Technical Support Document (TSD)

For the Modeling and Weight of Evidence (WOE) Portions of the Document
Entitled “Revision to the Allegheny County Portion of the Pennsylvania
State Implementation Plan Attainment Demonstration for the
Liberty-Clairton PM2.5 Nonattainment Area”

April 2011

TSD Prepared October 2011

Timothy A. Leon Guerrero

Office of Air Monitoring & Analysis, 3AP40

U.S. Environmental Protection Agency, Region 3

1650 Arch Street

Philadelphia, Pennsylvania 19103

/s/

____________________________________________

Reviewed by Andrew Hass, Acting Associate Director

Office of Air Monitoring & Analysis (3AP40)

October 24, 2011

_________________

Date Signed

Purpose of the Technical Support Document

This Technical Support Document (TSD) describes the Environmental
Protection Agency’s (EPA’s) evaluation of the modeling and
Weight-of-Evidence (WOE) portions of Pennsylvania’s State
Implementation Plan (SIP) entitled “Attainment Demonstration for the
Liberty-Clairton PM2.5 Nonattainment Area”.

The Liberty-Clairton, PA PM-2.5 nonattainment area is comprised of a
small portion of Allegheny County that includes the boroughs of
Glassport, Liberty, Lincoln, Port Vue, and the City of Clairton (see
figure below).  The area was designated as nonattainment for the 1997
PM-2.5 National Ambient Air Quality Standard (NAAQS) in the January 5,
2005 Federal Register [70 FR 944], effective April 5, 2005.

EPA requires states with nonattainment areas to submit an attainment
demonstration and associated air-quality modeling, adopted state
regulations to reduce emissions of PM-2.5 and its precursors, and other
supporting information demonstrating that the area will attain the
standards as expeditiously as practical.  The Allegheny County Health
Department (ACHD) prepared its PM-2.5 SIP in cooperation with the
Pennsylvania Department of Environmental Protection to comply with the
CAAA of 1990 and with EPA requirements for the nonattainment area as
stated in EPA’s 2005 designation of the PM-2.5 nonattainment area and
EPA’s Clean Air Fine Particle Implementation Rule [72 FR 20586,
25APR2007].

The purpose of this TSD is to provide more detailed information than can
be contained in the official notice published in the Federal Register. 
Readers who need more information than we provide in this TSD or want to
review the modeling should consult the ACHD’s Revision to the
Allegheny County Portion of the Pennsylvania State Implementation Plan
(April 2011) for the Liberty-Clairton, PA PM-2.5 nonattainment area.

The Regulatory Framework

EPA established the PM-2.5 NAAQS on July 18, 1997 (62 FR 36852).  This
included an annual PM-2.5 standard of 15.0 µg/m3 based on an annual
arithmetic mean, averaged over 3 years, and a 24-hour (or daily)
standard of 65 µg/m3 based on a 3-year average of the 98th percentile
of 24-hour concentrations. These standards were decided upon after
reviewing numerous health studies demonstrating that serious health
effects are associated with exposures to PM-2.5.

Following promulgation of a new or revised NAAQS, EPA is required by the
CAA to designate areas throughout the United States as attaining or not
attaining the NAAQS.  This designation process is described in section
107(d)(1) of the CAA.  In 1999, EPA and state air-quality agencies
initiated the monitoring process for the 1997 PM-2.5 NAAQS, and, by
January 2001, established a complete set of air-quality monitors. On
January 5, 2005, EPA published initial air-quality designations for the
1997 PM-2.5 NAAQS (70 FR 944), based on air-quality monitoring data for
calendar years 2001-03.

The nonattainment designation for an area starts the process whereby a
state or tribe must develop an implementation plan that includes, among
other things, a demonstration showing how it will attain the ambient
standards by the attainment dates required in the CAA.  Under section
172(b), states have up to three years after the EPA’s final
designations to submit their SIPs to EPA.  These SIPs were due on April
5, 2008, three years from the effective date of the designations.

Section 172(a)(2) of the Act requires states to attain the standards as
expeditiously as practicable but within 5 years of designation (i.e.
attainment date of April 5, 2010 based on air quality data for
2007–09), or within up to 10 years of designation (i.e. to April 2015)
if the EPA Administrator extends an area’s attainment date by 1–5
years based upon the severity of the nonattainment problem or the
feasibility of implementing control measures.

Introduction to the Liberty-Clairton, PA PM-2.5 State Implementation
Plan 

Pennsylvania’s SIP for the Liberty-Clairton, PA PM-2.5 nonattainment
area outlines plans to reduce emissions that contribute to PM-2.5
nonattainment.  The SIP includes a modeling demonstration that is used
to show that the Liberty-Clairton area will meet the 1997 annual PM-2.5
standard as expeditiously as practical.  The Liberty-Clairton
nonattainment area has some of the highest annual and 24-hour design
values in the United States.  This TSD reviews the modeling and WOE
portions of Pennsylvania’s SIP for the Liberty-Clairton, PA PM-2.5
nonattainment area.

ase year design values (2000-02) were 21.4 μg/m3 for the annual NAAQS
and 63 μg/m3 24-hour NAAQ at the Liberty monitor.

What Are The Components Of A Modeled Attainment Demonstration?

Modeling Process Overview                                               
                                                                        
                  

EPA’s Guidance on the Use of Models and Other Analyses for
Demonstrating Attainment of Air Quality Goals for Ozone, PM2.5, and
Regional Haze (2007) describes how to construct a model attainment
demonstration.  The air quality modeling used in the Liberty-Clairton
area’s attainment demonstration consisted of two components; a
regional photochemical grid model and a local-scale model with
sufficient resolution to examine the impacts of local emission sources. 
Model results were used in a “relative” rather than “absolute”
sense.  Following this methodology the ratio of the model’s future to
current (baseline) predictions at both of the nonattainment area’s
PM-2.5 monitors determines if the Liberty-Clairton SIP’s proposed
local and regional emission controls are likely to lead to compliance
with the 1997 PM-2.5 NAAQS.

The Liberty-Clairton modeling demonstration used a regional
photochemical grid model (CMAQ) for part of its modeling demonstration. 
Baseline and future CMAQ modeling scenarios were run by the Bureau of
Air Quality Analysis and Research Division of Air Resources New York
State Department of Environmental Conservation with emissions
information compiled by the Mid-Atlantic/Northeast Visibility Union
Regional Planning Organization (MANE-VU).  CMAQ provided results for the
chemically reactive PM-2.5 species such as sulfates and nitrates.

Local sources in the Liberty-Clairton nonattainment area create steep
gradients in the PM-2.5 concentrations.  These gradients cannot be
properly resolved in CMAQ whose model grid cells are roughly twelve
square kilometers.  Local scale meteorology is also not well simulated
by the regional model due to steep topography within the nonattainment
area that often contributes to strong temperature inversions and complex
flow patterns within the valleys. A local-scale model, CALPUFF, was used
in the attainment demonstration to better gauge the local impacts of
direct PM-2.5 emissions and provide some meteorological enhancements
within the nonattainment area.  ACHD was responsible for running
CALPUFF.

Steps Required in a Modeled Attainment Demonstration

EPA guidance encourages states to take certain steps when preparing a
modeling analysis to demonstrate attainment of the PM-2.5 NAAQS. 
Recommendations include the following nine steps: 

Formulate a conceptual description of an area’s nonattainment problem.

Develop a modeling/analysis protocol.

Select an appropriate air quality model to use.

Select appropriate meteorological episodes to model.

Choose a modeling domain with appropriate horizontal and vertical
resolution and establish the initial and boundary conditions to be used.

Generate meteorological and air quality inputs to the air quality model.

Generate emissions inputs to the air quality model.

Evaluate performance of the air quality model and perform diagnostic
tests, as necessary.

Perform future year modeling (including additional control strategies,
if necessary) and apply the attainment test.

How Did Allegheny County Address All of the Modeled Attainment
Demonstration Components?

ACHD’s PM-2.5 SIP for the Liberty-Clairton, PA PM-2.5 nonattainment
area addressed all of the modeled attainment demonstration elements. 
Each of these elements will be discussed in the following sections.

Conceptual Description of the Problem

"Particulate matter," also known as particle pollution or PM, is a
complex mixture of extremely small particles and liquid droplets. 
Particle pollution is made up of a number of components, including acids
(such as nitrates and sulfates), organic chemicals, metals, and soil or
dust particles.  Particle pollution exposure is linked to a variety of
health problems, including increased respiratory symptoms, such as
irritation of the airways, coughing, or difficulty breathing, decreased
lung function, aggravated asthma, development of chronic bronchitis,
irregular heartbeat, nonfatal heart attacks, and premature death in
people with heart or lung disease.

ACHD’s conceptual description of its PM-2.5 nonattainment problem is
detailed in Appendix C (Modeling Protocol) of its SIP document.  Two
additional reports prepared by the ACHD are included in Appendix E
present data from two speciated PM-2.5 monitors in Allegheny County.

The Liberty-Clairton Area is made up of complex river valley terrain
encompassing an area roughly three by five miles in size.  It includes a
four mile winding portion of the Monongahela River south of the City of
Pittsburgh and is bordered on the east by the Youghiogheny River. The
area includes rural land interspersed with densely populated residential
areas and industrial facilities.

There are significant differences in terrain elevations within the
nonattainment area; river valley elevations average just over 700 feet
above mean sea level while adjacent hilltops can rise up to 1250 feet. 
Under weak synoptic forcing these sharp elevation differences can
generate strong nocturnal temperature inversions.  Temperature
differences between the valley floors and nearby hilltops can range from
2 to 7° F under these conditions.  Cold air drainage flows often occur
during these periods of strong nocturnal cooling.  Under these
circumstances air flows within the valleys can become quite complex with
down slope flows reaching 3-4 mph and wind directions veering up to
180° within each valley’s vertical profile.  

Episodes of poor air quality often occur within the Liberty-Clairton, PA
PM-2.5 nonattainment area during periods of strong nocturnal inversions.
 Under these conditions, air dispersion is often minimized allowing
emissions to “build up” within the river valleys contributing to
episodes of poor air quality, which lead to the nonattainment area’s
high PM-2.5 design values.  On some occasions PM-2.5 concentrations
within the Liberty-Clairton area are significantly higher than
concentrations in the nearby City of Pittsburgh.  These differences in
air quality are one of the reasons why the Liberty-Clairton, PA PM-2.5
nonattainment area was separated from the surrounding Pittsburgh-Beaver
Valley, PA PM-2.5 nonattainment area.  

ACHD performed source apportionment modeling for the Liberty and
Lawrenceville monitors in Allegheny County using the Positive Matrix
Factorization Model (PMF Version 1.1).  The monitors are approximately
eleven miles apart from each other with the Liberty monitor located in
the Liberty-Clairton area and the Lawrenceville monitor located to the
north near Pittsburgh.

Local impacts were gauged by comparing both monitors’ speciation data
with each other using PMF.  ACHD’s analysis found that the Liberty
monitor’s PM-2.5 concentrations are impacted by regional loading
(mainly sulfates) based on similarities in both monitor’s speciation
data.  Sources near the Liberty monitor, both stationary and mobile,
were thought to be responsible for the speciation differences observed
between two Allegheny County monitors.  On peak concentration days at
Liberty, carbons and ammonium are prominent species while chorine is
more prominent on cold-weather peak days.

Models Used in the Attainment Demonstration

ACHD utilized two models in its attainment demonstration; a regional
photochemical grid model used to simulate regional PM-2.5 concentrations
and a local scale model used to simulate the effects of local sources of
direct PM-2.5 emissions.

The photochemical grid model used in ACHD’s modeling demonstration is
the CMAQ (version 4.5.1) model with CB4 chemistry, aerosol module for
PM-2.5 and RADM cloud scheme.  Photochemical modeling was performed by
the Bureau of Air Quality Analysis and Research, New York State
Department of Environmental Conservation (DEC).

The CMAQ modeling system has been designed to approach air quality as a
whole by including state-of-the-science capabilities for modeling
multiple air quality issues, including tropospheric ozone, fine
particles, toxics, acid deposition, and visibility degradation.  This
system includes input from both meteorological and emissions models,
which are detailed in the following sections.  

EPA recommends using air-quality models such as CMAQ in a “relative”
rather than “absolute” sense.  That is, take the ratio of the
model’s future to current (baseline) predictions at each monitor
within the nonattainment area using relative response factors (RRF), a
ratio of model projected concentrations over model base-line
concentrations.  Future PM-2.5 concentrations are estimated at existing
monitor sites by multiplying the modeled relative response factor at
locations “near” each monitor by the observation-based,
monitor-specific, “baseline” design value. The resulting predicted
“future concentrations” are then compared to the NAAQS.

ACHD’s analysis of its PM-2.5 monitors within Allegheny County showed
significant local impacts in the Liberty-Clairton PM-2.5 nonattainment
area.  A more detailed modeling analysis was needed to account for these
important local source impacts.  As noted previously, atmospheric
circulation in the Liberty-Clairton area during periods of high PM-2.5
concentrations is quite complex and beyond the resolution capabilities
of most regional photochemical models, which typically contain model
grid cells on the order of twelve square kilometers.  Additionally,
regional scale photochemical models are unable to fully resolve
important local topographic features that influence emission dispersion.
 Important spatial relationships between local sources and the affected
monitors within the Liberty-Clairton, PA PM-2.5 nonattainment area are
also not properly resolved in the regional photochemical grid model. 
For these reasons, ACHD decided to do local scale modeling using EPA’s
CALPUFF modeling system to help better resolve local atmospheric
circulation, local terrain features and local emission source
contributions.

The CALPUFF modeling system was used to create near-field and some long
range transport simulations for the Liberty-Clairton area. CALPUFF
Version 5.8 (dated June 29, 2007) is the most recent EPA-approved
version for regulatory applications. CALPUFF is currently not EPA’s
preferred near-field model to use for sources within 50 kilometers of
the affected monitor (see August 14, 2008 EPA Clarification Memo
entitled “Clarification of Regulator Status of CALPUFF for Near-field
Applications”).  The use of CALPUFF in this SIP, however, was granted
since the modeling protocol outlining the modeling methodology was
established before the issuance of this EPA memorandum.

The CALPUFF modeling system is comprised of CALMET (meteorological
model), CALPUFF (dispersion model), geophysical data preprocessors,
meteorological data preprocessors, and several postprocessors such as
CALPOST.  The CALPro GUI (dated June 29, 2007) comprises CALPUFF and
associated preprocessors and postprocessors in a central interface and
was utilized for the near field modeling analysis.

ACHD followed modeling procedures outlined in its PM-2.5 Model Protocol
(Appendix C) for the Liberty-Clairton, PA PM-2.5 nonattainment area and
EPA’s modeling guidance.  The modeling analysis utilized a combination
of regional CMAQ modeling using the MANE-VU inventory and local CALPUFF
modeling using a local revised ACHD inventory (for direct PM-2.5
emissions only).  ACHD listed several deviations from its original
modeling protocol in section 7.1.1 of its main SIP document.  These were
necessary to account for several issues that came up during the modeling
demonstration.

ACHD’s modeling methodology for its attainment demonstration followed
roughly the same methodology used in the Philadelphia Air Toxics study. 
Average impacts from the uncorrected MANE-VU inventory across an
“equivalent grid” were subtracted from the CMAQ impacts on a
concurrent basis.  The corrected ACHD inventory impacts for Liberty and
Clairton were then added to the CMAQ results.

Meteorological Time Periods Used in the Modeling

The EPA’s 2007 Modeling Guidance recommends modeling an entire year
or, at a minimum, several days in each quarter of a year, to adequately
represent the range of meteorological conditions that contribute to
elevated levels of fine particulate matter.  MANE-VU regional modeling
efforts were based on a base year of 2002.  This was the year used for
all PM-2.5 modeling efforts in EPA’s Region 3 and was also used in the
modeling demonstration submitted for the Liberty-Clairton, PA PM-2.5
nonattainment area.

Meteorological Models Used in the Air Quality Model

Meteorological data for the regional photochemical grid model and local
scale model were constructed for the modeling demonstration.  The
regional photochemical grid model was driven by output from the
Pennsylvania State University/National Center for Atmospheric Research
(PSU/NCAR) Mesoscale Meteorological Model, also known as MM5.  The
meteorological model for the local scale modeling was derived using
CALMET, which used a combination of MM5 data coupled with local
observations from nearby met stations.   The MM5 simulation included
“nudging”, which uses observations in the tendency equations to
force the model solution closer to observed values.

Additional information regarding the MM5 set up can be found in Appendix
K-2 of ACHD’s SIP document.

The following is a short list of MM5 options used in the modeling
demonstration simulation:

Kain-Fritsch (1993) convective scheme for both 36 and 12 km domains with
some adjustments to correct diurnal phase shifts in surface winds and
temperatures

Explicit moisture scheme (without the mixed phase) containing prognostic
equations for cloud water (ice) and rainwater (snow)

Modified version of the Blackadar PBL scheme

Simple radiative cooling scheme

Multi-layer soil model to predict land surface temperatures using the
surface energy budget equation 

MM5 met data was also used in the near field analysis using CALPUFF
though the CALMET processor.  CALMET was initialized with MM5 met fields
and supplemented with meteorological data from the Liberty monitoring
site, the Allegheny County Airport and USX Clairton.  The CALMET
processor is used to try and recreate some of the more complex
atmospheric flows in the Liberty-Clairton nonattainment area that cannot
be resolved in the MM5 simulation.  Further details of the
meteorological data processing for CALMET can be found in Appendix D of
ACHD’s SIP document.

Model Domain, Horizontal/Vertical Resolution and the Initial and
Boundary Conditions

MM5 (version 3.6) was run by the University of Maryland in consultation
with New York DEC staff.  The model was applied using a Lambert
conformal map projection.  The two nested domain consisted of coarse (36
km) and fine (12 km) mesh corresponding to 149 x 129 and 175 x 175 grids
respectively.  The Lambert projection used in this work followed the
Regional Planning Organization (RPO) national domain setup.

MM5 uses a terrain following σ-coordinate system with the pressure at
each σ-level determined from a reference state that is estimated using
the hydrostatic equation from a given sea-level pressure and temperature
with a standard lapse rate. There are 30 unevenly spaced σ-levels
yielding 29 vertical layers with higher resolution within the planetary
boundary layer (PBL).  The surface layer was set at about 10 m, the
level at which surface winds were typically observed, and the model top
was set at 50 hPa with a radiative top boundary condition.

CMAQ simulations were performed using a one-way nesting approach for a
continental simulation at 36 km grid spacing with a finer 12 km grid
centered over the eastern portion of the US.  Both domains utilized 22
layers in the vertical extending to about 16 km with 16 layers placed
within the lower three kilometers.

The continental CMAQ simulation utilized clean initial conditions with
boundary conditions extracted from a GEOS-CHEM global chemical model
simulation. The CMAQ 36 km simulation was initiated from December 15,
2001 with the first 15 days used as a spin up period and terminated on
December 31, 2002.  RPO emissions data from 2002 was used with
corresponding MM5 meteorological fields developed by the University of
Maryland.

Hourly boundary fields for the 12 km CMAQ domain were obtained from the
concentration fields generated by the 36 km CMAQ simulation.  The 12 km
simulations for both base and future year were assigned the boundary
conditions based on the 36 km CMAQ simulation with clean initial
conditions.  A more detailed description of the CMAQ runs can be found
in Appendix K-4 of ACHD’s SIP document.

Model resolution for CALPUFF, the near field model, was on a much finer
scale than the regional model.  CALPUFF grid spacing for the 150 km
regional source analysis domain was one kilometer and 100 meters for the
20 kilometer local scale analysis domain.  Both CALPUFF runs contained
11 vertical layers spaced within the first three km.  More information
on ACHD’s CALPUFF settings can be found in Appendix I of ACHD’s SIP
document.

Emissions Used in the Air Quality Models

Northeastern US emissions inventories for all source classifications
were developed by MANE-VU for use in regional analyses and SIPs. The
MANE-VU emission inventory was originally developed for regional haze
but was also used for ozone and PM-2.5 modeling efforts in the
Northeast.  A report by MARAMA located in Appendix G of the SIP document
provides additional details on the regional emission inventories.

The year 2002 was used for the baseline emissions inventory and 2014 for
the projected inventory for the Liberty-Clairton area.  Regional
projections utilized on-the-books/on-the-way (OTB/W) controls through
the 2012 timeframe. Since no additional projections were available at
the time, and since Liberty-Clairton controls focus on direct PM-2.5
emissions, regional projections for 2012 were used for precursors and
non-point PM-2.5 emissions in combination with local projections for
2014 for stationary point PM-2.5 emissions.

A more detailed inventory was developed by the ACHD for the extended
Liberty-Clairton area as part of its near scale modeling analysis.  This
inventory was limited to direct PM-2.5 emissions and was developed from
both the regional MANE-VU inventories and projections along with
ACHD’s inventories for stationary point sources.  These emissions
represent “actual” values based on pollutant emission factors and
throughputs or capacities of each emission source.  Emission values do
not necessarily represent permitted or “allowable” limits.

Regional controls for SO2 and NOx in the MANE-VU inventory were based on
CAIR.  On July 6, 2011, EPA finalized a rule that protects the health of
millions of Americans by helping states reduce air pollution and attain
clean air standards.  This rule, known as the Cross-State Air Pollution
Rule (CSAPR or Cross State Rule), requires 27 states to significantly
improve air quality by reducing power plant emissions that contribute to
ozone and/or fine particle pollution in other states.

Base Case Run Model Performance Evaluation

Model performance evaluations were done for MM5 and the regional CMAQ
runs as well as the local scale runs using CALPUFF.  Regional model
performance was based on matrices outlined in EPA’s Guidance on the
use of models and other analyses for demonstrating attainment of air
quality goals for ozone, PM-2.5, and regional haze (EPA-454/B-07-002)
and was summarized in Appendix K-1.

NY DEC examined the MM5 output to determine how well the simulation was
performing.  Only meteorological fields from the ozone season (May
through September) were examined.  The nudging process used National
Weather Service data so as expected the model did relatively well
compared with this data set.  Meteorological data from the Clean Air
Status and Trends Network or CASTNET was also used to assess model
accuracy since this data set was not included in the nudging process. 
Vertical parameters were also included in the analysis as well as cloud
cover data, precipitation and simulated calm periods.

The meteorological assessment showed that in general, MM5’s
performance was reasonable both at the surface and in the vertical.  It
was therefore concluded that MM5 performed adequately and the data could
be used in the CMAQ simulations.

Regional model evaluation statistics for the entire 2002 simulation were
calculated over a wide region for PM-2.5 mass and major species. In
general, CMAQ did a reasonable job simulating daily average PM-2.5 and
SO4 (sulfate) mass.  CMAQ, however, did not appear to be able to
reproduce the day-to-day or seasonal variation in other species like
NO3, OM, EC, or crustal mass.  The model appeared to meet model
performance criteria for PM-2.5 reasonably well over the Ozone Transport
Region (OTR) as a whole but not as well if one focused on an
emissions-dense, complex urban setting such as the New York City area. 
This finding appears to support ACHD’s local scale modeling efforts
for the Liberty-Clairton area.

ACHD examined CALPUFF’s ability to simulate local monitor
concentrations, specifically the model’s ability to simulate the
strong diurnal variations that contribute to the area’s high PM-2.5
design values.  A comparison of CALPUFF hourly averaged modeled and TEOM
(monitor) concentrations showed the local scale modeling generally
provided a better simulation of the strong diurnal contrasts (high
overnight concentrations) than did the regional model though the CALPUFF
model tended to over predict PM-2.5 concentrations.  This over
prediction, however, was not considered a problem since the local scale
modeling was used in a relative sense for attainment demonstration
purposes.  CALPUFF model performance using EPA matrices was judged to be
adequate overall with the 2nd and 3rd quarters showing better
performance than the 1st and 4th quarters.

2014 Control Case Modeling and the Modeled Attainment Test

The monitored attainment test for PM-2.5 utilizes both PM-2.5 and
individual PM-2.5 component species.  EPA refers to the attainment test
for PM-2.5 as the Speciated Modeled Attainment Test (SMAT).  A separate
relative response factor (RRF) is calculated for each PM-2.5 species.

In order to perform the recommended modeled attainment test, it is
recommended that states divide observed mass concentrations of PM-2.5
into seven (7) components plus passive mass:

mass associated with sulfates

mass associated with nitrates

mass associated with ammonium

mass associated with organic carbon

mass associated with elemental carbon

mass associated with particle bound water

mass associated with “other” primary inorganic particulate matter

and passively collected mass

To apply the test, states must first run an air quality model at least
twice to simulate current emissions and to simulate future year
emissions.   RRF values are then calculated for each PM-2.5 species. 
These RRF values are then multiplied by the base year concentrations for
each monitor within the nonattainment area.

Data analyses by Frank have noted that the FRM monitors do not measure
the same components and do not retain all of the PM-2.5 that is measured
by routine speciation samplers.  EPA recommends using the sulfate,
adjusted nitrate, derived water, inferred carbonaceous material balance
approach or SANDWICH to account for these differences in the SMAT.  A
more detailed description of this process can be found in EPA’s 
Guidance on the Use of Models and Other Analyses for Demonstrating
Attainment of Air Quality Goals for Ozone, PM-2.5, and Regional Haze,
EPA -454/B-07-002 April 2007.

ACHD’s modeled attainment test included an additional step to account
for the two models used in its analysis.  The direct PM-2.5 component
was taken from the CALPUFF modeling while CMAQ provided results for the
other PM-2.5 components.

The Liberty-Clairton nonattainment area has two PM-2.5 monitoring sites,
Liberty and Clairton.  Speciation data from Liberty was used for the
Clairton monitor since this site does not collect speciation data. 
Details on how the modeling attainment demonstration was conducted are
included in section 8 of the main PM-2.5 SIP document.

Projected monitor concentrations were calculated for both monitors in
the Liberty-Clairton area following EPA guidance.  EPA has developed the
Modeled Attainment Test Software (MATS) to automatically calculate
design values and interpolate for unmonitored areas.  Testing of MATS
version 2.2.1 showed inconsistent results with the methodology given in
the EPA Modeling Guidance.  EPA’s MATS software, therefore, was not
used by ACHD in this attainment demonstration though the general
methodology was followed to calculate the projected PM-2.5
concentrations.

Summary of the Combined Photochemical and CALPUFF Modeling Results

Annual and 24-hour PM-2.5 concentrations for both the Liberty and
Clairton monitors were calculated from the quarterly base-year averaged
monitor concentrations and the RRFs calculated from CMAQ and CALPUFF for
each PM-2.5 species components.  Results for the annual and 24-hour
PM-2.5 NAAQS are summarized in the following table and show that the
projected 2014 annual and 24-hour design values are below the 1997
PM-2.5 NAAQS.



Model Projected PM-2.5 Concentrations

For the Liberty-Clairton, PA PM-2.5 Nonattainment Area

Monitor	Annual Standard	24-Hour Standard

	Projected	1997 NAAQS	Projected	1997 NAAQS

Liberty	14.3 μg/m3	15.0 μg/m3	42 μg/m3	65 μg/m3

Clairton	11.8 μg/m3	15.0 μg/m3	27 μg/m3	65 μg/m3



EPA’s CSAPR rule was published after ACHD submitted their PM-2.5 SIP
for Liberty-Clairton.  ACHD did provide an analysis of EPA’s Transport
Rule, which had replaced EPA’s CAIR regional emission control
proposal, in its Weight of Evidence (WOE) section (13.3).  EPA’s
Transport Rule was the proposal in effect at the time of the SIP
document’s preparation.

EPA Region 3 suggests that ACHD submit an addendum to their current SIP
document that examines EPA’s most recent regional emission control
proposal (CSAPR) to ensure that its SIP modeling analysis is at least as
controlled as EPA’s current proposal.  This addendum should reexamine
the elements included in section 13.3 of ACHD’s main SIP document and
could additionally include a comparison of model results for the
Liberty-Clairton area from EPA’s CAIR and CSAPR modeling analyses.

Local Area Analysis

EPA guidance calls for additional analyses in areas where large sources
of primary particulates could cause substantial spatial gradients in
PM-2.5 concentrations.  These are often referred to as “hot spots”,
which could impact an area’s ability to comply with the NAAQS. 
Analysis of these hot spots could necessitate additional photochemical
grid modeling using a finer modeling grid scale, down to 1 to 2
kilometers, or alternatively the use of a Gaussian dispersion model.

ACHD conducted a local area analysis using EPA’s CALPUFF model.  This
analysis examined the impacts on direct PM-2.5 emissions in the
Liberty-Clairton, PA nonattainment area.  These impacts were coupled
with the regional CMAQ modeling analysis to demonstrate that the
nonattainment area would comply with the 1997 PM-2.5 NAAQS.

Unmonitored Area Analyses

The unmonitored area analysis is intended to be the primary means for
identifying high PM-2.5 concentrations outside of traditionally
monitored locations. The spatial resolution of the model that is the
underlying basis of the unmonitored area analysis will determine how
well it addresses primary PM-2.5 hotspots.  

The size of the Liberty-Clairton, PA PM-2.5 nonattainment area is much
smaller than most of the neighboring nonattainment areas.  As a
consequence of its size, the Liberty-Clairton, PA PM-2.5 nonattainment
area has a much higher density of monitors within it.  This partially
negates the necessity of doing this analysis though ACHD detailed an
assessment of the area in section 8.6 of its main SIP document.

ACHD examined modeled PM-2.5 reductions in the vicinity of the Liberty
monitor.  Several large sources of direct PM-2.5 emissions lie southwest
of the monitor near steep terrain adjacent to the Monongahela River. 
ACHD examined percent reductions in these areas and noted that shutdowns
and plant emission reductions would produce much greater improvements in
air quality in these areas than what are projected to occur at the
Liberty monitor itself.

Weight of Evidence Demonstration

ACHD conducted a Weight of Evidence (WOE) demonstration in accordance
with section 2.3 of EPA’s modeling guidance.  The WOE demonstration is
detailed in section 13 of ACHD’s main SIP document.  ACHD’s WOE
demonstration consisted of an analysis of monitor trends, local and
national emission control programs, population trends and monitoring
concentrations during periods of low production.  These results were
used as further evidence supporting the ACHD’s conclusion that their
SIP modeling analysis demonstrates compliance with the 1997 PM-2.5
NAAQS.

ACHD examined PM-2.5 monitoring trends within the nonattainment area and
the surrounding region.  PM-2.5 concentrations at most of the regional
monitors appear to be declining though ACHD noted these trends may not
be statistically significant at this time.

ACHD included a summary of various local and regional emission control
programs being implemented in the area.  Some of these control measures
may extend beyond the Liberty-Clairton, PA PM-2.5 nonattainment area and
therefore may have a lesser impact.  Emission control programs noted in
the WOE section include Pennsylvania’s enactment of a wood boiler
rule, a wood stove change out program in southwest Pennsylvania, EPA’s
recently proposed Transport Rule, Allegheny County’s diesel fuel
engine retrofit program, local and state anti-idling campaigns and
Allegheny County’s grant program to reduce diesel particulate
emissions.  ACHD contends that the combined effects of these controls
add additional support for their attainment demonstration.

Recent population trends are declining in the Pittsburgh region.  ACHD
noted that these population declines, if they continue, would help
reduce emissions in the area that contribute to the nonattainment
problem; fewer people equates to less driving and other activities that
contribute to emissions in the area.

ACHD also looked at the effect of lower emissions on its PM-2.5
monitoring network during the recent economic downturn.  This was done
by comparing TEOM data from 2009, a period of lower production due to
the economic downturn, to the unaffected 2000-2008 time period. 
Production levels at the local and regional level have fallen due to the
recent economic downturn.  These emission reductions appear to have
lowered monitor concentrations when compared to previous time periods. 
ACHD interprets these declines as demonstrating the future effectiveness
of their proposed local source emission controls. 

Recommendation of Conditional Approval of Liberty-Clairton PM-2.5
Attainment Demonstration

below 15.0 μg/m3 and 65 μg/m3 respectively indicating the
nonattainment area satisfies the Clean Air Act’s requirement that
State Implementation Plans provide for attainment of the NAAQS by the
applicable attainment date.

	

d

e

x

z

{

ˆ

”

Á

õ

ù

$

¢

£

 

¡

¢

£

¤

¥

¦

§

µ

Õ

Ö

í

£

¥

¦

§

h

h#W

h

6

hÄ

hx

h=

hg1

hx

hg1

kd.Œ

j

Anal CMAQ modeling included emission controls from EPA’s CAIR program.
 Since ACHD’s SIP submittal, EPA has proposed CSAPR to replace CAIR. 
To ensure that the modeling demonstration is still valid EPA is
requesting that ACHD redo the analysis it included in section 13.3 of
its main SIP document and consider examining the modeling technical
support documents for CAIR and the recently proposed Cross State Rule to
ensure that the more recently proposed rule provides at least the same
or possibly lower model concentrations in the Liberty-Clairton area as
the modeling analysis used in the April 2011 SIP modeling demonstration.

Based on the technical information provided in the Liberty-Clairton SIP
document, EPA concludes that the modeling and WOE analyses demonstrate
attainment of the 1997 PM-2.5 NAAQS.  EPA, therefore, should
conditionally approve the model attainment demonstration portion of
Liberty-Clairton, PA PM-2.5 SIP.  Final approval of the model attainment
demonstration portion of the SIP will be contingent on ACHD’s
reanalysis of the elements included in section 13.3 of its main SIP
document and an examination of modeled PM-2.5 concentrations from
EPA’s proposed CSAPR Rule.

 See section 10.0 of EPA’s Guidance on the Use of Models and Other
Analyses for Demonstrating Attainment of Air Quality Goals for Ozone,
PM2.5, and Regional Haze Attainment of Air Quality Goals for Ozone,
PM2.5, and Regional Haze, April 2007 (EPA -454/B-07-002)

 http://www.epa.gov/airtransport

   Frank, N., 2006: “Retained Nitrate, Hydrated Sulfates, and
Carbonaceous Mass in Federal Reference Method Fine Particulate Matter
for Six Eastern U.S. Cities” J. Air Waste Mange. Assoc., 56, 500-511.

EPA Technical Support Document for the Modeling Portion of
Pennsylvania’s 

 Revision to the Allegheny County Portion of the Pennsylvania State
Implementation Plan Entitled “Attainment Demonstration for the
Liberty-Clairton PM2.5 Nonattainment Area”

TSD Prepared October 2011

Page   PAGE  6  of   NUMPAGES  16 

