Attachment to:

Technical Support Document For the Modeling and Weight of Evidence (WOE)
Portions of the Document Entitled “Cecil County, Maryland 8-Hour Ozone
State Implementation Plan and Base Year Inventory SIP Revision: 07-05,
June 15, 2007

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”

Technical Support Document for the Modeling and Weight of Evidence (WOE)
Portions of the Commonwealth of Pennsylvania’s Ozone State
Implementation Plan (SIP) Entitled “Commonwealth of Pennsylvania
Department of Environmental Protection State Implementation Plan
Revision: Attainment Plan and Base Year Inventory Bucks, Chester,
Delaware, Montgomery and Philadelphia Counties located in the
Philadelphia-Wilmington-Atlantic City, PA-NJ-DE Eight-Hour Ozone
Nonattainment Area July 2007”

SCOPE:

This attachment supplies some estimates of additional reductions
anticipated to occur between 2007 and 2009 and between 2008 and 2009.

I.  Estimation of emission reductions from fleet turnover between 2007
and 2009.

A.  Background:

A typical State SIP submittal includes the 2002 base year emissions,
2008 projected controlled emissions in the RFP plan and 2009 projected
controlled emissions in the attainment demonstration.   The expected
decrease or reduction in emissions between 2008 and 2009 can be
determined directly.  One way to estimate the emissions reductions for
2007 to 2009 would be to double the reduction from 2008 to 2009.  This
method assumes the change in emissions is linear over time.  

Mobile Source Estimation Methodologies:

 ∙  VMT.

Because the EF generated by MOBILE6.2 is a function of speed, vehicle
type, vehicle age and other factors, the overall emissions are a
summation of many VMT values representing each link or roadway class
multiplied by many emission factors corresponding to the vehicle mix and
speed on that roadway class.  

	Emissions = ∑ (EF  ∙  VMT )

B.  On-Road Mobile Source Emissions Changes Over Time:

In future years the on-road mobile source emissions are affected by
changes in VMT and changes in emission factors.  The TDM projections
tend predict VMT will increase over time while speeds may or may not
change generally decreasing (due to congestion).   The VMT/speed changes
will cause future year projection VMT to be higher resulting in higher
emissions than if VMT remained constant.  The EF values decrease
dramatically because the future year fleet is predicted to be made up of
a higher percentage of newer vehicles then the 2002 fleet (motor
vehicles must be replaced over time as they wear out, corrode, are
involved in collisions, etc.).  These newer vehicles meet the more
stringent standards such as the Tier 2 standards which started only in
model year 2004 and the heavy-duty standards which took effect with the
model year 2002 standards with more stringent standards applying to
model year 2007 and newer vehicle.  The effect of the decreasing EFs is
a net overall decrease in projected on-road vehicle emissions in future
years (as evinced by the much lower 2009 emission estimates).  

Due to the competing effects of VMT (tending to increase future year
emissions) and of the EF value declines (tending to decrease future year
emissions) one should not expect the change over time to be linear
between 2002 and 2009.

Simple examination of plotting the emissions provided in the plan will
show that the emissions change over time is non-linear.

Consider for example the on-road NOx emissions for the Pennsylvania
portion of the Philadelphia area.  

On-road NOx emissions for the Pennsylvania (in tons per day (TPD))

Year	2002	2008	2009

Emissions of Nox (TPD)	184.66	108.78	101.6





	

Plotted graphically one sees this is the case:

 

If one linearly interpolates a 2008 value between 2002 and 2009 one
gets:

66)/7 ∙ 5], which is more than the value in the SIP.

This suggests that linear interpolation between 2002 and 2009 will over
estimate 2007 emissions levels.

C.  Nonroad Mobile Source Emissions Changes Over Time:

The nonroad mobile source emissions are determined conceptually in the
same manner – EFS are multiplied by an activity levels and these
products for all over all nonroad engine types are summed.  The NONROAD
model packages this process within one model.  

The NONROAD model will predict higher activity levels for future years
but lower EF values as the nonroad fleet turns over.  Generally, the net
effect is an overall decrease in nonroad mobile sources after the 2002
base year. 

Due to the competing effects of increased projected activity levels
(tending to increase future year emissions) and of the EF value declines
(tending to decrease future year emissions) one should not expect the
change over time to be linear between 2002 and 2009.

Simple examination of plotting the emissions provided in the plan will
show that the emissions change over time is non-linear.

Consider for example the non-road NOx emissions for the Pennsylvania
portion of the Philadelphia area.  

Non-road NOx emissions (in TPD)

Year	2002	2008	2009

Emissions of Nox 	70.95	62.67	61.13



If one linearly interpolates a 2008 value between 2002 and 2009 one
gets:

62.53 = 70.95 –  [(61.13 – 70.95)/7 ∙ 5], which is less than the
value in the SIP.

(The non-linearity is not as pronounced graphically for the nonroad case
as for the on-road case.)

D.  Conclusion:

Linear Interpolation between 2002 and 2009 may not be the best method to
estimate 2007 mobile source emissions.

E.  Non-Linear Interpolation:

Because there are three data points (except in the case for Delaware)
one can fit a quadratic equation of the form AY² + By + C = Eyyyy
where:

	A, B and C are constants, 

	Y is the number of years after 2002 (hence Y = 0 for 2002, 5 for 2007,
6 for 2008 and 7 for 2009) and 

	Eyyyy is the year of emissions (yyyy = 2002, 2007, etc.).  

Because Y = 0 for 2002, C = E2002 the base year level.

The formulae used to calculate A and B from E2002, the base year
emissions, E2008,the 2008 projected emissions and E2009, the 2009
projected emissions are derived in Section IV below.

Using the coefficients derived for each State and source sector, a
quadratic equation was used to estimate the 2007 on-road mobile source
and 2007 nonroad mobile source NOx emissions for each state (with the
exception of Delaware’s 2007 nonroad NOx emissions which were a linear
interpolation between 2002 and 2009).

F.  Summary of Interpolation Results:

∙ A + 5 ∙ B + C	50.5*	2.92	64.5	64.16

	Projected 2009 On-Road NOx from the SIP Submittal	49.59

	2.81	62.1	61.13

	Reductions 2007 to 2009	0.91	0.11	2.40	  3.03	6.46



Note:  * Linear Interpolation used:  E2007 = E2002 - (E2009 - E2002)/7 *
5

Summary of Estimated 2007 to 2009 On-Road Mobile Source NOx Reductions

	DE	MD	NJ	PA	Total

A	0.07142857	0.10326190	0.859523	0.780952

	B	-4.6185714	-1.6640476	-16.5738	-17.3323

	C	69.03	14.22	179.8	184.66

	Estimated 2007 Nonroad NOx = 25 ∙ A + 5 ∙ B + C	47.72	8.47	118.42
117.52

	Projected 2009 On-Road NOx from the SIP Submittal	40.2	7.60	105.9
101.60

	Reductions 2007 to 2009	  7.52	0.87	12.52	15.92	36.83



∙ A + 5 ∙ B + C	22.66	7.28	83.78	65.56

	Projected 2009 On-Road NOx from the SIP Submittal	20.98	7.23	76.2	60.11

	Reductions 2007 to 2009	1.68	0.05	7.59	5.45	14.77



Summary of Estimated 2007 to 2009 On-Road Mobile Source VOC Reductions

	DE	MD	NJ	PA	Total

A	0.1152381	0.0278571	0.538095	0.556904

	B	-2.448095	-0.452142	-10.3952	-9.61976

	C	32.38	4.00	91.8	98.76

	Estimated 2007 Nonroad NOx = 25 ∙ A + 5 ∙ B + C	23.02	2.44	53.28
64.58

	Projected 2009 On-Road NOx from the SIP Submittal	20.89	2.29	45.4	58.71

	Reductions 2007 to 2009	2.13	0.24	7.88	5.87	16.11



The expected magnitude of reductions remaining is simply the difference
between the 2007 estimates and the 2009 projections provided in the
State submittals.

G. Data and computations:

The emissions estimates and computations are in the attached printouts
of the worksheet titled “Mobile Emissions Estimates” in a
Microsoft® Excel 2003 spreadsheet, a copy of which is included in the
docket for this action.  

H.  Sources of Data for the Shreadsheets:

Tables 3-3, 3-4, Section 5.5, Tables 6-4, 6-10 and 6-11 from “Final
Submittal, Delaware State Implementation Plan For Attainment Of The
8-Hour Ozone National Ambient Air Quality Standard, Reasonable Further
Progress and Attainment Demonstration, Submitted To U.S. Environmental
Protection Agency, By Delaware Department of Natural Resources and
Environmental Control,” June 2007.

Tables 3-1, 4-3 and 4-4 to “Cecil County, Maryland 8-Hour Ozone State
Implementation Plan and Base Year Inventory,” SIP Revision: 07-05,
June 15, 2007.

Tables 4-1, 4-2, 4-7 to “Commonwealth of Pennsylvania, Department of
Environmental Protection, State Implementation Plan Revision: 
Attainment Demonstration and Base Year Inventory, Bucks, Chester,
Delaware, Montgomery and Philadelphia Counties located in the 

Philadelphia-Wilmington-Atlantic City, PA-NJ-DE Eight-Hour Ozone
Nonattainment Area,” August 2007.

Table 6.3: Projected Emissions and Control Measure Benefits New Jersey
Portion of Southern New Jersey/Philadelphia Nonattainment Area in
Chapter 6.0 “Reasonable Further Progress (RFP)” of “State
Implementation Plan (SIP) Revision for the Attainment and Maintenance of
the Ozone National Ambient Air Quality Standard 8-Hour Ozone Attainment
Demonstration, Final,” dated October 29, 2007.  Available at
http://www.nj.gov/dep/baqp/8hrsip/8hrsip.html#final (last checked
Thursday, December 11, 2008) and at http://www.regulations.gov, docket
EPA-R02-OAR-2008-0497.  

II. Remaining Reductions from Delaware’s Multi Pollutant Regulation:

A.  The Multi Pollutant Regulation and Requirements 

The implementing regulations for Delaware’s Multi Pollutant Regulation
are found in Regulation No. 1146, “Electric Generating Unit (EGU)
Multi-Pollutant Regulation” which was approved into the Maryland SIP
on August 28, 2008 (73 FR 50723). 

The relevant provisions of Regulation No. 1146 are as follows:

Regulation .01 (COMAR 26.11.27.01) provides the following pertinent
definitions in sections A and B as follows:  

Section 2.0 “Applicability” States:

This regulation applies to coal-fired and residual oil-fired electric
generating units located in Delaware with a nameplate capacity rating of
25 MW or greater that commenced operation on or before the effective
date of this regulation.

Section 3.0 “Definitions” conatins the following definitions:

“Unit” means, for the purposes of this regulation, a stationary,
fossil-fuel-fired boiler supplying all or part of its output to an
electric generating device.

Section 4.0 “NOx Emissions Limitations” sets the following
requirments for affected units:

4.1 From May 1, 2009 through December 31, 2011, no unit subject to this
regulation shall emit NOx at a rate exceeding 0.15 lb/MMBTU.

4.1.1 Compliance with the requirements of paragraph 4.1 of this section
shall be demonstrated on a rolling 24-hour average basis.

In Table 6-1 “Delaware 2009 VOC and NOx Emission Projections (TPD) for
Electric Generating Units (EGUs)” Delaware identified eight units
affected by regulation 1146.  These are 

Facility Name Unit Description

City Of Dover - McKee Run Generating Sta Boiler #3 	

Conectiv Delmarva Generation-Edge Moor Boiler #3	

Conectiv Delmarva Generation-Edge Moor Boiler # 4	

Conectiv Delmarva Generation-Edge Moor Boiler # 5	

Indian River Generating Station Boiler #1	

Indian River Generating Station Boiler # 2 	

Indian River Generating Station Boiler # 3	

Indian River Generating Station Boiler # 4	

B.  Actual 2007 Emissions and Preliminary 2008 Emissions Data

For EPA’s review, the 2007 actual emissions levels at the affected
units are known.  This data is available at the EPA’s “Clean Air
Markets - Data and Maps” web portal (  HYPERLINK
"http://camddataandmaps.epa.gov/gdm/index.cfm" 
http://camddataandmaps.epa.gov/gdm/index.cfm ) maintained by EPA’s
Clean Air Markets Division (CAMD).  Data for the eight affected units
was downloaded on December 12, 2009.  The data downloaded included
information identifying the facility and units, NOx mass emissions (in
tons) for each season, and the heat input for the unit (in mmBTU) for
each season.  EPA obtained the data for two seasons:  the regulatory
ozone season (May 1st through September 30th) and the peak ozone season
(June 1st through August 31st).  For the full ozone season, the sum of
the actual emissions for the regulatory ozone season was divided by 153
days in the period to obtain an average daily actual emissions amount
for the 2007 regulatory ozone season.  For the peak ozone season the
average daily emissions was divided by 92 days in the period to obtain
an average daily actual emissions amount for the 2007 peak ozone season.

To estimate the effect of regulation 1146, the heat input for each
season was multiplied by 0.15 pounds NOx/mmBTU to estimate the levels of
NOx emissions in tons for each season that would have been allowed if
Regulation 1148 had been in effect.  

The estimated allowable emissions for each unit were subtracted from the
actual emissions for each season to obtain the estimated reductions over
the period.  The estimated reduction each season was divided by the
number of days in each season (153 for the regulatory ozone season and
92 for the peak ozone season) to obtain an average reduction in tons per
day for that season.

C.  Summary of results 2007 to 2009:

Regulation No. 1146, “Electric Generating Unit (EGU) Multi-Pollutant
Regulation”  2007-2009

	Full Regulatory Ozone Season	Peak ozone Season

Seasonal  2007 NOx Emissions	3994.985	3491.336

Seasonal  2007 NOx Emissions at 0.15 lbs NOx/mmBTU Heat Input	1995.103
1742.436

Estimated seasonal reductions available	1999.882	1748.900

Estimated daily Reductions Between 2007 and 2009	   13.071	      19.010



D.  Summary of results 2008 to 2009:

A similar process to that described in section II. C above was applied
to the preliminary 2008 emissions date available at the “Clean Air
Markets - Data and Maps” web portal which was downloaded December 12,
2008.  The results are:

Regulation No. 1146, “Electric Generating Unit (EGU) Multi-Pollutant
Regulation”  2008-2009

	Full Regulatory Ozone Season	Peak ozone Season

Seasonal  2007 NOx Emissions	3008.733	2075.988

Seasonal  2007 NOx Emissions at 0.15 lbs NOx/mmBTU Heat Input	1519.851
1042.854

Estimated seasonal reductions available	1437.282	1033.134

Estimated daily Reductions Between 2007 and 2009	9.394     	    11.230



E.  Data and computations:

The CAMD data and computations are in the attached printouts of the
worksheets titled “2008 Full Ozone Season,” “2008 Peak Ozone
Season,” “2007 Ozone Full Season,” and “2007 Peak Ozone
Season” in a Microsoft® Excel 2003 spreadsheet, a copy of which is
included in the docket for this action.  

III. Other Regulations taking Effect after January 1, 2007.

A. NOx Measures:

New Jersey identified several additional measures scheduled to take
effect after January 1, 2007.

These include the Refinery Enforcement Initiative, and Certain
Categories of ICI Boilers for which New Jersey projected 0.3 tpd and 2.0
tpd in NOx emissions reductions, respectively, commencing in 2009.  See 
Table 6.3 “Projected Emissions and Control Measure Benefits New Jersey
Portion of Southern New Jersey/Philadelphia Nonattainment Area” in
Chapter 6.0 “Reasonable Further Progress (RFP)” of “State
Implementation Plan (SIP) Revision for the Attainment and Maintenance of
the Ozone National Ambient Air Quality Standard 8-Hour Ozone Attainment
Demonstration, Final,” dated October 29, 2007, of New Jersey’s SIP
submittal.  Additionally, New Jersey’s NOx RACT Rule 2006 (which
includes distributed generation) is discussed in section 4.3.1.4.3 in
Chapter 4.0 Control Measures of the New Jersey SIP submittal.  This rule
has a number of compliance dates in 2007.  See 72 FR 11812, March 14,
2007 and 72 FR 41626, July 31, 2007.  New Jersey estimated reductions in
NOx emissions of 2.1 TPD.   See Table 6.3 “Projected Emissions and
Control Measure Benefits New Jersey Portion of Southern New
Jersey/Philadelphia Nonattainment Area” of New Jersey’s SIP
submittal.

B.  VOC Measures:

The plans for the Philadelphia area identified VOC control regulations
for which the plan projected no reductions in 2008 but did project
reductions for 2009.   These three regulations are summarized in the
table below:

Estimate of VOPC Reductions From Stationary Sources 2007-2009

State	Measure	Reductions	Source

MD	OTC - Consumer Products Phase 2	0.02	Table 6.1: Control Measures
Summary

MD	OTC - PFC Phase 1	0.06	    Ditto

MD	OTC - PFC Phase 2	0.03	    Ditto

MD	OTC – Industrial Adhesives	0.10	    Ditto

NJ	EPA MACT Standards	1.1	Table 6.3: Projected Emissions and Control
Measure Benefits New Jersey Portion of Southern New Jersey/Philadelphia
Nonattainment Area

NJ	Refinery Enforcement Initiative	0.9	    Ditto

NJ	Consumer Products 2009 Amendments	0.4	    Ditto

NJ	Adhesives and Sealants	2.2	    Ditto

NJ	Asphalt Paving (Cutback and Emulsified Asphalt)	1.5	    Ditto

NJ	Portable Fuel Containers (2005 + 2009)	0.8	    Ditto

DE	Regulation 1124 Section 46, Crude Oil Lightering Operations

	2.58	Section 5.1.1 A. Crude Oil Lightering Operations



	Totals …………….	9.69

	

The portable fuel containers (PFC) rules are projected to achieve vastly
(in a relative sense) more reductions in 2009 than in 2008.  Some of
these reductions will only offset growth in activity levels, but for
this analysis EPA will assume all reductions from 2008 to 2009 are net
of growth as the amounts are small in relation to the reductions from
2002 through 2009.  The PFC rules achieve additional reductions after
the implementation date as new containers meeting the PFC rule replace
older ones.  The differences that could be readily discerned from the
plans are:  

Cecil County Maryland:  For the Phase I PFC rule, the State projected
0.26 tons per day reduction in 2008 and 0.32 for 2009 for difference of
0.06.  The M<Maryland Phase 2 PFC rule is an additional rule.

∙Y² + B ∙Y + C = Eyyyy where:

	A, B and C are constants, 

	Y is the number of years after 2002 (hence Y = 0 for 2002, 5 for 2007,
6 for 2008 and 7 for 2009) and 

	Eyyyy is the year of emissions (yyyy = 2002, 2007, etc.).  

Now, E2, E8 and E9 are known.

For the 2002 year, E2, Y is zero by definition.

Thus, C = E2002 because A ∙ 0² + B ∙ 0 + C = E2002.

For 2008, Y = 6 and 

A ∙6² + B ∙6 + C = E2008.           [Eqn 1]

For 2009 Y = 7, and ,   

A ∙7² + B ∙7 + C = E2009.            [Eqn 2]  



-



-



h%

kd™

@

@

@

@

@

@

@

@

@

@

@

@

@

1 by 7 and Eqn 2 by 6 to get:

7 ∙ A ∙6² + B ∙ 42 + 7 ∙ C = 7 ∙ E2008.   [Eqn 1A]

6 ∙ A ∙7² + B ∙ 42 + 6  ∙ C = 6 ∙ E2009.  [Eqn 2A]

Subtract Eqn 1B from Eqn 1A to get:

7 ∙ A ∙6² - 6 ∙ A ∙7² + + 7 ∙ C - 6  ∙ C = 7 ∙ E2008  -
6 ∙ E2009     [Eqn 3]

A ∙ (7  ∙ 6² - 6  A ∙7² ) + C = 7 ∙ E2008  - 6 ∙ E2009

Solving for A:

A =   (7 ∙ E2008  - 6 ∙ E2009 – C)

              (7  ∙ 6² - 6  A ∙7² )

Now, C = E2002 yielding:

A =   (7 ∙ E2008  - 6 ∙ E2009 – E2002)                      [Eqn
4]

              (7  ∙ 6² - 6  A ∙7² )

Taking equation Eqn 2A, A ∙6² + B ∙ 6 +  C = E2008, knowing C =
E2002 and solving for B yields:

B ∙ 6  = E2008 - A ∙6² - E2002 

B= (E2008 - A ∙6² - E2002)       where Eqn 4 is used to calculate A. 
 [Eqn 5]

                     6

PRINOUT OF WORKSHEETS (18 Pages):

“Mobile Emissions Estimates,”  “2008 Full Ozone Season,” “2008
Peak Ozone Season,” “2007 Ozone Full Season,” and “2007 Peak
Ozone Season” in a Microsoft® Excel 2003 spreadsheet, a copy of which
is included in the docket for this action.  

 

   Delaware’s RFP plan relied upon a demonstration that sufficient VOC
reductions would be achieved by 2008 to meet the 15% RFP requirement. 
Therefore, Delaware’s RFP plan did not need to (nor did it) include
2008 projections of NOx emissions for the source sectors of point, area
and nonroad mobile sources.  Delaware’s RFP plan did include 2008
projections of VOC emissions for all source sectors.  Delaware did
include 2008 NOx emissions projections for on-road sources to set NOx
motor vehicle emissions budgets (MVEBs) for 2008 in addition to 2008 VOC
MVEBs.  For this analysis, a linear interpolation was used to estimate
nonroad mobile source NOx emissions in Delaware for 2007.

  A copy of Regulation No. 1146: Electric Generating Unit (EGU)
Multi-Pollutant Regulation

 is attached.  The Delaware SIP can be viewed on line at
http://yosemite.epa.gov/r3/r3sips.nsf/SIPIndex!OpenForm (last accessed
December 12, 2008).  

1” of “State Implementation Plan (SIP) Revision for the Attainment
and Maintenance of the Ozone National Ambient Air Quality Standard
8-Hour Ozone Attainment Demonstration, Final,” dated October 29, 2007.
 

 “Final Submittal, Delaware State Implementation Plan For Attainment
Of The 8-Hour Ozone National Ambient Air Quality Standard, Reasonable
Further Progress and Attainment Demonstration, Submitted To U.S.
Environmental Protection Agency, By Delaware Department of Natural
Resources and Environmental Control,” June 2007.

 PAGE   

 PAGE   11 

PA Portion of Philadelphia Area On-Road NOx

0

50

100

150

200

2000

2002

2004

2006

2008

2010

Year

NOx (TPD)

