Technical Support Document for the Modeling and Weight of Evidence (WOE)
Portions

Of the State of Delaware’s Ozone State Implementation Plan (SIP)
Entitled “Delaware State Implementation Plan

For Attainment of the 8-Hour Ozone National

Ambient Air Quality Standard

Reasonable Further Progress and

Attainment Demonstration, June 2007”

Purpose of the Technical Support Document

This Technical Support Document (TSD) describes the Environmental
Protection Agency’s (EPA’s) evaluation of the modeling and WOE
portions of Delaware’s State Implementation Plan (SIP) revision
entitled “Delaware State Implementation Plan for Attainment of the
8-Hour Ozone National Ambient Air Quality Standard Reasonable Further
Progress and Attainment Demonstration, June 2007” (Delaware SIP
revision for the Philadelphia area).  Delaware is part of the
Philadelphia-Wilmington-Atlantic City ozone nonattainment area
(Philadelphia area).  The Philadelphia area is comprised of 18 counties
in Pennsylvania, New Jersey, Delaware and Maryland.  Classified as a
“moderate” nonattainment area, the Philadelphia area is required to
attain the NAAQS by June 2010, in effect, by the ozone season of 2009.  

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 in more detail should read the above referenced
Delaware SIP revision for the Philadelphia area.    

Introduction to the Delaware State Implementation Plan for Ozone

The Delaware SIP revision for the Philadelphia area includes plans to
reduce ozone-causing emissions for Delaware’s portion of the
Philadelphia Area.  The Plan also includes modeling that will predict
whether the area will meet the ozone standard by the due date.  This TSD
reviews modeling and WOE portions of the Delaware SIP revision for the
Philadelphia area.  This TSD reviews Delaware’s documentation of
modeling and WOE that attempts to show attainment of the eight-hour
ozone standard by the 2009 ozone season.  All of the states in the
northeastern United States cooperated via the Ozone Transport
Commission’s (OTC) Modeling Committee to prepare the modeling that was
performed by the New York State Department of Environmental Conservation
(NYSDEC) and supporting organizations.

Delaware believes that their SIP modeling when combined with additional
supporting analyses (Weight of Evidence) (WOE) “Demonstrates that the
entire Philadelphia-Wilmington-Atlantic City, PA-NJ-DE-MD moderate
non-attainment area will attain the 8-hour ozone NAAQS in 2009”.  

What Are The Components Of A Modeled Attainment Demonstration?

Modeling Process Overview                                               
                                                                        
                  

The OTC Modeling Committee used the Community Multi-scale Air Quality
model (CMAQ) version 4.4 as its photochemical grid model.  The model
uses simulations of chemical reactions, emission of ozone precursors and
a sophisticated meteorological model to produce ozone concentrations
over the eastern United States.  The meteorological data used in the
meteorological model was for the base year ozone season of 2002.  The
photochemical grid model was run with the base year meteorology and base
year emissions to determine if the model performance was satisfactory. 
Once the model performance was determined to be adequate, 2009 ozone
concentrations were modeled by running the model with projected
emissions for 2009 and the original 2002 meteorology.    The meteorology
was held constant so that the results of changing the emissions would
not be influenced by changing meteorology.  The EPA modeling guidance
recommends that states use the modeled ozone concentrations in a
relative sense and not rely solely on the concentrations the model
predicts for the attainment year (2009).  The relative response factor
(RRF) used in the modeled attainment test is computed by taking the
ratio of the mean of the 8-hour daily maximum predictions in the future
to the mean of the 8-hour daily maximum predictions with baseline
emissions, over all relevant days. The base year ozone design values at
each monitor are then multiplied by the monitor-specific RRF to produce
an attainment year design value.  If the result is less than or equal to
84 ppb, at all locations, the modeling system will have predicted
attainment of the ozone air quality standard. 

The Regulatory Framework

Section 110 (a) (2) (k) of the Clean Air Act requires states to prepare
air quality modeling to show how they will meet ambient air quality
standards.   EPA determined that areas classified as ‘moderate’ or
above must use photochemical grid modeling or any other analytical
method determined by the Administrator to be at least as effective to
demonstrate attainment of the ozone health-based standard by the
required attainment date (40 CFR 51.908, published at 70 FR 71612 on
November 29, 2005).  In 40 CFR 51.903, published at 69 FR 23951 on April
30, 2004, EPA specified how areas would be classified with regard to the
eight-hour ozone standard set by EPA in 1997.  At 69 FR 23858, on April
30, 2004, EPA followed these procedures and classified the
Philadelphia-Wilmington-Trenton, PA-NJ-DE-MD and ozone nonattainment
area as moderate, so they must attain the 1997 eight-hour ozone standard
by June 2010.  Since the attainment date is June 2010 for moderate
areas, states must achieve emission reductions by the ozone season of
2009 in order for ozone concentrations to be reduced, and show
attainment, during the last complete ozone season before the 2010
deadline.  

The modeling guidance lists nine steps for preparing modeling to
demonstrate attainment of the ozone standard.  

1. Develop a conceptual description of the problem to be addressed.

2. Develop a modeling/analysis protocol.

3. Select an appropriate model to support the demonstration.

4. Select appropriate meteorological time periods to model.

5. Choose an appropriate area to model with appropriate
horizontal/vertical resolution                                          
   and establish the initial and boundary conditions that are suitable
for the application.

6. Generate meteorological inputs to the air quality model.

7. Generate emissions inputs to the air quality model.

8. Run the air quality model with base case emissions and evaluate the
performance.

    Perform diagnostic tests to improve the model, as necessary.

9. Perform future year modeling (including additional control
strategies, if necessary) and

    apply the attainment test.

How Did Delaware Address All Of The Components Of A Modeled Attainment
Demonstration?

The Delaware SIP revision for the Philadelphia area addresses each of
the elements of a modeled attainment demonstration as follows:

Conceptual description of the problem

A conceptual model describes how weather patterns affect the formation
and transport of ozone, accounting for emissions and photochemistry. 
The Delaware SIP revision for the Philadelphia area describes the
climatology that produces high ozone concentrations in the Philadelphia
area.  On warm sunny days, winds from the southwest and west blow
emissions from sources of ozone-forming chemicals both at the surface
and aloft toward the Philadelphia area.  In addition, emissions from
large combustion sources in the Ohio Valley and other areas are brought
eastward by upper level winds to the east coast, augmenting the ozone
formed locally.  A complete conceptual description of the ozone problem
in the Philadelphia area is contained Delaware explains this in Section
7.7 of their SIP revision for the Philadelphia area.

Modeling/analysis protocol

The Ozone Transport Commission’s (OTC’s) Modeling Committee
developed a protocol for modeling the ozone problem in the northeastern
United States.  The modeling protocol is located in Appendix 7-1 of the
Delaware SIP revision for the Philadelphia area.  It adequately
describes how the photochemical grid model will be used in the ozone
attainment plans. 

The Model Used in the Attainment Demonstration

By agreement of the OTC, NYSDEC ran the Community Multi-scale Air
Quality Model version 4.4 (CMAQ) for the states in the northeast ozone
transport region, including Pennsylvania.  CMAQ is an acceptable model,
listed in the photochemical modeling guidance as a currently used
photochemical grid model.  EPA agrees CMAQ is appropriate for this
modeling demonstration.  The inputs to the model are described in
Section 7 of the Delaware SIP revision for the Philadelphia area.

Meteorological Time Periods Used in the Modeling

Section 7.3 of the Delaware SIP revision for the Philadelphia area notes
that the OTC Modeling Committee agreed to model the entire ozone season
of 2002.  Using the entire ozone season of 2002 covers many different
kinds of ozone episodes and exceeds EPA’s recommendations for episode
selection.  2002 was a good year to model because it was the base year
for the attainment plans and as a whole is representative of
meteorological conditions conducive to elevated concentrations of ozone.

Meteorological Data Used in the Air Quality Model

The OTC Modeling Committee decided to use a prognostic meteorological
model that provides life-like meteorological inputs to the photochemical
grid model.  The Pennsylvania State University/National Center for
Atmospheric Research Mesoscale Meteorological Model (MM5) version 3.6
was chosen for the modeling analysis.  The MM5 model provides a
reasonable representation of weather conditions at the surface and
aloft.  A summary of the MM5 application is contained in Section 7.4 of
the Pennsylvania SIP revision for the Philadelphia area. 

 

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

The modeling domain extends from Maine to Florida and out in the
Atlantic Ocean on the east and west to the Mississippi River.  The size
of the modeling domain was made large enough to include all emission
sources that affect ozone formation in the northeastern United States. 
Even this boundary is defined by a larger photochemical modeling domain
that covers much of North America.  Over the northeastern United States,
the model used 12 kilometer grid cells. The Philadelphia area is
included in the 12 kilometer grid cell area.  The OTC Modeling Committee
used a 12-kilometer grid size for the areas in and near its states to
provide a fine enough grid resolution to adequately capture the ozone
patterns experienced in the ozone transport region (OTR).  The OTC
Modeling Committee chose to use a 12 kilometer grid resolution based on
its states’ experiences with the one-hour ozone attainment modeling
that indicated that modeling at finer resolutions than 12 kilometers
would not result in increased model accuracy.  Outside the local areas
the grid resolution used in the modeling is 36 kilometers.  The
selection of model domains and horizontal grid resolution was deemed
acceptable to EPA.  The grid resolution, modeling domain and boundary
conditions are described in Appendix 7-2 of the Delaware SIP revision
for the Philadelphia area.

Vertical resolution is the number of layers and the size of each layer
in the model.  The layers in the photochemical grid model were set up to
be compatible with the model that produced weather conditions for the
photochemical grid model.  The vertical grid resolution used in the
modeling is described in Appendix 7-2 of the Delaware SIP revision for
the Philadelphia area.  The vertical resolution used in the modeling
exercise followed EPA’s modeling guidance and therefore adequately
represents the atmosphere where ozone forms and is transported.  

Emissions Used in the Air Quality Model

The emissions data for 2002 were generated by individual states within
the OTR and assembled and processed through the Mid-Atlantic Northeast
Visibility Union (MANE-VU), a Regional Planning Organization (RPO).
These emissions were then processed by NYSDEC using the SMOKE emissions
processor to provide CMAQ compatible inputs.   The 2002 emissions for
the non-OTR areas within the modeling domain were obtained from the
corresponding RPOs and were processed using SMOKE, in a manner similar
to that of the OTR emissions.

The OTR states, through MANE-VU, contracted MACTEC Federal Programs
(called Contractor) to develop 2009, 2012 and 2018 inventories based
upon 2002 inventories that the states had previously developed for use
in the base-year model work. The Contractor, in consultation with the
states, developed the necessary growth and control factors and applied
to the 2002 inventory. Emissions for mobile sources and the electric
energy generating units (EGUs) was not part of the Contractor’s
effort. .  To generate on-road mobile emissions, the states provided the
Virginia Department of Environmental Quality (VADEQ) and the Northeast
States for coordinated Air Use Management (NESCAUM) with appropriate
MOBILE 6 input files along with the projected vehicle miles traveled
(VMT), which was coupled with hourly “gridded temperature”
information. As for the emissions from the EGU sector, the inter-RPO
work group utilized the Integrated Planning Model (IPM) to develop the
state and unit-level emissions.  These inventories are identified as
2009 on the way (2009OTW), since they reflect all emission control
measures that were promulgated or would become effective on or before
2009.

Details of emissions processing are provided in Appendices 7-7, 7-8 and
7-9 of the Delaware SIP revision for the Philadelphia area.

Base Case Run Model Performance Evaluation

NY DEC performed a model evaluation for the OTC to determine how well
CMAQ reproduced the Philadelphia nonattainment area’s 2002 ozone
season concentrations.  Model evaluation followed performance statistics
outlined in EPA modeling guidance.

The OTC’s evaluation included performance statistics for all monitors
inside the Philadelphia nonattainment area.  The model performance
evaluation results are presented in included in Appendix 7-6 the
Delaware SIP.  The analysis indicates the modeling system does an
adequate job of estimating the eight-hour surface ozone concentrations
throughout the Philadelphia area.

2009 Control Case Modeling and the Modeled Attainment Test

The photochemical grid model used with projected emissions for 2009,
including emission changes due to regulations the states are planning to
implement and expected growth by the 2009 ozone season.   

Meteorological conditions from 2002, the same as the base year modeling,
were used in the 2009 modeling. Using the base case meteorology allows
the effect of changes in states’ emissions to be determined without
being influenced by yearly fluctuations in meteorology and is consistent
with EPA guidance.    

As described earlier in this TSD, the attainment test used in the
Philadelphia area                    modeling demonstration involved the
application of model-based RRFs to base year design values at each
monitor to produce projected future year design values (2009).  The
projected 2009 design values represent design values that should result
from emission controls Delaware and other states planned to have in
place in 2009.  The 2009 design values should be less than or equal to
84 ppb at all monitoring stations to meet the attainment test.  The SIP
modeling predicts that in 2009, Philadelphia area will not pass the
attainment test since design values are projected to be over the 84 ppb
standard (see Table 1 below). 

Table 1 Modeled vs. Monitored Design Value

Monitor	Site ID Number	2009 Modeled   DV

Fairhill - CECIL CO, MD	240150003	81

Colliers Mills - OCEAN CO, NJ	340290006	91

Rider - MERCER CO, NJ	340210005	86

Ancora State Hospital - CAMDEN CO, NJ	340071001	87

Camden - CAMDEN CO, NJ	340070003	88

Clarksboro - GLOUCESTER CO, NJ	340155001	88

Bristol - BUCKS CO, PA	420170012	88

Northeast Airport - PHILADELPHIA CO, PA	421010024	87



                                                                        
                                                       Summary of
Photochemical Grid Modeling Results

In summary, the basic photochemical grid modeling presented in the
Delaware SIP revision for the Philadelphia area meets EPA’s guidelines
and when used with the methods recommended in EPA’s modeling guidance,
is acceptable to EPA.  When EPA’s attainment test is applied to the
modeling results, the 2009 ozone design value is predicted to be 91 ppb
in the Philadelphia-Wilmington-Atlantic City, PA-NJ-DE-MD ozone
nonattainment area.  Thus, based on EPA’s modeled attainment test,
Philadelphia area has not demonstrated that it will reach attainment of
the ozone standard in the 2009 attainment year with the modeled emission
reduction strategies committed to by the OTC states.  To make a case for
attainment, Delaware decided to use WOE to demonstrate attainment of the
ozone standard in the Philadelphia area.   

Weight of Evidence Demonstration

EPA(s modeling guidance describes how to use a photochemical grid model
and additional analytical methods to complete a WOE analysis to estimate
if emissions control strategies will lead to attainment of the
eight-hour NAAQS for ozone.  A WOE analysis is a supporting analysis
that helps to determine if the results of the photochemical modeling
system are correctly (or not correctly) predicting future air quality. 

The WOE portion of the Delaware SIP revision for the Philadelphia area
describes the analyses performed, databases used, key assumptions and
outcomes of each analysis, and why the evidence, viewed as a whole,
supports a conclusion that the Philadelphia nonattainment area will
attain the NAAQS despite the model prediction that some monitors’
future design values exceed the current eight-hour ozone standard.

Table 2 outlines under what circumstances a WOE demonstration is needed.
 Model-predicted design values are summarized above in Table 1.  Of the
twenty-two (22) ozone monitors in the Philadelphia nonattainment area
only eight (8) exceed the threshold (82 ppb) where the EPA modeling
guidance recommends a WOE demonstration.  Three of the monitors fall
within the 82-87 ppb threshold outlined in Table 1 and four others fall
within the last category listed in the WOE table included in the US EPA
guidance.   The Roxboro monitor is excluded from the WOE analysis since
its current design value is significantly lower than the eight-hour
standard (modeled 82 ppb, actual monitored 78 ppb).

The WOE analysis for the remaining seven monitors includes the
following:

A comparison of model-predicted 2009 ozone design values and monitored
design values for 2006, 

An analysis of recent ozone trends in the Philadelphia nonattainment
area,

Alternative methods for calculating the 2009 ozone design value,

An analysis of model-predicted regional transport,

University of Maryland’s analysis of model sensitivity to emission
changes.

Table 2. EPA Guidelines for Supplemental Analyses and Weight of Evidence
Determinations

Results of Modeled Attainment Test	Supplemental Analyses

Future Design Value < 82 ppb, all monitor sites	Basic supplemental
analyses should be completed to confirm the outcome of the modeled
attainment test

Future Design Value 82 - 87 ppb, at one or more sites/grid cells	A
weight of evidence demonstration should be conducted to determine if
aggregate supplemental analyses support the modeled attainment test

Future Design Value > 88 ppb, at one or more sites/grid cells	More
qualitative results are unlikely to support a conclusion differing from
the outcome of the modeled attainment test.

Modeled Concentrations and Current Design Values

Table 3 below lists the OTC modeled 2009 design values and the 2006 and
2007 design values and 2008 preliminary design values.  The modeled 2009
and projected 2006 design values are surprisingly close to one another
with most modeled concentrations slightly lower than the projected 2006
design values.  The Delaware SIP revision for the Philadelphia area
states that the 2006 design values suggest additional reductions over
the next three ozone seasons might bring several of the monitors that
are currently just over the eight-hour ozone standard into compliance.
If this is true, air monitoring data from 2007 and 2008 should support
Delaware’s conclusion, but it does not.

Table 3.  Modeled vs. Monitored Design Values

Monitor	Site ID Number	2009 Modeled   DV	2006 Monitored DV	2007
Monitored DV 	Preliminary 2008  Monitored DV

Fairhill - CECIL CO, MD	240150003	81	90	93	90

Colliers Mills - OCEAN CO, NJ	340290006	91	93	92	87

Rider - MERCER CO, NJ	340210005	86	87	91	87

Ancora State Hospital - CAMDEN CO, NJ	340071001	87	89	88	86

Camden - CAMDEN CO, NJ	340070003	88	84	88	90

Clarksboro - GLOUCESTER CO, NJ	340155001	88	86	91	87

Bristol - BUCKS CO, PA	420170012	88	86	92	92

Northeast Airport - PHILADELPHIA CO, PA	421010024	87	90	91	89



Delaware did not have access to 2007 ambient data or preliminary 2008
ambient data when it wrote its SIP revision for the Philadelphia area. 
However, even at that time, information was available that showed a
comparison between a 2006 design value with modeled 2009 results was
biased toward an optimistic result because the summer of 2004 was a much
cooler than normal (according to the National Climatic Data Center, the
summer of 2004 was ranked as the 20th coolest of the last 110 years in
the Northeast).  Data from 2005 and 2006 showed that the number of 4th
high values over the ozone standard was much higher in 2005 and 2006
than 2004’s in the Philadelphia nonattainment area.  (See Figure
below)  We also know, based on the 2007 design values and the
preliminary 2008 design values, that the 2006 design values were biased
low due to the extremely low values measured in 2004 due to cool weather
conditions in 2004.   

Based on the 2007 and preliminary 2008 ambient data, the modeled 2009
modeled design values now appear to be either reasonable or maybe even
too low in some cases.  

When the 2007 and preliminary 2008 design values (92 ppb) and the
preliminary 2008 4th high ozone value (90 ppb) are examined, it is not
reasonable to assume the Philadelphia area will be in attainment of the
ozone standard after the 2009 ozone season.  The 2008 design values also
suggest that the 2009 model-predicted design values are very reasonable
and do not seem to be over-predicted by the model. 

Design Value Trends

The Delaware SIP revision for the Philadelphia area states that there
have been significant declines in the Philadelphia nonattainment
area’s eight-hour ozone design values over the last several decades. 
Significant declines did occur after enactment of the NOx SIP Call in
the 2003/2004 period.  The data in Table 4 show that ozone design values
have decreased sharply after 2003.  EPA’s study of the effect of the
NOx SIP Call has documented the beneficial effects of the NOx SIP Call
emission reductions that were in place by the 2004 ozone season across
the eastern United States 

 

Since 2004, design values have not continued to decrease as sharply for
several reasons.  First, the 2004 ozone season was cooler than normal,
so the decreases in ozone due to additional controls have been offset
somewhat by higher concentrations as the areas have returned to more
typical summertime conditions.  Second, current reductions in ozone and
precursors transported from the Midwest are not as large as the
decreases that occurred while the NOx SIP Call was implemented.  Even if
EPA’s Clean Air Implementation Rule (CAIR) is implemented, most of the
ozone season reductions due to CAIR are not likely to be implemented
until after 2009.  Any additional reductions in ozone in the 2008 and
2009 ozone seasons will be limited to mostly local controls in each
nonattainment area and a few reductions in major sources due to
States’ initiatives in the Ozone Transport Region.  

While the observed design values for 2007 are only 1 to 3 ppb higher
than the model-predicted value for 2009, an ozone reduction of 9 ppb
would be needed to reduce design values to attainment by 2009 and 6 ppb
to reduce the 4th high value to less than or equal to 84 ppb (level
needed to qualify for a one year extension of the attainment date). 
Unless emission reductions that will occur between now and 2009 are to
equal to, or more effective than the reductions from 2002 to 2007 when
ozone decreased by up to 10 ppb, the Philadelphia nonattainment area
will not reach attainment by the 2009 ozone season. Even if local
controls in 2008 and 2009 could reduce emissions by the amount of local
reductions observed during the NOx SIP Call implementation years,
comparable reductions in ozone needed to attain the standard in 2009
will not happen since the present ozone trends will be not be assisted
by the large reductions in transported ozone and precursors that
occurred before 2005.  The recent trend in 8-hour ozone concentrations
at the Philadelphia area ozone monitors presented in Table 4 indicate
that attainment of the 8-hour ozone standard by the end of the 2009
ozone season is unlikely.

Site	2003	2004	2005	2006	2007	Preliminary 2008

                         Camden	101	93	85	84	88	88

Ancora S.H. 	101	95	91	89	88	86

Clarksboro	98	94	88	86	91	87

Rider	99	91	86	87	91	87

Colliers Mills	106	99	94	93	92	87

Bristol	100 	93 	86	86	92	92

NE Airport	97	95	90	90	91	89

Fairhill	98	92	89	90	93	90

 Table 4. Recent Ozone Design Value Trends

                                                                        
                                                                        

                                                                        
                                                                        
                                                        

                                                                        
                                                                        
                                                                        
                                                                        
                                                            

 Alternative Baseline Design Value     

                                                                        
                                                                        
                          Seven monitors within the Philadelphia area
are projected to exceed the current eight-hour ozone standard following
the US EPA’s recommended method.  The recommended baseline
concentration used in the attainment demonstration is the average of the
three eight-hour ozone design values that include the emission base year
(2002).  Thus the baseline concentration is the average of the 2002,
2003 and 2004 eight-hour ozone design values.

Delaware asserts that by using the US EPA recommended method for
calculating a monitor’s baseline concentration, undue weight is placed
on the 2002 ozone season, one of the worst ozone seasons since the late
1990s.  As alternative to the US EPA’s baseline concentration
calculation, Delaware took the straight average of the 4th highs over
the same years (2000-2004).  This approach weighs each year equally. 
EPA did not recommend this method as it is not as robust as EPA’s
recommended method.  The Delaware method gives equal weight to the 2004
ozone season that has the advantage of emission reductions resulting
from the NOx SIP Call.  Additionally, in 2004 the Philadelphia area had
the fewest number of days with eight-hour ozone 85 ppb or above during
the period 1980 to 2007.  According to the National Climatic Data
Center, the summer of 2004 was ranked as the 20th coolest of the last
110 years in the Northeast.  In the Midwest, upwind from the northeast
and a significant source of ozone precursors, 2004 was ranked 5th
coolest.  2004 was significantly cooler than normal, which means 2004
was not typical according to the conceptual model described in the SIP
revision for the Philadelphia area.  With a cooler than usual summer and
lower concentrations due to implementation of the NOx SIP Call,
including 2004 in a pre-control strategy baseline creates a low bias for
any averaging method that includes it equally with other years.  An
examination of the recent design value trends presented earlier in Table
4 shows that the extraordinary decrease in ozone in 2004 does not
continue in the following years, when increasing ozone occurred while
emissions did not increase.  For these reasons, EPA believes the
recommended method for calculating the baseline design value provides
more stability to the baseline design value and better represents the
meteorology and emissions of the 2002 emissions baseline year.

Methods (see below) equally as reasonable as Delaware’s method that
were not evaluated in the Delaware SIP revision for the Philadelphia
area produce baseline design values the same as EPA’s recommended
method or even higher for the  peak ozone site of the Philadelphia area
ozone nonattainment area at Colliers Mills.  

EPA guideline method baseline = 105.7 ppb 

Delaware’s alternative baseline = 104.2 ppb

2002 design value used by EPA and the State for designation purposes =
113 ppb

2003 design value, centered on 2002 = 106 ppb.  

                                                                        
                                                                 The
baseline value calculated using EPA’s recommended method is well
within the range of values calculated by other methods, so using EPA’s
recommended method is reasonable and does not produce an anomalous
value.  

Table 5 lists the alternative baseline value and the projected 2009
concentration for the seven monitors that are projected to exceed the
current eight-hour ozone standard in 2009.  This reduces the modeled
2009 values slightly but still leaves them close to the preliminary 2008
design values.  Even with this adjustment the 2009 Colliers Mills
monitored design value (90 ppb) still remains above the highest
concentrations listed in EPA’s WOE cut-offs.

 

Table 5. Alternative Baseline Concentration Analysis

Site Name	State	Alternative Baseline	OTW/OTB V4

RRF	Alternate 2009

Camden	NJ	94.0	0.8996	84

Ancora S.H.	NJ	98.6	0.8733	86

Clarksboro	NJ	96.4	0.9004	86

Rider College	NJ	95.6	0.8908	85

Colliers Mills	NJ	104.2	0.8703	90

Bristol	PA	96.6	0.8976	86

NE Phila	PA	94.6	0.9035	85



  

                                             

Alternative RRF

Delaware calculated 2009 modeled ozone design values by multiplying the
modeling baseline design values with a relative response factor (RRF) as
recommended by EPA’s photochemical modeling guidance.  At each
monitoring site, the state calculates the RRF associated with the
maximum eight-hour ozone concentration in the nine grid cells nearest
with the monitoring site on high ozone days.  Lower RRFs produce larger 
 decreases in ozone between the base and future years.  

The Delaware ozone Plan recalculates RRFs for several different ozone
levels; 2002 baseline model concentrations ( 85 ppb, 2002 baseline model
concentrations ( 90 ppb and 2002 baseline model concentrations ( 95 ppb.
 The idea is to see if the air-quality model predicts more reductions on
days with higher ozone concentrations (more benefit on the worst days). 
Table 6 contains the different RRFs based on the 2002 baseline model
concentrations.

Delaware then recalculated the projected modeled 2009 design values
using the most beneficial alternative RRFs.  This lowered the projected
modeled design values at nearly all seven monitors by 1 ppb, except
Bristol and Colliers Mills which were unchanged. 

Despite the use of alternative methods for calculating the RRF, the
resulting future year concentrations for the Bristol and Colliers Mills
monitors still exceed EPA’s recommended WOE cut-off concentrations
(see above in Table 2). 

Table 6. Alternative Projected 2009 Modeled Values Using Alternative
RRFs

Site	Alt RRF	DV Base	Alt Projected 2009

Camden	0.8915	98.0	87

Ancora S.H.	0.8723	99.7	86

Clarksboro	0.8875	98.0	86

Rider	0.8908	97.7	86

Colliers Mills	0.8703	105.7	91

Bristol	0.8892	99.0	88

NE Airport	0.8991	96.7	86



Combining Alternative Baseline Concentrations and Alternative RRFs

Table 7 contains the projected 2009 modeled design values from combining
the alternative baseline concentrations and the alternative RRF
calculations described in the previous two sections.  The combination of
these two alternative approaches lowers the projected 2009-modeled
concentrations significantly, but still leaves the Colliers Mills
monitor above the highest concentrations listed in the U.S. EPA’s WOE
chart.  Five of the other six monitors still remain above the levels
needed for attainment of the ozone standard. 

 

Table 7. Combined Effects of Alternative Baseline Concentrations and
Alternative RRFs

Site	Alt RRF	Alt DV Base	Alt Projected 2009

Camden	0.8915	94.0	83

Ancora S.H.	0.8723	98.6	86

Clarksboro	0.8875	96.4	85

Rider	0.8908	95.6	85

Colliers Mills	0.8703	104.2	90

Bristol	0.8892	96.6	85

NE Airport	0.8991	94.6	85



Regional Transport Analysis

The NOx SIP Call reduced ozone precursor emissions over a large region
of the eastern US.  These reductions reduced regional transport from the
large power plants along the Ohio River into the Philadelphia
nonattainment area.  Pennsylvania maintains an elevated monitoring site
(Methodist Hill) on South Mountain in south-central Pennsylvania
approximately 40 miles southwest of the City of Harrisburg.  Methodist
Hill sits at approximately 1900 ft above mean-sea level and is in a good
position to sample ozone concentrations entering the eastern ozone
transport region (OTR). 

The Delaware SIP revision for the Philadelphia area contends that one
way to gauge how well the OTC air quality model is simulating regional
transport is to examine how well the modeled 2009 eight-hour ozone
design value compares to the 2006 actual design value.  Table 8 contains
the modeled 2009 concentration and the actual 2006 ozone design value. 
The model appears to be over predicting Methodist Hill’s design value
by approximately 6 ppb.  This seems to indicate the model may not be
adequately characterizing the effects of the NOx SIP Call on upwind
sources (under predicting the benefit).  Delaware contends that the 6
ppb difference represents an overestimation of background (regional)
concentrations entering the eastern OTR and suggests the modeled 2009
design values may be overestimated by as much as 6 ppb within the
Philadelphia Nonattainment Area.  This may be true for this particular
monitor, but it is inappropriate to generalize this statement to the
entire Philadelphia area based on one monitor. Additionally, when a
comparison is made using the 2007 and preliminary 2008 design values for
the Methodist Hill monitor, the difference is only 4 ppb, which is
closer to the projected 2009 model-predicted design value and not nearly
enough of a reduction to reduce modeled 2009 modeled ozone design values
to levels needed to demonstrate attainment.   

Table 8. Modeled 2009 vs. Monitored Ozone Design Values at Methodist
Hill

EPA Baseline	RRF OTB/OTW V4	Modeled 2009	2006 DV	2007 DV	Preliminary
2008 DV

90.6	0.8488	76	70	72	72



Analysis of Air-Quality Model Sensitivity to Emission Changes

The Delaware SIP revision for the Philadelphia area cites a study of the
2003 Northeast Blackout (Marufu et al., 2004) that shows that the
blackout caused a drop of at least 7 ppb ozone, and likely considerably
more, and a modeling study of the same event (Hu et al., 2006) that used
CMAQ to predicted  only a 2.2 ppb change.  The modeled reduction in
ozone due to the blackout on August 14, 2003 was compared to the
difference in aircraft measurements of ozone made by the University of
Maryland at Selinsgrove PA on the blackout day and on reference
non-blackout days with similar meteorology.  Modeled ozone reductions
due to the large number of power plant shutdowns during the blackout
were on the order of 2.2 ppb while reductions of up to 7 ppb were noted
in the aircraft data when measurements made on the blackout day were
compared to measurements made on non-blackout days (August 4, 2002 and
August 3, 2005) that had meteorology thought to be similar to the
meteorology on the blackout day.  The University of Maryland concluded
air-quality models such as CMAQ may be under predicting ozone reductions
due to control programs such as the NOx SIP Call by up to a factor of
two.  To be conservative, the University of Maryland lowered their
estimate of model under- prediction of ozone reductions from 100% to
50%.  Application of the 50 % under prediction results in adjusted
modeled 2009 design values of less than 85 ppb for all WOE monitors
inside the Philadelphia Nonattainment Area (see Table 9 below). 

Given the CMAQ 2009 design value predictions and the 2007 and
preliminary 2008  monitored 2008 design values shown earlier in Table 3,
it appears that the model is doing a good job predicting the future year
design values in the Philadelphia area.  CMAQ predicted a ~10-13 percent
ozone reduction for the 2002-2009  time period which is very close to
the measured ozone reductions for the same period.  For this reason and
for other reasons detailed below, we believe the 50 % adjustment to the
2009 design values discussed earlier to be inappropriate.   

∆ ppb	1.5 x ∆ ppb	Adjusted 2009 DV

Camden	98.0	88	-10.0	-15.0	83.0

Ancora S.H.	99.7	87	-12.7	-19.0	80.7

Clarksboro	98.0	88	-10.0	-15.0	83.0

Rider	97.7	87	-10.7	-16.0	81.7

Colliers Mills	105.7	91	-14.7	-22.0	83.7

Bristol	99.0	88	-11.0	-16.5	82.5

NE Airport	96.7	87	-9.7	-14.5	82.2

Table 9. Modeled Design Value Adjustments Due To Model Under-Prediction
of          Benefits

Analysis of the Blackout Study Adjustments

Measured changes in ozone at Selinsgrove PA during the August 14, 2003
blackout are used to argue that CMAQ does not respond adequately to
upwind emissions reductions.  In order to determine what ozone should
have been if the blackout had not occurred, the University of Maryland
(UM) paper compared ozone measurements on the blackout day with ozone
measurements on meteorologically similar days.  As a result of this
comparison, the UM paper concludes that ozone should have been reduced
by twice as much as the amount predicted by photochemical grid modeling
done by Georgia Tech (Hu et al., 2006) for the blackout day.

The vast majority of emissions reductions from the August 14, 2003
Northeast blackout were from NY and Ontario (not the Midwest). The ozone
and ozone precursors were likely reduced in central PA, but the
emissions reductions were generally not coming from the locations that
transport ozone into the Baltimore or Philadelphia Areas. 

Even if the blackout had a large impact in central PA, there is no way
to translate that impact on ozone to the entire Philadelphia area.  EGU
emissions likely account for a large fraction of the ozone in rural PA,
but a much smaller fraction of ozone in the urban Northeast corridor. 
In the urban corridor, transported ozone may be important, but NOx from
mobile sources is still the biggest source of ozone in the Philadelphia
area (See Tables 3-4, 6-10, 6-11 and 6-12 in the Delaware SIP revision
for the Philadelphia area).  .   

This is not the case in rural PA. 

The only modeling of the blackout study was done by Georgia Tech (Hu et
al., 2006). 

The Georgia Tech paper argues that because the emissions reductions were
coming from NY and Ontario (and not the Midwest), they would not expect
a large ozone reduction in the urban areas.  Georgia Tech's modeling
only saw a 2.2 ppb ozone reduction in the Selinsgrove PA area.  . 
Additionally, the Georgia Tech modeling used the SAPRC chemistry.  No
additional modeling was performed by the University of Maryland to
support their conclusions from the blackout study.  No effort was made
to separate out the impacts of transported ozone and precursors from
local sources.  

 

As mentioned above August 4, 2002 and August 3, 2005 were the reference
days with ostensibly similar meteorology to the August 15, 2003 blackout
meteorology that the University of Maryland used in their analysis.  EPA
understands that comparing different days with similar weather
conditions is only an approximation as no two weather patterns are
exactly alike.  It is almost impossible to find a period with identical
meteorological conditions, much less the same chemical state of the air
parcels entering the regions as occurred on August 15, 2003. 
Furthermore, the weather on the days used to compare with the blackout
day was significantly different because the wind trajectories on the
blackout day were mostly from Western New York and Northern Ontario, as
opposed to northern Ohio and Western Pennsylvania on the comparison
days.  The reduction in ozone on the blackout day would be less than if
the blackout occurred on the comparison days since the blackout day’s
emissions came from an area with lower emissions than the comparison
days.  For these reasons, and other discussed in the Georgia Tech paper,
the modeling may not be wrong when it predicts only a 2.2 ppb decrease
in ozone on the blackout day in the area upwind of the northeast
corridor’s ozone nonattainment areas.  Also, the studies compared the
model with observations in central Pennsylvania, not the northeast
corridor.  Central Pennsylvania is not representative of ozone impacts
affecting the urban areas of the northeast corridor, where mobile source
NOx emissions dominate (see Tables 3-3, 3-4 , 6-11 and 6-12 in the
Delaware SIP revision for the Philadelphia area).  Mobile source
emissions were not significantly affected during the blackout. 
Monitored ozone on the reference days was used to reach the conclusion
that the model is under-predicting the benefits of upwind NOx
reductions.  

Even though the recent measurements don’t support the claim of an
underestimate of ozone reduction from the model, there may still be a
valid issue with regard to underestimation of upwind ozone reductions. 
If the model underestimates the impact of upwind NOx reductions, then
there may be a compensating error which overestimates the ozone
reductions from other sources.  Therefore it is possible that the
modeling system overestimated the benefits of low level NOx reductions. 
This makes it inappropriate to make any across the board adjustments to
the DV.  Any reductions in DV from transport may need to be compensated
by increases to the DV from other (local low level) sources

Given all of the facts, there is no technical basis for the adjustment
factor that is used to adjust the design values.  Increasing the modeled
ozone reduction by 50% is not conservative and is not supported by the
current modeling contained in the Delaware SIP revision for the
Philadelphia area and ambient measurements.

                                                                        
                                                                        
                                    The Effect of 2007-2009 and
2008-2009 Emission Reductions

Any measures that reduce emissions that cause ozone pollution, even if
not committed to by Delaware will help the Philadelphia area meet the
Clean Air Act’s requirement in section 172 (c) that all nonattainment
areas reach attainment as soon as practicable.  Even if Delaware’s
attainment plan does not conclusively show attainment by 2009 and this
portion of their SIP revision for the Philadelphia area is not approved,
attainment must still occur as soon as possible.

Table 10 below contains NOx and VOC emission reductions that are
expected to occur between 2007-2009 and 2008-2009 in the Philadelphia
area.  A detailed discussion of how EPA estimated these emission
reductions is contained in Attachment 1 of this TSD. Using the
sensitivity of the CMAQ modeling contained in the Delaware SIP revision
for the Philadelphia area to reductions in emissions, EPA estimated the
potential reduction in ozone concentration that might be expected from
emission reductions that will occur between 2007- 2009 and 2008-2009 in
the Philadelphia area.  The modeled ozone reduction factor (0.021
ppb/ton) is estimated by dividing the average model predicted ppb
reduction in design values for all the Philadelphia area monitors by the
total 2002-2009 Philadelphia area emission reductions that were included
in the modeling.  If both VOC and NOx contribute proportionally to ozone
reductions (not necessarily true, but used here as a simplifying
assumption), then EPA’s modeled ozone reduction factor is multiplied
by the total expected 2007-2009 emission reduction (VOC +NOx) expected
to occur in the Philadelphia area.  The result is a 2.25 ppb reduction
in ozone for the 2007-2009 period. When this concentration is subtracted
from the 2007 design value (93 ppb) and the 2007 4th high ozone value
(102 ppb) for the Philadelphia area, the result is 91 ppb and 99 ppb
respectively, which does not show attainment.  When a similar process is
applied to the emission reductions expected to occur between 2008 and
2009 using the preliminary 2008 design value (92 ppb) and the
preliminary 2008 4th high ozone value (90 ppb), the result is projected
2009 design value of 90 ppb and a projected 2009 4th high ozone level of
88 ppb which again does not show attainment.  Additionally, the
potential 2.25 ppb and 1.21 ppb reductions in ozone are probably over
estimates of the actual ozone reductions that could result from
2007-2009 and 2008-2009 emission reductions.  Emission controls between
2007 and 2009 may have the potential to reduce ozone concentrations in
2009, however; the actual ozone reduction will most likely be less than
the ozone reduction predicted from the model based ozone reduction
factor.  The 2007-2009 emission reductions will not be assisted by the
large reductions in transported ozone and precursors that occurred
before 2005 due to the NOx SIP Call that are accounted for in the
modeling.  Any additional reductions in ozone in the 2008 and 2009 ozone
seasons will be limited to mostly local controls in the nonattainment
area and a few reductions in major sources due to States’ initiatives
in the Ozone Transport Region.  

Table 10. Estimated Ozone Benefit from 2007-2009 and 2008-2009 Emission
Reductions

Area Name 	Major NOx Emission Reduction In SIP Revision	NOx Emission
Reduction  (TPD)        (2007-2009)	NOx Emission Reduction  (TPD)       
(2008-2009)	Major VOC Reductions in SIP Revision	VOC Emission Reduction 
(TPD)      (2007-2009)	VOC Emission Reduction  (TPD)      (2008-2009)
Ozone  Benefit   (PPB)  ( .021 ppb/tpd) x (tpd reduction) (2007-2009)
Ozone  Benefit   (PPB)   (.021 ppb/tpd) x (tpd reduction) (2008-2009)

Philadelphia	Fleet Turnover 	43.3	20.31	 Fleet Turnover 	30.88	14.29



 	DE Multi-P	13.01 to 19.01  	9.394 to 11.23  	 





 	Certain Categories of ICI Boilers 	2  (NJ estimate)	2  (NJ estimate)	
The below are all 2008 to 2009





 	 Refinery Enforcement Initiative 	0.3  (NJ estimate)	0.3  (NJ
estimate)	 Refinery Enforcement Initiative	0.9  (NJ estimate)	0.9  (NJ
estimate)



	NOx RACT rule 2006 	2.1  (NJ estimate)

OTC - Consumer Products Phase 2	0.02 (MD estimate)	0.02 (MD estimate)







OTC - PFC Phase 1	0.06  (MD estimate)	0.06  (MD estimate)







OTC - PFC Phase 2	0.03  (MD estimate)	0.03  (MD estimate)







OTC – Industrial Adhesives	0.10  (MD estimate)	0.10  (MD estimate)







EPA MACT Standards	1.1  (NJ estimate)	1.1  (NJ estimate)







Consumer Products 2009 Amendments	0.4  (NJ estimate)	0.4  (NJ estimate)







Adhesives and Sealants	2.2  (NJ estimate)	2.2  (NJ estimate)







Asphalt Paving (Cutback and Emulsified Asphalt)	1.5  (NJ estimate)	1.5 
(NJ estimate)







Portable Fuel Containers (2005 + 2009)	0.8  (NJ estimate)	0.8  (NJ
estimate)







Regulation 1124 Section 46, Crude Oil Lightering Operations

	2.58 (DE estimate)	2.58 (DE estimate)



	Totals ….	66.71	33.84	Totals ….	40.57	23.98	2.25	1.21



In summary, for ozone reductions to continue as they have since 2002,
then emission reductions would have to continue as they have in the
past.  Even if reductions in local emissions continue to be about the
same as past years, reductions in transported emissions will not be at
the same level seen during the implementation of the NOx SIP Call in the
2003-2004 period.  Therefore, we should not assume that a continuation
of local emission reductions per the rate achieved during the earlier
part of this decade will mean the same reduction in ozone now, as
transported ozone will not be reduced at the same rate as in previous
years.  States have not provided quantitative information about how much
emissions in the Ozone Transport Region outside the Philadelphia
nonattainment area will be reduced compared to earlier years.  To make
an estimate of reductions, EPA used information in Delaware SIP revision
for the Philadelphia area and those of surrounding states to calculate
the amount of emission reductions in states’ plans for the
Philadelphia nonattainment area.  

      

Ozone design values in the past five years have decreased enough to get
the Philadelphia ozone nonattainment area more than halfway to attaining
the standard by 2009.  However, that large drop in ozone concentration
was aided by the NOx SIP Call.  The expected emission reductions over
the next two years are less than the past five year period.  There will
be no comparable reduction in long-range transported pollutants
scheduled to occur in the next two years, even if the Clean Air
Interstate Rule (CAIR) program was implemented on schedule. 

EPA believes that the arguments presented above indicate that it is not
likely that planned emission reductions between 2007 and 2009 are
sufficient for the Philadelphia area to attain the ozone standard.      


  

Summary of EPA’s Evaluation Pennsylvania’sDelaware’s WOE Analysis

Delaware’s photochemical grid modeling results predict a 2009
projected design value well above the air quality health standard for
the Philadelphia ozone nonattainment area.  Even after accounting for
the potential ozone benefits from additional measures not included in
the modeling, attainment of the ozone standard by 2009 for Philadelphia
area is not likely.  The Delaware SIP revision for the Philadelphia area
makes a possible, but ultimately, not convincing argument to show that
ozone will be less than or equal to 84 ppb in 2009.  With the modeled
result at least 7 ppb away from attainment, EPA’s modeling guidance
reasonably points out (see Table 2 above) that Delaware needs to supply
a very substantial amount of evidence that the model is significantly
over-estimating future ozone concentrations.  

Modeling studies give conflicting results as to whether Delaware will
get more reductions in ozone than predicted by the model.  Current air
quality data seem to indicate that the model is not significantly over
predicting ozone concentrations in 2009.  Recent air quality data also
confirm the large gap between present air quality and attainment by
2009.  

Air quality data through 2007  and preliminary data for 2008 are so far
above the level needed for attainment that it is doubtful if  those
states in the Philadelphia area will able to implement enough emission
controls to reach the standard by 2009.  The present air quality also
does not support the hypothesis that the models are incorrect.  Present
air quality concentrations should be closer to the standard since the
Philadelphia area is only one year from when it should be attaining the
standard.  Instead, present air quality is very similar to the 2009
model-predicted air quality.  Emission reductions anticipated to occur
between now and 2009 are also not enough to close the ambient air
quality gap.  

The information and calculations provided in the Delaware SIP revision
for the Philadelphia area emphasizes methods or data that support their
claims that is reasonable to assume the Philadelphia nonattainment area
could attain the ozone standard by the deadline.  EPA’s review of the
WOE analyses must evaluate a spectrum of likely alternative
calculations, not only those that tend to enhance arguments that the
area could attain the ozone standard.  As noted before, the method
recommended by EPA’s guidance and other reasonable variations on
EPA’s methods predict the area will not attain the ozone standard by
2009.  Delaware has not provided convincing evidence that the proposed
adjustments to the photochemical grid model’s attainment year forecast
will give a more accurate answer than the calculations based on EPA’s
recommendations in its modeling guidance. Additionally, recent air
quality trends when combined with an analysis of ozone benefits   from
emissions controls expected to come online between now and the start of
the 2009 ozone season do not support the case for attainment.    

  

Impact of the Vacatur of EPA’s Clean Air Interstate Rule (CAIR)

On June 11, 2008 the United States Court of Appeals for the District of
Columbia Circuit issued an opinion vacating the Clean Air Interstate
Rule (CAIR) in its entirety.  This opinion did not become final before
December 23, 2008, when the Court reconsidered its June 11 decision and
issued an opinion remanding CAIR to EPA without vacating.  North
Carolina v. EPA, Docket No. 05-1244 (D.C. Cir. Dec. 23, 2008). The Court
stated in its December 23 opinion that EPA was still under an obligation
to remedy what the Court characterized as “fundamental flaws” with
CAIR, though the Court declined to impose a deadline for EPA to finalize
that remedy.

Whether or not CAIR and its associated emissions reductions remain in
effect does not alter EPA’s analysis of the approvability of the
attainment demonstration.  The modeling and WOE analysis in the proposed
SIP revision took into account the anticipated CAIR reductions.  EPA has
determined that the Area is not likely to attain the standard even with
these reductions.  If additional proceedings were to result in vacatur
of CAIR, our conclusion that the Area us unlikely to attain the NAAQS
would not change; however, loss of the CAIR reductions would further
increase the odds of the Area reaching attainment by the applicable June
2010 deadline.  

                                                                        
                                                       Summary of
EPA’s Technical Findings

The result of the photochemical grid modeling analysis using EPA’s
recommended methods predicts that the Philadelphia area ozone
nonattainment area will not attain the standard in the attainment year
of 2009.  In response to this, Delaware has offered a number of WOE
arguments that when taken together present an outside chance of truly
attaining the ozone standard.  However, EPA’s review of the attainment
demonstration portion of the Delaware SIP revision for the Philadelphia
area determined the following:

The Delaware SIP revision for the Philadelphia area modeling uses an
appropriate photochemical grid model and follows EPA’s guidance
methods and does not predict attainment in 2009.

The Delaware SIP revision for the Philadelphia area relies on emission
reductions resulting from a commitment to adopt and implement a number
of regulations prior to the start of the 2009 ozone season.  Some of
these were included in the photochemical grid modeling.  These
regulations would provide for additional reductions from boilers,
refineries, power generation, consumer products and portable fuel
containers.  Delaware’s own analysis shows that Delaware must adopt
all these measures, and additional measures as well to show attainment
by 2009.   Delaware must adopt these measures as soon as possible to
insure that these emission reductions occur by the start of the 2009
ozone season to have any chance of attaining the standard in 2009.

Regardless of the issues raised by Delaware regarding the performance of
EPA’s recommended air quality models, the 2007 air quality data and
preliminary 2008 air quality data continue to exceed the ozone standard
by a significant margin.  Even a linear comparison of the percentage of
additional emission reductions planned by the state with the needed
improvement in air quality between 2007 and 2009 indicates it is
unlikely that air quality will improve enough for the Philadelphia area
to meet the ozone standard or qualify the area for a one year extension
of the attainment date in 2009.

When comparing the measured ozone concentrations in 2007 to the ozone
concentrations predicted for 2009 by using EPA’s recommended
application of the photochemical grid modeling, the photochemical grid
model does not exhibit the magnitude of inaccuracies suggested in the
Delaware SIP revision for the Philadelphia area.  

In order to insure attainment, Delaware refers to additional measures
which were not included in the original photochemical modeling analysis.
 Most of these measures are not contained in the SIP revision for the
Philadelphia area and are not committed to by Delaware as part of its
attainment demonstration.  While EPA encourages Delaware to continue to
promote these worthwhile and important emission reduction programs, the
amount of tangible air quality benefit is difficult to estimate with any
degree of certainty.  Based on EPA’s estimates of the ozone benefits
that the proposed additional measures could provide, it is not likely
that the additional ozone benefit will be sufficient for the
Philadelphia area to demonstrate attainment of the ozone standard or
qualify for a one year extension of the attainment date in 2009. 
However, implementing these measures will move the State’s
nonattainment areas closer to attainment.   

The Philadelphia area attainment demonstration relied on adjustments to
the baseline design values and on adjustments to the model predicted
relative response factors form the basis of the photochemical modeling
analysis. These adjustments to the base year starting value and the
amount of reduction in ozone from 2002 to 2009 differ from EPA’s
modeling guidance and, more importantly, are not sufficiently justified
and are artificially weighted toward a conclusion that Philadelphia area
will attain the ozone standard in 2009.

The Philadelphia area attainment demonstration relied on research which
evaluated the impact of a widespread power blackout to develop an
alternative approach to estimating anticipated air quality improvements
from upwind power plants.  While EPA believes that this approach
provides some insight into the transport of ozone precursors, a critical
review of all the research available to EPA and Delaware leads EPA to
disagree with the premise that the air quality modeling results should
be adjusted using Delaware’s alternative approaches.

EPA encourages Delaware to implement all of its planned measures as soon
as possible to meet the Clean Air Act’s requirement that all
nonattainment areas reach attainment as soon as practicable.  

Recommendation of Disapproval of Attainment Demonstration

EPA has carefully evaluated the information provided by Delaware and
other information it deems relevant to help understand what the air
quality is likely to be by the 2009 ozone season.  EPA concludes that
Delaware’s SIP revision for the Philadelphia does not satisfy the
Clean Air Act’s requirement that State Implementation Plans provide
for attainment of the National Ambient Air Quality Standard by the
applicable attainment date.

For the Delaware SIP revision for the Philadelphia area to show
attainment of the air quality standard, several factors would have to
coincide.  The photochemical grid model would have to be
under-predicting the change in ozone by at least 40 percent.  All of the
measures Delaware committed to implement as part of its SIP revision for
the Philadelphia area as well as measures not committed to as WOE would
have to be implemented by the beginning of the 2009 ozone season.  To
expect all of these factors to occur is not realistic.  In addition,
based on recent air quality data, Delaware and its surrounding states
would have to reduce emissions by levels equaling perhaps 50 percent of
the total emission reductions contained in the SIP revision for the
Philadelphia area.  These emission reductions would have to occur
between 2007 and the beginning of the 2009 ozone season.  Based on
EPA’s evaluation of Delaware’s rate of progress plans for the
Philadelphia area, it is unlikely that more than 20 percent of the
Philadelphia area’s emission reductions will occur in the next few
years, making it implausible that ozone in the Philadelphia area will be
reduced from 2007 design value of 93 ppb or the preliminary 2008 design
value of 92 ppb to 84 ppb by 2009.  

Based on EPA’s review of the technical information provided in the
Pennsylvania SIP revision for the Philadelphia area, EPA concludes that
the SIP revision for the Philadelphia area does not show attainment of
the ozone standard by 2009 and EPA should disapprove the attainment
demonstration portion of the SIP.

Attachments:

Attachment 1

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

 The guidance also states that additional analyses are recommended to
determine if attainment will be likely, even if the modeled attainment
test is “passed”.  The guidance recommends supplementary analyses in
all cases, and specifically recommends a WOE analysis if modeled ozone
is greater than or equal to 82 ppb.

&

B

n

€

—

ï

ð

摧異

*

¥

º

Â

Ã

Æ

ê

ì

ð

h

hpu

hpu

h-

hpu

hpu

hpu

hpu

ༀຄᄁ㟾$␸䠀$葞Ď葠ﻲ摧戂\	㜀$␸䠀$摧異

摧異

ht	

ht	

摧⧈c§欀幤

ht	

ht	

ht	

ht	

hS

hS

hS

h~V

hS

h~V

ht	

hS

hS

hS

gd~V

␃ጀ¤᐀¤嬀$⑜愀$摧卿

摧卿

摧噾

ht	

hS

hS

hS

hS

hS

摧卿

␃ጀ¤᐀¤嬀$⑜愀$摧卿

␃ጀ¤᐀¤嬀$⑜愀$摧卿

h

h

h

ht	

hÁ

hÁ

hÁ

ht	

 PAGE   8 

