Acid Rain Program; Continuous Emission Monitoring Rule Revisions

Related Material

Other Related Documents
Note: EPA no longer updates this information, but it may be useful as a reference or resource.
 


[Federal Register: May 21, 1998 (Volume 63, Number 98)]
[Proposed Rules]               
[Page 28031-28080]
From the Federal Register Online via GPO Access [wais.access.gpo.gov]
[DOCID:fr21my98-42]
 

[[Page 28031]]

_______________________________________________________________________

Part II





Environmental Protection Agency





_______________________________________________________________________



40 CFR Parts 72 and 75



Acid Rain Program; Continuous Emission Monitoring Rule Revisions; Acid 
Rain Program: Determinations Under EPA Study of Bias Test and Relative 
Accuracy and Availability Analysis; Proposed Rules


[[Page 28032]]



ENVIRONMENTAL PROTECTION AGENCY

40 CFR Parts 72 and 75

[FRL-6007-8]
RIN 2060-AG46

 
Acid Rain Program; Continuous Emission Monitoring Rule Revisions

AGENCY: Environmental Protection Agency (EPA).

ACTION: Proposed rule.

-----------------------------------------------------------------------

SUMMARY: Title IV of the Clean Air Act (CAA or the Act), as amended by 
the Clean Air Act Amendments of 1990, authorizes the Environmental 
Protection Agency (EPA or Agency) to establish the Acid Rain Program. 
The Acid Rain Program and the provisions in this proposed rule benefit 
the environment by preventing the serious, adverse effects of acidic 
deposition on natural resources, ecosystems, materials, visibility, and 
public health. The program does this by setting emissions limitations 
to reduce the acidic deposition precursor emissions of sulfur dioxide 
and nitrogen oxides. On January 11, 1993, the Agency promulgated final 
rules, including the final continuous emission monitoring (CEM) rule, 
under title IV. On May 17, 1995, the Agency published direct final and 
interim rules to make the implementation of the CEM rule simpler. 
Subsequently, on November 20, 1996, the Agency published a final rule 
in response to public comments received on the direct final and interim 
rules.
    These proposed revisions to the CEM rule would make a number of 
further minor changes to make the implementation of the CEM rule 
simpler, more streamlined, and more efficient for both EPA and the 
facilities affected by the rule. Furthermore, the proposed revisions 
would provide reduced monitoring burdens for affected facility units 
with low mass emissions. In addition, the proposed revisions would 
establish quality assurance requirements for moisture monitoring 
systems and add a new flow monitor quality assurance test to assure the 
accuracy of data reported from these types of monitoring systems. 
Finally, the proposed revisions would create a new monitoring option, 
the F-factor/fuel flow method, for certain units.

DATES: Comments. All public comments must be received on or before July 
20, 1998.
    Public Hearing. Anyone requesting a public hearing must contact EPA 
no later than May 31, 1998. If a hearing is held, it will take place 
June 8, 1998, beginning at 10:00 a.m.

ADDRESSES: Comments. Comments must be mailed (in duplicate if possible) 
to: EPA Air Docket (6102), Attention: Docket No. A-97-35, Room M-1500, 
Waterside Mall, 401 M Street, SW, Washington, DC 20460.
    Public Hearing. If a public hearing is requested, it will be held 
at the Environmental Protection Agency, 401 M Street, SW, Washington, 
DC 20460, in the Education Center Auditorium. Refer to the Acid Rain 
homepage at www.epa.gov/acidrain for more information or to determine 
if a public hearing has been requested and will be held.
    Docket. Docket No. A-97-35, containing supporting information used 
to develop the proposal is available for public inspection and copying 
from 8:00 a.m. to 5:30 p.m., Monday through Friday, excluding legal 
holidays, at EPA's Air Docket Section at the above address.

FOR FURTHER INFORMATION CONTACT: Jennifer Macedonia, Acid Rain Division 
(6204J), U.S. Environmental Protection Agency, 401 M Street, SW, 
Washington, DC 20460, telephone number (202) 564-9123 or the Acid Rain 
Hotline at (202) 564-9620. Electronic copies of this notice and 
technical support documents can be accessed through the Acid Rain 
Division website at http://www.epa.gov/acidrain.

SUPPLEMENTARY INFORMATION: The contents of the preamble are listed in 
the following outline:

I. Regulated Entities
II. Background and Summary of the Proposed Rule
III. Detailed Discussion of Proposed Revisions
    A. Use of Projections in the Definitions of Gas-fired, Oil-
fired, and Peaking Unit
    B. Wording Correction of the Applicability Provisions in Part 72
    C. Low Mass Emissions Excepted Methodology
    1. Applicability Criteria
    2. Method for Determining Emissions
    3. Cutoff Limit for Applicability
    4. Continuing Applicability Criteria
    5. Reduced Monitoring and Quality Assurance Requirements
    6. Reduced Reporting Requirements
    D. Quality Assurance Requirements for Moisture Monitoring 
Systems
    E. Certification/Recertification Procedural Changes
    1. Initial Certification versus Recertification
    2. Disapproval of an Incomplete Application
    3. Submittal Requirements for Certification and Recertification 
Applications
    4. Decertification Applicability
    5. Recertification Test Notice
    6. Monitoring Plans
    7. Submittal Requirements for Petitions and Other Correspondence
    F. Substitute Data
    1. Missing Data Procedures for CO2 and Heat Input
    2. Prohibition Against Low Monitor Data Availability
    G. General Authority to Grant Petitions Under Part 75
    H. NO_X Mass Monitoring Provisions for Adoption by 
NO_X Mass Reduction Programs
    I. Span and Range Requirements
    1. Maximum Potential Values
    2. Maximum Expected SO2 and NO_X 
Concentrations
    3. Span and Range Values
    4. Dual Span and Range Requirements for SO2 and 
NO_X
    5. Adjustment of Span and Range
    J. Quality Assurance/Quality Control (QA/QC) Program
    1. QA/QC Plan
    2. Flow Monitor Polynomial Coefficient
    K. Calibration Gas Concentration for Daily Calibration Error 
Tests
    L. Linearity Test Requirements
    1. Unit Operation During Linearity Tests
    2. Linearity Test Frequency
    3. Linearity Test Method
    4. Exemptions
    M. Flow-to-Load Test
    N. RATA and Bias Test Requirements
    1. RATA Frequency
    2. RATA Load Levels
    3. Flow Monitor Bias Adjustment Factors
    4. Number of RATA Attempts
    5. Concurrent SO2 and Flow RATAs
    6. SO2 RATA Exemptions and Reduced Requirements
    7. QA Provisions for SO2 Monitors, for Natural Gas 
Firing or Equivalent
    8. General RATA Test Procedures
    9. Reference Method Testing Issues
    10. Alternative Relative Accuracy Specifications and 
Specifications for Low-Emitters
    11. Bias Adjustment Factors for Low-Emitters
    12. Clarification of Diluent Monitor Certification Requirements
    13. Daily Calibration Requirements for Redundant Backup Monitors
    14. Daily Performance Specification and Control Limits for Low-
Span DP Flow Monitors
    O. CEM Data Validation
    1. Recalibration and Adjustment of CEMS
    2. Linearity Tests
    3. RATAs
    4. Recertification of Gas and Flow Monitors
    5. Recertification and QA
    6. Data from Non-Redundant Backup Monitors
    7. Missed QA Test Deadlines
    P. Appendix D
    1. Pipeline Natural Gas Definitions
    2. Fuel Sampling
    3. Sulfur, Density, and Gross Calorific Value Used in 
Calculations
    4. Missing Data Procedures for Sulfur Content, Density, and 
Gross Calorific Value
    5. Installation of Fuel Flowmeters for Recirculation
    6. Fuel Flowmeter Testing

[[Page 28033]]

    7. Use of Uncertified Commercial Gas Flowmeter
    Q. Appendix G
    1. Use of ASTM D5373-93 for Determining the Carbon Content of 
Coal
    2. Changes to Fuel Sampling Frequency
    3. Addition of Missing Data Procedures for Fuel Analytical Data
    R. Reporting Issues
    1. Partial Unit Operating Hours and Emission and Fuel Flow Rates
    2. Use of Bias-Adjusted Flow Rates in Heat Input Calculations
    3. Removing the Restriction of Using the Diluent Cap Only for 
Start-up
    4. Complex Stacks--General Issues
    5. Complex Stacks--Heat Input at Common Stacks
    6. Start-up Reporting--Units Shutdown Over the Compliance 
Deadline
    7. Start-up Reporting--New Units
    8. Recordkeeping and Reporting Provisions
    9. Electronic Transfer of Quarterly Reports
    S. Revised Traceability Protocol for Calibration Gases
    T. Appendix I--New Optional Stack Flow Monitoring Methodology
    U. The Use of Predictive Emissions Modeling Systems (PEMS)
IV. Administrative Requirements
    A. Public Hearing
    B. Public Docket
    C. Executive Order 12866
    D. Unfunded Mandate Reform Act
    E. Paperwork Reduction Act
    F. Regulatory Flexibility Act
    G. National Technology Transfer and Advancement Act

I. Regulated Entities

    Entities potentially regulated by this action are fossil fuel-fired 
boilers and turbines that serve generators producing electricity, 
generate steam, or cogenerate electricity and steam. While part 75 
primarily regulates the electric utility industry, today's proposal 
could potentially affect other industries. The proposal includes 
NO_X mass provisions for the purpose of serving as a model 
which could be adopted by a state, tribal, or federal NO_X 
mass reduction program covering the electric utility and other 
industries. Regulated categories and entities include:

------------------------------------------------------------------------
                                                Examples of regulated   
                 Category                             entities          
------------------------------------------------------------------------
Industry..................................  Electric service providers, 
                                             boilers and turbines from a
                                             wide range of industries.  
------------------------------------------------------------------------

This table is not intended to be exhaustive, but rather provides a 
guide for readers regarding entities likely to be regulated by this 
action. This table lists the types of entities which EPA is now aware 
could potentially be regulated by this action. Other types of entities 
not listed in the table could also be regulated. To determine whether 
your facility, company, business, organization, etc., is regulated by 
this action, you should carefully examine the applicability criteria in 
Secs. 72.6, 72.7, and 72.8 of title 40 of the Code of Federal 
Regulations. If you have questions regarding the applicability of this 
action to a particular entity, consult the person listed in the 
preceding FOR FURTHER INFORMATION CONTACT section of this preamble.

II. Background and Summary of the Proposed Rule

    Title IV of the Act requires EPA to establish an Acid Rain Program 
to reduce the adverse effects of acidic deposition. On January 11, 
1993, the Agency promulgated final rules implementing the program, 
including the CEM rule (58 FR 3590-3766). Technical corrections were 
published on June 23, 1993 (58 FR 34126) and July 30, 1993 (58 FR 
40746-40752). A notice of direct final rulemaking and of interim final 
rulemaking further amending the regulations was published on May 17, 
1995 (60 FR 26510 and 60 FR 26560). Subsequently, on November 20, 1996, 
a final rule was published in response to public comments received on 
the direct final and interim rules (61 FR 59142-59166) .
    The issues addressed by this proposed rule are: (1) revised 
definitions of gas-fired, oil-fired, and peaking unit to allow for 
changes in unit fuel usage and/or operation; (2) a minor wording 
correction of the applicability provisions in Part 72; (3) new excepted 
methodologies for units with low mass emissions; (4) new QA/QC 
requirements for moisture monitoring systems; (5) clarifying changes to 
the certification and recertification process; (6) substitute data 
requirements for CO2 and heat input, as well as a 
prohibition against low data availability; (7) clarifying revisions to 
the petition provisions for alternatives to part 75 requirements; (8) 
NO_X mass monitoring provisions provided as a model for 
adoption by state, tribal, or federal NO_X mass reduction 
programs; (9) clarifying changes to span and range requirements; (10) 
clarifying revisions to general QA/QC requirements; (11) calibration 
gas concentrations for daily calibration error tests; (12) linearity 
test requirements; (13) a new flow-to-load QA test for flow monitors; 
(14) reductions in and/or clarifications to the relative accuracy test 
audit (RATA) and bias test requirements; (15) clarifying revisions to 
the procedures for CEM data validation; (16) clarifying revisions to 
the SO2 emissions data protocol for gas-fired and oil-fired 
units (Appendix D); (17) determining CO2 emissions (Appendix 
G, sections 2.1 and 5); (18) recordkeeping and reporting changes to 
reflect the proposed revisions; (19) a revised traceability protocol 
(Appendix H); and (20) a new optional F-factor/fuel flow method 
(Appendix I). In addition, the preamble also includes a discussion on 
potential provisions to allow for the use of predictive emissions 
modeling systems (PEMS) as an alternative to CEMS for certain units.
    Many of the changes proposed today are minor technical revisions 
based on comments received from utilities following the initial 
implementation of part 75. Based on experience gained in the early 
years of the program, utilities have developed a number of suggestions 
that EPA believes would simplify and streamline the monitoring process 
without sacrificing data quality. In addition, the Agency is proposing 
to reduce the monitoring requirements for units with low mass emissions 
to reduce burdens on those types of units and to add new monitoring 
options for some units. The Agency has also proposed new quality 
assurance requirements based on gaps identified by EPA during 
evaluation of the initial implementation of part 75. Finally, several 
minor technical changes are also proposed in order to maintain 
uniformity within the rule itself and to clarify various provisions.

III. Detailed Discussion of Proposed Revisions

A. Use of Projections in the Definitions of Gas-Fired, Oil-Fired, and 
Peaking Unit

Background
    Section 72.2 of the January 11, 1993 rule provides definitions for 
the terms ``gas-fired,'' ``oil-fired,'' and ``peaking unit.'' Each 
definition provides a limit on the fuel usage or capacity factor 
averaged over a three year period, as well as an individual limit on 
each of the three years, in order to qualify under the definition. The 
May 17, 1995 revisions to part 75 amended those definitions by adding 
provisions for how a unit would initially qualify to meet the 
definition. Each definition provides for the case where a unit has 
three years of historical data demonstrating qualification, as well as 
the case where a unit does not have data for one or more of the three 
previous years (e.g., a new unit or a unit that has been in an extended 
shutdown). In addition, the gas-fired definition provides for the case 
where a unit's fuel usage is projected to change on or before January 
1, 1995 and the peaking unit definition provides for the case where a 
unit's capacity factor is projected to change on or before the 
certification deadline (either 1995 or 1996) for NO_X

[[Page 28034]]

monitoring in Sec. 75.4. In each case where historical data does not 
exist or is not representative based on projected change, the amended 
definitions set provisions for allowing projections of unit operation 
to be used in place of historical data in order to meet the criteria of 
the respective definition. However, none of the three definitions 
provides for the case where a unit's fuel usage or capacity factor is 
expected to change after initial classification.
    Under the existing rule, the importance of determining whether a 
unit qualifies under the definitions of gas-fired, oil-fired, and 
peaking unit, centers on the differences in regulatory requirements and 
options for different classifications of units. For example, under 
Sec. 75.11(d)(2), a unit that qualifies as gas-fired or oil-fired has 
an additional option for monitoring SO2 emissions using the 
excepted protocol of Appendix D, in lieu of an SO2 CEMS and 
flow monitor. Additionally, under Sec. 75.14(c), a unit that qualifies 
as gas-fired is exempt from opacity monitoring, and, under section 2.3 
of Appendix G to part 75, a gas-fired unit has an additional option for 
determining CO2 mass emissions in lieu of a CO2 
CEMS or using carbon sampling in conjunction with a fuel flowmeter. 
Qualifying under the definition of peaking unit also has the advantage 
of allowing additional regulatory options. For example, a peaking unit 
has the option of monitoring NO_X emission rate using the 
excepted protocol under Appendix E, in lieu of a NO_X CEMS. 
Further, under section 2.3.1 of Appendix B to part 75, a peaking unit 
is required to perform annual quality assurance flow monitor RATAs at a 
single load level instead of at three load levels.
    Utility representatives have contacted EPA for guidance about how a 
change in the manner of operation of the unit after certification and 
initial classification of the unit affects the status of the unit with 
respect to the definitions of gas-fired, oil-fired, and peaking unit. 
For example, a utility representative contacted the Agency about a unit 
designed to burn gas and/or oil that historically had burned primarily 
oil and was classified as an oil-fired unit. The utility had decided to 
switch from oil to burn almost entirely gas at the unit and asked 
whether it was necessary to wait three years after the switch to gas in 
order to gather three years of historical data, to qualify for the 
additional regulatory options available only for gas-fired units. The 
utility requested permission to use projections of fuel usage certified 
by the designated representative, to demonstrate that the unit would 
meet the gas-fired definition after the switch to gas, so that the unit 
could be exempt from opacity monitoring and qualify to use equation G-4 
to determine CO2 mass emissions. The existing rule would 
require such a unit to wait three years after the change in operation 
in order to qualify as gas-fired. Based on EPA's experience of 
implementing the provisions of Parts 72 and 75, the definitions of the 
terms gas-fired, oil-fired, and peaking unit are not sufficiently 
detailed or flexible to address situations where a permanent change in 
the manner of operation after the initial classification (i.e, capacity 
factor or fuel usage) affects the gas-fired, oil-fired, or peaking unit 
status.
Discussion of Proposed Changes
    Today's proposal would amend the definitions of the terms gas-
fired, oil-fired, and peaking unit, to add provisions for an existing 
unit that does not presently qualify under the definition but that 
experiences a permanent change in operation (i.e., fuel usage for the 
gas-and oil-fired definitions and capacity factor for the peaking unit 
definition).
    For the definition of gas-fired, the proposed revisions would allow 
an existing unit to qualify under the definition if the designated 
representative submits a minimum of 720 hours of unit operating data 
demonstrating that the unit meets the percentage criteria of a gas-
fired unit (i.e., no less than 90.0 percent of the unit's heat input 
from the combustion of gaseous fuels with a total sulfur content no 
greater than natural gas and the remaining heat input from the 
combustion of fuel oil), accompanied by a certification statement from 
the designated representative. The designated representative statement 
would certify that the changed pattern of fuel usage, represented in 
the 720 hours of data, is considered permanent and is projected to 
continue for the foreseeable future.
    The proposed definition of oil-fired unit would simplify the 
provisions for qualification, for purposes of part 75. The proposed 
definition would simply require that a unit burn only fuel oil and 
gaseous fuels with a total sulfur content no greater than natural gas 
and that the unit does not meet the definition of gas-fired, in order 
to qualify as oil-fired. With this simplification, a unit could qualify 
under any of the following circumstances: (1) a new unit projected to 
burn only fuel oil and gaseous fuels with a sulfur content no greater 
than natural gas but projected to burn too much oil to qualify as gas-
fired; (2) an existing gas-fired unit, which burns only fuel oil and 
natural gas, but which exceeds the gas-fired annual limit of 15 percent 
of the annual heat input from fuel oil; and (3) an existing coal-fired 
unit that is converted to only burn fuel oil and/or gas but which 
projects it will burn too much oil to qualify as gas-fired.
    The proposed definition of peaking unit would allow an existing 
unit whose capacity factor is projected to change, to qualify as a 
peaking unit if the designated representative submits a demonstration 
satisfactory to the Administrator that the unit will qualify as a 
peaking unit, using the three calendar years beginning with the first 
full year following the change in the unit's capacity factor as the 
three year period. This demonstration would need to show that the 
unit's capacity factor in the year following the permanent change in 
operation did not exceed 10.0 percent and that the projected average 
annual capacity factor for the unit in the three year period and the 
projected capacity for each of the two individual projected years will 
meet the definition of a peaking unit.
    Additionally, under today's proposal, the gas-fired definition 
would be revised to clarify the requirements as they apply for the 
purposes of part 75 versus the requirements for the purposes of all 
other Parts under the Acid Rain Program. This proposed revision is 
merely editorial and would not change the intent of the existing 
regulation.
Rationale
    The Agency proposes to allow projections of fuel usage or capacity 
factor in conjunction with some actual data to be used for the purpose 
of meeting the criteria of the gas- or oil-fired or peaking unit 
definitions, respectively. The Agency believes it is unnecessary to 
require three years to pass before a unit that the designated 
representative certifies has permanently changed its manner of 
operation is allowed to utilize the additional regulatory options 
allowed for units meeting the definitions of gas-fired, oil-fired, and 
peaking unit. The Agency believes it is sufficient to require the 
designated representative to submit representative data that the unit 
would qualify under the definition following the permanent change in 
operation or fuel usage (i.e., 720 hours for the gas-fired definition 
and a full year for the peaking unit definition) and to certify that 
the change in fuel usage or capacity factor is considered permanent and 
that the unit is expected to continue to meet the definition of gas-
fired, oil-fired, or peaking unit, as applicable, into the foreseeable 
future.
    Under the existing rule, the peaking unit definition does provide 
for the

[[Page 28035]]

situation where a unit's operation is projected to change and the unit 
will meet the peaking unit definition with those projections. However, 
this provision is limited to the case where a unit's operation has 
changed by the certification deadline for NO_X monitoring. 
The existing rule does not provide for the scenario where a change to 
the unit's operation after the certification deadline would affect the 
peaking unit status and where the designated representative might want 
to take advantage of regulatory options that are available under this 
new status.
    EPA believes that it is appropriate to allow a unit to use the 
regulatory options that are only allowed for peaking units, if a unit's 
operation permanently changes such that it meets the capacity factor 
definition with one year of actual data and two years of projections. 
If the projections are incorrect, the unit will lose its peaking unit 
status and will not be able to use projections again to qualify.
    Similarly, under the existing rule, the gas-fired definition does 
provide for the situation where an existing unit that does not qualify 
under the gas-fired definition experiences a change in operations or 
fuel usage that would result in the unit qualifying as gas-fired in 
future years. However, this provision is limited to the case where a 
unit's operation has changed by the certification deadline for 
SO2 and opacity monitoring, from 1995 through 1997. The 
existing rule does not provide for the scenario where a change to the 
unit's fuel usage after the certification deadline would affect the 
gas-fired status and that the designated representative might want to 
take advantage of regulatory options that are available under this new 
status.
    However, EPA believes that it is appropriate to allow a unit to use 
the regulatory options that are only allowed for gas-fired units, if a 
unit's fuel usage permanently changes such that it meets the gas-fired 
definition with 720 hours of actual data and projections of fuel usage 
to make up the remainder of the three year period. If the projections 
are incorrect, the unit will lose its gas-fired status and will not be 
able to use projections again to qualify.

B. Wording Correction of the Applicability Provisions in Part 72

Background
    Section 72.6(b)(1) currently includes, in the list of types of 
units that are unaffected units under the Acid Rain Program, ``[a] 
simple combustion turbine that commenced operation before November 15, 
1990.'' 40 CFR 72.6(b)(1). Title IV actually provides, through 
statutory definitions and provisions setting emission limitations, that 
a simple combustion turbine that commenced commercial operation before 
the enactment of title IV, i.e., November 15, 1990, is an unaffected 
unit. A simple combustion turbine commencing commercial operation on or 
after November 15, 1990 is an affected unit (unless it is exempt under 
some other provision, e.g., the new units exemption under Sec. 72.7).
    To begin, the definition of ``existing unit'' in section 402(8) of 
the Act excludes existing simple combustion turbines (i.e., those that 
commenced commercial operation prior to November 15, 1990) and so 
excludes them from being affected units subject to an SO2 
emission limitation under section 405(a)(1). As stated in that section 
402(8):

``existing unit'' means a unit * * * that commenced commercial 
operation before the date of enactment of the Clean Air Act 
Amendments of 1990 [i.e., November 15, 1990] * * * For purposes of 
this title, existing units shall not include simple combustion 
turbines * * * 42 U.S.C. 7651a(8).

In contrast, the statutory definition of ``new unit'' does not exclude 
any new simple combustion turbines, and under section 403(e), all new 
utility units are affected units subject to an SO2 emission 
limitation. As stated in section 402(10):

``new unit'' means a unit that commences commercial operation on or 
after the date of enactment of the Clean Air Act Amendments of 1990 
[i.e., November 15, 1990]. 42 U.S.C. 7651a(10).

A unit that commences commercial operation after November 15, 1990, and 
so does not meet the definition of ``existing unit'', is therefore a 
new unit and an affected unit subject to Acid Rain Program 
requirements.
    While Sec. 72.6(b)(1) states that a simple combustion turbine that 
``commenced operation'' before November 15, 1990 is not an affected 
unit, EPA interprets this provision, consistent with the Act, to refer 
to commencement of commercial operation. However, in order to remove 
any ambiguity and any possibility of erroneous application of the 
statutory exemption for simple combustion turbines, EPA believes that 
the regulatory provision should be corrected.
Discussion of Proposed Changes
    Today's proposal would revise the existing Sec. 72.6(b)(1) in order 
to make it consistent with title IV of the Act. EPA proposes to revise 
the language of the provision to refer expressly to ``commercial 
operation,'' rather than simply ``operation,'' of a simple combustion 
turbine.
Rationale
    EPA notes that the existing Sec. 72.6(b)(1) was not intended to 
deviate from the provisions in the Act concerning simple combustion 
turbines. In proposing the applicability provisions that were finalized 
(with changes) as Sec. 72.6, EPA explained that:

simple combustion turbines would be subject to Acid Rain Program 
requirements in Phase II (as new units) if such units commenced 
commercial operation on or after November 15, 1990, because the 
statutory exemption for simple combustion turbines is only 
applicable to existing units. 56 FR 63002, 63008 (1991).

In noting that new simple combustion turbines are affected units, EPA 
requested comment on whether a ``de minimis exclusion should be 
included in the final rule'' for ``very small units'' from the Acid 
Rain Program. Id. In response to comments supporting an exemption for 
simple combustion turbines and other units, EPA established in the 
final rule an exemption for new units (including new simple combustion 
turbines) serving generators with total capacity of 25 MWe or less. 58 
FR 3590, 3593-4 (1993); Response to Comment at P-22 and P-23 (1993). In 
the final rule preamble, EPA did not indicate any intention to make any 
other changes concerning the applicability of the Acid Rain Program to 
new simple combustion turbines.

C. Low Mass Emissions Excepted Methodology

Background
    In the January 11, 1993 Acid Rain permitting rule, EPA provided for 
a conditional exemption from the emissions reduction, permitting, and 
emissions monitoring requirements of the Acid Rain Program for new 
units having a nameplate capacity of 25 MWe or less that burn fuels 
with a sulfur content no greater than 0.05 percent by weight, because 
of the de minimis nature of their emissions (see 58 FR 3593-94 and 
3645-46). Moreover, in the January 11, 1993 monitoring rule, EPA 
allowed gas-fired and oil-fired peaking units to use the provisions of 
Appendix E, instead of CEMS, to determine the NO_X emission 
rate, stating that this was a de minimis exception. EPA allowed this 
exception from the requirements of section 412 of the Clean Air Act 
because the NO_X emissions from these units would be 
extremely low, both

[[Page 28036]]

collectively and individually, and because the cost of measuring a ton 
of NO_X with CEMS could be several hundred dollars per ton of 
NO_X monitored (see 58 FR 3644-45). One utility wrote to the 
Agency, suggesting that the Agency consider further regulatory relief 
for other units with extremely low emissions that do not fall under the 
categories of small new units burning fuels with a sulfur content less 
than or equal to 0.05 percent by weight or gas-fired and oil-fired 
peaking units (see Docket A-97-35, Item II-D-31). The utility 
specifically suggested that the Agency consider an exemption, the 
ability to use Appendix E, or some other simplified methods which are 
more cost effective.
    In the process of implementing part 75, other utilities also have 
suggested to EPA that it provide regulatory relief to low mass emitting 
units (see Docket A-97-35, Items II-D-29, II-E-25). These units might 
be low mass emitting because they use a clean fuel, such as natural 
gas, and/or because they operate relatively infrequently. Some 
utilities stated that they spend a great deal of time reviewing the 
emissions data when preparing quarterly reports for these units. Others 
indicated that it would be important to reduce monitoring and quality 
assurance (QA) requirements in order to save time and money currently 
devoted to units with minimal emissions (see Docket A-97-35, Item II-E-
25).
Discussion of Proposed Changes
    Today's proposal would incorporate optional reduced monitoring, 
quality assurance, and reporting requirements into part 75 for units 
that burn only natural gas or fuel oil, emit no more than 25 tons of 
SO2 and no more than 25 tons of NO_X annually, and 
have calculated annual SO2 and NO_X emissions 
(reflecting their potential emissions during actual operation) that do 
not exceed such limits.
    A unit would initially qualify for the reduced requirements by 
demonstrating to the Administrator's satisfaction that the unit meets 
the applicability criteria in proposed Sec. 75.19(a). Proposed 
Sec. 75.19(a) would require facilities to submit historical actual (or 
projections, as described below) and calculated emissions data from the 
previous three calendar years demonstrating that a unit falls below the 
25-ton cutoffs for SO2 and NO_X. The calculated 
emissions data for the previous three calendar years would be 
determined by applying the emission factors and maximum rated hourly 
heat input, under Sec. 75.19(c), to the hours of operation and fuel 
burned during the previous three calendar years. The data demonstrating 
that a unit meets the applicability requirements of Sec. 75.19(a) would 
be submitted in a certification application for approval by the 
Administrator to use the low mass emissions excepted methodology. The 
Agency requests comments on whether a unit that exceeded the 25-ton 
emissions cutoff for a part of the previous three years, but that has 
made a permanent change in the operation of the unit such that it would 
expect to meet the applicability criteria based on projections of 
future operation, should be allowed to use the excepted methodology.
    For units that lack historical data for one or more of the previous 
three calendar years (including new units that lack any historical 
data), proposed Sec. 75.19(a) would require the facility to provide (1) 
any historical emissions and operating data, beginning with the unit's 
first calendar year of commercial operation, that demonstrates that the 
unit falls under the 25-ton cutoffs for SO2 and 
NO_X, both with actual emissions and with calculated 
emissions using the proposed methodology, as described above; and (2) a 
demonstration satisfactory to the Administrator that the unit will 
continue to emit below the tonnage cutoffs (e.g., for a new unit, 
applying the emission rates and hourly heat input, under Sec. 75.19(c), 
to a projection of annual operation and fuel usage to determine the 
projected mass emissions).
    For units with historical actual (or projections, as described 
above) emissions and calculated emissions falling below the tonnage 
cutoffs, facilities would be allowed to use the optional methodology in 
proposed Sec. 75.19(c) in lieu of either CEMS or, where applicable, in 
lieu of the excepted methods under Appendix D, E, or G for the purpose 
of determining and reporting heat input, NO_X emission rate, 
and NO_X, SO2, and CO2 mass emissions. 
Under the optional methodology in proposed Sec. 75.19(c), a facility 
would calculate and report hourly SO2 and CO2 
mass emissions based on the unit's maximum rated hourly heat input and 
the appropriate emission factor, defined in Sec. 75.19(c), Tables 1a 
and 1c, for the fuel burned that hour. Similarly, a facility would 
calculate and report hourly NO_X mass emissions as the 
product of the maximum rated hourly heat input and the appropriate fuel 
and boiler type NO_X emission rate located in proposed Table 
1b. The facility would no longer be required to keep monitoring 
equipment installed on low mass emissions units, nor would it be 
required to meet the quality assurance test requirements or QA/QC 
program requirements of Appendix B to part 75. Moreover, emissions 
reporting requirements would be reduced by requiring only that the 
facility report the unit's hourly mass emissions of SO2, 
CO2, and NO_X, the unit's NO_X emission 
rate, and the fuel type burned for each hour of operation, and report 
the quarterly total and year-to-date cumulative mass emissions, heat 
input, and operating time, in addition to the unit's quarterly average 
and year-to-date average NO_X emission rate for each quarter. 
Facilities would continue to be required to monitor, record, and report 
opacity data for oil-fired units, as specified under Secs. 75.14(a), 
75.57(f), and 75.64(a)(iii) respectively. Under Sec. 75.14(c) and (d), 
however, gas-fired, diesel-fired, and dual-fuel reciprocating engine 
units would continue to be exempt from opacity monitoring requirements.
    If an initially qualified unit were subsequently to burn fuel other 
than natural gas or fuel oil, the unit would be disqualified from using 
the reduced requirements starting the first date on which the fuel 
(other than natural gas or fuel oil) was burned.
    In addition, if an initially qualified unit were to subsequently 
exceed the 25-ton cutoff for either SO2 or NO_X 
while using the proposed methodology, the facility would no longer be 
allowed to use the reduced requirements in proposed Sec. 75.19(c) for 
determining the affected unit's heat input, NO_X emission 
rate, or SO2, CO2, and NO_X mass 
emissions. Proposed Sec. 75.19(b) would allow the facility two quarters 
from the end of the quarter in which the exceedance of the relevant 25-
ton cutoff(s) occurred to install, certify, and report SO2, 
CO2, and NO_X data from a monitoring system that 
meets the requirements of Secs. 75.11, 75.12, and 75.13, respectively.
Rationale
    In addressing concerns from utilities about the cost of monitoring, 
quality assurance testing, and reporting emissions from low-emitting 
sources, EPA considered how to establish reduced requirements. 
Utilities have indicated to EPA that it would be more helpful for the 
Agency to reduce testing requirements for monitoring equipment than it 
would be to reduce only reporting requirements (see Docket A-97-35, 
Item II-E-25). The Agency considered whether a reduction in monitoring 
or reporting requirements might have unintended adverse consequences 
for the environment. In order to minimize this possibility, but still 
make the program more cost

[[Page 28037]]

effective for facilities, the Agency is proposing to allow an exception 
from full monitoring and reporting requirements for low mass emitting 
units. In proposing these reduced requirements, the Agency is 
exercising its discretion to allow de minimis exceptions from statutory 
requirements in administering the Clean Air Act (see, e.g., Alabama 
Power Co. v. Costle, 636 F.2d 323, 360-61 (D.C. Cir. 1979); and 58 FR 
3593-94 and 3645-46). The Agency, in exercising its discretion, 
believes that in light of the de minimis aggregate amount of emissions 
from low-emitting units as a group, little or no environmental benefit 
would be derived from continuing to require the additional accuracy of 
monitoring data from low-emitting units under the existing regulations, 
if such units are subjected instead to the proposed optional 
requirements. EPA also notes that any such benefit would be greatly 
outweighed by the cost of providing the more accurate data.
    In drafting today's proposal, the Agency considered six relevant 
questions: (1) What parameters should the applicability criteria be 
based on? (2) How should estimated emissions be calculated? (3) What 
cutoff emission level should be used to determine applicability of the 
reduced requirements? (4) What should the on-going applicability 
requirements be? (5) What should the reduced monitoring and quality 
assurance requirements be for these units? and (6) What should the 
recordkeeping and reporting requirements be for these units?
1. Applicability Criteria
    The Agency believes that the initial criteria for a unit to qualify 
for the excepted monitoring should be consistent with the on-going 
criteria for using such monitoring so that only units that can likely 
continue to use the methodology will qualify in the first place. With 
the reduced monitoring requirements under this exception, a unit will 
not need to install monitors. Consequently, the Agency believes that 
the on-going applicability criteria should not depend on measurements 
from emissions monitoring equipment and that actual emissions data or 
actual heat input data, which are measured by the monitoring equipment, 
would not be appropriate as the primary applicability criteria for 
initial qualification for the exception or as the criteria for on-going 
qualification.
    The Agency considered what criteria, other than actual 
measurements, should be used as a basis for determining applicability 
to use the reduced monitoring and reporting exception. EPA considered 
various parameters to use in the applicability criteria, including: 
estimated emissions or heat input, the fuel burned, the unit capacity 
factor, and annual generation measured in MW-hr. Because the Agency's 
objectives for the exception include ensuring that the total emissions 
from the group of units that would qualify under the exception are de 
minimis and allowing more cost effective monitoring for units in such a 
group, the Agency believes it would be preferable to base the 
applicability on estimated emissions. While it may be simpler to base 
qualification for reduced monitoring solely on the fuel burned, the 
unit capacity factor, or the annual generation than to estimate the 
emissions, the Agency believes that it would be more difficult under 
that approach to ensure that total emissions that qualify under the 
exception were de minimis. The Agency further believes that using any 
of the other parameters, while attempting to ensure that the total 
emissions from the group are de minimis, might exclude some units that 
actually have low emissions. For example, a unit that burns mostly 
natural gas with emergency oil would be excluded from an exception 
limited to units that burn only natural gas. The Agency believes that 
an applicability criteria based on emissions would relate more directly 
to the objectives behind the optional exception than would other 
operating factors that might serve as a proxy for emissions.
2. Method for Determining Emissions
    The Agency considered several methods for determining the estimated 
emissions as the basis for applicability of the reduced monitoring and 
reporting excepted methodology. For each of the methods considered, 
rather than using actual measured sulfur and carbon values, 
CO2, SO2, and flow CEM readings, NO_X 
CEM readings, or NO_X values from an Appendix E 
NO_X-versus-heat input correlation, a facility would 
calculate the unit's emissions based on an emission rate factor and 
default heat input. Since the units that would qualify for the excepted 
methodology would still be accountable for reporting emissions to the 
Agency and surrendering allowances based on those emissions, where 
applicable, the emissions estimations would not just be used to 
determine if the unit qualifies under the exception; the reported 
estimations would also be used to determine compliance. The Agency 
considered its goals for emissions accounting in order to establish the 
emission rate factors and default heat input. The Agency maintains that 
it would be inappropriate to select values that would potentially 
underestimate emissions, thereby undermining the Agency's ability to 
determine compliance and achieve emission reductions under title IV or 
any other regulatory program involving SO2, CO2, 
or NO_X. Some industry representatives suggested that 
facilities would be willing to use a conservative emission estimate, 
such as a maximum potential emission rate times the maximum heat input, 
if it would allow them to save time and money currently spent on 
monitoring and quality assurance (see Docket A-97-35, Items II-D-30, 
II-D-43, II-D-45, II-E-13, and II-E-25).
    The Agency explored basing the estimated emissions on a unit's 
maximum potential emissions, i.e., converting the unit's nameplate 
capacity (which assumes maximum possible operation) to a maximum annual 
heat input for the unit and multiplying by the unit's maximum emission 
rate (which assumes the highest emission rate of all fuels capable of 
being burned at the unit). This option would have several advantages. 
It would ensure that emissions are not underestimated, would allow for 
reduced monitoring requirements, and would ensure that a unit that 
initially qualifies for the exception would continue to qualify without 
having to reevaluate the unit's emissions each year (unless some 
modification was made to the unit to increase its nameplate capacity or 
allow a higher emitting fuel to be burned). This approach, however, 
would likely disqualify gas-fired units that sometimes burn oil or 
peaking units that operate infrequently, since maximum potential 
emissions would be substantially higher than their actual emissions and 
would likely exceed the applicability criteria limit. Using this method 
to estimate emissions for purposes of an applicability cutoff would 
greatly diminish the usefulness of the reduced requirements and would 
fail to fully meet the intended purpose of today's proposal.
    In place of using a heat input derived from maximum possible 
operation (i.e., nameplate capacity), the Agency considered estimating 
heat input by multiplying the actual operating hours times a maximum 
rated hourly heat input for the unit. While this would require re-
evaluation of a unit's eligibility each year, this would allow an 
infrequently operated peaking unit to qualify if its emissions are low, 
which EPA believes is worth the additional burden of annual re-
evaluation. Therefore, the Agency is proposing to use maximum rated 
hourly heat input as the heat input in the emissions

[[Page 28038]]

estimation. Maximum rated hourly heat input would be defined, in 
Sec. 72.2, as a unit-specific maximum hourly heat input (mmBtu) based 
on the manufacturer's rating of the unit or, if that value has been 
exceeded in practice, based on the highest observed hourly heat input. 
In addition, there would be provisions for a lower maximum hourly heat 
input to be used if the unit has undergone modifications which 
permanently limit its capacity.
    The Agency also considered what emission rate(s) to apply, instead 
of using the highest emission rate of all fuels capable of being burned 
at the unit, in order to avoid underestimation and to allow a unit that 
primarily burns gas but has the ability to burn oil to qualify for the 
reduced requirements. The Agency believes that it would be appropriate 
to use emission rates based on uncontrolled emissions for the actual 
fuel burned in any given hour to estimate emissions for purposes of the 
initial and on-going applicability cutoffs to qualify to use the low 
mass emissions excepted methodology and for purposes of emissions 
reporting, allowance accounting, and compliance. This approach would 
avoid disqualifying gas-fired units simply because of their occasional 
use of oil and would also avoid underestimating emissions.
    For determining SO2 mass emissions using the low mass 
emissions methodology, EPA proposes the use of emission factors in lb/
mmBtu based on its AP-42 air pollution emission rate factors, which are 
established from the sulfur content and gross calorific value of the 
fuel being burned (see Docket A-97-35, Items II-A-11, II-I-1). Since 
the SO2 emissions are directly proportional to the amount of 
sulfur in the fuel and in light of the limited variability in the 
sulfur content of natural gas and oil, the proposed SO2 mass 
emission factors should be fairly representative of uncontrolled, 
actual emissions. Because of the relatively low sulfur content of 
natural gas or oil, it is doubtful that any of such units have 
SO2 controls. The proposed factors fall within the typical 
range of sulfur content and gross calorific value for each fuel, 
although somewhat on the conservative side for sulfur content of diesel 
fuel and natural gas other than pipeline natural gas.
    For determining NO_X mass emissions and emission rate, 
EPA proposes using the fuel- and unit-type-specific NO_X 
emission rate factors based on 90th percentile emission rate data 
reported under part 75 generally for uncontrolled units (see Docket A-
97-35, Item II-A-9). While attempting to develop an accounting approach 
for NO_X emissions from low mass emission units, EPA 
encountered several issues. The first issue involves the use of AP-42 
factors. During the finalization of the core part 75 monitoring rule, 
EPA considered allowing peaking units with negligible emissions both 
individually and collectively to estimate NO_X emissions 
using AP-42 emission rate factors. EPA rejected this approach in the 
January 11, 1993 final rule preamble at 58 FR 3644-45 because the AP-42 
emission rate factors are derived from industry-wide average estimates 
of emissions for different fuel and boiler types and are not based on 
actual historical operating experience of the units to which the 
estimates would be applied. Applying AP-42 factors could result in 
underestimation of NO_X emissions because actual 
NO_X emissions can vary significantly from unit to unit. The 
formation of NO_X from the combustion of fossil fuels is 
dependent on the amount of nitrogen in the fuel being combusted and on 
the mix of nitrogen and oxygen in combustion air. Further, the 
NO_X formation process depends on unit-specific factors of 
combustion gas temperature and stoichiometry of fuel and air local to 
the flame. Consequently, there can be significant variations in the 
level of NO_X emissions from unit to unit due to variations 
in combustion conditions. Therefore, EPA is not proposing the use of 
AP-42 factors to estimate NO_X emissions from low mass 
emissions units. Instead, now that three years of actual historical 
operating data collected under part 75 are available, it was possible 
to develop the default NO_X emission rate factors being 
proposed today. Although the default NO_X emission rate 
factors in today's proposal are generic factors, they should not 
underestimate NO_X emissions because they are based on the 
90th percentile of actual annual average emission rates reported 
generally from uncontrolled units under part 75.
    The Agency also considered using site-specific NO_X 
emission rate factors based on historical emission data or emissions 
testing data for the unit. For example, a facility might use the 
maximum value ever recorded by the CEM for the unit, or it might use 
the highest NO_X emission rate value calculated from the 
unit's most recent Appendix E NO_X test, or it might use 
site-specific values similar to those discussed in the guidance manual 
for implementing the NO_X budget program in the OTR (see 
Docket A-97-35, Item II-I-7). The application of site-specific 
NO_X emission factors for low mass emission units raises 
several issues. First, for units with pollution controls where the 
emission factor is based on controlled emissions, the site-specific 
emission factor could underestimate actual emissions if the controls 
are not operating properly. EPA considered only allowing site-specific 
NO_X emission factors with units that do not utilize 
NO_X emission controls; however, EPA realizes that many units 
employ at least some form of NO_X emission controls (e.g., 
water or steam injection). EPA also considered allowing a source with 
controls to use a site-specific emission factor only if it could 
demonstrate that the pollution controls are operating properly. 
However, this would involve extensive, additional recordkeeping and 
tracking to verify the proper operation of pollution controls and 
ensure that emissions are not underestimated; this would run contrary 
to the general approach under the exception of reducing monitoring and 
reporting requirements. A second issue involves verifying that the 
site-specific NO_X emission factor is still representative 
over time or after unit modifications. This would require future 
NO_X emission rate testing. Therefore, for purposes of 
creating a methodology that is simple to implement and in order to 
reduce future testing requirements for facilities with low mass 
emitting units, the Agency proposes instead using NO_X 
emission rate factors based on fuel and unit type and reflecting 
uncontrolled emissions. EPA requests comments on this approach, whether 
other approaches should be used, and especially whether there are any 
additional boiler types not represented in today's proposed rule for 
which NO_X emission rates should be provided.
    For determining CO2 mass emissions, today's rule 
proposes to use CO2 emission rate factors in tons/mmBtu. The 
CO2 emission rate factors are derived based on ideal gas 
theory and standard Agency Fc factors for estimating the 
volume of CO2 to be emitted when a certain heat input of a 
particular fuel is burned (see Docket A-97-35, Item II-A-11). This 
resembles the approach currently used in Equation G-4 of Appendix G for 
gas-fired units.
    Therefore, the Agency believes that an appropriate method of 
estimating emissions for the purposes of qualifying for a reduced 
monitoring and reporting exception and for purposes of emissions 
accounting and compliance for units under the exception is to calculate 
emissions based on the actual number of operating hours and the actual 
fuel burned using maximum rated hourly heat input and fuel-based and, 
for NO_X unit-type-based, emission factors. The Agency 
requests comments on this approach and on whether an alternate

[[Page 28039]]

approach should be used. While the Agency believes that the resulting 
emissions estimates will in most, if not all, cases be conservative and 
result in an overestimation of emissions, it would be possible, however 
unlikely, that the estimate could underestimate the actual emissions 
for some types of units. Therefore, for existing units with historical 
emissions data available, the proposal would require that in addition 
to meeting the applicability criteria using the emissions estimates 
calculated as described above, the unit would have to meet the cutoffs 
for initial qualification for the exception using the actual annual 
emissions monitored during the three years prior to applying to use the 
exception.
3. Cutoff Limit for Applicability
    EPA began developing applicability criteria by first considering 
the level of projected aggregate emissions determined to be de minimis 
for purposes of developing the new unit exemption promulgated in the 
January 11, 1993 Acid Rain permitting rule (see 58 FR 3593-94 and 3645-
46). Aggregate emissions projected for units under the exemption were 
approximately 138 cumulative tons of SO2 and 1934 cumulative 
tons of NO_X emitted per year. The Agency then conducted a 
study of actual emissions data from 1996 quarterly reports under part 
75 and evaluated potential tonnage cutoffs for SO2 and 
NO_X. The Agency compared the cumulative mass emissions from 
groups of units emitting less than various specified amounts to the 
total emissions reported under the Acid Rain program during the year 
(see Docket A-97-35, Item II-A-10). For example, the study shows what 
proportion of total SO2 was emitted by units with both 
actual and potential ¹ emissions of 25 tons or less per 
year, 50 tons or less per year, 60 tons or less per year, and 75 tons 
or less per year. From these analyses, EPA also estimated how many 
units might be eligible for reduced requirements for determining 
emissions and how much of an impact the new emissions accounting option 
would have on nationwide emissions accounting.
---------------------------------------------------------------------------

    \1\ The terms ``potential emissions'' used in this section of 
the preamble have a different meaning than the terms ``potential to 
emit'' used elsewhere by the Agency.
---------------------------------------------------------------------------

    EPA is proposing cutoff values of 25 tons per year of 
SO2 and 25 tons per year of NO_X. In order to 
qualify as a low mass emissions unit, a unit would have to demonstrate 
that both actual historical emissions and potential emissions 
(calculated with maximum hourly heat input, emission factors and 
either, for existing units, actual historical number of operating hours 
or, for new units, projections of future annual operating hours) do not 
exceed 25 tons each for SO2 and NO_X on an annual 
basis. Based upon its analyses (see Docket A-97-35, Item II-A-10), EPA 
estimates that this tonnage cutoff level would mean that the group of 
units subject to the proposed reduced requirements, even after Acid 
Rain Program emission reductions are considered, would have total 
annual emissions of about 16 tons of SO2 and 90 tons of 
NO_X (less than a thousandth of a percent of total annual 
SO2 emissions and about 0.002 percent of total annual 
NO_X emissions for all affected units). Both amounts, 16 tons 
of SO2 and 90 tons of NO_X, are less than the 
total number of tons of those pollutants determined to be de minimis 
for purposes of the new unit exemption. Today's proposal to treat low 
mass emission units as de minimis is consistent with the de minimis 
conclusions reached for new units.
    While the reduced requirements are somewhat less accurate than the 
methodologies under the existing regulations, the reduced requirements 
are intended to yield emissions data that are conservative and that, to 
the extent they are inaccurate, are likely to overstate emissions. 
Moreover, EPA believes that the level of inaccuracy (i.e., 
overstatement of emissions) would similarly be extremely low (i.e., 
less than a thousandth of a percent). Both the total emissions subject 
to the reduced requirements and the potential amount of overstatement 
of emissions are de minimis. Moreover, any overstatement of regulated 
emissions would have the effect of tightening emission limits (e.g., by 
requiring surrender of more allowances for SO2 than 
otherwise). Any overstatement of other emissions would be too small to 
affect adversely the air quality related activities (e.g., air quality 
modeling) for which the emissions data would be used.
    EPA would, however, be concerned about extending today's proposed 
reductions in monitoring, quality assurance, and reporting requirements 
to units that exceed the 25-ton cutoffs for actual or potential 
emissions. Section 412 of the CAA requires all affected units to 
monitor SO2, volumetric flow, NO_X, and opacity 
using continuous emission monitoring systems or an alternative 
monitoring system approved by the Administrator as having the same 
precision, reliability, accessibility, and timeliness. In addition, 
section 412 of the Act requires that emissions data be quality-assured. 
Section 821 of the Clean Air Act Amendments of 1990 provides that, 
through regulations issued by the Administrator, all affected units 
must be required to monitor CO2 emissions in the same manner 
and to the same extent as SO2 and NO_X are 
monitored under section 412. Part 75 of EPA's rules requires monitoring 
of SO2, NO_X, and CO2 and allows 
certain exceptions to the statutory requirement for CEMS or CEMS-
equivalent alternative monitoring: in Appendix D because, inter alia, 
the information gathered using the Appendix D methods is as precise, 
reliable, accessible, and useful as that from CEMS, and compares 
acceptably with regard to timeliness; and in Appendix E because the 
emissions from all units eligible to use Appendix E are negligible and 
such units do not have emission limitations for NO_X under 
the Acid Rain Program (see 58 FR 3641-45). The proposed reduced 
monitoring and reporting requirements for low mass emissions units 
would not yield information equivalent to that from CEMS. EPA must 
balance the benefits of reduced monitoring, quality assurance, and 
reporting requirements for units against the intent of the statute that 
monitoring with CEMS or their equivalent be required so as to obtain 
reliable, precise, timely, and readily accessible information on 
emissions. EPA solicits comment on whether 25 tons is the appropriate 
cutoff level for applicability of the low mass emission excepted 
methodology.
    In particular, EPA is concerned that extending the proposed 
reduction in requirements to units with more than this de minimis level 
of emissions could have a negative impact on the environment. Emissions 
data from the Acid Rain Program are being used for a variety of 
efforts, including emissions modeling and establishing baseline 
emissions information (prior to any emission reductions) for new air 
pollution control programs. Using less accurate methods to monitor more 
than a de minimis amount of emissions could potentially undermine 
efforts to establish baseline emissions and to assess what emission 
reductions have already taken place and how much further emissions must 
be reduced in order to meet air quality standards.
    Furthermore, with regard to coal-fired units, such units account 
for the largest proportion of all emissions, tend to be operated more 
frequently, and generally have much higher emission rates in lb/mmBtu 
for SO2, NO_X and CO2, and the majority 
of the units have emission limitations and emission reduction

[[Page 28040]]

requirements for SO2 and NO_X. In addition, the 
sulfur content in coal and gaseous fuels other than natural gas is much 
more variable than for natural gas and oil, and the emission factors 
for coal or gaseous fuels other than natural gas, particularly an 
SO2 emission factor, are therefore less reliable and much 
more likely to understate, rather than overstate, emissions. Based on 
these considerations, the proposed rule would restrict the use of the 
reduced requirements to gas-fired units and oil-fired units that burn 
natural gas and/or fuel oil.
    In order to qualify for the proposed low mass emissions excepted 
methodology, the proposed applicability criteria would require a unit 
to meet annual tonnage cutoffs of 25 tons each for SO2 and 
NO_X. EPA considered whether the excepted methodology should 
be available on a pollutant specific level so that, for example, a unit 
which falls below the tonnage cutoff for SO2 but not for 
NO_X could use the proposed excepted methodology under 
Sec. 75.19 to measure SO2 emissions but use a NO_X 
CEM or the excepted methodology under Appendix E, where applicable, to 
measure NO_X emissions. EPA believes this approach would not 
be appropriate because some of the same monitoring equipment and 
reporting software is necessary for measuring and reporting both of the 
pollutants. One of the prime benefits of the low mass emissions 
excepted methodology would be the simplified reporting which would 
require less time and a less sophisticated Data Acquisition and 
Handling System. In particular, the need for a DAHS that could 
calculate substitute data using the missing data algorithms would be 
removed because there are no missing data algorithms for the low mass 
emissions excepted methodology. If the excepted methodology is only 
applied to one of the pollutants, much of the benefit would be negated 
because the DAHS would still need to be capable of calculating 
substitute data for the measured pollutant and close to the full 
quarterly report would still be required. Another prime benefit of the 
proposed low mass emissions excepted methodology would be the removal 
of monitoring and quality assurance requirements. However, EPA believes 
that almost all units that would qualify for a 25-ton cutoff for only 
one pollutant would meet the cutoff for SO2, not 
NO_X, and would already be using Appendices D and E. A unit 
using a fuel flowmeter to determine SO2 mass emissions under 
Appendix D likely uses the same fuel flowmeter to determine 
CO2 emissions and heat input. Additionally, the same fuel 
flowmeter is used to determine NO_X emissions under Appendix 
E. Even if the unit were allowed to use the proposed low mass emissions 
excepted methodology for SO2 in lieu of Appendix D, the unit 
would still have to install, certify, operate, maintain, quality 
assure, and report from a fuel flowmeter to determine NO_X 
emission rate and heat input. Accurate heat input is important since 
heat input is used to calculate NO_X mass emissions. In 
short, the cost of operation, maintenance, and quality assurance of the 
fuel flowmeter would not be removed simply by removing the requirement 
to monitor SO2. Even if a unit that qualified under the low 
mass emissions excepted methodology for SO2 but not for 
NO_X was currently monitoring with Appendix D, for 
SO2 and heat input, and using a NO_X CEM, for 
NO_X emission rate, using the excepted methodology for 
SO2 but not for NO_X would have little benefit 
since the installation, certification, and quality assurance testing of 
the fuel flowmeter would still be required to determine heat input. 
Therefore, today's proposed low mass emissions excepted methodology 
would be provided as an option only if the unit has low mass emissions 
of both SO2 and NO_X. EPA solicits comment on this 
approach and on whether any benefit of allowing the excepted 
methodology for one pollutant only would outweigh the added complexity 
in the excepted methodology.
    EPA also considered whether a tonnage cutoff for CO2 
emissions was appropriate as part of the proposed applicability 
criteria for low mass emissions units. However, the proposed excepted 
methodology under Sec. 75.19 would require the use of a standard 
emission factor (in lb of NO_X/mmBtu) for NO_X to 
determine eligibility for the exception. This would effectively 
establish an upper limit on the annual heat input for a given fuel and 
boiler type at the level that would allow the unit to meet the tonnage 
cutoff applicability requirements. Because CO2 emissions are 
directly proportional to heat input, there would be a built-in annual 
CO2 emissions cutoff inherent in the methodology.
4. Continuing Applicability Criteria
    In drafting today's proposal, EPA also considered how to ensure 
that after individual units initially qualified to use the reduced 
monitoring exception, they could continue to use the exception only if 
they continued to have de minimis emissions. Many of the units that 
would qualify as low mass emissions units under the proposal have low 
emissions either because they use pipeline natural gas and/or because 
they operate infrequently. In both of these situations, it is 
conceivable that a unit's emissions could become significant if the 
unit's fuel or hours of operation were to change. Most gas-fired units 
are capable of burning oil, but generally do so only when pipeline 
natural gas is not available. However, if the prices of gas and oil 
were to change such that oil became far more economical than gas, some 
gas-fired units might switch to burning high sulfur oil. Similarly, 
increases in demand for electricity could cause some peaking units to 
operate more frequently, thereby generating more emissions. Therefore, 
EPA is proposing that in order to ensure that emissions from units 
using the reduced requirements would remain de minimis, units would 
have to continue to meet the applicability criteria in order to qualify 
as low mass emissions units. Because of the conservative heat input and 
in some cases, conservative emission factors, the Agency believes that 
meeting the applicability criteria of less than 25 tons of both 
SO2 and NO_X when calculating the emissions using 
the low mass emissions excepted methodology, will ensure that the 
actual emissions of the low mass emission units will be below those 
levels. Therefore, once the methodology is implemented, the on-going 
applicability would only require that the limits be met with the 
calculated mass emissions, i.e., the facilities would be required to 
continue to meet the 25-ton cutoffs on an annual basis, as determined 
using the emission calculation procedures in proposed Sec. 75.19.
    It would, therefore, be necessary for low mass emissions units to 
report NO_X mass emissions, in addition to the required 
SO2 mass emissions and NO_X emission rate, in 
order to determine continuing applicability. A continuing applicability 
provision of this nature would prevent a unit from continuing to use 
the reduced requirements when its emissions were no longer negligible. 
If a unit initially met the applicability criteria but failed to meet 
one or both of the annual 25-ton cutoffs in a future year, the unit 
would become disqualified from using the exception. Sufficient time 
would be necessary to purchase, install, and certify CEMS or the 
equipment necessary for monitoring under Appendices D and/or E. 
Therefore, a unit would not be disqualified until two calendar quarters 
after the quarter in which the 25-ton cutoff is exceeded and would not 
be required to certify and report from

[[Page 28041]]

monitoring systems until then. If that unit changes, or is projected to 
change, its fuel or amount of operation in the future so that it would 
again meet the 25-ton SO2 and NO_X cutoffs, the 
unit could again qualify as a low mass emissions unit. However, if the 
unit initially qualified based on projected operating hours and fuel 
usage and then was disqualified the unit could not use projected data 
to qualify again. The unit would need to monitor using CEMS, an 
approved alternative monitoring system, or an optional protocol under 
Appendices D and/or E, where applicable, for at least an additional 
three years in order to accumulate three years of actual data.
5. Reduced Monitoring and Quality Assurance Requirements
    As discussed above, today's proposed rule would allow facilities to 
use a maximum rated hourly heat input value and an emission rate factor 
to determine the mass emissions from a low-emitting unit for each hour 
of actual operation. This approach would involve no actual emissions 
monitoring and no quality assurance activities. Instead, the facility 
would only need to keep track of whether the unit combusted any fuel 
for a particular hour and what type of fuel was combusted. In this way, 
the proposed revisions would significantly reduce the burden on 
affected facilities, while still ensuring that emissions are not 
underreported.
6. Reduced Reporting Requirements
    Some utilities have mentioned that they find it troublesome to 
spend as much time or more reviewing quarterly report submissions for 
small, infrequently operating gas-fired units as they spend reviewing 
quarterly report submissions for large coal-fired units (see Docket A-
97-35, Items II-D-75, II-E-25). EPA agrees that facility environmental 
personnel should be able to spend a greater percentage of their time 
focusing on units with higher emissions than on low mass emissions 
units, which, as discussed above, account for such a small portion of 
total emissions. Thus, today's proposed rule would simplify the 
reporting requirements for low-emitting units so that facilities could 
spend less of their environmental department resources on units with 
negligible emissions. For units that rely on the procedures in proposed 
Sec. 75.19(c), the owner or operator would have no requirements related 
to records or reports of certification testing and would be exempt from 
all of the specific recordkeeping requirements in Secs. 75.54(b) 
through (e) or 75.57(b) through (e) relating to operating parameter and 
emissions records. Instead, the rule would require only that an initial 
certification application, containing data supporting the applicability 
demonstration, and a monitoring plan be submitted and that limited 
hourly, quarterly, and year-to-date cumulative data be reported on a 
quarterly basis. The hourly record would only be reported for hours of 
unit operation, and an hour in which the unit combusted fuel for any 
portion of the hour would be considered a full hour, for simplicity.
    One utility has suggested that it would be less burdensome if it 
could simply report its quarterly cumulative emissions, without 
reporting any supporting hourly data; other utility representatives 
have indicated that it would be no more burdensome to report an hourly 
default emission value if the utility were already reporting hourly 
operating information (see Docket A-97-35, Item II-E-25). For purposes 
of modeling air quality, the Agency considers hourly operating 
information far more valuable (e.g., for modeling discrete periods of 
ozone exceedance) than just a quarterly emission value with no time or 
date mentioned. Furthermore, because facilities already keep track of 
the operation of their units for business purposes, keeping track of 
and reporting hourly operating information should not be a substantial 
burden. According to industry representatives, however, allowing 
facilities to record and report default emission values instead of 
hourly measured values would significantly speed up their review of 
quarterly reports prior to submission to the Agency (see Docket A-97-
35, Item II-E-25). Thus, requiring facilities to report hourly 
operational data and the default emissions data for the fuel burned 
that hour, but not hourly measured emissions or heat input in 
additional record types, would preserve the Agency's ability to model 
air quality while imposing far less burden upon facilities than the 
current part 75 requirements. Furthermore, because hourly default 
values would be employed, the need for missing data procedures would be 
eliminated and the Data Acquisition and Handling System (DAHS) could be 
greatly simplified. In fact, the reporting requirements for a low mass 
emissions unit could most likely be fulfilled with the use of a 
commercially available spreadsheet software package. EPA has 
incorporated this approach into today's proposed rule.

D. Quality Assurance Requirements for Moisture Monitoring Systems

Background
    Section 75.11(b) of the original January 11, 1993 Acid Rain rule 
requires the owner or operator to continuously (or on an hourly basis) 
account for the moisture content of the stack gas when SO2 
concentration is measured on a dry basis. The moisture content is 
needed to correct the measured hourly stack gas volumetric flow rates 
to a dry basis when calculating SO2 mass emission rates in 
lb/hr. Section 75.13(a) of the rule, as amended on May 17, 1995, 
contains provisions for CO2 monitoring paralleling the 
provisions of Sec. 75.11(b); that is, when CO2 concentration 
is measured on a dry basis, a correction for stack gas moisture content 
is needed to accurately determine the CO2 mass emissions. 
The stack gas moisture content is also needed when a dry-basis 
O2 monitor is used to account for CO2 emissions 
and, in some instances, when accounting for unit heat input (see 
Secs. 75.13(c), 75.16(e), and Equations F-14b, F-16, F-17 and F-18 in 
Appendix F) or when determining NO_X emission rate in lb/
mmBtu (see section 3.2 in Appendix F, and Equations 19-3 through 19-5, 
19-8, and 19-9 in Method 19 of Appendix A to part 60).
    As presently codified, part 75 does not specify any quality 
assurance requirements for moisture measurement devices. Implementation 
has shown this to be an unfortunate omission in the rule, since 
approximately 5 to 10 percent of the continuous emission monitors in 
the Acid Rain Program require moisture corrections to accurately 
measure SO2, CO2, or NO_X emissions or 
heat input (see Docket A-97-35, Item II-I-6). The accuracy of the stack 
gas moisture measurements directly affects the accuracy of the reported 
SO2 mass emission rates, CO2 mass emission rates, 
NO_X emission rates and heat input values. An error of 1.0 
percent H2O in measured moisture content causes a 1.0 
percent error in the reported emission rate or heat input value. 
Failure to quality assure the moisture data can therefore result in 
significant under-reporting of SO2, CO2, and 
NO_X emissions and heat input. The Agency does not know the 
extent of inaccuracy that currently exists in the measurement of 
moisture by affected units but believes it is important to require 
certification and quality assurance of moisture monitors--just as is 
required for other CEMS used under part 75--because the success of the 
SO2 trading system depends on accurate monitoring.

[[Page 28042]]

Discussion of Proposed Changes
Today's proposal would incorporate into part 75 quality assurance 
requirements for moisture monitoring systems. Section 75.11(b) would be 
revised to require the owner or operator to install, maintain, operate, 
and quality assure a moisture monitoring system. Proposed Sec. 75.11(b) 
also specifies that a moisture monitoring system may either consist of: 
(1) a continuous moisture sensor; (2) an oxygen analyzer (or analyzers) 
capable of measuring O2 on both a wet basis and on a dry 
basis; or (3) a system consisting of a temperature sensor and a 
certified DAHS component capable of determining moisture from a lookup 
table, i.e., a psychrometric chart (this third option would apply only 
to saturated gas streams following wet scrubbers). Corresponding 
changes would be made to Secs. 75.12, 75.13(c) and 75.16(e) to require 
that a quality assured moisture monitoring system be used whenever 
moisture corrections are needed to accurately account for 
NO_X emissions, CO2 emissions, or heat input.
    Requirements for the initial certification of moisture monitoring 
systems are proposed in three new sections, Secs. 75.20(c)(5), (c)(6), 
and (c)(7). To make room for the new sections, existing 
Sec. 75.20(c)(3) would be deleted; existing Secs. 75.20(c)(4) and 
(c)(5) would be redesignated as Secs. 75.20(c)(3) and (c)(4); and 
existing Secs. 75.20(c)(6), (c)(7), and (c)(8) would be redesignated, 
respectively, as Secs. 75.20(c)(8), (c)(9), and (c)(10). The 
certification requirements for continuous moisture sensors are found in 
proposed Sec. 75.20(c)(6) and include a 7-day calibration error test 
and a relative accuracy test audit (RATA). For moisture monitoring 
systems consisting of one or more wet- and dry-basis oxygen analyzers, 
the proposed certification requirements are found in Sec. 75.20(c)(5) 
and include a 7-day calibration error test, a linearity test and a 
cycle time test of each O2 analyzer, and a RATA of the 
moisture measurement system. Corresponding revisions to 
Sec. 75.22(a)(4) are proposed, specifying that EPA Method 4 (either the 
standard procedure or the midget impinger procedure) would be used as 
the reference method for the moisture RATAs. For saturated gas streams, 
if a lookup table is used to determine the hourly stack gas moisture 
content, the certification requirement in proposed Sec. 75.20(c)(7) 
would consist of a DAHS verification. At a minimum, the DAHS 
verification would have to demonstrate, at three temperatures covering 
the normal range of stack temperatures, that the software extracts the 
proper moisture value from the lookup table and applies it correctly to 
the emission calculations. In today's proposal, a new Sec. 75.4(i) 
would also be added, requiring owners or operators to complete all of 
the applicable moisture monitoring system certification tests specified 
in proposed Secs. 75.20(c)(5), (c)(6), and (c)(7) no later than January 
1, 2000.
    Proposed performance specifications for moisture monitoring systems 
are found in sections 3.1, 3.2, 3.3, and 3.5 of Appendix A to part 75. 
These specifications would apply to continuous moisture sensors and to 
wet- and dry-basis oxygen analyzers. The proposed calibration error 
specification in section 3.1 for continuous moisture sensors is 3.0 
percent of span. A new section, 2.1.5, would be added to Appendix A, 
defining the span of a moisture sensor as equal to the full-scale range 
of the instrument and requiring that the range be consistent with 
section 2.1 of Appendix A. For moisture monitoring systems consisting 
of wet- and dry-basis O2 analyzers, the proposed span values 
and performance specifications for calibration error, linearity, and 
cycle time in sections 2.1.3, 3.1, 3.2, and 3.5 of Appendix A would be 
the same as the current specifications for O2 monitors. The 
proposed relative accuracy (RA) specification for moisture monitoring 
systems is found in a new section, 3.3.6, in Appendix A and would be 
equal to 10.0 percent. An alternative RA specification would also be 
provided in section 3.3.6, i.e., the relative accuracy would also be 
acceptable if the difference between the mean difference of the 
reference method measurements and the moisture monitoring system 
measurements is within  1.0 percent H2O. A 
relative accuracy specification of 10.0 percent is being proposed in 
order to maintain consistency with the relative accuracy requirements 
for the other program monitors (SO2, NO_X, flow 
rate, and CO2). The Agency notes that moisture RATAs have 
not previously been required by any other EPA continuous monitoring 
regulation, and therefore there is no relative accuracy database upon 
which to draw. However, moisture data are sometimes collected using EPA 
Method 4 during each run of a part 75 gas monitor RATA to convert the 
gas reference method readings from a dry basis to a wet basis. 
Therefore, some part 75 sources that currently account for moisture 
using wet- and dry-basis oxygen analyzers or a moisture sensor should 
be able to construct moisture RATAs from previous test data by 
comparing the Method 4 moisture data from the gas monitor RATAs against 
the readings recorded by the moisture sensor or O2 analyzers 
at the time of the gas RATAs. EPA encourages those facilities that 
currently make moisture corrections in their emission equations to 
perform this type of data analysis, if possible, and to provide comment 
on the appropriateness of the proposed moisture relative accuracy 
specification.
    On-going QA requirements for moisture monitoring systems are also 
proposed in sections 2.1.1, 2.1.4, 2.2.1, 2.3.1.1, and 2.3.1.2 of 
Appendix B to part 75. Proposed section 2.1.1 of Appendix B would 
require daily calibrations of moisture monitoring systems. Continuous 
moisture sensors would be calibrated in accordance with the 
manufacturers' recommended procedures. Proposed section 2.1.4 would 
give control limits for the daily calibrations (i.e.,  1.0 
percent O2 for oxygen analyzers and  6.0 percent 
of span for continuous moisture sensors). Proposed section 2.2.1 would 
require quarterly linearity checks of wet- and dry-basis oxygen 
analyzer(s). Proposed section 2.3.1.1 would require semiannual RATAs of 
moisture monitoring systems, and proposed section 2.3.1.2 would specify 
that if a moisture monitoring system achieves a relative accuracy of 
 7.5 percent or if the mean difference between the CEMS and 
reference method values is within  0.7 percent 
H2O, the system qualifies for an annual, rather than 
semiannual RATA frequency.
    Missing data procedures for moisture are included in today's 
proposal in a new section, Sec. 75.37. The proposed missing moisture 
data procedures are as follows:
    (1) Begin by using the following ``initial'' missing data 
procedures as of the date and time of provisional certification of the 
moisture monitoring system or as of January 1, 2000 (whichever is 
earlier). Substitute 0.0 percent moisture for each hour of missing data 
if no prior quality assured data exist, and for the first 720 hours of 
quality assured monitor operating data, substitute, for each hour of 
each missing data period, the average of the ``hour before'' and ``hour 
after'' moisture values.
    (2) After 720 hours of quality assured data have been obtained, 
provided that the moisture data availability is  90.0 
percent, substitute the average of the ``hour before'' and ``hour 
after'' values for each hour of the missing data period.
    (3) When the percent data availability for moisture is below 90.0 
percent, substitute 0.0 percent moisture for each hour of the missing 
data period.

[[Page 28043]]

    These proposed missing data procedures are considerably simpler 
than the corresponding procedures for SO2, NO_X, 
CO2, and flow rate, in that they do not include the concepts 
of lookback periods, 90th, or 95th percentile values. However, the 
procedures are also somewhat less representative than the missing data 
procedures for SO2, NO_X, CO2, and flow 
rate, because the most conservative possible value (0.0 percent 
moisture) is substituted when the moisture monitor data availability 
drops below 90.0 percent. The Agency solicits comment on whether the 
simpler (but less accurate) missing data procedures or the more complex 
(but more representative) procedures are more appropriate.
    Finally, Secs. 75.57(c) and 75.59(a) (revised versions of 
Secs. 75.54(c) and 75.56(a)) would be added in today's proposal to 
require that records be kept of the following: (1) Component-system 
identification code for the moisture monitoring system; (2) hourly 
average moisture readings (including, if applicable, hourly averages 
from each wet- and dry-basis O2 analyzer); (3) percent data 
availability for the moisture monitoring system; (4) daily and 7-day 
calibrations of moisture monitoring systems; (5) linearity tests of 
each wet and dry oxygen analyzer used to determine moisture; and (6) 
relative accuracy tests of moisture monitoring systems.
    In summary, EPA is proposing quality assurance (QA) procedures for 
moisture monitoring systems because the Agency believes that 
continuous, quality assured, direct measurement of the stack gas 
moisture content or continuous measurement of surrogate parameters, 
such as wet- and dry-basis oxygen concentrations, is the best way to 
ensure the accuracy of the reported emission data when moisture 
corrections must be applied. However, the Agency is willing to consider 
and solicits comment on simpler alternative methods of accounting for 
the stack gas moisture content, such as using a conservative default 
moisture value. Any proposed alternative methodology submitted to the 
Agency for consideration would have to provide a comparable level of 
accuracy and would have to ensure that emissions and heat input are not 
under-reported.

E. Certification/Recertification Procedural Changes

Background
    Currently, Sec. 75.20 lays out the process for certifying 
monitoring systems. Section 75.20(a) specifies the requirements for 
initial certification, including the contents of a certification 
application, when the application must be submitted and the process for 
reviewing and acting on an application. Sections 75.20(a)(3) and (4) of 
the existing rule establish a certification application review period 
of 120 days (after receipt of a complete application) for EPA to review 
an application and issue an approval or disapproval. For a continuous 
emission monitor (CEM), initial certification includes the following 
tests: relative accuracy, bias, linearity (pollutant monitors only), 7-
day calibration error, cycle response time (pollutant monitors only), 
missing data, and formula verification. All of these tests must be 
passed for a CEM to be certified and produce valid quality assured 
data. Once a CEMS is certified, Sec. 75.20(b) specifies that if 
something changes that significantly affects the ability of the CEM to 
accurately measure concentration or volumetric flow, the affected 
monitoring system(s) must be recertified. Recertification includes one 
or more of the initial certification tests. All required 
recertification tests must be passed, and a recertification application 
must be submitted in order for a CEM to be recertified. Section 
75.20(b)(5) of the existing rule establishes a 60 day review period for 
recertification applications. Separate but similar certification and 
recertification test requirements apply for a monitoring system other 
than a CEM, i.e., an excepted monitoring system under Appendix D or E, 
an alternative monitoring system under subpart E, or a system under 
proposed Appendix I.
    Submittal requirements for certification and recertification 
applications are included in Secs. 75.60 and 75.63 of the current part 
75. Generally, these provisions require submittal of certification test 
results in electronic formats, with some information required to be 
submitted in hardcopy format. Certification or recertification test 
results also must be submitted electronically in quarterly reports 
under Sec. 75.64. Finally, Sec. 75.61 requires the designated 
representative to provide advance notice to the applicable state or 
local agency and EPA Regional Office of certification and 
recertification testing.
    In many respects, monitoring plan requirements are tied to the 
certification/recertification process because a modification to the 
monitoring system that requires a recertification application also 
usually requires a monitoring plan update. In addition, because it 
contains the information about what type of equipment is located where, 
the monitoring plan is an essential tool in the review of a 
certification or recertification application. Section 75.53 specifies 
the content of monitoring plans and when changes to the plan are 
required. Section 75.62(a) specifies the submission requirements for 
monitoring plans.
    Based on EPA's initial experience with part 75 implementation and 
the numerous questions and problems encountered in the review of 
certification and recertification applications and monitoring plans, 
the Agency believes that the certification and recertification 
provisions and the related sections of the rule are possibly neither 
sufficiently detailed nor clear. Therefore, in today's rulemaking, EPA 
is proposing to revise those provisions and sections in order to 
improve the certification/recertification process. The issues addressed 
in today's proposed rule include the following: (1) whether a 
particular provision applies to initial certification, recertification, 
or both; (2) the scope of events that require submittal of a 
recertification application; (3) the review period lengths for initial 
certification and recertification applications; (4) the criteria 
governing disapproval of an incomplete certification or recertification 
application; (5) the format (electronic or hardcopy) in which test 
notifications, certification and recertification applications, and 
monitoring plans are to be submitted; (6) which EPA Regional Offices 
and state and local agency offices must receive test notifications, 
certification and recertification applications, and monitoring plans, 
and whether the submittal and notice requirements can be waived; and 
(7) when a monitoring plan needs to be revised. The proposed revisions 
on these topics and the rationale for the changes are discussed below.
    The Agency notes that today's package of proposed revisions to part 
75 includes other substantive revisions to the certification and 
recertification provisions in part 75. These are discussed elsewhere in 
this preamble. The provisions of most significance are related to 
certain proposed QA/QC revisions, back-up monitoring systems, CEM data 
validation issues, and the new Appendix I procedures. See sections 
III.D, O, R and T of this preamble for further discussion.
Discussion of Proposed Changes
    The proposed revisions discussed in this section affect Sec. 75.20 
generally, as well as specific aspects of Secs. 75.20(a)(4), (b)(1), 
(b)(5), and (g)(6); 75.21(e)(1); 75.53(b); new Sec. 75.53(e) and (f); 
75.60(b); 75.61(a); 75.62(a); 75.63(a) and

[[Page 28044]]

(b); 75.64(a), (b) and (d) and the addition of Sec. 75.59 as a revised 
version of Sec. 75.56. Proposed revisions to Sec. 75.20 would clarify 
which provisions apply to initial certification, recertification, or 
both. Proposed revisions to Sec. 75.20(b)(1) and (g)(6) would provide a 
narrow definition of recertification events, thereby significantly 
reducing the number of monitoring system changes, configuration changes 
or changes in the manner of operation that would require submission of 
a recertification application. Proposed revisions to Sec. 75.20(b)(5) 
would make the lengths of the review periods the same for initial 
certification and recertification applications. Proposed revisions to 
Sec. 75.20(a)(4) would clarify what constitutes a complete 
certification or recertification application and also would more 
clearly define EPA's authority to disapprove an incomplete application.
    Proposed revisions to Sec. 75.53(b) would expand the universe of 
monitoring system changes that require monitoring plan revisions to 
include any change that would make the information in the current plan 
inaccurate (currently, only changes that require recertification 
require monitoring plan changes). Sections 75.53(e) and (f), which are 
revised versions of existing Sec. 75.53(c) and (d), would clarify which 
elements of a monitoring plan must be submitted in electronic format 
and which elements must be submitted in hardcopy format. Section 
75.53(e) would revise existing Sec. 75.53(c) so that after January 1, 
2000 an owner or operator would have to report the unit stack height in 
the monitoring plan. Section 75.59 (a revised version of Sec. 75.56) 
would specify the minimum required content (as of January 1, 2000) for 
the hardcopy portion of a certification or recertification application. 
Section 75.60(b) would more clearly define the general requirements for 
submittal of reports and petitions. Section 75.61(a) would allow for 
certification and recertification test notices to be sent in various 
alternative media and would allow for EPA or a State or local agency to 
waive test notices in some circumstances. Section 75.62(a) would be 
revised to clarify when monitoring plans are to be submitted and to 
whom elements of the monitoring plan must be submitted. Similarly, 
Sec. 75.63(a) would be revised to detail which elements of a 
certification or recertification application are to be submitted 
electronically, which elements are to be submitted in hard copy, and to 
whom the various elements would be submitted. Section 75.63(b) would 
clarify when and how failed tests are to be reported in a certification 
or recertification application. Finally, Sec. 75.64(a) would specify 
that the hardcopy monitoring plan is not to be submitted with a 
quarterly report. The rationale for these changes is discussed below.
Rationale
1. Initial Certification Versus Recertification
    Several provisions in the current rule refer either to 
certifications or to certification applications; however, it is not 
always clear whether these provisions apply solely to initial 
certifications or whether they also apply to recertifications. 
Therefore, today's proposed revisions would make a number of minor text 
edits throughout Sec. 75.20 for clarification. There are, however, some 
events that do not fit neatly under the definition of initial 
certification or recertification (e.g., construction of a new stack 
with a new CEM at an existing unit when a scrubber is installed). This 
element of subjectivity in classifying an event as a certification or 
recertification makes it desirable for the certification and 
recertification processes to be as similar as possible. Having one 
general process with one set of rules rather than having two separate 
processes also makes program implementation easier. Currently, the main 
differences between initial certifications and recertifications are the 
types of tests required and the lengths of the application review 
periods. Today's proposed rule revisions would attempt to minimize 
these differences to the extent possible in order to bring greater 
uniformity and consistency to the certification and recertification 
process.
    (a) Scope of Recertification Events. The proposed revisions would 
narrow the scope of the types of changes to a monitoring system that 
would be classified as ``recertification events'' and would require 
submittal of a recertification application. Sections 75.20(b)(1) and 
(g)(6) would define a recertification event as any change that requires 
the performance of an accuracy test of a monitoring system, i.e., 
either a relative accuracy test audit (RATA) of a CEMS, an accuracy 
test of a fuel flowmeter, or a retest to develop the Appendix E 
NO_X correlation curve. For changes to a monitoring system or 
process that do not require a system accuracy test but require one or 
more of the other (lesser) quality assurance tests to be performed 
(e.g., linearity test or 7-day calibration error test), those other 
required tests would be classified as diagnostic tests rather than as 
recertification tests in Sec. 75.20(b)(1) of the proposal. For 
instance, a source would be required to conduct a linearity check after 
replacing a capillary tube in a gas analyzer with a tube from a like 
model and manufacturer (see Docket A-97-35, Item II-I-9, Policy Manual, 
Question 13.13). However, because this change to the CEMS does not 
require a RATA, it would not be considered a recertification event. 
Therefore, no recertification application would be required, and the 
linearity test would be considered a diagnostic test. Note that even 
though diagnostic tests would not be classified as recertifications, 
the recertification data validation procedures in proposed 
Sec. 75.20(b)(3) of today's rule would apply to these tests. EPA 
believes that the proposed narrowing of the definition of a 
recertification event will significantly reduce the number of required 
recertification applications and will make the submittal requirements 
for initial certifications and recertifications more consistent.
    (b) Recertification Review Period. Consistent with the proposed 
narrowing of the definition of a recertification event, EPA also 
proposes to revise Sec. 75.20(b)(5) by increasing the recertification 
application review period from 60 days to 120 days to make it the same 
as the review period for initial certifications. The advantage of 
making the two review periods consistent is that there would be no need 
to distinguish which requirements are applicable to which events. Some 
events combine aspects of initial certification and of recertification. 
For example, the certification of a new CEMS on a new stack at an 
existing unit when a scrubber is installed can be thought of as initial 
certification because it is an entirely new system in a new location; 
however, this event also involves aspects of recertification because it 
is an existing unit which has been reporting emissions from certified 
systems. Therefore, the Agency believes that making the review periods 
the same would reduce confusion and case-by-case determination of how 
long the review period should be for a given application. The Agency 
believes that it would be more effective to establish consistent 
procedural requirements for both initial certification and 
recertification events, rather than attempting to classify each event 
as an initial certification or recertification.
    In making the review periods consistent, EPA considered reducing 
the length of the review period for initial certifications. EPA 
considered both the

[[Page 28045]]

time it takes to complete a thorough technical review of an application 
and the time it takes to resolve issues raised during that technical 
review. The resolution of issues raised during a review can take a 
significant amount of time because it involves coordination between the 
source submitting the application, the applicable state and/or local 
air agency, the applicable EPA Regional Office, and the Acid Rain 
Division at EPA headquarters. Therefore, even though EPA would 
anticipate receiving fewer recertification applications under today's 
proposed revisions, EPA believes that a 120-day review period is 
necessary for recertifications (which, according to today's proposed 
definition of a recertification event, would involve the review of 
monitoring system accuracy tests) in order to coordinate resolution of 
issues raised during the technical review of an application.
    EPA recognizes that there are concerns with increasing the 
recertification review period to longer than 60 days, as more hours of 
data could be invalidated if an application were disapproved. However, 
EPA believes that the criteria for approval of monitoring system 
certification tests are clear and that when an application is 
submitted, the owner/operator should know whether or not the 
performance specifications of part 75 have been met. In EPA's 
experience of four years of implementation, disapprovals are rarely 
issued; in fact, less than 2 percent of all monitoring system 
applications submitted between 1993 and September 1997 were disapproved 
(see Docket A-97-35, Item II-A-4). In most cases where applications 
have been disapproved, the owner or operator should have been aware of 
the deficiencies before the application was submitted. Additionally, 
EPA has found that a longer review period has allowed more time to 
resolve minor deficiencies which could have served as grounds for 
disapproval, but which, given sufficient time, were often resolved 
without issuing a notice of disapproval and without invalidating any 
hourly emissions data.
2. Disapproval of an Incomplete Application
    Section 75.20(a)(4) of the existing rule requires EPA to issue a 
``notice of approval or disapproval of the certification application 
within 120 days of receipt of the complete certification application.'' 
This provision implies that an application must be complete in order to 
issue a disapproval. In attempting to implement this provision, EPA has 
encountered the problem of incomplete applications. The Agency has, in 
most of these instances, issued a notice of incompleteness to the 
source. However, affected sources have not always complied with the 
incomplete notices and have sometimes failed to submit the information 
requested to complete the application in a timely manner. Therefore, 
EPA proposes to clarify that EPA may disapprove an incomplete 
certification or recertification application if the submittal deadline 
is passed. Before a disapproval would be issued for an incomplete 
application, the designated representative would receive a notice of 
insufficiency and be given a reasonable period of time to complete the 
application. If the complete application was not received by this 
extended deadline, EPA could issue a notice of monitoring system 
disapproval. The Agency believes that this provision will result in 
faster resolution of incomplete certification or recertification 
applications, thereby eliminating extended periods of uncertainty about 
data validation status.
3. Submittal Requirements for Certification and Recertification 
Applications
    The current rule requires the owner or operator to submit 
certification and recertification applications to the Administrator 
(i.e., the Acid Rain Division of EPA) and to the appropriate EPA 
Regional Office and state or local air agency. Hardcopy test results 
must be submitted, as well as an updated monitoring plan and electronic 
test results. The electronic test results must also be submitted to the 
Administrator as part of the next quarterly report.
    Sections 75.20(a)(4)(ii), 75.59, and 75.63 of today's proposal 
would revise and clarify the completeness, format, and submittal 
requirements for certification and recertification applications. For a 
certification or recertification application to be considered complete, 
the appropriate information specified in proposed Sec. 75.63 would be 
sent to the Administrator, to the EPA Regional Office, and to the state 
and local air agency. Under proposed Sec. 75.63, the Administrator 
would receive only a hardcopy application form and would not receive 
any hardcopy test results, unless specifically requested. The 
Administrator would, however, receive certification and recertification 
test results electronically in the quarterly report. In most cases, the 
electronic test results would be submitted in the quarter in which the 
testing is completed. However, there may be occasional exceptions to 
this, for initial certification testing and for recertification 
testing, when a series of tests spans two consecutive calendar 
quarters.
    The local and State agencies, as well as the EPA Regional Office 
would receive a hardcopy application form, electronic test results, and 
hardcopy test results. For recertification tests, today's proposal 
would allow the EPA Regional Office or the state or local air agency to 
waive the requirement for a hardcopy recertification test report for 
their respective offices. The EPA Regional Office or the state or local 
agency could also reinstate that requirement at a later date. EPA 
Regional Offices and state and local agencies have historically 
received hardcopy certification and recertification reports with 
varying contents and formats. Section 75.59(a)(10) would specify the 
minimum content for hardcopy certification and recertification reports 
for gas and stack flow CEMS. Section 75.63(a)(2)(iii) would limit the 
amount of reporting for ``non-recertification events'' that require 
diagnostic tests. For a diagnostic test, the only reporting requirement 
would be to submit the applicable electronic test results in the next 
quarterly report. For DAHS verifications, no reporting would be 
required; instead, records of the tests would be maintained on-site in 
a manner suitable for inspection.
    This series of revisions is intended both to clarify the elements 
of a complete application, and to clarify how and to whom the essential 
information should be submitted. By not requiring hardcopy test reports 
to be sent to the Administrator and by allowing the EPA Regional Office 
or state or local agencies to waive hardcopy recertification test 
reports, the Agency believes that unnecessary hardcopy reporting to 
offices that do not intend to review the reports will be eliminated.
    Finally, Sec. 75.63(b) would clarify that for failed certification 
or recertification tests, only tests that affect data validation would 
need to be reported. For example, if the ordinary rules of data 
validation, rather than the retrospective validation procedures, were 
applied and a test failure occurred during the initial certification 
testing for a new unit, only the passed test would be reported if the 
test was subsequently repeated and passed. However, if the conditional 
data validation procedures set forth in Sec. 75.20(b)(3) of today's 
proposal had been utilized during that same initial certification, the 
failed test would have to be reported because it would affect the data 
validation of hourly emissions.

[[Page 28046]]

4. Decertification Applicability
    The proposed revisions to Sec. 75.21(e)(1) would clarify that 
excepted monitoring systems under Appendix D, E, or I or an alternative 
monitoring system under subpart E may be decertified in accordance with 
Sec. 75.21(e)(1). The proposed revisions would also clarify that 
decertification would apply to both an initial certification and a 
recertification. EPA believes that logic and consistency dictate the 
need for these changes.
5. Recertification Test Notice
    Section 75.61(a) would be revised to reduce the burdens associated 
with submitting notices of recertification tests. The proposed 
revisions would allow EPA or the state agency to waive notification 
requirements for recertification tests. Currently, a designated 
representative must notify EPA and the state agency prior to commencing 
certification or recertification testing so that EPA or a state 
representative has an opportunity to observe the testing. Allowing the 
recertification notification requirement to be waived and providing 
more media options for notifications will help conserve paper, reduce 
the reporting burden, and provide more flexibility to facilities when 
scheduling tests. In addition, the Agency solicits comment on whether 
Sec. 75.61 should be revised to state that the requirement for written 
notification could be satisfied by mail, facsimile, or e-mail, subject 
to approval by the agency receiving the notification.
6. Monitoring Plans
    In Secs. 75.53(e) and (f), which are revised versions of 
Sec. 75.53(c) and (d), and Sec. 75.62, today's proposal clarifies 
completeness and formatting requirements for monitoring plans. In 
Sec. 75.53(e), the existing provisions would be separated into two 
separate paragraphs (e)(1) and (e)(2) to clarify which parts of the 
monitoring plan must be submitted in electronic format and which 
elements must be submitted in hardcopy format. In addition, a number of 
minor changes would be made to clarify the actual required content of 
the plan. Similarly, in Sec. 75.53(f), the same type of revisions would 
be made to clarify the electronic versus hardcopy elements of 
monitoring plans for specific situations (Appendix D, E, and I units, 
units claiming an opacity exemption, and units with add-on emission 
controls). These proposed revisions are generally consistent with 
existing implementation of the monitoring plan reporting requirements 
and primarily would serve to clarify possibly ambiguous elements of the 
current rule. The revisions reflected in Sec. 75.53(e) would add a 
requirement to electronically report in the monitoring plan the unit 
stack height above ground level and the stack base elevation above sea 
level. EPA understands that these data are readily available to unit 
owners and operators. EPA collects stack heights for some units, e.g., 
for new or modified sources subject to 40 CFR Sec. 51.166. However, 
stack height data is not currently collected for all of the units 
affected under title IV of the Act. Moreover, the stack height data 
that the Agency has is inconsistent, i.e., some of the data are for 
stack height above sea level, some are for above ground level, and some 
are undefined. Stack height data is necessary to improve the modeling 
of plume height and transport of sulfates and nitrates as part of 
acidic deposition and other atmospheric modeling. EPA conducts 
atmospheric modeling as part of the congressionally-mandated program of 
air pollution monitoring, analysis, modeling, and inventory research 
under section 103 of the Act. Such modeling is also used to analyze the 
impact of the Act on the public health, economy, and environment, 
pursuant to section 312 of the Act. (See also, e.g., Human Health 
Benefits From Sulfate Reductions Under Title IV of the 1990 Clean Air 
Act Amendments at 3-6 through 3-11 (EPA, 1995)). EPA is also proposing 
to collect the Energy Information Administration (EIA) flue 
identification numbers associated with each unit. While this data is 
already reported to EIA, it is difficult to correlate it with the unit 
and stack level data reported to EPA. By having sources specify for 
each unit and stack the corresponding flue identification number 
reported to EIA, it will be easier to correlate the emissions data 
reported to EPA to other data that is reported to EIA and is used for 
atmospheric modeling purposes, such as stack exit temperature and 
velocity.
    Section 75.62 would be revised to clarify which parts of the 
monitoring plan must be submitted to the EPA Regional Office and state 
and local agencies, and when such submittals are required. The 
Administrator would receive an electronic monitoring plan at the 
following times: (1) no later than 45 days prior to the initial 
certification application; (2) at the time of a recertification 
application, if a change in the hardcopy monitoring plan information is 
associated with the recertification event; and (3) in each electronic 
quarterly report. The EPA Regional Office and state and local agency 
would receive the required hardcopy monitoring plan 45 days prior to an 
initial certification. Thereafter, hardcopy monitoring plan information 
(changed portions, only) would be submitted as follows: (1) with a 
recertification application, if a change in the hardcopy monitoring 
plan information is associated with the recertification event; and (2) 
within 30 days of any other event with which a hardcopy monitoring plan 
change is associated. Finally, today's proposed rule would require a 
complete monitoring plan to be kept on-site in a form suitable for 
inspection (this could include an electronic portion which could be 
printed out for inspection). These revisions are intended to clarify 
the monitoring plan format and submission requirements, but are 
generally consistent with existing practices.
    Today's proposal would also clarify when revisions must be made to 
the monitoring plan. Currently, only changes that require 
recertification require monitoring plan revisions. The EPA recognizes, 
however, that many changes affecting the information in a monitoring 
plan would not require recertification. Therefore, Sec. 75.53(b) would 
be revised to require that the owner or operator update a monitoring 
plan whenever information in the monitoring plan changes (e.g., a 
change to a serial number for a component of a monitoring system), and 
Sec. 75.62 would require submission of the revised monitoring plan in 
the next quarterly report or, for hardcopy portions, within 30 days of 
the change. This revision would assure that the monitoring plan does 
not contain outdated, erroneous information.
    Section 75.64(a) would clarify that no hardcopy monitoring plan is 
to be submitted with a quarterly report.
7. Submittal Requirements for Petitions and Other Correspondence
    Section 75.60(b)(5) would clarify what hardcopy information is sent 
to the Administrator for petitions and other communications. These 
revisions would clarify the existing rule, but would not represent a 
significant change in the requirements for these types of submittals.

F. Substitute Data

1. Missing Data Procedures for CO2 and Heat Input
Background
    In the May 17, 1995 rule, two new sections, Secs. 75.35 and 75.36, 
were added to part 75. These two new sections provided, respectively, 
missing data procedures for CO2 and heat input,

[[Page 28047]]

which were not provided in the original January 11, 1993 rule. Section 
75.35 specifies that for CO2, the initial missing data 
procedures of Sec. 75.31 are to be followed for the first 720 quality 
assured monitor operating hours following initial certification. 
Thereafter, provided that the CO2 data availability (as of 
the last hour of the previous quarter) is maintained above 90.0 percent 
and provided that the length of any CO2 missing data period 
does not exceed 72 consecutive hours, a simple average of the ``hour 
before'' and ``hour after'' CO2 concentrations is used to 
fill in missing data periods. However, if the monitor availability as 
of the last hour in the previous quarter is below 90.0 percent or if a 
CO2 missing data period exceeds 72 consecutive hours in 
length (regardless of the percent monitor availability), then the fuel 
sampling procedures of Appendix G must be used to provide substitute 
CO2 data.
    Section 75.36 has a parallel structure to Sec. 75.35. For units 
that determine unit heat input by using a flow monitor and a diluent 
(CO2 or O2) monitor, the initial missing data 
procedures of Sec. 75.31 are to be followed for the first 720 quality 
assured monitor operating hours (for the diluent monitor) and for the 
first 2,160 quality assured monitor operating hours (for the flow 
monitor), following initial certification. Thereafter, the standard 
missing data procedures of Sec. 75.33 are to be followed for the flow 
monitor. For the diluent monitor, the on-going missing data provisions 
of Sec. 75.36 are nearly identical to those for CO2 in 
Sec. 75.35 (i.e., use an ``hour before hour after'' missing data 
algorithm, provided that the monitor availability is  90.0 
percent and the missing data period length is  72 hours). 
However, when the diluent monitor availability is < 90.0 percent or 
when the diluent missing data period exceeds 72 hours, Sec. 75.36 
specifies that the owner or operator must use the procedures in section 
5.5 of Appendix F to determine the hourly heat input.
    Utility representatives have asked EPA to consider revising the 
missing data procedures for CO2 and heat input (see, e.g., 
Docket A-97-35, Items II-D-20, II-D-30, II-E-13, and II-E-14). The 
utilities object to several elements of the current procedures. They 
suggest that the Appendix G procedures are burdensome and that the 
missing data procedures are considerably different from the standard 
missing data procedures for SO2, NO_X, and flow 
rate, which are based solely on historical data and monitor 
availability and require no additional procedures such as fuel 
sampling.
Discussion of Proposed Changes
    EPA has reconsidered the provisions of Secs. 75.35 and 75.36 in 
light of the concerns raised by the regulated community, and is 
proposing revisions to the diluent gas missing data procedures for 
CO2 and for heat input determinations. The Agency proposes 
that the same missing data routines prescribed in Sec. 75.33(b) for 
SO2 pollutant concentration monitors also be applied to the 
CO2 and O2 data streams that are used to 
determine CO2 emissions and heat input. The diluent gas 
substitute data values would therefore be determined in a purely 
mathematical way, based on historical data and the percent monitor data 
availability; no fuel sampling procedures would be required.
    Note that these proposed revisions would require the percent 
monitor data availability to be known on an hourly basis. This would 
require the percent availability for CO2 and O2 
monitors to be updated hourly within the data acquisition system. EPA 
realizes that this would involve software modifications, and in cases 
where the unit heat input is determined using a flow monitor and an 
O2 diluent monitor in accordance with Equation F-17 or F-18, 
some new recordkeeping provisions would also be required. The necessary 
recordkeeping provisions have been proposed in Sec. 75.57(g). To allow 
time for software revisions to be made, the revised missing data 
procedures in Secs. 75.35 and 75.36 would not take effect until January 
1, 2000. The owner or operator could, however, opt to use the new 
procedures prior to January 1, 2000.
    EPA believes that today's proposed revisions to the missing data 
procedures for CO2 and heat input determinations would be 
relatively easy to implement because the missing data routines for 
SO2 monitors are well-established and are familiar to both 
the regulated community and to software vendors. The Agency believes 
that the proposed revised missing data procedures would ensure that 
data availability remains high and would, over time, reduce the cost of 
compliance with the requirements of part 75.
2. Prohibition Against Low Monitor Data Availability
Background
    Under the current rule, when a unit uses SO2, flow rate, 
and NO_X monitoring systems to account for its emissions, for 
each clock hour in which a CEMS fails to provide quality assured data, 
a substitute data value must be reported to EPA in accordance with the 
standard missing data procedures of Sec. 75.33. The method required for 
determining the appropriate substitute data values under Sec. 75.33 
depends on several factors, such as the overall monitor data 
availability and the length of the missing data period. For monitor 
data availabilities  90.0 percent, the substitute data value 
(which is reported for each clock hour of the missing data period) will 
normally be the arithmetic average of the readings from the hour before 
and the hour after the missing data period. At other times, it will be 
the 90th (or 95th) percentile value from a lookback period of 720 (for 
SO2) or 2,160 (for NO_X and flow rate) quality 
assured monitor operating hours. When the data availability drops below 
90.0 percent, the substitute data value for SO2 will be the 
maximum concentration recorded in the last 720 quality assured monitor 
operating hours, and for flow rate and NO_X, the substitute 
data value will be the maximum flow rate or NO_X emission 
rate recorded in the last 2,160 quality assured monitor operating hours 
at the corresponding load range.
    Based on four years of program implementation, EPA believes that 
the standard missing data procedures need to be strengthened. As 
presently written, the missing data algorithms lack a safeguard which 
will ensure that high monitor data availability continues to be 
maintained in future years. In the current version of Sec. 75.33, no 
distinction is made between data availabilities of 89.0 percent, 50.0 
percent or 10.0 percent. For all three of these data availability 
percentages, the substitute data value is the same (i.e., the maximum 
value in a lookback period of 720 or 2,160 quality-assured monitor 
operating hours). This has potentially serious consequences. For 
example, if the substitute data value from the lookback period is non-
punitive or perhaps is even favorable to the facility (e.g., if a low-
sulfur fuel was burned during the lookback period), there would be 
little incentive to repair a malfunctioning CEMS in a timely manner and 
emissions could possibly be under-reported for a long period of time. 
Currently, part 75 does not specifically address this ``gaming 
activity.''
Discussion of Proposed Changes
    In order to maintain the credibility of the SO2 
allowance accounting system and to ensure that affected units continue 
to comply with their part 76 NO_X emission limits, monitor 
data availability must not be allowed to deteriorate indefinitely 
without clear and significant consequence to the facility. Therefore, 
in today's rulemaking, EPA is proposing to add a

[[Page 28048]]

safeguard to part 75 to ensure that this does not happen. A new 
paragraph 75.33(d) would be added, which would make it a violation of 
the primary measurement requirement of Sec. 75.10(a) to allow the 
annual monitor data availability to drop below 80.0 percent for 
SO2, NO_X, flow rate, or CO2. Based on 
an analysis conducted on data availability information for the third 
quarter of 1996, EPA believes that affected facilities will easily be 
able to comply with the 80.0 percent data availability criterion (see 
analyses in Docket A-97-35, Item II-B-16). The results of that analysis 
indicated a mean percent monitor data availability of 96.9 percent for 
SO2, 95.0 percent for NO_X, and 96.6 percent for 
flow rate. Although there were 13 (out of 995 total) SO2 
monitors, 21 (out of 997 total) flow monitors, and 46 (out of 1365 
total) NO_X monitoring systems with percent monitor 
availabilities below 80.0 percent in the 4th quarter of 1996, the 
Agency expects that many of these systems would be exempt from the 
prohibition based on a limited number of operating hours in the 
previous year (see Docket A-97-35, Item II-A-8).
    The proposed prohibition would not apply to units that have only a 
limited number of operating hours (less than 3000 hours of operation in 
the previous 12 calendar quarters) because such units can have a low 
data availability percentage without necessarily having extended 
monitor downtime incidents. In addition, no violation would occur if 
the low monitor availability is caused by a sudden and reasonably 
unforeseeable event beyond the control of the owner or operator (such 
as destruction of monitoring equipment by fire or flood). The owner or 
operator would, however, be required to notify the Administrator, in 
writing, within 7 days of the occurrence of such catastrophic events 
and also to provide notification to the EPA Regional Office and to the 
appropriate State agency. The owner or operator would be further 
required to submit a corrective action plan, including an 
implementation schedule. Thus, this proposed prohibition should not 
result in violations of part 75, except for situations involving poor 
operation and maintenance practices, which are clearly not beyond the 
control of the owner or operator.
    Another option considered by the Agency was to modify the standard 
missing data algorithms for SO2, NO_X, and flow 
rate as follows. Under this option, the algorithms for monitor data 
availabilities of 90.0 percent to 100.0 percent would remain unchanged. 
The algorithms currently used for all monitor data availabilities below 
90.0 percent would be retained, but these would apply only to data 
availabilities between 80.0 percent and 89.9 percent. Finally, a new 
algorithm would be added for monitor data availabilities below 80.0 
percent. When the data availability drops below 80.0 percent, the 
appropriate maximum substitute data value would have to be used (i.e., 
the maximum potential concentration for SO2 or 
CO2, the maximum NO_X emission rate, or the 
maximum potential flow rate). EPA believes that requiring maximum 
values to be reported when the data availability drops below 80.0 
percent would provide incentive to the affected sources to keep their 
monitors well-maintained. Because any changes to the standard missing 
data algorithms would require software modifications, this option, if 
adopted, would not take effect until January 1, 2000. The Agency has 
not proposed this option because it would require software changes for 
all affected units even though very few units have data availabilities 
that fall below 80.0 percent. The Agency seeks comment, however, on 
whether this option should be used instead of the proposed prohibition 
given that it is more consistent with the structure of the missing data 
requirements in part 75 and would be self-implementing without any need 
to initiate enforcement actions to achieve the desired result of 
continued high data availabilities that assure accurate reporting of 
emissions.
    The Agency also emphasizes that the required data availability for 
the Acid Rain Program would remain at 100.0 percent even if the 
proposed prohibition is adopted, meaning that substitute data would 
have to be supplied for any periods in which data from a certified 
monitoring system are not available. This approach is in sharp contrast 
to most other CEMS programs that do not rely on substitute data. In 
those programs, the Agency, as well as State and local agencies, expect 
and often require much higher data availabilities than 80.0 percent. 
Based on the number of units with data availability higher than 95.0 
percent under the Acid Rain Program, CEMS data availability less than 
95.0 percent may well indicate a failure to properly operate and 
maintain a CEMS. Many agencies rely on that 95.0 percent availability 
level to target systems for inspection and other compliance-related 
follow-up actions. In addition, agencies have adopted various required 
minimum data availabilities for CEMS that far exceed the 80.0 percent 
level selected for the prohibition proposed in today's rulemaking.
    It is also important to note that monitor availability under part 
75 and monitor downtime under other programs are not always the same. 
Under part 75, a source may have actual monitoring data that are 
suspect, based on an evaluation of various quality assurance 
activities. In this situation, the owner or operator may, as a 
conservative measure, report substitute data rather than the actual 
data. In contrast, this type of missing data substitution does not 
occur under most other programs. In most programs, the suspect data 
would simply be invalidated and no emission data would be reported for 
those hours.
    Therefore, because of the structure of the missing data provisions 
in the Acid Rain Program and the generally applicable economic 
incentive to achieve high data availabilities under part 75, it would 
be improper to equate the proposed prohibition in today's rulemaking 
with a required minimum data availability requirement established for 
other programs that do not have the same features. The Agency does not 
intend that this proposed provision should serve as a precedent for 
evaluating the appropriate achievable data availability for other 
programs. Consistent with current practices, the Agency would continue 
to expect CEMS to achieve high data availability and that, generally, 
monitor downtime in excess of 5.0 percent may warrant appropriate 
investigation and follow-up activities.

G. General Authority to Grant Petitions Under Part 75

Background
    Section 75.66(a) provides generally that a designated 
representative of a unit subject to part 75 may submit a petition to 
the Administrator. Sections 75.66(b) through (h) address petitions to 
the Administrator on the specified topics of alternative flow 
monitoring methods, alternatives to standards incorporated by 
reference, alternative monitoring systems, parametric monitoring 
procedures, missing data for units with add-on emission controls, 
emission or heat input apportionments, and the partial recertification 
process. Each of these subsections set forth the items which must be 
included with a particular type of petition. In addition, Sec. 75.66(i) 
states that, for any other petition to the Administrator under part 75, 
the designated representative for an affected unit shall include 
sufficient information for the evaluation of such petition.

[[Page 28049]]

Discussion of Proposed Changes
    Today's proposal would revise Sec. 75.66(a) to state clearly that 
the designated representative of an affected unit may petition the 
Administrator for authorization to apply an alternative to any 
requirement under part 75 or incorporated by reference in part 75, 
regardless of whether another section of part 75 explicitly allows such 
a petition concerning the particular requirement. EPA views this change 
as a clarification to the general authority already provided by 
Secs. 75.66(a) and (i). The proposed rule would also be amended to 
include new paragraphs (i) through (l), which would set forth the 
specific requirements for other petitions that are explicitly allowed 
by other sections of the rule but which are not currently included in 
this section. In addition, the proposed rule, at Sec. 75.66(m), would 
also indicate the appropriate documentation to be submitted for 
petitions under subsection (a), except those under subsections (b) 
through (l), where the required documentation is already specified. The 
required documentation in subsection (m) would be: (1) Identification 
of the unit; (2) information explaining why the proposed alternative 
should be used instead of the existing part 75 provision; (3) 
descriptions and, if applicable, diagrams of the equipment and 
procedures to be used in the proposed alternative; and (4) information 
demonstrating that the proposed alternative is consistent with the 
purposes of the provision for which an alternative is requested and is 
consistent with the purposes of part 75 and of section 412 of the Act.
Rationale
    As presently codified, EPA is concerned that the rule does not 
state clearly what types of petitions may be submitted under 
Sec. 75.66. In particular, existing subsection (i) could be interpreted 
as referring only to petitions that are mentioned in other sections of 
part 75 and that are not specifically listed in Sec. 75.66(b) through 
(h). EPA has not interpreted Sec. 75.66(i) in this manner. In 
administering the Act, EPA has inherent discretion to grant de minimis 
exceptions from statutory or regulatory requirements, where EPA 
determines that holding the regulated entity to the applicable 
requirement would yield a gain of trivial or no benefit, provided 
Congress has not unambiguously demonstrated its intent to foreclose 
such exceptions. See, e.g., Public Citizen v. Young, 831 F.2d 1108, 113 
(D.C. Cir. 1987); Alabama Power Co. v. Costle, 636 F.2d 323, 360-61 
(D.C. Cir. 1979). Since the issuance of part 75 in 1993, EPA has 
accepted, and, in some cases exercised its discretion and granted, 
petitions under Sec. 75.66 that requested exceptions and that were not 
specifically referenced in Sec. 75.66(b) through (h) or elsewhere in 
part 75 (see Docket A-97-35, Item II-B-17). Such petitions have 
included, for example, a request to set a CO2 span lower 
than that required by part 75 in order to more accurately quality 
assure the CO2 monitor. Another petition requested an 
alternative to the requirement to perform an annual RATA on a unit that 
was scheduled to shutdown, prior to the deadline for performing the 
RATA, in order to install a scrubber, construct a new stack, and 
install and certify new CEMS. A petition was also submitted for 
permission to use a propane sampling frequency as specified in the 
State operating permit and to then calculate SO2 emissions 
by using the highest sulfur content recorded during the previous 365 
days and report these data in quarterly reports. These petitions were 
submitted for the purpose of requesting alternatives to various 
requirements of part 75, even though the ability to petition the Agency 
on these issues was not referenced explicitly in other sections of part 
75 or in Sec. 75.66(b) through (h). In most cases, the circumstances 
leading to the request for an alternative to a part 75 requirement were 
not anticipated during the drafting of part 75 regulations. In fact, 
today's proposal revises several part 75 requirements to allow for 
alternatives that were originally requested and approved through the 
petition process set forth in Sec. 75.66. The Agency continues to 
believe that the general provision allowing petitions for alternatives 
to part 75 requirements is necessary to enable EPA to address 
circumstances that were not foreseen during the development of such 
requirements. This is important since circumstances can sometimes vary 
significantly from boiler to boiler. While the response to comment 
document for the January 11, 1993 rule (see Docket A-91-69, Item V-C-1, 
Issue # M-8.8.2) might be read to bar petitions for exceptions from any 
provision of part 75, EPA maintains that such a reading would be 
inconsistent with the regulatory language of Secs. 75.66(a) and (i) 
that allow such petitions, and with the established practice of the 
Agency in administering part 75.
    The existing Sec. 75.66(i) states that for petitions other than 
Sec. 75.66(b) through (h) petitions submitted under the section, the 
designated representative should include sufficient information for the 
evaluation of the petition. No other information is provided concerning 
the contents of such petitions. As Secs. 75.66(b) through (h) all 
provide a list of the type of information that should be included in 
petitions submitted under the respective sections, the Agency believes 
that, in addition to amending Sec. 75.66(a) to clarify that petitions 
may be submitted for circumstances that may not be covered by other 
sections authorizing petitions to the Administrator, it is appropriate 
to provide units with a list of the type of information that should be 
included with the petition. Similarly, EPA believes that it is 
appropriate to add to the section provisions setting forth the 
information requirements for those petitions that are explicitly 
allowed under other sections of part 75 but that are not listed in the 
existing Sec. 75.66. All these revisions would make the petition 
process more uniform and minimize confusion regarding what information 
EPA would require in order to accept and consider any petition for an 
alternative to a part 75 requirement.

H. NO_X Mass Monitoring Provisions for Adoption by 
NO_X Mass Reduction Programs

Background
    Part 75 contains requirements for monitoring NO_X 
emissions with a continuous emission monitoring system or other 
approved method. Owners and operators are required to calculate hourly, 
quarterly average, and annual average NO_X emission rates (in 
lb/mmBtu). Part 75, however, currently contains no requirements for 
reporting NO_X mass emissions (in tons). Other NO_X 
emission reduction programs being developed pursuant to title I of the 
Act (such as the NO_X Budget Program in the Ozone Transport 
Region) are expected to require reporting of NO_X mass 
emissions from many of the units affected under the Acid Rain Program. 
To streamline reporting burdens under multiple programs and to allow 
for the administration of multi-state NO_X mass trading 
programs, the Agency believes it appropriate to amend part 75 to 
include provisions for monitoring, recording, and reporting 
NO_X mass emissions that could apply to such trading 
programs. These provisions would provide standard procedures--resulting 
in precise, reliable, accessible, and timely emissions data--that could 
be adopted under a state or federal NO_X mass emission 
reduction program. To the extent that these standard provisions are 
adopted, the burden on industry would be reduced and the administration 
of the programs would be facilitated, in

[[Page 28050]]

that the Agency or implementing states would not need to develop 
NO_X mass monitoring provisions anew and industry would not 
need to become familiar with multiple approaches to NO_X mass 
monitoring.
Discussion of Proposed Changes
    The proposed NO_X mass emissions provisions would apply 
only where EPA, states, or groups of states incorporate them and 
mandate their use through a separate regulatory action. The proposed 
amendments would make changes to Secs. 75.1, 75.2, 75.4, 75.16, 75.17, 
Appendix D, section 2.1.2.2, and Appendix F, section 5.5. They would 
also add a new subpart H containing new Secs. 75.70, through 75.73 and 
a new section 8 in Appendix F containing sections 8.1, 8.1.1, 8.1.2, 
8.1.3, 8.1.4, 8.2, 8.3, 8.3.1, and 8.3.2.
    Section 75.1, the purpose and scope section, would be amended to 
broaden the scope by adding that part 75 will also set forth provisions 
for monitoring and reporting NO_X mass emissions that EPA, 
states, or groups of states may require sources to use to demonstrate 
compliance with a NO_X mass emission reduction program. 
Section 75.2 would be amended to add that the provisions of part 75 may 
also apply to sources subject to a state or federal NO_X mass 
emission reduction program.
    The compliance date section, Sec. 75.4(a), would be altered to 
state that the provisions relating to monitoring and reporting of 
NO_X mass emissions become applicable on the deadlines 
specified in the applicable state or federal NO_X mass 
emission reduction program requiring the use of part 75 to monitor and 
report NO_X mass emissions.
    Section 75.16 would be amended to state that title IV affected 
units using the provisions of part 75 to monitor and report 
NO_X mass emissions under a state or federal NO_X 
mass emission reduction program would have to meet the heat input 
monitoring and determination requirements in both Sec. 75.16 and in 
subpart H, Secs. 75.71 and 75.72. Section 75.17 would be amended to 
state that title IV affected units using the provisions of part 75 to 
monitor and report NO_X mass emissions under such a program 
would have to meet the NO_X emission monitoring and 
determination requirements in both Sec. 75.17 and subpart H, 
Secs. 75.71 and 75.72.
    The applicable procedures for the monitoring and determination of 
NO_X mass emissions would be added in proposed subpart H, 
Secs. 75.70, 75.71, and 75.72 and corresponding recordkeeping and 
reporting requirements would be set forth in Sec. 75.73.
    Section 75.70 would set forth the general requirements including: 
definitions, compliance dates, incorporation by reference, initial 
certification and recertification procedures, quality assurance and 
quality control requirements, substitute data requirements, and 
requirements regarding petitions. In general these provisions for 
monitoring NO_X mass would mirror the provisions for 
monitoring of SO2, NO_X, and CO2 for 
compliance with title IV. However, because the program would be a state 
program, rather than a federal program, there would be some differences 
in the administrative requirements. These differences would be most 
pronounced for units that were not subject to Acid Rain emission 
limitations and were not already subject to the provisions of part 75. 
The major differences in administrative requirements would involve the 
process for petitioning under Sec. 75.66 and the process for certifying 
and recertifying monitors. Under the existing Acid Rain Program, the 
Administrator must approve all petitions under Sec. 75.66. Under this 
proposal, petitions for units that were only subject to the provisions 
of part 75 because they were subject to a state or federal 
NO_X mass emission reduction program, would have to be 
approved by both the permitting authority for the applicable 
NO_X mass program and the Administrator. The permitting 
authority would also be responsible for reviewing and approving or 
disapproving certification and recertification applications for such 
units.
    Section 75.71 sets forth the general monitoring methodologies that 
would be allowed for different types of units. The proposal would 
require units to determine hourly NO_X mass emissions (in lb) 
by monitoring NO_X emission rate (in lbs/mmBtu) and heat 
input (in mmBtu/hr) on an hourly basis and by multiplying those two 
values and the hourly unit operating time (in hour or fraction of an 
hour) together. Coal units and other units that burn solid fuel and 
that are covered by subpart H would be required to measure 
NO_X emission rate using a NO_X emission rate CEM 
consisting of a NO_X concentration CEM and a diluent CEM 
(CO2 or O2 CEM) and to measure heat input using a 
diluent CEM and a continuous volumetric flow monitor. All gas- and oil-
fired units covered by subpart H would be allowed to use that approach 
or, alternatively, could measure NO_X emission rate using a 
NO_X emission rate CEM and heat input by using a fuel 
flowmeter and performing fuel sampling and analysis. This alternative 
for determining heat input from gas- and oil-fired units is set forth 
in Appendix D of part 75. Gas and oil units that qualify as either 
peaking units or low mass emission units under part 75 would also have 
additional lower cost monitoring methodologies available to them. 
Peaking units, for example, would have the option to do source testing 
to create heat input versus NO_X emission rate correlation 
curves. Then, based on hourly measurement of heat input from a fuel 
flowmeter and fuel sampling and analysis using the provisions in 
Appendix D to part 75, the heat input vs NO_X emission rate 
correlation curves would be used to estimate the hourly NO_X 
emission rate. This rate would be used in conjunction with hourly 
measured heat input to determine NO_X mass. A unit that 
qualifies as a low mass emission unit would have the option to use a 
fuel-type and boiler-type specific default NO_X emission rate 
and the unit's maximum rated hourly heat input to determine 
NO_X mass emissions. The low mass emissions unit provisions 
are in proposed Sec. 75.19.
    Section 75.72 sets forth the specific requirements for monitoring 
emissions at units that share common stacks and/or common pipes, for 
units that emit to multiple stacks and for units that receive fuel from 
multiple pipes. These provisions mirror similar provisions in 
Sec. 75.16 for monitoring SO2 mass emissions from similar 
units and groups of units.
    Appendix D, section 2.1.2.2 would indicate that the heat input 
apportionment procedures of that section would not be applicable for 
units whose compliance with this part is required under a 
NO_X mass emissions reduction program. Instead, the unit 
would have to meet the heat input monitoring and determination 
requirements in subpart H, Secs. 75.71 and 75.72.
    The applicable procedures for calculating NO_X mass 
emissions would be added in proposed section 8 of Appendix F. Section 
8.1 of Appendix F contains proposed equations for determining hourly 
NO_X mass emissions, section 8.2 contains proposed equations 
for determining quarterly, cumulative annual and ozone season 
NO_X mass emissions, and section 8.3 contains specific 
provisions for monitoring NO_X emissions from a common stack. 
Additionally, revisions to section 5.5 of Appendix F would indicate 
that the heat input calculation procedures of section 5.5.3 would not 
be applicable for units whose compliance with this part is required 
under a NO_X mass emissions reduction program.

[[Page 28051]]

Rationale
    (a) Authority to Propose NO_X Mass Provisions. The 
authority for the proposed NO_X mass provisions rests in two 
separate portions of the Act. First, section 412(a) states that the 
owner or operator of an affected source under title IV must monitor and 
quality assure data for sulfur dioxide and nitrogen oxide for each 
affected unit at the source. 42 U.S.C. 7651k(a). This section does not 
limit the nitrogen oxide data requirement to emission rate data in lb/
mmBtu or to data necessary for compliance with emission limits 
established under title IV. Indeed, oil-and gas-fired units have been 
required to report NO_X emission rate data under part 75 even 
though only existing coal units are subject to NO_X emission 
limits under title IV. (See 58 FR 3590, 3644, January 11, 1993). Thus, 
the Agency believes that providing for reporting NO_X mass 
emissions under part 75 is an appropriate exercise of the authority 
under section 412, particularly since NO_X mass emissions 
reporting may be required under a separate applicable requirement.
    Second, independently of the authority granted by section 412, 
section 114(a) of the Act gives the Administrator broad authority to 
collect data for ``the purpose of developing or assisting in the 
development of any implementation plan under section 110 or 111(d)'', 
``of determining whether any person is in violation of any such 
standard or a requirement of such a plan'', or ``carrying out any other 
provision of [the] Act'' (except certain provisions of title II 
concerning mobile sources). Section 114 is, of course, not limited to 
sources that are affected units under title IV. Moreover, section 
301(a)(1) authorizes the Administrator ``to prescribe such regulations 
as are necessary to carry out his functions'' under the Act, including 
the functions specified in section 114. Thus, EPA maintains that the 
Agency is authorized to adopt provisions in part 75 that could govern 
monitoring of NO_X mass emissions, especially where such 
information is expected to support States' efforts to attain ambient 
air quality standards.
    From a policy perspective, now is the appropriate and most 
efficient time to adopt these changes. In July 1997, EPA Administrator 
Carol Browner announced a series of initiatives to reform environmental 
data management and collection (see Docket A-97-35, Item II-I-21). The 
new initiatives are intended to streamline reporting requirements and 
increase coordination across different programs that affect the same 
sources. There are a number of examples of ongoing efforts to 
streamline the reporting of emissions for utility units. One example is 
a proposal to revise the NSPS NO_X standards for utility and 
industrial boilers subject to reporting under 40 CFR part 60. That 
proposal would allow facilities to submit NSPS reports through part 75 
reporting (see 62 FR 36948, July 9, 1997). Another example is the Ozone 
Transport Commission's NO_X Budget program. That program is 
expected to require utility sources and certain industrial sources in 
the northeast to reduce emissions of NO_X through a trading 
program similar to the Acid Rain SO2 trading program. On 
January 31, 1996, the OTC released the Model Rule which outlines 
procedures for the monitoring and reporting of NO_X mass 
emissions; these procedures are based on the monitoring and reporting 
requirements set forth in part 75 (see Docket A-97-35, Items II-I-7 and 
II-I-22). Today's proposal would facilitate the coordination of 
reporting under the Acid Rain Program and NO_X mass programs 
like the OTC NO_X Budget Program.
    In addition, the Agency believes it is appropriate to include these 
requirements in the current proposal because the Acid Rain affected 
units may be undertaking DAHS software changes to respond to the other 
proposed revisions to part 75 if they are adopted. The Agency would 
enable facilities to coordinate the necessary software changes by 
proposing the revised reporting requirements to allow for 
NO_X mass emission reporting at this time along with the 
other part 75 revisions. Although EPA is proposing this requirement now 
to facilitate software changes, the requirement to actually record and 
report NO_X mass emission data under part 75 generally would 
not become effective for any unit unless and until a program requiring 
such recording and reporting is implemented for that particular unit 
(EPA notes that, as discussed elsewhere in Section III.C.4. of this 
preamble, a limited group of title IV affected units (i.e., low mass 
emissions units) would be required to record and report NO_X 
mass emissions for purposes of the Acid Rain Program.) In addition, if 
a state elected to require the use of these requirements to support a 
state NO_X mass emission monitoring and reporting 
requirement, these requirements would not become federally enforceable 
until those requirements were approved by EPA as part of the SIP.
    (b) Monitoring Methodology. The proposed requirement would require 
sources to determine NO_X mass as a function of hourly 
average NO_X emission rates, heat input rates, and unit 
operating time. EPA is proposing this approach because it accurately 
accounts for NO_X mass emissions without requiring any 
changes to the current missing data routines and quality assurance 
requirements in part 75. An alternative to this approach, not included 
in today's proposal, would be to measure total mass emissions using a 
NO_X pollutant concentration monitor, a volumetric flow 
monitor and unit operating time, analogous to the approach taken 
currently for SO2 emissions. This methodology would have two 
advantages: first, there would be less missing data from a 
NO_X pollutant concentration monitor than from a 
NO_X CEMS which (under the existing and proposed rule) 
contains both a NO_X pollutant concentration monitor and a 
diluent monitor; and second, it would avoid possible overestimation 
from a bias adjustment factor applied to the NO_X system to 
correct bias in the diluent monitor (see Docket A-97-35, Item II-D-96).
    However, this methodology would also have a number of 
disadvantages. In order to monitor NO_X as total mass 
emissions using a NO_X pollutant concentration monitor and a 
volumetric flow monitor, several major changes would need to be made to 
part 75. The entire concept of a NO_X CEMS--and the quality 
assurance tests and missing data procedures associated with the 
NO_X CEMS--might need to be revised, to include either a 
NO_X CEMS with only a NO_X pollutant concentration 
monitor and a DAHS (in which case, a separate flow monitoring system 
would also be required in order to determine NO_X mass), or a 
NO_X CEMS with a NO_X pollutant concentration 
monitor, a volumetric flow monitor, and a DAHS. Since the relative 
accuracy standard currently in part 75 for NO_X systems is in 
lb/mmBtu, it would be necessary to establish a new relative accuracy 
standard for NO_X concentration in ppm if the NO_X/
flow method described above were incorporated into the final rule. Bias 
adjustment would also have to occur on the newly defined NO_X 
CEMS. It would also be necessary to create a missing data procedure 
either for NO_X concentration in ppm or for hourly 
NO_X mass emission rate in lb/hr. Hourly NO_X mass 
emission rate would be calculated using the same formula as for 
SO2 mass emission rate (Equation F-1 or F-2), only using a 
constant of 1.194 x 10^-7(lb/scf)/ppm NO_X. In 
addition, this methodology would not easily support the monitoring and 
reporting of NO_X emission rate data in lb/mmBtu.

[[Page 28052]]

Therefore, in order to meet the emission rate reporting requirements, 
affected sources under title IV would still be required to maintain a 
diluent CEMS and the current NO_X emission rate missing data 
procedures. The Agency has not proposed this approach because it does 
not believe that the benefits of slightly reduced amounts of missing 
data for NO_X mass and removal of the bias adjustment factor 
for the diluent monitor justify the complication of having two separate 
procedures for monitoring NO_X emissions from a given unit. 
Nevertheless, the Agency requests comment on whether this approach to 
measuring mass emissions should be used in lieu of the proposed heat 
input and emission rate approach for sources required to report 
NO_X mass.
    (c) Common Stack and Pipe Monitoring. The Agency notes that the 
proposed procedures for monitoring NO_X emission rate at a 
common stack to determine NO_X mass emissions under the 
proposed Sec. 75.72 procedures are different than the procedures 
currently allowed for monitoring NO_X emission rate in 
Sec. 75.17. The Agency is concerned that the Sec. 75.17 provisions 
would be too imprecise for measuring NO_X mass emissions 
because the two values used to determine NO_X mass emissions 
(NO_X emission rate and heat input) are not required to be 
measured at the same location. In the existing rule, NO_X 
emission rate may be monitored at the unit level in the duct leading to 
the common stack and heat input can be determined from measurements at 
the common stack and then apportioned to the individual units using 
unit load. While this heat input apportionment method has been allowed 
for Acid Rain purposes, it is not accurate in all cases because it does 
not account for different heat rates from the units exhausting to the 
common stack and does not account for differences in operating time at 
the units. It has been allowed by the Agency for Acid Rain purposes 
because apportioned heat input determined under Sec. 75.16 (e) had only 
a limited effect on emissions trading (i.e., on the SO2 
allowance program). Although apportioned heat input determined under 
Sec. 75.16(e) is used to determine compliance with the reduced 
utilization provisions of the Acid Rain Program, the apportioned heat 
input estimate was deemed accurate enough for that purpose and for the 
relatively small number of units and short period involved. 
Determinations of reduced utilization are required only for Phase I 
units during 1995-1999 and for opt-in units. However, for purposes of a 
NO_X mass trading program, the heat input value would be used 
in the calculation to determine NO_X mass, and an imprecise 
unit level heat input value could cause the NO_X mass 
emissions from some units to be underestimated. The NO_X mass 
trading program could be undermined by the lack of a consistent 
emissions value for each NO_X allowance. Therefore, the 
proposed provisions for monitoring heat input and NO_X 
emission rate from units in a NO_X mass trading program would 
be similar to the provisions that are currently used for monitoring 
SO2 mass emissions at a common stack at Sec. 75.16. The 
provisions for monitoring SO2 mass emissions require that 
the two values needed to determine SO2 mass emissions, stack 
flow rate and SO2 concentration, be monitored at the same 
location. The Agency is proposing that, for purposes of determining 
NO_X mass emissions, a facility could use the same location 
options currently available for SO2: the facility could 
either monitor both NO_X emission rate and heat input at the 
common stack level or monitor them both at the unit level. The Agency 
is also proposing a third option: heat input could be monitored at the 
unit level and summed to the common stack level, while NO_X 
emission rate could be monitored at the common stack level. Even though 
this option would allow NO_X emission rate and heat input to 
be measured at different locations, it does not have the inherent 
inaccuracies described above because it does not require heat input 
apportionment.
    Similarly, the optional procedures currently allowed for the 
apportionment of heat input measured at a common pipe in Appendix D, 
section 2.1.2.2 are not available for units with a common pipe under 
subpart H. As discussed above for common stacks, the Agency is 
concerned that the heat input apportionment under Appendix D, section 
2.1.2.2 provisions would be too imprecise for the purpose of 
calculating NO_X mass emissions. In the existing rule, heat 
input can be determined from measurements at the common pipe and then 
apportioned to the individual units using unit load. For purposes of 
calculating NO_X mass emissions under subpart H for a unit 
which is supplied fuel from a common pipe, the measurement of fuel flow 
rate would have to be made at the pipe leading to the individual unit 
in order to determine unit level heat input.
    The Agency solicits comment on the proposed approach for monitoring 
NO_X mass emissions at a common stack or pipe and whether it 
is appropriate to mirror the common stack and pipe provisions for 
SO2 mass emissions.
    (d) Multiple duct/stack monitoring. The current provisions for 
monitoring NO_X emission rate, in Secs. 75.17(c)(1) and (2), 
allow the owner or operator to determine NO_X emission rate 
for a unit that exhausts through multiple ducts or stacks using a Btu-
weighted sum of the NO_X emission rates measured in each duct 
or stack or by monitoring NO_X emission rate in only one duct 
or stack. The new proposed Sec. 75.72 would set forth specific 
requirements for monitoring NO_X mass in multiple ducts or 
stacks and would in some cases place a number of limits on the options 
in Sec. 75.17(c) and in some cases not allow the options in 
Sec. 75.17(c). The proposed options for monitoring NO_X mass 
are similar to the existing provision in Sec. 75.16(d) for monitoring 
SO2 mass emissions at multiple ducts/stacks. They are also 
similar to the provisions being used in the OTC NO_X Budget 
Program to determine NO_X mass in similar situations.
    The new proposed Sec. 75.72 does not contain an option for any 
units to use a Btu-weighted sum of the NO_X emission rates 
measured in each duct or stack. The reason that this option is not 
appropriate is that in order to use this option to determine a unit's 
NO_X emission rate, the owner or operator of the unit would 
have to monitor both NO_X emission rate and heat input in 
each duct or stack. (As discussed above, the heat input apportionment 
method allowed under Sec. 75.17 is not sufficiently accurate for a 
NO_X mass program.) These two values allow the calculation of 
NO_X mass and, therefore, there is no reason to determine a 
Btu-weighted sum for purposes of this subpart.
    The new proposed Sec. 75.72 would not allow coal units to monitor 
NO_X emission rate in only one duct or stack. The proposal 
would also not allow gas and oil units to monitor the NO_X 
emission rate in only one duct or stack, unless heat input is 
determined using the provisions of Appendix D to this part and the 
owner or operator makes a demonstration that the emission rate would 
always be the same in both ducts or stacks. Reasons that the emission 
rate might vary include the use of add-on emission controls in the 
ducts or stacks or venting of emissions to one duct or stack and not 
the other.
    These limitations are required for monitoring mass emissions (in 
lbs), but are not necessary for monitoring emission rate (in lbs/mmBtu) 
at coal units or gas and oil units that use continuous volumetric flow 
monitors, because, for reasons discussed above, monitoring mass 
requires the monitoring of both emission rate and heat input. Since the 
amount of stack

[[Page 28053]]

flow that is vented to each duct or stack could vary significantly 
depending upon the location and use of dampers and induction fans in 
the ducts or stacks, it is necessary to measure volumetric flow in both 
ducts or stacks in order to determine heat input for the unit(s). In 
order to accurately use these heat input values to determine 
NO_X mass, it is also necessary to measure NO_X 
emission rate in both ducts or stacks. Therefore, proposed Sec. 75.72 
would require monitoring of heat input and NO_X emission rate 
in both ducts or stacks for coal units and gas-and oil-fired units that 
use continuous volumetric flow monitors and exhaust to multiple ducts 
or stacks.
    Since gas-and oil-fired units that are using the procedures in 
appendix D of part 75 to determine heat input based on fuel consumption 
do not have to measure volumetric flow in the duct or stack in order to 
determine heat input, EPA believes it is appropriate to allow these 
units to measure NO_X emission rate in only one duct or stack 
if they can demonstrate to both the permitting authority and the 
Administrator that the NO_X emission rate in either duct or 
stack is representative of the NO_X emission rate in each 
duct or stack. Therefore, proposed Sec. 75.72 allows gas-and oil-fired 
units that are using the procedures in appendix D of part 75 to measure 
NO_X emission rate in only one duct or stack if they can 
demonstrate to both the permitting authority and the Administrator that 
the NO_X emission rate in either duct or stack is 
representative of the NO_X emission rate in each duct or 
stack.
    (e) Reporting of NO_X Mass Emissions. The Agency also 
notes that the proposed procedures differ in two key respects from the 
way data is currently reported under part 75. The first difference is 
that the proposal would require reporting of hourly NO_X mass 
emissions, in lbs, (instead of hourly mass emission rate, in lb/hr, as 
is currently required for the reporting of SO2 under part 
75). The OTC NO_X Budget Program is expected to require the 
reporting of hourly mass emissions, in lb, rather than hourly mass 
emission rates, in lb/hr, because of experience under the Acid Rain 
Program with reporting hourly SO2 and CO2 mass 
emission rates. As discussed in Section III.R.1 of this preamble, the 
reporting of hourly SO2 and CO2 mass emission 
rates has been a source of some confusion in the implementation of the 
Acid Rain Program. For the reasons presented in Section III.R.1 of this 
preamble, EPA is not proposing to change the existing SO2 
and CO2 reporting requirements. However, the existing part 
75 does not require any NO_X mass emission reporting, and in 
order to avoid the problems experienced under the Acid Rain Program and 
to be consistent with the OTC NO_X Budget Program, EPA 
proposes here to base the new NO_X reporting on mass 
emissions in pounds. Maintaining consistency with the provisions 
expected to be adopted for the OTC NO_X Budget Program is 
important to ensure that a central body such as EPA would be able to 
effectively administer the program if states opted to participate in a 
multi-state NO_X trading program larger than the Ozone 
Transport Region covered by the OTC NO_X Budget Program.
    The second key difference is that, in addition to reporting a 
quarterly and cumulative annual total emissions value, the proposed 
revisions would also require reporting of a cumulative ozone season 
total value. Generally, the ozone season extends from May 1 to 
September 30 of every year. The cumulative ozone season emissions would 
be reported with the second quarter and third quarter reports submitted 
to EPA. The reason that reporting would be required on an ozone season 
basis is that one of the main reasons the data is being collected is to 
support other programs designed to control emissions during the ozone 
season.
    (f) Role of EPA and States/Localities in Administering the 
Monitoring Portion of a NO_X Trading Program. The Agency also 
notes that another important potential difference between the use of 
this part to support the Acid Rain Program under Title IV of the CAA 
and the use of this part to support other NO_X mass emission 
reduction programs is the role that EPA and the state or local 
permitting authority that may establish such a program will play. Under 
the Acid Rain Program, even though many states have assumed the role of 
the permitting authority under Phase II of the program, EPA still 
retains authority to issue approvals and disapprovals related to all of 
the monitoring and reporting issues, such as certification of 
monitoring systems under Sec. 75.20, approval of petitions under 
Sec. 75.66 and approvals of alternate monitoring petitions under 
Sec. 75.48. EPA believes that if a NO_X mass emission 
reduction program is approved as part of a SIP or if EPA agrees to work 
with individual or groups of states to help administer the monitoring 
and reporting portion of a NO_X mass emission reduction 
program, EPA would still have to be involved in the approval process.
    The level of this involvement might vary depending upon the 
specific type of approval or disapproval. It also would vary depending 
upon whether or not the unit had an Acid Rain emission limitation. For 
instance, EPA would play a significant role in the approval of an 
alternate monitoring petition under Sec. 75.48 or any other petitions 
under Sec. 75.66. For a unit with an Acid Rain emission limitation, any 
petition would already have to be approved by EPA. In order to 
streamline the process for these sources, EPA believes that EPA should 
continue to issue approvals and disapprovals of petitions. However, 
since sources would also be using the monitored data to meet SIP 
requirements, EPA would take this action in consultation with the 
applicable state. For units that are not subject to an Acid Rain 
emission limitation, EPA would still need to be involved in petition 
determinations. There are two primary reasons that this involvement 
would be necessary. The first would be as part of EPA's typical role in 
assuring that any alternative to the approved SIP will still result in 
the air quality benefit that would have been derived if the permitting 
authority had not deviated from the SIP. The second would be as part of 
EPA's role in administering the emissions tracking portion of a 
NO_X mass emission reduction program. If EPA was not involved 
and a state approved, for a unit, an alternative that allowed 
variations to the reporting requirements, EPA might not be able to 
administer the emissions tracking portion of the program for that unit. 
Similarly, for approval and disapproval of certification applications 
and recertification applications, EPA believes that there should be two 
separate requirements; one for units subject to an Acid Rain emission 
limitation, and one for units not subject to an Acid Rain emission 
limitation. For units subject to an Acid Rain emission limitation, EPA 
would still approve or disapprove certification and recertification 
applications. This would streamline the process for units since they 
would only have to deal with one regulatory agency for both programs. 
For units not subject to an Acid Rain emission limitation, the 
permitting authority would approve certification and recertification 
applications. EPA requests comment on this approach and whether the 
respective roles of the Administrator and the permitting authority 
should be different for units that are subject to both an Acid Rain 
emission limitation and to a NO_X mass emission reduction 
program and for units that are subject solely to a NO_X mass 
emission reduction program.

[[Page 28054]]

I. Span and Range Requirements

Background
    The span and range requirements for part 75 continuous emission 
monitoring systems are found under section 2.1 of Appendix A to the 
January 11, 1993, rule, as amended on May 17, 1995. Sections 2.1.1, 
2.1.2, 2.1.3 and 2.1.4 of Appendix A give the specific span and range 
requirements for SO2 monitors, NO_X monitors, 
diluent (O2 and CO2) monitors, and flow rate 
monitors, respectively.
    The span of a CEMS provides an estimate of the highest expected 
value for the parameter being measured by the CEMS. For instance, the 
span value of an SO2 monitor should be an approximation, 
based on the type of fuel being combusted, of the highest 
SO2 concentration likely to be recorded by the CEMS during 
operation of the affected unit. The range of a CEMS is the full-scale 
setting of the instrument. Under part 75, the range of a monitor must 
be equal to or greater than the span value. Section 2.1 of Appendix A 
further specifies that the range must be chosen such that the majority 
of the readings during normal operation fall between 25.0 and 75.0 
percent of full-scale. Part 75 span values are used to determine the 
appropriate reference gas concentrations and reference signals for 
daily calibration of the CEMS; the reference concentrations and signal 
values are expressed as percentages of the span value. The allowable 
daily calibration error for a CEMS is also expressed as a percentage of 
span.
    Sections 2.1.1 through 2.1.4 of Appendix A to the January 11, 1993 
rule specified procedures for determining the span values for four 
parameters: SO2, NO_X, diluent gas (O2 
or CO2), and volumetric flow rate. For SO2, the 
``maximum potential concentration'' (MPC) was first calculated based on 
fuel sampling results from the previous 12 months (using the highest 
sulfur content and lowest heating value in Equation A-1a or A-1b). The 
SO2 span value was then obtained by multiplying the MPC by 
1.25 and rounding the result upward to the next highest multiple of 
100.0 ppm. The MPC values for NO_X were specified in the rule 
and were based on the type of fuel being combusted (e.g., 800.0 ppm for 
coal-firing and 400.0 ppm for oil-firing). The NO_X span 
value was then determined by multiplying the MPC by 1.25 (e.g., 1000.0 
ppm for coal-firing and 500.0 ppm for oil-firing). For CO2 
and O2, a span value of 20.0 percent CO2 or 
O2 was required for all diluent monitors. For flow rate, the 
``maximum potential velocity'' (MPV) was first determined either using 
Equation A-3a (or A-3b) or from historical test data (i.e., from 
velocity traverses conducted at or near maximum load). Then, the span 
value was obtained by multiplying the MPV by 1.25 and rounding the 
result upward to the next highest multiple of 100 feet per minute 
(fpm).
    In the January 11, 1993 rule, the SO2 or NO_X 
monitor range derived from the MPC was referred to as the ``high-
scale.'' The rule further specified that whenever the majority of the 
readings during normal operation were expected to be less than 25.0 
percent of the high full-scale range value (e.g., if a scrubber were 
used to reduce SO2 emissions), a second, ``low-scale'' span 
and range would be required. The low scale of the CEMS would be defined 
as 1.25 times the ``maximum expected concentration'' (MEC). The 
original rule was prescriptive regarding the method of determining the 
MEC. For SO2, the MEC was to be calculated using Equation A-
2; for NO_X, an MEC value of 320.0 ppm was to be used for 
coal-firing and 160.0 ppm for oil-or gas-firing.
    In the first two years of Acid Rain Program implementation, it 
became increasingly clear to both the regulated community and to EPA 
that the span and range provisions of part 75 lacked sufficient 
flexibility and clarity. The NO_X provisions were 
particularly problematic, being overly prescriptive in some instances 
and sometimes requiring two spans and ranges when a single, 
appropriately-sized range would suffice. Also, the units of the flow 
rate span were expressed in terms of velocity (i.e., feet per minute), 
and this was not consistent with either the units of measure used for 
daily monitor calibrations or the units used for electronic reporting 
of flow rate data.
    The May 17, 1995 rule attempted to address these deficiencies, as 
follows. For SO2, an alternative means of determining the 
MPC, in lieu of using historical fuel sampling data, was added; the MPC 
could be based upon 30 days of historical CEMS data. The use of 
historical CEMS data was also allowed as an option for MEC 
determinations, instead of using Equation A-2. For NO_X, the 
method of determining the MPC was made less prescriptive. First, a 
comprehensive list of MPC values was promulgated (Tables 2-1 and 2-2 in 
Appendix A), taking into consideration the unit type in addition to the 
fuel type. The MPC value from this list could be used in lieu of the 
fuel-based MPC prescribed in the original rule. Second, two alternative 
methods of determining the MPC or MEC were added, i.e., from historical 
CEMS data or from emission test results. Finally, flexibility was added 
to the dual-range requirements for NO_X monitors so that, in 
many instances, the span and range requirements of part 75 could be met 
on a site-specific basis, using a single span and range.
    The span provisions for CO2 and O2 were not 
significantly changed in the May 17, 1995 rule. For flow rate, however, 
a more detailed procedure for determining the span value was added. 
This addition was considered necessary because during the first year of 
program implementation it came to light that there are actually two 
important span values associated with flow rate: (a) the 
``calibration'' span value used for daily calibrations, and (b) the 
``flow rate'' span value in units of standard cubic feet per hour 
(scfh). These two span values are both derived from the MPV, but are 
almost invariably expressed in different units of measure, and, 
therefore, the two spans are generally not equal numerically. For 
instance, the calibration span value for the daily calibration of a 
differential pressure-type flow monitor, expressed in units of inches 
of water, is a small number (generally less than 5.0 in. 
H2O); while the flow rate span value, in scfh, is a very 
large number, usually in the tens or hundreds of millions.
    The May 17, 1995 rule also revised the procedures for adjusting the 
span and range of SO2, NO_X, and flow monitors. 
Sections 2.1.1.4, 2.1.2.4, and 2.1.4 of Appendix A to the original rule 
had specified that span and range adjustments were required whenever 
the MPC, the MEC, or the MPV changed significantly. When a significant 
change in the MPC, MEC, or MPV occurred, a new range setting was to be 
established and a new span value defined, equal to 80.0 percent of the 
adjusted range value. The revised sections 2.1.1.4, 2.1.2.4, and 2.1.4 
of Appendix A to the May 17, 1995 rule changed this procedure, 
requiring the new span value to be determined first, followed by the 
new range. The May 17, 1995 rule also added procedures for addressing 
full-scale exceedances, specifying that the full-scale value is to be 
reported for an exceedance of one hour and that a range adjustment is 
required for an exceedance greater than one hour. Finally, the May 17, 
1995 rule specified that whenever the range of a gas monitor is 
adjusted, a linearity test is required, and a calibration error test 
must be done when the range of a flow monitor is adjusted.
Discussion of Proposed Changes
    Since promulgation of the May 17, 1995 rule, EPA has continued to 
receive questions and comments about the span and range sections of 
part 75. Many of

[[Page 28055]]

the questions and comments have centered on the adjustment of span and 
range. The following questions are typical: When must the span and 
range be changed? What constitutes a ``significant'' change in the MPC, 
MEC, or MPV? When a span and range adjustment is required, what are the 
deadlines for making the changes and for completing the required 
linearity test? How should full-scale exceedances be reported? There 
also appears to be some lingering confusion and misunderstanding about 
how to determine the flow rate span values and how to calculate the 
maximum potential flow rate (MPF) and the NO_X maximum 
emission rate (MER) (see Docket A-97-35, Items II-B-8, II-D-67, and II-
E-31). In view of this, EPA believes that the span and range sections 
of the rule are still not sufficiently clear, flexible, or detailed and 
are in need of further revision. In June, 1996, a national part 75 CEM 
Implementation Workgroup meeting was held in Washington D.C. to discuss 
possible revisions to part 75. One of the principal topics of 
discussion was span and range (see Docket A-97-35, Item II-E-32). 
Today's rulemaking proposes comprehensive revisions to sections 2.1 
through 2.1.4 of Appendix A, based in part on the discussions of the 
June, 1996 meeting. The principal changes are described in paragraphs 
(1) through (5), below.
1. Maximum Potential Values
    The basic procedure for determining the maximum potential of 
SO2 concentration would be unchanged by today's proposal. 
However, two new provisions would be added to section 2.1.1.1 of 
Appendix A to prevent overestimation of the MPC. The first of these 
provisions would allow the exclusion of clearly anomalous fuel sampling 
results when determining the MPC. The second provision would apply to 
units for which the designated representative certifies that the 
highest sulfur fuel is never combusted alone, but is always blended or 
co-fired with other fuel(s) during normal operation. For such units, 
the MPC would be calculated using best estimates of the highest sulfur 
content and lowest gross calorific value expected for the blend or fuel 
mixture and inserting these values into Equation A-1a or A-1b. The best 
estimates of the highest percent sulfur and lowest GCV for a blend or 
fuel mixture would be derived from weighted-average values based upon 
the historical composition of the blend or mixture in the previous 12 
(or more) months.
    The alternative procedure for determining the MPC of SO2 
based upon quality assured historical CEMS data would be retained, but 
it is proposed that the MPC be based, at a minimum, upon the previous 
720 quality assured monitor operating hours, rather than the previous 
30 unit operating days. This is to ensure that a sufficient quantity of 
valid data is used for the MPC determination. Making the determination 
based on 30 unit operating days does not provide that assurance, 
particularly for units that may only operate for a few hours a day 
(e.g., peaking units). Revised section 2.1.1.1 would also specify that 
for a unit with add-on SO2 emission controls, the historical 
CEMS data option may only be selected if the certified SO2 
monitor used to determine the MPC is located at the control device 
inlet.
    For NO_X, the general procedures for determining the MPC 
would also remain the same, i.e., either: (1) use the MPC value 
prescribed in the original rule, (2) use the unit-specific value listed 
in Table 2-1 or 2-2, or (3) determine the MPC by emission testing or 
from historical CEM data. However, the following changes to section 
2.1.2.1 of Appendix A are proposed. First, a statement would be added 
that the MPC would have to be based upon the combustion of whichever 
fuel or blend combusted at the unit produces the highest level of 
NO_X emissions. Second, an advisory statement would be added, 
noting that the initial MPC value determined for a unit that is not 
equipped with low-NO_X burners (LNB) would have to be re-
evaluated if a low-NO_X burner system is subsequently 
installed and optimized. Third, if historical CEMS data are used to 
determine the MPC, the determination would have to be based on the 
previous 720 (or more) quality assured monitor operating hours (instead 
of the previous 30 unit operating days). Fourth, units with add-on 
NO_X emission controls could only use the historical CEM data 
option if the historical data represented uncontrolled emissions (e.g., 
if the certified CEMS used to collect the data were located prior to 
the control device inlet or, for a unit with seasonal NO_X 
controls, if the historical data were from a period when the controls 
were not operating). Fifth, if emission testing is used for the MPC 
determination, sufficient tests would have to be performed at various 
loads and excess oxygen levels to ensure that a credible MPC value is 
obtained. For units with add-on NO_X emission controls, the 
emission test data would have to be collected upstream of all controls, 
or, for a unit with seasonal controls, during a period when the 
controls were not operating. Finally, a specific requirement to 
calculate the maximum potential NO_X emission rate (MER) 
would be added to section 2.1.2.1 of Appendix A. The May 17, 1995 rule 
had provided a definition of the MER in Sec. 72.2; however, a 
corresponding requirement to calculate the MER was not included in part 
75 at that time. The MER is occasionally needed to provide substitute 
NO_X emission rates during missing data periods. The owner or 
operator would be permitted to use the diluent cap value of 5.0 percent 
CO2 or 14.0 percent O2 for boilers (or 1.0 
percent CO2 or 19.0 percent O2 for turbines) in 
the NO_X MER calculation.
    For CO2, today's proposed rule would add a new section 
2.1.3.1 to Appendix A, which provides a definition of the MPC. The MPC 
for CO2 pollutant concentration monitors would be 14.0 
percent for boilers and 6.0 percent CO2 for combustion 
turbines. Alternatively, the MPC could be based on a minimum of 720 
hours of representative quality assured historical CEM data.
    For flow rate, the procedure for determining the MPV would be 
essentially unchanged by today's proposed rule, i.e., the MPV would 
either be determined from Equation A-3a (or A-3b, as applicable) in 
Appendix A, or it would be based on velocity traverse data taken at or 
near maximum load. However, a procedure for calculating the maximum 
potential flow rate (MPF) would be added to section 2.1.4.1 of Appendix 
A. The MPF is occasionally used to provide substitute flow rate data; 
therefore, a clear, consistent method of determining the MPF is needed.
2. Maximum Expected SO2 and NO_X Concentrations
    Today's proposal would significantly change the procedures for 
determining the maximum expected concentration (MEC) of SO2. 
The purpose of the revisions would be to ensure that the proper span(s) 
and range(s) are selected for SO2 measurement. Proposed 
section 2.1.1.2 of Appendix A would require the MEC to be determined 
for units with SO2 controls and also for uncontrolled units 
that burn both high- and low-sulfur fuels (or blends) as primary or 
backup fuels (e.g., high- and low-sulfur coal or different grades of 
fuel oil).
    The revised procedures for determining the MEC for SO2 
would be as follows. For units with emission controls, Equation A-2 in 
Appendix A would be used to calculate the MEC. For uncontrolled units 
that burn both high-sulfur and low-sulfur fuels or blends as primary or 
backup fuels, Equation A-1a or A-1b in Appendix A (which in the

[[Page 28056]]

current rule is reserved for MPC calculations) would be used to 
determine an MEC value for each fuel or blend, with three important 
exceptions. The MEC would not be calculated for: (1) the highest-sulfur 
fuel or blend (because it would be duplicative of the MPC calculation); 
(2) fuels or blends with a total sulfur content no greater than the 
total sulfur content of natural gas, i.e.,  0.05 percent 
sulfur by weight, because Sec. 75.11(e)(3)(iv) of the current rule 
specifies that natural gas combustion does not trigger a dual span and 
range requirement for the SO2 monitor (for gas firing, the 
MEC and low-scale span values would be too low to be practical for 
quality assurance purposes, e.g., < 5 ppm for pipeline natural gas); 
and (3) fuels or blends that are combusted only during unit startup, 
because such fuels are infrequently used and are not representative of 
normal unit operation.
    Today's proposal would continue to allow the same flexibility in 
the SO2 MEC determination that was introduced in the May 17, 
1995 rule. That is, if a certified SO2 CEMS is already 
installed, the owner or operator could determine the MEC based upon 
historical continuous monitoring data, in lieu of using mathematical 
equations. If this option were chosen for a unit with SO2 
controls, the MEC would be the maximum SO2 concentration 
measured at the control device outlet by the CEMS over the previous 720 
or more quality assured monitor operating hours with the unit and the 
control device both operating normally. For units that burn both high- 
and low-sulfur fuels or blends as primary and backup fuels and have no 
SO2 controls, the MEC for each fuel would be the maximum 
SO2 concentration measured by the CEMS over the previous 720 
or more quality assured monitor operating hours in which that fuel or 
blend was the only fuel being burned in the unit.
    Today's rule also proposes to change the way in which the MEC is 
determined for NO_X. Revised section 2.1.2.2 of Appendix A 
would require a determination of the MEC during normal operation for 
units with add-on NO_X controls capable of reducing 
NO_X emissions to 20.0 percent or less of the uncontrolled 
level (i.e., steam injection, water injection, selective catalytic 
reduction or selective non-catalytic reduction). A separate MEC 
determination would be required for each type of fuel combusted, except 
for fuels that are only used for unit startup or for flame 
stabilization. The MEC would be determined in one of three ways: (1) 
using Equation A-2 in Appendix A; or, if that equation is not 
appropriate, (2) by emission testing or (3) by using historical CEMS 
data from the previous 720 (or more) quality assured monitor operating 
hours. Revised section 2.1.2.2 would give specific guidelines and 
procedures by which to obtain the MEC when the emission testing or CEMS 
data options are selected. All CEMS or emission test data used for the 
MEC determination would be taken under stable operating conditions with 
all control devices and methods operating properly.
3. Span and Range Values
    For SO2, NO_X, and flow rate, respectively, 
revised sections 2.1.1.3, 2.1.2.3 and 2.1.4.2 of Appendix A would allow 
the high-scale span value to be between 100.0 and 125.0 percent of the 
maximum potential value (i.e., the MPC or MPV), rounded off 
appropriately. This is a change from the current rule which requires 
the high span to be set at 125.0 percent of MPC or MPV, rounded off 
appropriately. However, the change is not expected to be disruptive, 
because properly sized span values previously determined by multiplying 
the MPC or MPV by 1.25 could continue to be used. The change would 
allow the owner or operator to set the span value in such a way that a 
small exceedance of MPC or MPV would not require a span change (see 
paragraph 5, ``Adjustment of Span and Range,'' below). The added 
flexibility in span selection would also allow different units with 
similar (but not identical) MPCs for SO2 and/or 
NO_X to use the same span value and to use the same 
calibration gas concentrations, which could result in cost savings for 
some facilities. In 1996, EPA received and approved a petition from one 
utility to equalize the SO2 span values at several of its 
coal-fired units (see Docket A-97-35, Items II-C-23, II-D-71).
    For CO2 and O2 monitors, today's proposal 
would revise section 2.1.3 of Appendix A to allow the owner or operator 
maximum flexibility in selecting an appropriate span value. The 
CO2 or O2 span value would not be determined in 
the same way as an SO2, NO_X, or flow rate span 
value. Rather, for CO2 monitors installed on boilers, any 
convenient span value between 14.0 percent and 20.0 percent 
CO2 representing the percent diluent in the flue gas would 
be acceptable. For combustion turbines, any CO2 span value 
between 6.0 and 14.0 percent CO2 could be used. For 
O2 monitors, a span value between 15.0 percent and 25.0 
percent O2 could be selected. However, if the O2 
concentrations are expected to be consistently below 15.0 percent, an 
alternative span value of less than 15.0 percent could be used, 
provided that an acceptable technical justification was included in the 
monitoring plan. The proposed rule would also allow purified instrument 
air containing 20.9 percent O2 to be used as the high level 
calibration gas for oxygen monitors having span values greater than or 
equal to 21.0 percent O2.
    There are two principal reasons why EPA is proposing increased 
flexibility in the selection of the CO2 and O2 
span values. The first is to encourage greater accuracy in the diluent 
gas measurements. The revisions would allow the span value to be 
customized so that the concentration of the upscale calibration gas 
used for daily calibrations can be as close as possible to the actual 
average CO2 or O2 concentrations in the stack. In 
1996, EPA received and approved a petition from one utility to use a 
CO2 span value of 15.0 percent for its coal-fired units, 
rather than the 20.0 percent span value required by part 75 (see Docket 
A-97-35, Items II-C-20, II-D-68). The second reason for revising the 
CO2 and O2 span requirements is to eliminate 
unnecessary high-level span and range requirements. The current rule 
requires a high span value of 20.0 percent for all CO2 and 
O2 monitors. However, there are many units (e.g., combustion 
turbines) for which the diluent gas concentrations are so low that the 
guideline in the current section 2.1 of Appendix A (i.e., that the 
majority of the readings be within 25.0 to 75.0 percent of full-scale) 
cannot be met unless a second, low-scale span and range are used. For 
most of these units, there are technical and safety reasons why the 
diluent concentrations must remain low; therefore, it is unreasonable 
to require a high range to be maintained if a lower range will suffice 
and can never be exceeded. During the Phase II certification process, 
EPA approved CO2 span values of 10.0 percent for a number of 
combustion turbines and waived the high-scale range requirement (see 
Docket A-97-35, Items II-C-19, II-C-21, II-D-64).
    Today's proposal would not change the basic way in which the full-
scale range setting of a monitor is determined. The range would still 
have to be set greater than or equal to the span value. However, the 
guideline for selecting an appropriate full-scale range in section 2.1 
of Appendix A would be revised as follows. With few exceptions, the 
full-scale range would be selected so that, to the extent practicable, 
the readings during typical unit operation fall between 20.0 and 80.0 
percent of full-scale; this represents a slight increase in flexibility 
from the ``25-to-75 percent of

[[Page 28057]]

full-scale'' guideline in the current rule. Today's proposal would also 
emphasize that section 2.1 is only a guideline and would cite three 
specific cases in which it is inapplicable. Specifically, the guideline 
would not apply to: (1) quality assured SO2 readings 
obtained during the combustion of natural gas or fuel with equivalent 
total sulfur content (because the resulting SO2 emissions 
are too low to be subject to the span and range requirements); (2) 
quality assured SO2 or NO_X readings on the high 
range for an affected unit with SO2 or NO_X 
emission controls and two span values (because the high range is not 
the normal operating range for the unit); and (3) quality assured 
SO2 or NO_X readings less than 20.0 percent of the 
low measurement range for a dual-span unit with SO2 or 
NO_X emission controls, provided that the low readings are 
associated with periods of high control device efficiency (because it 
is not necessary to re-range a monitor based on non-representative 
hours of exceptional control performance).
    For flow monitors, today's rule proposes to revise section 2.1.4.2 
of Appendix A to more clearly define the ``calibration span value'' 
(which is the span expressed in the units of measure used for the daily 
calibrations) and the ``flow rate span value'' (which is the span 
expressed in the units used for electronic data reporting, i.e., scfh). 
The proposed rule defines these two span values in considerable detail 
and outlines how to use them. EPA believes that this will result in 
greater consistency in implementation of the part 75 flow rate 
monitoring requirements.
4. Dual Span and Range Requirements for SO2 and 
NO_X
    In today's rule, revisions are proposed to the dual span and range 
requirements for SO2 and NO_X monitors in sections 
2.1.1.4 and 2.1.2.4 of Appendix A. The revised provisions are 
essentially the same for both pollutants. To determine whether a 
second, low-scale span is required in addition to the high-scale span 
based on the MPC, each of the maximum expected concentration (MEC) 
values determined under revised section 2.1.1.2 or 2.1.2.2 of Appendix 
A would be compared against the maximum potential concentration (MPC) 
determined under proposed sections 2.1.1.1 or 2.1.2.1. If this 
comparison shows any of the MEC values to be < 20.0 percent of the MPC, 
a low-scale span would be required. If several of the MEC values are 
found to be < 20.0 percent of the MPC, then the low-scale span would be 
based upon whichever MEC value is closest to 20.0 percent of the MPC. 
The low-scale span value would be determined in a manner similar to the 
high-scale span, i.e., by multiplying the MEC by a factor between 1.00 
and 1.25 and rounding off the result appropriately.
    When both a high-scale span and a low-scale span are required for 
SO2 or NO_X, proposed sections 2.1.1.4 and 2.1.2.4 
would allow the owner or operator to use either of the following 
monitor configurations to meet the dual-range requirement: (1) a single 
analyzer with two ranges, or (2) two separate analyzers connected to a 
common probe and sample interface. The use of other monitoring 
configurations would be subject to the approval of the Administrator. 
The monitor configurations would be represented in the monitoring plan 
as follows: (a) the high and low ranges could be designated as two 
separate, primary monitoring systems; (b) the high and low ranges could 
be designated as separate components of a single, primary monitoring 
system; or (c) one range (the ``normal'' range) could be designated as 
a primary monitoring system, and the other range as a non-redundant 
backup monitoring system. The high and low ranges would be quality 
assured according to their designation in the monitoring plan. Primary 
monitoring systems would have to meet the QA requirements for primary 
systems in Sec. 75.20(c), Appendix A, and Appendix B, with the 
following exception: relative accuracy test audits (RATAs) would be 
required only on the normal range. For units with emission controls, 
the low range would be considered normal; for other units, the range in 
use at the time of the scheduled RATA would be considered normal. Non-
redundant backup systems would have to meet the applicable QA 
requirements for ``like-kind replacement analyzers'' in proposed 
Sec. 75.20(d).
    Today's rule would add a new alternative provision under sections 
2.1.1.4 and 2.1.2.4 of Appendix A for dual-span units with 
SO2 or NO_X emission controls. The new provision 
would allow the owner or operator to use a ``default high-range value'' 
in lieu of operating, maintaining, and quality assuring a high-scale 
monitor range. The default high-range value would be 200.0 percent of 
the MPC (based on uncontrolled emissions). This value would be reported 
whenever the SO2 or NO_X concentration exceeded 
the full-scale of the low-range analyzer. The default high-range value 
is being proposed for controlled units that seldom, if ever, experience 
full-scale exceedances of the low monitor range during normal operation 
(e.g., units that have a permit condition requiring cessation of unit 
operation when a full-scale exceedance occurs or units that experience 
low-range exceedances only during startup). EPA solicits comment on the 
proposed approach of using a default high-range value in lieu of a high 
range monitor and on the value of the default.
    EPA specifically requests comment on whether the proposed dual-span 
monitoring configurations, monitoring system designations, and quality 
assurance requirements are adequate, or whether there are additional 
configurations (e.g., one range with two spans, two separate analyzers 
with separate probes, etc.) that should be included in the rule.
    Finally, when two spans and ranges are required, proposed revised 
sections 2.1.1.4 and 2.1.2.4 of Appendix A would specify that the low 
range would have to be used to record emission data when the 
SO2 or NO_X concentrations are expected to be 
consistently below 20.0 percent of the MPC (i.e., when a fuel or blend 
with a MEC value < 20.0 percent of the MPC is combusted). And if the 
full-scale of the low range is exceeded, the high range would be used 
to record data (or, if applicable, the default high range value would 
be reported).
5. Adjustment of Span and Range
    In today's rule, detailed guidelines and procedures are proposed 
for adjusting the span and range of the CEMS in revised sections 
2.1.1.5, 2.1.2.5, 2.1.3.2 and 2.1.4.3 of Appendix A. The intent of 
these provisions is to ensure that each owner or operator assesses the 
adequacy of all CEMS span values on at least a quarterly basis (and 
whenever operational changes are planned) and, based on that 
assessment, makes any necessary adjustments to the spans or ranges in a 
timely manner. EPA believes that the proposed procedures are 
sufficiently flexible so that frequent span and range adjustments will 
not be necessary. The procedures are primarily directed at CEMS with 
improperly-sized spans and ranges, to bring them into full conformance 
with part 75 requirements or for future changes in unit operation 
(e.g., fuel switch or low-NO_X burner installation) that may 
significantly affect the level of emissions or flow. All required span 
or range adjustments would have to be made no later than 45 days after 
the end of the quarter in which the need to adjust the span or range is 
identified, unless the span change would require new calibration gases 
to be ordered for daily calibration error and linearity tests, in which 
case, the owner or operator would have up to

[[Page 28058]]

90 days after the end of the quarter to make the span adjustment.
    The revised procedures for span and range adjustment would be as 
follows. First, if the maximum value upon which the high span value is 
based (i.e., the MPC or, for flow rate, the MPF) is exceeded during a 
calendar quarter, but the span is not exceeded, the span or range would 
not have to be adjusted. However, for missing data purposes, if any 
quality assured hourly concentration or flow rate exceeds the MPC or 
MPF by  5.0 percent during the quarter, a new MPC or MPF 
would have to be defined, equal to the highest value recorded during 
the quarter, and a monitoring plan update would be required. Second, 
for the high measurement range, if any quality assured reading exceeded 
the span value by  10.0 percent during the quarter but did 
not exceed the range, a new MPC or MPF (as applicable) would have to be 
defined, equal to the highest on-scale reading recorded during the 
quarter, and the span value would also have to be changed. If the new 
span value exceeded the current full-scale range setting, then a new 
range setting would also be required. Similar span adjustment 
requirements would apply to the low scale if the two measurement ranges 
are used separately for distinctly different modes of operation (e.g., 
during the combustion of different fuels), rather than being used in 
combination to provide a continuum of measurement range capability.
    The proposed procedures for responding to full-scale exceedances 
are as follows. Whenever the full-scale of a high monitor range is 
exceeded, excluding hours of non-representative operating conditions 
(e.g., a trial burn of a new fuel), corrective action would be required 
to adjust the span and range. In addition, any time the range is 
exceeded, a value of 200.0 percent of the current full-scale range 
would be reported to EPA for each hour of each full-scale exceedance. 
The Agency believes that 200.0 percent of the range is sufficiently 
conservative to ensure that emissions would not be under-reported. One 
utility that experienced a full-scale exceedance of the high 
SO2 monitor range estimated from the results of fuel 
sampling that the SO2 concentration was approximately 150.0 
percent of full-scale during the incident (see Docket A-97-35, Item II-
D-24).
    For units with two span values and two measurement ranges for a 
particular parameter (e.g., SO2), when the full-scale of the 
low range is exceeded, provided that the high monitor range is 
available to record emission data, no corrective actions would be 
required. However, if, at the time of the low-range exceedance or 
during the continuation of the low-range exceedance, the high range is 
either out-of-service or out-of-control for any reason (and therefore 
is not available to record quality assured data), the MPC would have to 
be reported until the readings either returned to the low scale or 
until the high scale returned to service and was able to provide 
quality assured data. However, if the reason the high scale is 
unavailable is because of a high scale exceedance, 200.0 percent of the 
high range value would be reported for each hour of the exceedance.
    Proposed sections 2.1.1.5(e), 2.1.2.5(e), and 2.1.4.3(e) of 
Appendix A would require that the monitoring plan be updated whenever 
changes are made in the maximum potential values, maximum expected 
values, span values, or full-scale range settings. The updates would be 
made in the quarter in which the changes become effective. The proposed 
sections 2.1.1.5(e) and 2.1.2.5(e) of Appendix A would further require 
a linearity test to be done whenever the span of a gas monitor is 
adjusted, if the span change is significant enough to require new 
calibration gases for daily calibration error tests and linearity 
checks. Finally, proposed sections 2.1.4.3(c) and (d) of Appendix A 
would require a calibration error test to be done whenever a flow 
monitor span or range is adjusted (unless the adjustment requires a 
significant change to the flow monitor that would require 
recertification under Sec. 75.20(b)).

J. Quality Assurance/Quality Control (QA/QC) Program

1. QA/QC Plan
Background
    Section 1 of Appendix B to part 75 as originally promulgated on 
January 11, 1993 sets forth provisions for developing and implementing 
a quality control program. As part of the quality control program, 
section 1 requires that the source develop and maintain a quality 
control plan that documents how the equipment used to report emissions 
data for part 75 is maintained and quality assured. While the 
provisions in sections 1.1, 1.2, and 1.4 of Appendix B to part 75 are 
applicable only to continuous emissions monitoring systems, the 
provisions in sections 1.3 and 1.5 of the existing rule are more 
generally applicable to all monitoring systems under part 75. The 
quality assurance requirements for excepted monitoring systems under 
Appendices D and E and for alternative monitoring systems under subpart 
E are provided in the respective Appendices or subpart of part 75, as 
revised; however, specific guidelines for the quality control plans for 
these systems are not given.
    Based on the experience of state and EPA inspectors at Acid Rain 
field audits, there has been confusion and inconsistency among industry 
sources regarding the contents of the quality control plan. In some 
cases, utility staff have requested further guidance from EPA on what 
the quality control plan should contain. Based on this experience, the 
Agency believes that the quality control program provisions in section 
1 of Appendix B need to be revised. Specifically, the rule needs to be 
clarified in two areas: (1) the applicability of the QA/QC program 
(i.e., do the provisions apply to all monitoring systems, only to CEMS, 
or only to specific excepted or alternative monitoring systems?); and 
(2) the recordkeeping requirements for repair and maintenance events. 
In addition, several utilities have asked EPA to consider deleting the 
requirement to maintain an inventory of spare parts, which they believe 
to be unnecessary and burdensome.
Discussion of Proposed Changes
    The proposed revisions discussed in this section affect section 1 
of Appendix B to part 75. The terms ``quality control program and 
plan'' would be changed to ``quality assurance/quality control program 
and plan.'' The scope of section 1 would be expanded to include QA/QC 
program provisions for excepted monitoring systems under Appendices D, 
E, and I and alternative monitoring systems under subpart E. Section 1 
would also be reordered to separate the requirements applicable to all 
monitoring systems (section 1.1) from the requirements specific to CEMS 
(section 1.2). The preventative maintenance provisions, in section 1.3 
of the existing rule, would be moved to section 1.1.1 of the proposal, 
and would be revised to delete the requirement to maintain an inventory 
of spare parts. A new section 1.1.3 would be added to specify the 
requirements for maintaining records of testing, maintenance, and 
repair activities. QA/QC program requirements specific to excepted 
monitoring systems under Appendices D, E, and I would be added in 
section 1.3. These provisions would require written procedures to be 
maintained for fuel flowmeter testing, primary element inspection, and 
fuel sampling and analysis as well as requiring a description of 
equipment and records of testing to be maintained. Section 1.3.6 would 
make the

[[Page 28059]]

recordkeeping requirements consistent with the quality assurance 
requirements of section 2.3.1 of Appendix E. Section 1.3.7 would 
specify which QA/QC program requirements apply for excepted monitoring 
systems under Appendix I. Finally, section 1.4 would define the QA/QC 
program requirements for alternative monitoring systems approved under 
subpart E, based on the quality assurance requirements of subpart E.
Rationale
    The Agency believes that the manner in which quality assurance/
quality control (QA/QC) and maintenance-related activities are 
performed can have a significant effect upon the accuracy of the data 
reported by a monitoring system. Therefore, today's proposal seeks to 
ensure that adequate records are kept to document that each monitoring 
system and its ancillary components is being maintained and operated in 
a proper manner. Section 1 in Appendix B to part 75 would, therefore, 
be amended to provide sources with General guidance regarding QA/QC 
program requirements. However, the Agency recognizes that QA/QC 
programs may vary from site to site and that many sources have already 
developed and implemented an effective QA/QC program. It is the 
Agency's intent to allow each source the flexibility to develop and 
implement a QA/QC program that will result in the reporting of accurate 
emissions data through proper equipment calibration, maintenance and 
troubleshooting procedures.
    (a) Inventory of Spare Parts. Section 1.3 of Appendix B to part 75 
in the January 11, 1993 rule requires that an inventory of spare parts 
be maintained as part of the QA/QC program. The intent of this 
requirement is one of the fundamental goals of a QA/QC program, i.e., 
to maximize the availability of quality-assured data from the 
monitoring system. Since maintenance and repairs are required in order 
to keep the monitoring system operating properly, the need for 
replacement parts will arise over the term of use of the monitoring 
equipment. In order to minimize the amount of time when the system is 
unable to provide data because a new part is needed, the existing rule 
requires that the source maintain an inventory of spare parts. The 
Agency has received comments on this requirement from both affected 
utilities and from state inspectors arguing that it is unnecessary and 
cumbersome (see Docket A-97-35, Item II-D-49, II-E-28). Commenters have 
suggested that different approaches have been effectively employed to 
ensure that spare parts are available in a timely manner; however, not 
all of these approaches require that an inventory of spare parts be 
kept on-site. For example, some spare parts may be available on a very 
timely basis from a local supplier, making it unnecessary to maintain 
spare parts on-site. The Agency believes that these different 
approaches may be adequate substitutes for keeping an on-site inventory 
of spare parts. Therefore, the requirement to maintain an inventory of 
spare parts would be removed in today's proposal, although the 
objective of an effective QA/QC program, i.e., to maximize data 
availability, would not change.
    (b) Maintenance Records. The Agency believes that maintaining 
records of monitoring system maintenance and repairs is an essential 
component of an effective QA/QC program. Several utilities have 
indicated that they agree and have instituted QA/QC programs which 
include maintaining such records (see, e.g., Docket A-97-35, Item II-D-
88). However, some EPA and state inspectors have found that not all 
sources keep adequate records of maintenance and repairs in their QA/QC 
program. EPA believes that this failure to keep adequate records 
compromises the effectiveness of the QA/QC program. Therefore, today's 
proposal would require each source to maintain proper records of all 
testing, maintenance, or repair activities performed on any monitoring 
system or component. Additionally, today's proposal would require that 
these records and any additional supporting documentation be made 
available for review during an audit.
    (c) Excepted Monitoring System Requirements. The required quality 
assurance activities for excepted monitoring systems are set forth in 
the respective Appendices D, E, or I. Today's proposed revisions in 
section 1.3 of Appendix B would specify that information on the 
approved methods, test procedures and test results must be maintained 
on-site suitable for inspection as part of the QA/QC program. The 
proposed revisions would consolidate all of the QA/QC requirements in 
Appendix B rather than having them spread out in Appendices D, E, and 
I.
2. Flow Monitor Polynomial Coefficient
Background
    Many of the stack gas volumetric flow rate monitors currently in 
use by affected sources use software polynomial coefficients to convert 
electrical signals from the monitors into flow rate values that are 
electronically reported to the Acid Rain Division. The flow rate values 
generated from these monitors are used by the source's data acquisition 
and handling system (DAHS) to compute hourly mass emission rates of 
SO2, CO2, and hourly heat input rates. Currently, 
affected sources are not specifically required to report, record, or 
document the numerical values of the polynomial coefficients used by 
their flow monitors.
Discussion of Proposed Changes
    Proposed Sec. 75.59(a)(5)(vi) and proposed revisions to section 
1.1.3 of Appendix B would require the current values of the flow 
monitor coefficients to be recorded and would require records to be 
kept of any changes or adjustments to the coefficient values. The 
proposed revisions in Sec. 75.20(b) define flow monitor coefficient 
adjustment as an event which requires recertification.
Rationale
    (a) Recordkeeping of Coefficients. The agency has recently become 
aware (by a comment received in response to a request for review of the 
Acid Rain Audit Manual) of a potentially serious omission in the flow 
monitor recordkeeping requirements of part 75 (see Docket A-97-35, Item 
II-D-92). The commenter indicated that part 75 lacks a requirement to 
document the values of the polynomial coefficients which are programmed 
into the software of most flow monitoring devices, and that the Acid 
Rain CEM audit manual does not recommend that Agency or state auditors 
check the coefficient values. The values of the polynomial coefficients 
are important because they are directly related to the accuracy of a 
flow monitor. The coefficient values are usually established at three 
different load levels (low, mid, and high), in a process called 
``linearization'' or ``characterization'' of the monitor. Linearization 
is done in an attempt to ensure that the flow monitor reads accurately 
across all load levels. The Agency agrees with the commenter that the 
flow monitor variables are a critical component of the flow monitoring 
system and that the adjustment of those variables represents a 
significant change to the flow monitoring system. Therefore, today's 
rulemaking proposes to add Sec. 75.59(a)(5)(vi) to require owners and 
operators of affected sources to record the numerical values of the 
flow monitor polynomial coefficients used during initial certification 
of the monitor and during each subsequent relative accuracy test audit 
(RATA). In

[[Page 28060]]

addition, section 1 of Appendix B to part 75 would be revised to 
require that any changes to the flow monitor polynomial coefficients be 
documented and maintained as part of the QA/QC program maintenance 
records. Section 1 of Appendix B would also be changed to require the 
source to document procedures related to the adjustment of flow monitor 
variables in its QA/QC plan. The values of the flow monitor 
coefficients and the related adjustment procedures would be required to 
be kept on-site, in a format suitable for review by an inspector during 
an audit.
    (b) Recertification After Adjustment of Coefficients. Since 
changing the flow monitor polynomial constants relinearizes the 
instrument, significantly altering the monitored reading, today's 
proposed rule would amend Sec. 75.20(b) to require recertification 
subsequent to any flow monitor polynomial coefficient change. Since a 
three level RATA is the only part 75 quality assurance test that checks 
the linearity of a flow monitor, the recertification would require a 
three level RATA.

K. Calibration Gas Concentration for Daily Calibration Error Tests

Background
    All part 75 gas monitoring systems are required by section 2.1.1 of 
Appendix B of the current rule to pass daily calibration error tests, 
in order to validate emission data from the CEMS. The procedures for 
conducting the daily calibration error tests are found in section 6.3.1 
of Appendix A. Each daily calibration error test consists of injecting 
two protocol gases of known concentration into the CEMS and comparing 
the responses of the instrument to the tag values of the protocol 
gases. The two required gas concentrations for the calibration error 
tests are zero-level (i.e., 0.0 to 20.0 percent of the span value of 
the instrument) and high-level (80.0 to 100.0 percent of span).
    The span values of part 75 SO2 and NO_X 
monitors are determined by multiplying the maximum potential 
concentration (MPC) by 1.25 and rounding the result upward to the 
nearest 100.0 ppm. For CO2 and O2 monitors, a 
span value of 20.0 percent O2 or CO2 is 
prescribed. These span values have been deliberately oversized to 
prevent full-scale exceedances from occurring. Consequently, the 
SO2, NO_X, CO2, and O2 
readings obtained during normal unit operation are generally well below 
the span values and typically range from about 25.0 to 75.0 percent of 
full-scale. Because of the oversized span values, the concentrations of 
the high-level calibration gases used for daily calibration error tests 
are often much higher than the actual pollutant and diluent gas 
concentrations in the stack. As a result, the representativeness of the 
daily calibration error test can be questioned, because the test does 
not always check the accuracy of an analyzer on the part of the scale 
where most of the readings occur. For instance, typical CO2 
concentrations for many part 75 units range from about 10.0 to 12.0 
percent CO2 (i.e., 50.0 to 60.0 percent of the span value). 
However, when CO2 analyzers are calibrated, the high-level 
calibration gas concentrations (i.e., 16.0 to 20.0 percent 
CO2 ) are considerably higher than normal stack emissions. 
In view of this, EPA believes it would be appropriate to allow the 
owner or operator to have greater flexibility in selecting a 
representative upscale gas for daily calibrations. One State agency has 
successfully implemented this type of flexibility in its CEM program. 
The State's CEM rule specifies the acceptable range of values for the 
upscale calibration gas, but adds the following qualifying statement, 
``* * *unless an alternative concentration can be demonstrated to 
better represent the normal source operating levels *-*-*'' (see Docket 
A-97-35, Item II-D-72).
Discussion of Proposed Changes
    Today's rule proposes to add flexibility to the procedures for 
conducting the calibration error tests of part 75 gas monitors to 
encourage daily calibrations to be done more representatively. Section 
6.3.1 of Appendix A would be revised so that, beginning on January 1, 
2000, either the mid-level gas (50.0 to 60.0 percent of span) or the 
high-level gas (80.0 to 100.0 percent of span) could be used as the 
upscale calibration gas for daily calibration error tests. A 
corresponding change would be made to the procedure for calculating the 
calibration error in section 7.2.1 of Appendix A. Prior to January 1, 
2000, the owner or operator would have the option of using the mid-
level calibration gas for daily calibrations if it better represents 
the typical stack gas concentrations than the high-level gas.

L. Linearity Test Requirements

Background
    Section 75.20(c) of the current part 75 rule requires a 3-point 
linearity test of each SO2 and NO_X pollutant 
concentration monitor and each diluent gas (O2 or 
CO2) monitor, as part of the initial certification process. 
A linearity test consists of a series of nine reference calibration gas 
injections at three different known concentration levels (low, mid, and 
high) to establish the accuracy of a gas analyzer across its 
measurement range. The procedures for conducting linearity tests are 
found in section 6.2 of Appendix A to part 75. Section 6.1 of Appendix 
A specifies that linearity tests must be done while the unit is 
operating.
    After the initial certification of a gas monitoring system, section 
2.2 of Appendix B to part 75 requires periodic linearity tests to be 
performed. A linearity check is required during each unit operating 
quarter or, for bypass stacks, during each quarter in which flue gases 
are discharged through the stack. For units with two span values for a 
particular parameter (e.g., units with add-on SO2 controls), 
linearity tests must be conducted on both the ``low'' and ``high'' 
monitor ranges. Successive linearity tests are, to the extent 
practicable, to be conducted no less than 2 months apart.
    Utility representatives have asked EPA to consider changing the 
requirement for the unit to be operating when linearity tests are done 
(see Docket A-97-35, Items II-D-20, II-D-65, II-E-13, II-E-14). This 
has been requested because owners and operators of peaking units and 
other units that operate on an ``on-call'' basis have experienced 
difficulty in complying with the requirement for the unit to be on-line 
during linearity tests. For instance, a unit may only operate for a few 
hours in a quarter and not be needed again until the next quarter. In 
such a situation, the utility might be forced to re-start and operate 
the unit (whether or not it is needed) to comply with the linearity 
test requirement. Some of the utility representatives have also 
expressed the opinion that for certain monitoring technologies (e.g., 
dry extractive), on-line and off-line linearity tests are essentially 
equivalent.
Discussion of Proposed Changes
1. Unit Operation During Linearity Tests
    Today's rule proposes to revise the linearity test requirements of 
part 75 to make them easier with which to comply. EPA agrees that the 
current linearity test requirements of part 75 lack flexibility and 
that compliance with the requirements is particularly difficult for 
infrequently operated units. However, the Agency does not agree with 
the utility representatives that have suggested allowing off-line 
linearity tests as the best solution to the problem. Nor is the Agency 
proposing to allow technology-specific exemptions to the on-line 
linearity test requirement.

[[Page 28061]]

Rather, today's proposal would retain the requirement for linearity 
tests to be performed while the unit is combusting fuel at conditions 
of typical stack temperature and pressure. A clarifying statement would 
be added to section 6.2 of Appendix A, indicating that the unit does 
not have to be generating electricity during the test. But EPA would 
continue to require that a linearity test be performed while the unit 
is combusting fuel at conditions of typical stack temperature and 
pressure in order to test the monitoring system under the same 
conditions as when the monitor is measuring emissions, in order to 
account for any temperature and pressure effects. An on-line linearity 
test challenges a CEMS while it is in equilibrium with the stack 
environment and has been sampling stack gas continuously for a period 
of time.
2. Linearity Test Frequency
    The Agency proposes instead to add flexibility to the linearity 
test requirements by changing the basis upon which the frequency of 
linearity tests is determined and by providing a linearity grace 
period. In today's proposal, section 2.2 of Appendix B would be revised 
to require that a linearity test be performed in each ``QA operating 
quarter'' rather than in each ``unit operating quarter'' or ``bypass 
stack operating quarter.'' For linearity tests, a QA operating quarter 
would be defined in the same way as for RATAs, i.e., as a calendar 
quarter in which the unit operates for at least 168 hours (or, for 
common stacks, a quarter in which effluent gases discharge through the 
stack for at least 168 hours). EPA believes that the QA operating 
quarter methodology would, in most instances, enable the owner or 
operator of a peaking unit or other infrequently operated unit to 
complete an on-line linearity test within the calendar quarter in which 
it is due. However, the following additional changes would be made to 
further ensure that the linearity test requirements can be met: (1) the 
requirement to perform successive linearity tests at least 2 months 
apart would be reduced to allow successive tests to be done one month 
(30 days) apart; and (2) a new section, 2.2.4, would be added to 
Appendix B, providing a 168 unit operating hour grace period after the 
end of each QA operating quarter in which to complete the required 
test. Thus, to make it easier for infrequently operated units to 
complete the required linearity tests in the quarters in which they are 
due, the required waiting time between successive linearity tests would 
be reduced. And, if circumstances should prevent a linearity test from 
being completed in the QA operating quarter in which it is due, the 
test could be done during the grace period. If the required linearity 
test were not completed by the end of the grace period, data from the 
monitor would be considered invalid from the hour after the grace 
period expires until the hour of completion of a subsequent successful 
linearity test.
    For infrequently operated units, certain calendar quarters would 
not qualify as QA operating quarters. Therefore, in accordance with 
today's proposed rule, no linearity tests would be required in those 
quarters. However, this exemption from linearity testing would not be 
without limit. Proposed section 2.2.2 of Appendix B would allow no more 
than four consecutive calendar quarters to elapse following the quarter 
in which the last linearity test was conducted, without a subsequent 
linearity test having to be performed. That is, a linearity test would 
either have to be done by the end of the fourth consecutive elapsed 
calendar quarter since the last test or within a 168 unit operating 
hour grace period after the end of the fourth consecutive elapsed 
quarter. Data from the monitor would become invalid if the linearity 
test was not completed by the end of the grace period and would remain 
invalid until a linearity test was successfully completed.
    Today's proposal would also change the requirement for units with 
two span values for a particular parameter (e.g., units with add-on 
SO2 controls) to perform quarterly linearity tests on both 
the low and high monitor ranges. Section 2.2.1 of Appendix B would be 
revised to require a linearity test of a monitor range only if that 
range is used to report data during the QA operating quarter. However, 
under proposed section 2.2.3(e) of Appendix B, at least one linearity 
test of each range would still be required every four calendar quarters 
to maintain data validation on the range.
3. Linearity Test Method
    Today's proposal would add two new requirements to section 6.2 of 
Appendix A: (1) that all linearity tests must be done ``hands-off,'' 
meaning that no adjustments of the CEMS other than certain calibration 
error adjustments would be permitted prior to or during the linearity 
test period; and (2) to the extent practicable, each linearity test 
would have to be completed within a period of 24 unit operating hours. 
These proposed provisions are intended to ensure greater consistency in 
the way in which linearity tests are conducted and to ensure that the 
tests are completed in a timely manner. The allowable calibration 
adjustments prior to and during a linearity test would be defined in 
proposed section 2.1.3 of Appendix B. For a further discussion, see 
Section O of this preamble, ``CEM Data Validation,'' below.
4. Exemptions
    Finally, section 6.2 of Appendix A would be revised to exempt 
SO2 and NO_X monitors with span values of 30 ppm 
or less from the linearity test requirements of part 75. At these low 
span values, the linearity test begins to lose its significance. For 
example, typical low, mid, and high calibration gases for a span value 
of 30.0 ppm would be 24.0 ppm, 18.0 ppm, and 9.0 ppm, respectively. The 
appropriate linearity performance specification in section 3.2 of 
Appendix A is  5.0 ppm at each calibration gas level. 
Therefore, in this illustration, the monitor reading could be 14.0 ppm 
for both the ``low'' and ``mid'' gases or 20.0 ppm for both the ``mid'' 
and ``high'' gases. Even though a valid straight line comparing the 
reference gas concentrations and the monitor readings cannot be 
constructed from such data, the monitor would still appear to pass the 
linearity test.

M. Flow-to-Load Test

Background
    The current quality assurance requirements for flow rate monitoring 
systems in Appendices A and B to part 75 include daily calibration 
error tests, daily interference checks, quarterly leak checks (for 
differential pressure type monitors only), and semiannual or annual 
relative accuracy test audits. Of these required QA tests, only the 
RATA provides a true evaluation of a flow monitor's measurement 
accuracy by direct comparison against an independent reference method. 
The daily calibration error test purports to check flow monitor 
accuracy, but, as explained below, the ability of the test to 
accomplish this objective is somewhat questionable.
    There is a distinct difference between the daily calibration error 
test of a flow rate monitor and the calibration error test of a gas 
monitor. To calibrate a gas monitor, a protocol gas of known 
concentration is sent through the monitoring system and analyzed. This 
generally serves as a reliable indicator of the system's ability to 
accurately measure pollutant or diluent gas concentrations, because the 
calibration closely simulates the sampling and analysis of stack gas by 
the monitoring

[[Page 28062]]

system. A flow monitor calibration error test, on the other hand, does 
not provide the same level of assurance of data quality. Generally, a 
flow monitor calibration checks the system's internal electronic 
components by means of reference signals. The calibration error test is 
useful in that it can diagnose certain types of monitor problems, but 
it is not a ``true'' calibration of the monitor, since it does not 
evaluate the system's ability to measure an actual stack gas flow rate. 
In order to perform true daily flow monitor calibrations, two reference 
stack gas flow rates would have to be generated and measured. Practical 
considerations preclude such calibrations from being done, however, 
because the unit load level would have to be significantly varied 
during each operating day, and suitable reference method measurements 
(e.g., velocity traverses using EPA Method 2) would have to be made 
daily at each calibration load level.
    Because of the limited usefulness of the flow monitor daily 
calibration error test, EPA believes that a more substantive, periodic 
QA test is needed to ensure that the accuracy of the reported flow rate 
data is maintained in the interval between successive RATAs. The Agency 
is particularly concerned about the potential for poor data quality 
from flow monitors that are not properly maintained. For instance, the 
sensors of DP and thermal-type monitors are subject to plugging and/or 
fouling, which will cause the monitors to read lower than true and can 
result in under-reporting of emissions. One utility observed a 
substantial increase in the readings from its flow monitor after the 
sensors were cleaned during a unit outage. Apparently, the sensor 
problems had not been detected by the daily calibration error tests 
(see Docket A-97-35, Item II-E-29). A second utility experienced a 
gradual deterioration of the monitor's performance in the 9-month 
period following the RATA. By the sixth month (at load levels and 
CO2 concentrations virtually identical to the conditions at 
the time of the RATA), the flow monitor readings were consistently 15.0 
to 20.0 percent lower than the baseline average flow rate measured by 
EPA Reference Method 2 during the RATA. However, during the 9-month 
period, the flow monitor had consistently passed its daily calibration 
error tests (see Docket A-97-35, Item II-B-11). During a State 
inspection of a third utility, the inspector observed a consistent 20.0 
to 30.0 percent difference between the hourly flow rates measured by 
the primary and redundant backup flow monitors even though both 
monitors had been passing their daily calibration error tests. In this 
instance, the primary flow monitor was being used for data reporting 
and was reading higher than the redundant backup monitor; therefore, it 
is unlikely that emissions were being under-reported. Had the primary 
monitor malfunctioned and the redundant backup been used, however, 
emissions would have been significantly under-reported (see Docket A-
97-35, Item II-B-10).
Discussion of Proposed Changes
    In view of the apparent shortcomings of the flow monitor daily 
calibration error test, EPA proposes to add a new flow monitor quality 
assurance test, the ``flow-to-load test,'' to part 75. The flow-to-load 
test, which would be performed quarterly, is described in proposed 
sections 7.7 of Appendix A and 2.2.5 of Appendix B. The proposed 
quarterly flow-to-load test would be required beginning in the first 
quarter of the year 2000.
    The basic premise of the flow-to-load test is that a meaningful 
correlation exists between the stack gas volumetric flow rate and unit 
load. In general, for a single unit discharging to a single stack, as 
the load increases, the flow rate increases proportionally, and the 
flow rate at a given load should remain relatively constant if the same 
type of fuel is burned (see Docket A-97-35, Items II-B-9, II-D-69). 
Common stacks are somewhat less predictable, because the same combined 
unit load can be produced in a number of ways by using different 
combinations of boilers. Despite this, if the diluent gas concentration 
is properly taken into account, the flow-to-load characteristics of 
common stacks often become more normalized (see Docket A-97-35, Items 
II-B-9, II-D-73, II-D-74, II-D-76, II-D-83, II-D-84). The flow-to-load 
ratio, or a normalized ratio, can thus serve as a quantitative 
indicator of flow monitor accuracy from quarter to quarter until the 
next RATA is performed.
    The quarterly flow-to-load ratio test would be conducted as 
follows. The owner or operator would be required to determine 
Rref, a reference value of the ratio of flow rate to unit 
load, each time that a successful normal-load flow RATA is performed. 
The value of Rref would be reported in the electronic 
quarterly report required under Sec. 75.64, along with the completion 
date of the associated RATA. If two load levels (e.g., mid and high) 
are designated as normal, the owner or operator would determine a 
separate Rref value for each normal load level. The 
reference flow-to-load ratio would be calculated as follows:
[GRAPHIC] [TIFF OMITTED] TP21MY98.059

    In the equation above, Rref is the reference value of 
the flow-to-load ratio from the most recent normal-load flow RATA; 
Qref is the average stack gas volumetric flow rate (in scfh) 
measured by the reference method during the normal-load RATA; and 
Lavg is the average unit load during the normal-load flow 
RATA. For a common stack, Lavg would be the sum of the 
operating loads of all units that discharge through the stack. For a 
unit that discharges its emissions through multiple stacks or ducts, 
Qref would be the sum of the total volumetric flowrates that 
discharge through all of the stacks (or ducts). The reference flow-to-
load ratio would be rounded off to 2 decimal places.
    As an alternative, the owner or operator could calculate a 
reference value of the gross heat rate (GHR) in lieu of 
Rref. In order to exercise this option, quality assured 
diluent gas (CO2 or O2) data would have to be 
available for each hour of the most recent normal-load flow RATA. The 
reference value of the GHR would be determined as follows:
[GRAPHIC] [TIFF OMITTED] TP21MY98.060

    In the equation above, (GHR)ref is the reference value 
of the gross heat rate at the time of the most recent normal-load flow 
RATA; (Heat Input)avg is the arithmetic average hourly heat 
input during the normal-load flow RATA; and Lavg is the 
average unit load during the normal-load flow RATA. In calculating 
(Heat Input)avg, the average volumetric flow rate measured 
by the reference method during the RATA would be used in conjunction 
with the average diluent gas concentration measured during the RATA, 
substituting these values into the applicable heat input equation in 
Appendix F.
    After establishing the reference flow-to-load or GHR value, an 
evaluation of the flow-to-load ratio or GHR would be required for each 
primary and redundant backup flow monitor on a quarterly basis. The 
owner or operator would be required to evaluate the flow-to-load ratio 
in each ``QA operating quarter'' (i.e., each quarter in which the unit 
or stack operates for at least 168 hours). At the end of each QA 
operating quarter, the owner or operator would calculate the flow-to-
load ratio for every hour during the quarter in which: (1) the unit (or 
combination of units, for a common stack) operated within 
10.0 percent of Lavg, the average load during 
the most recent normal-load flow

[[Page 28063]]

RATA; and (2) a quality assured hourly average flow rate was obtained 
with a certified flow rate monitor. The owner or operator would have 
the option of using either bias-adjusted flow rates or unadjusted flow 
rates in the hourly flow-to-load ratios, provided that all of the 
ratios were calculated the same way. EPA had originally considered 
proposing that only unadjusted flow rates should be used to calculate 
the flow-to-load ratios. However, in response to comments received from 
CEMS Utility Workgroup members, the Agency is proposing to allow either 
unadjusted or bias-adjusted flow rates to be used, on the condition 
that the acceptance criteria for the flow-to-load test would be more 
stringent if bias-adjusted flow rates are used (see Docket A-97-35, 
Item II-D-82).
    For a common stack, the ``load'' in each hourly flow-to-load ratio 
would be the sum of the hourly operating loads of all units that 
discharge through the stack. For a unit that discharges its emissions 
through multiple stacks (or for a unit that monitors total flow rate in 
multiple ducts or breechings), the ``flow'' in the flow-to-load ratio 
would be the combined hourly volumetric flow rate through all of the 
stacks (or ducts). Each hourly flow-to-load ratio would be rounded off 
to 2 decimal places.
    Alternatively, the owner or operator could calculate the hourly 
gross heat rate (GHR) values in lieu of the hourly flow-to-load ratios. 
However, an hourly GHR could only be determined for those hours within 
10.0 
 
  
for which quality assured flow rate and diluent gas (CO2 or 
O2) concentration data are available from a certified CEMS 
or reference method. The owner or operator could use either bias-
adjusted flow rates or unadjusted flow rates to determine the hourly 
GHR values.
    The calculated hourly flow-to-load ratios (or gross heat rates) 
would be analyzed at the end of the quarter. A separate data analysis 
would be performed for each primary and each redundant backup flow rate 
monitor used to record and report data during the quarter. Each 
analysis would be based on a minimum of 168 hours of data. If two RATA 
load levels are designated as normal, the analysis would be performed 
at the higher load unless fewer than 168 data points were available at 
that load, in which case, the analysis would be performed at the lower 
load. If, for a particular flow monitor, fewer than 168 hourly flow-to-
load ratios (or GHR values) were available at any normal load level, a 
flow-to-load (or GHR) evaluation would not be required for that monitor 
for that calendar quarter.
    For each flow monitor, Eh, the difference (absolute 
value) between each hourly flow-to-load ratio and Rref, 
would be expressed as a percentage of Rref (or, if the GHR 
is used, the absolute difference between each hourly GHR value and 
(GHR)ref would be expressed as a percentage of 
(GHR)ref). Then, Ef, the arithmetic average of 
all of the Eh values, would be calculated. Note that 
Rref would always be based upon the most recent normal-load 
RATA, even if that RATA was performed in the calendar quarter being 
evaluated.
    The owner or operator would be required to report the results of 
each quarterly flow-to-load (or GHR) evaluation in the electronic 
quarterly report required under Sec. 75.64. The results of a quarterly 
flow-to-load (or GHR) evaluation would be considered acceptable, and no 
further action would be required if the average absolute percentage 
difference (Ef) did not exceed the following limits:
    (i) 15.0 percent, if Lavg for the most recent normal 
load flow RATA is  50 megawatts (or  500 klb/hr 
of steam) and if unadjusted flow rates were used in the calculations;
    (ii) 10.0 percent, if Lavg for the most recent normal 
load flow RATA is  50 megawatts (or  500 klb/hr 
of steam) and if bias-adjusted flow rates were used in the 
calculations;
    (iii) 20.0 percent, if Lavg for the most recent normal 
load flow RATA is < 50 megawatts (or < 500 klb/hr of steam) and if 
unadjusted flow rates were used in the calculations;
    (iv) 15.0 percent, if Lavg for the most recent normal 
load flow RATA is < 50 megawatts (or < 500 klb/hr of steam) and if 
bias-adjusted flow rates were used in the calculations.
    If Ef exceeded the applicable limit, the owner or 
operator would have two available options: (1) perform a RATA, as 
described in proposed section 2.2.5.2 of Appendix B, unless a monitor 
malfunction is diagnosed and corrected, in which case an abbreviated 
flow-to-load test could be performed, in lieu of a RATA, in accordance 
with section 2.2.5.3 of Appendix B and discussed below; or (2) re-
examine the hourly data used for the flow-to-load or GHR analysis and 
recalculate Ef, after excluding all non-representative 
hourly flow rates. If the owner or operator were to choose option (2), 
i.e., to recalculate Ef, only the flow rates for the 
following hours would be considered non-representative and could be 
excluded from the data analysis:
    (1) Any hour in which the type of fuel combusted was different from 
the fuel burned during the most recent normal-load RATA. The type of 
fuel would be different if the fuel is in a different state of matter 
(i.e., solid, liquid, or gas) or is a different classification of coal 
(e.g., bituminous versus sub-bituminous) than the fuel burned during 
the RATA;
    (2) Any hour in which an SO2 scrubber was bypassed;
    (3) Any hour in which ``ramping'' occurred, i.e., the hourly load 
differed by more than + 15.0 percent from the load during the preceding 
hour or the subsequent hour;
    (4) If a normal-load flow RATA was performed and passed during the 
quarter being analyzed, any hour prior to completion of that RATA; and
    (5) If a problem with the accuracy of the flow monitor was 
discovered during the quarter and corrected, any hour prior to 
completion of the subsequent diagnostic test described in proposed 
section 2.2.5.3 of Appendix B, confirming that the corrective actions 
were successful.
    After identifying and excluding any non-representative hourly data 
in accordance with (1) through (5) above, the owner or operator could 
analyze the remaining data a second time. At least 168 representative 
hourly ratios or GHR values at normal load would have to remain in 
order to perform the analysis; otherwise, the flow-to-load (or GHR) 
analysis would not be required for that monitor for that calendar 
quarter.
    If, after re-analyzing the data, Ef is found to be 
within the applicable limit in (i), (ii), (iii), or (iv), above, then 
no further action would be required. However, if Ef is still 
outside the applicable limit, the monitor would be declared out-of-
control as of the first hour of the quarter following the quarter in 
which the flow-to-load test was failed. The owner or operator would 
then perform a RATA as described in proposed section 2.2.5.2 of 
Appendix B, unless, as the result of an investigation, an instrument 
malfunction is discovered and corrected as described in proposed 
section 2.2.5.1 of Appendix B.
    If a problem with the monitor is identified, all corrective actions 
(e.g., non-routine maintenance, repairs, major component replacements, 
re-linearization of the monitor, etc.) would have to be documented in 
the operation and maintenance records for the monitor. Data from the 
monitor would remain invalid until a ``probationary'' calibration error 
test of the monitor was passed following completion of all corrective 
actions, at which point data from the monitor would be assigned a 
``conditionally valid'' status. The owner or operator would then 
perform an abbreviated flow-to-load test (found in proposed section 
2.2.5.3 of Appendix B) to verify that the corrective actions were

[[Page 28064]]

effective, unless the linearity of the flow monitor was affected by the 
corrective actions (e.g., by the changing of its polynomial 
coefficients). If the flow monitor linearity was affected, the owner or 
operator would no longer have the option of performing the abbreviated 
flow-to-load test in section 2.2.5.3 of Appendix B, but would instead 
be required to perform a 3-load recertification RATA in accordance with 
the recertification test period and data validation procedures of 
Sec. 75.20(b)(3).
    The abbreviated flow-to-load test in proposed section 2.2.5.3 of 
Appendix B is based on a recertification policy developed jointly by 
EPA, several utility representatives, and one flow monitor vendor (see 
Docket A-97-35, Items II-B-1, II-D-70, II-I-9, and II-I-16). Use of the 
abbreviated flow-to-load test would not be limited to situations in 
which a quarterly flow-to-load test has been failed. Rather, the test 
could be performed after any documented repair, component replacement, 
or other corrective maintenance to a flow monitor (except for changes 
affecting the linearity of the flow monitor, such as adjusting the flow 
monitor coefficients) to demonstrate that the repair, replacement, or 
other corrective maintenance has not significantly affected the 
monitor's ability to accurately measure the stack gas volumetric flow 
rate. Data from the monitoring system would be considered invalid from 
the hour of commencement of the repair, replacement, or other 
corrective maintenance until the hour in which a ``probationary'' 
calibration error test is passed following completion of the repair, 
replacement, or other corrective maintenance and any associated 
adjustments to the monitor. The abbreviated flow-to-load test would 
have to be completed within 168 unit operating hours of the 
probationary calibration error test (or, for peaking units, within 30 
unit operating days, if that is less restrictive). Data from the 
monitor would be considered ``conditionally valid'' (as defined in 
Sec. 72.2) beginning with the hour of the probationary calibration 
error test.
    Following a flow-to-load test failure, the abbreviated flow-to-load 
test could be performed if the investigation into the cause of the test 
failure revealed a problem with the flow monitor and the problem was 
subsequently corrected without having to re-linearize the flow monitor. 
The test procedures would be as follows. The unit(s) would be operated 
in such a way as to reproduce, as closely as practicable, the exact 
conditions at the time of the most recent normal load flow RATA. To 
achieve this, the load should be held constant to within  
5.0 percent of the average load during the RATA, and the diluent gas 
(CO2 or O2) concentration should be maintained 
within  0.5 percent CO2 or O2 of the 
average diluent concentration during the RATA. For common stacks, to 
the extent possible, the same combination of units and load levels that 
were used during the RATA should be used. When the process parameters 
have been set, a minimum of 6 and a maximum of 12 consecutive hourly 
average flow rates would be recorded using the flow monitor(s) for 
which Ef was outside the applicable limit. For peaking 
units, a minimum of 3 and a maximum of 12 consecutive hourly average 
flow rates would be required. The corresponding hourly load values and, 
if applicable, the hourly diluent gas concentrations would also be 
recorded. The flow-to-load ratio or the GHR would be calculated for 
each hour in the test hour period using proposed Equation B-1 or B-1a 
in Appendix B. Then, Eh would be determined for each hourly 
flow-to-load ratio or GHR using proposed Equation B-2 in Appendix B. 
Finally, Ef , the arithmetic average of the Eh 
values, would be determined.
    The results of the abbreviated flow-to-load test would be 
considered acceptable, and no further action would be required if the 
value of Ef did not exceed the applicable limit specified in 
proposed section 2.2.5.1 of Appendix B. All conditionally valid data 
recorded by the flow monitor would then be considered quality assured, 
beginning with the hour of the probationary calibration error test that 
preceded the abbreviated flow-to-load test. However, if Ef 
was found to be above the applicable limit, all conditionally valid 
data recorded by the flow monitor would be considered invalid back to 
the hour of the probationary calibration error test that preceded the 
abbreviated flow-to-load test, and a single-load RATA would be 
required, in accordance with proposed section 2.2.5.2 of Appendix B.
    When a single-load RATA is performed because the owner or operator 
is unable to reconcile a quarterly flow-to-load test failure, either by 
excluding non-representative hours and recalculating Ef or 
by passing the abbreviated flow-to-load test after performing component 
replacement or other corrective maintenance on the flow monitor, then 
data from the monitor would remain invalid until the hour of successful 
completion of the single-load RATA.
Rationale
    EPA believes that the proposed methodology for the quarterly flow-
to-load test is fundamentally sound. It has been developed through a 
series of teleconferences and face-to-face meetings between EPA, 
members of the regulated community, and State and local agency 
personnel (see Docket A-97-35, Items II-D-77, II-D-80, II-D-81, II-D-
82, II-D-85, II-E-23, II-E-24, II-E-25, II-E-26, and II-E-28). In 
addition, some provisions of the flow-to-load test were revised 
following pre-proposal comment. Specifically, the proposal reflects, in 
section 2.2.5.1 (b) of Appendix B to part 75, a commenter's request 
that if a quarterly flow-to-load test is failed and the monitor 
malfunction is discovered and corrected (without the need to 
relinearize the monitor), the correction could be verified using the 
abbreviated flow-to-load test in lieu of performing a single load RATA 
(see Docket A-97-35, Item II-D-42).
    The proposed tolerance limits set forth in paragraphs (i), (ii), 
(iii), and (iv) of section 2.2.5 of Appendix B are believed to be both 
reasonable and achievable. When these tolerance limits are met, it 
provides a strong indication that the flow monitor is still accurate to 
within 10.0 percent of the reference method baseline established during 
the last normal-load flow RATA and would, therefore, appear to be in 
control with respect to the relative accuracy requirements of part 75. 
An extra tolerance of 5.0 percent has been incorporated into the limits 
to account for imprecision in the flow-to-load methodology. An extra 
5.0 percent tolerance has also been added for smaller units (i.e., 
normal load less than 50 megawatts or 500 klb/hr of steam), because the 
flow-to-load ratio or GHR for such units is very sensitive to small 
variations in load (see Docket A-97-35, Item II-B-7).
    To test the viability of the proposed tolerance limits, EPA 
analyzed quarterly flow rate and load data from the third quarter of 
1996 for 21 units and stacks, including 9 single units, 11 common 
stacks, and 1 multiple-stack unit (see Docket A-97-35, Items II-A-1, 
II-A-2, II-A-3). The units chosen for this analysis were selected as a 
representative sample of units that would be affected by this QA test 
requirement and included various operational circumstances (e.g., 
baseloaded and peaking units, single fuel units, and units that burn 
multiple fuels). The flow-to-load test was applied to each unit or 
stack in the manner described above, except that no hours within 
 10.0 percent of Lavg were excluded from the 
data analysis. The data from these same units plus one additional 
multiple-stack unit were

[[Page 28065]]

analyzed a second time, with each flow-to-load ratio being multiplied 
by the diluent gas concentration. This is similar, but not identical, 
to calculating the GHR. Once again, no hours within  10.0 
percent of Lavg were excluded. In both analyses, unadjusted 
flow rates were used in the ratios. The results of the two data 
analyses were nearly the same. Only one failure of the quarterly flow-
to-load test was observed in each analysis (i.e., the failure rate was 
< 5.0 percent). The average value of Ef was 6.1 percent for 
the analysis without the diluent gas corrections and 6.4 percent for 
the analysis with the diluent gas corrections. A few units and stacks 
had a much lower Ef value when the diluent correction was 
applied, but in most cases, the diluent correction had relatively 
little effect. These results suggest that the flow-to-load test can 
provide EPA with the necessary assurance that flow monitors continue to 
generate accurate data from one RATA to the next. The results also 
indicate that the test should be relatively easy to pass if flow 
monitors are properly maintained and operated.
    Because of the added quality assurance that would be provided by 
performing the flow-to-load or GHR test each quarter, EPA has 
reconsidered the scope of the other quality assurance tests for flow 
monitors. In today's proposed rule, the Agency is proposing to reduce 
the annual 3-load flow RATA requirement to a 2-load RATA and to reduce 
the frequency of 3-load RATAs to once every five years (and whenever a 
flow monitor is re-linearized). In addition, single-load flow RATA 
testing would be allowed in lieu of the annual 2-load test if the 
facility could demonstrate that a unit has operated at a single load 
level for at least 85.0 percent of the time in the four ``QA operating 
quarters'' prior to the scheduled RATA. (See Section N.2 of this 
preamble, below, for further discussion.) The Agency believes that, 
taken together, these proposed changes will reduce the cost and burden 
of quality assurance testing for flow monitors, while ensuring high 
data quality. The proposed reduction in the amount of required RATA 
testing is considered feasible because of the increased quality 
assurance provided by the quarterly flow-to-load test. EPA requests 
comment on the proposed revisions to flow monitor quality assurance 
requirements.

N. RATA and Bias Test Requirements

Background
    Section 6.5 of Appendix A to the January 11, 1993 rule, as amended 
on May 17, 1995 and November 20, 1996, requires relative accuracy test 
audits of all primary and redundant backup SO2, 
NO_X, CO2, and flow monitoring systems to be 
performed during the initial certification of the CEMS. A RATA consists 
of a series of 9 or more simultaneous test runs, comparing measurements 
made by the continuous monitoring system against an EPA reference test 
method. The procedures for conducting RATAs are found in section 6.5 of 
Appendix A to part 75.
    Following the initial certification of a CEMS, section 2.3 of 
Appendix B to part 75 requires that periodic RATAs of gas and flow 
monitors be performed to quality assure the data from the CEMS on an 
on-going basis. The frequency at which relative accuracy testing is 
required depends upon the results of the last RATA of a monitoring 
system. Part 75 currently requires RATAs to be performed semiannually, 
unless a monitoring system achieves a low enough relative accuracy to 
qualify for an annual test frequency. The Agency has always interpreted 
``semiannually'' to mean that the deadline for the next RATA is the end 
of the second calendar quarter following the quarter in which a RATA is 
successfully completed, and ``annually'' to mean that the next RATA is 
due by the end of the fourth calendar quarter following the quarter in 
which a RATA is successfully completed. For monitors installed on 
peaking units and bypass stacks, however, the RATA deadlines are based 
on operating quarters, not calendar quarters. That is, the next RATA is 
due either at the end of the second or fourth unit operating quarter 
(for peaking units) or bypass stack operating quarter following the 
quarter in which a RATA is successfully completed.
    For SO2, NO_X, and CO2 monitors, 
the RATAs are to be conducted while the unit is operating at normal 
load and while combusting the fuel that is normal for the unit. Flow 
monitor RATAs are to be conducted at three different loads, evenly 
spaced over the operating range of the unit. When a flow monitor is on 
a semiannual RATA frequency, a normal-load RATA rather than a 3-load 
RATA may be conducted to satisfy the semiannual test requirement, but a 
3-load RATA is still required annually. Note that for flow monitors 
installed on peaking units and bypass stacks, 3-level flow RATAs are 
not required; RATAs are performed only at the normal load.
    For SO2, NO_X, and flow monitoring systems, 
section 7.6 of Appendix A requires that each time a RATA is 
successfully completed, a bias test be performed to determine if the 
system has a low measurement bias. If a monitoring system fails the 
bias test, a ``bias adjustment factor'' (BAF) must be applied to all 
subsequent emission data reported from that monitoring system. For 3-
load flow RATAs, the bias test is done at the normal load. If a flow 
monitor fails the normal-load bias test, then a BAF must be calculated 
at each of the three load levels, and the highest of the three BAFs is 
applied to all flow data reported from the monitor.
    When a RATA is due, section 2.3.1 in Appendix B of the rule allows 
the owner or operator two attempts to achieve an annual RATA frequency 
and/or a favorable BAF. If a second attempt is made, the RATA frequency 
and BAF obtained in the second RATA supersede the results of the first 
RATA. Once the RATA frequency has been established as semiannual or 
annual, section 2.3.1 of Appendix B specifies that (to the extent 
practicable) the next RATA of the CEMS may not be done until at least 
four months have elapsed.
    Finally, Sec. 75.21(a)(6) of the November 20, 1996 rule provides an 
exemption from the RATA requirements of part 75 for SO2 
monitors installed on units that burn only natural gas or fuel with a 
sulfur content no greater than natural gas. For units that burn both 
gas and higher-sulfur fuel, such as oil, as primary or backup fuels, 
Sec. 75.21(a)(5) requires that the RATA of the SO2 monitor 
be done when the higher-sulfur fuel is burned. Section 75.21(a)(7) 
further states that calendar quarters in which only fuel with a sulfur 
content no greater than natural gas is burned are to be excluded in 
determining the deadline for the next SO2 monitor RATA.
    Two utility groups, UARG and the Class of '85, have requested that 
EPA consider revising the RATA requirements of part 75 to make them 
more flexible, easier with which to comply, and less costly. Some of 
the possible changes suggested by these groups are as follows: (1) 
reduce the frequency of required RATAs; (2) determine RATA deadlines 
based on the amount of unit operation since the last RATA, rather than 
the number of calendar quarters that have elapsed; (3) remove the 
requirement to achieve a more stringent relative accuracy standard in 
order to obtain an annual RATA frequency; (4) except for initial 
certification, allow flow RATAs to be done at a single load; (5) allow 
single-point sampling during gas RATAs; and (6) allow a grace period in 
which to complete a RATA whenever a deadline is not met (see Docket A-
97-35, items II-D-20, II-D-30, II-D-65, II-E-13, II-E-14).

[[Page 28066]]

Discussion of Proposed Changes
    EPA is proposing revisions to the RATA requirements of part 75 
based upon experience gained through implementation of the rule and in 
light of the recommendations made by the utility groups. Today's 
rulemaking sets forth the proposed changes, which are intended to make 
the RATA requirements less burdensome without sacrificing data quality.
1. RATA Frequency
    EPA does not propose to revise the basic semiannual and annual RATA 
requirements of part 75 or the incentive system by which to obtain an 
annual RATA frequency (i.e., to obtain the reduced frequency, a better 
percentage relative accuracy is required). Instead, the Agency proposes 
to re-define the terms ``semiannual RATA frequency'' and ``annual RATA 
frequency,'' and to change the method by which RATA deadlines are 
determined.
    Today's rule proposes to amend section 2.3 of Appendix B so that 
the deadline for the next RATA is determined on the basis of ``quality 
assurance operating quarters,'' rather than calendar quarters. This 
change would apply, with few exceptions, to all primary and redundant 
backup monitoring systems, including monitors installed on peaking 
units and bypass stacks. A ``QA operating quarter'' would be defined as 
a calendar quarter in which a unit operates for at least 168 hours or, 
for common-stacks and bypass stacks, a quarter in which flue gases 
discharge through the stack for at least 168 hours.
    Any calendar quarter that does not qualify as a QA operating 
quarter would be excluded in determining the deadline for the next 
RATA. EPA therefore proposes to re-define the term ``semiannual RATA 
frequency'' to mean that the next RATA is due at the end of the second 
QA operating quarter following the quarter in which a RATA is 
successfully completed. Similarly, ``annual RATA frequency'' would mean 
that the next RATA is due at the end of the fourth QA operating quarter 
following the quarter in which a RATA is successfully completed.
    The QA operating quarter methodology has been proposed principally 
for the benefit of cycling and peaking units to make the part 75 RATA 
requirements easier to meet. The proposed methodology will not greatly 
affect base-loaded units, since they seldom operate for less than 168 
hours in a quarter. For base-loaded units, the QA operating quarter 
method is, in most instances, equivalent to the familiar calendar 
quarter scheme for determining RATA deadlines. Note, however, that on 
occasion a base-loaded unit may obtain an extended RATA deadline by the 
QA operating quarter methodology, e.g., when the unit goes into an 
extended outage (planned or forced) and experiences one or more 
quarters in which the unit operates for less than 168 hours.
    Although the QA operating quarter method allows RATA deadlines to 
be extended by the exclusion of quarters in which the unit(s) operate 
for less than 168 hours, such exclusion of calendar quarters is not 
without limit. Section 2.3.1.1 of Appendix B proposes to allow a 
maximum of eight consecutive calendar quarters to elapse following the 
quarter in which the last RATA was performed. A RATA would either have 
to be performed by the end of the eighth consecutive elapsed calendar 
quarter since the last RATA or within a 720 unit operating hour ``grace 
period'' following the end of the eighth consecutive elapsed quarter. 
Failure to complete a RATA within the grace period would cause data 
from the monitoring system to become invalid from the hour of 
expiration of the grace period until the hour of completion of a 
successful RATA.
    Although the proposed QA operating quarter methodology would serve 
as the basis for determining the RATA deadline for most routine quality 
assurance RATAs, there are five notable instances in the current rule 
or in today's proposal where the RATA deadline is either not determined 
solely on that basis or is determined entirely on another basis. The 
first instance is for a unit that burns both natural gas (or fuel with 
equivalent total sulfur content) and other higher-sulfur fuels as 
primary or backup fuels and that uses an SO2 monitor to 
account for SO2 mass emissions. Section 75.21(a)(7) of the 
current part 75 (redesignated as Sec. 75.21(a)(9) in today's proposal) 
specifies that irrespective of the number of hours of unit operation in 
the quarter, any calendar quarter in which natural gas (or fuel with a 
total sulfur content no greater than the total sulfur content of 
natural gas) is the only fuel combusted in the unit (i.e., a ``gas-
only'' quarter) is to be excluded in determining the deadline for the 
next RATA of the SO2 monitoring system. Section 75.21(a)(5) 
of the current rule further states that for such units, the RATA of an 
SO2 monitoring system is to be performed only when the 
higher-sulfur fuel is being combusted. Second, as discussed in section 
III.N.6 of this preamble, Sec. 75.21(a)(7) of today's proposed rule 
would conditionally exempt from SO2 RATA requirements any 
unit certified by the designated representative to burn fuel(s) with a 
sulfur content greater than natural gas only as emergency backup fuel 
or for short-term testing, provided that the annual usage of the 
higher-sulfur fuel(s) is kept below 480 hours. However if, during any 
quarter, the annual usage of the higher-sulfur fuel exceeded 480 hours, 
an SO2 RATA would be required either in that quarter or 
during a subsequent grace period. Thus, for RATAs of SO2 
monitoring systems, it is evident that the number of unit operating 
hours in a calendar quarter is not the only consideration that 
determines the deadline for the next RATA; the total sulfur content of 
the fuel being combusted must also be considered. Third, as discussed 
in section III.O.6 of this preamble, for certain non-redundant backup 
monitoring systems, Sec. 75.20(d) of today's proposal would require a 
periodic RATA every eight calendar quarters (rather than QA operating 
quarters). Fourth, as discussed in section III.N.2 of this preamble, 
under section 2.3.1.3 of Appendix B in today's proposal, 3-level flow 
RATAs would have to be performed once in every period of five 
consecutive calendar years (e.g., prior to permit renewal) and whenever 
a flow monitor is re-linearized. Fifth, as discussed in section III.O.4 
of this preamble, for recertification RATAs, which are not regularly 
scheduled tests, but are done on an ``as-required'' basis, 
Sec. 75.20(b)(3) of today's proposal specifies that the deadline for 
completing such RATAs would be 720 unit operating hours after the start 
of the recertification test period.
2. RATA Load Levels
    Today's proposed rule would more clearly define the load levels at 
which RATAs are done in order to provide greater consistency in the way 
that RATAs are performed. The current provisions of part 75 are neither 
sufficiently standardized nor clear in defining the appropriate RATA 
load levels, particularly for flow RATAs. For example, section 6.5.2 of 
Appendix A specifies that the ``low'' load audit point for a 3-level 
flow RATA can be located anywhere from the minimum safe, stable load to 
50.0 percent of the maximum load. Also, there is no minimum required 
load separation between the audit points at adjacent load levels. If 
adjacent audit points are too close together, a 3-level flow evaluation 
loses its significance. Finally, while the current rule requires gas 
and flow RATAs to be conducted at normal

[[Page 28067]]

load, no definition of normal load is provided. It could be inferred 
from the current section 6.5.2 of Appendix A that the ``mid'' load 
level is considered normal because it requires the 3-load RATA to be 
done at a frequently used low load, a frequently used high operating 
load, and a normal load. However, experience in implementing the 
program has shown that for many units, the high load level is 
considered normal by the facility. For a few units, low load is 
considered normal, and for still others, the normal load can depend 
upon the time of day or the season of the year.
    Proposed section 6.5.2.1 of Appendix A would therefore require the 
owner or operator first to define the ``range of operation'' for each 
unit or common stack equipped with hardware CEMS. The range of 
operation would extend from the minimum safe, stable load to the 
``maximum sustainable load,'' which is the higher of: (a) the nameplate 
capacity of the unit (less any physical or regulatory deratings), or 
(b) the highest sustainable load, based on at least four quarters of 
representative historical data. For a common stack, the lower boundary 
of the range of operation would be the lowest minimum safe, stable load 
for any of the individual units using the stack. The upper boundary of 
the range would be obtained by adding together the maximum sustainable 
loads of all units using the stack, or if that combined load is 
unattainable in practice, by using the highest sustainable combined 
load based on at least four quarters of representative historical data. 
Three load levels would then be defined in terms of the range of 
operation. The ``low'' level would be the lower 30.0 percent of the 
range; the ``mid'' level would be the central portion (30.0 percent to 
60.0 percent) of the range; and the ``high'' level would be 60.0 
percent to 100.0 percent of the range. Proposed section 6.5.2 of 
Appendix A would specify that for multi-level flow RATAs, the audit 
points at adjacent load levels (e.g., low and mid, or mid and high) 
must be separated by no less than 25.0 percent of the range of 
operation. The owner or operator would be required to report the upper 
and lower boundaries of the range of operation in the electronic 
quarterly report required under Sec. 75.64.
    Section 6.5.2.1 of Appendix A in today's proposal would further 
require the owner or operator to determine, for each unit or common 
stack on which CEMs are installed (except for peaking units), the two 
load levels (low, mid, or high) that are the most frequently used. The 
two-fold purpose of this determination, which would be required, at a 
minimum, annually (just prior to the annual quality assurance RATAs and 
in the same calendar quarter as the RATAs), would be to identify the 
normal load level(s) and to identify the two load levels that are the 
most appropriate for annual 2-level flow monitor audits and for flow 
monitor bias adjustment factor calculations. To make the determination, 
the owner or operator would construct an historical load frequency 
distribution (e.g., histogram), depicting the relative number of 
operating hours at each of the three load levels, low, mid, and high. 
The frequency distribution would be based upon all available data from 
the four most recent QA operating quarters, as defined in proposed 
section 2.3.1.1 of Appendix B. The load frequency distribution would be 
used to determine the percentage of the time (to the nearest 0.1 
percent) that each load level (low, mid, and high) has been used in 
recent history and thereby to identify the two most frequently used 
load levels. A summary of the data used for these determinations would 
be maintained on-site in a format suitable for inspection, and the 
results of the determinations would be included in the electronic 
quarterly report under Sec. 75.64. The proposed revisions discussed in 
this paragraph would become effective as of January 1, 2000.
    The owner or operator would be required under proposed section 
6.5.2.1 of Appendix A to designate the most frequently used load level 
(low, mid, or high) as the normal load level for each unit or common 
stack (except for peaking units). The owner or operator would also have 
the option of designating the second most frequently used load level as 
an additional normal load level. Today's proposal would, therefore, not 
limit normal load to a single load level. This way of defining normal 
load is particularly appropriate for units that operate on a diurnal 
cycle and units that operate at distinctly different load levels during 
different seasons of the year due to ambient conditions, electrical 
demand, etc. EPA believes that the added flexibility in the definition 
of normal load (i.e., not confining it to a single load level) will 
allow the normal-load RATA requirements of part 75 to be more easily 
met. The owner or operator would be required to identify the selected 
normal load level(s) in the electronic quarterly report required under 
Sec. 75.64. For peaking units, the entire range of operation would, for 
simplicity, be considered normal.
    Revisions to section 2.3.1.3 of Appendix B are proposed in today's 
rule, requiring the routine quality assurance RATAs of flow monitors to 
be done as follows. For flow monitors installed on peaking units and 
bypass stacks, no changes are proposed; the requirement to perform only 
single-load flow RATAs at normal load would be retained. For all other 
flow monitors, the routine semiannual and annual RATAs would be done at 
2 loads (i.e., the two most frequently used load levels, as identified 
in section 6.5.2.1 of Appendix A), with two exceptions: (1) the 2-load 
flow RATA could be performed alternately with a single-load flow RATA 
at the most frequently used (normal) load level, if the flow monitor is 
on a semiannual RATA frequency; and (2) a single-load flow RATA at the 
most frequently used load level could be performed in lieu of the 2-
load RATA if, for the four QA operating quarters prior to the quarter 
in which the RATA is conducted, the historical load frequency 
distribution constructed under section 6.5.2.1 of Appendix A shows that 
the unit has operated at the most frequently used load level for 
 85.0 percent of the time. For all units, the requirement to 
perform periodic 3-load flow RATAs would be retained, but the frequency 
would be changed from annual to once every five calendar years. A 3-
load RATA would also be required whenever a flow monitor is re-
linearized (i.e., when its polynomial coefficients are changed). EPA is 
proposing to reduce the required frequency of 3-load RATAs and to allow 
limited use of single-load flow RATA testing principally because of the 
added assurance of data quality that will be provided by the proposed 
quarterly flow-to-load test.
3. Flow Monitor Bias Adjustment Factors
    Today's rulemaking proposes to change the method of determining the 
bias adjustment factor for multiple-load flow RATAs. For 2-load RATAs 
(which would be done at the two most frequently used load levels as 
identified in proposed section 6.5.2.1 of Appendix A), the bias test 
would be done at the load level (or levels) designated as normal. If 
the monitor were to fail the bias test at any load level designated as 
normal, a bias adjustment factor (BAF) would be calculated at both load 
levels, and the higher of the two BAFs would then be applied to the 
subsequent flow data. For 3-load RATAs, the bias test would be required 
at each load level designated as normal under proposed section 6.5.2.1 
of Appendix A. If the bias test were failed at any load level 
designated as normal, BAFs would be calculated only at the two most 
frequently used load levels (not all three

[[Page 28068]]

levels), and the higher of the two BAFs would be applied to subsequent 
flow data. Thus, for all multiple-load flow RATAs, the appropriate BAF 
would be determined in the same way. For 3-load RATAs, this methodology 
for determining the BAF when the normal-load bias test is failed 
differs from the current rule, which requires the highest BAF from any 
of the three levels to be applied to subsequent data. Experience gained 
in the first few years of program implementation has shown that in many 
instances, the highest BAF has been from a load level that is seldom 
used (generally the low load level), which can result in an 
unrepresentatively high BAF being applied to the normal-load flow rate 
data.
4. Number of RATA Attempts
    Section 2.3.1.4 of Appendix B to today's proposed rule would remove 
the restriction limiting to two the number of RATA attempts that may be 
done to achieve an annual RATA frequency. In addition, the requirement 
that successive RATAs be conducted no less than 4 months apart would be 
removed from section 2.3.1 of Appendix B. The proposed rule would 
conditionally allow the owner or operator to perform as many RATAs as 
are necessary to achieve a better relative accuracy percentage or a 
more favorable bias adjustment factor, the condition being that the 
data validation procedures for RATAs in proposed section 2.3.2 of 
Appendix B would have to be followed (these procedures are discussed in 
detail in Section II.O of this preamble, ``CEM Data Validation''). The 
Agency believes that this extra flexibility will provide an incentive 
for owners or operators to optimize CEMS performance and to eliminate 
bias from their monitoring systems and to reduce the frequency of the 
required RATAs.
5. Concurrent SO2 and Flow RATAs
    Today's proposed rulemaking would delete the requirement for 
concurrent SO2 and flow RATA testing from Sec. 6.5 of 
Appendix A. This requirement was included in the January 11, 1993 rule 
in order to generate a data base from which EPA could determine the 
appropriateness of setting a combined flow rate-SO2 system 
relative accuracy specification. Section 3.3.5 of Appendix A was 
reserved for this future standard, which, if promulgated, would have 
become effective on January 1, 2000. After three years of program 
implementation, data collection, and evaluation, however, the Agency 
believes it is not appropriate or necessary to propose a combined flow 
rate-SO2 system relative accuracy standard. Instead, EPA 
believes it would be more appropriate to retain the individual relative 
accuracy specifications for the SO2 and flow monitors. 
Because the historical relative accuracy percentages of the individual 
component monitors have proven to be so low (i.e., average relative 
accuracy less than 5.0 percent for the period from the first quarter of 
1995 through the second quarter of 1996), the Agency believes that it 
is not necessary to promulgate the combined standard (see Docket A-97-
35, Item II-I-27). Data analysis from an EPA study (see Docket A-97-35, 
Item II-I-14) indicates that quality assuring the individual component 
monitors to 7.5 percent relative accuracy (the RA value needed to 
qualify for an annual RATA frequency) effectively ensures that a 
combined flow rate-SO2 standard of 10.0 to 15.0 percent 
relative accuracy will be consistently achieved. That same study also 
indicates that meeting a combined flow rate-SO2 standard of 
10.0 percent does not necessarily ensure that the individual component 
monitor relative accuracies will be  10.0 percent. In view 
of this and given that flow monitors are also used to calculate heat 
input and CO2 mass emissions, the Agency believes it is 
appropriate to maintain individual relative accuracy standards for the 
flow monitor and SO2 monitor. EPA solicits comment on its 
proposed treatment of this issue.
6. SO2 RATA Exemptions and Reduced Requirements
    Today's proposed rulemaking would clarify the RATA requirements for 
units that burn principally natural gas and other very low-sulfur 
fuels. In Sec. 75.21(a)(6) of the November 20, 1996 rule, an exemption 
from SO2 RATA requirements was provided for units that have 
SO2 monitors and exclusively burn natural gas (or fuels with 
a sulfur content no greater than natural gas). Today's proposed rule 
would clarify this exemption from SO2 RATAs by interpreting 
the term ``fuel with a total sulfur content no greater than the total 
sulfur content of natural gas'' to mean any type of fuel that has a 
total sulfur content of less than or equal to 0.05 percent sulfur by 
weight. The rationale for this is as follows. In order to meet the 
definition of natural gas in Sec. 72.2, the total sulfur content of the 
gas cannot exceed 20 grains/100 scf. When this sulfur content is 
converted to a weight percentage, it comes out slightly higher than 
0.05 percent sulfur by weight (see Docket A-97-35, Item II-B-14). 
Consequently, for a unit that has an SO2 monitor and for 
which the designated representative certifies that the unit burns only 
fuels (whether solid, liquid, or gaseous) with a total sulfur content 
of > 0.05 percent sulfur by weight, the SO2 monitor would be 
exempted from the part 75 RATA requirements. The Agency takes comment 
on this approach and on whether 0.05 percent sulfur by weight is an 
appropriate applicability threshold for fuels other than natural gas.
    Finally, Sec. 75.21(a)(7) of today's rule proposes reduced RATA 
requirements for units with SO2 monitors for which the 
designated representative certifies that the units burn fuel(s) with a 
total sulfur content greater than the total sulfur content of natural 
gas (e.g., distillate oil) only as emergency backup fuel(s) and/or for 
short-term testing. For such units, RATA testing of the SO2 
monitor would only be required if fuel with a total sulfur content 
greater than the total sulfur content of natural gas (i.e., > 0.05 
percent sulfur by weight) is combusted for more than 480 hours in a 
calendar year. If the higher-sulfur fuel usage were to exceed 480 hours 
in a particular year, then an SO2 RATA, conducted while 
burning the higher-sulfur fuel, would be required either by the end of 
the quarter in which the exceedance occurred or within a 720 unit 
operating hour grace period following that calendar quarter. In this 
instance, if the grace period were used, proposed section 2.3.3 in 
Appendix B would specify that it would begin with the first unit 
operating hour in which the higher-sulfur fuel is combusted in the 
unit, following the calendar quarter in which the annual usage of the 
higher-sulfur fuel exceeded 480 hours. The 480-hour criterion for 
maintaining an SO2 RATA exemption is consistent with many 
state and local air permits which contain a similar exemption from 
particulate emission testing for gas-fired units that burn oil for only 
400 to 500 hours per year (see Docket A-97-35, Item II-E-23). EPA 
believes that these provisions would effectively eliminate the need to 
start up a unit and/or to burn an infrequently used, uneconomical, and 
higher-emitting fuel solely for the purpose of performing a RATA of the 
SO2 monitor.
7. QA Provisions for SO2 Monitors, for Natural Gas Firing or 
Equivalent
    In Sec. 75.11(e) of the November 20, 1996 revisions to part 75, 
three SO2 compliance options were promulgated for units with 
SO2 CEMS during hours in which only natural gas (or gaseous 
fuel with a total sulfur content no greater than the total sulfur 
content of natural gas) is burned. One of the compliance options was to 
allow the use of an SO2 monitoring system, subject to

[[Page 28069]]

certain restrictions and quality assurance provisions. The restrictions 
and QA provisions, which are found at Secs. 75.11(e)(3)(i) through 
(iv), are as follows: (i) a calibration gas with a concentration of 0.0 
percent of span must be used for daily calibration error tests of the 
CEMS; (ii) the response of the monitoring system to the 0.0 percent 
calibration gas must be adjusted to read exactly 0.0 ppm each time that 
a daily calibration error test is passed; (iii) any hourly average of 
less than 2.0 ppm recorded by the SO2 monitor while fuel is 
being combusted in the unit(s) (including zero and negative averages) 
must be reported as a default value of 2.0 ppm; and (iv) if a unit 
combusts only natural gas (or gaseous fuel with a total sulfur content 
no greater than the total sulfur content of natural gas) and never 
combusts any other type of fuel, the SO2 monitor span must 
be set to a value not exceeding 200.0 ppm. Compliance with conditions 
(i) through (iv) is required by January 1, 1999, except that conditions 
(i) and (ii) are always optional for units that combust natural gas 
only during unit startup.
    The provisions in Secs. 75.11(e)(3)(i) through (iv), as presently 
codified, apply only to the combustion of gaseous fuel with a total 
sulfur content no greater than the total sulfur content of natural gas. 
However, as noted above (under ``SO2 RATA Exemptions and 
Reduced Requirements''), today's proposed rulemaking would add an 
interpretation of the term ``fuel with a total sulfur content no 
greater than the total sulfur content of natural gas'' to 
Sec. 75.21(a)(6). The term would include any fuel (whether solid, 
liquid, or gaseous) with a total sulfur content of  0.05 
percent by weight. EPA believes that it is appropriate to apply the 
quality assurance and reporting provisions in Secs. 75.11(e)(3)(i) 
through (iv) to the combustion of all fuels with a total sulfur content 
 0.05 percent by weight. Therefore, in today's proposed 
rule, a new section, Sec. 75.21(a)(8) would be added, extending the QA 
provisions of Secs. 75.11(e)(3)(i) through (iv) to the combustion of 
all types of fuels with a total sulfur content no greater than the 
total sulfur content of natural gas. The new requirements would become 
effective on January 1, 2000.
    Note that EPA has reconsidered one of the four QA provisions for 
the use of an SO2 monitor during natural gas (or fuel with 
equivalent total sulfur content) combustion in Secs. 75.11(e)(3)(i) 
through (iv). Specifically, the Agency believes that 
Sec. 75.11(e)(3)(ii), which requires a daily adjustment of the 
monitor's calibration to read exactly 0.0 ppm, may be too stringent 
because in practice it can be very difficult to attain a reading of 
exactly 0.0 ppm with a zero-level calibration gas, particularly when 
manual calibration adjustments are made. Therefore, today's rulemaking 
proposes to revise Sec. 75.11(e)(3)(ii) as follows. Rather than 
requiring a daily adjustment of the SO2 monitor's 
calibration, an adjustment would only be required when the ``as-found'' 
response of the monitor to the zero gas during a daily calibration 
error test exceeded the performance specification of the instrument 
(i.e., 2.5 percent of span). And instead of requiring the 
calibration to be adjusted to exactly 0.0 ppm, the procedures for 
routine calibration adjustments in proposed section 2.1.3 of Appendix B 
would be followed, to bring the ``as-left'' response of the instrument 
(i.e., the response during the additional calibration error test 
required by proposed section 2.1.3 of Appendix B) ``as close as 
practicable'' to the true value of the zero gas (0.0 ppm).
    The Agency solicits comment on the proposed approach for QA 
provisions for SO2 CEMS for gas-firing or equivalent.
8. General RATA Test Procedures
    Under today's proposal, sections 6.5, 6.5.1, and 6.5.2 of Appendix 
A, which describe the general requirements for RATAs, would be 
extensively revised. Some of the proposed changes are simply 
structural, but others are substantive. For instance, as previously 
discussed above under ``Concurrent SO2 and Flow RATAs,'' the 
requirement to perform concurrent SO2 and flow RATAs would 
be deleted from the regulation. Further, section 6.5 would now 
recognize that more than one type of fuel and more than one monitor 
range may be considered normal for a particular unit. Also, the 
requirement to complete each RATA within 7 consecutive calendar days 
would be modified to require that the RATA be completed within 168 unit 
operating hours (for single-load flow RATAs and, to the extent 
practicable, for 2-load and 3-load flow RATAs). However, for the 
multiple-load flow RATAs, up to 720 unit operating hours would be 
allowed, if necessary, to complete the testing. This is consistent with 
Agency guidance published in March, 1995, Policy Question 8.15 of the 
Acid Rain Policy Manual, which discusses allowing up to 30 calendar 
days to complete all three levels of a 3-load flow RATA (see Docket A-
97-35, Item II-I-9). Even though the policy says the RATAs at the 
individual load levels should be completed within 7 days, thirty days 
are acceptable to complete the 3-load RATA in order to account for the 
possibility that the unit might shut down in between levels of the RATA 
or that certain load levels may be difficult to attain and to hold. 
Today's proposal would allow 720 unit operating hours (irrespective of 
the number of calendar days) to complete a multiple-load flow RATA. EPA 
believes that this proposed requirement provides greater flexibility 
than currently allowed.
    Sections 6.5.1 and 6.5.2 of Appendix A would be re-titled ``Gas 
Monitoring Systems (Special Considerations)'' and ``Flow Monitor RATAs 
(Special Considerations),'' respectively. Proposed section 6.5.1 
contains a recommendation that, for initial monitor certifications, the 
RATA not be commenced until all of the other certification tests have 
been completed. Section 6.5.2 would be amended, as previously discussed 
under ``Flow RATA Load Levels.'' The definition of normal load would be 
revised and the number of loads and the load levels at which flow RATAs 
are to be performed would be more clearly defined.
    Today's rule proposes changes to section 6.5.6 of Appendix A, which 
pertains to RATA traverse point selection. Proposed section 6.5.6 would 
allow the following alternative reference method measurement point 
locations. For all moisture determinations, a single reference method 
point, located at least 1.0 meter from the stack wall, could be used. 
For gas RATAs, the owner or operator would have four options: (1) at 
any location (including locations where stratification is expected), a 
minimum of six traverse points along a diameter, located in accordance 
with Method 1 in Appendix A to part 60, could be used; (2) at locations 
where stratification is not expected and section 3.2 of Performance 
Specification No. 2 (``PS No. 2'') in Appendix B to part 60 allows the 
use of a short reference method measurement line (with three points 
located at 0.4, 1.0, and 2.0 meters from the stack wall), the owner or 
operator could use an alternative 3-point measurement line, locating 
the three points 4.4 percent, 14.6 percent and 29.6 percent of the way 
across the stack, in accordance with Method 1 in Appendix A to part 60; 
(3) at locations where stratification is expected (i.e., after a wet 
scrubber or when dissimilar gas streams are combined), the short 
measurement line from section 3.2 of PS No. 2 (or the alternative line 
described in option (2) above) could be used in lieu of the ``long'' 
measurement line prescribed in section 3.2 of PS No. 2, provided that a 
stratification test is performed prior to each RATA at the location and 
certain acceptance criteria

[[Page 28070]]

are met; and (4) a single reference method measurement point, located 
no less than 1.0 meter from the stack wall, could be used at any test 
location if a stratification test is performed prior to each RATA at 
the location and certain acceptance criteria are met. EPA's Office of 
Air Quality Planning and Standards (OAQPS) has endorsed the use of the 
Method 1 traverse points as an alternative to the points prescribed by 
PS No. 2 (see Docket A-97-35, Item II-C-22).
    Regarding option (3) above, one utility and one stack testing firm 
have requested that EPA allow the short measurement line to be used at 
scrubbed unit stacks, citing logistical difficulties and safety 
concerns associated with using the long measurement line prescribed by 
PS No. 2 for sampling locations following wet scrubbers (see Docket A-
97-35, Items II-D-66, II-D-78). Both parties appeared willing to 
perform stratification testing to demonstrate that the gas streams are 
not significantly stratified. EPA responded to these requests by 
issuing policy guidance which discusses allowing the short measurement 
line to be used for scrubbed units, provided that stratification test 
results show the stratification at the sampling location to be minimal 
(see Docket A-97-35, Item II-I-9, Policy Manual, Question 8.25). 
Regarding single-point RATA testing (option (4), above), which utility 
groups asked EPA to consider, today's proposed rule would allow it on 
the condition that a stratification test at the sampling location 
demonstrates stratification to be essentially absent.
    Sections 6.5.6.1 and 6.5.6.2 of Appendix A in today's proposed rule 
provide two stratification test protocols which may be used to 
demonstrate that a sampling location qualifies for the alternative RM 
measurement point locations allowed under proposed section 6.5.6 (i.e., 
options (3) and (4), above). The first stratification test protocol, in 
proposed section 6.5.6.1, is based upon technical guidance issued by 
OAQPS (see Docket A-97-35, Item II-I-3) and would consist of measuring 
the SO2, NO_X, and diluent gas concentrations at a 
minimum of 12 traverse points, located in accordance with Method 1 in 
Appendix A to part 60. The gas concentration measurements would be made 
using Reference Methods 6C, 7E, and 3A in Appendix A to part 60. The 
average NO_X, SO2, and CO2 (or 
O2) concentration at each of the individual traverse points 
would be determined, and the arithmetic average NO_X, 
SO2, and CO2 (or O2) concentrations 
for all traverse points calculated. This 12-point test would have to be 
passed one time at the sampling location under consideration. Once the 
12-point test has been passed at the candidate sampling location, the 
second (abbreviated) stratification test protocol, in proposed section 
6.5.6.2, could be done prior to subsequent RATAs at the location in 
lieu of the 12-point test. The abbreviated test would be done either at 
3 points (located in accordance with the long measurement line in PS 
No. 2) or at 6 points along a diameter (located according to EPA Method 
1 in Appendix A to part 60).
    The acceptance criteria for the stratification test results are 
given in proposed section 6.5.6.3 of Appendix A. For each pollutant or 
diluent gas, the short 3-point reference method measurement line 
specified in section 3.2 of PS No. 2 (or the alternative 3-point line 
described in proposed section 6.5.6 of Appendix A) could be used for 
that pollutant or diluent gas in lieu of the long measurement line in 
section 3.2 of PS No. 2, if the concentration at each individual 
traverse point differed by no more than 10.0 percent from 
the arithmetic average concentration for all traverse points. The 
results would also be acceptable if the concentration at each 
individual traverse point differed by no more than 5.0 ppm 
or 0.5 percent CO2 (or O2) from the arithmetic 
average concentration for all traverse points. Further, for each 
pollutant or diluent gas, a single reference method measurement point 
located at least 1.0 meter from the stack wall could be used for that 
pollutant or diluent gas, if the concentration at each individual 
traverse point differed by no more than 5.0 percent from 
the arithmetic average concentration for all traverse points. The 
results would also be acceptable if the concentration at each 
individual traverse point differed by no more than 3.0 ppm 
or 0.3 percent CO2 (or O2) from the arithmetic 
average concentration for all traverse points. Finally, proposed 
section 6.5.6.3 would require the owner or operator to keep the results 
of all stratification tests on-site, suitable for inspection, as part 
of the supplementary RATA records required under Sec. 75.56(a)(7) and 
Sec. 75.59(a)(7).
    Today's rule also proposes to clarify the sampling strategy for 
RATAs in section 6.5.7 of Appendix A. The proposed revisions make it 
clear that for gas monitor RATAs, the minimum time per run is 21 
minutes, and all of the necessary data for each run (i.e., pollutant 
concentration measurements and, if applicable, diluent concentration 
data and moisture measurements) would have to be collected, to the 
extent practicable, within a 60-minute period. The proposed revisions 
would also require the pollutant and diluent concentration measurements 
to be made simultaneously during RATAs of SO2/diluent and 
NO_X/diluent monitoring systems. For flow monitor RATAs, the 
minimum time per run would be 5 minutes. A requirement to properly 
account for flow pulsations (e.g., by sight-weighted averaging) at each 
velocity traverse point would be added, as well as a clear statement 
that successive flow RATA runs may be done as rapidly as practicable, 
with no required waiting period between runs. Proposed section 6.5.7 of 
Appendix A states that a minimum of one set of auxiliary data (moisture 
and diluent gas measurements) would have to be collected for every 
three RATA runs or for every clock hour of a flow RATA (whichever is 
less restrictive). A related change to Sec. 75.22(a)(4) is also 
proposed, which would allow the alternative moisture measurement 
techniques described in section 1.2 of Method 4 in Appendix A to part 
60 to be used for stack gas molecular weight determinations.
9. Reference Method Testing Issues
Discussion of Proposed Changes
    Currently, Sec. 75.22 specifies several reference methods 
(Reference Methods 2, 2A, 2C, or 2D) as appropriate methods for 
determining volumetric flow under part 75. The Agency is currently 
conducting a study of the accuracy of Reference Method 2 to determine 
whether changes to Method 2 or the addition of other alternatives to 
the Method are appropriate. Thus, the Agency anticipates that, in the 
future, revisions to Method 2 in part 60 may create alternatives beyond 
the specific reference methods specified in Sec. 75.22(a)(2). 
Therefore, in Sec. 75.22(a)(2), EPA proposes to add: ``or its allowable 
alternatives, except for 2B and 2E'' to Method 2 to automatically 
incorporate into part 75 anticipated future revisions to the Method 2 
requirements in Appendix A to part 60.
    Section 75.22 specifies a number of instrumental reference methods 
from Appendix A to part 60 (Reference Methods 3A, 6C, 7E, and 20) as 
appropriate test methods for conducting CEMS performance tests under 
part 75. These methods require the use of calibration gases to 
calibrate the reference analyzers. Currently, however, part 60 does not 
require that EPA protocol gas be used when performing instrumental 
reference methods. The Agency believes that protocol gas should be used 
when performing instrumental reference methods in order

[[Page 28071]]

to achieve accurate results. Therefore, proposed Sec. 75.22(c)(1) would 
state that, for purposes of part 75, instrumental reference methods 
must be performed using calibration gases as defined in section 5 of 
Appendix A to part 75.
10. Alternative Relative Accuracy Specifications and Specifications for 
Low-Emitters
    One utility group has suggested to EPA (see Docket A-97-35, Item 
II-E-13) that there is inconsistency and apparent inequity in the 
relative accuracy specifications for units that qualify as low emitters 
of NO_X and SO2 (i.e., sources with average 
SO2 concentrations of 250.0 ppm or less and/or average 
NO_X emission rates of 0.20 lb/mmBtu or less). Specifically, 
they have questioned the appropriateness of the alternative relative 
accuracy specifications used to determine the RATA frequency (i.e., 
semiannual or annual). Under section 3.3 of Appendix A and section 
2.3.1 of Appendix B to the current part 75 rule, the RATA frequency for 
an SO2 monitor installed on a low-emitting SO2 
source may be determined in either of two ways: by the normal relative 
accuracy specification (i.e. the RATA frequency is semiannual if the 
relative accuracy is > 7.5 percent but  10.0 percent, and 
annual if  7.5 percent relative accuracy is achieved), or by 
the alternative specification (i.e., the RATA frequency is semiannual 
if the reference method mean value and CEMS mean value differ by > 8.0 
ppm but  15.0 ppm, and annual if the two mean values differ 
by  8.0 ppm). For low-emitting NO_X sources, the 
RATA frequency for the NO_X monitoring system is determined 
in the identical manner to SO2 when the normal specification 
is applied. For the alternative specification, the NO_X RATA 
frequency is semiannual if the CEMS and reference method mean values 
differ by  0.01 lb/mmBtu but  0.02 lb/mmBtu, and 
annual if the mean values differ by > 0.01 lb/mmBtu. The 8.0 ppm value 
for SO2 was originally determined based on the performance 
of a single set of monitors at a facility regulated under subpart Da of 
the NSPS in part 60. However, in the first few years of Acid Rain 
Program implementation, many part 75 utilities with wet scrubbers have 
found it difficult to consistently meet the 8.0 ppm criterion for 
obtaining an annual RATA frequency.
    The utility group maintains that since, when the normal relative 
accuracy (RA) specification is applied, the criterion for obtaining an 
annual RATA frequency is to achieve a relative accuracy 25.0 percent 
below the RA specification in section 3.3 of Appendix A (i.e., 7.5 
percent RA is 25.0 percent below the specification of 10.0 percent), 
the criterion for an annual RATA frequency should be essentially the 
same when the alternative specification is applied. Under the current 
rule, the alternative SO2 specification requires that the 
mean CEMS and reference method values differ by no more than 8.0 ppm in 
order to obtain an annual RATA frequency. This is 47.0 percent below 
the 15.0 ppm alternative RA specification. Similarly for 
NO_X, the alternative NO_X specification for an 
annual RATA frequency requires the difference between the CEMS and 
reference method mean values to be  0.01 lb/mmBtu, or 50.0 
percent below the 0.02 lb/mmBtu alternative RA specification.
    EPA agrees that the alternate RA specifications for low emitters of 
SO2 and NO_X appear to be somewhat inequitable, 
and today's rulemaking proposes changes to these specifications. In 
proposed section 2.3.1 of Appendix B, the alternative relative accuracy 
specification for low emitters of SO2, (i.e., the difference 
between the reference method and CEMS mean values) that must be met by 
an SO2 monitor in order to obtain an annual RATA frequency 
would be changed from 8.0 ppm to 12.0 ppm. For low emitters of 
NO_X, the alternative low emitter relative accuracy 
specification that must be met by a NO_X-diluent monitoring 
system in order to obtain an annual RATA frequency would be changed 
from 0.01 lb/mmBtu to 0.015 lb/mmBtu.
    In today's rule, EPA is also proposing an alternative relative 
accuracy specification of 0.025 lb/mmBtu for SO2-diluent 
monitoring systems to obtain an annual RATA frequency and an 
alternative relative accuracy specification of 0.7 percent 
CO2 or O2, by which CO2 and 
O2 monitors could obtain an annual RATA frequency. During 
the investigation of the alternative RA specifications for the 
SO2 and NO_X-diluent monitoring systems, the 
Agency noted that for SO2-diluent systems, part 75 specifies 
only an alternative RA criterion of 0.030 lb/mmBtu for a semiannual 
RATA frequency, but fails to specify a corresponding alternative RA 
criterion for obtaining an annual RATA frequency. Similarly, for 
CO2 and O2 monitors, EPA noted that an 
alternative relative accuracy specification of 1.0 percent 
CO2 or O2 (in terms of the mean difference 
between the reference method and CEM values during the RATA) is given 
for obtaining a semiannual RATA frequency, but no corresponding 
alternative criterion is given for obtaining an annual frequency.
    EPA notes that in order to make the annual RATA frequency criteria 
for NO_X-diluent and SO2-diluent monitoring 
systems more equitable, a third decimal place is required. However, 
Secs. 75.54 and 75.55 currently require NO_X and 
SO2 emission rates in lb/mmBtu to be reported only to 2 
decimal places. Therefore, revisions are being proposed, see 
Secs. 75.57(d)(6) and 75.58(a)(1)(iv), to require that, beginning on 
January 1, 2000, all NO_X emission rates in lb/mmBtu must be 
reported to three decimal places. Prior to January 1, 2000, the owner 
or operator would have the option of reporting NO_X emission 
rates to either two or three decimal places. Note that no corresponding 
change is being proposed for the reporting of SO2 emission 
rates in lb/mmBtu, since such emission rates will only be reported to 
EPA by units that have installed Phase I Qualifying Technologies for a 
three-year period (1997-1999), and are not required to be reported 
thereafter. EPA solicits comments on the appropriateness of requiring 
all NO_X lb/mmBtu emission rates to be reported to three 
decimal places. The Agency favors this approach, particularly for 
quality assurance purposes, due to increased precision in the 
calculation of RATA results. The Agency notes that this proposed change 
would not affect the way in which compliance with the NO_X 
emission limits under part 76 is determined. Compliance with part 76 
NO_X limits, in lb/mmBtu, would still be based on two decimal 
places.
    All of the proposed revisions to the part 75 relative accuracy 
specifications in today's rulemaking are summarized in proposed Figure 
2 of Appendix B.
    11. Bias Adjustment Factors for Low Emitters
    As discussed in the preceding section, sources that qualify as low 
emitters of SO2 and/or NO_X have two ways to 
evaluate the relative accuracy of SO2 and NO_X 
monitoring systems: (a) by the normal relative accuracy specification 
(i.e., 10.0 percent RA), and (b) by the alternative RA specification 
(i.e., the difference between the mean CEMS and reference method values 
is within 15.0 ppm for SO2 low emitters, or 
within 0.02 lb/mmBtu for NO_X low emitters).
    The normal RA is determined by a statistical analysis of the 
reference method and CEMS data from the RATA. Mathematically, the 
normal RA is the sum of the absolute values of the mean difference 
(dmean) and the confidence coefficient (cc), expressed as a 
percentage of the mean reference method value (RM)avg. The 
mean difference indicates how closely the CEMS measurements agree with 
the

[[Page 28072]]

reference method and is generally the principal contributor to the 
percentage relative accuracy in the RA equation. The confidence 
coefficient (cc) is a statistical term related to the standard 
deviation and is an indicator of the amount of scatter in the data.
    Section 7.6 of Appendix A requires a bias test of each 
SO2 and NO_X monitoring system whenever a RATA of 
the CEMS is performed. If the mean difference is greater than the 
absolute value of the confidence coefficient, the CEMS measurements are 
systematically lower than the corresponding references method 
measurements, i.e., the monitoring system has a low bias. In such 
cases, sources are given two options. The first, preferred by EPA, is 
to locate and eliminate the source of the measurement bias in the 
instrument. The second option is to apply a bias adjustment factor 
(BAF). This alternative was developed in response to an industry 
request to provide an alternative for sources that choose not to expend 
the effort to locate and eliminate the technical problem causing the 
systematic measurement error. The BAF is equal to 1.000 + 
|dmean| /(CEM)avg, where (CEM)avg is 
the mean value of the CEMS measurements from the RATA.
    At least one utility has questioned whether it is appropriate for 
low emitters to calculate a BAF in the usual way when a CEMS fails a 
RATA by the normal RA specification, but passes by the alternative 
specification, because in such cases the BAF can become inordinately 
high, particularly at very low emission levels (see Docket A-97-35, 
Items II-D-62 and II-E-23). Since both the percent relative accuracy 
and the BAF are based upon the same statistical terms (dmean 
and cc), the utility questions whether the standard calculation 
procedure for the BAF is adequate to determine a meaningful BAF for low 
emitters. Just as the value obtained from the standard relative 
accuracy equation tends to become large for low emitters, so, too, the 
BAF is seen as becoming inordinately large for low emitters which use 
the current BAF equation.
    As this comment suggests, it is not uncommon for an SO2 
or NO_X CEMS installed on a low-emitting unit to fail a RATA 
by the normal specification of 10.0 percent RA and to pass the same 
RATA by the alternative RA specification. For instance, suppose that 
the mean RM and CEMS values during an SO2 RATA of a low 
emitter are 51.0 ppm and 40.0 ppm, respectively, and that 
dmean is 11.0 ppm and the confidence coefficient is 0.50. 
Suppose further that the bias test is failed. Then, the percent RA by 
the normal specification (i.e.,  |dmean| + |cc |  / 
(RM)avg) would exceed 20.0 percent, indicating a failed 
RATA, but the alternative RA specification would indicate a pass (i.e., 
(CEMS)avg is within 15.0 ppm of 
(RM)avg). In this same illustration, the BAF would be 1 + 11 
/ 40 = 1.275.
    In fact, if it is assumed that the difference between the CEMS and 
the reference method measurements does not decrease as emissions 
decline, then the lower the SO2 or NO_X emissions, 
the more likely it is for the CEMS to fail the normal relative accuracy 
specification because the mean difference becomes a larger percentage 
of the average reference method value. It was precisely in response to 
such concerns that the alternative relative accuracy specifications 
were originally included in part 75.
    Today's rule proposes to provide an option in the way the BAF is 
determined for low emitters of SO2 and NO_X. Low 
emitters of SO2 and NO_X would be given the choice 
of using either: (a) the normal BAF calculation procedure described 
above and found in Equation A-12, section 7.6.5 of Appendix A, or (b) 
an alternative default bias adjustment factor of 1.111.
    The justification is as follows: for units that meet the normal 
relative accuracy standard of RA   10.0 percent, the 
theoretically maximum possible Bias Adjustment Factor is 1.111 (see 
Docket A-97-35, Item II-B-2). Therefore, low-emitting units meeting the 
alternative relative accuracy standards (|dmean|  
15.0 ppm for SO2 low emitters and |dmean| 
 0.02 lb/mmBtu for NO_X low emitters) should not 
have to apply a bias adjustment any higher than the maximum BAF value 
applicable to units meeting the normal relative accuracy standard. EPA 
solicits comment on allowing the alternative BAF of 1.111 for low-
emitting units.
12. Clarification of Diluent Monitor Certification Requirements
    Today's proposed rule would clarify the certification requirements 
for diluent gas (CO2 and O2) monitors, in 
response to comments received on the pre-proposal draft of the rule 
(see Docket A-97-35, Item II-D-52). Section 75.20(c)(1)(iii) of the 
current rule requires a RATA of each NO_X continuous 
monitoring system to be done for initial certification. Even though the 
NO_X system consists of two component monitors 
(NO_X concentration and diluent gas), the required RATA is 
done on a system basis in units of lb/mmBtu. Separate RATAs of the 
individual component monitors are not required, except when the diluent 
component monitor is also used as a CO2 pollutant 
concentration monitor or to account for unit heat input, in which case 
Sec. 75.20(c)(5)(iii) in the current rule requires a RATA of the 
diluent monitor. To be sure that this is clear, today's proposed rule 
would add a statement to Sec. 75.20(c)(1)(iii), indicating that the 
RATA for the NO_X-diluent system shall be done on a system 
basis (i.e., individual component RATAs are unnecessary for 
certification of a NO_X-diluent system). Therefore, units 
that have installed NO_X monitoring systems, but that use 
Appendix D for SO2 emission accounting and Appendix G for 
CO2 accounting, would not be required to submit separate 
RATA results for the diluent monitor.
    A second point of clarification would be added in proposed 
Sec. 75.20(c)(3), which was previously designated as Sec. 75.20(c)(4). 
The new section would make it clear that when a diluent monitor 
(O2 or CO2) is used both as a CO2 
pollutant concentration monitor and for heat input determinations, only 
one set of diluent monitor certification test results would have to be 
submitted under the component and system ID codes of the CO2 
monitoring system. This is appropriate because there is no such thing 
as a ``heat input monitoring system'' or an ``oxygen monitoring 
system'' under part 75.
13. Daily Calibration Requirements for Redundant Backup Monitors
    Section 75.20(d)(1) of the current rule requires redundant backup 
(``hot-standby'') monitoring systems to be operated during all periods 
of unit operation and to meet all of the quality assurance requirements 
of Appendix B, including daily calibrations and interference checks, 
quarterly linearity checks and leak checks, and semiannual or annual 
RATAs. One commenter on a pre-proposal draft of today's proposed rule 
requested that EPA consider changing the daily calibration requirement 
for redundant backup monitors (see Docket A-97-35, Item II-D-35). The 
commenter recommended that the daily calibrations be made mandatory 
only for days on which the redundant backup monitoring system is 
actually used to report emission data to EPA. Daily calibrations would 
be optional on all other days. Fewer calibrations of redundant backup 
systems would considerably reduce calibration gas consumption. The 
commenter estimated that this change could result in an annual savings 
of more than $100,000 for his company. EPA agrees that the request is 
reasonable, provided that the redundant

[[Page 28073]]

backup systems are kept on hot-standby and are calibrated prior to each 
use for reporting. The Agency therefore proposes to amend 
Sec. 75.20(d)(1) accordingly.
14. Daily Performance Specification and Control Limits for Low-Span DP 
Flow Monitors
    Section 3.1 of Appendix A of the current rule gives the calibration 
error performance specification for flow monitors. Section 2.1.4 of 
Appendix B gives the calibration error limits for daily operation of 
flow monitors. For initial certification, a flow monitor is required to 
meet a calibration error specification of  3.0 percent of 
the span value. For daily operation of the flow monitor, the 
calibration error must not exceed 6.0 percent of span. These 
specifications are both reasonable and achievable for the vast majority 
of flow monitors. However, when a differential pressure (DP) type flow 
monitor is used to measure stack gas flow rate in a stack that has low 
exit velocities, it can be very difficult for the monitor to pass its 
daily calibration error tests. This is because the daily calibration 
span value for a DP flow monitor is expressed in units of inches of 
water. For stack exit velocities less than 2000 feet per minute, the 
calibration span value will be a very small number (0.20 inches of 
water or less). When performing a daily calibration error test of a 
flow monitor with a span value of 0.20 inches of water, the test would 
be failed (i.e., the calibration error would exceed 6.0 percent of 
span) if the response of the monitor deviated from either the zero or 
high reference signal by 0.02 inches of water. For span values of 0.15 
inches of water or less, the calibration error test would be failed if 
the monitor's response deviated from the reference signals by 0.01 
inches of water. One utility with a DP type flow monitor with a span 
value less than 0.15 inches of water has indicated to EPA that it 
cannot pass daily calibrations unless the monitor responses exactly 
equal the reference signal values (see Docket A-97-35, Item II-E-30). 
Clearly, these daily calibration specifications are too stringent for 
low span DP-type flow monitors. In view of this, EPA is proposing 
alternative calibration error specifications for DP type flow monitors 
with low span values, with ``low'' span value meaning a span value of 
0.20 inches of water or less. The alternative performance specification 
for initial certification, given in proposed section 3.1 of Appendix A, 
would be  0.01 inches of water, rather than  
3.0 percent of span. The alternative specification for daily operation 
of the monitor, given in proposed section 2.1.4 of Appendix B, would be 
 0.02 inches of water, rather than  6.0 percent 
of span. Since the results of a calibration error test of a DP type 
flow monitor are reported to 2 decimal places, the performance 
specification of  0.01 inches of water, is the tightest 
specification that could be imposed, short of requiring the monitor to 
read exactly the reference value with zero tolerance (which is what the 
current specification of  3.0 percent of span essentially 
imposes on a DP flow monitor with very low span). The Agency solicits 
comment on this proposed approach and on the value of the alternate 
specification.

O. CEM Data Validation

Background
    The current requirements of part 75 regarding CEM data validation 
are as follows. Section 75.10 specifies that a valid hourly average 
from a CEMS must be based on a minimum of four evenly spaced data 
points (i.e., one point in each 15-minute quadrant of the clock hour), 
except that two evenly spaced data points separated by at least 15 
minutes are sufficient to validate an hourly average when daily 
calibration error tests and/or other required quality assurance 
activities are conducted during the hour. Data from a CEMS are 
considered to be quality assured, provided that the monitoring system 
has passed all of the initial certification tests required under 
Sec. 75.20(c) and provided that the CEMS is not ``out-of-control,'' as 
a result of having failed any of the daily, quarterly, semiannual, and/
or annual quality assurance tests required in sections 2.1 through 2.3 
of Appendix B. Out-of-control periods extend from the hour of failure 
of a QA test until the hour of completion of a subsequent successful QA 
test of the same type. For instance, if a linearity check of a gas 
monitor is failed, the monitor is considered out-of-control from the 
hour of completion of the failed test until the hour of completion of a 
subsequent successful linearity test.
    Finally, Sec. 75.20(b)(3) specifies that when a change is made to a 
CEMS such that recertification of a monitor becomes necessary, data 
from the CEMS are invalid from the hour in which the change is made to 
the system until the hour of completion of all required recertification 
tests.
    In the first three years of implementing part 75, EPA has received 
numerous requests from the utilities for guidance concerning CEM data 
validation. This has prompted the Agency to re-examine these provisions 
of the rule. From this re-examination, the Agency believes that the 
current data validation provisions of part 75 are neither sufficiently 
detailed nor flexible to address the complex realities of daily 
operation of utility boilers and continuous emission monitoring 
systems. Therefore, today's proposed rule would set forth more 
comprehensive data validation criteria.
Discussion of Proposed Changes
    Today's proposed rule would set forth proposed guidelines for the 
validation of CEM data, attempting to take into account the realities 
associated with the operation and maintenance of electric utility steam 
generating units and continuous emission monitoring systems. The 
proposed guidelines would govern CEM data validation as it pertains to 
six principal areas: (1) calibration error tests and adjustment of gas 
and flow monitors; (2) linearity tests of gas monitors; (3) relative 
accuracy test audits of gas and flow monitoring systems; (4) 
recertifications of gas or flow monitors; (5) data from non-redundant 
backup monitoring systems; and (6) missed QA test deadlines. These 
proposed guidelines for data validation are discussed in detail below.
1. Recalibration and Adjustment of CEMS
    Today's proposed rule would revise section 2.1.3 of Appendix B, the 
``recalibration'' section. The May 17, 1995 rule recommends (but does 
not require) the calibration of a monitor to be adjusted whenever the 
daily calibration error exceeds the performance specification in 
Appendix A. For example, if the calibration error of a gas monitor 
exceeds 2.5 percent of span, but does not exceed the daily control 
limit of 5.0 percent of span, the monitor is considered to be out-of-
adjustment but not out-of-control, and EPA recommends that calibration 
of the monitor be adjusted.
    Today's proposal would re-title section 2.1.3 as ``Additional 
Calibration Error Tests and Calibration Adjustments.'' The 
recommendation to adjust the monitor when the calibration error exceeds 
the Appendix A performance specification would be retained, but 
definitions of ``routine calibration adjustments'' and ``non-routine 
calibration adjustments'' would be added. Routine calibration 
adjustments would be defined as adjustments made to a CEMS following a 
successful calibration error test. The purpose of these adjustments 
would be to bring the monitor readings as close as practicable to the 
tag values of the reference calibration gases or to the

[[Page 28074]]

known values of the flow monitor reference signals. Non-routine 
calibration adjustments would be adjustments in either direction 
(toward or away from the reference value), but within the performance 
specifications of the monitor (i.e., within  2.5 percent of 
span for an SO2 or NO_X monitor,  0.5 
percent CO2 or O2 for a diluent monitor, or 
 3.0 percent of span for a flow monitor). Non-routine 
calibration adjustments would be permitted, provided that an acceptable 
technical justification is included in the QA/QC program required under 
section 1 of Appendix B. An additional calibration error test would be 
required following non-routine adjustments, to demonstrate that the 
instrument is still operating within its performance specifications.
    In addition to the daily calibration error requirements in section 
2.1.1 of Appendix B, today's proposed rule would require a calibration 
error test in four specific instances: (1) whenever a daily calibration 
error test is failed; (2) when a CEMS is returned to service following 
routine or corrective maintenance that may affect the ability of the 
CEMS to accurately measure and record emissions data; (3) following 
routine calibration adjustments in which the monitor's calibration is 
physically adjusted, e.g., by means of a potentiometer (however, an 
additional calibration error test would not be required if a 
mathematical algorithm in the DAHS is used to make the routine 
adjustments); and (4) following non-routine calibration adjustments. 
Data from the CEMS would be considered invalid until the required 
additional calibration error test had been successfully completed.
    EPA is proposing to allow non-routine calibration adjustments 
within the performance specifications of an instrument for two 
principal reasons. First, commenters have expressed concern that 
restricting allowable adjustments to routine calibration adjustments 
would limit their ability to make adjustments within the acceptable 
plus or minus control limits of a monitor, particularly prior to 
linearity tests and RATAs. They have indicated that this flexibility is 
necessary because the tag values of reference gases are not 100.0 
percent accurate and adjustments of the analyzer may be needed to 
account for these inaccuracies (see Docket A-97-35, Item II-I-15). EPA 
agrees that this is a legitimate concern. Because there is a tolerance 
of  2.0 percent on the different reference gases used for 
daily calibration error tests, linearity tests, and RATAs, it may be 
necessary to adjust toward or away from the tag value in order to make 
sure that the test specifications are met. The Agency believes, 
however, that it is appropriate to limit the calibration adjustments to 
within the instrument's performance specifications (i.e.,  
2.5 percent of span (for SO2 and NO_X), 
 3.0 percent of span (for flow rate), and  0.5 
percent CO2 or O2) in order to provide an on-
going demonstration that the CEMS can simultaneously comply with the 
applicable daily, quarterly, semiannual, or annual performance 
specifications in Appendix A. One utility has expressed concern about 
its vendor's practice of making large calibration adjustments to the 
CO2 monitor prior to RATA testing (see Docket A-97-35, Item 
II-D-63).
    The second reason for proposing to allow non-routine calibration 
adjustments is the sensitivity of dilution-extractive monitors to 
changes in barometric pressure, temperature, and molecular weight. EPA 
believes that the best way to deal with this deficiency in the 
dilution-extractive monitoring technology is to develop a mathematical 
algorithm (site-specific, if necessary) that continuously applies a 
correction to the measurement in order to compensate for pressure, 
temperature, and molecular weight, as necessary, and to program the 
algorithm into the DAHS. However, in commenting on a pre-proposal draft 
of today's proposed rule, a number of utilities indicated that they 
prefer to account for dilution probe pressure effects by manually 
adjusting the monitor's calibration in anticipation of barometric 
pressure changes (e.g., approaching weather fronts) (see Docket A-97-
35, Items II-D-41, II-D-55). After much deliberation, the Agency is 
proposing to allow such adjustments, provided that: (1) the calibration 
of the monitor is not adjusted outside of its performance 
specifications; (2) an additional calibration error test is done to 
verify that the adjustments have been properly made; and (3) the 
procedures used for the adjustments are included in the QA/QC program 
for the CEMS. Despite this, EPA still prefers that automatic pressure, 
temperature, and molecular weight compensation be used, where 
necessary, and would strongly encourage all facilities with dilution-
extractive monitors to develop and apply the necessary mathematical 
algorithm(s).
2. Linearity Tests
    Today's proposal would provide rules for data validation during 
linearity tests, in proposed section 2.2.3 of Appendix B. A routine 
quality assurance linearity test could not be commenced if the CEMS 
were operating ``out-of-control'' with respect to any of its other 
daily, semiannual, or annual quality assurance tests. Linearity tests 
would be done ``hands-off,'' as follows. Prior to the test, both 
routine and non-routine calibration adjustments, as defined in proposed 
section 2.1.3 of Appendix B, would be permitted. During the linearity 
test period, however, no adjustment of the monitor would be permitted 
except for routine daily calibration adjustments following successful 
daily calibration error tests (the Agency notes that it is unlikely for 
calibration error tests to be done during a linearity test period 
except when two or more operating days are required to complete the 
test, e.g., for a peaking unit).
    Proposed section 2.2.3 of Appendix B would specify that when a 
linearity check is failed or aborted due to a problem with the monitor, 
the monitor would be declared out-of-control as of the hour in which 
the test is failed or aborted. Data from the monitor would remain 
invalid until the hour of completion of a subsequent successful hands-
off linearity test. This proposed requirement is not substantially 
different from the out-of-control provision in the current rule. It 
would merely extend the definition of out-of-control to include 
linearity tests that are aborted prior to completion due to a problem 
with the monitor. The underlying assumption is that the aborted 
linearity test would not have been passed if all nine gas injections 
had been completed. However, a linearity test that is aborted for a 
reason unrelated to a monitor malfunction (e.g., an unplanned or forced 
unit outage) would not trigger an out-of-control period.
    Finally, a new section, 2.2.4, would be added to Appendix B, 
providing a linearity test grace period of 168 unit operating hours. 
The purpose of the grace period would be to give the owner or operator 
a window of opportunity in which to perform a linearity test, when 
either: (1) the required linearity test cannot be completed within the 
QA operating quarter in which it is due, or (2) four consecutive 
calendar quarters have elapsed since the end of the calendar quarter in 
which a linearity test of a monitor (or range) was last done. Data 
validation during a grace period would be done according to the 
applicable provisions of proposed section 2.2.3 of Appendix B. Proposed 
section 2.2.4 of Appendix B would specify that if the required 
linearity test has not been completed within the grace period, data 
from the monitor would become invalid, beginning with the first hour 
following the expiration of the grace period and would remain invalid 
until the hour of completion of a

[[Page 28075]]

subsequent successful, hands-off linearity test. Proposed section 2.2.4 
would further specify that a linearity test done during a grace period 
could only be used to meet the linearity test requirement of a previous 
QA operating quarter, not the requirement of the quarter in which the 
grace period is used. Note that proposed sections 2.2.3 and 2.2.4 of 
Appendix B would also extend the 168 unit operating hour grace period 
to apply to the quarterly leak checks of differential pressure-type 
flow monitors.
3. RATAs
    Today's proposal would provide rules for data validation during gas 
and flow monitor RATA tests, in section 2.3.2 of Appendix B. Proposed 
section 2.3.2 would specify that a routine quality assurance RATA could 
not be commenced if the monitoring system is out-of-control with 
respect to any of its daily quality assurance assessments, including 
the additional calibration error test requirements of proposed section 
2.1.3 of Appendix B. All RATAs would be done ``hands-off,'' as follows. 
Prior to the RATA , both routine and non-routine calibration 
adjustments would be permitted, in accordance with proposed section 
2.1.3 of Appendix B. During the RATA test period, however, only routine 
calibration adjustments (as defined in proposed section 2.1.3 of 
Appendix B) would be permitted. For 2-level and 3-level flow RATAs, no 
linearization of the monitor would be permitted between load levels.
    Note that EPA is proposing to allow pre-RATA adjustments and 
linearization of a CEMS, principally to encourage facilities to 
optimize the performance of their CEMS by achieving the best possible 
relative accuracy results in a cost-effective manner with little or no 
data loss. The Agency believes that there is no significant risk in 
allowing pre-RATA adjustments, provided that the monitor's continued 
accuracy between successive RATAs can be reasonably established. For 
gas monitors, EPA believes that the daily calibration error tests and 
quarterly linearity tests, which challenge the analyzers with protocol 
gases of known concentration, provide that assurance. For flow 
monitors, however, the daily calibration error tests, which check the 
internal electronics of the flow monitor but do not evaluate the actual 
flow measurement capability of the instrument, do not provide the 
necessary assurance. Therefore, in today's rulemaking, EPA is proposing 
a new flow monitor quality assurance requirement, the ``flow-to-load 
test,'' to provide a reasonable indicator of continued flow monitor 
accuracy between successive RATAs. The flow-to-load test has been 
discussed in detail under section III.M. of this preamble.
    If a RATA is failed or aborted due to a problem with the CEMS, 
proposed section 2.3.2 of Appendix B would specify that the monitoring 
system is out-of-control as of the hour in which the test is failed or 
aborted. Data from the monitoring system would remain invalid until the 
hour of completion of a subsequent successful hands-off RATA. This 
proposed requirement is essentially the same as the out-of-control 
provision in the current rule, except that it would extend the 
definition of out-of-control to include RATAs that are aborted prior to 
completion due to a problem with the CEMS. Note, however, that a RATA 
which is terminated for a reason unrelated to monitor malfunction 
(e.g., process operating problems or unit outage) would not trigger an 
out-of-control period.
    For multiple-load flow RATAs, each load level would be treated as a 
separate RATA. Therefore, if a flow RATA is failed at a particular load 
level, previously-passed RATAs at the other loads would not have to be 
repeated unless the flow monitor has to be re-linearized. In that case, 
a subsequent 3-load RATA would be required.
    If a daily calibration error test is failed during a RATA test 
period, proposed section 2.3.2 of Appendix B would require invalidation 
of the RATA, and an out-of-control period would begin with the hour of 
the failed calibration error test. The RATA could not to be re-started 
until a subsequent calibration error test had been passed, following 
corrective actions.
    Proposed section 2.3.2 of Appendix B further specifies that when 
the RATA of a CO2 pollutant concentration monitor (or an 
O2 monitor used to measure CO2 emissions) is 
failed and that same CO2 (or O2) monitor also 
serves as the diluent component in a NO_X-diluent (or 
SO2-diluent) monitoring system, then both the CO2 
(or O2) monitor and the associated NO_X-diluent (or 
SO2-diluent) system would be considered to be out-of-control 
until the hour of completion of subsequent hands-off RATAs which 
demonstrate that both systems are in-control and have met the 
applicable relative accuracy specifications in sections 3.3.2 and 3.3.3 
of Appendix A. The beginning of the out-of-control period for each 
monitoring system would be the hour of completion of the failed or 
aborted RATA of the CO2 (or O2) monitor. The 
lengths of the out-of-control periods would, therefore, be determined 
from the same reference point for both the CO2 (or O2) 
monitoring system and the NO_X-diluent (or SO2-
diluent) monitoring system.
    Today's proposal would clarify the way in which RATA results are to 
be reported to EPA in the electronic quarterly report required under 
Sec. 75.64. Proposed section 2.3.2 of Appendix B specifies that only 
the results of completed and partial RATAs that affect data validation 
would have to be reported. That is, all completed passed RATAs, all 
completed failed RATAs, and all RATAs aborted due to a problem with the 
CEMS would have to be included in the quarterly report. Therefore, 
aborted RATA attempts followed by corrective maintenance, re-
linearization of the monitor, or any other adjustments other than those 
allowed under proposed section 2.1.3 of Appendix B would have to be 
reported. RATAs which are aborted or invalidated due to problems with 
the reference method or due to operational problems with the affected 
unit(s) would not need to be reported, because such runs do not affect 
the validation status of emission data recorded by the CEMS. In 
addition, aborted RATA attempts which are part of the process of 
optimizing a monitoring system's performance would not have to be 
reported, provided that in the period from the end of the aborted test 
to the commencement of the next RATA attempt: (1) no corrective 
maintenance or re-linearization of the CEMS is performed, and (2) no 
adjustments other than the calibration adjustments allowed under 
proposed section 2.1.3 of Appendix B are made. However, such RATA runs 
would still have to be documented and kept on-site as part of the 
official test log.
    Whenever a required RATA has not been completed by its deadline, 
section 2.3.3 of Appendix B of today's proposed rulemaking would 
provide a grace period of 720 unit operating hours in which to complete 
the test. Data validation during a grace period would be done according 
to the applicable provisions of proposed section 2.3.2 of Appendix B. 
Proposed section 2.3.3 would specify that if the RATA is not completed 
by the end of the grace period, data from the CEMS would become invalid 
upon expiration of the grace period and remain invalid until the hour 
of completion of a subsequent successful hands-off RATA.
    EPA has proposed a 720 unit operating hour RATA grace period 
because the Agency believes this will allow the facility sufficient 
time to schedule the RATA, to provide all required test notifications, 
and to complete the testing. The proposed grace period would be based 
on unit

[[Page 28076]]

operating hours rather than clock hours, because this is believed to be 
more equitable for peaking and cycling units. Data validation during 
the grace period would be prospective, i.e., data from the monitoring 
system would be considered valid during the grace period until the time 
of the RATA. If the RATA is failed or aborted due to a problem with the 
CEMS, data would be invalidated from the hour in which the test is 
failed or aborted, forward. Data would not be invalidated 
retrospectively back to the beginning of the grace period. Several 
utilities have expressed a preference for a grace period with 
prospective data invalidation, because it is simple to implement and is 
consistent with other part 75 provisions for which data invalidation is 
prospective when a test is failed (see Docket A-97-35, Item II-E-23).
4. Recertification of Gas and Flow Monitors
    Today's proposed rule would revise Sec. 75.20(b)(3) concerning data 
validation during recertification test periods. In the January 11, 1993 
rule, as amended on May 17, 1995, Sec. 75.20(b)(3) specifies that for 
any replacement, change, or modification to a monitoring system 
requiring recertification of the CEMS, all data from the CEMS are 
considered invalid from the hour of that replacement, change, or 
modification until the hour of completion of all required 
recertification tests. Today's rulemaking proposes to conditionally 
allow emission data generated by the CEMS during a recertification test 
period to be used for part 75 reporting, provided that the required 
tests are successfully completed in a timely manner and that certain 
data validation rules are followed during the recertification test 
period. Proposed sections 6.2, 6.3.1, and 6.5 of Appendix A would allow 
these new data validation procedures to also be applied to the initial 
certification of monitoring systems. The proposed revisions are based, 
in part, on policy guidance issued by EPA to address the initial 
certification of CEMS when a wet scrubber is installed on an affected 
unit (see Docket A-97-35, Item II-I-9, Policy Manual, Question 16.10). 
The intent of that policy guidance and of today's proposal is to 
minimize the number of hours of substitute data or maximum potential 
values that must be reported during a monitor certification or 
recertification period.
    In proposed Sec. 75.20(b)(3), specific rules are provided for data 
validation during the recertification test period. The recertification 
test period would begin with the first successful calibration error 
test after making the change to the CEMS and completing all necessary 
post-change adjustments, re-programming, linearization, etc. of the 
CEMS. The post-change activities could also include preliminary tests 
such as trial RATA runs or a challenge of the monitor with calibration 
gases. The first successful calibration error test following all of 
these activities would be known as a probationary calibration error 
test. Data from the CEMS would be considered invalid from the hour in 
which the replacement, modification, or change to the system is 
commenced until the hour of completion of the probationary calibration 
error test, at which point, the data status would become conditionally 
valid.
    Today's proposal would place a specific time limit on the length of 
the recertification test period, depending upon the type(s) of test(s) 
required. If a linearity test or cycle time test is required, the test 
would have to be completed within 168 unit operating hours of the hour 
in which the probationary calibration error test was passed, marking 
the beginning of the recertification test period. If a RATA is 
required, it would have to be completed within 720 unit operating 
hours. If a 7-day calibration error test were required, it would have 
to be completed within 21 unit operating days. Routine daily 
calibration error tests would continue to be done as required by part 
75 throughout the recertification test period. If a particular 
recertification test is not completed within the specified number of 
hours, data validation would be done as follows. For a late linearity 
test, RATA, or cycle time test that is passed on the first attempt, or 
for a late 7-day calibration error test (whether or not it is passed on 
the first attempt), data from the monitoring system would be 
invalidated from the hour of expiration of the recertification test 
period until the hour of completion of the late test. However, for a 
late linearity test, RATA, or cycle time test that is failed on the 
first attempt or aborted on the first attempt due to a problem with the 
monitor, all conditionally valid data from the monitoring system would 
be invalidated from the hour of the probationary calibration error test 
that initiated the original recertification test period to the hour of 
completion of the late recertification test. Data would remain invalid 
until successful completion of the failed/aborted test and any 
additional recertification or diagnostic tests that are required as a 
result of changes made to the monitoring system to correct the 
problem(s) that caused failure of the late recertification test.
    A conditionally valid status would be assigned to emission data 
generated by a CEMS during a recertification test period. The 
conditionally valid data status would begin with the first hour of the 
recertification test period (i.e., the hour in which the probationary 
calibration error test is passed, following completion of all necessary 
monitor adjustments, preliminary tests, etc.). The conditionally valid 
status of the CEMS data would continue throughout the recertification 
test period, provided that the required recertification tests are done 
``hands-off'' (i.e., with no adjustments, reprogramming, etc. of the 
CEMS other than the calibration adjustments allowed under proposed 
section 2.1.3 of Appendix B) and provided that the recertification 
tests and required daily calibration error tests continue to be passed. 
If all of the required recertification tests and calibration error 
tests are passed hands-off, with no failures and within the required 
time period, then all of the conditionally valid emission data recorded 
by the CEMS during the recertification test period would be considered 
quality assured and suitable for part 75 reporting. Note, however, that 
if a required recertification test has not been completed by the end of 
a calendar quarter, the owner or operator would indicate this by using 
a suitable conditional data flag in the electronic quarterly report for 
that quarter. The owner or operator would be required to resubmit the 
report for that quarter if the required recertification test is 
subsequently failed. In the resubmitted report, the owner or operator 
would use the appropriate missing data routine in Sec. 75.31 or 
Sec. 75.33 to replace each hour of conditionally valid data that was 
invalidated by the failed recertification test with substitute data. In 
addition, if conditionally valid data is submitted to the Agency in any 
quarterly report, the owner or operator would have to indicate in the 
end of the year compliance report required under Sec. 72.90 whether the 
final status of the conditionally valid data has been determined. Note 
that in certain instances where a recertification test period spans two 
calendar quarters, it may be possible to avoid use of the conditional 
data flag and quarterly report resubmittal. If a required 
recertification test(s) is completed no later than 30 days after the 
end of a calendar quarter (i.e., prior to the quarterly report 
submittal deadline), the test data and results may be submitted

[[Page 28077]]

with the quarterly report, even though the test dates are from the next 
calendar quarter. If the recertification test(s) is passed, this would 
allow the ``conditionally valid'' data to be reported as quality 
assured, in lieu of using a conditional data flag. If the test(s) is 
failed, conditionally valid data could be replaced with substitute 
data, as appropriate, and resubmittal of the quarterly report would not 
be necessary.
    If a recertification test is failed or aborted due to a problem 
with the CEMS or if a routine daily calibration error test is failed 
during a recertification test period, proposed Sec. 75.20(b)(3) 
specifies that data validation would be done as follows:
    (1) If any required recertification test is failed, the test would 
have to be repeated. If any recertification test, other than a 7-day 
calibration error test, is failed or aborted due to a problem with the 
CEMS, the original recertification test period would end and any 
necessary maintenance activities, adjustments, linearizations, and 
reprogramming of the CEMS would need to be completed before a new 
recertification test period could begin. The new recertification test 
period would begin with a probationary calibration error test. The 
tests that would be required in this new recertification test period 
would include any tests that were required for the initial 
recertification event which were not successfully completed and any 
recertification or diagnostic tests required as a result of changes 
that were made to the monitoring system to correct the problems that 
caused failure of the recertification test;
    (2) If a linearity test, RATA, or cycle time test is failed or 
aborted due to a problem with the CEMS, all conditionally valid 
emission data recorded by the CEMS would be invalidated from the hour 
of commencement of the original recertification test period to the hour 
in which the test is failed or aborted. Data from the CEMS would remain 
invalid until the hour in which a new probationary calibration error 
test is passed following all of the necessary maintenance procedures, 
diagnostic tests, etc., at which time the conditionally valid status of 
emission data from the CEMS would begin;
    (3) If a 7-day calibration error test is failed within the 
recertification test period, the test would have to be re-started. 
Previously-recorded conditionally valid emission data from the CEMS 
would not be invalidated by a failed 7-day calibration error test 
unless the calibration error on the day of the failed 7-day calibration 
error test exceeded twice the performance specification in section 3 of 
Appendix A (causing the monitor to be considered out-of-control); and
    (4) If a calibration error test is failed during a recertification 
test period, the CEMS would be considered out-of-control as of the hour 
in which the calibration error test is failed. Emission data from the 
CEMS would be invalidated prospectively from the hour of the failed 
calibration error test until the hour of completion of a subsequent 
successful calibration error test following corrective action, at which 
time the conditionally valid data status would resume. Failure to 
perform a required daily calibration error test during a 
recertification test period would also cause data from the CEMS to be 
invalidated prospectively from the hour in which the calibration error 
test was due until the hour of completion of a subsequent successful 
calibration error test. Following a failed or missed calibration error 
test, no recertification tests could be performed until the required 
subsequent calibration error test had been passed.
5. Recertification and QA
    In today's proposed rule, a new section, 2.4, entitled 
``Recertification, Quality Assurance, and RATA Deadlines'' would be 
added to Appendix B. The purpose of this section would be to clarify 
the inter-relationship between normal quality assurance testing of CEMS 
and recertification events and to further clarify how RATA deadlines 
are determined. Appendix B to part 75 currently requires periodic 
(daily, quarterly, and semiannual or annual) quality assurance tests of 
all CEMS. The required daily QA tests include calibration error tests 
of all monitors and interference checks of flow monitors. Quarterly QA 
tests include linearity checks of gas monitors and leak checks of 
differential pressure-type flow monitors. The required semiannual or 
annual QA tests for all types of CEMS are RATAs.
    Under the current rule, when a significant change is made to a CEMS 
which affects the ability of the monitoring system to accurately read 
and record emissions data, Sec. 75.20(b) specifies that the CEMS must 
be recertified. To recertify a monitoring system, one or more of the 
following tests that were performed for initial certification of the 
CEMS must be repeated. That is, depending upon the nature of the change 
made to a CEMS, one or more of the following tests may be required for 
recertification: (1) calibration error test, (2) cycle time test, (3) 
linearity check, (4) RATA, or (5) DAHS verification. Notice that 
recertification tests (1), (3), and (4) are the same types of tests 
that are done for routine daily, quarterly, and semiannual or annual 
QA. There is, therefore, a connection between routine QA tests and 
recertification tests. Proposed Sec. 75.20(b) would further clarify 
that any change to a CEMS that does not require a RATA would not be 
considered a recertification event, and, therefore, would not require a 
recertification application. In such cases, the required tests would be 
considered diagnostic tests.
    Routine QA tests are generally planned and scheduled in advance, 
while recertification tests are performed on an as-required basis. 
Despite this, it is sometimes possible to coordinate component 
replacements or other changes to a CEMS with the QA test schedule for 
the CEMS. For instance, suppose that in a particular quarter, a CEMS 
component is replaced, and a RATA is required to recertify the 
monitoring system. Suppose, further, that in the quarter of the 
component replacement, the annual RATA is due, but has not yet been 
conducted. In this case, the recertification RATA could serve a dual 
purpose, i.e., to recertify the CEMS and to meet the annual RATA 
requirement. For this reason, EPA proposes to recommend in today's rule 
that, to the extent practicable, component replacements, system 
upgrades, and other events that require recertification be coordinated 
with the periodic (daily, quarterly, and semiannual or annual) QA 
testing required under Appendix B. Proposed section 2.4 of Appendix B 
clarifies that when a particular test is done for the dual purpose of 
recertification and routine QA, the data validation rules in 
Sec. 75.20(b)(3) pertaining to recertification would take precedence 
and would be followed. In a similar manner, a required diagnostic test 
(e.g., linearity check) could also be used to satisfy a quarterly 
linearity test requirement.
    Proposed section 2.4 of Appendix B emphasizes that, in general, 
whenever a RATA is performed, whether for QA purposes, recertification 
purposes, or both, the projected deadline for the next RATA (i.e., 
whether the next test is due in 2 or 4 QA operating quarters) would be 
established based upon the percentage relative accuracy obtained. For 
2-load and 3-load flow RATAs, the projected deadline for the next RATA 
would be established according to the highest relative accuracy at any 
of the loads tested. There would, however, be two important exceptions 
to this for single-load flow RATAs. Irrespective of

[[Page 28078]]

the relative accuracy percentage obtained, the results of a single-load 
flow RATA could only be used to establish an annual RATA frequency if: 
(1) the single-load flow RATA is specifically required under section 
2.3.1.3(b) of Appendix B for flow monitors installed on peaking units 
and bypass stacks, or (2) the single-load RATA is allowed under 
proposed section 2.3.1.3(c) of Appendix B for  85.0 percent 
historical unit operation at a single-load level. No other single-load 
flow RATA could be used to establish an annual frequency; however, a 2-
load flow RATA could be performed in place of any required single-load 
RATA, in order to achieve an annual frequency.
6. Data From Non-Redundant Backup Monitors
    Today's rule proposes to revise the quality assurance and data 
validation requirements in Sec. 75.20(d) for non-redundant backup 
monitoring systems. Under the May 17, 1995 rule, a ``non-redundant 
backup monitoring system'' is defined as a ``cold'' backup monitoring 
system which is brought into service on an as-needed basis, rather than 
being operated continuously. Non-redundant backup monitors must be 
initially certified at each location at which they are to be used, but 
unlike ``redundant backup'' monitors which are operated continuously 
and kept on ``hot-standby,'' non-redundant backup systems are not 
required to meet the daily and quarterly quality assurance requirements 
of Appendix B, except when they are actually used for data reporting. A 
linearity test of each non-redundant backup gas monitor is required 
before it is placed in service, and each non-redundant backup flow 
monitor must pass a calibration error test before being used to report 
data. The use of non-redundant backup monitors is restricted to 720 
hours a year at a particular unit or stack, unless a 7-day calibration 
error test is passed. A periodic recertification RATA of each non-
redundant backup monitor is required at least once every two years, at 
each location where it is to be used.
    Section 75.20(d) of today's proposal would clarify and expand the 
definition of a non-redundant backup monitoring system. Under the 
proposal, two distinct types of non-redundant backup systems would be 
defined: (1) type-1 is a system that has its own separate probe, sample 
interface, and analyzer (e.g., a portable gas monitoring system), and 
(2) type-2 is a system consisting of one or more like-kind replacement 
analyzers that use the same sample probe and interface as the primary 
monitoring system. This would include non-redundant backup analyzers 
that are used to meet the dual span and range requirements for 
SO2 or NO_X under proposed sections 2.1.1.4 and 
2.1.2.4 of Appendix A.
    The ``type-1'' system is the familiar non-redundant backup system 
described in the current version of part 75. However, the ``type-2'' 
system is a new kind of non-redundant backup monitoring system. EPA 
believes that allowing limited use of type-2 monitoring systems will 
encourage facilities that do not have redundant backup monitors to 
perform better maintenance on their primary analyzers. The Agency is 
concerned that primary analyzers with excessive, recurring daily 
calibration drift (i.e., monitors that fail calibration error tests 
more often than expected) are sometimes kept in service to avoid using 
substitute data, when the analyzers should be in the shop for 
maintenance. If the monitor readings tend to drift low from day to day, 
this can result in under-reporting of emissions, because data 
validation for daily calibrations under part 75 is prospective. That 
is, data are invalidated from the hour of a failed calibration error 
test forward, while data recorded from the hour of the previous 
successful calibration to the hour of the failed calibration are 
considered valid. EPA believes that allowing limited use of type-2 non-
redundant backup monitoring systems would provide a simple way (i.e., 
like-kind analyzer replacement) for primary analyzers to be properly 
maintained and repaired with minimal data loss.
    Today's proposal would retain the requirement for type-1 non-
redundant backup monitoring systems to be initially certified (except 
for a 7-day calibration error test) at each location at which they are 
to be used. However, type-2 systems would require no initial 
certification. Both types of systems would have to pass a linearity 
test (for gas monitors) or a calibration error test (for flow monitors) 
each time that they were used to report emission data. For a type-2 
``mix-and-match'' NO_X monitoring system consisting of one 
primary analyzer and one like-kind replacement analyzer, only the like-
kind replacement analyzer would have to pass a linearity test, provided 
that the primary analyzer is operating and not out-of-control with 
respect to any of its quality assurance requirements. When a non-
redundant backup monitoring system is brought into service, emission 
data from the non-redundant backup system could be deemed conditionally 
valid during the linearity test period, as follows. After making the 
like-kind replacement and prior to conducting the linearity test, a 
probationary calibration error test could be done to begin the period 
of conditionally valid data. If the linearity test is then passed 
within 168 unit operating hours of the probationary calibration error 
test, the conditionally valid data would be validated. However, if the 
linearity test is either failed, aborted due to a problem with the 
CEMS, or not completed as required, then all of the conditionally valid 
data would be invalidated beginning with the hour of the probationary 
calibration error test, and data from the non-redundant backup CEMS 
would remain invalid until the hour of completion of a successful 
linearity test.
    Under today's proposal, when a non-redundant backup system is used 
for part 75 reporting, the bias adjustment factor (BAF) from the most 
recent RATA of the system would be applied to the data generated by the 
system. If no RATA results were available for a type-2 system, the 
primary monitoring system BAF would be applied to the data generated by 
the type-2 system.
    Today's proposal would retain the restrictions of the current rule, 
which limit the annual usage of a non-redundant backup monitoring 
system to 720 hours at a particular location (unit or stack). To use a 
non-redundant backup system for more than 720 hours per year at a 
particular location would require a RATA of the system at that 
location. For type-1 systems, a recertification RATA would be required 
at least once every eight calendar quarters at each location at which 
the system is to be used. All non-redundant backup monitoring systems 
(type-1 and type-2) would have to be assigned unique system and 
component identification numbers and would have to be included in the 
monitoring plan for the unit or stack.
7. Missed QA Test Deadlines
    As discussed above under the subsections on ``Linearity Tests'' and 
``Relative Accuracy Test Audits,'' proposed sections 2.2.4 and 2.3.3 of 
Appendix B to today's rulemaking would allow a grace period in which to 
perform required linearity tests and RATAs whenever a test cannot be 
completed by the end of the quarter in which it is due. EPA believes it 
is appropriate to allow a grace period because circumstances beyond the 
control of the owner or operator (e.g., unplanned unit outages) 
sometimes arise which prevent the deadline for a quality assurance test 
from being met.
    The proposed linearity grace period is 168 unit operating hours, 
and the proposed RATA grace period is 720 unit operating hours. A 
linearity grace period

[[Page 28079]]

could only be used to satisfy the linearity requirement from a previous 
quarter. For any RATA (or RATAs, if more than one attempt is made) 
conducted during a grace period, the deadline for the next RATA would 
be calculated from the quarter in which the RATA was originally due, 
not from the quarter in which the RATA is actually completed.
    Data validation during a grace period would be done according to 
the applicable provisions in proposed section 2.2.3 of Appendix B (for 
linearities) or section 2.3.2 of Appendix B (for RATAs). Data from a 
CEMS would become invalid upon expiration of a grace period if the 
required linearity test or RATA had not been completed. Data from the 
CEMS would remain invalid after the expiration of the grace period 
until the required test is successfully completed.

P. Appendix D

1. Pipeline Natural Gas Definitions
Background
    Appendix D provides an optional protocol by which oil-fired and 
gas-fired units may account for their SO2 mass emissions. 
Under the definitions of ``oil-fired'' and ``gas-fired'' in Sec. 72.2, 
Appendix D may be used to measure SO2 emissions from gaseous 
fuels only if the gaseous fuel's sulfur content is less than or equal 
to that of natural gas.
    In developing Appendix D, EPA assumed that virtually all of the 
gaseous fuel combusted by affected units in the Acid Rain Program would 
be pipeline natural gas. Section 2.3 of Appendix D of the January 11, 
1993 rule allowed for accounting for SO2 emissions from 
gaseous fuel using EPA's ``National Allowance Database (NADB) emission 
rate.'' The NADB was used to establish a baseline of historical 
SO2 emissions in order to allocate allowances. For the vast 
majority of units combusting pipeline natural gas, NADB used the 
historical heat input from gas and an emission rate of 0.0006 pounds of 
SO2 per measured million British thermal units (lb/mmBtu) 
(see Docket A-92-06; Docket A-94-16, Item II-F-2). This default factor 
is derived from EPA Publication AP-42 and is based on a sulfur content 
of 0.2 grains per 100 standard cubic feet of gaseous fuel (gr/100 scf) 
(see Docket A-97-35, Item II-I-1). Use of this default SO2 
emission rate factor for pipeline natural gas was clarified by EPA in 
its Acid Rain Policy Manual (see Docket A-97-35, Item II-I-9, Policy 
Manual, Question 2.4).
    Section 2.3.2 of Appendix D, as revised by the May 17, 1995 direct 
final rule, explicitly allows owners or operators to use a default 
emission factor of 0.0006 (lb/mmBtu) to estimate SO2 
emissions during hours in which pipeline natural gas is combusted. 
Alternatively, section 2.3.1 of Appendix D, also as revised by the May 
17, 1995 direct final rule, allows for determining SO2 
emissions from any gaseous fuel with a sulfur content no greater than 
natural gas by performing daily fuel sampling, analyzing the sulfur 
content of the gaseous fuel, and multiplying that sulfur content in 
grains per 100 standard cubic feet (gr/100scf) times the volume of 
gaseous fuel combusted. Units combusting gaseous fuels with a total 
sulfur content greater than natural gas (i.e., > 20 gr/100scf) are not 
allowed to use the procedures of Appendix D and must instead use an 
SO2 CEMS and a flow monitor to determine SO2 mass 
emissions. This limitation is explicitly stated in Sec. 75.11(e)(4), as 
revised on November 20, 1996.
    The definition of ``natural gas'' in Sec. 72.2, as revised by the 
May 17, 1995 direct final rule, indicates that the sulfur content of 
natural gas is ``1 grain or less hydrogen sulfide per 100 standard 
cubic feet, and 20 grains or less total sulfur per 100 standard cubic 
feet.'' This definition was taken from Requirements of the Federal 
Energy Regulatory Commission (FERC) for regulation of the transmission 
of natural gas. ``Pipeline natural gas'' is also defined in Sec. 72.2. 
However, the definition is simply ``natural gas that is provided by a 
supplier through a pipeline,'' and provides no specifications for 
sulfur content or hydrogen sulfide content.
    Section 2.3.2.2 of Appendix D requires documentation of the 
contractual sulfur content of pipeline natural gas from the supplier. 
This documentation was intended to demonstrate that the natural gas is 
supplied through a pipeline, as well as that it meets the sulfur 
content definition for natural gas.
    Questions over the applicability of Appendix D and the apparent 
inconsistencies between the definitions ``natural gas'' and ``pipeline 
natural gas'' in Sec. 72.2 and the provisions of section 2.3 of 
Appendix D have caused confusion during program implementation since 
the May 17, 1995 direct final rule. Some utilities have interpreted 
section 2.3.2.2 of Appendix D to allow pipeline natural gas to have a 
sulfur content as high as 20 gr/100 scf, which is one hundred times 
higher than the sulfur content upon which the 0.0006 lb/mmBtu emission 
factor is based. During the process of applying for certification of 
monitoring equipment for six gas-fired units, one utility indicated to 
the Agency that it intended to use a default emission rate of 0.0006 
lb/mmBtu and heat input to account for SO2 mass emissions 
from propane liquefied petroleum gas (see Docket A-97-35, Item II-D-6). 
Based upon the information provided by the utility in its monitoring 
plan for the units, the sulfur content of propane was several times 
higher than that of pipeline natural gas, with a range of sulfur 
content between 0.08 and 2.72 gr/100 scf, compared to a typical sulfur 
content of 0.2 gr/100 scf for pipeline natural gas, upon which the 
default SO2 emission rate of 0.0006 lb/mmBtu is based. Later 
information submitted by the utility indicated that during the previous 
three years, the sulfur content of propane combusted at that plant had 
an average value of 0.83 gr/100 scf and a maximum value of 2.20 gr/100 
scf (see Docket A-97-35, Item II-D-60). EPA rejected the utility's 
monitoring approach using the default emission rate for pipeline 
natural gas because it would have resulted in an underestimation of 
SO2 emissions, as well as not following the procedures of 
Appendix D (see Docket A-97-35, Item II-C-2).
    Other utilities have tried to use the default SO2 
emission rate of 0.0006 lb/mmBtu for higher sulfur gaseous fuels, such 
as digester gas (see Docket A-94-16, Item II-D-71). EPA issued policy 
guidance to ensure that other utilities were aware that the default 
SO2 emission rate of 0.0006 lb/mmBtu should only be used for 
pipeline natural gas with a low sulfur content of 0.2 gr/100 scf (see 
Docket A-97-35, Item II-I-9, Policy Manual, Question 2.15, as 
originally published in March 1996). However, several utilities were 
concerned that this excluded some pipeline natural gas (see Docket A-
97-35, Items II-B-3, II-E-16). As stated in the technical support 
document for the May 17, 1995 direct final rule, EPA had intended that 
all pipeline natural gas would qualify for use of the default 
SO2 emission rate of 0.0006 lb/mmBtu. Therefore, the Agency 
revised its guidance to clarify that a facility needed only to document 
that it was using pipeline natural gas, without documenting a sulfur 
content of 0.2 gr/100 scf (see Docket A-97-35, Item II-I-9, Policy 
Manual, Question 2.15, as revised in June 1996). During this process, 
the Agency became concerned that the definition of pipeline natural gas 
in Sec. 72.2 was not clear enough and that the sulfur content 
documentation required for pipeline natural gas in section 2.3.2.2 of 
Appendix D was confusing and possibly inappropriate.

[[Page 28080]]

Discussion of Proposed Changes
    For the definition of pipeline natural gas in Sec. 72.2, today's 
proposal includes a revised definition that would indicate pipeline 
natural gas is low in the sulfur-bearing compound hydrogen sulfide 
(H2S). The proposed revised definition would specifically 
include the maximum hydrogen sulfide content for pipeline natural gas 
permitted by fuel purchase or transportation contracts. The hydrogen 
sulfide content of pipeline natural gas is proposed to be up to 0.3 gr/
100 scf.
    In addition, section 2.3 of Appendix D would be revised. As under 
the current rule provisions, sources would be allowed to use a default 
SO2 emission rate of 0.0006 lb SO2/mmBtu in 
conjunction with unit heat input to calculate the SO2 mass 
emission rate during the combustion of pipeline natural gas. In order 
to demonstrate that the pipeline natural gas qualifies to use the 
default SO2 emission rate of 0.0006 lb/mmBtu, it would be 
necessary for the designated representative to provide information in 
the monitoring plan on the gas's maximum hydrogen sulfide content from 
the facility's purchase contract with the pipeline gas supplier or from 
the pipeline natural gas supplier's transportation contract. In such 
contracts, or in the tariff sheets associated with them, the pipeline 
gas supplier typically agrees to provide natural gas with a maximum 
hydrogen sulfide content of 0.25 gr/100 scf or 0.30 gr/100 scf. If a 
facility has previously submitted contract information from its 
pipeline gas supplier containing a limit on the sulfur content, this 
information typically also verifies the limit on the hydrogen sulfide 
content. For pipeline natural gas, it would not be necessary to provide 
sampling information to verify that the hydrogen sulfide content 
actually meets the quality specification limit on the hydrogen sulfide 
content stated in the definition of pipeline natural gas.
    If a facility wanted to demonstrate that another gaseous fuel had 
an SO2 emission rate no greater than pipeline natural gas, 
and thus, could use the default emission rate factor of 0.0006 lb/
mmBtu, the designated representative would provide sulfur content and 
GCV information in the monitoring plan for the unit or could petition 
under Sec. 75.66(i) after initial certification for the unit. It would 
be necessary for the designated representative to demonstrate that the 
gaseous fuel has an SO2 emission rate no greater than 0.0006 
lb/mmBtu. The designated representative would need to provide at least 
720 hours of data for the demonstration. The data could come from the 
fuel supplier, if the fuel came from a gas supplier.
    For all units using Appendix D, proposed section 2.3.3 would 
require the designated representative to provide information to the 
Agency demonstrating that the total sulfur content of the gaseous fuel 
meets the requirements of Appendix D and that the unit meets the 
Sec. 72.2 definition of ``gas-fired'' or ``oil-fired.'' Additionally, 
the gas-fired definition would be revised to indicate that the 
restriction of burning gaseous fuels containing no more sulfur than 
natural gas is actually a restriction on the total sulfur in the fuel. 
The gaseous fuel's total sulfur content would have to be shown to be 
less than or equal to 20 grains total sulfur per 100 standard cubic 
feet of gaseous fuel.
Rationale
    The Agency proposes to introduce specific hydrogen sulfide content 
values into the definition of pipeline natural gas in order to provide 
a guideline that will separate gaseous fuels with a higher sulfur 
content from low sulfur pipeline natural gas. The maximum hydrogen 
sulfide content of 0.3 gr/100 scf is being proposed for two reasons. 
First, hydrogen sulfide contents of 0.25 or 0.3 gr/100 scf are 
typically required under pipeline gas transmission contracts, and 
should be relatively easy to document (see Docket A-97-35, Item II-E-
19). In addition, 0.2 gr/100 scf is the sulfur content equivalent to 
the default emission rate factor of 0.0006 lb/mmBtu from the Agency's 
AP-42 emission factors that may be used by units combusting pipeline 
natural gas under section 2.3.2 of Appendix D (see Docket A-97-35, Item 
II-A-6). A maximum hydrogen sulfide content of 0.3 gr/100 scf 
corresponds to this default emission rate far more closely than a total 
sulfur content of 20.0 gr/100 scf or a hydrogen sulfide content of 1.0 
gr/100 scf and, yet, would allow for some variability in the hydrogen 
sulfide content above a 0.2 gr/100 scf average. EPA believes that all 
or virtually all pipeline natural gas that is supplied through a 
pipeline for commercial use can meet these qualifications.
    Pipeline natural gas is composed predominantly of methane 
(CH4). Hydrogen sulfide is the predominant molecule 
containing sulfur in pipeline natural gas. Therefore, restricting the 
hydrogen sulfide content of pipeline natural gas to 0.3 gr/100 scf 
serves as a proxy for a limit on the total sulfur content, while being 
relatively easy to document. This revised definition of pipeline 
natural gas would also serve to restrict the default emission rate 
factor from being inappropriately applied to higher sulfur gaseous 
fuels, such as liquefied petroleum gas (see Docket A-97-35, Item II-D-
6) or digester gas (see Docket A-94-16, Item II-D-71).
    Appendix D of today's proposed rule would be revised to clarify the 
documentation requirements for sulfur content and hydrogen sulfide 
content of gaseous fuel, including pipeline natural gas. The original 
wording of section 2.3.2.2 implied that pipeline natural gas only need 
to have a total sulfur content of 20 gr/100 scf, roughly 100 times the 
sulfur content associated with the default emission rate of 0.0006 lb/
mmBtu. Some utilities found this confusing (see Docket A-97-35, Items 
II-D-6, II-E-10). Therefore, EPA issued guidance to clarify that the 
default emission rate factor was only intended to apply to lower sulfur 
pipeline natural gas (see Docket A-97-35, Item II-I-9, Policy Manual, 
Question 2.15).
    However, some utilities using pipeline natural gas were concerned 
that because their fuel suppliers were not willing to certify or agree 
to a sulfur content of 0.3 gr/100 scf by contract, they might be 
required to perform daily gas sampling (see Docket A-97-35, Items II-B-
3, II-E-15, II-E-16). This was not the Agency's intent. The Agency 
merely wishes to ensure that facilities provide adequate documentation 
to demonstrate that the unit will not be underestimating SO2 
emissions for a high sulfur gaseous fuel by using an inappropriate 
default emission rate factor that applies to extremely low sulfur gas. 
Similar to EPA's Policy Manual Question 2.15 referred to above, a 
facility would need only to provide the fuel quality specification for 
total sulfur content and hydrogen sulfide from the pipeline supplier, 
or from the tariff sheet for the pipeline, in order to qualify to use 
the default emission rate.
    If a facility intends to use the default emission rate factor for a 
gaseous fuel other than pipeline natural gas, sulfur content and GCV 
data would have to be provided and analyzed to demonstrate that the 
fuel has an SO2 emission rate no greater than 0.0006 lb/
mmBtu. A minimum of 720 hours of data would be required for the 
demonstration. Each hourly value of the total sulfur content (in gr/100 
scf) would be divided by the GCV value (in Btu/100 scf) and then 
multiplied by a conversion factor of 10⁶ Btu/mmBtu. This 
would provide a ratio of the number of grains of sulfur in the fuel to 
the heat content of the fuel. For pipeline natural gas with an assumed 
SO2 emission rate of 0.0006 lb/mmBtu, a sulfur content of 
0.2 gr/100 scf and a


[[Continued on page 28081]]
Notices For
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
Acid Rain Program; Continuous Emission Monitoring Rule Revisions

Related Material

Local Navigation
