In accordance with the Regional Haze Rule (40 CFR 51.308(i)(3)), the
Pennsylvania Department of Environmental Protection (DEP) is making
available the comments received by the Federal Land Managers on the
draft proposed Regional Haze (RH) State Implementation Plan (SIP)
revision, and the DEP’s responses to those comments.  The following is
a reproduction of a comment letter received by the Department from the
United States Fish and Wildlife Service (FWS) and the National Park
Service (NPS), with DEP‘s responses inserted:

U.S. Fish and Wildlife Service and National Park Service Comments

Pennsylvania Draft Regional Haze Rule State Implementation Plan

August 2, 2010

On April 22, 2008, the Commonwealth of Pennsylvania (PA) submitted a
draft Regional Haze Rule State implementation plan (SIP), pursuant to
the requirements codified in federal rule at 40 CFR 51.308(i)(2), to the
U.S. Department of the Interior, U.S. Fish and Wildlife Service (FWS)
and National Park Service (NPS).  Shortly after receiving the draft SIP,
the FWS and NPS Air Quality staff discussed concerns with the draft
document during a telephone conference attended by PA Department of
Environmental Protection (PADEP) staff.  Subsequently, PA decided that
the FWS and NPS concerns would be considered and another draft document
would be sent to FWS and NPS at a future date.

We received the new draft on June 2, 2010.  The FWS Branch of Air
Quality and the NPS Air Resources Division staffs have conducted a
substantive review of this most recent draft and our comments follow.

We look forward to the PADEP response as per section 40 CFR
51.308(i)(3), and as always, we are willing to work with the PADEP staff
towards resolving the issues discussed below. For further information,
please contact Tim Allen (FWS) at (303) 914-3802 or Pat Brewer (NPS) at
(303) 969-2153.

Overall Comments

1.  Comment:  We commend PA for working with us to revise its April 2008
draft SIP to address many of the original concerns identified by
previous consultation with the FWS and NPS.  There is significant
improvement in this revised draft in comparison to the previous
document.  

DEP Response:  The Department appreciates the reviewer’s assessment
that the Department made significant improvement in developing the
revised draft proposed Regional Haze (RH) State Implementation Plan
(SIP) revision.

2.  Comment:  We continue to have concern over PA’s approach towards
meeting the controls identified by the Mid Atlantic Northeast Visibility
Union (MANE-VU) Ask (“Ask”).  Although we appreciate the
Commonwealth’s need to thoroughly review the effects of adopting
additional major air quality controls, MANE-VU-generated progress goals
are dependent on successful adoption and implementation of these
controls.  We believe that the Commonwealth‘s contribution to the
success of the “Ask” is critical to achieving success in the region.
 

DEP Response:  The Department intends to pursue the control strategies
identified in the MANE-VU “Ask” in accordance with Pennsylvania law.
 Page 95 of the proposed RH SIP revision contains the following
paragraph:  “MANE-VU identified emissions from 167 stacks at EGU
facilities as having visibility impacts in MANE-VU Class I areas that
make controlling emissions from those stacks crucial to improving
visibility at MANE-VU Class I areas.  MANE-VU’s agreed regional
approach for this source sector is to pursue a 90 percent control level
on SO2 emissions from these 167 stacks by 2018 as appropriate and
necessary.  MANE-VU has concluded that pursuing this level of sulfur
reduction is both reasonable and cost-effective.  Table 10.4.4-1 below
lists the EGU stacks in Pennsylvania identified on the list of 167
stacks in all of MANE-VU, with their current and anticipated controls,
the permit status of the SO2 controls, and the anticipated SO2 emissions
reductions achieved by 2018.”  Additionally, page 97 of the proposed
RH SIP revision states the following:  “Based on the controls
proposed, constructed and under construction in Pennsylvania, the
Department has concluded that at this time, the Department’s CAIR
regulation constitutes a reasonable measure for EGUs in Pennsylvania. 
Therefore, consistent with the MANE-VU ‘Ask’ Statement, the
Department projects a 94.5% reduction in SO2 emissions from the EGUs
listed above in Table 10.4.4-1 (or from alternative measures, as
appropriate and necessary).”  Also, page 120 of the proposed RH SIP
revision states the Department has developed a proposed low-sulfur
heating and distillate oil regulation, in response to the MANE-VU
‘Ask’ low-sulfur fuel oil strategy.  The Environmental Quality Board
(EQB) adopted the Department’s proposed low-sulfur fuel oil regulation
at its July 13, 2010, meeting.  The public comment period began on
September 25th and ends on November 29th, 2010.  Additionally, the
Department has revised the proposed RH SIP to address the status of the
regulation development that pertains to the source categories considered
in the Mactec Report (Appendix O).  (Please see page 83 of the proposed
RH SIP).  These examples clearly indicate the Commonwealth‘s
dedication to pursue the goals of the “Ask” in order to achieve
success in the region.

3.  Comment:  PA’s Reasonable Progress consists of “pursuing, as
appropriate and necessary, the four goals of the “Ask” Statement”.
 We are concerned that commitments are only to “pursue” emission
reductions, and there is no commitment to an enforceable mechanism that
would ensure that the Commonwealth achieves these emission reductions.  

DEP Response:  It is not the Department’s practice to commit to
adopting measures that must go through the notice and comment process
under the Pennsylvania Air Pollution Control Act (APCA).  Furthermore,
the EQB has the statutory authority to adopt regulations, not the
Department.  In addition to plan approvals, regulations undergo the
review of several advisory committees before EQB action, and undergo
consideration by the Independent Regulatory Review Commission (IRRC) as
well as the standing committees of the General Assembly after EQB
consideration.  A commitment to adopt could nullify the value of a
public comment and board review process.  The Department intends to
pursue these goals through the regulatory process in accordance with
Pennsylvania law.

4.  Comment:  We complement PADEP’s discussion of the five factors in
the Best Available Retrofit Technology (BART) analyses.  However, it was
not clear how PADEP applied the results of those five steps in making
its final BART determination because all of the BART determinations led
to conclusions that no additional controls represented BART, even when
the five factor analysis led to results that have been accepted as BART
by many other states.  No indications were given, nor could any be
derived from the information presented, as to what level of cost and
benefit, if any, would be acceptable as BART by PADEP.

DEP Response:  In making BART determinations for the sources subject to
BART in Pennsylvania, the Department did not establish or utilize bright
line thresholds for cost or for visibility improvement.  Instead, the
Department employed an approach that considered the multiple BART
Guideline factors.  As a result, sources with a higher degree of
potential visibility improvement from control would justify higher cost
controls.  Conversely, only low cost controls would be justified for
sources with a lower degree of potential visibility improvement.  The
BART determinations are consistent with the BART Guidelines.

5.  Comment:  PA does not specifically identify the inconsistency of
using MANE-VU based Reasonable Progress Goal calculations.  These final
runs are based on Clean Air Interstate Rule (CAIR) and MANE-VU “Ask”
control assumptions.  These controls are presently not realized, nor are
there commitments with implementation plans specified to accomplish
these controls in the SIP.  Therefore, more information should
supplement the document that fully describes the uncertainty and whether
the Commonwealth or the Regional Planning Organization has any efforts
planned (or in progress) to minimize these uncertainties.  Specifically
addressing these future estimates with more specific projections in the
Commonwealth’s 2013 mid-term review is imperative.

DEP Response:  Plan approvals for controls at a permitted source are
subject to the notice and comment provisions under the APCA.  A
commitment to adopt controls could nullify the value of a public comment
and board review process.  As stated in a previous response above, it is
not the Department’s practice to commit to adopting measures that must
go through a public comment process.  The EQB has the statutory
authority to adopt regulations, not the Department.  In addition to plan
approvals, regulations undergo the review of several advisory committees
before EQB action, and undergo consideration by the IRRC as well as the
standing committees of the General Assembly after EQB consideration. 
The Department intends to re-evaluate the projections of SO2 reductions
due to CAIR and the “Ask” control assumptions at the time of
Pennsylvania’s first periodic report to determine whether the
predicted reductions are realized.  The reasonable progress report will
evaluate the progress made toward the reasonable progress goal for each
Class I area located outside Pennsylvania that may be affected by
emissions from within Pennsylvania.

Specific Comments

Section 5.0, Baseline and Natural Conditions

Section 6.0, Monitoring Strategy

6.  Comment:  PA does not have Class I areas, but does list Class I
areas in the MANE-VU and Visibility and Tribal Association of the
Southeast (VISTAS) States to which air pollution emissions from PA’s
sources contribute towards visibility impairment.  It would be helpful
to include a brief summary of IMPROVE data for the Class I areas that PA
influences for the 20% worst and 20% best days.  These data are
available at the VIEWS website (http://views.cira.colostate.edu/web/). 
It is not clear if these data are present in section 6.2 that was
omitted to reduce file size.  It is also not clear if Section 6.4 is
intended to meet the requirement to commit to future monitoring.

DEP Response:  Pages 23-24 of the proposed RH SIP revision include
summary tables that display the IMPROVE data that shows baseline
visibility for the 20% worst and the 20% best visibility days based on
the five-year average for 2000-2004, natural background visibility for
the 20% worst and the 20% best visibility days, and the difference
between baseline and natural visibility conditions for each MANE-VU
Class I area and for nearby Class I areas.  The Department considers the
IMPROVE monitors located at Class I areas in MANE-VU to be
representative of those Class I areas and the monitoring data from those
monitors to be a reasonable basis for assessing progress toward the
regional haze program goals.  Section 6.2 that was omitted to reduce
file size contains photographic information about the MANE-VU Class I
area IMPROVE monitors, and will be inserted back into the proposed RH
SIP revision for the public comment period.  Section 6.4 describes
additional monitoring information the Department plans to use to improve
its understanding of visibility impairment and to assess visibility and
fine particle pollution in the region.  Because there are no Class I
areas located in Pennsylvania, a monitoring plan is not required under
40 CFR § 51.308(d)(4).

Section 8.0, BART

7.  Comment:  Due to the number of BART Sources in PA, the reviews were
divided among the separate NPS and FWS Air Quality Offices.  Please
refer to Attachment 1 for the NPS BART review and Attachment 2 for the
BART review of the sources done by FWS.  As noted above, it was not
clear to us how PADEP concluded what emission controls were feasible and
cost-effective, and those that were not.

DEP Response:  The Department did not establish or utilize bright line
thresholds for cost or for visibility improvement in making BART
determinations for the sources subject to BART.  The Department
considered all of the BART Guideline factors, and determined that a BART
source analysis resulting in significant visibility improvement from
controls would consequently justify higher cost controls.  The
Department maintains that a low $/ton control cost should not supersede
an excessively high cost of visibility improvement.  The Department does
not believe that requiring additional controls on a BART source based
solely on a high cost of visibility improvement for very small
visibility improvement, albeit a low $/ton control cost, is justified or
appropriate.

Section 9.0, Reasonable Progress Goals

8.  Comment:  PA and other MANE-VU states are using the modeled results
for the “Ask” to set reasonable progress goals for 2018.  It would
be preferable to set reasonable progress goals based on the On the Books
(OTB)/On the Way (OTW) assumptions.  The MANE-VU scenario is problematic
because it includes several control assumptions that are not reasonably
OTW to implementation in the MANE-VU, MWRPO (Midwest Regional Planning
Organization), and VISTAS states.  Most problematic is MANE-VU’s
choice to unilaterally increase Electric Generating Unit (EGU) emissions
in MRPO and VISTAS States without consideration of legally enforceable
controls already in place in those States.  Also of concern are the
assumptions that MANE-VU States will reduce SO2 from non-EGU sectors
though low sulfur fuel requirements by 2018 and that MWRPO and VISTAS
States will reduce SO2 from their non-EGU sectors by comparable amounts.
 These reductions are not in progress.  PA needs to address this
discrepancy in its discussion of Reasonable Progress, and it is good to
read that PA will also be addressing the differences between reality and
reasonable progress goals at the mid-course review.  The discussion PA
supplied regarding the differences between the MANE-VU and VISTAS
control assumptions used in setting 2018 reasonable progress goals was
helpful.  PA should discuss how uncertainty in the federal CAIR program
could affect the reasonable progress goals for the affected Class I
areas.  Please clarify whether PA’s CAIR rule requires specific SO2
and NOx reductions independent of the status of the federal CAIR rule.

DEP Response:  The MANE-VU “Ask” modeling scenario ensures that the
modeled results are indicative of the agreed upon strategy by the
MANE-VU states that is reflected in the “Ask” statement.  MANE-VU
accounted for emission controls already in place and emission controls
that are not yet finalized but likely to achieve additional reductions
by 2018 in the “Ask” (beyond on the way) emissions modeling
inventory for 2018.  The Department relied on technical analyses,
including the modeling scenario, developed by MANE-VU to demonstrate
that Pennsylvania’s emission reductions, when coordinated with those
of other states and tribes, are sufficient to achieve the reasonable
progress goals in Class I areas affected by Pennsylvania’s emissions. 
The Department believes these modeled results are reasonable to achieve
the reasonable progress goals in Class I areas affected by
Pennsylvania’s emissions.  The Department has developed a proposed
low-sulfur heating and distillate oil regulation, in response to the
MANE-VU “Ask” low-sulfur fuel oil strategy.  Pennsylvania has
proposed a single strategy for the state.  Pennsylvania is pursuing a
strategy that is not less stringent than the outer zone strategy and one
that would meet the sulfur content emission limits listed above by 2018.
 Please see the modifications to page 120 in the proposed RH SIP
revision for the updated status of this low sulfur fuel oil regulation. 
CAIR has been remanded, but not vacated and remains in place until
replaced.  Pennsylvania’s CAIR regulation will require emissions
reductions prior to implementation of EPA’s Transport Rule (TR), when
final.  The proposed TR is the replacement rule to the remanded federal
CAIR rule.  EPA is expected to finalize the TR in mid 2011; EPA also
anticipates promulgating a second TR at the same time and finalizing it
in mid-2012.  

Section 10.0, Long Term Strategy

9.  Comment:  The Long Term Strategy Section is well written in that it
defines the MANE-VU “Ask” and the major contributions to visibility
impairment.  There is a good discussion of related regulations and
enforcement actions that will result in substantive emissions reductions
in the Commonwealth.  Although PA has committed to address this in the
mid-course review, this SIP should still discuss the Commonwealth’s
efforts to implement the low sulfur fuel strategy of the MANE-VU
“Ask”.  We concur with PA’s assessment that prescribed fire is a
comparatively small contributor to visibility impairment in its
geographic region and that residential wood smoke is the more important
contributor.  Does PA have plans to address residential wood smoke?

DEP Response:  The Department has developed a proposed low-sulfur
heating and distillate oil regulation, in response to the MANE-VU
‘Ask’ low-sulfur fuel oil strategy.  Pennsylvania has proposed a
single strategy for the state.  Pennsylvania is pursuing a strategy that
is not less stringent than the outer zone strategy and one that would
meet the sulfur content emission limits listed above by 2018. 
Consideration of the regulation is anticipated to proceed on a schedule
in accordance with the Pennsylvania Regulatory Review Act, the
Pennsylvania Air Pollution Control Act, the Commonwealth Documents Law,
the Commonwealth Attorneys’ Act and the procedures established by the
Commonwealth’s Environmental Quality Board and by the Department.  The
proposed RH SIP revision has been modified on page 120 to include the
following:  “The EQB adopted the Department’s proposed low-sulfur
fuel oil regulation at its July 13, 2010, meeting.  The public comment
period began on September 25th and ends on November 29th, 2010.  The
proposed regulation would reduce the allowable sulfur content limits of
commercial fuel oil to 15 parts per million (ppm) for Number (No.) 2 and
lighter commercial fuel oils and to 0.25% sulfur content by weight for
No. 4 commercial fuel oil, and 0.5% sulfur content by weight for No. 5,
6 and heavier commercial fuel oils beginning May 1, 2012.”  This
statement will be updated at the time of final SIP submittal to account
for further progress.

Regarding residential wood smoke, the proposed RH SIP revision has been
modified on page 83 to include the following:  “New indoor wood stoves
in Pennsylvania are regulated by EPA’s Residential Woodstoves NSPS. 
The NSPS for residential woodstoves was part of the area source
inventory developed by MANE-VU to model the 2018 inventory for area
sources (see page 106 of this SIP revision).  A final-form outdoor wood
boiler (OWB) regulation was approved by the Commonwealth’s Independent
Regulatory Review Commission (IRRC) on August 19, 2010.  The
Department’s final-form regulation would require EPA’s Phase 2
emission standards, including the particulate matter standard of 0.32
lb/MMBtu, for all new OWBs in the Commonwealth manufactured after June
1, 2011.”   This statement will be updated at the time of final SIP
submittal to indicate publication date of the OWB regulation as a final
regulation.



Attachment 1

To FWS/NPS Comments – Pennsylvania Draft Regional Haze SIP

NPS Best Available Retrofit Technology Comments

General Comments

10.  Comment:  We complement PA DEP for the clarity of its discussions
of how it applied the five factors in the BART analyses in making its
BART determinations.  However, it was not clear how PA DEP applied the
results of those five steps in making its final BART determinations
because all of the BART determinations led to conclusions that no
additional controls represented BART, even when the five factor analysis
led to results that have been accepted as BART by many other states.  No
indications were given, nor could any be derived from the information
presented, as to what level of cost and benefit, if any, would be
acceptable as BART by PA DEP.

DEP Response:  The Department did not establish or utilize bright line
thresholds for cost or for visibility improvement in making BART
determinations for the sources subject to BART.  The Department
considered all of the BART Guideline factors, and determined that a BART
source analysis resulting in significant visibility improvement from
controls would consequently justify higher cost controls.  The
Department maintains that for a BART determination, a low $/ton control
cost should not supersede an excessively high cost of very low
visibility improvement.

11.  Comment:  The core purpose of the BART program is to improve
visibility in federal Class I areas, and BART is not necessarily the
most cost-effective solution.  Instead, BART represents a broad
consideration of technical, economic, energy, and environmental
(including visibility improvement) factors.  We believe that it is
essential to consider both the degree of visibility improvement in a
given Class I area as well as the cumulative benefits of improving
visibility across all of the Class I areas affected.

There are several Class I areas impacted by Pennsylvania’s BART
sources.  We believe that it is appropriate to consider both the degree
of visibility improvement in a given Class I area as well as the
cumulative effects of improving visibility across all of the Class I
areas affected.  The same metric should not be used to evaluate the
effects of reducing emissions from a BART source that impacts only one
Class I area as for a BART source that impacts multiple Class I areas. 
Also, evaluating impacts at one Class I area, while ignoring others that
are similarly significantly impaired, should not be done.  Emissions
savings from a source are benefits that will be spread well beyond only
the most-impacted Class I area, and should be considered.  While
Pennsylvania presented data describing improvements to visibility at a
specific Class I area that would result from the various control
scenarios it investigated, the Commonwealth has not explained how it
incorporated this information about impacts upon all Class I areas into
its BART decision.

For example, Wyoming evaluated cumulative visibility improvement for
both its BART and reasonable progress determinations—following are
excerpts from those Wyoming determinations (with emphasis added):

• Visibility impacts were addressed in a comprehensive visibility
analysis covering all

three visibility impairing pollutants and associated control options. 
The cumulative 3-

year averaged visibility improvement from the baseline summed across the
three ClassI areas achieved with LNB with separated OFA, upgraded wet
FGD, and FGC for enhanced ESP (Post-Control Scenario A) was 1.070 _dv
from Unit 1, 0.199 _dv from Unit 2, 1.068 _dv from Unit 3, and 0.892 _dv
from Unit 4.1

• Visibility impacts were addressed in a comprehensive visibility
analysis covering all three visibility impairing pollutants and
associated control options.  The cumulative 3-

year averaged 98th percentile visibility improvement from the baseline
summed across

all four Class I areas achieved with LNB with advanced OFA, dry FGD, and
a new fullscale fabric filter, Post-Control Scenario A for each unit,
was 3.558 _dv from Unit 3 and 1.963 _dv from Unit 4.2

• Visibility impacts were addressed in a comprehensive visibility
analysis covering three visibility impairing pollutants and the
associated control options. The cumulative

visibility improvement as compared to the baseline across Wind Cave NP
and

Badlands NP achieved with new LNB with OFA at the 30-day limit of 0.23
lb/MMBtu

(based on the 98th percentile modeled results) was 0.14 _dv from each of
the three units.  The expected visibility improvement over the course of
a full annual period would be even greater due to the annual BART limit
that is based on 0.19 lb/MMBtu.3

• Visibility impacts were addressed in a comprehensive visibility
analysis covering all

three visibility impairing pollutants and associated control options. 
The cumulative 3-

year averaged 98th percentile visibility improvement from the baseline
summed across

both Class I areas achieved with LNB with advanced OFA, wet FGD, and
existing ESP with FGC (Post-Control Scenario A) was 1.716 _dv from Unit
1 and 1.934 _dv from Unit 2. 4

• Visibility impacts were addressed in a comprehensive visibility
analysis covering all

three visibility impairing pollutants and associated control options. 
The cumulative 3-

year averaged 98th percentile visibility improvement from the baseline
summed across

both Class I areas achieved with LNB with advanced OFA, upgrading the
existing dry

FGD, and a new full-scale fabric filter, Post-Control Scenario A for
Unit 1, was 0.996 _dv.5

Oregon considered cumulative benefits for the Boardman Power Plant SCR
addition for

reasonable progress:

Table 22: Visibility Modeling Results (percent improvement) 

Total visibility impacts (sum of 98th percentile for all Class I areas)

The BART guidelines recommend analyzing visibility improvement for the
highest

impacted Class I area with the assumption that any improvement in the
worse impacted

area would result in improvement in the lesser impacted areas.  However,
since the

Boardman Plant significantly impacts 14 Class I Areas within 300
kilometers, the

Department tried to include other parameters that would assess the
significance of

the improvements for all Class I areas impacted.  Therefore, the
Department added the number of Class I areas with impacts greater than
1.0 delta deciview, the total delta deciviews for all Class I areas
(98th percentile), and the average delta deciview for all Class I areas
(98th percentile).  As can be seen in Table 21, any one of the
parameters is fairly representative of the other parameters perhaps with
the exception of WFGD.  Given these results, the Department does not
believe that adding additional parameters, such as total deciview days,
would result in any other conclusions and would probably just add
confusion to the analysis (e.g., more days of impacts than are in a
year).  Using the results of the visibility modeling, the cost
effectiveness of the control technologies is recalculated by relating
the costs to deciview improvement (Mt. Hood and all Class I areas) as
shown in the following 2 tables.6

Pennsylvania has ignored the other Class I areas where a given BART
source is also causing or contributing to visibility impairment.  The
dollar cost per increment of visibility improvement would be
substantially lower if full consideration is given to all affected Class
I areas that would benefit from emission reductions. While we recognize
that EPA has provided no guidance on this issue of assessing visibility
benefits that would result in multiple Class I areas when emissions are
reduced from a given BART source, we commend Wyoming and Oregon for
their initiative in addressing the issue. We also recognize that there
is no “perfect” method for addressing cumulative benefits, but we
firmly believe that Pennsylvania must show how it considered the
cumulative impact of the BART sources the affected Class I areas.  We
have suggested an approach to Pennsylvania that is consistent with
available information and with the approach used by Wyoming and Oregon,
and again request that Pennsylvania show how it has considered the
cumulative benefits of potential BART reductions.

DEP Response:  The EPA’s Guidelines for BART Determinations, Final
Rule (40 CFR Part 51, July 6, 2005) does not stipulate a requirement for
a “cumulative” impact analysis from one BART source on multiple
Class I areas.  Therefore, and as stated by the commentator, since EPA
has provided no guidance on the issue of assessing visibility benefits
that may result in multiple Class I areas when emissions are reduced
from a given BART source, the Department maintains that such an analysis
is unwarranted.  

The Department also maintains that requiring additional controls on a
BART source to provide very small visibility improvements at multiple
Class I areas due to an apparent low $/deciview visibility improvement
cost is nether justified or appropriate.  The summation of very small
delta deciviews serves the mathematical purpose of increasing the delta
deciview value with respect to the $/delta deciview equation and thereby
lowering the apparent cost of controls per delta deciview, while not
providing any significant visibility improvement to any of the affected
Class I areas.  The examples of BART source impacts described in the
comment are examples of large deciview visibility impacts at many Class
I Areas within 300 kilometers of the Wyoming and Oregon BART sources. 
Very small delta deciview values from a BART source should not be summed
to determine the cumulative effect on all lesser affected Class 1 areas.
 EPA’s BART guidelines recommend analyzing visibility improvement for
the highest impacted Class I area, with the assumption that any
improvement in the worse impacted area would result in improvement in
the lesser impacted areas.  The Department applied the guidelines for
the determination of BART for the affected sources. 

12.  Comment:  Based upon our reviews of BART analyses across the U.S.,
we believe that cost-per-deciview ($/dv) of visibility improvement is
the most-common and most-useful parameter for assessing the
cost-effectiveness of strategies to improve visibility in Class I areas.
 Our compilation7 of BART analyses across the U.S. reveals that the
average cost/dv proposed by either a state or a BART source is $13 - $20
million,8 with a maximum of almost $50 million/dv proposed by Colorado
at the Martin Drake power plant in Colorado Springs.

DEP Response:  The Department agrees that the cost per-deciview of
visibility improvement is the most-common and most-useful parameter for
assessing the cost-effectiveness of strategies to improve visibility in
Class I areas, because the Regional Haze program was designed to improve
visibility in Class I areas, which is measured in deciviews.  However,
the Department did not establish or utilize bright line thresholds for
cost or for visibility improvement in making BART determinations.  The
Department considered all of the BART Guideline factors, and determined
that a BART source analysis resulting in significant visibility
improvement from controls would consequently justify higher cost
controls.  It should also be noted that the $/deciview cost in the
example provided in the comment is an incremental cost effectiveness
rather than the average cost effectiveness.

13.  Comment:  Comments on the BART determinations for individual
facilities that are subject to BART follow.  We are focusing our
comments on the BART determinations for the cement and paper facilities
because they have larger impacts than the other BART sources. (Also,
Pennsylvania is the only state that has not proposed Selective
Non-Catalytic Reduction (SNCR), or any additional controls, for all of
its BART cement kilns.)  We are also providing comments on some of the
other BART sources.  More detailed comments regarding our BART reviews
follow.

DEP Response:  The Department notes that the baseline visibility impacts
from the cement kilns in other states are significantly higher than
those from kilns in Pennsylvania.  The total maximum 24-hour visibility
impacts, due to all pollutants for all BART sources at a facility, from
the cement kilns in Pennsylvania range from 0.0008dv to 0.608dv.  For
each individual BART stack, the maximum 24-hour visibility impact due to
NO3 (nitrates) range from 0.03dv to 0.39dv.  Therefore it was
appropriate to require SNCR control systems as BART for the cement kilns
in other states.



14.  Comment:  AK Steel Corporation, Butler Works

BART 5 Factor Analysis:

STEP – 1: Identify All Available Retrofit Control Technologies

PA DEP: Ultra low NOx burners is the available retrofit control option
with the practical potential for application to the miscellaneous
natural gas burners for the control of NOx.

NPS: PA DEP should have included SCR in its analysis for the annealing
furnace.9

STEP – 4: Evaluate Impacts and Document the Results

PA DEP: Cost of Compliance: Ultra Low NOx Burner $12,800/ton, the
Annualized Cost is $520,000. These calculations are based upon
information obtained from EPA’s AP42 Manual. The potential emissions
reduction for this control was estimated to be 41 tons.

NPS: PA DEP should better explain these estimates.

DEP Response:  Based on NESCAUM modeling results on A.K Steel
facility’s NOx emissions, the following are the visibility impacts on
the Class 1 areas:

Class I area	Visibility Impact due to all sources	Visibility Impact due
to EAF	Visibility Impact due to remaining sources

Shenandoah	0.041 dv	0.034 dv	0.007 dv

Dolly Sods     	0.035 dv	0.024 dv	0.011 dv

Otter Creek    	0.047 dv	0.036 dv	0.011 dv

Total Visibility Impacts	0.123 dv	0.094 dv	0.029 dv



As acknowledged in the comment, there is no feasible NOx control
technology for EAFs.

As per AK Steel's submittal, the facility recently installed ULNB on two
process furnaces.  For the six processes that could utilize similar
ULNBs the average cost would be $3.2 million dollars as per the AK Steel
submittal to replace all of the burners.  The NOx reduction from the
installation of ULNBs at a 65% reduction would be 41 tons per year as
per the AK Steel submittal.  AK Steel arrived at the 65% reduction
factor by comparing the uncontrolled and ULNB emission factors depicted
in AP-42, Chapter 1, Natural Gas Combustion.   AK Steel provided an
explanation of their recent costs and the averaging process that they
used to estimate their possible future costs on page 3 of their January
12, 2007 submittal.  AK Steel provided the NESCAUM modeling results in
Table 1 of their submittal.  AK Steel provided the necessary source
information for all of their sources in Table 2 of their submittal.  AK
Steel provided a cost summary estimation in Table 3 of their submittal. 
The Department concurs with the technical and cost analysis that AK
Steel provided.  When summed the total existing deciview effect for the
three Class 1 areas from AK Steel's remaining controllable sources is
0.029 deciviews.  However, the Department does not believe that the
thousandths of a deciview improvement that could be achieved at any
individual Class 1 area would provide true visibility relief, and that a
summation is not warranted.

AK Steel did not evaluate SCR as an option.  SCR may be technically
feasible, but it is not economically feasible.  It is the Department's
experience that SCR will typically cost more than ULNB, usually at least
50% more.  The BART determination is consistent with the BART
Guidelines.  The Department considered all of the BART Guideline
factors, and determined that a BART source analysis resulting in
significant visibility improvement from controls would consequently
justify higher cost controls.  The maximum modeled deciview impact due
to this BART source on the most impacted Class I area is considerably
less than 0.5 dv, the threshold suggested in the BART Guidelines to
determine if a source contributes to visibility impairment.

15.  Comment:  Allegheny Ludlum Corporation, Brackenridge Facility

STEP – 4: Evaluate Impacts and Document the Results

PA DEP: Cost of Compliance: (2) Ultra Low NOx Burner (for Loftus Soaking
Pits):

$12,800/ton, the Annualized Cost is $182,000. These calculations are
based upon information obtained from EPA’s AP42 Manual. The potential
emissions reduction for this control was estimated to be 14 tons.

NPS: PA DEP should better explain these estimates.

DEP Response:  In its May 4, 2009, “Review Memo for BART Application
for the Allegheny Ludlum Corporation Brackenridge Facility”, the
Allegheny County Health Department (ACHD) conducted an engineering
analysis under the BART determination process of 40 CFR 51.308(e) for
each source subject to BART at Allegheny Ludlum Brackenridge.  The
review memo was submitted to the Department as part of its BART
submittal for Pennsylvania.  The $12,800/ton estimated cost of Ultra Low
NOx Burners is based on the costs associated with the recent
installation of ULNB at a similar facility in Pennsylvania, the AK Steel
Corporation, Butler Works, which recently installed ULNB on two process
furnaces.  A 65% emission reduction factor for ULNBs was used in the
ACHD calculation.  This same number was used in the BART analysis by the
similar facility in Pennsylvania noted above.  As indicated by the PA
DEP in its submittal for that facility, the number was derived from
AP-42, Chapter 1, Natural Gas Combustion.  ACHD used the 65% reduction
effectiveness figure to be consistent with the PA DEP use of that value
in its AK Steel BART analysis.  The $182,000 “Annualized Cost” is
arrived at as follows:  The value of the 2002 NOx emission inventory
used in NESCAUM modeling for the Loftus Soaking Pits Numbers 9 through
23, and Numbers 43 through 46, is 21.9 tons per year.  The NOx reduction
from the installation of the ULNB s is figured at 65% reduction factor,
resulting in emissions reductions of 14.2 tpy.  14.2 tpy, at $12,800 per
ton NOx removed, results in the $182,000 annualized cost ($181,760).

The BART determination is consistent with the BART Guidelines.  The
Department considered all of the BART Guideline factors, and determined
that a BART source analysis resulting in significant visibility
improvement from controls would consequently justify higher cost
controls.   The maximum modeled deciview impact due to this BART source
on the most impacted Class I area is considerably less than 0.5 dv, the
threshold suggested in the BART Guidelines to determine if a source
contributes to visibility impairment.

16.  Comment:  Appleton Papers Inc. /Spring Mill

BART 5 Factor Analysis:

STEP – 1: Identify All Available Retrofit Control Technologies

NPS: PA DEP is incorrect in omitting Flue Gas Recirculation as a NOX
control option.10

STEP – 2: Eliminate Technically Infeasible Options

NPS: PA DEP is incorrect in stating that wet scrubbing is technically
infeasible due to the size of Power Boiler #3. While the size of the
boiler, and its emissions, will affect the economic feasibility of
installing a wet scrubber, it will not affect the technical feasibility.
Georgia Pacific is installing a wet scrubber on Power Boiler #4 at its
Big Island, Virginia, paper mill. Mead Westvaco is upgrading the wet
scrubber on the power boilers at its Covington, Virginia, paper mill.

STEP – 4: Evaluate Impacts and Document the Results

PA DEP: It was determined that the cost of control for this device was
not cost effective considering the commensurate visibility improvement.

Power Boiler #3 (SO2) Cost Effectiveness ($/ton) Cost of Visibility
Improvement ($/dv) $3,487/ton $51,200,000/dv

NPS: PA DEP should better explain these estimates.

STEP – 5: Evaluate Visibility Impacts

PA DEP: The total deciview impact of this facility, including all BART
eligible units, was modeled to be 0.089 dv. The cost in terms of dollars
per deciview for installing a dry flue gas desulfurization system at
this facility was calculated to be $51,200,000/dv.

Several control options were considered for SO2 control for the power
boiler. The most cost effective means of control considered was an SDA
system. The resulting average cost effectiveness for installing the SDA
system based on 2002 emissions was calculated to be $3,487/ton. The
corresponding visibility improvement based on the installation of this
technology over the baseline was estimated to be 0.048 dv. The cost of
this control equipment in terms of visibility improvement was determined
to be $51,200,000/dv.

NPS: PA DEP should better explain these estimates.

DEP Response:  The Department received updated data from NESCAUM that
modeled a 0.087 total deciview effect from Appleton's power boiler,
0.051 deciview from SO2 and 0.057 from NOx, with these numbers not
adding up do to non-linearity of CALPUFF modeling.  In Appendix A of
Appleton's August 2007 BART submittal they cited "Sargent and Lundy Dry
Flue Gas Desulfurization; Costs for retrofit Units.  Project 11311-000.
9/26/02."  The levelized annual costs in Appleton's submittal came out
to be $865,779.  The Department calculates that at 100 % efficiency the
$865,779 / 0.051 deciview = $16,976,450 per delta deciview.  The actual
delta deciview will be significantly lower than 100% depicted here. 
Appleton's power boiler burns a combination of coal, bark and sludge as
does the No. 5 boiler at the Big Island, Virginia Georgia Pacific
facility that is referenced in this comment.  BART control for this
multi-fuel boiler was determined to be limits on the % sulfur content of
the coal and limits on the percentage of coal versus the total amount of
other fuel burned. The effectiveness of FGD was deemed to be
questionable because of the multiple fuels.  The No. 4 Georgia Pacific
boiler that installed wet FGD burns only coal.  The No. 4 boiler is a
280 MMBtu/hr boiler where wet FGD is typically used.  The Appleton
boiler is 180 MMBtu/hr where wet FGD is typically not used.  The
required BART efficiency on the Georgia Pacific boiler FGD is 90% which
may be achieved by dry FGD as Appleton proposed for this evaluation. 
The Department concurs that FGR may be a technically feasible control
option for a stoker boiler.  However, the Department expects that a NOx
cost analysis divided by some fraction of the total 0.057 deciview
effect will yield similar results as the SOx calculations shown here. 
The Appleton boiler already has over fired air for NOx control and
through material balance has indicated an approximate 40% existing SO2
control rate.  The Appleton facility has agreed to reduce their
permitted allowable SO2 emission rate from 4.0 to 3.2 lbs/MMBtu.  The
Department proposes that these requirements represent BART for the
Appleton Facility especially when the fact that only a few hundredths of
a delta deciview improvement could be achieved.

The BART determination is consistent with the BART Guidelines.  The
Department considered all of the BART Guideline factors, and determined
that a BART source analysis resulting in significant visibility
improvement from controls would consequently justify higher cost
controls.   The maximum modeled deciview impact due to this BART source
on the most impacted Class I area is less than 0.5 dv, the threshold
suggested in the BART Guidelines to determine if a source contributes to
visibility impairment.

17.  Comment:  Essroc Cement Corporation

Step 2 – Eliminate Technically Infeasible Options.

PA DEP: For Selective Non-Catalytic Reduction (SNCR), ammonia or urea
would be injected into the rotating kiln at a location where the gas is
within the temperature range of 1600 to 2000 Deg. F. On long kilns, this
temperature may move along the axis of the kiln with time, causing
injection to take place outside of the temperature range. SNCR has not
been used full time on long wet or long dry kiln systems.

NPS: PA DEP is incorrect. As noted in our e-mail of 9/07/07 to PA DEP,
Ash Grove Cement has installed a full-scale SNCR system on one of its
Midlothian, Texas, kilns and, according to Ash Grove, “is achieving a
35% - 40% NOX reduction on a consistent basis.”11

Step 3 – Evaluate Control Effectiveness of Remaining Control
Technologies.

PA DEP: Removal efficiencies of the technically feasible NOx control
technologies range from 18% to 35%.

NPS: PA DEP is incorrect. Both Colorado and Washington have proposed 40%
NOx reductions for SNCR.

Step 4 – Evaluate Impacts and Document the Results.

PA DEP: The estimated cost of SNCR is $1,014 per ton of NOx removed.

NPS: PA DEP should better explain this estimate.

Step 5 – Evaluate Visibility Impacts.

PA DEP: The maximum visibility improvement due to the most effective NOx
control in the most affected Class I area (Shenandoah National Park) was
0.076 dv. The minimum cost of improvement was $7,494,026 annually per
deciview. Therefore, PA DEP does not recommend any additional control of
NOx as a result of the BART analysis. The current operating permit
limitation for NOx emissions from Clinker Kiln Number 5 is 476 pounds
per hour.

5. Conclusion: The estimated visibility improvement is too low and the
cost of additional air emission control too high to warrant additional
emission control to meet the BART requirements. Thus, this reviewer
concludes that no additional emission control equipment for BART is
warranted at this location. Therefore, the existing permit limits will
meet the requirements for BART.

NPS: PA DEP is the only state to date to conclude that SNCR is not BART
for a cement kiln. The results for the four other cement kiln BART
analyses that we have reviewed are summarized below:

• Ash Grove has proposed to add SNCR at 35% control to its wet kiln in
Montana City,

MT.

• Holcim has proposed to add SNCR at 30% control to its wet kiln in
Trident, MT.

• Colorado has proposed that CEMEX add SNCR at 40% control to its kiln
near Lyons,

CO.

• Washington Ecology has proposed that LeFarge add SNCR at 40% control
to its wet kiln near Seattle, WA.

PA DEP has underestimated the effectiveness of SNCR and has provided no
information on how it arrived at its cost estimates. Finally, even with
these short-comings, the $1,014 per ton is lower than the $4,200/ton
cost at Lefarge (WA) and the $7.5 million per deciview cost is below the
national average of $13 million/dv for NOX BART.

DEP Response:  The Department has recently finalized a statewide amended
NOx control regulation for Portland cement kilns which is viewable here:


 HYPERLINK
"http://www.pacode.com/secure/data/025/chapter145/subchapCtoc.html"
http://www.pacode.com/secure/data/025/chapter145/subchapCtoc.html 

The Department projects that Essroc’s NOx emissions from kiln #4 will
be reduced approximately 20% by 2013 as a result of the amended Portland
cement kiln regulation. 

The BART determination is consistent with the BART Guidelines.  The
Department considered all of the BART Guideline factors, and determined
that a BART source analysis resulting in significant visibility
improvement from controls would consequently justify higher cost
controls.   The maximum modeled deciview impact due to this BART source
on the most impacted Class I area is less than 0.5 dv, the threshold
suggested in the BART Guidelines to determine if a source contributes to
visibility impairment.

18.  Comment:  Lehigh Cement/York Operations

Step 1 – Identify All Available Retrofit Control Technologies.

PA DEP: Table 3 lists all available control technologies on the
Lehigh\York White Cement Kiln for control of SO2 and NOx. Lehigh Cement
conducted ambient modeling and visibility analysis for the period 2001
through 2003 at Class 1 areas affected by this facility. Visibility
improvement is based on the company analysis.

NPS: PA DEP’s Table 3 is incomplete and does not include all of the
eight NOX control technologies identified by Lehigh in its January 2006
BART analysis. We are especially concerned that PA DEP omitted SNCR,
which has been proposed as BART by every cement plant outside of PA that
is subject to BART. PA DEP should explain why it omitted this critical
technology.

Step 2 – Eliminate Technically Infeasible Options.

PA DEP: SNCR has not been used full time on long wet or long dry kiln
systems.

NPS: PA DEP is incorrect. As noted in our e-mail of 9/07/07 to PA DEP,
Ash Grove Cement has installed a full-scale SNCR system on one of its
Midlothian, Texas, kilns and, according to Ash Grove, “is achieving a
35% - 40% NOX reduction on a consistent basis.”12

PA DEP: It has been reported that SNCR has been proposed for NOx control
as BART on a long cement kiln in the US. While there several technical
issues associated with the installation of SNCR for long kilns, economic
and impact analyses were performed for this control option. A NOx
reduction of 35% was used for SNCR on long kilns based on the July 2006
ERG report to the Texas Commission on Environmental Quality. They used
this reduction since SNCR control on long kilns is considered Innovative
and available data is limited.

NPS: PA DEP is incorrect. Both Colorado and Washington have proposed 40%
NOx reductions for SNCR.

Step 4 – Evaluate Impacts and Document the Results.

PA DEP: The estimated cost of a Cadence fan combined with a SNCR system
is $ 2,623 per ton of NOx removed and also determined to be Economically
Infeasible. The estimated cost of a Cadence fan system is $ 1,118 per
ton of NOx removed.

NPS: PA DEP failed to evaluate addition of SNCR without the Cadence fan
system.

Conclusion:

PA DEP: The estimated visibility improvement is too low and the cost of
additional air emission control too high to warrant additional emission
control to meet the BART requirements. Thus, this reviewer concludes
that no additional emission control equipment for BART is warranted at
this location. Therefore, the existing permit limits will meet the
requirements for BART.

Step 5 – Evaluate Visibility Impacts.

PA DEP: The maximum visibility improvement due to the most effective NOx
control in the most affected Class I area (Shenandoah National Park) was
0.017 dv. The minimum cost of improvement was $ 10,606,000 annually per
deciview, for lesser improvement. Therefore, I do not recommend any
additional control of NOx as a result of the BART analysis. The current
operating permit limitation for NOx emissions from the kiln is 8.2
pounds per ton of cement clinker produced.

Conclusion:

The estimated visibility improvement is too low and the cost of
additional air emission control too high to warrant additional emission
control to meet the BART requirements. Thus, this reviewer concludes
that no additional emission control equipment for BART is warranted at
this location. Therefore, the existing permit limits will meet the
requirements for BART.

NPS: PA DEP is the only state to date to conclude that SNCR is not BART
for a cement kiln.

The results for the four other cement kiln BART analyses that we have
reviewed are summarized below:

• Ash Grove has proposed to add SNCR at 35% control to its wet kiln in
Montana City,

MT.

• Holcim has proposed to add SNCR at 30% control to its wet kiln in
Trident, MT.

• Colorado has proposed that CEMEX add SNCR at 40% control to its kiln
near Lyons,

CO.

• Washington Ecology has proposed that LeFarge add SNCR at 40% control
to its wet kiln near Seattle, WA.

PA DEP has not evaluated SNCR and has provided no information on how it
arrived at its cost estimates. Finally, even with these short-comings,
the $10.1 million per deciview cost for the Cadence fan plus SNCR is
below the national average of $13 million/dv for NOX BART.

DEP Response:  The Department has recently finalized a statewide NOx
control regulation for portland cement kilns which is viewable here: 

 HYPERLINK
"http://www.pacode.com/secure/data/025/chapter145/subchapCtoc.html"
http://www.pacode.com/secure/data/025/chapter145/subchapCtoc.html 

The Department projects that Lehigh Cement/York’s NOx emissions from
their white cement kiln will be reduced approximately 30% by 2013 as a
result of the amended Portland cement kiln regulation.

The BART determination is consistent with the BART Guidelines.  The
Department considered all of the BART Guideline factors, and determined
that a BART source analysis resulting in significant visibility
improvement from controls would consequently justify higher cost
controls.   The maximum modeled deciview impact due to this BART source
on the most impacted Class I area is less than 0.5 dv, the threshold
suggested in the BART Guidelines to determine if a source contributes to
visibility impairment.

19.  Comment:  P.H. Glatfelter Company

STEP – 4: Evaluate Impacts and Document the Results

PA DEP: Based on a series of inquiries this reviewer made about the
basis of their BART analysis, P.H. Glatfelter provided two revisions to
their cost data and the basis of their cost calculations. Their most
current cost analysis breakdown for the wet scrubber is in Table 4 of
the third version of their cost analysis.

NPS: The third version revisions cited above were not received by NPS
until they were

requested on 7/28/10. Our review of those revisions has determined that
PA DEP has

overestimated the costs of wet scrubbing SO2 from Power Boiler #1. The
over-estimation results from several deviations from the OAQPS Control
Cost Manual approach:9

• PA DEP did not follow the Cost Manual in estimating installation
costs. The Cost manual recommends multiplying the Purchased Equipment
Cost (PEC) by 0.85 to estimate the Direct Installation Cost. Instead, PA
DEP multiplied the PEC by a factor of 1.65.

• PA DEP assumed a 10% interest rate instead of the 7% rate
recommended by the Cost

Manual.

• PA DEP estimated operating labor time at 3 hours/shift versus the
0.5 hours/shift

estimated by the Cost Manual.

• PA DEP estimated maintenance labor time at 1 hour/shift versus the
0.5 hours/shift

estimated by the Cost Manual.

• The solid waste generation rate used by PA DEP is double the
8/17/2007 vendor estimate.

As a result of these deviations from the Cost Manual, PA DEP estimated a
Total Annual Cost (TAC) of $5.4 million and cost-effectiveness =
$1,667/ton of SO2 removed. Our application of the Cost Manual (see
electronic attachment) yielded a TAC = $3.6 million and
cost-effectiveness = $1,127/ton of SO2 removed.

STEP – 5: Evaluate Visibility Impacts

PA DEP: The 98th percentile deciview improvement expected by the
installation of a wet scrubber system on the Number 1 Power Boiler was
found to be 0.219 dv. The cost in terms of dollars per deciview at this
facility for the installation of the wet scrubber was calculated to be
$24,545,196/dv. The 98th percentile deciview improvement expected by
operating the Number One Power Boiler existing OEC year round was found
to be 0.010 dv. The cost in terms of dollar per deciview for this
control approach was calculated to be $41,629,300/dv.

NPS: In addition to improving visibility at Shenandoah National Park
(the Class I areas used by PA DEP), a similar visibility improvement was
modeled at Brigantine Wildlife Refuge. PA DEP did not model additional
visibility improvements at Dolly Sods and Otter Creek Wilderness Areas
which had visibility impacts about half of those at Shenandoah and
Brigantine.) The cumulative benefits of reducing SO2 from Power Boiler
#1 would be 0.44 dv at Shenandoah and Brigantine, with additional
benefits at Dolly Sods and Otter Creek.

Conclusion:

PA DEP: The cost effectiveness of installing a wet scrubber system for
SO2 control on Number 1 Power Boiler, taking into account visibility
improvement, was $24,545,196/dv. The cost effectiveness of operating the
OEC system year-round for NOx control on the Number 1 Power Boiler,
taking into account visibility improvement, was $41,629,300/dv. This
data, in addition to cost effectiveness values for emissions reduced and
modeled visibility impacts, are shown in Table III.

NPS: Using the Cost Manual approach to estimate the cost of adding a 90%
efficient wet scrubber to Power Boiler #1, the resulting
cost-effectiveness at Shenandoah National Park is $17 million/dv is
which is lower than the $20 million/dv average cost-effectiveness for
all of the SO2 controls we have seen proposed as BART nationwide. When
one considers the cumulative benefits of improving visibility at the
four Class I areas modeled by PA DEP, the cost-effectiveness drops below
$9 million/dv, which is less than half of the $20 million/dv average
cost-effectiveness for all of the SO2 controls we have seen proposed as
BART nationwide. We conclude that addition of a 90% efficient wet
scrubber to Power Boiler #1 is BART.

DEP Response:  The Department explained on page seven of the BART review
memo that a wall would have to be demolished and an addition would have
to be built.  Additionally, an asbestos abatement program would be
required to allow for the installation of FGD.  The facility added these
costs to the Direct Installation calculation by using a factor of 1.65
instead of 0.85.  The Department visited the facility and agrees that
these additional costs will be incurred and should be accounted for as
additional direct installation costs.  

Attachment 1 Footnotes

1 DEPARTMENT OF ENVIRONMENTAL QUALITY AIR QUALITY DIVISION BART
Application Analysis AP-6040 May 28, 2009 NAME OF FIRM: PacifiCorp NAME
OF FACILITY: Jim Bridger Power Plant

2 DEPARTMENT OF ENVIRONMENTAL QUALITY AIR QUALITY DIVISION BART
Application Analysis AP-6041 May 28, 2009 NAME OF FIRM: PacifiCorp NAME
OF FACILITY: Dave Johnston Plant

3 DEPARTMENT OF ENVIRONMENTAL QUALITY AIR QUALITY DIVISION BART
Application Analysis AP-6047 May 28, 2009 NAME OF FIRM: Basin Electric
Power Cooperative NAME OF FACILITY: Laramie River Station

4 DEPARTMENT OF ENVIRONMENTAL QUALITY AIR QUALITY DIVISION BART
Application Analysis AP-6042 May 28, 2009 NAME OF FIRM: PacifiCorp NAME
OF FACILITY: Naughton Power Plant

5 DEPARTMENT OF ENVIRONMENTAL QUALITY AIR QUALITY DIVISION BART
Application Analysis AP-6043 May 28, 2009 NAME OF FIRM: PacifiCorp NAME
OF FACILITY: Wyodak Plant

6 DEQ BART Report for the Boardman Power Plant Updated December 19, 2008

7 See http://www.wrapair.org/forums/ssjf/bart.html

8 For example, PacifiCorp has stated in its BART analysis for its
Bridger Unit #2 that “The incremental cost effectiveness for Scenario
1 compared with the baseline for the Bridger WA, for example, is
reasonable at $580,000 per day and $18.5 million per deciview.”

9 EPA ALTERNATIVE CONTROLTECHNIQUES DOCUMENT -- NOX EMISSIONS FROM IRON
AND STEEL MILLS At iron and steel mills, there are 2 SCR units in the
U.S. An SCR unit is being used to control NOX emissions from a
gas-fired, radiant tube, continuous annealing furnace at a steel
mini-mill in the United States. This furnace also has LNB's. Controlled
emissions from this unit are about 33 ppm at 3 percent O2. A second SCR
unit, currently under construction, will be used to control NOX
emissions from an annealing furnace at an integrated steel plant in the
United States. This furnace does not have LNB's. The unit has a
guaranteed NO reduction of 90 percent.

10 BEST AVAILABLE RETROFIT TECHNOLOGYAT NON-EGU FACILITIES, April 19,
2010

WISCONSIN DEPARTMENT OF NATURAL RESOURCES The Georgia Pacific Green Bay
facility has a system of six boilers supplying power and electricity
used in the manufacturing of consumer paper products. Boiler B26 is a
spreader stoker-fired unit manufactured by Babcock and Wilcox installed
in 1962. It is a two drum, balanced draft furnace, with a maximum rated
heat input capacity of 350 mmBtu/hr. Boiler B26 burns washed coal
(eastern high and low fusion and western coals) and petroleum coke.
After reviewing these potential options, the

NOx control approach with the largest emission reduction potential for
each boiler is a combination of flue gas recirculation and over-fire air
(FGR/OFA) for boiler B26 and over-fire air (OFA) for boiler B27 followed
by a full-sized selective catalytic reduction (SCR) system for treating
the combined flue gas of both boilers. [emphasis added]

11 Overland Park, Kan., October 27, 2008, Ash Grove Cement Company Earns
Working for Cleaner Air Award from North Texas Clean Air Coalition, Ash
Grove Is First Texas Cement Manufacturer to Receive Award.

12 Overland Park, Kan., October 27, 2008, Ash Grove Cement Company Earns
Working for Cleaner Air Award from North Texas Clean Air Coalition, Ash
Grove Is First Texas Cement Manufacturer to Receive Award Recognizing
Businesses That Do Their Share for Cleaner Air. Ash Grove’s Midlothian
plant is the first Texas cement manufacturer recognized in the five-year
history of the program and received the award for the company’s
successful efforts to control nitrogen oxide (NOx) emissions. Ash
Grove’s Midlothian facility was one of the first

wet process cement plants in the world to install Selective
Non-Catalytic Reduction (SNCR) technology to successfully control ozone
forming emissions.

ATTACHMENT 2

To NPS/FWS Comments – Pennsylvania Draft Regional Haze SIP

FWS Comments on Best Available Retrofit Technology (Section 8)

General Comments

20.  Comment:  No draft BART determination performed by the Pennsylvania
Department of Environmental Protection (PADEP) resulted in a
recommendation that BART controls be implemented on any emission unit. 
PADEP stated in Section 8.5.2 of the Proposed Revision to the State
Implementation Plan for Regional Haze (June 2010) that, “The
Department did not establish or utilize bright line thresholds for cost
or for visibility improvement.  Instead, the Department employed an
approach that considered the multiple BART Guideline factors.  As a
result, sources with a higher degree of potential visibility improvement
from control would justify higher cost controls.  Conversely, only low
cost controls would be justified for sources with a lower degree of
potential visibility improvement.”  It seems that PADEP did not
establish any objective criteria for determining the acceptability of a
given control technology’s cost effectiveness or cost of visibility
improvement.  The above PADEP statement would seem to imply that in the
absence of absolute bright line thresholds, given all the BART sources
in the State, visibility improvement at the relatively lowest cost
facilities would result in BART controls at some subset of the
BART-eligible units.  Using that premise, it would seem reasonable that,
at a minimum, the following facilities would qualify as the relatively
lowest cost facilities for BART controls, in order of preference:

Lehigh Cement –Evansville Kilns #1 and #2 SNCR

CEMEX – Wampum Kiln #3 SNCR

Carmeuse Lime Kiln #5 SNCR

Carmeuse Lime Kiln #5 LNB

Sunoco Philadelphia Refinery Process Heater 2H-3 ULNB

CEMEX – Wampum Kiln #3 Water Injection

Sunoco Philadelphia Refinery Process Heater 2H-5 ULNB

It is interesting that in Pennsylvania (PA), NOx controls on cement
plants came out high in the BART cost rankings.  Nationwide, many cement
plants are controlling NOx from their kilns in their BART
determinations, so in this sense if PA were to control NOx from its
cement plant kilns, the Commonwealth would be consistent with the rest
of the nation.  Given the proximity of Class I areas to PA BART sources,
many BART determinations showed the cost-effectiveness of visibility
improvement was relatively expensive, even though the cost per ton of
NOx or SO2 controlled was, in our opinion, very reasonable.  In judging
cost-effectiveness of a given control technology as defined under the
EPA BART Guidelines1 in Step 4 (cost per ton of NOx, SO2 or particulate
matter) and Step 5 (cost per deciview improvement), it is the position
of the Federal Land Managers that if either of the cost-effective
measures are reasonable then there is sufficient justification for
implementing the control technology.

DEP Response:  The Department does not agree.  The Department considered
all of the BART Guideline factors, and determined that a BART source
analysis resulting in significant visibility improvement from controls
would consequently justify higher cost controls.  The Department
maintains that for a BART determination, a low $/ton control cost should
not supersede an excessively high cost of very low visibility
improvement.

21.  Comment:  It’s important to note that two facilities in PA
(Lehigh Cement Company/Evansville and Glatfelter Pulp & Paper Mill) each
have visibility impact on Brigantine National Wildlife Refuge and
Shenandoah National Park of about 0.6 deciviews.  Though PADEP concluded
that cost per deciview of visibility improvement at one or the other
Class 1 area exceeded what it considered reasonable for a single
facility, deference should be given to the total impact on multiple
Class 1 areas by a facility.  In this particular case the additive cost
per deciview improvement could bring a control technology within
acceptable cost range, since cost per ton was already reasonable for
several particular control technologies at these facilities.

DEP Response:  The EPA’s Guidelines for BART Determinations, Final
Rule (40 CFR Part 51, July 6, 2005) does not stipulate a requirement for
a “cumulative” impact analysis from one BART source on multiple
Class I areas.  Therefore, and as stated by the commentator, since EPA
has provided no guidance on the issue of assessing visibility benefits
that may result in multiple Class I areas when emissions are reduced
from a given BART source, the Department maintains that such an analysis
is unwarranted.  

The Department also maintains that requiring additional controls on a
BART source to provide very small visibility improvements at multiple
Class I areas due to an apparent low $/deciview visibility improvement
cost is nether justified or appropriate.  The summation of insignificant
delta deciviews serves the mathematical purpose of increasing the delta
deciview value with respect to the $/delta deciview equation and thereby
lowering the apparent cost of controls per delta deciview, while not
providing any significant visibility improvement to any of the affected
Class I areas.  Very small delta deciview values from a BART source
should not be summed to determine the cumulative effect on all lesser
affected Class 1 areas.  EPA’s BART guidelines recommend analyzing
visibility improvement for the highest impacted Class I area, with the
assumption that any improvement in the worse impacted area would result
in improvement in the lesser impacted areas.  The Department applied the
guidelines for the determination of BART for the affected sources.

22.  Comment:  Appendix J contains all of the PADEP Review Memos, but
the original company BART determinations should also be in the record. 
Please make these available in an Appendix.

DEP Response:  The Department is making the original company BART
submittals available to the public upon request during the public
comment period.

23.  Comment:  Please provide a discussion of how the five factors were
used in making the BART determinations.  Since there are so many
sources, a summary based on source category may be sufficient with
detailed information that could be included as an appendix.

DEP Response:  Summaries of BART determinations by source categories do
not convey the technical and detailed analyses the Department performed
to make the BART determinations for each individual BART unit.  The BART
analysis review memos convey all the information the Department relied
on to make its determinations of BART, including how the five factors
were used in making the BART determinations.  The Department’s review
memos are included in Appendix J (relating to Pennsylvania’s BART
analysis review memos) of the proposed Regional Haze (RH) State
Implementation Plan (SIP) revision.

24.  Comment:  On page 42, section 8.6, a large table lists BART
eligible sources and their corresponding emission levels. It is not
clear if these controls document existing levels or are implemented for
BART.  Please add a column that indicates whether these constitute new
BART or existing controls.

DEP Response:  The Department modified the proposed RH SIP revision on
page 58 (relating to BART determinations, control levels and schedules)
by including this statement: “The Department proposes to determine
that existing controls satisfy BART for the BART sources listed below in
Table 8.6.”

Refineries

PADEP has declared in the RH SIP that the Refinery consent decree
controls represent state-of-the- art-control and that this level of
control constitutes BART.

25.  Comment:  Sunoco Marcus Hook Refinery

It is noted that the 2.197 deciview impact at the Brigantine Wilderness
Class I area (using the MM5 modeling platform) by this refinery is the
largest visibility impact of any facility in PA on a Class I area.  The
visibility improvement due to installation of the controls to be
installed as part of the Environmental Protection Agency (EPA) Consent
Decree (which is claimed to be BART) is not quantified.  Please quantify
this visibility improvement. The installation of Selective Catalytic
Reduction (SCR) on the Fluidized Catalytic Cracking Unit (FCCU) under
the Consent Decree should provide for excellent visibility improvement.

In comparing the Review Memo dated September 25, 2007, to the Review
Memo dated June 10, 2008, the control efficiency assumption for using
Ultra Low NOx Burners (ULNB) on the process heaters went from 68% to
73%, respectively.  However, the cost per ton and cost per deciview
improvement shown under STEP 4 on page 8 did not change.  A pro-rata
calculation reflecting this change would show $8,532,138 per deciview
improvement. This is not an unreasonable cost of visibility improvement.
Also, the uncorrected $4,791 per ton of NOx control seems to be about
double the amount of some other such installations.

DEP Response:  The Department agrees that the dollar/ton estimate should
be adjusted to $4,462.00/ton for installation of ULNB.  The annual cost
will remain at $316,919 / year.  Sunoco conducted modeling that
predicted an approximate 0.035 delta deciview improvement from the
installation of ULNB, maximum daily impact scenario.  $316,919 / 0.035
dv = $9,054,828.57.71/dv.  In concurrence with the commentator, the
Department maintains that compliance with the consent decree will
achieve a significant visibility improvement and further analysis at
this facility for BART is not warranted at this time.

26.  Comment:  United Refining Company

In comparing the Review Memo dated September 25, 2007, to the Review
Memo dated June 11, 2008, the only difference is that the latter added a
Table 3 showing the visibility impact at the Presidential Range to be
about twice the impact at the Mingo Wilderness.  However, cost per
deciview improvement calculations went unchanged. It is, however,
recognized that even a 2x reduction in the cost per deciview improvement
would still result in very high values and would not change the
conclusion.

The cost effectiveness figures for NOx control via Ultra Low NOx Burners
at the Crude Heater are $3,266 per ton (PADEP 6/11/08 analysis).  This
seems high when compared with the $750 - $1,110 per ton costs developed
in Table 3-6 in the MARAMA Assessment of Control Technology Options For
Petroleum Refineries in the Mid-Atlantic Region (January 2007).  The
reason for such a discrepancy should be explained.

The Flue Gas Recirculation (FGR) alternative at $2,200 per ton of NOx
reduction might be considered to be reasonable if PADEP would consider
the cost per ton to supersede consideration of a high cost per deciview
improvement, as discussed in the General Comments section above.

DEP Response:  The Department concluded in the review memo that if the 2
best feasible technologies were employed that a delta deciview of
0.00525 + 0.0046 = 0.00985 would be achieved at the worst affected Class
1 area.  The Department concludes that this best achievable improvement
would not be cost-effective.  The BART determination is consistent with
the BART Guidelines.  The Department considered all of the BART
Guideline factors, and determined that a BART source analysis resulting
in significant visibility improvement from controls would consequently
justify higher cost controls.   The maximum modeled deciview impact due
to this BART source on the most impacted Class I area is considerably
less than 0.5 dv, the threshold suggested in the BART Guidelines to
determine if a source contributes to visibility impairment.

27.  Comment:  Sunoco, Inc. – Philadelphia Refinery

Ultra Low NOx burners on Heaters 2H-3 and 2H-5 seem reasonable on a cost
per ton basis ($1,775 and 2,148, respectively) and the cost per deciview
($7.0MM and $8.5MM, respectively) is not considered as excessive.  These
controls should be implemented as BART.

DEP Response:  The Department concluded in the review memo that the
impact benefit from Ultra Low NOx Burners at the worst Class 1 area
would be a delta deciview of 0.0135 for the 2H-3 Heater and 0.0149 for
the 2H-5 Heater.  The Department concludes that this best available
achievement would be minimal.  Additionally, the Department concluded in
the review memo that compliance with a federally enforceable consent
decree satisfies BART for the facility.  This consent decree requires
over 2,000 tons per year of NOx reductions from a 2001-2002 baseline
from boilers and heaters at the facility through the installation of
controls or shut downs.  The BART determination is consistent with the
BART Guidelines.  The Department considered all of the BART Guideline
factors, and determined that a BART source analysis resulting in
significant visibility improvement from controls would consequently
justify higher cost controls.   The maximum modeled deciview impact due
to this BART source on the most impacted Class I area is considerably
less than 0.5 dv, the threshold suggested in the BART Guidelines to
determine if a source contributes to visibility impairment.

28.  Comment:  ConocoPhillips Trainer Refinery

The Review Memo dated June 10, 2008 makes two references on pages 3 and
8 to setting a NOx emission limit for the FCCU by May 2009.  Since this
date is past, the actual emission limit should be inserted or another
date should be set. 

DEP Response:  The Department agrees with the commentator.  The BART
table on page 58 and the discussion regarding consent decrees on page
103 in the proposed RH SIP revision have been modified to correctly
describe the proposed final NOx limits for the FCCU from the enhanced
SNCR installed in 2006.  EPA proposed final NOx limits for the Trainer
FCC unit pursuant to paragraphs 50 and 51 of the referenced Consent
Decree on January 5, 2010.  Subsequently, ConocoPhillips (COPC) accepted
EPA’s limits on August 5, 2010.  Based on COPC’s acceptance of the
NOx limits on August 5, 2010, COPC will act to incorporate the limits
into a federally enforceable permit pursuant to paragraph 257 of the
referenced Consent Decree by November 2010.  The final FCC NOx limits
are:  121.1 ppmvd (365-day) and 155.3 ppmvd (7-day).  

Electric Generating Units

Pursuant to earlier comments by the Federal Land Managers, PADEP
performed an additional BART analysis on behalf of most of the Electric
Generating Units (EGU). This additional BART analysis considered an
Electrostatic Precipitator enhancement alternative, along with an
objective visibility improvement analysis of cost per deciview, rather
than relying on the subjective argument of “imperceptibility to the
human eye” of any visibility improvement.  The PADEP analysis lacked
backup data/information and references for arriving at the conclusions,
but it could be argued that relatively small adjustments to the
supporting data would likely not result in a different overall outcome.

29.  Comment:  The PADEP Review Memos for the EGUs (except for the
Cheswick Plant) generally did not contain a final ‘Conclusion’
section as many other Review Memos did.  We interpret the
‘Conclusion’ section as being a confirmation by PADEP management
that the staff reviewer’s recommendation was accepted as PADEP
findings and conclusions. A more definitive statement by PADEP in
Section 8.0 of the Proposed Revision to the State Implementation Plan
for Regional Haze that all Review Memos are confirmed as PADEP
conclusions could address this comment.

DEP Response:  The Department agrees with the commentator.  A
‘Conclusion’ section is a confirmation by the Department that the
staff reviewer’s recommendation was accepted as the Department’s
findings and conclusions.  A statement was added to the BART Section 8.6
(relating to BART determinations, control levels and schedules) on page
57 of the proposed RH SIP revision to indicate that all review memos are
confirmed as the Department’s conclusions.

30.  Comment:  Reliant Energy/Portland Generating Station

The Company conclusion is that BART is the existing electrostatic
precipitator.  BART for NOx and SO2 is complying with requirements of
the Clean Air Interstate Rule (CAIR).  An assertion is made on page 1-1
that CAIR will also reduce emissions of sulfates/inorganic condensable
PM10 emissions. CAIR does not regulate sulfates and no further
reductions will occur at this plant anyway.

DEP Response:  While the Department concurs that CAIR does not directly
regulate sulfates, the SO2 reductions due to CAIR would reduce emissions
of sulfates.

31.  Comment:  Allegheny Energy Supply/Hatfield Station Units 1, 2, and
3

The Company conclusion is that BART is existing control equipment, which
is an electrostatic precipitator (ESP) and an under-construction
Flue-gas desulfurization (FGD) system.  BART for NOx and SO2 is CAIR. 
We commend Allegheny Energy Supply and its environmental consultant,
EnviroMet for developing a complete BART determination with dominant
alternatives and a least-cost envelope.  It follows Appendix Y to Part
51 – Guidelines for BART Determinations Under the Regional Haze Rule
– better than any other analysis that we have reviewed for PA.

The ESP Upgrade alternative of replacing T-R Sets and Controls looks to
be a reasonable selection for BART given the $1,734 cost per ton figure.
The $39 million cost per deciview at Otter Creek Wilderness may be high,
but if visibility improvement at multiple Class 1 areas (i.e., Dolly
Sods Wilderness, Shenandoah National Park and James River Face
Wilderness) is considered, the cost of overall visibility improvement
warrants further consideration.

DEP Response:  The Department has concluded that a maximum of 0.006
delta deciview improvement due to an ESP upgrade at any one Class 1 area
would not be cost-effective.  Even with a cumulative maximum delta
deciviews for all four Class 1 areas of 0.018 delta deciviews, ESP
upgrade would be cost prohibitive for a BART determination,

32.  Comment:  PPL Generation LLC/Martins Creek SES Units 3 & 4

The Company conclusion is that BART is the use of existing #6 Fuel Oils.
BART for NOx and SO2 is CAIR. Low annual capacity factors of units (21%
& 15%, respectively) result in large control costs on a per ton basis
when examining ESP and venturi scrubber technology.  PADEP should
develop emission limits commensurate with the low capacity use, given
that the lower emissions were used, in part, to evade emission controls.
 The assumption of only 80% control efficiency for an ESP could be
challenged with a more realistic 98% control efficiency, even though the
Company made an argument for 80%.  Nevertheless, it would not change the
final conclusion that controls would be too expensive for the benefit. 
Neither the PPL Generation BART determination nor the PADEP BART
determination indicates that the facility is closed.

However, PADEP indicated on the FLM BART consultation conference call
that the facility was closed.  This should be documented in the BART
determinations and in the permit limits for this facility.

DEP Response:  The Department meant to convey during the BART conference
call the information that the coal fired boilers at Martin's Creek,
Units 1 & 2, are permanently shut down.  The oil fired boilers, Units 3
& 4, are still in operation and are operated at the limited capacities
as noted.

33.  Comment:  Orion Power – Cheswick Plant

The PADEP Review Memo did not consider an upgrade to the existing ESP,
but considered only the cost of a new baghouse.  The EPA BART
Guidelines1 state that, “. . . you should consider ways to improve the
performance of existing control devices, particularly when a control
device is not achieving the level of control that other similar sources
are achieving in practice with the same device.  For example, you should
consider requiring those sources with electrostatic precipitators (ESPs)
performing below currently achievable levels to improve their
performance.”

DEP Response:  The Department has concluded that removal of a small
percentage of the maximum 0.03 deciview impact due to particulate matter
at any one Class 1 area would not be cost-effective.  

34.  Comment:  Allegheny Energy – Mitchell Power Station

The reviewers did not locate a Company BART determination in the record.
 The PADEP Review Memo addressed the excessive cost of an ESP upgrade
alternative.

DEP Response:  The Department has concluded that a most maximum 0.001
delta deciview improvement at any one Class 1 area would not be
cost-effective.  A copy of Allegheny Energy's Mitchell Power Station
BART submittal will be made available.  The Mitchell Station has one 288
MW boiler that is subject to BART requirements.

35.  Comment:  Exelon Power – Eddystone Generating Station

The Exelon Power BART Evaluation document in section 5.5.1, and again on
pages 9 and 10 of Appendix E, seemed to indicate that only one year of
meteorological data was used in the meteorological modeling.  Section 3
of PADEP’s Review Memo seems to determine the visibility impact costs
on a 98th percentile basis.  If only one year of meteorological data was
used, visibility impact costs should be based on the maximum 24-hour
impact, rather than the 98th percentile value.  We realize that the $141
million cost per deciview of improvement, even if modified by the above
comment, would likely still be excessive.  Nonetheless, it should still
be corrected.

DEP Response:  The Department concurs that the facility did use one
year's worth of meteorological data, and revised the BART review memo to
reflect the visibility impact costs based on the maximum 24-hour impact.

Cement Companies

36.  Comment:  Lehigh Cement Company/Evansville Pennsylvania Facility

PADEP’s conclusion is that BART for particulate matter is the existing
control equipment, which is a fabric filter meeting the National
Emissions Standards for Hazardous Air Pollutants requirements (commonly
refered to as “MACT” standards). PADEP concludes that no additional
control equipment is justified for NOx or SO2 control.

A Selective Non-Catalytic Reduction (SNCR) system for the combined NOx
emissions from long dry preheater Kilns #1 and #2 should be given more
consideration as being cost-effective.  Certainly, the $627 per ton of
NOx removal is reasonable.  The $14,267,800 per deciview improvement as
shown in Table 3a and 3b of the PADEP Review Memo for BART Application
is not unreasonable in the context of visibility improvement costs
undertaken by some other BART determinations that have been made.  Our
most extensive information about the visibility costs of NOx control for
BART come from electric generation units.  There are many instances of
company-proposed BART NOx controls costing between $12 million and $35
million per deciview for visibility improvement at a single Class I
area.  In addition, as discussed in the General Comments above, a cost
of $14,267,800 per deciview of visibility improvement and $627 per ton
of NOx control using SNCR for the Lehigh Cement – Evansville plant is
one of the lowest costs-per-ton and visibility costs encountered among
all of the BART determinations in PA.  Therefore, if any BART controls
are to be undertaken in the State, SNCR at this plant should be one of
them. 

An additional point should be made regarding visibility impact as
measured by the calculation of cost per deciview.  We continue to
believe that it is appropriate to consider both the degree of visibility
improvement in a given Class I area as well as the cumulative effects of
improving visibility across all of the Class I areas affected. It simply
does not make sense to use the same metric to evaluate the effects of
reducing emissions from a BART source that impacts only one Class I area
as for a BART source that impacts multiple Class I areas.  And, it does
not make sense to evaluate impacts at one Class I area, while ignoring
others that are similarly significantly impaired.  In this case, the
visibility improvement of deploying SNCR on Kiln #1 and Kiln #2 should
be aggregated over both Shenandoah National Park and Brigantine
Wilderness Area.  Using Table 3-5 of Lehigh Cement Company’s original
BART determination (January 2006), the visibility improvement (in 98th
Percentile deciviews) of SNCR deployment was shown as follows:

		Shenandoah	 Brigantine

Kiln #1 	0.02 		0.04

Kiln #2 	0.02 		0.04

The result is a cumulative 98th Percentile deciview impact due to the
deployment of a common SNCR on both units of 0.12 deciviews.  If the
$971,310 annual cost of SNCR is divided by the 0.12 deciviews of
visibility improvement, the result is $8,094,250 per deciview.  This is
well within a range deemed to be reasonable.

DEP Response:  The Department has recently finalized a statewide NOx
control regulation for portland cement kilns which is viewable here: 

 HYPERLINK
"http://www.pacode.com/secure/data/025/chapter145/subchapCtoc.html"
http://www.pacode.com/secure/data/025/chapter145/subchapCtoc.html 

The Department projects that Lehigh Cement/Evansville’s NOx emissions
from kilns #1 and #2 will be reduced approximately 40% by 2013 as a
result of the amended Portland cement kiln regulation. 

37.  Comment:  Lafarge Corporation/Whitehall Plant

PADEP’s conclusion is that BART for particulate matter is the existing
control equipment, which is a fabric filter meeting MACT, and that no
additional control equipment is justified for NOx or SO2 control.

A Selective Non-Catalytic Reduction (SNCR) system for NOx emissions from
the dry preheater Kiln #2 should be given more consideration as being
cost-effective. A cost of $1,804 per ton of NOx removal might be
considered to be reasonable.  The $27,177,065 per deciview improvement
as shown in the PADEP Review Memo for BART Application is somewhat high,
but in consideration of the discussion in the General Comment section
above, a reasonable cost per ton might be cause to recommend that the
technology be deployed as BART, regardless of the cost per deciview.

DEP Response:  The Department has recently finalized a statewide NOx
control regulation for portland cement kilns which is viewable here: 

 HYPERLINK
"http://www.pacode.com/secure/data/025/chapter145/subchapCtoc.html"
http://www.pacode.com/secure/data/025/chapter145/subchapCtoc.html 

The Department does not believe that requiring additional controls on a
BART source based on a very small visibility improvement, albeit a low
$/ton control cost, is justified or appropriate.  

The Department projects that Lafarge’s NOx emissions from kiln-3 will
be reduced approximately 25% by 2013 as a result of the amended Portland
cement kiln regulation. 

The BART determination is consistent with the BART Guidelines.  The
Department considered all of the BART Guideline factors, and determined
that a BART source analysis resulting in significant visibility
improvement from controls would consequently justify higher cost
controls.   The maximum modeled deciview impact due to this BART source
on the most impacted Class I area is less than 0.5 dv, the threshold
suggested in the BART Guidelines to determine if a source contributes to
visibility impairment.

38.  Comment:  Cemex/Wampum Cement Plant Kiln #3

Cemex determined that it could commit to water injection and process
controls for 7% NOx control and an as-yet undetermined technology for a
10% reduction in SO2. The Company admitted that this is not a
“typical” BART analysis and that CALPUFF modeling was not performed.
 In the absence of a full-company BART determination, PADEP performed a
BART determination. PADEP concluded that SNCR at $1,014/ton NOx
reduction and $4,678,401 per deciview of visibility improvement is too
expensive and no additional control of NOx is warranted.  The conclusion
is incorrect. Both the $/ton and $/deciview are within an acceptable
range for BART and should not be dismissed – especially in the absence
of any state-defined guideline as to what cost ranges are considered
acceptable for BART, per our discussion in the General Comments Section
above.  In addition, it seems that at a minimum, PADEP should

accept the company’s BART offer to install the water injection
technology for a 7% NOx reduction and an “as yet” determined 10% SO2
control strategy, rather than concluding that no control technology is
necessary.

DEP Response:  The Department has recently finalized a statewide NOx
control regulation for portland cement kilns which is viewable here: 

 HYPERLINK
"http://www.pacode.com/secure/data/025/chapter145/subchapCtoc.html"
http://www.pacode.com/secure/data/025/chapter145/subchapCtoc.html 

The Department projects that Cemex’s NOx emissions from each of their
three kilns will be reduced approximately 20% by 2013 as a result of the
amended Portland cement kiln regulation. 

The BART determination is consistent with the BART Guidelines.  The
Department considered all of the BART Guideline factors, and determined
that a BART source analysis resulting in significant visibility
improvement from controls would consequently justify higher cost
controls.   The maximum modeled deciview impact due to this BART source
on the most impacted Class I area is considerably less than 0.5 dv, the
threshold suggested in the BART Guidelines to determine if a source
contributes to visibility impairment.

39.  Comment:  Keystone Cement Company – Bath Facility

It was stated in the Keystone BART Proposal (page 2-5) that the
anticipated shutdown of Kiln 2 was to occur no later than 2009.  This
shutdown is a result of a newpreheater/precalciner kiln that Keystone is
constructing.  However, PADEP listed proposed emission limits for
Keystone Kiln 2 in the listing of proposed emission limits for each BART
facility.  Please confirm whether or not Kiln 2 has been shutdown, and,
if so, the emission limits for Keystone Kiln 2 should be zero.

The following discussion of SNCR for Keystone Kiln 2 may be moot if it
is to be replaced by a new kiln before the five year BART deadline.  A
Selective Non-Catalytic Reduction (SNCR) system for NOx emissions from
Kiln 2 should be given more consideration as being cost-effective.  A
cost of $1,014 per ton of NOx removal might be considered to be
reasonable.  The $23,431,248 per deciview improvement as shown in the
PADEP Review Memo for BART Application is somewhat high, but in
consideration of the discussion in the General Comment section above, a
reasonable cost per ton might be cause to recommend that the technology
be deployed as BART, regardless of the cost per deciview.

DEP Response:  The Department has recently finalized a statewide NOx
control regulation for portland cement kilns which is viewable here: 

 HYPERLINK
"http://www.pacode.com/secure/data/025/chapter145/subchapCtoc.html"
http://www.pacode.com/secure/data/025/chapter145/subchapCtoc.html 

The Department does not believe that requiring additional controls on a
BART source based on a very small visibility improvement, albeit a low
$/ton control cost, is justified or appropriate. 

The Department projects that Keystone’s NOx emissions from kilns #1
will be reduced approximately 40% by 2013 as a result of the amended
Portland cement kiln regulation. 

The BART determination is consistent with the BART Guidelines.  The
Department considered all of the BART Guideline factors, and determined
that a BART source analysis resulting in significant visibility
improvement from controls would consequently justify higher cost
controls.   The maximum modeled deciview impact due to this BART source
on the most impacted Class I area is less than 0.5 dv, the threshold
suggested in the BART Guidelines to determine if a source contributes to
visibility impairment.

Other Facilities

40.  Comment:  Trigen-Philadelphia Energy Corp./Edison Boiler #3 & #4 &
Schuylkill Boiler #26

Trigen used a PADEP Reasonably Achievable Control Technology (RACT)
benchmark limit of $1,500 per ton of NOx as a bright line determination
for judging the acceptability of BART control technologies in section
2.2.2.  This is contrary to a statement in section 8.5.2 of the draft
Pennsylvania Regional Haze SIP which states, “The Department did not
establish or utilize bright line thresholds for cost or for visibility
improvement.”  The $1,500 per ton cost would be within BART cost
limits observed in national experience.  The control effectiveness costs
($/ton) for various NOx control alternatives are well developed, but
seem consistently higher than commensurate values shown in EPA
AirControlNET.  Please discuss any differences or better document the
references used to arrive at the stated costs.  Lower, more reasonable
costs could cause the Flue Gas Recirculation alternative to be a viable
BART control.  Trigen did not further consider wet scrubber systems for
SO2 control as a BART alternative because it was not demonstrated in
practice.  Actually, wet scrubbing is successfully deployed on oil-fired
boilers being operated in Japan, Cyprus and Korea.  This should provide
reason enough to perform cost analysis on this alternative.

DEP Response:  The Department did not use the $1,500 per ton of NOx
proposed by Trigen as a bright line determination.  The Department
concluded in the review memo that the impact benefit from all controls
for these units at the worst Class 1 area would be less than a delta
deciview of 0.01 for NOx and less than or equal to 0.0123 for switching
to ultra low sulfur diesel fuel for Boiler #26 and assuming 99% SO2
reduction.  A wet gas scrubber would not have impact benefits that
exceeded these levels.  The Department concludes that the best available
achievements would be minimal.

The BART determination is consistent with the BART Guidelines.  The
Department considered all of the BART Guideline factors, and determined
that a BART source analysis resulting in significant visibility
improvement from controls would consequently justify higher cost
controls.   The maximum modeled deciview impact due to this BART source
on the most impacted Class I area is considerably less than 0.5 dv, the
threshold suggested in the BART Guidelines to determine if a source
contributes to visibility impairment.

41.  Comment:  Carmeuse Lime, Inc./Annville Operation Lime Kiln #5

PADEP proposed that no additional equipment be installed on Lime Kiln #5
for NOx control, even though both low NOx burners (LNB) and Selective
Non-catalytic Reduction (SNCR) are technically feasible and
cost-effective NOx control alternatives for a long dry kiln.  The
Lafarge Cement Plant in Alpena, MI has proposed to install both
technologies on five long dry kilns.  Very recently, the literature has
begun to accept that SNCR is a technically feasible alternative for NOx
control in long dry kilns. The most effective NOx control is SNCR at a
cost of $1,014 per ton of NOx removed and $6,398,357 per deciview
improvement at Dolly Sods Wilderness Area. LNB was shown to cost $1,318
per ton of NOx removed and $8,315,000 annually per deciview improvement.
 All of the above figures are within reasonable cost per ton and cost
per deciview improvement ranges for BART.  Therefore, these control
technologies should be considered.

The Portland Cement Association2 noted that a relatively inexpensive,
but effective NOx control technique, is a ‘high pressure air injection
system’ (also called a mixing air system) that can be installed on the
kiln.  Mixing air systems have shown significant emissions reduction up
to 48% on the 13 kilns operating with this technology. This should have
been considered among the BART NOx control alternatives.  

The proposed NOx emission limit of 6.0 lb NOx/ton of lime should be
reconsidered if a NOx control technology is accepted as BART.

DEP Response:  The Department has stated in the review memo that a
maximum delta deciview improvement of 0.016 dv is all that would be
achieved from the original 0.059 deciview impact at the worst case Class
1 area.  The Department does not agree that such a minimal improvement
would be cost-effective.

The BART determination is consistent with the BART Guidelines.  The
Department considered all of the BART Guideline factors, and determined
that a BART source analysis resulting in significant visibility
improvement from controls would consequently justify higher cost
controls.   The maximum modeled deciview impact due to this BART source
on the most impacted Class I area is considerably less than 0.5 dv, the
threshold suggested in the BART Guidelines to determine if a source
contributes to visibility impairment.

42.  Comment:  United States Steel/Clairton Coke Works

Section 4 of the Review Memo stated that ACHD performed a BART analysis
and the results are presented in that document.  However, the record
does not contain the detailed BART determination performed by ACHD on
behalf of the Clairton Coke Works.  Please provide this documentation in
the record.  Step 4 of the Review Memo states that annualized cost
information was obtained from EPA’s AP42 Manual.  The EPA BART
Guidelines state that, “The basis for equipment cost estimates also
should be documented either with data supplied by an equipment vendor
(i.e., budget estimates or bids) or by a referenced source (such as the
OAQPS Control Cost Manual).  In order to maintain and improve
consistency, cost estimates should be based on the OAQPS Control Cost
Manual, where possible. The Control Cost Manual addresses most control
technologies in sufficient detail for a BART analysis.”3 You should
assure that the cost estimates that were developed are accurate. 
Nevertheless, the cost per ton is very high, so minor changes in the
costs will not likely change the final conclusions.  In terms of cost
per deciview improvement, please consider the effect on multiple Class I
areas, rather than just Otter Creek, to determine the overall cost per
deciview improvement.

DEP Response:  In its June 23, 2009 “Review Memo for BART Application
for the United States Steel Corporation Clairton Works Facility”, the
ACHD conducted an engineering analysis under the BART determination
process of 40 CFR 51.308(e) for each source subject to BART at the
Clairton Coke Works.  The review memo was submitted to the Department as
part of its BART submittal for Pennsylvania.  The details of the BART
analysis consist of what was presented in Section 4 of the review memo,
under the Subheading “BART 5 Factor Analysis,” and the following
additional details are presented in this response to the comment:

For the USS Clairton Works facility, Step 1 of the analysis identified
Ultra Low NOx Burners (ULNBs) as the available retrofit control
technologies for the sources identified as Boiler #2 and R1 Boiler. 
Steps 2 and 3 stated that ULNBs provide a 65% improvement over
conventional natural gas burners.  The ACHD believes instead a value of
50% is a more appropriate emission reduction factor for use with the
boilers under consideration.  This is based on information found in
AP-42, Chapter 1.4, “Natural Gas Combustion,”  Table 1.4-1, which
provides emission factors for Large Wall-Fired Boilers greater than 100
MMBtu/hr, for Uncontrolled (Pre-NSPS) boilers = 280 LB/106 scf, and
Controlled –Low NOx burners  = 140 280 LB/106 scf, for a reduction
factor of 50%.  All of the boilers are larger than 100 MMBtu/hr and are
pre-NSPS.

In Step 4, the ACHD used a value of $12,800 per ton for the cost of
ULNBs, the same value used for ULNBs in the analysis for another
Allegheny County source (Allegheny Ludlum).  In response to the comment,
ACHD reviewed the Sixth Edition of the EPA OAQPS Control Cost Manual
found at   HYPERLINK "http://www.epa.gov/ttncatc1/products.html#cccinfo"
 http://www.epa.gov/ttncatc1/products.html#cccinfo .  However,
information is only available for NOx post-combustion controls, and
information related to NOx combustion controls, including low NOx
burners, is identified as “planned.”  Therefore, ACHD utilized the
original $12,800 per ton figure for the cost of ULNBs. 

The “Annualized Cost” is arrived at as follows:  

For Boiler #2:  The value of the 2002 NOx emission inventory used in
NESCAUM modeling for the Boiler #2 is 290 tons per year.  The NOx
reduction from the installation of the ULNBs is now figured at 50%
instead of the previous 65% reduction factor, resulting in emissions
reductions of 145 tpy.  145 tpy, at $12,800 per ton NOx removed, results
in a $1,856,000 annualized cost for Boiler #2.

For the R1 Boiler:  The value of the 2002 NOx emission inventory used in
NESCAUM modeling for the R1 Boiler is 8.8 tons per year.  The NOx
reduction from the installation of the ULNBs is now figured at 50%
instead of the previous 65% reduction factor, resulting in emissions
reductions of 4.4 tpy.  4.4 tpy, at $12,800 per ton NOx removed, results
in a $56,320 annualized cost for the R1 Boiler.

Step 5, “Evaluate Visibility Impacts” is now revisited:

As before, using the CALPUFF NWS platform computer modeling the total
visibility impact of this facility, including all BART eligible units,
on the Otter Creek Class 1 area was found to be 0.0897dv.  The costs of
possible new controls in terms of dollars per deciview for this facility
were calculated to be:

The impact of Boiler #2 on Otter Creek is 0.0605 deciviews. 
$1,856,000/0.0605dv = $30,677,686/dv (instead of the previously
calculated $39,775,000/dv) for the ultra low NOx Burners for Boiler #2.

The impact of the R1 Boiler on Otter Creek is 0.0017 deciviews. 
$56,320/0.0017 = $33,129,412/dv (instead of the previously calculated
$43,144,000/dv) for the ultra low NOx Burners for the R1 Boiler. 

The majority of the visibility impairing emissions are associated with
the emissions from Boiler #2.  The greatest visibility impact
individually was 0.0605 dv attributed to the NOx emissions from Boiler
#2.  

The other emission source, the Desulfurization Plant, has existing
controls.  Based on a review of current emissions reduction approaches
in the RACT/BACT/LAER Clearinghouse (RBLC) there does not appear to be
adequate alternatives available for the Desulfurization Plant, beyond
the currently installed afterburner.  Thus, no cost effectiveness
calculations were performed for SOx from this unit.  

5. Conclusion:  The conclusion is unchanged, even with the revised
analysis.  Based on the five-factor analysis, the impact of this
facility does not warrant additional control.  The Department concurred
with the ACHD recommendation that compliance with the existing operating
permits for this facility is BART.

The BART determination is consistent with the BART Guidelines.  The
Department considered all of the BART Guideline factors, and determined
that a BART source analysis resulting in significant visibility
improvement from controls would consequently justify higher cost
controls.   The maximum modeled deciview impact due to this BART source
on the most impacted Class I area is considerably less than 0.5 dv, the
threshold suggested in the BART Guidelines to determine if a source
contributes to visibility impairment.

43.  Comment:  ISG Plate, LLC – Coatesville

Section 4 of the Review Memo stated that PADEP performed a BART analysis
and the results are presented in that document.  However, the record
does not contain the detailed BART determination performed by the
Department on behalf of the ISG Plate – Coatesville Plant.  Please
provide that documentation in the record.  PADEP should assure that the
cost estimates that were developed are accurate as discussed above for
the Clairton Coke Works.  The cost per ton is very high, so relatively
minor changes in the costs will not likely change the final conclusions.

Wet FGD for SO2 control was analyzed in Step 3 of the Review Memo, but
in Step 5 ‘dry’ FGD is mentioned in the visibility analysis.  This
very possibly could be a typographical error.  Wet FGD is capable of 98%
removal, rather than the 90% assumed in the analysis, but again, even
this change would not likely change the final conclusions.

In terms of cost per deciview improvement, please consider the effect on
multiple Class I areas (i.e., Shenandoah National Park, Dolly Sods
Wilderness Area and Otter Creek Wilderness Area), rather than just the
Brigantine Wilderness Area, to determine the overall cost per deciview
improvement.

DEP Response:  The Department reviewed the NESCAUM CALPUFF data.  As
stated in the review memo the maximum impact at the Brigantine
Wilderness Area was modeled to be 0.055 deciview.  This is the worst
Class 1 area case.  Of this, 0.022 deciview was attributable to NOx from
the electric arc furnace.  The Department concurs that there are no
feasible controls for NOx for EAFs.  The Department does not concur that
control of some portion of the remaining 0.033 deciview would be
cost-effective.  

The BART determination is consistent with the BART Guidelines.  The
Department considered all of the BART Guideline factors, and determined
that a BART source analysis resulting in significant visibility
improvement from controls would consequently justify higher cost
controls.   The maximum modeled deciview impact due to this BART source
on the most impacted Class I area is considerably less than 0.5 dv, the
threshold suggested in the BART Guidelines to determine if a source
contributes to visibility impairment.

44.  Comment:  Sunoco Chemicals – Frankford Plant

The cost analysis for NOx and SO2 control alternatives lacked detail in
constructing the basis for Total Capital Investment.  This information
should be provided to allow a third party to check for reasonableness of
the estimates.  However, relatively minor adjustments to the NOx
alternatives would not likely result in a change to the final conclusion
that the alternatives are too expensive.  In the case of wet FGD for SO2
control, $2,836 is not an unreasonable cost per ton of SO2 removal,
especially if a 98% control efficiency were used rather than the 90%
figure assumed by the analysis. Again, if PADEP is willing to allow a
reasonable cost per ton to supersede an excessively high cost of
visibility improvement, then the wet FGD alternative could be
considered.

DEP Response:  The Department concluded in the review memo that the
impact benefit from a wet FGD at the worst Class 1 area would be a
deciview improvement of 0.0301 for the boiler.  The Department concludes
that this best available achievement would be minimal.  Since the
baseline impact for the boiler is only 0.1379 deciview, an analysis
using a control efficiency of 98% would not result in a significant
impact benefit.

The BART determination is consistent with the BART Guidelines.  The
Department considered all of the BART Guideline factors, and determined
that a BART source analysis resulting in significant visibility
improvement from controls would consequently justify higher cost
controls.  The maximum modeled deciview impact due to this BART source
on the most impacted Class I area is considerably less than 0.5 dv, the
threshold suggested in the BART Guidelines to determine if a source
contributes to visibility impairment.

No comments were made on:

Dyno Nobel, Inc. Nitric Acid Plant

First Energy Generation Corp – Bruce Mansfield Plant

PPL Generation LLC – Montour

Exelon Power - Eddystone Generating Station

ATTACHMENT 2 Footnotes

1 See 40 CFR Part 51, Appendix Y. See Section IV.D.Step 3.4.

2 “Summary of Control Techniques for Nitrogen Oxide” by Zephyr
Environmental Corporation for the Portland Cement Association, 2008, p.
2.

3 See 40 CFR Part 51, Appendix Y. The U.S. Environmental Protection
Agency finalized its BART Guidelines on June 15, 2005, and published the
preamble and final rule text in the Federal Register on July 6, 2005.
The rulemaking action added Appendix Y to Part 51, titled “Guidelines
for BART Determinations Under the Regional Haze Rule.” See Section
IV.D.Step 4.a.5.



In accordance with the Regional Haze Rule (40 CFR 51.308(i)(3)), the
Pennsylvania Department of Environmental Protection (DEP) is making
available the comments received by the Federal Land Managers on the
draft proposed Regional Haze (RH) State Implementation Plan (SIP)
revision, and the DEP’s responses to those comments.  The following is
a reproduction of a comment letter received by the Department from the
United States Department of Agriculture (USDA) Forest Service, with
DEP‘s responses inserted:

USDA Forest Service Comments Regarding 

Pennsylvania Draft Regional Haze Rule State Implementation Plan 

7/30/2010

On June 2, 2010, the Commonwealth of Pennsylvania submitted a draft
Regional Haze Rule State implementation plan (SIP), pursuant to the
requirements codified in federal rule at 40 CFR 51.308(i)(2), to the
USDA Forest Service (FS).  The air program staff of the FS conducted a
substantive review of the Pennsylvania draft plan, and have provided the
comments listed below.  We look forward to the Pennsylvania Department
of Environmental Protection (PADEP) response as per section 40 CFR
51.308(i)(3).  For further information regarding these comments, please
contact Ralph Perron at (802) 222-1444.

1.	Comment:  This draft Pennsylvania  RH SIP has progressed nicely from
its earlier version, however it still heavily relies on MANE-VU work
without clear specifics on Pennsylvania efforts.

	DEP Response:  The Department appreciates the reviewer’s
acknowledgment of the Department’s progress in developing the proposed
Regional Haze (RH) State Implementation Plan (SIP).  MANE-VU was formed
by the Mid-Atlantic and Northeastern states, tribes, and federal
agencies to coordinate regional haze planning efforts and activities for
the region as a group.  MANE-VU was formed to encourage a coordinated
approach to meeting the requirements of EPA’s Regional Haze Rule (RHR)
and reducing visibility impairment in major national parks and
wilderness areas in the Northeast and Mid-Atlantic.  With regard to
MANE-VU work and Pennsylvania efforts, the Department clearly and
specifically indicates Pennsylvania’s efforts throughout the proposed
RH SIP as they relate to the stated goals and emission management
strategies and measures pursued by the MANE-VU regional planning
organization to reduce the visibility impairment in the Class I areas of
the Northeast and Mid-Atlantic.  For example, page 95 of the proposed RH
SIP revision contains the following paragraph:  “MANE-VU identified
emissions from 167 stacks at EGU facilities as having visibility impacts
in MANE-VU Class I areas that make controlling emissions from those
stacks crucial to improving visibility at MANE-VU Class I areas. 
MANE-VU’s agreed regional approach for this source sector is to pursue
a 90 percent control level on SO2 emissions from these 167 stacks by
2018 as appropriate and necessary.  MANE-VU has concluded that pursuing
this level of sulfur reduction is both reasonable and cost-effective. 
Table 10.4.4-1 below lists the EGU stacks in Pennsylvania identified on
the list of 167 stacks in all of MANE-VU, (including for reference
purposes and where applicable, the BART facility identification number
associated with the stack modeled in the MANE-VU Class I area deciview
impact spreadsheets contained in Appendix I) with their current and
anticipated controls, the permit status of the SO2 controls, and the
anticipated SO2 emissions reductions achieved by 2018.”  As is
discussed further on page 97 of the proposed RH SIP revision, the
Department has concluded that at this time and based on the controls
proposed, constructed and under construction in Pennsylvania, the
Department’s CAIR regulation constitutes a reasonable measure for EGUs
in Pennsylvania.  Consistent with the MANE-VU ‘Ask’ Statement, the
Department projects a 94.5% reduction in SO2 emissions from the EGUs
listed in Table 10.4.4-1 (or from alternative measures, as appropriate
and necessary) by 2018.  Also, page 120 of the proposed RH SIP revision
contains the statement that the Department has developed a proposed
low-sulfur heating and distillate oil regulation, in response to the
MANE-VU ‘Ask’ low-sulfur fuel oil strategy.  The Environmental
Quality Board (EQB) adopted the Department’s proposed low-sulfur fuel
oil regulation at its July 13, 2010, meeting.  The public comment period
began on September 25th and ends on November 29th, 2010.  Additionally,
the Department has revised the proposed RH SIP to address the status of
the regulation development that pertains to the source categories
considered in the Mactec Report (Appendix O).  (Please see page 83 of
the proposed RH SIP).  

2.	Comment:  Overarching comment is that it remains very difficult to
separate out how many facilities are in each category (EGU(CAIR),
non-EGU BART, 167 stack) and exactly which controls are being installed
vs which are modeled, with associated reductions in NOx, SO2 and PM. 
For example, page 80: “Because of the SO2 controls installed in 2009
and 2010 …, no additional measures will be needed to meet the
reasonable progress goals.”  Are these 167 stack controls operational
now?  Appendix I lists large deciview impacts from “RELIANT ENERGY
NORTHEAS” on Dolly Sods and Otter Creek Wilderness, and while this
facility is covered under the CAIR=BART umbrella, it is not clear what
controls are anticipated for this facility (or other top contributors)
or when they will be installed. 

	DEP Response:  The Department modified Tables 10.4.4-1 (relating to EGU
stacks in Pennsylvania and controls identified from the MANE-VU 167
stack list) and 10.4.4-2 (relating to additional EGU stacks and controls
in Pennsylvania) in the proposed RH SIP in response to the comment.  The
modifications indicate the permit status of the SO2 controls at the
Pennsylvania EGU stacks listed in the two tables, and provide BART
facility identification numbers associated with the stacks modeled in
the MANE-VU Class I area deciview impact spreadsheets contained in
Appendix I (relating to MANE-VU CALPUFF modeling results for
Pennsylvania BART sources), where applicable, for easy cross-reference. 
Table 10.4.4-1 now lists the EGU stacks in Pennsylvania that were
identified on the list of 167 stacks in all of MANE-VU, with their
current and anticipated controls, the permit status of the SO2 controls,
and the anticipated SO2 emissions reductions to be achieved by 2018. 
Table 10.4.4-2 provides similar information for several EGU stacks in
Pennsylvania not identified on the MANE-VU 167 stack list that the
Department expects to be controlled as a result of the Department’s
CAIR regulation.  

3.	Comment:  Reasonable Progress consists of “pursuing, as appropriate
and necessary, the four goals of the MANE-VU ‘Ask’ Statement”.  
These are not insignificant, but are commitments to pursue, not
enforceable reductions, yet.  Also a four factor analysis was conducted
(Appendix O “the Mactec Report”), but no controls adopted beyond
pursuit of “ask” even though it appears that several sources
categories have reasonable cost controls available.

	DEP Response:  The Department is pursuing proposed regulations to
address other source categories as described in Appendix O.  The
Department has revised the proposed RH SIP in response to the comment to
address the status of the regulation development that pertains to the
source categories considered in the Mactec Report (Appendix O).  (See
page 83 of the proposed RH SIP).  It is not the Department’s practice
to commit to adopting measures that must go through a public comment
process.  Furthermore, the EQB has the statutory authority to adopt
regulations, not the Department.  In addition, regulations undergo the
review of several advisory committees before EQB action, and undergo
consideration by the Independent Regulatory Review Commission as well as
the standing committees of the General Assembly after EQB consideration.
 Plan approvals for controls as a permitted source are also subject to a
public comment period.  A commitment to adopt could nullify the value of
a public comment and board review process.

4.	Comment:  Tables 7.2-1 and 7.3.1-1 show projected ~700,000 tpy of SO2
reductions under CAIR, yet there is no accounting for where these
reductions will occur or if they are still projected to occur. 
Pennsylvania needs to ensure it has a mechanism to make the facilities
operate at the required higher efficiencies, even if the market is not
providing sufficient incentives.

	DEP Response:  Table 7.2-1 (relating to Pennsylvania’s 2002 emissions
in tons per year) summarizes the Pennsylvania emissions, including
emissions from stationary point sources, for 2002.  Table 7.3.1-1
(relating to Pennsylvania’s 2018 projected emissions) summarizes the
projected emissions, including emissions from stationary point sources,
in 2018 for Pennsylvania sources.  Comparing the SO2 totals for
stationary point sources for 2002 and 2018 shows an approximate
reduction in SO2 from all stationary point sources of 700,000 tons per
year.  These reductions are projections from all stationary point source
sectors, and not solely from the CAIR-affected electric generating units
(EGUs) in Pennsylvania.  The Department has concluded that at this time
and based on the controls proposed, constructed and under construction
in Pennsylvania, the Department’s CAIR regulation constitutes a
reasonable measure for EGUs in Pennsylvania.  Consistent with the
MANE-VU ‘Ask’ Statement, the Department projects a 94.5% reduction
in SO2 emissions from the EGUs identified from the MANE-VU 167 stack
list, or from alternative measures, as appropriate and necessary, by
2018.  The Department is relying on implementation of EPA’s CAIR
replacement rule, the proposed Transport Rule, for achieving the
longer-term (2015-2018) SO2 reductions from the EGUs.  The Department
intends to re-evaluate the projections of SO2 reductions due to CAIR at
the time of Pennsylvania’s first periodic report to determine whether
the predicted reductions are realized. 

5.	Comment:  It would be helpful to include a summary of the non-EGU
BART determinations as was in the previous draft.

	DEP Response:  The lengthy summaries of BART determinations by source
categories contained in previous RH SIP drafts did not convey the
technical and detailed analyses the Department performed to make the
BART determinations for each individual BART unit.  The BART analysis
review memos convey all the information the Department relied on to make
its determinations of BART, and the review memos are included in
Appendix J (relating to Pennsylvania’s BART analysis review memos).

6.	Comment:  There is still no demonstration that Pennsylvania is doing
its share of emissions reductions at non MANE-VU class I areas.

	DEP Response:  The measures described in the proposed SIP revision
demonstrate that the Department is pursuing the adoption and
implementation of emission management strategies to obtain its share of
emissions reductions at the MANE-VU Class I areas and the non-MANE-VU
Class I areas.  Shenandoah National Park in Virginia and the Dolly
Sods/Otter Creek Wilderness Area in West Virginia are Class I areas in
the Visibility Improvement State and Tribal Association of the Southeast
(VISTAS) region that are impacted by emission sources operating in
Pennsylvania and other MANE-VU states.  The Department consulted with
the planning staff in West Virginia and Virginia when they were
establishing their reasonable progress goals for the Class I areas in
their states.  Neither of these two States intends to request that
Pennsylvania make additional reductions at emission sources located in
Pennsylvania (see page 19 of the proposed SIP revision).  

	The most recent modeling completed by VISTAS for the Class I areas in
VISTAS showed the uniform rate of progress in 2018 would be met at both
Shenandoah and Dolly Sods.  VISTAS modeling results indicate that the
projected level of emissions controls for Pennsylvania’s sources by
2018 will allow the Shenandoah National Park and Dolly Sods Class I area
to meet the reasonable progress goals established by Virginia and West
Virginia, respectively, by 2018.  The emissions control strategy for the
MANE-VU Class I areas’ reasonable progress analysis is, therefore, a
reasonable emissions control strategy for the reasonable progress
analysis for the Shenandoah National Park and Dolly Sods Class I areas.
(Please see pages 64-66 of the proposed SIP revision for the complete
discussion.)

7.	Comment:  Pennsylvania contains many stationary area sources related
to oil and gas development and long term operation.  This should be
addressed, at a minimum, on pages 33-34 (Section 7.3.2, Stationary Area
Sources) along with dry cleaners and service stations.

	DEP Response:  The development and operation of oil and gas reserves in
Pennsylvania requires stationary area sources (mostly stationary
internal combustion engines) and non-road engines located at oil and gas
sites.  Stationary internal combustion engines and non-road engines were
part of the MANE-VU SIP Modeling Inventory, prepared in 2002, as
described on pages 29-34 of the proposed RH SIP revision.  In the
Version 3 MANE-VU SIP Modeling Inventory, the non-road inventory was
completely redone because of changes that the EPA made to the
NONROAD2005 model.  

	The Department is presently examining and evaluating practices at oil
and gas sites due to the increase in oil and gas development in
Pennsylvania, to ensure that oil and gas development and operation
proceed in an environmentally sound manner.  When the MANE-VU SIP
modeling inventory was being prepared, much of the oil and gas drilling
activities in Pennsylvania had not begun.  The emissions from the recent
increase in oil and gas drilling, therefore, were not occurring and not
part of the 2002 baseline inventory.  Currently, the Department does not
have adequate information to compile an accurate inventory for oil and
gas exploration and production operations in future years.  In
accordance with Section 135.4 (relating to report format) contained in
25 Pa. Code Chapter 135 (relating to reporting of sources), the
Department intends to advise owners and operators of oil and gas
exploration and production operations in Pennsylvania by December 2010
that they must submit a source report for air contamination sources by
March 1 of each year, and that the reports must contain sufficient
information to enable the Department to complete its emissions
inventory.  The Department intends to examine the emissions projections
at oil and gas sites at the time of Pennsylvania’s first periodic
regional haze progress report to determine whether they are properly
characterized in all future-year modeling inventories.

	The Department requires permits for the operation of the types of
diesel and nonroad engines typically located at oil and gas sites.  As
explained on page 112 of the proposed RH SIP revision in Section 10.5.2
(Measures to Mitigate the Impacts of Construction Activities), Section 2
of both the General Plan Approval and/or General Operating Permit
(BAQ-GPA/GP 9) and the General Plan Approval and/or General Operating
Permit (BAQ-GPA/GP 11), states that nonroad and diesel engines must have
the best available technology (BAT) installed and in operation so that
the engine is in compliance with regulated emissions standards; these
general permits are federally enforceable.  Both general permits require
the permittee to maintain accurate records of the amount of time the
engine is in operation per month and the amount of fuel used.  In
addition, the proposed RH SIP revision has been modified to include this
paragraph on page 112:  “The Department also requires a General Plan
Approval and/or General Operating Permit (BAQ-GPA/GP – 5) for natural
gas, coal bed methane or gob gas production or recovery facilities. 
This General Permit authorizes construction and/or operation of a
natural gas, coal bed methane or gob gas production or recovery
facility.  This permit authorizes the construction of internal
combustion engine(s); dehydrator(s) and associated equipment that meet
the best available technology (BAT) required under 25 Pa. Code §§
127.1 and 127.12(a)(5).  Once authorization to use GP – 5 is granted,
construction of the natural gas, coal bed methane or gob gas production
or recovery facility designated in the application may proceed.”

8.	Comment:  There are many references to Dolly Sods Wilderness Area,
while there are few references to Otter Creek Wilderness Area, both of
which are Class I areas in West Virginia.  One example of this is on
page 19, first paragraph, which seems to incorrectly imply that Dolly
Sods and Otter Creek are the same Wilderness area.  In paragraph 2, on
the page 19, reference to Otter Creek Wilderness Area should be
included, but it is missing.  Another example of this is on page 60
(last paragraph), where Dolly Sods is mentioned, but Otter Creek is
omitted.  This is in contrast to sentences in same paragraph that state
Great Gulf IMPROVE monitor is representative of Presidential Range/Dry
River Wilderness and also that Moosehorn IMPROVE monitor is
representative of Roosevelt Campobello International Park.  A suggestion
would be to say that Dolly Sods IMPROVE monitor is representative of
Otter Creek.  And on page 61, Table 10.2-1, add Otter Creek to Dolly
Sods column. 

	DEP Response:  The Department has corrected the references, or lack
thereof, to Otter Creek Wilderness Area in response to this comment. 
Specifically, the Department has revised pages 19 and 76 to include the
reference that the Dolly Sods IMPROVE monitor is representative of the
Otter Creek Wilderness Area.  The Department has also revised Table
10.2-1 (relating to percent of modeled sulfate due to emissions from
listed states) on page 77 of the proposed RH SIP to include Otter Creek
in the Dolly Sods column.

 PAGE   

 PAGE   46 

