that each source must meet the specified floor every day and under all operating conditions. Therefore, the commenter argued that the 99.9 percent UCL procedure used by EPA is deficient and must be revised.

The other commenter stated that EPA's use of a 99.9 percent UCL to estimate individual units' variability marks a sharp departure from EPA's approach in other rulemakings (

e.g.,

90 percent and 95 percent UCL), and said that EPA offers no real explanation for this departure from past practice or why a 99.9 percent UCL would account for variability but a lower UCL, such as 99 percent or 95 percent or 90 percent, would not. The commenter recommended that EPA correct its floor approach to avoid the overcompensation for variability seen with some of the floors for new units.

Two commenters stated that a more realistic assessment of an individual unit's ability to meet an emissions limit during a compliance test would use the 99.9 percent UCL for that unit/pollutant instead of the average value.

Four commenters disagreed with EPA's decision to use individual test run results to account for variability in setting MACT floors for new and existing sources. The commenters urged EPA to use complete performance test results instead. One of the commenters argued that EPA is arbitrarily using different measures of performance for establishing emissions standards on the one hand (using test runs) and measuring compliance with these standards on the other (using whole tests), without explaining why different measurement approaches are appropriate. According to the commenter, it appears likely that disaggregating test results leads to less protective floors by creating false variability in individual units' performance. The commenter recommended that EPA calculate the floors with and without disaggregating individual test runs to ensure that its floors are not less stringent as a result of that approach. The other commenters noted that data limitations may not leave EPA an alternative to using test run results in some cases, but they recommended that EPA use complete test results where enough data exist to characterize emissions variability.

Response:

Based on the responses to our waste segregation practices questionnaire, we believe that most HMIWI are practicing (or encouraging the practice of) waste segregation of materials such as batteries, fluorescent bulbs, paper, glass, plastics, and metals-containing materials, which we expect

to impact the emissions of CDD/CDF, mercury, and other pollutants and be reflected in the actual emissions data we use in our analysis. (

See

2008 memorandum “Summary of Industry Responses to HMIWI Waste Segregation Information Collection Request,” which is included in the docket.) Consequently, we believe that using actual emissions data sufficiently and inherently accounts for non-technology factors such as feed material quantity or composition which influence the level and composition of emissions. We also believe that our use of multiple emissions tests and individual test runs for each HMIWI, where possible, and our estimation of 99 percent confidence intervals for MACT floor data sufficiently accounts for variability. The use of multiple emissions tests allows us to evaluate “between-test variability,” which can occur even where conditions appear to be the same when two or more tests are conducted. As we noted in the preamble to the December 1, 2008 re-proposal (73 FR 72976, 72980), variations in emissions may be caused by different settings for emissions testing equipment, different field teams conducting the testing, differences in sample handling, or different laboratories analyzing the results. Identifying an achieved emissions level needs to account for these differences between tests, in order for “a uniform standard [to] be capable of being met under most adverse conditions which can reasonably be expected to recur[.]” (

See NLA I,

627 F.2d at 431, n. 46.) (

See

also

Portland Cement Ass'n,

486 F.2d at 396 (noting industry point that “a single test offered a weak basis” for inferring that plants could meet the standards).) The use of individual test runs (as opposed to test averages or unit averages) allows us to evaluate “within-test variability.” A single test at a unit usually includes at least three separate test runs. (

See

§ 63.7(e)(3) (for MACT standards under Section 112 of the CAA), and § 60.8(f) (for NSPS under CAA Section 111).) Each data point should be viewed as a snapshot of actual performance. Along with an understanding of the factors that may affect performance, each of these snapshots gives information about the normal, and unavoidable, variation in emissions that would be expected to recur over time. To account for pollutant-specific variability at the best-performing unit (for new source MACT) or best-performing 12 percent of units (for existing source MACT), we used emissions data for each test run conducted by those units. The amount of pollutant-specific test data for those HMIWI varies widely for each size category. Given the limited amount of test data and the uncertainty regarding that short-term emissions test data, we have decided that using the 99 percent UCL is an appropriate method of estimating variability. The UCL represents the statistical likelihood that a value, in this case an emissions value from the best performing source, will fall at or below the UCL value. (Further discussion regarding the 99 percent UCL is provided later in this section.)

After reviewing the commenters' suggestion that we characterize emissions data distributions before calculating statistics, we took a closer look at our statistical approach. In statistics, skewness is a measure of the degree of asymmetry of a distribution. Normal distributions typically have a skewness of zero. Consequently, to determine whether the emissions test data used in our UCL calculations had a normal or lognormal distribution, we estimated the skewness of the data using the SKEW function in Excel. Except as specified below, those datasets with a skewness value greater than zero (when rounded to a whole number) were categorized as lognormal, and all other datasets were categorized as normal. Those data categorized as lognormal were transformed (by taking the natural log of the data) prior to the calculation of UCL values. When there were only a few data points (

e.g.,

one emissions test with three test runs), which is the case for most datasets for small HMIWI, it was not possible to make a definitive determination that the data were distributed normally or lognormally. (In fact, assuming a lognormal distribution for those data often resulted in UCL values that were substantially higher than the 1997 promulgated limits.) In those cases, we decided to use the normal distribution in calculating UCL values, a conservative assumption which provided a more protective emissions limit. When we had more data and could make a more definitive determination about a dataset's distribution, we treated the data as noted previously. In most cases, we found that the larger datasets are lognormally distributed, although there are some cases where they appear to be distributed normally, and we treated the data as such when doing our UCL calculations. We believe this approach is more accurate and obtained more representative results than those at re-proposal.

Regarding the commenters' suggestion about using Student's t-statistics in calculating the UCL values, we also decided to revisit our statistical approach. We agree that we have only a relatively small, random sample of emissions data available for our analysis, which calls for the use of the Student's t-test, in accordance with standard statistical practice. Consequently, we have decided to use the TINV function in Excel (specifically the one-tailed t-value), rather than the NORMSINV function, to calculate the UCL values. This approach (using the Student's t-test) is consistent with approaches being taken in other EPA rulemakings, such as Portland Cement.

In response to public comments on the size of the confidence limits used at re-proposal and in light of the aforementioned changes in our statistical approach, we also decided to reevaluate the percentiles used in the UCL values. We evaluated four different percentiles (90, 95, 99, and 99.9 percent). The 99.9 percent UCL values estimated for the 2009 final rule are substantially higher than the highest test runs for the MACT floor units and are frequently higher than the emissions limits in the September 15, 1997 promulgated standards, indicating the 99.9th percentile overcompensates for variability. Lower percentiles (

e.g.,

90, 95, and 99 percent) are inherently more stable than the 99.9th percentile, with less uncertainty (less variability) than the 99.9th percentile from a statistical standpoint. However, the 90 and 95 percent UCL values are frequently lower than the highest test runs for the MACT floor units and the stringent emissions limits in the December 1, 2008 re-proposal, indicating that those percentiles provide insufficient compensation for variability.

The 99 percent UCL values are somewhat higher than the emissions limits in the December 1, 2008 re-proposal but are well below the emissions limits in the September 15, 1997 promulgated standards. The 99 percent UCL values are more in line with the highest test runs for the MACT floor units than the other percentiles, indicating that the 99 percent UCL provides a more reasonable compensation for variability. This approach results in standards more representative of the level of emissions reduction that the best performing sources are actually achieving. Accordingly, we have decided to use the 99 percent UCL to estimate emissions limits for the 2009 final rule.

We disagree with one commenter's argument that the 99.9 percent UCL must provide for the floor to be met every day and under all operating conditions. The UCL is not about time, but about the population of data. Accounting for variability using the 99.9 percent UCL goes beyond the absolute

average but does not produce expectations of 0.1 percent noncompliance. Setting the emissions limit at the UCL accounts for the possibility of variability and the possibility that the average is outside the range. These statistical procedures are used to help us identify the average emissions limitation achieved by the best performing units, as Section 129(a)(2) of the CAA requires. Also, there is no practical upper limit as to what a facility can emit, so the argument that that EPA must set a floor at a level that equates to what a facility can meet at all times is not consistent with the CAA's requirement that EPA estimate the emissions levels achieved by best performing units.

Regarding the comment about our decision to use individual test run results to account for variability, we felt it was necessary to use test run results when we had data limitations (

e.g.,

for small HMIWI) and for consistency decided to take the same approach where data were more plentiful. As noted previously, we believe that each data point should be viewed as a snapshot of actual performance, which gives information about the variation in emissions that would be expected to recur over time.

D. Emissions Limits

1. HCl, CDD/CDF, and Metals Emissions Limits

Comment:

One commenter argued that EPA's proposed HCl standards of 2.4 parts per million by volume (ppmv) for existing sources and 0.75 ppmv for new sources are based on biased data of indeterminate quality and are unachievable. The commenter also claimed that setting the HCl standards at such low levels will negatively impact the development and application of CEMS, due to the lack of correlation between Method 26A and CEMS at concentrations comparable to the proposed standards. According to the commenter, the test results (Methods 26 and 26A and RCRA SW 846 Method 0050) that EPA used to set the HCl standards contain a known bias at low levels of HCl, varying widely with temperature and moisture at HCl levels below 20 ppmv (all three methods), and having a negative bias at HCl levels below 5 ppmv (Method 26A). The commenter noted that all of the top performers in the large, medium, and small non-rural categories use wet scrubbers to control HCl emissions, and will have considerable moisture in the stack gas. Thus, the data from every one of these sources has the potential to be biased. The commenter argued that HCl data below 20 ppmv are not usable and/or representative and are technically indefensible. The commenter recommended that EPA follow the example of Office of Solid Waste (OSW), which corrected all HCl values below 20 ppmv to 20 ppmv, used a statistical method to impute a standard deviation for these test runs, and calculated a floor standard based on those values.

Response:

We are basing the HCl standards in this rulemaking on the data we have available to us from the HMIWI source category, and can base them only on that data. The sensitivity of Method 26A for HCl is 0.04 ppmv. Moisture is only an issue with Method 26A if the testing contractor does not perform the method correctly. Unless we are given data to the contrary, we assume that the HCl data in our dataset are correct. These data, for this particular rulemaking, support the HCl standards being adopted today.

Nonetheless, we acknowledge that the HCl standards in our re-proposal were very close to the method detection limit for HCl. The changes in statistical approach for the final rule have resulted in increases to the HCl standards above 5 ppmv, which should address some of the concerns listed above. Furthermore, based on reported HCl emissions data for all HMIWI, we estimate that 64 percent of large, 82 percent of medium, and 100 percent of small/small rural HMIWI will be capable of meeting the revised HCl standards, on average, based on their currently used control measures. It should also be noted that HMIWI subject to the 1997 NSPS have been meeting the 15 ppmv HCl standard in that rule, which is below the 20 ppmv threshold level that the commenter cited.

Comment:

One commenter recommended that EPA set beyond-the-floor standards for both HCl and chlorinated organic pollutants (including CDD/CDF) based on removing chlorinated plastics from the waste stream. According to the commenter, it is well established that the combustion of chlorinated plastics increases emissions of HCl as well as CDD/CDF and other chlorinated pollutants. The commenter stated that it is achievable for HMIWI to remove chlorinated plastics from the waste stream that they burn. The commenter said that EPA can gather data that will quantify the total amount of HCl that is attributable to the combustion of chlorinated plastics and set a standard reflecting the maximum degree of reduction that is achievable through the removal of chlorinated plastics from the waste stream.

The same commenter also recommended that EPA set beyond-the-floor standards for metals based on removing all metals from the waste stream before combustion, consistent with the requirements under Section 129(a)(2) and (3), which obligate EPA to require the maximum degree of reduction in emissions that is achievable through the use of methods and technologies before, during, and after combustion. The commenter stated that metals do not belong in an incinerator because they cannot be destroyed by incineration and are especially dangerous to public health and deleterious to the environment. As far as the commenter knew, EPA has never disagreed that removing metals from the HMIWI waste stream is achievable technically and economically, and the commenter noted that EPA has data from the MWC rulemaking that show materials separation requirements are effective and cost-effective. (

See

Docket A-89-08, various items.)

Given the language of Section 129 that requires the maximum degree of reduction in emissions that is achievable through the use of pre-combustion measures, the commenter argued that EPA has a duty to gather information on these measures and evaluate such measures in its beyond-the-floor analysis. According to the commenter, EPA's failure to gather information about the precise reduction of emissions that will result from such measures and failure to provide any explanation for rejecting such a standard is unlawful and arbitrary. The commenter noted that EPA has committed to set final standards by September 2009, and stated that EPA should not delay issuance of final standards to conduct this data gathering, but should commence data gathering now and revise the HMIWI regulations to include beyond-the-floor standards in the future.

Response:

As we explained in the 2008 re-proposal, the identified beyond-the-floor add-on control measures we analyzed were not reasonable on a cost-effectiveness basis, especially in light of the significantly more stringent floor levels as compared to the 1997 rule's standards. We read the commenter's suggestion that we examine additional beyond-floor measures but without delaying final action on the re-proposal as recommending that we conduct the requested data gathering and analysis for those measures in a subsequent rulemaking action. A possible opportunity for that would be the next review of the rule under Sections 129(a)(5) and (h)(3). In the interim,

however, we have decided to revise the waste management plan provisions in §§ 60.35e and 60.55c to promote the segregation of chlorinated plastics and metals to the extent possible.

2. CO Emissions Limits

Comment:

One commenter argued that the proposed CO emissions limits will be unattainable by many applicable units, based on the emissions data provided in the docket. The commenter stated that the add-on controls evaluated by EPA do not reduce CO emissions, and that CO emissions can be a function of the feed material composition (which the commenter stated EPA did not evaluate). As a result, the commenter stated, HMIWI operators will have very little latitude or options to meet the proposed CO limits. Three other commenters stated that historical CO CEMS data from well-performing commercial HMIWI demonstrate that the proposed CO emissions limit is not achievable on a continuous basis and argued that the existing 40 ppmv emissions limit must be retained. The commenters further stated that the proposed CO standards must include a reasonable, extended averaging period (

e.g.,

24 hours) that accounts for the variability of the waste stream and waste characteristics. The commenters noted that the proposed standards are currently based on discrete 3-hour average data developed during performance test conditions, which they said do not account for the typical operational variability. According to the commenters, such snapshot data are also not representative of long-term continuous monitoring, placing facilities with CO CEMS at a competitive disadvantage with any revisions to the CO standard.

The same three commenters also stated that the proposed CO standard in combination with the 7 percent oxygen (O

) diluent correction factor will pose technological monitoring challenges to HMIWI that either choose or will be required to use CO CEMS, especially given the variability of HMIWI operations and waste feed streams. According to the commenters, costly monitoring systems (

e.g.,

dual range or ambient level monitors) will be needed, resulting in additional QA activities. The commenters further stated that the application of an O

correction factor to the measured CO concentration CEMS data may cause artificial exceedances of the CO emissions standard at higher O

operating scenarios.

Response:

Based on our review of CO emissions data for all HMIWI, we have found many HMIWI outperforming the existing 40 ppmv CO limit. We believe that the CO limits developed using the revised statistical approach are more representative of actual operation, and we estimate that a substantial percentage of HMIWI with their current controls will still be capable of meeting the revised limits (89 percent of large, 76 percent of medium, and 100 percent of small/small rural HMIWI, on average). Therefore, we disagree that the 40 ppmv CO limit must be retained.

Regarding the comment about the 3-hour average basis for the CO limit, it should be noted that the 2008 re-proposal included an amendment to § 60.56c allowing sources using CEMS to demonstrate compliance with the applicable emissions limit on a 24-hour block average, instead of a 12-hour rolling average (as specified in the 1997 final rule). This amended provision should address concerns about the ability of sources equipped with CEMS to demonstrate compliance with emissions limits on a continuous basis (as opposed to a 3-hour annual test) and would be consistent with past rulemakings for incineration units (

e.g.,

large and small MWCs).

Regarding the comment about the application of an O

correction factor to the CO CEMS data, it should be noted that correction to consistent standards (

e.g.,

percent O

) is necessary in order to compare to other units and to an emissions limit. Applying an O

correction factor to CO CEMS should only be a problem at O

levels greater than 15 percent. For comparison purposes, we reviewed the O

levels recorded in initial test reports, and found only about 7 of 57 HMIWI reported O

levels above 15 percent during at least one pollutant test run, and we estimate that 6 of those 7 with their current equipment will still meet the revised CO emissions limits, based on a comparison of the revised limits to the average CO concentrations for those HMIWI.

3. Opacity Limits

Comment:

Three commenters noted that EPA requested facility test data from 2003 through 2006 for all pollutants except opacity, even though annual opacity testing is required for all units. According to the commenters, if EPA wanted to review and revise the opacity limit pursuant to Section 129(a)(5), it should have requested opacity data and should have used those data in the re-establishment of the MACT standards. Instead, the commenters said, the proposed opacity limit was inappropriately established from a single continuous opacity monitoring system (COMS) located at a single HMIWI. The commenters argued that data from a single unit are insufficient to set an emissions limit that must be continuously achieved, and they said that EPA must seek additional monitoring data. The commenters also noted that compliance with the proposed opacity limit established by COMS is demonstrated using a different measurement methodology (Method 9).

The same three commenters, plus a fourth commenter, stated that the methodology that EPA used to establish the 2 percent opacity limit fails to account for actual opacity monitoring capabilities and normal operational variability, such as that included in PS-1 (40 CFR part 60, appendix B). According to the commenters, the inherent potential error of a COMS meeting PS-1 could greatly exceed the proposed opacity limit value. The fourth commenter argued that opacity under the worst foreseeable circumstances for the best-performing units would thus easily violate the MACT floor, which the commenter said would violate

Sierra Club.

167 F.3d at 665.

All four commenters noted that, similar to COMS accuracy, Method 9 calls for recording visual observations to the nearest 5 percent at 15-second intervals. The commenters stated that using a compliance method with inherent potential accuracy levels exceeding the proposed 2 percent opacity limit appears problematic.

Given the limitations of Method 9 and the variability of all the HMIWI subject to the revised opacity standard, the first three commenters recommended that EPA establish an opacity standard based on Method 9 data instead of COMS data from a single unit. All four commenters argued that the current 10 percent opacity limit is reasonable, and would allow conventional compliance determination methods to be used, accounting for their limitations.

Response:

The commenters' argument about how we established the proposed opacity limit is somewhat misleading. While we acknowledge that opacity data were inadvertently not included in the 2007-08 test data request, we already had opacity data for nearly 90 percent of all HMIWI from their initial compliance tests, and our initial opacity MACT floor analysis was based on the best-performing 12 percent of sources for opacity. As we stated in the preamble to the December 1, 2008 re-proposal (73 FR 72983), based on the opacity averages alone, without any accounting for variability, the MACT floor for opacity for existing and new units would have been 0 percent. We tried to account for variability by looking at the single highest opacity reading for HMIWI in the MACT floor

for PM, based on opacity being an appropriate surrogate for PM. We based our MACT floor opacity limit on the single highest COMS reading (1.1 percent) for one of the HMIWI in the MACT floor for PM. Because we commonly set opacity standards based on whole numbers and could not round down without risking having the MACT floor unit not meet the standard, we rounded up and proposed an opacity limit of 2 percent for both new and existing HMIWI. However, we now believe this analysis was incomplete. The analysis did not account for two other HMIWI in the MACT floor for PM that could more effectively account for variability for opacity. The maximum opacity averages for these two HMIWI are 5.87 and 4.17 percent. (

See

2008 memorandum entitled “Documentation of HMIWI Test Data Database,” which is included in the docket.) The opacity data for these two HMIWI were measured using Method 9. Using the same approach that we used at re-proposal, we are establishing an opacity limit of 6 percent, by rounding up the highest opacity average of 5.87 percent to the nearest whole number.

Regarding the commenters' arguments that the inherent potential error of a COMS meeting PS-1 could exceed the proposed opacity limits, the potential error (about 4 percent opacity at the highest) is not the same as expected error (more on the order of 0.5 percent). Nonetheless, the increase in the opacity limit to 6 percent should address the commenters' concerns on this issue.

We disagree with the commenters' argument that a 10 percent opacity limit be used to allow conventional compliance determination methods. While opacity is read in 5 percent increments, average opacity can be any number above 0. Method 9 values are averages of 24 readings, which can include readings of 0 and an occasional 5 or 10 percent.

Regarding the commenters' argument that only Method 9 data should be used to establish the opacity standard because that is the measurement method that would be used to demonstrate compliance, the commenters' argument is moot, since the revised opacity standard is now based on Method 9 results.

4. Percent Reduction Limits

Comment:

One commenter agreed with EPA's proposed elimination of percent reduction alternatives. According to the commenter, EPA correctly noted that standards based only on control technology performance do not reflect the effects of non-technology factors and, therefore, do not reflect the best units' actual performance. Therefore, the commenter said, allowing units the option to meet these percent reduction limits instead of emissions standards contravenes Section 129, and EPA appropriately proposed to delete the percent reduction limits.

Three other commenters argued that the percent reduction compliance option that was available in the 1997 rule and in the 2007 proposed rule should be re-evaluated and retained for commercial HMIWI, since the ability for such units to reduce emissions is due almost exclusively to the effectiveness of the control equipment (and not waste segregation). According to the commenters, commercial HMIWI facilities, unlike captive units, cannot practically control the waste that is put in the containers they process, and applicable regulations from the U.S. Occupational Safety and Health Administration (OSHA) preclude them from practicing waste segregation at the time of treatment. Thus, the commenters noted, they experience extreme variability during stack tests (especially for volatile metals Cd, Pb, and Hg) and will experience higher inlet concentrations than captive units; since they operate at the same control efficiency, they will exhibit higher stack emissions. The commenters stated that the percent reduction option is a better assessment of the performance of the control system for commercial units.

Response:

We have decided not to include percent reduction limits in the final rule. In addition to the reasons we provided in the re-proposal, while commercial HMIWI facilities face greater challenges in controlling the waste they receive, compared to “captive” units, they are nonethele

This text is long and has been trimmed here. Open the source document for the complete record.

This is a copy of a public record, reproduced as it was published. It is not legal advice, and it may not be the version a court would rely on. Check the official source before you cite it.