[Code of Federal Regulations]
[Title 40, Volume 14]
[Revised as of July 1, 2003]
From the U.S. Government Printing Office via GPO Access
[CITE: 40CFR75.33]
[Page 264-271]
TITLE 40--PROTECTION OF ENVIRONMENT
CHAPTER I--ENVIRONMENTAL PROTECTION AGENCY (CONTINUED)
PART 75--CONTINUOUS EMISSION MONITORING--Table of Contents
Subpart D--Missing Data Substitution Procedures
Sec. 75.33 Standard missing data procedures for SO2,
NOX and flow rate.
(a) Following initial certification of the required SO2,
NOX, and flow rate monitoring system(s) at a particular unit
or stack location (i.e., the date and time at which quality-assured data
begins to be recorded by CEMS(s) at that location) and upon completion
of the first 720 quality-assured monitor operating hours (for
SO2) or the first 2,160 quality-assured monitor operating
hours (for flow, NOX emission rate, or NOX
concentration), the owner or operator shall provide substitute data
required under this subpart according to the procedures in paragraphs
(b) and (c) of this section and depicted in Table 1 (SO2) and
Table 2 of this section (NOX, flow). The owner or operator
may either implement the provisions of paragraphs (b) and (c) of this
section on a non-fuel-specific basis, or may, as described in paragraphs
(b)(5), (b)(6), (c)(7) and (c)(8) of this section, provide fuel-specific
substitute data values. Notwithstanding these requirements, if three
years (26,280 clock hours) have elapsed since the date and hour of
initial certification, and fewer than 720 (or 2,160, as applicable)
quality-assured monitor operating hours have been recorded, the owner or
operator shall begin using the missing data procedures of this section.
The owner or operator of a unit shall substitute for missing data using
quality-assured monitor operating hours of data from no earlier than
three years (26,280 clock hours) prior to the date and time of the
missing data period.
(b) SO2 concentration data. For each hour of missing
SO2 concentration data,
[[Page 265]]
(1) Whenever the monitor data availability is equal to or greater
than 95.0 percent, the owner or operator shall calculate substitute data
by means of the automated data acquisition and handling system for each
hour of each missing data period according to the following procedures:
(i) For a missing data period less than or equal to 24 hours,
substitute the average of the hourly SO2 concentrations
recorded by an SO2 pollutant concentration monitor for the
hour before and the hour after the missing data period.
(ii) For a missing data period greater than 24 hours, substitute the
greater of:
(A) The 90th percentile hourly SO2 concentration recorded
by an SO2 pollutant concentration monitor during the previous
720 quality-assured monitor operating hours; or
(B) The average of the hourly SO2 concentrations recorded
by an SO2 pollutant concentration monitor for the hour before
and the hour after the missing data period.
(2) Whenever the monitor data availability is at least 90.0 percent
but less than 95.0 percent, the owner or operator shall calculate
substitute data by means of the automated data acquisition and handling
system for each hour of each missing data period according to the
following procedures:
(i) For a missing data period of less than or equal to 8 hours,
substitute the average of the hourly SO2 concentrations
recorded by an SO2 pollutant concentration monitor for the
hour before and the hour after the missing data period.
(ii) For a missing data period of more than 8 hours, substitute the
greater of:
(A) the 95th percentile hourly SO2 concentration recorded
by an SO2 pollutant concentration monitor during the previous
720 quality-assured monitor operating hours; or
(B) The average of the hourly SO2 concentrations recorded
by an SO2 pollutant concentration monitor for the hour before
and the hour after the missing data period.
(3) Whenever the monitor data availability is at least 80.0 percent
but less than 90.0 percent, the owner or operator shall substitute for
each missing data period the maximum hourly SO2 concentration
recorded by an SO2 pollutant concentration monitor during the
previous 720 quality-assured monitor operating hours.
(4) Whenever the monitor data availability is less than 80.0
percent, the owner or operator shall substitute for each missing data
period the maximum potential SO2 concentration, as defined in
section 2.1.1.1 of appendix A to this part.
(5) For units that combust more than one type of fuel, the owner or
operator may opt to implement the missing data routines in paragraphs
(b)(1) through (b)(4) of this section on a fuel-specific basis. If this
option is selected, the owner or operator shall document this in the
monitoring plan required under Sec. 75.53.
(6) Use the following guidelines to implement paragraphs (b)(1)
through (b)(4) of this section on a fuel-specific basis:
(i) Separate the historical, quality-assured SO2
concentration data according to the type of fuel combusted;
(ii) For units that co-fire different types of fuel, either group
the co-fired hours with the historical data for the fuel with the
highest SO2 emission rate (e.g., if diesel oil and pipeline
natural gas are co-fired, count co-fired hours as oil-burning hours), or
separate the co-fired hours from the single-fuel hours;
(iii) For the purposes of providing substitute data under paragraph
(b)(4) of this section, determine a separate, fuel-specific maximum
potential SO2 concentration (MPC) value for each type of fuel
combusted in the unit, in a manner consistent with section 2.1.1.1 of
appendix A to this part. For fuel that qualifies as pipeline natural gas
or natural gas (as defined in Sec. 72.2 of this chapter), the owner or
operator shall, for the purposes of determining the MPC, either
determine the maximum total sulfur content and minimum gross calorific
value (GCV) of the gas by fuel sampling and analysis or shall use a
default total sulfur content of 0.05 percent by weight (dry basis) and a
default GCV value of 950 Btu/scf. For co-firing, the MPC value shall be
based on the fuel with the highest SO2 emission rate. The
exact methodology used to determine each fuel-specific MPC
[[Page 266]]
value shall be documented in the monitoring plan for the unit or stack;
and
(iv) For missing data periods that require 720-hour (or, if
applicable, 3-year) lookbacks, use historical data for the type of fuel
combusted during each hour of the missing data period to determine the
appropriate substitute data value for that hour. For co-fired missing
data hours, if the historical data are separated into single-fuel and
co-fired hours, use co-fired data to provide the substitute data values.
Otherwise, use data for the fuel with the highest SO2
emission rate to provide substitute data values for co-fired missing
data hours.
(7) Table 1 summarizes the provisions of paragraphs (b)(1) through
(b)(6) of this section.
(c) Volumetric flow rate, NOX emission rate and NOX
concentration data. Use the procedures in this paragraph to provide
substitute NOX and flow rate data for all affected units for
which load-based ranges have been defined in accordance with section 2
of appendix C to this part. For units that do not produce electrical or
thermal output (i.e., non-load-based units), use the procedures in this
paragraph only to provide substitute data for volumetric flow rate, and
only if operational bins have been defined for the unit, as described in
section 3 of appendix C to this part. Otherwise, use the applicable
missing data procedures in paragraph (d) or (e) of this section for non-
load-based units. For each hour of missing volumetric flow rate data,
NOX emission rate data, or NOX concentration data
used to determine NOX mass emissions:
(1) Whenever the monitor data availability is equal to or greater
than 95.0 percent, the owner or operator shall calculate substitute data
by means of the automated data acquisition and handling system for each
hour of each missing data period according to the following procedures:
(i) For a missing data period less than or equal to 24 hours,
substitute, as applicable, for each missing hour, the arithmetic average
of the flow rates or NOX emission rates or NOX
concentrations recorded by a monitoring system during the previous 2,160
quality-assured monitor operating hours at the corresponding unit load
range or operational bin, as determined using the procedure in appendix
C to this part.
(ii) For a missing data period greater than 24 hours, substitute, as
applicable, for each missing hour, the greater of:
(A) The 90th percentile hourly flow rate or the 90th percentile
NOX emission rate or the 90th percentile NOX
concentration recorded by a monitoring system during the previous 2,160
quality-assured monitor operating hours at the corresponding unit load
range or operational bin, as determined using the procedure in appendix
C to this part; or
(B) The average of the recorded hourly flow rates, NOX
emission rates or NOX concentrations recorded by a monitoring
system for the hour before and the hour after the missing data period.
(2) Whenever the monitor data availability is at least 90.0 percent
but less than 95.0 percent, the owner or operator shall calculate
substitute data by means of the automated data acquisition and handling
system for each hour of each missing data period according to the
following procedures:
(i) For a missing data period of less than or equal to 8 hours,
substitute, as applicable, the arithmetic average hourly flow rate or
NOX emission rate or NOX concentration recorded by
a monitoring system during the previous 2,160 quality-assured monitor
operating hours at the corresponding unit load range or operational bin,
as determined using the procedure in appendix C to this part.
(ii) For a missing data period greater than 8 hours, substitute, as
applicable, for each missing hour, the greater of:
(A) The 95th percentile hourly flow rate or the 95th percentile
NOX emission rate or the 95th percentile NOX
concentration recorded by a monitoring system during the previous 2,160
quality-assured monitor operating hours at the corresponding unit load
range or operational bin, as determined using the procedure in appendix
C to this part; or
(B) The average of the hourly flow rates, NOX emission
rates or NOX concentrations recorded by a monitoring system
for the hour before and the hour after the missing data period.
(3) Whenever the monitor data availability is at least 80.0 percent
but less
[[Page 267]]
than 90.0 percent, the owner or operator shall, by means of the
automated data acquisition and handling system, substitute, as
applicable, for each hour of each missing data period, the maximum
hourly flow rate or the maximum hourly NOX emission rate or
the maximum hourly NOX concentration recorded during the
previous 2,160 quality-assured monitor operating hours at the
corresponding unit load range or operational bin, as determined using
the procedure in appendix C to this part.
(4) Whenever the monitor data availability is less than 80.0
percent, the owner or operator shall substitute, as applicable, for each
hour of each missing data period, the maximum potential flow rate, as
defined in section 2.1.4.1 of appendix A to this part, or the maximum
NOX emission rate, as defined in section 2.1.2.1 of appendix
A to this part, or the maximum potential NOX concentration,
as defined in section 2.1.2.1 of appendix A to this part. In addition,
when non-load-based operational bins are used, the owner or operator
shall substitute the maximum potential flow rate for any hour in the
missing data period in which essential operating or parametric data are
unavailable and the operational bin cannot be determined.
(5) This paragraph, (c)(5), does not apply to non-load-based,
affected units using operational bins. Whenever no prior quality-assured
flow rate data, NOX concentration data or NOX
emission rate data exist for the corresponding load range, the owner or
operator shall substitute, as applicable, for each hour of missing data,
the maximum hourly flow rate or the maximum hourly NOX
concentration or maximum hourly NOX emission rate at the next
higher level load range for which quality-assured data are available.
(6) Whenever no prior quality-assured flow rate data, NOX
concentration data or NOX emission rate data exist at either
the corresponding load range (or a higher load range) or at the
corresponding operational bin, the owner or operator shall substitute,
as applicable, either the maximum potential NOX emission rate
or the maximum potential NOX concentration, as defined in
section 2.1.2.1 of appendix A to this part or the maximum potential flow
rate, as defined in section 2.1.4.1 of appendix A to this part.
(7) This paragraph (c)(7) does not apply to affected units using
non-load-based operational bins. For units that combust more than one
type of fuel, the owner or operator may opt to implement the missing
data routines in paragraphs (c)(1) through (c)(6) of this section on a
fuel-specific basis. If this option is selected, the owner or operator
shall document this in the monitoring plan required under Sec. 75.53.
(8) This paragraph, (c)(8), does not apply to affected units using
non-load-based operational bins. Use the following guidelines to
implement paragraphs (c)(1) through (c)(6) of this section on a fuel-
specific basis:
(i) Separate the historical, quality-assured NOX emission
rate, NOX concentration, or flow rate data according to the
type of fuel combusted;
(ii) For units that co-fire different types of fuel, either group
the co-fired hours with the historical data for the fuel with the
highest NOX emission rate, NOX concentration or
flow rate, or separate the co-fired hours from the single-fuel hours;
(iii) For the purposes of providing substitute data under paragraph
(c)(4) of this section, a separate, fuel-specific maximum potential
concentration (MPC), maximum potential NOX emission rate
(MER), or maximum potential flow rate (MPF) value (as applicable) shall
be determined for each type of fuel combusted in the unit, in a manner
consistent with Sec. 72.2 of this chapter and with section 2.1.2.1 or
2.1.4.1 of appendix A to this part. For co-firing, the MPC, MER or MPF
value shall be based on the fuel with the highest emission rate or flow
rate (as applicable). The exact methodology used to determine each fuel-
specific MPC, MER or MPF value shall be documented in the monitoring
plan for the unit or stack.
(iv) For missing data periods that require 2,160-hour (or, if
applicable, 3-year) lookbacks, use historical data for the type of fuel
combusted during each hour of the missing data period to determine the
appropriate substitute data value for that hour. For co-fired
[[Page 268]]
missing data hours, if the historical data are separated into single-
fuel and co-fired hours, use co-fired data to provide the substitute
data values. Otherwise, use data for the fuel with the highest
NOX emission rate, NOX concentration or flow rate
(as applicable) to provide substitute data values for co-fired missing
data hours. Tables 1 and 2 follow.
Table 1--Missing Data Procedure for SO2 CEMS, CO2 CEMS, Moisture CEMS and Diluent (CO2 or O2) Monitors for Heat
Input Determination
----------------------------------------------------------------------------------------------------------------
Trigger conditions Calculation routines
----------------------------------------------------------------------------------------------------------------
Duration (N) of CEMS
Monitor data availability (percent) outage (hours) \2\ Method Lookback period
----------------------------------------------------------------------------------------------------------------
95 or more.......................... N [le] 24 Average................ HB/HA.
N 24 For SO2, CO2, and H2O
**, the greater of:.
Average.............. HB/HA.
90th percentile...... 720 hours *.
For O2 and H2Ox , the
lesser of:.
Average.............. HB/HA.
10th percentile...... 720 hours *.
90 or more, but below 95............ N [le] 8 Average................ HB/HA.
N 8 For SO2, CO2, and
H2O**, the greater of:.
Average.............. HB/HA.
95th percentile...... 720 hours *.
For O2 and H2Ox, the
lesser of:.
Average.............. HB/HA.
5th percentile....... 720 hours *.
80 or more, but below 90............ N 0 For SO2, CO2, and
H2O**,.
Maximum value \1\.... 720 hours *.
For O2 and H2Ox:.......
Minimum value \1\.... 720 hours *.
Below 80............................ N 0 Maximum potential
concentration or %
(for SO2, CO2, and H2O
**) or.
Minimum potential None.
concentration or %
(for O2 and H2Ox).
----------------------------------------------------------------------------------------------------------------
HB/HA = hour before and hour after the CEMS outage.
*Quality-assured, monitor operating hours, during unit operation. May be either fuel-specific or non-fuel-
specific. For units that report data only for the ozone season, include only quality-assured monitor operating
hours within the ozone season in the lookback period. Use data from no earlier than 3 years prior to the
missing data period.
\1\ Where a unit with add-on SO2 emission controls can demonstrate that the controls are operating properly, as
provided in Sec. 75.34, the unit may, upon approval, use the maximum controlled emission rate from the
previous 720 operating hours.
\2\ During unit operating hours.
\x\ Use this algorithm for moisture except when Equation 19-3, 19-4 or 19-8 in Method 19 in appendix A to part
60 of this chapter is used for NOX emission rate.
**Use this algorithm for moisture only when Equation 19-3, 19-4 or 19-8 in Method 19 in appendix A to part 60 of
this chapter is used for NOX emission rate.
Table 2--Load-Based Missing Data Procedure for NOX-Diluent CEMS, NOX Concentration CEMS and Flow Rate CEMS
----------------------------------------------------------------------------------------------------------------
Trigger conditions Calculation routines
----------------------------------------------------------------------------------------------------------------
Duration (N) of
Monitor data availability CEMS outage Method Lookback period Load ranges
(percent) (hours) \2\
----------------------------------------------------------------------------------------------------------------
95 or more..................... N [le] 24 Average........... 2160 hours *...... Yes.
N 24 The greater of:...
Average......... HB/HA............. No.
90th percentile. 2160. hours *..... Yes.
90 or more, but below 95....... N [le] 8 Average........... 2160 hours *...... Yes.
N8 The greater of....
Average......... HB/HA............. No
95th percentile. 2160 hours *...... Yes.
80 or more, but below 90....... N 0 Maximum value \1\. 2160 hours *...... Yes.
Below 80....................... N 0 Maximum NOX None.............. No.
emission rate; or
maximum potential
NOX
concentration; or
maximum potential
flow rate.
----------------------------------------------------------------------------------------------------------------
HB/HA = hour before and hour after the CEMS outage.
* Quality-assured, monitor operating hours, using data at the corresponding load range (``load bin'') for each
hour of the missing data period. May be either fuel-specific or non-fuel-specific. For units that report data
only for the ozone season, include only quality-assured monitor operating hours within the ozone season in the
lookback period. Use data from no earlier than three years prior to the missing data period.
[[Page 269]]
\1\ Where a unit with add-on NOX emission controls can demonstrate that the controls are operating properly, as
provided in Sec. 75.34, the unit may, upon approval, use the maximum controlled emission rate from the
previous 720 operating hours. Alternatively, units with add-on controls that report NOX mass emissions on a
year-round basis under subpart H of this part may use separate ozone season and non-ozone season databases to
provide substitute data values, as described in Sec. 75.34(a)(2).
\2\ During unit operating hours.
(9) The load-based provisions of paragraphs (c)(1) through (c)(8) of
this section are summarized in Table 2 of this section. The non-load-
based provisions for volumetric flow rate, found in paragraphs (c)(1)
through (c)(4), and (c)(6) of this section, are presented in Table 4 of
this section.
(d) Non-load-based NO X emission rate and NOX
concentration data. Use the procedures in this paragraph to provide
substitute NOX data for affected units that do not produce
electrical output (in megawatts) or thermal output (in klb/hr of steam).
For each hour of missing NOX emission rate data, or
NOX concentration data used to determine NOX mass
emissions:
(1) Whenever the monitor data availability is equal to or greater
than 95.0 percent, the owner or operator shall calculate substitute data
by means of the automated data acquisition and handling system for each
hour of each missing data period according to the following procedures:
(i) For a missing data period less than or equal to 24 hours,
substitute, as applicable, for each missing hour, the arithmetic average
of the NOX emission rates or NOX concentrations
recorded by a monitoring system in a 2,160 hour lookback period. The
lookback period may be comprised of either:
(A) The previous 2,160 quality-assured monitor operating hours, or
(B) The previous 2,160 quality-assured monitor operating hours at
the corresponding operational bin, if operational bins, as defined in
section 3 of appendix C to this part, are used.
(ii) For a missing data period greater than 24 hours, substitute,
for each missing hour, the 90th percentile NOX emission rate
or the 90th percentile NOX concentration recorded by a
monitoring system during the previous 2,160 quality-assured monitor
operating hours (or during the previous 2,160 quality-assured monitor
operating hours at the corresponding operational bin, if operational
bins are used).
(2) Whenever the monitor data availability is at least 90.0 percent
but less than 95.0 percent, the owner or operator shall calculate
substitute data by means of the automated data acquisition and handling
system for each hour of each missing data period according to the
following procedures:
(i) For a missing data period of less than or equal to eight hours,
substitute, as applicable, the arithmetic average of the hourly
NOX emission rates or NOX concentrations recorded
by a monitoring system during the previous 2,160 quality-assured monitor
operating hours (or during the previous 2,160 quality-assured monitor
operating hours at the corresponding operational bin, if operational
bins are used).
(ii) For a missing data period greater than eight hours, substitute,
for each missing hour, the 95th percentile hourly flow rate or the 95th
percentile NOX emission rate or the 95th percentile
NOX concentration recorded by a monitoring system during the
previous 2,160 quality-assured monitor operating hours (or during the
previous 2,160 quality-assured monitor operating hours at the
corresponding operational bin, if operational bins are used).
(3) Whenever the monitor data availability is at least 80.0 percent
but less than 90.0 percent, the owner or operator shall, by means of the
automated data acquisition and handling system, substitute, as
applicable, for each hour of each missing data period, the maximum
hourly NOX emission rate or the maximum hourly NOX
concentration recorded during the previous 2,160 quality-assured monitor
operating hours (or during the previous 2,160 quality-assured monitor
operating hours at the corresponding operational bin, if operational
bins are used).
(4) Whenever the monitor data availability is less than 80.0
percent, the owner or operator shall substitute, as applicable, for each
hour of each missing data period, the maximum NOX emission
rate, as defined in Sec. 72.2 of this chapter, or the maximum potential
[[Page 270]]
NOX concentration, as defined in section 2.1.2.1 of appendix
A to this part. In addition, when operational bins are used, the owner
or operator shall substitute (as applicable) the maximum potential
NOX emission rate or the maximum potential NOX
concentration for any hour in the missing data period in which essential
operating or parametric data are unavailable and the operational bin
cannot be determined.
(5) If operational bins are used and no prior quality-assured
NOX concentration data or NOX emission rate data
exist for the corresponding operational bin, the owner or operator shall
substitute, as applicable, either the maximum potential NOX
emission rate, as defined in Sec. 72.2 of this chapter, or the maximum
potential NOX concentration, as defined in section 2.1.2.1 of
appendix A to this part.
(6) Table 3 of this section summarizes the provisions of paragraphs
(d)(1) through (d)(5) of this section.
(e) Non-load-based volumetric flow rate data. (1) If operational
bins, as defined in section 3 of appendix C to this part, are used for a
unit that does not produce electrical or thermal output, use the missing
data procedures in paragraph (c) of this section to provide substitute
volumetric flow rate data for the unit.
(2) If operational bins are not used, modify the procedures in
paragraph (c) of this section as follows:
(i) In paragraphs (c)(1) through (c)(3), the words ``previous 2,160
quality-assured monitor operating hours'' shall apply rather than
``previous 2,160 quality-assured monitor operating hours at the
corresponding unit load range or operational bin, as determined using
the procedure in appendix C to this part;''
(ii) The last sentence in paragraph (c)(4) does not apply;
(iii) Paragraphs (c)(5), (c)(7), and (c)(8) are not applicable; and
(iv) In paragraph (c)(6), the words, ``for either the corresponding
load range (or a higher load range) or at the corresponding operational
bin'' do not apply.
(3) Table 4 of this section summarizes the provisions of paragraphs
(e)(1) and (e)(2) of this section. Tables 3 and 4 follow:
Table 3--Non-load-based Missing Data Procedure for NOX-Diluent CEMS and NOX Concentration CEMS
----------------------------------------------------------------------------------------------------------------
Trigger conditions Calculation routines
----------------------------------------------------------------------------------------------------------------
Duration (N) of CEMS
Monitor data availability (percent) outage (hours)\1\ Method Lookback period
----------------------------------------------------------------------------------------------------------------
95 or more........................... N [le] 24 Average................. 2160 hours*
N 24 90th percentile......... 2160 hours*
90 or more, but below 95............. N [le] 8 Average................. 2160 hours*
N 8 95th percentile......... 2160 hours*
80 or more, but below 90............. N 0 Maximum value........... 2160 hours*
Below 80, or operational bin N 0 Maximum NOX emission None
indeterminable. rate or maximum
potential NOX
concentration.
----------------------------------------------------------------------------------------------------------------
* If operational bins are used, the lookback period is 2,160 quality-assured, monitor operating hours, and data
at the corresponding operational bin are used to provide substitute data values. If operational bins are not
used, the lookback period is the previous 2,160 quality-assured monitor operating hours. For units that report
data only for the ozone season, include only quality-assured monitor operating hours within the ozone season
in the lookback period. Use data from no earlier than three years prior to the missing data period.
\1\During unit operation.
Table 4--Non-load-based Missing Data Procedure for Flow Rate CEMS
----------------------------------------------------------------------------------------------------------------
Trigger conditions Calculation routines
----------------------------------------------------------------------------------------------------------------
Duration (N) of CEMS
Monitor data availability (percent) outage (hours)\1\ Method Lookback period
----------------------------------------------------------------------------------------------------------------
95 or more........................... N [le] 24 Average................. 2160 hours*
N 24 The greater of:.........
Average................. HB/HA
90th percentile......... 2160 hours*
90 or more, but below 95............. N [le] 8 Average................. 2160 hours*
[[Page 271]]
N 8 The greater of:.........
Average................. HB/HA
95th percentile......... 2160 hours*
80 or more, but below 90............. N 0 Maximum value........... 2160 hours*
Below 80, or operational bin N 0 Maximum potential flow None
indeterminable. rate.
----------------------------------------------------------------------------------------------------------------
* If operational bins are used, the lookback period is the previous 2,160 quality-assured, monitor operating
hours and data at the corresponding operational bin are used to provide substitute data values. If operational
bins are not used, the lookback period is the previous 2,160 quality-assured, monitor operating hours. For
units that report data only for the ozone season, include only quality-assured monitor operating hours within
the ozone season in the lookback period. Use data from no earlier than three years prior to the missing data
period.
\1\ During unit operation.
[58 FR 3701, Jan. 11, 1993, as amended at 60 FR 26529, May 17, 1995; 61
FR 25582, May 22, 1996; 64 FR 28602, May 26, 1999; 67 FR 40434, June 12,
2002; 67 FR 53505, Aug. 16, 2002; 67 FR 57274, Sept. 9, 2002]