Appendix A - Draft Attainment Demonstration for the 1997 8 Hour Ozone National Ambient Air Quality Standard for the Chicago Nonattainment Area

Regional Air Quality Analyses for Ozone, PM2.5, and Regional Haze:
Final Technical Support Document
Table of Contents
12 km
Visibility Improvement (dv)

Regional Air Quality Analyses for

Ozone, PM

2.5

, and Regional Haze:

Final Technical Support Document

April 25, 2008

States of Illinois, Indiana, Michigan, Ohio, and Wisconsin

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Table of Contents

Section-Title

Page

Executive Summary

iii

1.0

Introduction

1.1

SIP Requirements

1.2

Organization

1.3

Technical Work: Overview

2.0

Ambient Data Analyses

2.1

Ozone

2.2

2.5

2.3

Regional Haze

3.0

Air Quality Modeling

3.1

Selection of Base Year

3.2

Future Years of Interest

3.3

Modeling System

3.4

Domain/Grid Resolution

3.5

Model Inputs: Meteorology

3.6

Model Inputs: Emissions

3.7

Base Year Modeling Results

4.0

Attainment Demonstration for Ozone and PM

2.5

4.1

Future Year Modeling Results

4.2

Supplemental Analyses

4.3

Weight of Evidence Determination for Ozone

4.4

Weight of Evidence Determination for PM

2.5

5.0

Reasonable Progress Assessment for Regional Haze

5.1

Class I Areas Impacted

5.2

Future Year Modeling Results

5.3

Weight of Evidence Determination for Regional Haze

104

6.0

Summary

111

7.0

References

114

Appendix I Ozone and PM

2.5

Modeling Results

Appendix II Ozone Source Apportionment Modeling Results

Appendix III PM

2.5

Source Apportionment Modeling Results

Appendix IV Haze Source Apportionment Modeling Results

iii

EXECUTIVE SUMMARY

States in the upper Midwest face a number of air quality challenges. More than 50 counties are

currently classified as nonattainment for the 8-hour ozone standard and 60 for the fine particle

(PM

2.5

) standard (1997 versions). A map of these nonattainment areas is provided in the figure

below. In addition, visibility impairment due to regional haze is a problem in the larger national

parks and wilderness areas (i.e., Class I areas). There are 156 Class I areas in the U.S.,

including two in northern Michigan.

Figure i. Current nonattainment counties for ozone (left) and PM

2.5

(right)

To support the development of State Implementation Plans (SIPs) for ozone, PM

2.5

, and

regional haze in the States of Illinois, Indiana, Michigan, Ohio, and Wisconsin, technical

analyses were conducted by the Lake Michigan Air Directors Consortium (LADCO), its member

states, and various contractors. The analyses include preparation of regional emissions

inventories and meteorological data, evaluation and application of regional chemical transport

models, and collection and analysis of ambient monitoring data.

Monitoring data were analyzed to produce a conceptual understanding of the air quality

problems. Key findings of the analyses include:

Ozone

•

Current monitoring data (2005-2007) show about 20 sites in violation of the 8-hour

ozone standard of 85 parts per billion (ppb). Historical ozone data show a steady

downward trend over the past 15 years, especially since 2001-2003, due likely to

federal and state emission control programs.

•

Ozone concentrations are strongly influenced by meteorological conditions, with

more high ozone days and higher ozone levels during summers with above normal

temperatures.

•

Inter- and intra-regional transport of ozone and ozone precursors affects many

portions of the five states, and is the principal cause of nonattainment in some areas

far from population or industrial centers.

2.5

•

Current monitoring data (2005-2007) show 30 sites in violation of the annual PM

2.5

standard of 15 ug/m

. Nonattainment sites are characterized by an elevated

regional background (about 12 – 14 ug/m

) and a significant local (urban) increment

(about 2 – 3 ug/m

). Historical PM

2.5

data show a slight downward trend since

deployment of the PM

2.5

monitoring network in 1999.

•

2.5

concentrations are also influenced by meteorology, but the relationship is

more complex and less well understood compared to ozone.

•

On an annual average basis, PM

2.5

chemical composition consists mostly of sulfate,

nitrate, and organic carbon in similar proportions.

Haze

•

Current monitoring data (2000-2004) show visibility levels in the Class I areas in

northern Michigan are on the order of 22 – 24 deciviews. The goal of EPA’s visibility

program is to achieve natural conditions, which is about 12 deciviews for these

Class I areas, by the year 2064.

•

Visibility impairment is dominated by sulfate and nitrate.

Air quality models were applied to support the regional planning efforts. Two base years were

used in the modeling analyses: 2002 and 2005. Basecase modeling was conducted to evaluate

model performance (i.e., assess the model's ability to reproduce observed concentrations). This

exercise was intended to build confidence in the model prior to its use in examining control

strategies. Model performance for ozone and PM

2.5

was found to be generally acceptable.

Future year strategy modeling was conducted to determine whether existing (“on the books”)

controls would be sufficient to provide for attainment of the standards for ozone and PM

2.5

and if

not, then what additional emission reductions would be necessary for attainment. Based on the

modeling and other supplemental analyses, the following general conclusions can be made:

•

Existing controls are expected to produce significant improvement in ozone and

2.5

concentrations and visibility levels.

•

The choice of the base year affects the future year model projections. A key

difference between the base years of 2002 and 2005 is meteorology. 2002 was

more ozone conducive than 2005. The choice of which base year to use as the

basis for the SIP is a policy decision (i.e., how much safeguard to incorporate).

•

Modeling suggests that most sites are expected to meet the current 8-hour ozone

standard by the applicable attainment date, except for sites in western Michigan

and, possibly, in eastern Wisconsin and northeastern Ohio.

•

Modeling suggests that most sites are expected to meet the current PM

2.5

standard by the applicable attainment date, except for sites in Detroit, Cleveland,

and Granite City.

The regional modeling for PM

2.5

does not include air quality benefits expected

from local controls. States are conducting local-scale analyses and will use

these results, in conjunction with the regional-scale modeling, to support their

attainment demonstrations for PM

2.5

•

These findings of residual nonattainment for ozone and PM

2.5

are supported by

current (2005 – 2007) monitoring data which show significant nonattainment in

the region (e.g., peak ozone design values on the order of 90 – 93 ppb, and peak

2.5

design values on the order of 16 - 17 ug/m

). It is unlikely that sufficient

emission reductions will occur in the next couple of years to provide for

attainment at all sites.

•

Attainment at most sites by the applicable attainment date is dependent on actual

future year meteorology (e.g., if the weather conditions are consistent with [or

less severe than] 2005, then attainment is likely) and actual future year

emissions (e.g., if the emission reductions associated with the existing controls

are achieved, then attainment is likely). If either of these conditions is not met,

then attainment may be less likely.

•

Modeling suggests that the new PM

2.5

24-hour standard and the new lower

ozone standard will not be met at several sites, even by 2018, with existing

controls.

•

Visibility levels in a few Class I areas in the eastern U.S. are expected to be

greater than (less improved than) the uniform rate of visibility improvement

values in 2018 based on existing controls, including those in northern Michigan

and some in the northeastern U.S. Visibility levels in many other Class I areas in

the eastern U.S. are expected to be less than (more improved than) the uniform

rate of visibility improvement values in 2018. These results, along with

information on the costs of compliance, time necessary for compliance, energy

and non air quality environmental impacts of compliance, and remaining useful

life of existing sources, should be considered by the states in setting reasonable

progress goals for regional haze.

Section 1.0 Introduction

This Technical Support Document summarizes the final air quality analyses conducted by the

Lake Michigan Directors Consortium (LADCO)

and its contractors to support the development

of State Implementation Plans (SIPs) for ozone, fine particles (PM

2.5

), and regional haze in the

States of Illinois, Indiana, Michigan, Ohio, and Wisconsin. The analyses include preparation of

regional emissions inventories and meteorological modeling data for two base years (2002 and

2005), evaluation and application of regional chemical transport models, and analysis of

ambient monitoring data.

Two aspects of the analyses should be emphasized. First, a regional, multi-pollutant approach

was taken in addressing ozone, PM

2.5

, and haze for technical reasons (e.g., commonality in

precursors, emission sources, atmospheric processes, transport influences, and geographic

areas of concern), and practical reasons (e.g., more efficient use of program resources).

Furthermore, EPA has consistently encouraged multi-pollutant planning in its rule for the haze

program (64 FR 35719), and its implementation guidance for ozone (70 FR 71663) and PM

2.5

(72 FR 20609). Second, a weight-of-evidence approach was taken in considering the results of

the various analyses (i.e., two sets of modeling results -- one for a 2002 base year and one for a

2005 base year -- and ambient data analyses) in order to provide a more robust assessment of

expected future year air quality.

The report is organized in the following sections. This Introduction provides an overview of

regulatory requirements and background information on regional planning. Section 2 reviews

the ambient monitoring data and presents a conceptual model of ozone, PM

2.5

, and haze for the

region. Section 3 discusses the air quality modeling analyses, including development of the key

model inputs (emissions inventory and meteorological data), and basecase model performance

evaluation. A modeled attainment demonstration for ozone and PM

2.5

is presented in Section 4,

along with relevant data analyses considered as part of the weight-of-evidence determination.

Section 5 documents the reasonable progress assessment for regional haze, along with

relevant data analyses considered as part of the weight-of-evidence determination. Finally, key

study findings are reviewed and summarized in Section 6.

1.1 SIP Requirements

For ozone, EPA promulgated designations on April 15, 2004 (69 FR 23858, April 30, 2004). In

the 5-state region, more than 100 counties were designated as nonattainment.

The

designations became effective on June 15, 2004. SIPs for ozone were due no later than three

years from the effective date of the nonattainment designations (i.e., by June 2007). The

attainment date for ozone varies as a function of nonattainment classification. For the region,

the attainment dates are either June 2007 (marginal nonattainment areas), June 2009 (basic

nonattainment areas), or June 2010 (moderate nonattainment areas).

A sub-entity of LADCO, known as the Midwest Regional Planning Organization (MRPO), is responsible

for the regional haze activities of the multi-state organization.

Based on more recent air quality data, many counties in Indiana, Michigan, and Ohio were

subsequently redesignated as attainment. As of December 31, 2007, there are 53 counties designated

as nonattainment in the region.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

For PM

2.5

, EPA promulgated designations on December 17, 2004 (70 FR 944, January 5, 2005).

In the 5-state region, 70 counties were designated as nonattainment.

The designations became

effective on April 5, 2005. SIPs for PM

2.5

are due no later than three years from the effective

date of the nonattainment designations (per section 172(b) of the Clean Air Act) (i.e., by April

2008) and for haze no later than three years after the date on which the Administrator

promulgated the PM

2.5

designations (per the Omnibus Appropriations Act of 2004) (i.e., by

December 2007). The applicable attainment date for PM

2.5

nonattainment areas is five years

from the date of the nonattainment designation (i.e., by April 2010).

For haze, the Clean Air Act sets “as a national goal the prevention of any future, and the

remedying of any existing, impairment of visibility in Class I areas which impairment results from

manmade air pollution.” There are 156 Class I areas, including two in northern Michigan: Isle

Royale National Park and Seney National Wildlife Refuge

. EPA’s visibility rule (64 FR 35714,

July 1, 1999) requires reasonable progress in achieving “natural conditions” by the year 2064.

As noted above, the first regional haze SIP was due in December 2007 and must address the

initial 10-year implementation period (i.e., reasonable progress by the year 2018). SIP

requirements (pursuant to 40 CFR 51.308(d)) include setting reasonable progress goals,

determining baseline conditions, determining natural conditions, providing a long-term control

strategy, providing a monitoring strategy (air quality and emissions), and establishing BART

emissions limitations and associated compliance schedule.

1.2 Organization

LADCO was established by the States of Illinois, Indiana, Michigan, and Wisconsin in 1989. The

four states and EPA signed a Memorandum of Agreement (MOA) that initiated the Lake

Michigan Ozone Study (LMOS) and identified LADCO as the organization to oversee the study.

Additional MOAs were signed by the States in 1991 (to establish the Lake Michigan Ozone

Control Program), January 2000 (to broaden LADCO’s responsibilities), and June 2004 (to

update LADCO’s mission and reaffirm the commitment to regional planning). In March 2004,

Ohio joined LADCO. LADCO consists of a Board of Directors (i.e., the State Air Directors), a

technical staff, and various workgroups. The main purposes of LADCO are to provide technical

assessments for and assistance to its member states, and to provide a forum for its member

states to discuss regional air quality issues.

MRPO is a similar entity led by the five LADCO States and involves the federally recognized

tribes in Michigan and Wisconsin, EPA, and Federal Land Managers (i.e., National Park

Service, U.S. Fish & Wildlife Agency, and U.S. Forest Service). In October 2000, the States of

Illinois, Indiana, Michigan, Ohio, and Wisconsin signed an MOA that established the MRPO. An

operating principles document for MRPO, which describe the roles and responsibilities of states,

tribes, federal agencies, and stakeholders, was issued in March 2001. MRPO has a similar

purpose as LADCO, but is focused on visibility impairment due to regional haze in the Federal

Class I areas located inside the borders of the five states, and the impact of emissions from the

five states on visibility impairment due to regional haze in the Federal Class I areas located

outside the borders of the five states. MRPO works cooperatively with the Regional Planning

Organizations (RPOs) representing other parts of the country. The RPOs sponsored several

USEPA subsequently adjusted the final designations, which resulted in 63 counties in the region being

designated as nonattainment (70 FR 19844, April 15, 2005).

Although Rainbow Lake in northern Wisconsin is also a Class I area, the visibility rule does not apply

because the Federal Land Manager determined that visibility is not an air quality related value there.

joint projects and, with assistance by EPA, maintain regular contact on technical and policy

matters.

1.3 Technical Work: Overview

To ensure the reliability and effectiveness of its planning process, LADCO has made data

collection and analysis a priority. More than $7M in RPO grant funds were used for special

purpose monitoring, preparing and improving emissions inventories, and conducting air quality

analyses

. An overview of the technical work is provided below.

Monitoring: Numerous monitoring projects were conducted to supplement on-going state and

local air pollution monitoring. These projects include rural monitoring (e.g., comprehensive

sampling in the Seney National Wildlife Refuge and in Bondville, IL); urban monitoring (e.g.,

continuation of the St. Louis Supersite); aloft (aircraft) measurements; regional ammonia

monitoring; and organic speciation sampling in Seney, Bondville, and five urban areas.

Emissions: Baseyear emissions inventories were prepared for 2002 and 2005. States provided

point source and area source emissions data, and MOBILE6 input files and mobile source

activity data. LADCO and its contractors developed the emissions data for other source

categories (e.g., select nonroad sources, ammonia, fires, and biogenics) and processed the

data for input into an air quality model. To support control strategy modeling, future year

inventories were prepared. The future years of interest include 2008 (planning year to address

the 2009 attainment year for basic ozone nonattainment ares), 2009 (planning year to address

the 2010 attainment year for PM

2.5

and moderate ozone nonattainment areas), 2012 (planning

to address a 2013 alternative attainment date), and 2018 (first milestone year for regional haze).

Air Quality Analyses: The weight-of-evidence approach relies on data analysis and modeling.

Air quality data analyses were used to provide both a conceptual model (i.e., a qualitative

description of the ozone, PM

2.5

, and regional haze problems) and supplemental information for

the attainment demonstration. Given uncertainties in emissions inventories and modeling,

especially for PM

2.5

, these data analyses are a necessary part of the overall technical support.

Modeling includes baseyear analyses for 2002 and 2005 to evaluate model performance and

future year strategy analyses to assess candidate control strategies. The analyses were

conducted in accordance with EPA’s modeling guidelines (EPA, 2007a). The PM/haze

modeling covers the full calendar year (2002 and 2005) for an eastern U.S. 36 km domain, while

the ozone modeling focuses on the summer period (2002 and 2005) for a Midwest 12 km

subdomain. The same model (CAMx) was used for ozone, PM

2.5

, and regional haze.

Since 1999, MRPO has received almost $10M in RPO grant funds from USEPA.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Section 2.0 Ambient Data Analyses

An extensive network of air quality monitors in the 5-state region provides data for ozone (and

its precursors), PM

2.5

(both total mass and individual chemical species), and visibility. These

data are used to determine attainment/nonattainment designations, support SIP development,

and provide air quality information to public (see, for example, www.airnow.gov

Analyses of the data were conducted to produce a conceptual model, which is a qualitative

summary of the physical, chemical, and meteorological processes that control the formation and

distribution of pollutants in a given region. This section reviews the relevant data analyses and

describes our understanding of ozone, PM

2.5,

and regional haze with respect to current

conditions, data variability (spatial, temporal, and chemical), influence of meteorology (including

transport patterns), precursor sensitivity, and source culpability.

2.1 Ozone

In 1979, EPA adopted an ozone standard of 0.12 ppm, averaged over a 1-hour period. This

standard is attained when the number of days per calendar year with maximum hourly average

concentrations above 0.12 ppm is equal to or less than 1.0, averaged over a 3-year period,

which generally reflects a design value (i.e., the 4

highest daily 1-hour value over a 3-year

period) less than 0.12 ppm.

In 1997, EPA tightened the ozone standard to 0.08 ppm, averaged over an 8-hour period

. The

standard is attained if the 3-year average of the 4th-highest daily maximum 8-hour average

ozone concentrations (i.e., the design value) measured at each monitor within an area is less

than 0.08 ppm (or 85 ppb).

Current Conditions:

A map of the 8-hour ozone design values at each monitoring site in the

region for the 3-year period 2005-2007 is shown in Figure 1. The “hotter” colors represent

higher concentrations, where yellow and orange dots represent sites with design values above

the standard. Currently, there are 19 sites in violation of the 8-hour ozone NAAQS in the 5-state

region, including sites in the Lake Michigan area, Detroit, Cleveland, Cincinnati, and Columbus.

Table 1 provides the 4

-highest daily 8-hour ozone values and the associated design values

since 2001 for several high monitoring sites throughout the region.

On March 12, 2008, USEPA further tightened the 8-hour ozone standard to increase public health

protection and prevent environmental damage from ground-level ozone. USEPA set the primary (health)

standard and secondary (welfare) standard at the same level: 0.075 ppm (75 ppb), averaged over an 8-

hour period.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Figure 1. 8-hour ozone design values (2005-2007)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Key Sites

'01

'02

'03

'04

'05

'06

'07

'01-'03

'02-'04 '03-'05 '04-'06 '05-'07

Lake Michigan Area

Chiwaukee

99 116

101

Racine

92 111

Milwaukee-Bayside

Harrington Beach

102

Manitowoc

Sheboygan

102 105

100

Kewaunee

Door County

78 101

Hammond

90 101

Whiting

Michigan City

90 107

Ogden Dunes

85 101

Holland

92 105

Jenison

Muskegon

Indianapolis Area

Noblesville

88 101 101

Fortville

89 101

Fort B. Harrison

87 100

Detroit Area

New Haven

95 102

Warren

92 101

Port Huron

84 100

Cleveland Area

Ashtabula (Conneaut)

97 103

Notre Dame (Geauga)

99 115

103

Eastlake (Lake)

89 104

Akron (Summit)

98 103

Cincinnati Area

Wilmington (Clinton)

Sycamore (Hamilton)

88 100

Hamilton (Butler)

83 100

Middleton (Butler)

Lebanon (Warren)

Columbus Area

London (Madison)

New Albany (Franklin)

90 103

Franklin (Franklin)

Ohio Other Areas

Marietta (Washington)

St. Louis Area

W. Alton (MO)

Orchard (MO)

Sunset Hills (MO)

Arnold (MO)

Margaretta (MO)

Maryland Heights (MO)

4th High 8-hour Value

Design Values

Table 1. Ozone Data for Select Sites in 5-State Region

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Meteorology and Transport:

Most pollutants exhibit some dependence on meteorological

factors, especially wind direction, because that governs which sources are upwind and thus

most influential on a given sample. Ozone is even more dependent, since its production is

driven by high temperatures and sunlight, as well as precursor concentrations (see, for

example, Figure 2).

Figure 2. Number of hot days and 8-hour “exceedance” days in 5-state region

Qualitatively, ozone episodes in the region are associated with hot weather, clear skies

(sometimes hazy), low wind speeds, high solar radiation, and southerly to southwesterly winds.

These conditions are often a result of a slow-moving high pressure system to the east of the

region. The relative importance of various meteorological factors is discussed later in this

section.

Transport of ozone (and its precursors) is a significant factor and occurs on several spatial

scales. Regionally, over a multi-day period, somewhat stagnant summertime conditions can

lead to the build-up in ozone and ozone precursor concentrations over a large spatial area. This

pollutant air mass can be advected long distances, resulting in elevated ozone levels in

locations far downwind. An example of such an episode is shown in Figure 3.

Figure 3. Example of elevated regional ozone concentrations (June 23 – 25, 2005)

Note: hotter colors represent higher concentrations, with orange representing concentrations above the 8-

hour standard

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Locally, emissions from urban areas add to the regional background leading to ozone

concentration hot spots downwind. Depending on the synoptic wind patterns (and local land-

lake breezes), different downwind areas are affected (see, for example, Figure 4).

Figure 4. Examples of recent high ozone days in the Lake Michigan area

Note: hotter colors represent higher concentrations, with orange representing concentrations above the 8-

hour standard

Aloft (aircraft) measurements in the Lake Michigan area also provide evidence of elevated

regional background concentrations and “plumes” from urban areas. For one example summer

day (August 20, 2003 – see Figure 5), the incoming background ozone levels were on the order

of 80 – 100 ppb and the downwind ozone levels over Lake Michigan were on the order of 100 -

150 ppb (STI, 2004).

Figure 5. Aircraft ozone measurements over Lake Michigan (left) and along upwind boundary

(right) – August 20, 2003 (Note: aircraft measurements reflect instantaneous values)

As discussed in Section 4, residual nonattainment is projected in at least one area in the 5-state

region –i.e., western Michigan. To understand the source regions likely impacting high ozone

concentrations in western Michigan and estimate the impact of these source regions, two simple

transport-related analyses were performed.

First, back trajectories were constructed using the HYSPLIT model for high ozone days (8-hour

peak > 80 ppb) during the period 2002-2006 in western Michigan to characterize general

transport patterns. Composite trajectory plots for all high ozone days based on data from three

sites (Cass County, Holland, and Muskegon) are provided in Figure 6. The plots point back to

areas located to the south-southwest (especially, northeastern Illinois and northwestern Indiana)

as being upwind on these high ozone days.

Figure 6 Back trajectory analysis showing upwind areas associated with high ozone

concentrations

Second, to assess the impact from Chicago/NW Indiana, Blanchard (2005a) compared ozone

concentrations upwind (Braidwood, IL), within Chicago (ten sites in the City), and downwind

(Holland and Muskegon) for days in 1999 – 2002 with southwesterly winds - i.e., transport

towards western Michigan. Figure 7 shows the distribution of daily peak 8-hour ozone

concentrations by day-of-week, with a line connecting the mean values. The difference

between day-of-week mean values at downwind and upwind sites indicates that Chicago/NW

Indiana contributes about 10-15 ppb to downwind ozone levels.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Figure 7. Mean day-of-week peak 8-hour ozone concentrations at sites upwind, within, and

downwind of Chicago, 1999 – 2002 (southwesterly wind days)

Based on this information, the following key findings related to transport can be made:

•

Ozone transport is a problem affecting many portions of the eastern U.S. The Lake

Michigan area (and other areas in the LADCO region) both receive high levels of

incoming (transported) ozone and ozone precursors from upwind source areas on many

hot summer days, and contribute to the high levels of ozone and ozone precursors

affecting downwind receptor areas.

•

The presence of a large body of water (i.e., Lake Michigan) influences for the formation

and transport of ozone in the Lake Michigan area. Depending on large-scale synoptic

winds and local-scale lake breezes, different parts of the area experience high ozone

concentrations. For example, under southerly flow, high ozone can occur in eastern

Wisconsin, and under southwesterly flow, high ozone can occur in western Michigan.

•

Downwind shoreline areas around Lake Michigan are affected by both regional transport

of ozone and subregional transport from major cities in the Lake Michigan area.

Counties along the western shore of Michigan (from Benton Harbor to Traverse City, and

even as far north as the Upper Peninsula) are impacted by high levels of incoming

(transported) ozone.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Data Variability:

Since 1980, considerable progress has been made to meet the previous 1-

hour ozone standard. Figure 8 shows the decline in both the 1-hour and 8-hour design values

for the 5-state LADCO region over the last 25 years.

Figure 8 Ozone design value trends in 5-State region

The trend is more dramatic for the higher ozone sites in the 5-state region (see Figure 9). This

plot shows a pronounced downward trend in the design value since the 2001-2003 period, due,

in part, to the very low 4

high values in 2004.

Figure 9. Trend in ozone design values and 4

high values for higher ozone sites in region

The improvement in ozone concentrations is also seen in the decrease in the number of sites

measuring nonattainment over the past 15 years in the Lake Michigan area (see Figure 10).

105

'95-'97

'96-'98

'97-'99

'98-'00

'99-'01

'00-'02

'01-'03

'02-'04

'03-'05

'04-'06

'05-'07

105

115

'95

'96

'97

'98

'99

'00

'01

'02

'03

'04

'05

'06

'07

Design Values

High Values

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Figure 10. Ozone design value maps for 1995-1997, 2000-2002, and 2005-2007

Given the effect of meteorology on ambient ozone levels, year-to-year variations in meteorology

can make it difficult to assess trends in ozone air quality. Two approaches were considered to

adjust ozone trends for meteorological influences: an air quality-meteorology statistical model

developed by EPA (i.e., Cox method), and statistical grouping of meteorological variables

performed by LADCO (i.e., Classification and Regression Trees, or CART).

Cox Method

: This method uses a statistical model to ‘remove’ the annual effect of meteorology

on ozone (Cox and Chu, 1993). A regression model was fit to the 1997-2007 data to relate daily

peak 8-hour ozone concentrations to six daily meteorological variables plus seasonal and

annual factors (Kenski, 2008a). Meteorological variables included were daily maximum

temperature, mid-day average relative humidity, morning and afternoon wind speed and wind

direction. The model is then used to predict 4

high ozone values. By holding the

meteorological effects constant, the long term trend can be examined independently of

meteorology. Presumably, any trend reflects changes in emissions of ozone precursors.

Figure 11a shows the meteorologically-adjusted 4

high ozone concentrations for several

monitors near major urban areas in the region. The plots indicate a general downward trend

since the late 1990s for most cities, indicating that recent emission reductions have had a

positive effect in improving ozone air quality.

A similar model was run to examine meteorologically adjusted trends in seasonal average

ozone. This model incorporates more meteorological variables, including rain and long-distance

transport (direction and distance). Model development was documented in Camalier et al.,

2007. The seasonal average trends are shown in Figure 11b. Trends determined by seasonal

model for the same set of sites examined above are consistent with those developed by the 4

high model.

Chiwaukee, WI

Sheboygan, WI

Cleveland (Ashtabula), OH

Cincinnati (Sycamore), OH

Detroit (New Haven), MI

St. Louis, MO

Indianapolis, IN

Figure 11a. Trends in meteorologically

adjusted 4

high 8-hour ozone

concentrations for seven Midwestern sites

(1997 – 2007)

Chiwaukee, WI

Sheboygan, WI

1998

2000

2002 2004

2006

Seasonal Average

Adjusted

Ozone

for Weather

(ppb)

Unadjusted for Weather

1998

2000

2002

2004

2006

Seasonal Average

Adjusted

Ozone

for Weather

(ppb)

Unadjusted for Weather

Cleveland (Ashtabula), OH

Cincinnati (Sycamore), OH

1998

2000

2002

2004

2006

Seasonal Average

Adjusted

Ozone

for Weather

(ppb)

Unadjusted for Weather

1998

2000

2002

2004

2006

Seasonal Average

Adjusted

Ozone

for Weather

(ppb)

Unadjusted for Weather

Detroit (New Haven), MI

St. Louis, MO

1998

2000

2002 2004

2006

Seasonal Average

Adjusted

Ozone

for Weather

(ppb)

Unadjusted for Weather

1998

2000

2002 2004

2006

Seasonal Average

Adjusted

Ozone

for Weather

(ppb)

Unadjusted for Weather

Indianapolis, IN

1998

2000

2002 2004

2006

Seasonal Average

Adjusted

Ozone

for Weather

(ppb)

Unadjusted for Weather

Figure 11b. Trends in seasonal 8-hour ozone

concentrations for seven Midwestern sites

(1997 – 2007)

CART

: Classification and Regression Tree (CART) analysis is another statistical technique

which partitions data sets into similar groups (Breiman et al., 1984). CART analysis was

performed using data for the period 1995-2007 for 22 selected ozone monitors with current 8-

hour design values close to or above the standard (Kenski, 2008b). The CART model searches

through 60 meteorological variables to determine which are most efficient in predicting ozone.

Although the exact selection of predictive variables changes from site to site, the most common

predictors were temperature, wind direction, and relative humidity. Only occasionally were

upper air variables, transport time or distance, lake breeze, or other variables significant. (Note,

the ozone and meteorological data for the CART analysis are the same as used in the EPA/Cox

analysis.)

For each monitor, regression trees were developed that classify each summer day (May-

September) by its meteorological conditions. Similar days are assigned to nodes, which are

equivalent to branches of the regression tree. Ozone time series for the higher concentration

nodes are plotted for select sites in Figure 12. By grouping days with similar meteorology, the

influence of meteorological variability on the trend in ozone concentrations is partially removed;

the remaining trend is presumed to be due to trends in precursor emissions or other non-

meteorological influences. Trends over the 13-year period at most sites were found to be

declining, with the exception of Detroit which showed fairly flat trends. Comparison of the

average of the high concentration node values for 2001-2003 v. 2005-2007 showed an

improvement of about 5 ppb across all sites (even Detroit).

The effect of meteorology was further examined by using an ozone conduciveness index

(Kenski, 2008b). This metric reflects the variability from the 13-year average in the number of

days in the higher ozone concentration nodes (see Figure 13). Examination of these plots

indicates:

•

2002 and 2005 were both above normal, with 2002 tending to be more severe; and

•

2001-2003 and 2005-2007 were both above normal, with no clear pattern in which

period was more severe (i.e., ozone conduciveness values were similar at most sites,

2001-2003 values were higher at a few sites, and 2005-2007 values were higher at a

few sites).

Given the similarity in ozone conduciveness between 2001-2003 and 2005-2007, the

improvement in ozone levels noted above is presumed to be due to non-meteorological factors

(i.e., emission reductions).

In conclusion, all three statistical approaches (CART and the two nonlinear regression models)

show a similar result; ozone in the urban areas of the LADCO region has declined during the

1997-2007 period, even when meteorological variability is accounted for. The decreases are

present whether seasonal average ozone, peak values (annual 4

highs), or a subset of high

days with similar meteorology are considered. The consistency in results across models is a

good indication that these trends reflect impacts of emission control programs.

Chiwaukee, WI

Sheboygan, WI

Cleveland (Ashtabula), OH

Cincinnati (Sycamore), OH

Detroit (New Haven), MI

St. Louis, MO

Indianapolis, IN

Figure 12. Trends for higher ozone CART

groups (average ozone > 65 ppb) for seven

Midwestern sites (1995 – 2007)

Note: line represents linear best fit

Figure 13. Ozone conduciveness index (and

number of high ozone days) for seven

Midwestern site (1995 – 2007)

Precursor Sensitivity

: Ozone is formed from the reactions of hydrocarbons and nitrogen oxides

under meteorological conditions that are conducive to such reactions (i.e., warm temperatures

and strong sunlight). In areas with high VOC/NOx ratios, typical of rural environments (with low

NOx), ozone tends to be more responsive to reductions in NOx. Conversely, in areas with low

VOC/NOx ratios, typical of urban environments (with high NOx), ozone tends to be more

responsive to VOC reductions.

An analysis of VOC and NO

-limitation was conducted with the ozone MAPPER program, which

is based on the Smog Production (SP) algorithm (Blanchard, et al., 2003). The “Extent of

Reaction” parameter in the SP algorithm provides an indication of VOC and NOx sensitivity:

Extent Range

Precursor Sensitivity

< 0.6

VOC-sensitive

0.6 – 0.8

Transitional

> 0.8

NOx-sensitive

A map of the Extent of Reaction values for high ozone days is provided in Figure 14. As can be

seen, ozone is usually VOC-limited in cities and NOx-limited in rural areas. (Data from aircraft

measurements suggest that ozone is usually NO

-limited over Lake Michigan and away from

urban centers on days when ozone in the urban centers is VOC-limited.) The highest ozone

days were found to be NO

-limited. This analysis suggests that a NOx reduction strategy would

be effective in reducing ozone levels. Examination of day-of-week concentrations, however,

raises some question about the effectiveness of NOx reductions.

Figure 14. Mean afternoon extent of reaction (1998 – 2002)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Blanchard (2004 and 2005a) examined weekend-weekday differences in ozone and NO

in the

Midwest. All urban areas in these two studies exhibited substantially lower (40-60%) weekend

concentrations of NO

compared to weekday concentrations. Despite lower weekend NO

concentrations, weekend ozone concentrations were not lower; in fact, most urban sites had

higher concentrations of ozone, although the increase was generally not statistically significant

(see Figure 15). This small but counterproductive change in

local

ozone concentrations

suggests that

local

urban-scale NO

reductions alone may not be very effective.

Figure 15. Weekday/weekend differences in 8-hour ozone – number of sites with weekend

increase (positive values) v. number of sites with weekend decreases (negative values)

Two additional analyses, however, demonstrate the positive effect of NOx emission reductions

on downwind ozone concentrations. First, Blanchard (2005a) looked at the effect of changes in

precursor emissions in Chicago on downwind ozone levels in western Michigan. For the

transport days of interest (i.e., southwesterly flow during the summers of 1999 – 2002), mean

NOx concentrations in Chicago are about 50% lower and mean ozone concentrations at the

(downwind) western Michigan sites are about 1.5 – 5.2 ppb (3 – 8 %) lower on Sunday

compared to Wednesday. This degree of change in downwind ozone levels suggests a

positive, albeit non-linear response to urban area emission reductions.

Second, Environ (2007a) examined the effect of differences in day-of-week emissions in

southeastern Michigan on downwind ozone levels. This modeling study found that weekend

changes in ozone precursor emissions cause both increases and decreases in Southeast

Michigan ozone, depending upon location and time:

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

•

Weekend increases in 8-hour maximum ozone occur in and immediately downwind of

the Detroit urban area (i.e., in VOC-sensitive areas).

•

Weekend decreases in 8-hour maximum ozone occur outside and downwind of the

Detroit urban area (i.e., in NOx-sensitive areas).

•

At the location of the peak 8-hour ozone downwind of Detroit, ozone was lower on

weekends than weekdays.

•

Ozone benefits (reductions) due to weekend emission changes in Southeast Michigan

can be transported downwind for hundreds of miles.

•

Southeast Michigan benefits from lower ozone transported into the region on Saturday

through Monday because of weekend emission changes in upwind areas.

In summary, these analyses suggest that urban VOC reductions and regional (urban and rural)

NOx reductions will be effective in lowering ozone concentrations. Local NOx reductions can

lead to local ozone increases (i.e., NOx disbenefits), but this effect does not appear to pose a

problem with respect to attainment of the standard. It should also be noted that urban VOC and

regional NOx reductions are likely to have multi-pollutant benefits (e.g., both lower ozone and

2.5

impacts).

2.2 PM

2.5

In 1997, EPA adopted the PM

2.5

standards of 15 ug/m

(annual average) and 65 ug/m3 (24-hour

average). The annual standard is attained if the 3-year average of the annual average PM

2.5

concentration is less than or equal to the level of the standard. The daily standard is attained if

the 98th percentile of 24-hour PM

2.5

concentrations in a year, averaged over three years, is less

than or equal to the level of the standard.

In 2006, EPA revised the PM

2.5

standards to 15 ug/m

(annual average) and 35 ug/m

(24-hour

average).

Current Conditions:

Maps of annual and 24-hour PM

2.5

design values for the 3-year period

2005-2007 are shown in Figure 16. The “hotter” colors represent higher concentrations, where

red dots represent sites with design values above the annual standard. Currently, there are 30

sites in violation of the annual PM

2.5

standard.

Table 2 provides the annual PM

2.5

concentrations and associated design values since 2003 for

several high monitoring sites throughout the region.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Figure 16. PM

2.5

design values - annual average (top) and 24-hour average (bottom) (2005-2007)

2005 BY

2002 BY

Key Site

County

Site ID

'03

'04

'05

'06

'07

'03 - '05 '04 - '06 '05 - '07

Average

w/ 2007

Average

Chicago - Washington HS

Cook

170310022

15.6 14.2 16.9 13.2 15.7

15.6

14.8

15.3

15.2

15.9

Chicago - Mayfair

Cook

170310052

15.9 15.3 17.0 14.5 15.5

16.1

15.6

15.7

15.8

17.1

Chicago - Springfield

Cook

170310057

15.6 13.8 16.7 13.5 15.1

15.4

14.7

15.1

15.0

15.6

Chicago - Lawndale

Cook

170310076

14.8 14.2 16.6 13.5 14.3

15.2

14.8

14.9

15.6

Blue Island

Cook

170312001

14.9 14.1 16.4 13.2 14.3

15.1

14.6

14.8

15.6

Summit

Cook

170313301

15.6 14.2 16.9 13.8 14.8

15.6

15.0

15.2

15.2

16.0

Cicero

Cook

170316005

16.8 15.2 16.3 14.3 14.8

16.1

15.3

15.1

15.5

16.4

Granite City

Madison

171191007

17.5 15.4 18.2 16.3 15.1

17.0

16.6

16.5

16.7

17.3

E. St. Louis

St. Clair

171630010

14.9 14.7 17.1 14.5 15.6

15.6

15.4

15.7

15.6

16.2

Jeffersonville

Clark

180190005

15.8 15.1 18.5 15.0 16.5

16.5

16.2

16.7

16.4

17.2

Jasper

Dubois

180372001

15.7 14.4 16.9 13.5 14.4

15.7

14.9

15.2

15.5

Gary

Lake

180890031

16.8 13.3 14.5

16.8

15.1

14.9

15.6

Indy - Washington Park

Marion

180970078

15.5 14.3 16.4 14.1 15.8

15.4

14.9

15.4

15.3

16.2

Indy - W 18th Street

Marion

180970081

16.2 15.0 17.9 14.2 16.1

16.4

15.7

16.1

16.0

Indy - Michigan Street

Marion

180970083

16.3 15.0 17.5 14.1 15.9

16.3

15.5

15.8

15.9

16.6

Allen Park

Wayne

261630001

15.2 14.2 15.9 13.2 12.8

15.1

14.4

14.0

14.5

15.8

Southwest HS

Wayne

261630015

16.6 15.4 17.2 14.7 14.5

16.4

15.8

15.5

15.9

17.3

Linwood

Wayne

261630016

15.8 13.7 16.0 13.0 13.9

15.2

14.2

14.3

14.6

15.5

Dearborn

Wayne

261630033

19.2 16.8 18.6 16.1 16.9

18.2

17.2

17.5

19.3

Wyandotte

Wayne

261630036

16.3 13.7 16.4 12.9 13.4

15.5

14.3

14.2

14.7

16.6

Middleton

Butler

390170003

17.2 14.1 19.0 14.1 15.4

16.8

15.7

16.2

16.2

16.5

Fairfield

Butler

390170016

15.8 14.7 17.9 14.0 14.9

16.1

15.5

15.6

15.8

15.9

Cleveland-28th Street

Cuyahoga

390350027

15.4 15.6 17.3 13.0 14.5

16.1

15.3

14.9

15.4

16.5

Cleveland-St. Tikhon

Cuyahoga

390350038

17.6 17.5 19.2 14.9 16.2

18.1

17.2

16.8

17.4

18.4

Cleveland-Broadway

Cuyahoga

390350045

16.4 15.3 19.3 14.0 15.3

17.0

16.2

16.5

16.7

Cleveland-E14 & Orange

Cuyahoga

390350060

17.2 16.4 19.4 15.0 15.9

17.7

16.9

16.8

17.1

17.6

Newburg Hts - Harvard Ave Cuyahoga

390350065

15.6 15.2 18.6 13.1 15.8

16.5

15.6

15.8

16.0

16.2

Columbus - Fairgrounds

Franklin

390490024

16.4 15.0 16.4 13.6 14.6

15.9

15.0

14.9

15.3

16.5

Columbus - Ann Street

Franklin

390490025

15.3 14.6 16.4 13.6 14.7

15.4

14.9

15.1

16.0

Columbus - Maple Canyon Franklin

390490081

14.9 13.6 14.6 12.9 13.1

14.4

13.7

13.5

13.9

16.0

Cincinnati - Seymour

Hamilton

390610014

17.0 15.9 19.8 15.5 16.5

17.6

17.1

17.3

17.3

17.7

Cincinnati - Taft Ave

Hamilton

390610040

15.5 14.6 17.5 13.6 15.1

15.9

15.2

15.4

15.5

15.7

Cincinnati - 8th Ave

Hamilton

390610042

16.7 16.0 19.1 14.9 15.9

17.3

16.7

16.6

16.9

17.3

Sharonville

Hamilton

390610043

15.7 14.9 16.9 14.5 14.8

15.8

15.4

15.6

16.0

Norwood

Hamilton

390617001

16.0 15.3 18.4 14.4 15.1

16.6

16.0

15.9

16.2

16.3

St. Bernard

Hamilton

390618001

17.3 16.4 20.0 15.9 16.1

17.9

17.4

17.3

17.6

17.3

Steubenville

Jefferson

390810016

17.7 15.9 16.4 13.8 16.2

16.7

15.4

15.5

15.8

17.7

Mingo Junction

Jefferson

390811001

17.3 16.2 18.1 14.6 15.6

17.2

16.3

16.1

16.5

17.5

Ironton

Lawrence

390870010

14.3 13.7 17.0 14.4 15.0

15.0

15.4

15.2

15.7

Dayton

Montgomery 391130032

15.9 14.5 17.4 13.6 15.6

15.9

15.2

15.5

15.5

15.9

New Boston

Scioto

391450013

14.7 13.0 16.2 14.3 14.0

14.6

14.5

14.8

14.7

17.1

Canton - Dueber

Stark

391510017

16.8 15.6 17.8 14.6 15.9

16.7

16.0

16.1

16.3

17.3

Canton - Market

Stark

391510020

15.0 14.1 16.6 11.9 14.4

15.2

14.2

14.3

14.6

15.7

Akron - Brittain

Summit

391530017

15.4 15.0 16.4 13.5 14.4

15.6

15.0

14.8

15.1

16.4

Akron - W. Exchange

Summit

391530023

14.2 13.9 15.7 12.8 13.7

14.6

14.1

14.3

15.6

Annual Average Conc.

Design Values

Table 2. PM2.5 Data for Select Sites in 5-State Region

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

When

EPA initially set the 24-hour standard at 65 μg/m

, it also adopted the following

concentration ranges for its Air Quality Index (AQI) scale:

Good

< 15 ug/m

Moderate

15-40 μg/m

Unhealthy for Sensitive Groups (USG) 40-65 μg/m

Unhealthy

65-150 μg/m

Figure 17 shows the frequency of these AQI categories for major metropolitan areas in the

region. Daily average concentrations are often in the moderate range and occasionally in the

USG range. Moderate and USG levels can occur any time of the year.

Figure 17. Percent of days in AQI categories for PM

2.5

(2002-2004)

Data Variability

: PM

2.5

concentrations vary spatially, temporally, and chemically in the region.

This variability is discussed further below.

On an annual basis, PM

2.5

exhibits a distinct and consistent spatial pattern. As seen in Figure

16, across the Midwest, annual concentrations follow a gradient from low values (5-6 μg/m

) in

northern and western areas (Minnesota and northern Wisconsin) to high values (17-18 μg/m

) in

Ohio and along the Ohio River. In addition, concentrations in urban areas are higher than in

upwind rural areas, indicating that local urban sources add a significant increment of 2-3 μg/m

to the regional background of 12 - 14 μg/m

(see Figure 18).

Figure 18. Regional (lighter shading) v. local components (darker shading) of annual average PM

2.5

concentrations

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Because monitoring for PM

2.5

only began in earnest in 1999, after promulgation of the PM

2.5

standard, limited data are available to assess trends. Time series based on federal reference

method (FRM) PM

2.5

-mass data show a downward trend in each state (see Figure 19)

Figure 19. PM

2.5

trends in annual average (top) and daily concentrations (bottom)

Despite the general downward trend since 1999, all states experienced an increase during 2005.

Further analyses are underway to understand this increase (e.g., examination of meteorological and

emissions effects).

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

A statistical analysis of PM

2.5

trends was performed using the nonparametric Theil test for slope

(Hollander and Wolfe, 1973). Trends were generally consistent around the region, for both PM

mass and for the individual components of mass. Figure 20 shows trends for PM

2.5

based on

FRM data at sites with six or more years of data since 1999. The size and direction of each

arrow shows the size and direction of the trend for each site; solid arrows show statistically

significant trends and open arrows show trends that are not significant. Region-wide decreases

are widespread and consistent; all sites had decreasing concentration trends (13 of the 38 were

statistically significant). The average decrease for this set of sites is -0.24 ug/m

/year.

Figure 20. Annual trends in PM

2.5

mass (1999 – 2006)

Seasonal trends show mostly similar patterns (Figure 21). Trends were downward at most sites

and seasons, with overall seasonal averages varying between -0.15 to -0.56 ug/m

/year. The

strongest and most significant decreases took place during the winter quarter (January - March).

No statistically significant increasing trends were observed.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Figure 21. Seasonal trends in PM

2.5

mass (1999 – 2006)

2.5

shows a slight variation from weekday to weekend, as seen in Figure 22. Although most

cities have slightly lower concentrations on the weekend, the difference is usually less than 1

μg/m

. There is a more pronounced weekday/weekend difference at monitoring sites that are

strongly source-influenced. Rural monitors tend to show less of a weekday/weekend pattern

than urban monitors.

Figure 22 Day-of-week variability in PM

2.5

(2002-2004)

In the Midwest, PM

2.5

is made up of mostly ammonium sulfate, ammonium nitrate, and organic

carbon in approximately equal proportions on an annual average basis. Elemental carbon and

crustal matter (also referred to as soil) contribute less than 5% each.

Figure 23. Spatial map of PM

2.5

chemical composition in the Midwest (2002-2003)

The three major components vary spatially (Figure 23), including notable urban and rural

differences (Figure 24). The components also vary seasonally (Figure 25). These patterns

account for much of the annual variability in PM

2.5

mass noted above.

Figure 24. Average regional (lighter shading) v. local (darker shading) of PM

2.5

chemical species

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Figure 25 Seasonal and spatial variability in PM

2.5

components

Ammonium sulfate peaks in the summer and is highest in the southern and eastern parts of the

Midwest, closest to the Ohio River Valley. Sulfate is primarily a regional pollutant;

concentrations are similar in rural and urban areas and highly correlated over large distances. It

is formed when sulfuric acid (an oxidation product of sulfur dioxide) and ammonia react in the

atmosphere, especially in cloud droplets. Coal combustion is the primary source of sulfur

dioxide; ammonia is emitted primarily from animal husbandry operations and fertilizer use.

Ammonium nitrate has almost the opposite spatial and seasonal pattern, with the highest

concentrations occurring in the winter and in the northern parts of the region. Nitrate seems to

have both regional and local sources, because urban concentrations are higher than rural

upwind concentrations. Ammonium nitrate forms when nitric acid reacts with ammonia, a

process that is enhanced when temperatures are low and humidity is high. Nitric acid is a

product of the oxidation of nitric oxide, a pollutant that is emitted by combustion processes.

Organic carbon is more consistent from season to season and city to city, although

concentrations are generally slightly higher in the summer. Like nitrate, organic carbon has

both regional and local components. Particulate organic carbon can be emitted directly from

cars and other fuel combustion sources or formed in a secondary process as volatile organic

gases react and condense. In rural areas, summer organic carbon has significant contributions

from biogenic sources.

Precursor Sensitivity:

Data from the Midwest ammonia monitoring network were analyzed with

thermodynamic equilibrium models to assess the effect of changes in precursor gas

concentrations on PM

2.5

concentrations (Blanchard, 2005b). These analyses indicate that

particle formation responds in varying degrees to reductions in sulfate, nitric acid, and ammonia.

Based on Figure 26, which shows PM

2.5

concentrations as a function of sulfate, nitric acid

(HNO3), and ammonia (NH3), several key findings should be noted:

•

2.5

mass is sensitive to reductions in sulfate at all times of the year and all parts of the

region. Even though sulfate reductions cause more ammonia to be available to form

ammonium nitrate (PM-nitrate increases slightly when sulfate is reduced), this increase

is generally offset by the sulfate reductions, such that PM

2.5

mass decreases.

•

2.5

mass is also sensitive to reductions in nitric acid and ammonia. The greatest PM

2.5

decrease in response to nitric acid reductions occurs during the winter, when nitrate is a

significant fraction of PM

2.5

•

Under conditions with lower sulfate levels (i.e., proxy of future year conditions), PM

2.5

more sensitive to reductions in nitric acid compared to reductions in ammonia.

•

Ammonia becomes more limiting as one moves from west to east across the region.

Examination of weekend/weekday difference in PM-nitrate and NOx concentrations in the

Midwest demonstrate that reductions in local (urban) NOx lead to reductions, albeit non-

proportional reductions, in PM-nitrate (Blanchard, 2004). This result is consistent with analyses

of continuous PM-nitrate from several US cities, including St. Louis (Millstein, et al, 2007).

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Figure 26. Predicted mean PM fine mass concentrations at Bondville, IL (top) and Detroit (Allen Park), MI

(bottom) as functions of changes in sulfate, nitric acid (HNO3), and ammonia (NH3)

Note: starting at the baseline values (represented by the red star), either moving downward (reductions in nitric

acid) or moving leftward (reductions in sulfate or ammonia) results in lower PM

2.5

values

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Meteorology

: PM

2.5

concentrations are not as strongly influenced by meteorology as ozone, but

the two pollutants share some similar meteorological dependencies. In the summer, conditions

that are conducive to ozone (hot temperatures, stagnant air masses, and low wind speeds due

to stationary high pressure systems) also frequently give rise to high PM

2.5

. In the case of PM,

the reason is two-fold: (1) stagnation and limited mixing under these conditions cause PM

2.5

build up, usually over several days, and (2) these conditions generally promote higher

conversion of important precursors (SO

to SO

) and higher emissions of some precursors,

especially biogenic carbon. Wind direction is another strong determinant of PM

2.5

; air

transported from polluted source regions has higher concentrations.

Unlike ozone, PM

2.5

has occasional winter episodes. Conditions are similar to those for summer

episodes, in that stationary high pressure and (seasonally) warm temperatures are usually

factors. Winter episodes are also fueled by high humidity and low mixing heights.

2.5

chemical species show noticeable transport influences. Trajectory analyses have

demonstrated that high PM-sulfate is associated with air masses that traveled through the

sulfate-rich Ohio River Valley (Poirot, et al, 2002 and Kenski, 2004). Likewise, high PM-nitrate

is associated with air masses that traveled through the ammonia-rich Midwest. Figure 27

shows results from an ensemble trajectory analysis of 17 rural eastern IMPROVE sites.

Figure 27. Sulfate and nitrate source regions based on ensemble trajectory analysis

When these results are considered together with analyses of precursor sensitivity (e.g., Figure

26), one possible conclusion is that ammonia control in the Midwest could be effective at

reducing nitrate concentrations. The thermodynamic equilibrium modeling shows that ammonia

reductions would reduce PM concentrations in the Midwest, but that nitric acid reductions are

more effective when the probable reductions in future sulfate levels are considered.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Source Culpability:

Three source apportionment studies were performed using speciated PM

2.5

monitoring data and statistical analysis methods (Hopke, 2005, STI, 2006, and STI, 2008).

Figure 28 summarizes the source contributions from these studies. The studies show that a

large portion of PM

2.5

mass consists of secondary, regional impacts, which cannot be attributed

to individual facilities or sources (e.g., secondary sulfate, secondary nitrate, and secondary

organic aerosols). Nevertheless, wind analyses (e.g., Figure 27) provide information on likely

source regions. Regional- or national-scale control programs may be the most effective way to

deal with these impacts. EPA's CAIR, for example, will provide for substantial reductions in

SO2 emissions over the eastern half of the U.S., which will reduce sulfate (and PM

2.5

)

concentrations and improve visibility levels.

The studies also show that a smaller, yet significant portion of PM

2.5

mass is due to emissions

from nearby (local) sources. Local (urban) excesses occur in many urban areas for organic and

elemental carbon, crustal matter, and, in some cases, sulfate. The statistical analysis methods

help to identify local sources and quantify their impact. This information is valuable to states

wishing to develop control programs to address local impacts. A combination of

national/regional-scale and local-scale emission reductions may be necessary to provide for

attainment.

The carbon sources are not easily identified in complex urban environments. LADCO’s Urban

Organics Study (STI, 2006) identified four major sources of organic carbon: mobile sources,

burning, industrial sources, and secondary organic aerosols. Additional sampling and analysis

is underway in Cleveland and Detroit to provide further information on sources of organic

carbon.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Coal

Combustion

Secondary

Nitrate

Mobile

Industrial

Soil

Burning

ug/m3

Northbrook (Jan 03-Mar 05)

Cincinnati (Dec 03-Mar 05)

Indianapolis (Jan 02-Mar 05)

Allen Park (2002-2004)

Dearborn (Jun 02-May 05)

Bondville (2002-2004)

St. Louis (2004)

Granite City (2002-2004)

Tikhon (Jan 03-Jun05)

Orange (Apr 01-Mar 05)

Lorain (Feb 02-Dec04)

sulfate

nitrate

diesel

soil

steel

ind. zinc

ind. copper

mixed ind.

Allen Park

Dearborn

Lawndale

Springfield

Tikhon

Craig

Taft

Figure 28. Major Source Contributions in the Midwest based on Hopke, 2005 (upper left), STI, 2006 (upper right), and STI, 2008 (lower left)

(Note: the labeling of similar source types varies between studies – e.g., organic carbon/mobile sources are named gasoline and diesel by

Hopke, mobile by STI 2006, and OM and diesel by STI 2008)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

2.3 Haze

Section 169A of the Clean Air Act sets as a national goal “the prevention of any future, and the

remedying of any existing, impairment of visibility in mandatory Class I Federal areas which

impairment results from manmade air pollution”. To implement this provision, in 1999, EPA

adopted regulations to address regional haze visibility impairment (USEPA, 1999). EPA’s rule

requires states to “make reasonable progress toward meeting the national goal”. Specifically,

states must establish reasonable progress goals, which provide for improved visibility on the

most impaired (20% worst) days sufficient to achieve natural conditions by the year 2064, and

for no degradation on the least impaired (20% best) days.

The primary cause of impaired visibility in the Class I areas is pollution by fine particles that

scatter light. The degree of impairment, which is expressed in terms of visual range, light

extinction (1/Mm), or deciviews (dv), depends not just on the total PM

2.5

mass concentration, but

also on the chemical composition of the particles and meteorological conditions.

Current Conditions

: A map of the average light extinction values for the most impaired (20%

worst) visibility days for the 5-year baseline period (2000-2004) is shown in Figure 29.

Figure 29. Baseline Visibility Levels for 20% Worst Days (2000 – 2004), units: Mm

-1

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Initially, the baseline (2000 – 2004) visibility condition values were derived using the average for

the 20% worst and 20% best days for each year, as reported on the VIEWS website:

http://vista.cira.colostate.edu/views/Web/IMPROVE/SummaryData.aspx . These values were

calculated using the original IMPROVE equation for reconstructed light extinction.

Three changes were made to the baseline calculations to produce a new set of values. First,

the reconstructed light extinction equation was revised by the IMPROVE Steering Committee in

2005. The new IMPROVE equation was used to calculate updated baseline values.

Second, due to sampler problems, the 2002-2004 data for Boundary Waters were invalid for

certain chemical species. (Note, sulfate and nitrate data were valid.) A “substituted” data set

was developed by using values from Voyageurs for the invalid species.

Third, LADCO identified a number of days during 2000-2004 where data capture at the Class I

monitors was incomplete (Kenski, 2007b). The missing data cause these days to be excluded

from the baseline calculations. However, the light extinction due to the remaining measured

species is significant (i.e., above the 80

percentile). It makes sense to include these days in

the baseline calculations, because they are largely dominated by anthropogenic sources. (Only

one of these days is driven by high organic carbon, which might indicate non-anthropogenic

aerosol from wildfires.) As seen in Table 3, inclusion of these days in the baseline calculation

results in a small, but measurable, effect on the baseline values (i.e., values increase from 0.2

to 0.8 dv).

Table 3. Average of 20% worst days, with and without missing data days

Average Worst Day

DV, per RHR

Average Worst Day DV,

with Missing Data Days

Difference

BOWA

19.59

19.86

0.27

ISLE

20.74

21.59

0.85

SENE

24.16

24.38

0.22

VOYA

19.27

19.48

0.21

A summary of the initial and updated baseline values for the Class I areas in northern Michigan

and northern Minnesota are presented in Table 4. The updated baseline values reflect the most

current, complete understanding of visibility impairing effects and, as such, will be used for SIP

planning purposes.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Table 4. Summary of visibility metrics (deciviews) for northern Class I areas

Old IMPROVE Equation (Cite: VIEWS, November 2005)

20% Worst Days

2000

2001

2002

2003

2004

Baseline

Value

2018

URI Value

Natural

Conditions

Voyageurs

18.50

18.00

19.00

19.20

17.60

18.46

16.74

11.09

BWCA

19.85

19.99

19.68

19.73

17.65

19.38

17.47

11.21

Isle Royale

20.00

22.00

20.80

19.50

19.10

20.28

18.17

11.22

Seney

22.60

24.90

24.00

23.80

22.60

23.58

20.73

11.37

20% Best Days

2000

2001

2002

2003

2004

Baseline

Value

Natural

Conditions

Voyageurs

6.30

6.20

6.70

7.00

5.40

6.32

3.41

BWCA

5.90

6.52

6.93

6.67

5.61

6.33

3.53

Isle Royale

5.70

6.40

6.30

5.30

6.02

3.54

Seney

5.80

6.10

7.30

7.50

5.80

6.50

3.69

New IMPROVE Equation (Cite: VIEWS, March 2006)

20% Worst Days

2000

2001

2002

2003

2004

Baseline

Value

2018

URI Value

Natural

Conditions

Voyageurs

19.55

18.57

20.14

20.25

18.87

19.48

17.74

12.05

BWCA

20.20

20.04

20.76

20.13

18.18

19.86

17.94

11.61

Isle Royale

20.53

23.07

21.97

22.35

20.02

21.59

19.43

12.36

Seney

22.94

25.91

25.38

24.48

23.15

24.37

21.64

12.65

20% Best Days

2000

2001

2002

2003

2004

Baseline

Value

Natural

Conditions

Voyageurs

7.01

7.12

7.53

7.68

6.37

7.14

4.26

BWCA

6.00

6.92

7.00

6.45

5.77

6.43

3.42

Isle Royale

6.49

7.16

7.07

6.99

6.12

6.77

3.72

Seney

6.50

6.78

7.82

8.01

6.58

7.14

3.73

Notes: (1) BWCA values for 2002 - 2004 reflect "substituted" data.

(2) New IMPROVE equation values include Kenski, 2007 adjustment for missing days

URI = uniform rate of improvement

As noted above, the goal of the visibility program is to achieve natural conditions. Initially, the

natural conditions values for each Class I area were taken directly from EPA guidance (EPA,

2003). These values were calculated using the original IMPROVE equation. This equation was

revised by the IMPROVE Steering Committee in 2005, and the new IMPROVE equation was

used to calculate updated natural conditions values. The updated values are reported on the

VIEWS website.

A summary of the initial and updated natural conditions values are presented in Table 4. The

updated natural conditions values (based on the new IMPROVE equation) will be used for SIP

planning purposes.

Data Variability:

For the four northern Class I areas, the most important PM

2.5

chemical species

are ammonium sulfate, ammonium nitrate, and organic carbon. The contribution of these

species on the 20% best and 20% worst visibility days (based on 2000 – 2004 data) is provided

in Figure 30. For the 20% worst visibility days, the contributions are: sulfate = 35-55%, nitrate =

25-30%, and organic carbon = 12-22%. Although the chemical composition is similar, sulfate

increases in importance from west to east and concentrations are highest at Seney (the

easternmost site). It should also be noted that sulfate and nitrate contribute more to light

extinction than to PM

2.5

mass because of their hygroscopic properties.

20% Best Days

20% Worst Days

100

120

VOYA BOWA ISRO Seney

sea salt

soil

elem. carbon

coarse mass

organic carbon

nitrate

sulfate

Figure 30. Chemical composition of light extinction for 20% best visibility days (left) and 20%

worst visibility days (right) in terms of Mm

-1

Analysis of PM

2.5

mass and chemical species for rural IMPROVE (and IMPROVE-protocol) sites

in the eastern U.S. showed a high degree of correlation between PM

2.5

-mass, sulfate, and

nitrate levels (see Figure 31). The Class I sites in northern Michigan and northern Minnesota, in

particular, are highly correlated for PM

2.5

mass, sulfates, and organic carbon mass (AER, 2004).

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Figure 31. Correlations among IMPROVE (and IMPROVE-protocol) monitoring sites in Eastern U.S.

Long-term trends at Boundary Waters (the only regional site with a sufficient data record) show

significant decreases in total PM

2.5

(-0.005 ug/year) and SO4 (-0.04 ug/year) and an increase in

NO3 (+0.01 ug/year). These PM

2.5

and SO4 trends are generally consistent with long-term

trends at other IMPROVE sites in the eastern U.S., which have shown widespread decreases in

SO4 and PM

2.5

(DeBell, et al, 2006). Detecting changes in nitrate has been hampered by

uncertainties in the IMPROVE data for particular years and, thus, this estimate should be

considered tentative.

Haze in the Midwest Class I areas has no strong seasonal pattern. Poor visibility days occur

throughout the year, as indicated in Figure 32. (Note, in contrast, other parts of the country,

such as Shenandoah National Park in Virginia, show a strong tendency for the worst air quality

days to occur in the summer months.) This figure and Figure 33 (which presents the monthly

average light extinction values based on all sampling days) also show that sulfate and organic

carbon concentrations are higher in the summer, and nitrate concentrations are higher in the

winter, suggesting the importance of different sources and meteorological conditions at different

times of the year.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Figure 32. Daily light extinction values for 20% worst days at Boundary Waters (2000 – 2004)

Figure 33. Monthly average light extinction values for northern Class I areas

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Precursor Sensitivity:

Results from two analyses using thermodynamic equilibrium models

provide information on the effect of changes in precursor concentrations on PM

2.5

concentrations (and, in turn, visibility levels) in the northern Class I areas. First, a preliminary

analysis using data collected at Seney indicated that PM

2.5

there is most sensitive to reductions

in sulfate, but is also sensitive to reductions in nitric acid (Blanchard, 2004).

Second, an analysis was performed using data from the Midwest ammonia monitoring network

for a site in Minnesota -- Great River Bluffs, which is the closest ammonia monitoring site to the

northern Class I areas (Blanchard, 2005b). Figure 34 shows PM

2.5

concentrations as a function

of sulfate, nitric acid (HNO3), and ammonia (NH3). Reductions in sulfate (i.e., movement to the

left of baseline value [represented by the red star]), as well as reductions in nitric acid (i.e.,

movement downward) and NH3 (i.e., movement to the left), result in lower PM

2.5

concentrations.

Thus, reductions in sulfate, nitric acid, and ammonia will lower PM

2.5

concentrations and

improve visibility in the northern Class I areas.

Figure 34. Predicted PM

2.5

mass concentrations at Great River Bluffs, MN as functions of changes

in sulfate, nitric acid, and ammonia

Meteorology and Transport:

The role of meteorology in haze is complex. Wind speed and wind

direction govern the movement of air masses from polluted areas to the cleaner wilderness

areas. As noted above, increasing humidity increases the efficiency with which sulfate and

nitrate aerosols scatter light. Temperature and humidity together govern whether ammonium

nitrate can form from its precursor gases, nitric acid and ammonia. Temperature and sunlight

also play an indirect role in emissions of biogenic organic species that condense to form

particulate organic matter; emissions increase in the summer daylight hours.

Trajectory analyses were performed to understand transport patterns for the 20% worst and

20% best visibility days. The composite results for the four northern Class I areas are provided

in Figure 35. The orange areas are where the air is most likely to come from, and the green

areas are where the air is least likely to come from. As can be seen, bad air days are generally

associated with transport from regions located to the south, and good air days with transport

from Canada.

Figure 35. Composite back trajectories for light extinction- 20% best visibility days (left) and

20% worst visibility days (right) (2000 – 2005)

Source Culpability:

Air quality data analyses (including the trajectory analyses above) and

dispersion modeling were used to provide information on source region and source sector

contributions to regional haze in the northern Class I areas (see MRPO, 2008). Based on this

information, the most important contributing states are Michigan, Minnesota, and Wisconsin, as

well as Missouri, North Dakota, Iowa, Indiana and Illinois (see, for example, Figure 35 above).

The most important contributing pollutants and source sectors are SO2 emissions from

electrical generating units (EGUs) and certain non-EGUs, which lead to sulfate formation, and

NOx emissions from a variety of source types (e.g., motor vehicles), which lead to nitrate

formation. Ammonia emissions from livestock waste and fertilizer applications are also

important, especially for nitrate formation.

A source apportionment study was performed using monitoring data from Boundary Waters and

statistical analysis methods (DRI, 2005). The study shows that a large portion of PM

2.5

mass

consists of secondary, regional impacts, which cannot be attributed to individual facilities or

sources (e.g., secondary sulfate, secondary nitrate, and secondary organic aerosols). Industrial

sources contribute about 3-4% and mobile sources about 4-7% to PM

2.5

mass.

A special study was performed in Seney to identify sources of organic carbon (Sheesley, et al,

2004). As seen in Figure 36, the highest PM

2.5

concentrations occurred during the summer,

with organic carbon being the dominant species. The higher summer organic carbon

concentrations were attributed mostly to secondary organic aerosols of biogenic origin because

of the lack of primary emission markers, and concentrations of know biogenic-related species

(e.g., pinonic acid – see Figure 36) were also high during the summer.

Figure 36. Monthly concentrations of PM

2.5

species (top), and secondary and biogenic-related

organic carbon species in Seney (bottom)

Although the Seney study showed that biomass burning was a relatively small contributor to

organic carbon on an annual average basis, episodic impacts are apparent (see, for example,

high organic carbon days in Figure 32). To assess further whether burning is a significant

contributor to visibility impairment in the northern Class I areas, the PM

2.5

chemical speciation

data were examined for days with high organic carbon and elemental carbon concentrations,

which are indicative of biomass burning impacts. Only a handful of such days were identified:

Table 5. Days with high OC and EC concentrations in northern Class I areas

Site

2000

2001

2002

2003

2004

Voyageurs

---

Jun 1

Aug 25

Jul 17

Jun 28

Jul 19

Boundary Waters

---

Jun 28

Aug 25

Jul 17

Jul 19

Isle Royale

---

Jun 1

Aug 25

---

Jun 28

Seney

---

Jun 28

---

Back trajectories on these days point mostly to wildfires in Canada. Elimination of these high

organic carbon concentration days has a small effect in lowering the baseline visibility levels in

the northern Class I areas (i.e., Minnesota Class I areas change by about 0.3 deciviews and

Michigan Class I areas change by less than 0.2 deciviews). This suggests that fire activity,

although significant on a few days, is on average a relatively small contributor to visibility

impairment in the northern Class I areas.

In summary, these analyses show that organic carbon in the northern Class I is largely

uncontrollable.

Section 3.0 Air Quality Modeling

Air quality models are relied on by federal and state regulatory agencies to support their

planning efforts. Used properly, models can assist policy makers in deciding which control

programs are most effective in improving air quality, and meeting specific goals and objectives.

For example, models can be used to conduct “what if” analyses, which provide information for

policy makers on the effectiveness of candidate control programs.

The modeling analyses were conducted in accordance with EPA’s modeling guidelines (EPA,

2007a). Further details of the modeling are provided in two protocol documents: LADCO, 2007a

and LADCO, 2007b.

This section reviews the development and evaluation of the modeling system used for the multi-

pollutant analyses. Application of the modeling system (i.e., attainment demonstration for ozone

and PM

2.5

, and reasonable progress assessment for haze) is covered in the following sections.

3.1 Selection of Base Year

Two base years were used in the modeling analyses: 2002 and 2005. EPA’s modeling

guidance recommends using 2002 as the baseline inventory year, but also allows for use of an

alternative baseline inventory year, especially a more recent year. Initially, LADCO conducted

modeling with a 2002 base year (i.e., Base K/Round 4 modeling, which was completed in 2006).

A decision was subsequently made to conduct modeling with a 2005 base year (i.e., Base

M/Round 5, which was completed in 2007). As discussed in the previous section, 2002 and

2005 both had above normal ozone conducive conditions, although 2002 was more severe

compared to 2005. Examination of multiple base years provides for a more complete technical

assessment. Both sets of model runs are discussed in this document.

3.2 Future Years of Interest

To address the multiple attainment requirements for ozone and PM

2.5

, and reasonable progress

goals for regional haze, several future years are of interest:

2008 Planning year for ozone basic nonattainment areas (attainment date 2009)

2009 Planning year for ozone moderate nonattainment areas and PM

2.5

nonattainment

areas (attainment date 2010)

2012 Planning year for ozone moderate nonattainment areas and PM

2.5

nonattainment

areas, with 3-year extension (attainment date 2013)

2018 First milestone year for regional haze planning

According to USEPA’s ozone implementation rule (USEPA, 2005), emission reductions needed for

attainment must be implemented by the beginning of the ozone season immediately preceding the area’s

attainment date. The PM2.5 implementation rule contains similar provisions – i.e., emission reductions

should be in place by the beginning of the year preceding the attainment date (USEPA, 2007c). The logic

for requiring emissions reductions by the year (or season) immediately preceding the attainment year

follows from language in the Clean Air Act, and the ability for an area to receive up to two 1-year

extensions. Therefore, emissions in the year preceding the attainment year should be at a level that is

consistent with attainment. It also follows that the year preceding the attainment year should be modeled

for attainment planning purposes.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Detailed emissions inventories were developed for 2009 and 2018. To support modeling for

other future years, less rigorous emissions processing was conducted (e.g., 2012 emissions

were estimated for several source sectors by interpolating between 2009 and 2018 emissions).

3.3 Modeling System

The air quality analyses were conducted with the CAMx model, with emissions and meteorology

generated using EMS (and CONCEPT) and MM5, respectively. The selection of CAMx as the

primary model is based on several factors: performance, operator considerations (e.g., ease of

application and resource requirements), technical support and documentation, model

extensions (e.g., 2-way nested grids, process analysis, source apportionment, and plume-in-

grid), and model science. CAMx model set-up for Base M and Base K is summarized below:

Base M (2005)

Base K (2002)

• CAMx v4.50

* CAMx 4.30

• CB05 gas phase chemistry

* CB-IV with updated gas-phase chemistry

• SOA chemistry updates

* No SOA chemistry updates

• AERMOD dry deposition scheme

* Wesley-based dry deposition

• ISORROPIA inorganic chemistry

• SOAP organic chemistry

• RADM aqueous phase chemistry

• PPM horizontal transport

3.4 Domain/Grid Resolution

The National RPO grid projection was used for this modeling. A subset of the RPO domain was

used for the LADCO modeling. For PM

2.5

and haze, the large eastern U.S. grid at 36 km (see

box on right side of Figure 36) was used. A PM

2.5

sensitivity run was also performed for this

domain at 12 km. For ozone, the smaller grid at 12 km (see shaded portion of the box on the

right side of Figure 37) was used for most model runs. An ozone sensitivity run was also

performed with a 4km sub-grid over the Lake Michigan area and Detroit/Cleveland.

The vertical resolution in the air quality model consists of 16 layers extending up to 15 km, with

higher resolution in the boundary layer.

Figure 37. Modeling grids – RPO domain (left) and LADCO modeling domain (right)

12 km

36 km

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

3.5 Model Inputs: Meteorology

Meteorological inputs were derived using the Fifth-Generation NCAR/Penn State Meteorological

Model (MM5) – version 3.6.3 for the years 2001–2003, and version 3.7 for the year 2005. The

MM5 modeling domains are consistent with the National RPO grid projections (see Figure 38).

Figure 38. MM5 modeling domain for 2001-2003 (left) and 2005 (right)

The annual 2002 36 km MM5 simulation was completed by Iowa DNR. The 36/12 km 2-way

nested simulation for the summers of 2001, 2002, and 2003 were conducted jointly by Illinois

EPA and LADCO. The 36 km non-summer portion of the annual 2003 simulation was conducted

by Wisconsin DNR. The annual 2005 36/12 km (and summer season 4 km) MM5 modeling was

completed by Alpine Geophysics. Wisconsin DNR also completed 36/12 km MM5 runs for the

summer season of 2005.

Model performance was assessed quantitatively with the METSTAT tool from Environ. The

metrics used to quantify model performance include mean observation, mean prediction, bias,

gross error, root mean square error, and index of agreement. Model performance metrics were

calculated for several sub-regions of the modeling domain (Figure 39) and represent hourly

spatial averages of multiple monitor locations. Additional analysis of rainfall is done on a

monthly basis.

Figure 39. Sub-domains used for model performance for 2001-2003 (left) and 2005 (right)

A summary of the performance evaluation results for the meteorological modeling is provided

below. Further details are provided in two summary reports (LADCO, 2005 and LADCO, 2007c).

Temperature

: The biggest issue with the performance in the upper Midwest is the existence of a

cool diurnal temperature bias in the winter and warm temperature bias over night during the

summer (see Figure 40). These features are common to other annual MM5 simulations for the

central United States and do not appear to adversely affect model performance.

Figure 40. Daily temperature bias for 2002 (left) and 2005 (right) with hotter colors

(yellow/orange/red) representing overestimates and cooler colors (blues) representing

underestimates

Note: months are represented from left to right (January to December) and days are represented

from top to bottom (1 to 30(31) – i.e., upper left hand corner is January 1 and lower right hand

corner is December 31

Wind Fields

: The wind fields are generally good. Wind speed bias is less than 0.5 m/sec and

wind speed error is consistently between 1.0 and 1.5 m/sec. Wind direction error is generally

within 15-30 degrees.

Mixing Ratio:

The mixing ratio (a measure of humidity) is over-predicted in the late spring and

summer months, and mixing ratio error is highest during this period. There is little bias and

error during the cooler months when there is less moisture in the air.

Rainfall:

The modeled and observed rainfall totals show good agreement spatially and in

terms of magnitude in the winter, fall, and early spring months. There are, however, large over-

predictions of rainfall in the late spring and summer months (see Figure 41). These over-

predictions are seen spatially and in magnitude over the entire domain, particularly in the

Southeast United States, and are likely due to excessive convective rainfall being predicted in

MM5. This over-prediction of rainfall in MM5 does not necessarily translate into over-prediction

of wet deposition in the photochemical model. CAMx does not explicitly use the convective and

non-convective rainfall output by MM5, but estimates wet scavenging by hydrometeors using

cloud, ice, snow, and rain water mixing ratios output by MM5. Nevertheless, this could have an

effect on model performance for PM

2.5

, as discussed in Section 3.7, and may warrant further

attention.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Figure 41. Comparison of observed (left column) and modeled (right column) monthly rainfall for

July 2002 (top) and July 2005 (bottom)

3.6 Model Inputs: Emissions

Emission inventories were prepared for two base years: 2002 (Base K) and 2005 (Base M), and

several future years: 2008, 2009, 2012, and 2018. Further details of the emission inventories

are provided in two summary reports (LADCO, 2006a and LADCO, 2008a) and the following

pages of the LADCO web site:

http://www.ladco.org/tech/emis/basek/BaseK_Reports.htm

http://www.ladco.org/tech/emis/r5/round5_reports.htm

For on-road, nonroad, ammonia, and biogenic sources, emissions were estimated by models.

For the other sectors (point sources, area sources, and MAR [commercial marine, aircraft, and

railroads]), emissions were prepared using data supplied by the LADCO States and other

RPOs.

Base Year Emissions

: State and source sector emission summaries for 2002 (Base K) and

2005 (Base M) are compared in Figure 42. Additional detail is provided in Tables 6a (all sectors

– tons per day) and 6b (EGUs – tons per year).

VOC

NOx

SO2

2000

4000

6000

8000

10000

12000

2002

(K)

2005

(M)

2002

(K)

2005

(M)

2002

(K)

2005

(M)

2000

4000

6000

8000

10000

12000

2002

(K)

2005

(M)

2002

(K)

2005

(M)

2002

(K)

2005

(M)

Area

Nonroad+MAR

Onroad

Non-EGU

EGU

Figure 42. Base K and Base M emissions for 5-state LADCO region by state (top) and source

sector (bottom), units: tons per summer weekday

A summary of the base year emissions by sector for the LADCO States is provided below.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

VOC Base M BaseK Base M BaseK BaseK Base M

NOx Base M BaseK Base M BaseK BaseK Base M

SOX Base M BaseK Base M BaseK BaseK Base M

PM2.5 Base M BaseK Base M BaseK BaseK Base M

July

2002

2005

2009

2012

2018

2002

2005

2009

2012

2018

2002

2005

2009

2012

2018

2002

2005

2009

2012

2018

Nonroad

224

321

164

257

149

130

213

324

333

263

275

224

154

155

0.6

0.4

125

195

160

128

178

191

142

158

141

0.3

0.2

348

414

307

350

276

222

271

205

239

159

197

133

112

0.5

0.3

222

356

161

294

145

126

238

253

304

195

246

162

109

135

0.5

0.3

0.4

214

238

194

203

175

140

157

145

157

114

129

0.3

0.2

5-State Total

1133

1524

920

1264

840

713

1007

1105

1224

873

1005

757

566

568

103

118

4.9

1.5

110

U.S. Total

8463

9815

5442

8448

5244

6581

6041

9060

6057

8120

5832

5100

505

654

117

153

104

573

750

475

MAR

277

246

201

228

195

186

165

123

0.2

114

112

111

110

0.6

0.7

0.8

177

134

128

126

122

0.4

0.3

12.7

9.5

8.7

5-State Total

770

618

589

577

578

559

430

13.9

10.7

U.S. Total

307

317

321

157

329

346

334

4968

4515

4002

1813

3964

3919

3812

620

512

509

122

509

503

290

147

165

OtherArea

679

675

688

594

700

738

582

354

391

365

358

373

398

384

158

150

151

153

518

652

516

562

520

541

549

111

114

120

546

604

550

506

558

593

487

108

458

315

467

290

474

506

293

5-State Total

2555

2637

2586

2310

2625

2776

2295

255

283

278

301

284

293

304

271

273

105

276

279

103

183

227

237

U.S. Total

17876 21093 18638 18683

20512 24300

3856

4899

4100

4220

4418

5357

2075

2947

2062

2559

2189

2709

2735

2621

2570

On-Road

446

341

314

268

260

197

151

890

748

578

528

474

300

201

405

282

237

235

193

150

138

703

541

425

402

313

187

173

522

351

335

269

303

217

163

926

722

680

501

619

385

204

574

680

365

424

340

238

242

1035

934

609

693

512

270

274

238

175

144

119

117

481

457

303

322

226

118

138

5-State Total

2185

1829

1395

1315

1213

890

762

4035

3402

2595

2446

2144

1260

990

U.S. Total

14263

7825

23499

13170

EGU

712

305

227

275

244

231

224

1310

1158

944

958

789

810

869

830

393

406

370

424

283

255

2499

2614

1267

1033

1263

1048

1036

448

393

218

242

219

247

243

1103

1251

1022

667

1031

1058

725

1139

408

330

280

322

271

285

3131

3405

1463

1326

994

701

983

293

213

146

165

139

147

177

602

545

512

460

492

500

435

5-State Total

3422

1712

1327

1332

1348

1179

1184

8645

8973

5208

4444

4569

4117

4048

248

285

U.S. Total

214

140

195

124

197

215

138

14371 10316

7746

7274

7721

7007

6095

31839 34545 20163 16903 17629 14727 14133

685

1131

1571

Non-EGU

313

221

286

218

305

350

258

356

330

334

218

338

343

235

373

423

251

335

257

249

346

150

130

160

137

170

199

167

238

179

212

175

216

225

178

292

218

270

216

274

290

180

123

116

115

119

122

139

140

216

240

208

242

214

229

271

162

158

166

148

171

185

163

104

177

175

157

166

160

167

178

240

289

231

288

210

216

293

104

120

106

163

156

154

152

155

156

0.1

5-State Total

751

635

733

648

780

898

775

1085

1021

1002

894

1020

1058

943

1230

1244

1072

1139

1067

1096

1067

102

121

U.S. Total

4087

3877

4409

4700

5378

6446

6730

6129

6435

6952

5759

5630

6093

6340

6970

1444

1777

1681

1576

1470

1353

1432

1434

1217

2621

2010

1671

1572

1545

1287

1029

1725

1656

1212

1337

1059

1072

1251

119

155

189

1045

1009

867

901

843

853

826

2134

1453

1339

1250

1248

989

819

2966

2902

1690

1294

1691

1492

1256

133

131

1530

1546

1291

1311

1239

1139

1134

1958

1730

1429

1314

1349

1118

946

1356

1495

1260

865

1271

1312

927

183

190

1432

1735

1165

1323

1137

1062

1082

2831

2048

1478

1619

1342

1001

1074

3416

3761

1732

1650

1240

953

1304

121

195

166

1005

821

909

705

878

862

630

1128

1019

747

800

647

520

551

800

750

687

635

667

675

540

5-State Total

6693

6687

5702

5593

5529

5350

4889

10672

8260

6664

6555

6131

4915

4419

10263 10564

6581

5781

5928

5504

5280

539

727

723

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Heat Input (MMBTU/year)

Scenario

SO2 (tons/year)

SO2 (lb/MMBTU)

NOx (tons/year)

NOx (lb/MMBTU)

980,197,198

2001 - 2003 (average)

362,417

0.74

173,296

0.35

IPM 2.1.9

241,000

73,000

1,310,188,544

IPM3.0 (base)

277,337

0.423

70,378

0.107

IPM3.0 - will do

140,296

0.214

62,990

0.096

IPM3.0 - may do

140,296

0.214

62,990

0.096

1,266,957,401

2001 - 2003 (average)

793,067

1.25

285,848

0.45

IPM 2.1.9

377,000

95,000

1,509,616,931

IPM3.0 (base)

361,835

0.479

90,913

0.120

IPM3.0 - will do

417,000

0.552

94,000

0.125

IPM3.0 - may do

417,000

0.552

94,000

0.125

756,148,700

2001 - 2003 (average)

346,959

0.92

132,995

0.35

IPM 2.1.9

399,000

100,000

1,009,140,047

IPM3.0 (base)

244,151

0.484

79,962

0.158

IPM3.0 - will do

244,151

0.484

79,962

0.158

IPM3.0 - may do

244,151

0.484

79,962

0.158

1,306,296,589

2001 - 2003 (average)

1,144,484

1.75

353,255

0.54

IPM 2.1.9

216,000

84,000

1,628,081,545

IPM3.0 (base)

316,883

0.389

96,103

0.118

IPM3.0 - will do

348,000

101,000

IPM3.0 - may do

348,000

101,000

495,475,007

2001 - 2003 (average)

191,137

0.77

90,703

0.36

IPM 2.1.9

155,000

46,000

675,863,447

IPM3.0 (base)

127,930

0.379

56,526

0.167

IPM3.0 - will do

150,340

0.445

55,019

0.163

IPM3.0 - may do

62,439

0.185

46,154

0.137

390,791,671

2001 - 2003 (average)

131,080

0.67

77,935

0.40

IPM 2.1.9

147,000

51,000

534,824,314

IPM3.0 (base)

115,938

0.434

59,994

0.224

IPM3.0 - will do

115,938

0.434

59,994

0.224

IPM3.0 - may do

100,762

0.377

58,748

0.220

401,344,495

2001 - 2003 (average)

101,605

0.50

85,955

0.42

IPM 2.1.9

86,000

42,000

447,645,758

IPM3.0 (base)

61,739

0.276

41,550

0.186

IPM3.0 - will do

54,315

0.243

49,488

0.221

IPM3.0 - may do

51,290

0.229

39,085

0.175

759,902,542

2001 - 2003 (average)

241,375

0.63

143,116

0.37

IPM 2.1.9

281,000

78,000

893,454,905

IPM3.0 (base)

243,684

0.545

72,950

0.163

IPM3.0 - will do

237,600

0.532

72,950

0.163

IPM3.0 - may do

237,600

0.532

72,950

0.163

339,952,821

2001 - 2003 (average)

145,096

0.85

76,788

0.45

IPM 2.1.9

109,000

72,000

342,685,501

IPM3.0 (base)

41,149

0.240

44,164

0.258

IPM3.0 - will do

56,175

0.328

58,850

0.343

IPM3.0 - may do

56,175

0.328

58,850

0.343

39,768,357

2001 - 2003 (average)

12,545

0.63

15,852

0.80

IPM 2.1.9

12,000

15,000

44,856,223

IPM3.0 (base)

4,464

0.199

2,548

0.114

IPM3.0 - will do

4,464

0.199

2,548

0.114

IPM3.0 - may do

4,464

0.199

2,548

0.114

Table 6b. EGU Emissions for Midwest States (2018)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

On-road Sources: For 2002, EMS was run by LADCO using VMT and MOBILE6 inputs supplied

by the LADCO States. EMS was run to generate 36 days (weekday, Saturday, Sunday for each

month) at 36 km, and 9 days (weekday, Saturday, Sunday for June – August) at 12 km. For

2005, CONCEPT was run by a contractor (Environ) using transportation data (e.g., VMT and

vehicle speeds) supplied by the state and local planning agencies in the LADCO States and

Minnesota for 24 networks. These data were first processed with T3 (Travel Demand Modeling

[TDM] Transformation Tool) to provide input files for CONCEPT to calculate link-specific, hourly

emission estimates (Environ, 2008). CONCEPT was run with meteorological data for a July and

January weekday, Saturday, and Sunday (July 15 – 17 and January 16 – 18). A spatial plot of

emissions is provided in Figure 43.

VOC Emissions

NOx Emissions

Figure 43. Motor vehicle emissions for VOC (left) and NOx (right) for a July weekday (2005)

Off-road Sources: For 2002 and 2005, NMIM and NMIM2005, respectively, were run by

Wisconsin DNR. Additional off-road sectors (i.e., commercial marine, aircraft, and railroads

[MAR]) were handled separately. Local data for agricultural equipment, construction equipment,

commercial marine, recreational marine, and railroads were prepared by contractors (Environ,

2004, and E.H. Pechan, 2004). For Base M, updated local data for railroads and commercial

marine were prepared by a contractor (Environ, 2007b, 2007c). Table 7 compares the Base M

2005 and Base K 2002 emissions. Compared to 2002, the new 2005 emissions reflect

substantially lower commercial marine emissions and lower locomotive NOx emissions.

Table 7. Locomotive and commercial marine emissions for the five LADCO States (2002 v. 2005)

Railroads (TPY)

Commercial Marine (TPY)

2002

2005

2002

2005

VOC

7,890

7,625

1,562

828

20,121

20,017

8,823

6,727

NOx

182,226

145,132

64,441

42,336

5,049

4,845

3,113

1,413

SO2

12,274

12,173

25,929

8,637

NH3

----

Area Sources: For 2002 and 2005, EMS was run by LADCO using data supplied by the LADCO

States to produce weekday, Saturday, and Sunday emissions for each month. For 2005,

special attention was given to two source categories: industrial adhesive and sealant solvents

(which were dropped from the inventory to avoid double-counting) and outdoor wood boilers

(which were added to the inventory).

Point Sources: For 2002 and 2005, EMS was run by LADCO using data supplied by the LADCO

States to produce weekday, Saturday, and Sunday emissions for each month. For EGUs, the

annual and summer season emissions were temporalized for modeling purposes using profiles

prepared by Scott Edick (Michigan DEQ) based on CEM data.

Biogenics: For Base M, a contractor (Alpine) provided an updated version of the

CONCEPT/MEGAN biogenics model. Compared to the previous (EMS/BIOME) emissions,

there is more regional isoprene using MEGAN compared to the BIOME estimates used for Base

K (see Figure 44). Also, with the secondary organic aerosol updates to the CAMx air quality

model, Base M includes emissions for monoterpenes and sesquiterpenes, which are pre-

cursors of secondary PM

2.5

organic carbon mass.

Figure 44. Isoprene emissions for Base M (left) v. Base K (right)

Ammonia: For Base M, the CMU-based 2002 (Base K) ammonia emissions were projected to

2005 using growth factors from the Round 4 emissions modeling. These emissions were then

adjusted by applying temporal factors by month based on the process-based ammonia

emissions model (Zhang, et al, 2005, and Mansell, et al, 2005). A plot of average daily

emissions by state and month is provided in Figure 45. A spatial plot of emissions is provided in

Figure 46, which shows high emissions densities in the central U.S.

300

600

900

1200

1500

1800

Jan

Fed March Apr

May

Jun

Jul

Aug Sept

Oct

Nov Dec

Illinois

Indiana

Iowa

Michigan

Minnesota

Ohio

Wisconsin

Figure 45. Average daily ammonia emissions for Midwest States by month (2005) - (units: average

daily emissions – tons per day)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Figure 46. Ammonia emissions for a July weekday (2005) – 12 km modeling domain

Canadian Emissions: For Base M, Scott Edick (Michigan DEQ) processed the 2005 Canadian

National Pollutant Release Inventory, Version 1.0 (NPRI). Specifically, a subset of the NPRI

data (emissions and stack parameters) relevant to the air quality modeling were reformatted.

The resulting emissions represent a significant improvement in the base year emissions.

A spatial plot of point source SO2 and NOx emissions is provided in Figure 47. Additional plots

and emission reports are available on the LADCO website

(http://www.ladco.org/tech/emis/basem/canada/index.htm

Figure 47. Canadian point source emissions for SO2 (left) and NOx (right)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Fires: For Base K, a contractor (EC/R, 2004) developed a 2001, 2002, and 2003 fire emissions

inventory for eight Midwest States (five LADCO states plus Iowa, Minnesota, and Missouri),

including emissions from wild fires, prescribed fires, and agricultural burns. Projected emissions

were also developed for 2010 and 2018 assuming “no smoke management” and “optimal smoke

management” scenarios. An early model sensitivity run showed very little difference in modeled

2.5

concentrations. Consequently, the fire emissions were not included in subsequent

modeling runs (i.e., they were not in the Base K or Base M modeling inventories).

Future Year Emissions

: Complete emission inventories were developed for several future years:

Base K – 2009, 2012, and 2018, and Base M – 2009 and 2018. In addition, 2008 (Base K and

Base M) and 2012 (Base M) proxy inventories were estimated based on the 2009 and 2018

data. (Note, the EGU emissions for the Base M 2012 inventory were based on EPA’s IPM3.0

modeling.)

Source sector emission summaries for the base years and future years are shown in Figure 48.

Additional detail is provided in Tables 6a and 6b.

VOC

NOx

SO2

3000

6000

9000

12000

2002 2005 2009 2012 2015 2018

Area

On-road

Nonroad

Non-EGU

EGU

Figure 48. Base year and future year emissions for 5-State LADCO Region (TPD, July weekday)

For on-road, and nonroad, the future year emissions were estimated by models (i.e.,

EMS/CONCEPT and NMIM, respectively). One adjustment was made to the 2009 and 2018

motor vehicle emission files prepared by Environ with CONCEPT. To reflect newer

transportation modeling conducted by CATS for the Chicago area, emissions were increased by

9% in 2009 and 2018. The 2005 base year and adjusted 2009 and 2018 motor vehicle

emissions are provided in Table 8.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Table 8. Motor Vehicle Emissions Produced by CONCEPT Modeling (July weekday – tons per day)

Year

State

Sum of CO

Sum of TOG

Sum of NOx

Sum of PM2.5

Sum of SO2

Sum of NH3

Sum of VMT

2005

3,684.3

341.5

748.2

12.9

9.6

35.9

344,087,819.6

3,384.9

282.0

541.1

8.9

11.1

25.7

245,537,231.9

4,210.3

351.9

722.0

12.4

13.9

35.3

340,834,025.9

2,569.1

218.7

380.5

6.3

7.6

17.7

170,024,599.7

6,113.4

679.8

933.6

16.2

18.8

36.5

360,521,068.6

2,206.0

175.1

457.5

7.8

9.2

19.7

189,123,964.3

Total

22,168.0

2,049.0

3,782.9

64.5

70.2

170.8

1,650,128,709.9

2009

2,824.4

268.0

527.8

10.1

4.2

38.9

372,132,591.1

2,839.5

234.9

401.9

6.7

2.8

26.1

249,817,026.3

3,172.0

269.2

500.9

9.2

4.0

37.1

356,347,010.5

2,256.8

206.3

307.5

5.1

2.3

21.5

204,443,017.8

4,619.2

423.7

693.5

11.8

4.7

39.5

387,428,127.2

1,673.4

119.4

322.1

5.7

2.3

20.6

197,729,964.9

Total

17,385.3

1,521.5

2,753.6

48.7

20.3

183.6

1,767,897,737.8

2018

2,084.7

151.5

200.7

6.3

3.7

43.1

413,887,887.3

2,217.3

138.4

173.0

4.4

2.6

30.2

288,042,232.1

2,434.3

163.5

204.1

5.9

3.6

40.5

388,128,431.8

1,799.6

123.1

137.1

3.6

2.2

24.9

237,022,213.7

3,361.5

242.5

274.1

6.8

4.0

43.1

421,694,093.4

1,255.5

68.4

138.5

3.9

2.0

22.2

218,277,167.5

Total

13,152.9

887.5

1,127.5

30.8

18.1

203.9

1,967,052,025.8

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

For EGUs, future year emissions were based on IPM2.1.9 modeling completed by the RPOs in

July 2005 Base K and IPM3.0 completed by EPA in February 2007 for Base M. Several CAIR

scenarios were assumed:

Base K

1a: IPM2.1.9, with full trading and banking

1b: IPM2.1.9, with restricted trading (compliance with state-specific emission budgets) and full trading

1d: IPM2.1.9, with restricted trading (compliance with state-specific emission budgets)

Base M

5a: EPA’s IPM3.0 was assumed as the future year base for EGUs.

5b: EPA’s IPM3.0, with several “will do” adjustments identified by the States. These adjustments should

reflect a legally binding commitment (e.g., signed contract, consent decree, or operating permit).

5c: EPA’s IPM3.0, with several “may do” adjustments identified by the States. These adjustments reflect

less rigorous criteria, but should still be some type of public reality (e.g., BART determination or press

announcement).

For other sectors (area, MAR, and non-EGU point sources), the future year emissions for the

LADCO States were derived by applying growth and control factors to the base year inventory.

These factors were developed by a contractor (E.H. Pechan, 2005 and E.H. Pechan, 2007).

For the non-LADCO States, future year emission files were based on data from other RPOs.

Growth factors were based initially on EGAS (version 5.0), and were subsequently modified (for

select, priority categories) by examining emissions activity data. Due to a lack of information on

future year conditions, the biogenic VOC and NOx emissions, and all Canadian emissions were

assumed to remain the constant between the base year and future years.

A “base” control scenario was prepared for each future year based on the following “on the

books” controls:

On-Highway Mobile Sources

•

Federal Motor Vehicle Emission Control Program, low-sulfur gasoline and ultra-low sulfur diesel fuel

•

Inspection - maintenance programs, including IL’s vehicle emissions tests (NE IL), IN’s vehicle

emissions testing program (NW IN), OH’s E-check program (NE OH), and WI’s vehicle inspection

program (SE WI) – note: a special emissions modeling run was done for the Cincinnati/Dayton area to

reflect the removal of the state’s E-check program and inclusion of low RVP gasoline

•

Reformulated gasoline, including in Chicago-Gary,-Lake County, IL,IN; and Milwaukee, Racine, WI

Off-Highway Mobile Sources

•

Federal control programs incorporated into NONROAD model (e.g., nonroad diesel rule), plus the

evaporative Large Spark Ignition and Recreational Vehicle standards

•

Heavy-duty diesel (2007) engine standard/Low sulfur fuel

•

Federal railroad/locomotive standards

•

Federal commercial marine vessel engine standards

Area Sources (Base M only)

•

Consumer solvents

•

AIM coatings

•

Aerosol coatings

•

Portable fuel containers

Power Plants

•

Title IV (Phases I and II)

•

NOx SIP Call

•

Clean Air Interstate Rule

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Other Point Sources

•

VOC 2-, 4-, 7-, and 10-year MACT standards

•

Combustion turbine MACT

Other controls included in the modeling include: consent decrees (refineries, ethanol plants, and

ALCOA)

, NOx RACT in Illinois and Ohio

, and BART for a few non-EGU sources in Indiana

and Wisconsin.

For Base K, several additional control scenarios were considered:

Scenario 2 – “base” controls plus additional controls recommended in LADCO White

Papers for stationary and mobile sources

Scenario 3 – Scenario 2 plus additional White Papers for stationary and mobile sources

Scenario 4 – “base” controls plus additional candidate control measures under

discussion by State Commissioners

Scenario 5 – “base” controls plus additional candidate control measures identified by the

LADCO Project Team

3.7 Basecase Modeling Results

The purpose of the basecase modeling is to evaluate model performance (i.e., assess the

model's ability to reproduce the observed concentrations). The model performance evaluation

focused on the magnitude, spatial pattern, and temporal of modeled and measured

concentrations. This exercise was intended to assess whether, and to what degree, confidence

in the model is warranted (and to assess whether model improvements are necessary).

Model performance was assessed by comparing modeled and monitored concentrations.

Graphical (e.g., side-by-side spatial plots, time series plots, and scatter plots) and statistical

analyses were conducted. No rigid acceptance/rejection criteria were used for this study.

Instead, the statistical guidelines recommended by EPA and other modeling studies (e.g.,

modeling by the other RPOs) were used to assess the reasonableness of the results. The

model performance results presented here describe how well the model replicates observed

ozone and PM

2.5

concentrations after a series of iterative improvements to model inputs.

Ozone

: Spatial plots are provided for high ozone periods in June 2002 and June 2005 (see

Figures 49a and 49b). The plots show that the model is doing a reasonable job of reproducing

the magnitude, day-to-day variation, and spatial pattern of ozone concentrations. There is a

tendency, however, to underestimate the magnitude of regional ozone levels. This is more

apparent with the 2002 modeling; the regional concentrations in the 2005 modeling agree better

with observations due to model and inventory improvements.

E.H. Pechan’s original control file included control factors for three sources in Wayne County, MI.

These control factors were not applied in the regional-scale modeling to avoid double-counting with the

State’s local-scale analysis for PM2.5

NOx RACT in Wisconsin is included in the 2005 basecase (and EGU “will do” scenario). NOx RACT in

Indiana was not included in the modeling inventory.

Figure 49a. Modeled (top) v. monitored (bottom) 8-hour ozone concentrations: June 20 – 25, 2002

Figure 49b Modeled (top) v. monitored (bottom) 8-hour ozone concentrations: June 23– 28 2005

Standard model performance statistics were generated for the entire 12 km domain, and by day

and by monitoring site. The domain-wide mean normalized bias for the 2005 base year is

similar to that for the 2002 base year and is generally within 30% (see Figure 50).

Figure 50. Mean bias for summer 2005 (Base M) and summer 2002 (Base K)

Station-average metrics (over the entire summer) are shown in Figure 51. The bias results

further demonstrate the model’s tendency to underestimate absolute ozone concentrations.

Figure 51. Mean bias (left) and gross error (right) for summer 2005

A limited 4 km ozone analysis was performed by LADCO to address the effect of grid spacing.

For this modeling, 4 km grids were placed over Lake Michigan and the Detroit-Cleveland area

(see Figure 52). Model inputs included 4 km emissions developed by LADCO (consistent with

Base K/Round 4) and the 4 km meteorology developed by Alpine Geophysics.

Figure 52. 4 km grids for Lake Michigan region and Detroit-Cleveland region

Hourly time series plots were prepared for several monitors (see Figure 53). The results are

similar at 12 km and 4 km, with some site-by-site and day-by-day differences.

Figure 53. Ozone time series plots for 12 km and 4 km modeling (June 17-29, 2002)

An additional diagnostic analysis was performed to assess the response of the modeling system

to changes in emissions (Baker and Kenski, 2007). Specifically, the 2002-to-2005 change in

observed ozone concentrations was compared to the change in modeled ozone concentrations

based on the 95

percentile(and above) concentration values for each monitor. This analysis

was also done with the inclusion of model performance criteria which eliminated poorly

performing days (i.e., error > 35%). The results show good agreement in the modeled and

monitored ozone concentration changes (e.g., ozone improves by about 9-10 ppb between

2002 and 2005 according to the model and the measurements) – see Figure 54. This provides

further support for using the model to develop ozone control strategies.

Figure 54. Comparison of change in predicted and observed ozone concentrations (2002 v. 2005)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

2.5

: Time series plots of the monthly average mean bias and annual fractional bias for Base

M and Base K are shown in Figure 55. As can be seen, Base M model performance for most

species is fair (i.e., close to “no bias” throughout most of the year), with two main exceptions.

First, the Base M and Base K results for organic carbon are poor, suggesting the need for more

work on primary organic carbon emissions. Second, the Base M results for sulfate, while

acceptable (i.e., bias values are within 35%), are not as good as the Base K results (e.g.,

noticeable underprediction during the summer months).

Figure 55. PM

2.5

Model performance - monthly average mean bias and annual fractional bias for

Base M (left column) and Base K (right column)

Base M

Base K

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Two analyses were undertaken to understand sulfate model performance for 2005:

•

Assess Meteorological Influences: The MM5 model performance evaluation showed that

rainfall is over-predicted by MM5 over most of the domain during the summer months

(LADCO, 2007c). Because CAMx does not explicitly use the rainfall output by MM5, this

may or may not result in over-prediction sulfate wet deposition (and under-prediction of

sulfate concentrations). A sensitivity run was performed with no wet deposition for July,

August, and September. The resulting model performance (see green line in Figure 56)

showed a noticeable difference from the basecase (i.e., higher sulfate concentrations),

and suggests that further evaluation of MM5 precipitation fields may be warranted.

•

Assess Emissions Influences: The major contributor to sulfate concentrations in the

region is SO2 emitted from EGUs. The basecase modeling inventory for EGUs is based

on annual emissions, which were allocated to a typical weekday, Saturday, and Sunday

by month using CEM-based temporal profiles. A sensitivity run was performed using

day-specific emissions. The resulting model performance (see purple line in Figure 56)

showed little difference from the basecase.

Figure 56. Monthly sulfate bias for Base M (MRPO EGU) v. two sensitivity analyses (Note: positive

values indicate over-prediction, negative values indicate under-prediction)

To assess the effect of the wet deposition issue on future year modeled values, another

sensitivity run was conducted with no wet deposition in Quarters 2-3 for the base year

(2005) and 2018. The resulting future year values were only slightly different from the

current base strategy run. In general, the future year values (without wet deposition)

were a little higher (+0.15 ug/m

or less) in the Ohio Valley and a little lower (-.10 ug/m

of less) in the Great Lakes region. This sensitivity run provides a bound for sulfate wet

deposition issue in terms of the attainment test, given that having no wet deposition is

unrealistic. The results suggest that even with an improved wet deposition treatment,

the Base M strategy results are not expected to change very much.

Time series plots of daily sulfate, nitrate, elemental carbon, and organic carbon concentrations

for three Midwestern locations are presented in Figures 57 (2002) and 58 (2005). These results

are consistent with the model performance statistics (i.e., good agreement for sulfates and

nitrates and poor agreement [large underprediction] for organic carbon).

Monthly Sulfate Bias (ug/m3)

-3

-2

-1

MRPO EGU

MOG daily EGU

No Wet Deposition

SULFATE

NITRATE

ORGANIC CARBON

Seney

Detroit

Chicago

Figure 57. Time series of sulfate, nitrate, and organic carbon at three Midwest sites for 2005

SULFATE

NITRATE

ORGANIC CARBON

Seney

Detroit

Chicago

Figure 58. Time series of sulfate, nitrate, and organic carbon at three Midwest sites for 2005

In summary, model performance for ozone and PM

2.5

is generally acceptable and can be

characterized as follows:

Ozone

•

Good agreement between modeled and monitored concentration for higher

concentration levels (> 60 ppb) – i.e., bias within 30%

•

Regional modeled concentrations appear to be underestimated in the 2002 base

year, but show better agreement (with monitored data) in the 2005 base year due to

model and inventory improvements.

•

Day-to-day and hour-to-hour variation in and spatial patterns of modeled

concentrations are consistent with monitored data

•

Model accurately simulates the change in monitored ozone concentrations due to

reductions in precursor emissions.

2.5

•

Good agreement in the magnitude of fine particle mass, but some species are

overestimated and some are underestimated (during periods of the year when it is

important)

•

Sulfates: good agreement in the 2002 base year, but underestimated in

the summer in the 2005 base year due probably to meteorological factors

•

Nitrates: slightly overestimated in the winter in the 2002 base year, but

good agreement in the 2005 base year as a result of model and inventory

improvements

•

Organic Carbon: grossly underestimated in the 2002 and 2005 base

years due likely to missing primary organic carbon emissions and,

possibly, other factors (e.g., grid resolution and model chemistry).

•

Temporal variation and spatial patterns of modeled concentrations are consistent

with monitored data

Several observations should be noted on the implications of these model performance findings

on the attainment modeling presented in the following section. First, it has been demonstrated

that model performance overall is acceptable and, thus, the model can be used for air quality

planning purposes. Second, consistent with EPA guidance, the model is used in a relative

sense to project future year values. EPA suggests that this approach “should reduce some of

the uncertainty attendant with using absolute model predictions alone” (EPA, 2007a).

Furthermore, the attainment modeling is supplemented by additional information to provide a

weight of evidence determination.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Section 4.0 Attainment Demonstration for Ozone and PM

2./5

Air quality modeling and other information were used to determine whether existing (“on the

books”) controls would be sufficient to provide for attainment of the NAAQS for ozone and PM

2.5

and if not, then what additional emission reductions would be necessary for attainment.

Traditionally, attainment demonstrations involved a “bright line” test in which a single modeled

value was compared to the ambient standard. To provide a more robust assessment of

expected future year air quality, EPA’s modeling guidelines call for consideration of

supplemental information. This section summarizes the results of the primary (guideline)

modeling analysis and a weight of evidence determination based on the modeling results and

other supplemental analyses.

4.1 Future Year Modeling Results

The purpose of the future year modeling is to assess the effectiveness of existing and possible

additional control programs. The model was used in a relative sense to project future year

values, which are then compared to the standard to determine attainment/nonattainment.

Specifically, the modeling test consists of the following steps:

(1) Calculate base year design values: For ozone and PM

2.5

, the base year design

values were derived by averaging the three 3-year periods centered on the

emissions base year:

2002 base year: 2000-2002, 2001-2003, and 2002-2004

2005 base year: 2003-2005, 2004-2006, and 2005-2007

(2) Estimate the expected change in air quality: For each grid cell, a relative

reduction factor (RRF) is calculated by taking the ratio of the future year and

baseline modeling results.

(3) Calculate future year values: For each grid cell (with a monitor), the RRFs are

multiplied by the base year design values to project the future year values

(4) Assess attainment: Future year values are compared to the NAAQS to assess

attainment or nonattainment.

A comparison of the 2002 and 2005 base year design values for ozone and PM

2.5

is provided in

Figure 59. In general, the figure shows that the 2005 base year design values are much lower

than the 2002 base year design values, especially for ozone.

A handful of source-oriented PM2.5 monitors in Illinois and Indiana were excluded from the annual

attainment test, because these monitors are not to be used to judging attainment of the annual standard.

Figure 59. 2002 v. 2005 base year design values for ozone (top) and PM

2.5

(bottom)

2002

2005

Statistical Summary

# Sites > NAAQS

Peak Value

99.0 ppb

90.0 ppb

Ave Exceedance Amount

7 ppb

2 ppb

2002

2005

Statistical Summary

# Sites > NAAQS

Peak Value

19.3 ug/m

17.7 ug/m

Ave Exceedance Amount

1.2 ug/m

0.9 ug/m

Ozone results are provided for those grid cells with ozone monitors. The RRF calculation

considers all nearby grid cells (i.e., 3x3 for 12 km modeling) and a threshold of 85 ppb. (If there

were less than 10 days above this value, then the threshold was lowered until either there were

10 days or the threshold reached 70 ppb.) PM

2.5

results are provided for those grid cells with

FRM (PM

2.5

-mass) monitors. Spatial mapping was performed to extrapolate PM

2.5

-speciation

data from STN and IMPROVE sites to FRM sites. RRF values for PM

2.5

were derived as a

function of quarter and chemical species.

Additional, hot-spot modeling will be performed by the states for certain PM

2.5

nonattainment

areas (e.g., Detroit, Cleveland, and Granite City) to address primary emissions from local point

sources which may not be adequately accounted for by the regional grid modeling. This

modeling will consist of Gaussian dispersion modeling (e.g., AERMOD) performed in

accordance with EPA’s modeling guidance (see Section 5.3 of the April 2007 guidance

document). Further analyses will need to be undertaken to determine how to best combine the

regional modeling and the hot-spot modeling. This could mean some adjustment to the model

results presented in this document to reflect better the regional component.

The ozone and PM

2.5

modeling results are provided in Appendix I for select monitors (high

concentration sites) in the 5-state region for the following future years of interest: 2008 (ozone

only), 2009, 2012, and 2018. (Note, RRF values for ozone, and for PM

2.5

by season and

chemical species are also included in Appendix I for key monitoring sites.) A summary of the

modeling results is provided in Table 9 (ozone) and Table 10 (PM

2.5

), and spatial maps of the

Base M future year concentrations are provided in Figures 60-62.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Key Sites

2018

Round 5

Round 4

Round 5

Round 4

Round 5

Round 4

Round 5

Lake Michigan Area

Chiwaukee

550590019

82.0

93.0

82.3

92.0

80.9

90.3

76.2

Racine

551010017

77.6

85.9

77.5

84.9

76.1

82.9

71.2

Milwaukee-Bayside

550190085

79.6

85.4

79.8

84.9

78.0

82.3

72.7

Harrington Beach

550890009

80.0

86.7

80.1

85.4

78.3

82.9

72.5

Manitowoc

550710007

81.3

80.3

80.8

78.9

78.6

76.3

72.5

Sheboygan

551170006

84.4

90.0

84.0

88.9

81.8

86.4

75.4

Kewaunee

550610002

78.9

82.5

78.1

81.0

75.9

79.1

69.9

Door County

550290004

84.8

83.6

83.9

81.8

81.5

79.3

74.7

Hammond

180892008

75.4

86.9

75.4

86.6

74.6

86.3

71.6

Whiting

180890030

77.0

76.2

73.1

Michigan City

180910005

74.2

87.4

73.9

86.5

72.5

85.4

68.1

Ogden Dunes

181270020

75.7

82.3

75.6

82.8

74.5

82.0

70.8

Holland

260050003

85.6

84.9

85.3

83.4

82.8

81.0

76.1

Jenison

261390005

77.9

78.7

77.1

77.6

74.5

75.5

68.7

Muskegon

261210039

80.8

82.7

80.5

81.5

78.0

79.4

71.9

Indianapolis Area

Noblesville

189571001

78.0

85.2

78.1

83.7

75.6

82.0

68.7

Fortville

180590003

73.9

85.1

73.9

83.8

71.4

82.1

65.1

Fort B. Harrison

180970050

74.8

84.8

75.1

83.7

73.2

82.4

69.1

Detroit Area

New Haven

260990009

82.7

86.3

81.4

85.3

80.2

83.5

76.1

Warren

260991003

82.5

84.3

81.3

83.3

80.7

81.9

77.6

Port Huron

261470005

79.0

80.5

77.5

79.1

75.5

77.0

70.9

Cleveland Area

Ashtabula

390071001

84.9

84.7

83.4

82.7

81.0

80.2

75.1

Geauga

390550004

75.7

90.3

74.7

88.8

72.7

86.2

67.3

Eastlake

390850003

82.8

84.2

81.9

82.8

80.5

80.6

76.2

Akron

391530020

79.3

83.0

78.1

81.4

75.6

78.5

68.7

Cincinnati Area

Wilmington

390271002

77.8

84.8

77.5

83.5

74.9

81.1

68.3

Sycamore

390610006

81.7

85.4

81.9

84.7

80.3

82.9

74.6

Lebanon

391650007

83.6

80.1

83.0

79.0

80.7

77.0

74.2

Columbus Area

London

390970007

75.4

79.9

75.0

78.4

72.6

76.5

66.3

New Albany

390490029

82.4

84.1

81.8

82.6

79.6

80.2

73.0

Franklin

290490028

77.0

77.7

75.9

76.5

74.1

74.7

69.0

St. Louis Area

W. Alton (MO)

291831002

82.4

86.1

81.0

85.2

78.6

84.0

74.9

Orchard (MO)

291831004

83.3

82.0

82.2

80.0

80.4

76.2

Sunset Hills (MO)

291890004

79.5

82.8

78.7

81.9

77.1

80.6

73.9

Arnold (MO)

290990012

78.7

78.4

77.2

77.4

75.6

75.8

72.0

Margaretta (MO)

295100086

79.8

84.0

79.3

83.4

77.9

82.5

74.4

Maryland Heights (MO)

291890014

84.5

83.4

81.7

78.1

2008

2009

2012

Table 9. Summary of Ozone Modeling Results

County

Site ID

Site

Round 5

Round4

Round 5

Round4

Round 5

Round4

Cook

170310022

Chicago - Washington HS

14.1

14.8

14.0

14.6

13.9

14.4

Cook

170310052

Chicago - Mayfair

14.4

15.8

14.2

15.5

13.9

15.0

Cook

170310057

Chicago - Springfield

13.9

14.5

13.8

14.3

13.7

14.1

Cook

170310076

Chicago - Lawndale

13.8

14.5

13.7

14.3

13.6

14.1

Cook

170312001

Blue Island

13.7

14.5

13.6

14.3

13.4

14.1

Cook

170313301

Summit

14.2

14.8

14.0

14.6

13.9

14.4

Cook

170316005

Cicero

14.4

15.3

14.3

15.1

14.2

14.9

Madison

171191007

Granite City

15.1

16.0

14.9

15.8

14.3

15.5

St. Clair

171630010

E. St. Louis

14.1

14.9

13.9

14.7

13.4

14.5

Clark

180190005

Jeffersonville

13.8

15.5

13.7

15.0

13.4

14.4

Dubois

180372001

Jasper

12.4

13.8

12.2

13.5

11.8

13.0

Lake

180890031

Gary

13.0

12.8

12.4

Marion

180970078

Indy-Washington Park

12.8

14.5

12.6

14.2

12.0

13.7

Marion

180970083

Indy- Michigan Street

13.4

14.8

13.1

14.9

12.6

14.0

Wayne

261630001

Allen Park

13.0

14.5

12.8

14.1

12.4

13.3

Wayne

261630015

Southwest HS

14.2

15.8

13.9

15.3

13.5

14.4

Wayne

261630016

Linwood

13.1

14.1

12.8

13.7

12.5

13.0

Wayne

261630033

Dearborn

15.8

17.7

15.5

17.1

15.1

16.1

Wayne

261630036

Wyandotte

13.1

15.1

12.8

14.7

12.5

13.9

Butler

390170003

Middleton

13.5

14.2

13.2

13.7

12.8

13.1

Butler

390170016

Fairfield

13.1

13.5

12.9

12.5

12.2

Cuyahoga 390350027

Cleveland-28th Street

13.5

14.4

13.2

13.8

12.7

12.9

Cuyahoga 390350038

Cleveland-St. Tikhon

15.2

16.1

14.8

15.4

14.3

14.4

Cuyahoga 390350045

Cleveland-Broadway

14.4

14.6

14.0

13.5

13.1

Cuyahoga 390350060

Cleveland-GT Craig

15.0

15.3

14.6

14.7

14.1

13.7

Cuyahoga 390350065

Newburg Hts - Harvard Ave

14.0

14.1

13.6

13.5

13.1

12.6

Franklin

390490024

Columbus - Fairgrounds

12.9

14.6

12.6

14.0

12.0

13.0

Franklin

390490025

Columbus - Ann Street

12.7

14.1

12.4

13.5

11.9

12.5

Franklin

390490081

Columbus - Maple Canyon

11.7

14.0

11.4

13.4

10.9

12.5

Hamilton

390610014

Cincinnati - Seymour

14.5

15.5

14.3

14.8

13.8

14.0

Hamilton

390610040

Cincinnati - Taft Ave

12.8

13.6

12.6

13.0

12.2

12.3

Hamilton

390610042

Cincinnati - 8th Ave

14.0

14.6

13.8

14.0

13.4

13.2

Hamilton

390610043

Sharonville

12.9

13.6

12.7

13.0

12.3

12.2

Hamilton

390617001

Norwood

13.4

14.2

13.2

13.6

12.8

Hamilton

390618001

St. Bernard

14.7

15.2

14.4

14.6

14.0

13.8

Jefferson

390810016

Steubenville

12.8

16.3

12.5

15.9

12.7

16.2

Jefferson

390811001

Mingo Junction

13.5

15.5

13.2

15.0

13.4

15.3

Lawrence

390870010

Ironton

12.8

14.2

12.5

13.7

12.3

13.2

Montgomery 391130032

Dayton

13.2

13.7

12.9

13.2

12.4

12.3

Scioto

391450013

New Boston

12.1

15.4

11.9

14.8

11.6

14.2

Stark

391510017

Canton - Dueber

14.0

15.0

13.6

14.3

13.3

13.6

Stark

391510020

Canton - Market

12.6

13.6

12.3

13.0

11.9

12.2

Summit

391530017

Akron - Brittain

13.0

14.4

12.7

13.6

12.3

12.9

Summit

391530023

Akron - W. Exchange

12.3

13.6

12.0

13.0

11.5

12.2

2009

2012

2018

Table 10. Summary of PM2.5 Modeling Results

2005 (observed)

2009

2012

2018

Figure 60. Observed base year and projected future year design values for ozone – Base M

2005 (observed)

2009

2012

2018

Figure 61. Observed base year and projected future year design values for PM

2.5

(annual average)–Base M

2005 (observed)

2009

2012

2018

Figure 62. Observed base year and projected future year design values for PM

2.5

(24-hr average)-Base M

The number of monitors with design values above the standard are as follows:

Table 11. Number of sites above standard

Ozone (8 hour: 85 ppb)

State 2002

2005

2009

2012

2018

BaseK Base M

Total

PM2.5 (Annual: 15 ug/m

)

State 2002

2005

2009

2012

2018

BaseK Base M

Total

The modeling results above reflect the “base” controls identified in Section 3.6, with EGU

emissions based on IPM modeling (i.e., Round 4 – IPM2.1.9, and Round 5 – IPM3.0). In

addition, two sets of alternative future year EGU emissions were examined in Round 5. First,

alternative control assumptions were provided for several facilities by the states (i.e., “will do”

and “may do” scenarios). In general, these scenarios produced a small change in future year

ozone and PM

2.5

concentrations (i.e., about 0.1 ug/m

for PM

2.5

and 0.1-0.2 ppb for ozone).

Second, EPA suggested adjustments to the 2010 IPM emissions to reflect 2009 conditions. The

revised (2009) SO2 emissions represent a 5-6% increase in domainwide SO2 emissions. The

increased SO2 emissions result in slightly greater annual average PM

2.5

concentrations (on the

order of 0.1 – 0.2 ug/m

), but do not produce any new residual nonattainment areas.

The limited 4 km ozone modeling (based on Base K) performed by LADCO included a future

year analysis for 2009. The figure below shows the 2009 values with 12 km and 4 km grid

spacing for the LADCO modeling and similar modeling conducted by a stakeholder group

(Midwest Ozone Group).

Cook

Lake

LaPorte

Porter

Allegan

Muskegon

Kenosha

Milwaukee

Ozaukee

Racine

Sheboygan

MOG-12km MOG-4km LADCO-12km LADCO-4km

New Haven

Warren

Geauga

Eastlake

Portage

Akron

MOG-12km MOG-4km LADCO-12km LADCO-4km

Figure 63. Future year (2009) values for Lake Michigan area (top) and Detroit-Cleveland region

(bottom)

These results show that the 12 km and 4 km values are similar, with the most notable changes

in northwestern Indiana and northeastern Illinois (e.g., 4 km values are as much as 4 ppb lower

than 12 km values). The differences in the southern part of the Lake Michigan area are

plausible, given the tight emissions gradient there (i.e., finer grid resolution appears to provide

more appropriate representation).

In light of these findings, 12 km grid spacing can continue to be used for ozone modeling, but

the Base K/Round 4 results for northwestern Indiana/northeastern Illinois should be viewed with

caution (i.e., probably 1 – 4 ppb too high).

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

In summary, the ozone modeling provides the following information for the nonattainment areas

in the region (see Table 12):

Table 12. Ozone Nonattainment Areas in the LADCO Region (as of December 31, 2007)

Area Name

Category

Number of

Counties

Attainment

Deadline

Detroit-Ann Arbor, MI

Marginal

2007

Chicago-Gary-Lake County, IL-IN

Moderate

2010

Cleveland-Akron-Lorain, OH

Moderate

2010

Milwaukee-Racine, WI

Moderate

2010

Sheboygan, WI

Moderate

2010

St Louis, MO-IL

Moderate

2010

Allegan Co, MI

Subpart 1

2009

Cincinnati-Hamilton, OH-KY-IN

Subpart 1

2009

Columbus, OH

Subpart 1

2009

Door Co, WI

Subpart 1

2009

Kewaunee Co, WI

Subpart 1

2009

Manitowoc Co, WI

Subpart 1

2009

Marginal Areas (2007 attainment date): No modeling was conducted for the 2006 SIP planning

year. Rather, 2005 – 2007 air quality data are available to determine attainment.

Basic (Subpart 1) Areas (2009 attainment date): The modeling results for the 2008 SIP planning

year show:

•

Base K: all areas in attainment, except Cincinnati and Indianapolis

•

Base M: all areas in attainment, except Holland (Allegan County)

Moderate Areas (2010 attainment date): The modeling results for the 2009 SIP planning year

show:

•

Base K: all areas still in nonattainment

•

Base M: all areas in attainment

The PM

2.5

modeling results show:

•

Base K: all areas in attainment, except for Chicago, Cincinnati, Cleveland, Detroit,

Granite City (IL), Louisville, Portsmouth (OH), and Steubenville

•

Base M: all areas in attainment, except for Cleveland, Detroit, and Granite City (IL)

With respect to the new lower 8-hour ozone standard, the modeling about 30 sites in 2012 and

5 sites in 2018 with design values greater than 75 ppb. With respect to the new lower 24-hour

2.5

standard, the modeling shows 13 sites in 2012 and 10 in 2018 with design values greater

than 35 ug/m

4.2 Supplemental Analyses

EPA’s modeling guidelines recommend that attainment demonstrations consist of a primary

(guideline) modeling analysis and supplemental analyses. Three basic types of supplemental

analyses are recommended:

•

additional modeling

•

analyses of trends in ambient air quality and emissions, and

•

observational models and diagnostic analyses

Furthermore, according to EPA’s guidelines, if the future year modeled values are “close” to the

standard (i.e., 82 – 87 ppb for ozone and 14.5 – 15.5 ug/m

for PM

2.5

), then the results of the

primary modeling should be reviewed along with the supplemental information in a “weight of

evidence” assessment of whether each area is likely to achieve timely attainment.

A WOE determination for ozone and PM

2.5

is provided in the following sections. Special

attention is given to the following areas with future year modeled values that exceed or are

“close” to the ambient standard (see Appendix I):

Ozone

PM2.5

Lake Michigan area

Chicago, IL

Cleveland, OH

Cincinnati, OH

Granite City, IL

Detroit, MI

4.3 Weight-of-Evidence Determination for Ozone

The WOE determination for ozone consists of the primary modeling and other supplemental

analyses (some of which were discussed in Section 2). A summary of this information is

provided below.

Primary (Guideline) Modeling

: The guideline modeling is presented in Section 4.1. Key findings

from this modeling include:

•

Base M regional modeling shows attainment by 2008 and 2009 at all sites, except

Holland (MI), and attainment at all sites by 2012.

•

Base K modeling results reflect generally higher future year values, and show more

sites in nonattainment compared to the Base M modeling. The difference in the two

modeling analyses is due mostly to lower base year design values in Base M.

•

Base K and Base M modeling analyses are considered “SIP quality”, so the

attainment demonstration for ozone should reflect a weight-of-evidence approach,

with consideration of monitoring based information.

•

Base M modeling also shows that the proposed lower 8-hour standard will not be

met at many sites, even by 2018, with existing controls.

Additional Modeling:

Four additional modeling analyses were considered: (1) re-examination of

the primary modeling to estimate attainment probabilities, (2) remodeling with different

assumptions, (3) an unmonitored area analysis, and (4) EPA’s latest regional ozone modeling.

Each of these analyses is described below.

First, the primary modeling results (which were initially processed using EPA’s attainment test)

were re-examined to estimate the probability of attaining the ozone standard (Lopez, 2007, and

LADCO, 2008b). Seven estimates of future year ozone concentrations were calculated based

on model-based RRFs and appropriate monitor-based concentrations for each year between

2001 and 2007. RRF values for 2001, 2003, 2004, 2006, and 2007 were derived based on the

2002 and 2005 modeling results. Monitor-based concentrations reflect 4

high values, design

values, or average of three design values centered on the year in question. The probability of

attainment was determined as the percentage of these seven estimates below the standard.

The results indicate that sites in the Lake Michigan area (Chiwaukee, Sheboygan, Holland,

Muskegon), Cleveland (Ashtabula), and St. Louis (W Alton) have a fairly low probability of

attainment by 2009 (i.e., about 50% or less).

Second, the primary modeling analysis was redone with different types of assumptions for

calculating base year design values (i.e., using the 3-year period centered on base year, and

using the highest 3-year period that includes the base year), and for calculating RRFs (i.e.,

using all days with base year modeled value > 70 ppb, and using all days with base year

modeled value > 85 ppb, with at least 10 days and “acceptable” model performance). The

results for several high concentration sites are presented in Tables 13a and 13b for 2009. The

different modeling assumptions produce eight estimates of future year ozone concentrations.

The highest estimates are associated with base year design values representing the 3-year

average for 2001-2003, and the lowest estimates are associated with base year design values

representing the 3-year average 2004-2006. The different RRF approaches produce little

change in future year ozone concentrations. This suggests that future year concentration

estimates are most sensitive to the choice of the base year and the methodology used to derive

the base year design values.

Third, EPA’s modeling guidelines recommend that an “unmonitored area analysis” be included

as a supplemental analysis, particularly in nonattainment areas where the monitoring network

just meets or minimally exceeds the size of the network required to report data to EPA’s Air

Quality System. The purpose of this analysis is to identify areas where future year values are

predicted to be greater than the NAAQS.

Based on examination of the spatial plots in Figures 49a and 49b, the most notable areas of

high modeled ozone concentrations are over the Great Lakes. Over-water monitoring, however,

is not required by EPA

. A cursory analysis of unmonitored areas for ozone was performed by

LADCO using an earlier version of the 2002 base year modeling (i.e, Base I) (Baker, 2005).

Base year and future year “observed” values were derived for unmonitored grid cells using the

absolute modeled concentrations (in all grid cells) and the observed values (in monitored grid

cells). A spatial map of the estimated 2009 values is provided in Figure 64. As can be seen,

there are very few (over land) grid cells where additional monitors may be desirable. This

indicates that the current modeling analysis, which focuses on monitored locations, is

addressing areas of high ozone throughout the region.

Air quality measurements over Lake Michigan were collected by LADCO previously to understand

ozone transport in the area (see, for example, Figure 5). Due to cut-backs in USEPA funding, however,

these measurements were discontinued in 2003.

Table 13a. Primary and Additional Ozone Modeling Results – Lake Michigan and Cleveland Areas (2009)

2009 Modeling Results

Lake Michigan Area

Cleveland Area

Chiwaukee

Harr.Beach Sheboygan

DoorCounty

Holland

Hammond

MichiganCity

Ashtabula

Geauga

Eastlake

550590019

550890009

551170006

550290004

260050003

180892008

180910005

390071001 390550004

390850003

Attainment Test

(based on EPA guidance-2002 baseyear)

Base Year Design Value

(average of three 3-year periods)

98.3

93.0

97.0

91.0

94.0

88.3

90.3

95.7

99.0

92.7

RRF (all days > 85 ppb, or at least 10 days)

0.935

0.918

0.916

0.899

0.888

0.980

0.958

0.865

0.897

0.894

Future Year Design Value

91.9

85.4

88.9

81.8

83.5

86.5

82.8

88.8

82.9

Attainment Test

(based on EPA guidance-2005 baseyear)

Base Year Design Value

(average of three 3-year periods)

84.7

83.3

88.0

88.7

90.0

77.7

77.0

89.0

79.3

86.3

RRF (all days > 85 ppb, or at least 10 days)

0.972

0.961

0.955

0.946

0.948

0.971

0.960

0.937

0.942

0.949

Future Year Design Value

82.3

80.1

84.0

83.9

85.3

75.4

73.9

83.4

74.7

81.9

Weight of Evidence

(alternative approaches-2002baseyear)

Alt 1 - Base Year Des. Value

(3-year period centered on 2002)

101.0

98.0

100.0

94.0

97.0

90.0

93.0

99.0

103.0

95.0

Alt 2 - Base Year Des. Value

(Highest 3-year period including 2002 )

101.0

98.0

100.0

94.0

97.0

92.0

93.0

99.0

103

95.0

RRF (all days > 85 ppb, or at least 10 days)

0.935

0.918

0.916

0.899

0.888

0.980

0.958

0.865

0.897

0.894

Alt 1 - Future Year Projected Value

94.4

90.0

91.6

84.5

86.1

88.2

89.1

85.6

92.4

84.9

Alt 2 - Future Year Projected Value

94.4

90.0

91.6

84.5

86.1

90.2

89.1

85.6

92.4

84.9

Alt 1 - RRF (all days > 70 ppb)

0.933

0.918

0.912

0.907

0.893

0.969

0.947

0.876

0.907

0.900

Alt 1 - Future Year Projected Value

94.2

90.0

91.2

85.3

86.6

87.2

88.1

86.7

93.4

85.5

Alt 2 - Future Year Projected Value

94.2

90.0

91.2

85.3

86.6

89.1

88.1

86.7

93.4

85.5

Alt 2 - RRF (all days > 85 ppb, or at least 10

days; with acceptable model performance)

0.945

0.904

0.910

0.904

0.887

0.976

0.964

0.866

0.896

0.894

Alt 1 - Future Year Projected Value

95.4

88.6

91.0

85.0

86.0

87.8

89.7

85.7

92.3

84.9

Alt 2 - Future Year Projected Value

95.4

88.6

91.0

85.0

86.0

89.8

89.7

85.7

92.3

84.9

Weight of Evidence

(alternative approaches-2005baseyear)

Alt 1 - Base Year Des. Value

(3-year period centered on 2005)

83.0

79.0

86.0

88.0

76.0

86.0

77.0

86.0

Alt 2 - Base Year Des. Value

(Highest 3-year period including 2005)

86.0

88.0

89.0

90.0

93.0

79.0

78.0

91.0

86.0

89.0

Alt 1 - Future Year Projected Value

80.7

75.9

82.1

81.4

83.4

73.8

73.0

80.6

72.5

81.6

Alt 2 - Future Year Projected Value

83.6

84.6

85.0

85.1

88.2

76.7

74.9

85.3

81.0

84.5

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Table 13b. Primary and Additional Ozone Modeling Results – Cincinnati, Columbus, St. Louis, Indianapolis, and Detroit (2009)

2009 Modeling Results

Cincinnati Area

Columbus

St. Louis Area

Indianapolis Area

Detroit Area

Wilmington

Lebanon

Sycamore

NewAlbany

W. Alton

OrchardFarm

Noblesville

Fortville

New Haven

390271002

39165007 390610006

390490029

291831002

291831004

180571001 18059003

260990009

Attainment Test

(based on EPA guidance-2002 baseyear)

Base Year Design Value

(average of three 3-year periods)

94.3

90.7

94.0

90.0

93.7

91.3

92.3

RRF (all days > 85 ppb, or at least 10 days)

0.885

0.908

0.938

0.888

0.947

0.914

0.894

0.918

0.924

Future Year Design Value

83.5

82.4

85.1

83.5

85.2

82.3

83.8

85.3

Attainment Test

(based on EPA guidance-2005 baseyear)

Base Year Design Value

(average of three 3-year periods)

82.3

87.7

84.3

86.3

87.0

83.3

78.7

86.0

RRF (all days > 85 ppb, or at least 10 days)

0.941

0.947

0.967

0.947

0.938

0.942

0.945

0.947

Future Year Design Value

77.4

83.1

81.5

81.7

80.9

82.0

78.7

74.5

81.4

Weight of Evidence

(alternative approaches-2002baseyear)

Alt 1 - Base Year Des. Value

(3-year period centered on 2002)

96.0

92.0

93.0

95.0

91.0

92.0

96.0

94.0

97.0

Alt 2 - Base Year Des. Value

(Highest 3-year period including 2002 )

96.0

92.0

93.0

96.0

91.0

92.0

96.0

94.0

97.0

RRF (all days > 85 ppb, or at least 10 days)

0.885

0.908

0.938

0.888

0.947

0.914

0.894

0.918

0.924

Alt 1 - Future Year Projected Value

85.0

83.5

87.2

84.4

86.2

84.1

85.8

86.3

89.6

Alt 2 - Future Year Projected Value

85.0

83.5

87.2

85.2

86.2

84.1

85.8

86.3

89.6

Alt 1 - RRF (all days > 70 ppb)

0.885

0.914

0.940

0.901

0.945

0.911

0.912

0.907

0.918

Alt 1 - Future Year Projected Value

85.0

84.1

87.4

85.6

86.0

83.8

87.6

85.3

89.0

Alt 2 - Future Year Projected Value

85.0

84.1

87.4

86.5

86.0

83.8

87.6

85.3

89.0

Alt 2 - RRF (all days > 85 ppb, or at least 10 days;

with acceptable model performance)

0.880

0.911

0.940

0.886

0.951

0.913

0.894

0.916

0.935

Alt 1 - Future Year Projected Value

84.5

83.8

87.4

84.2

86.5

84.0

85.8

86.1

90.7

Alt 2 - Future Year Projected Value

84.5

83.8

87.4

85.1

86.5

84.0

85.8

86.1

90.7

Weight of Evidence

(alternative approaches-2005baseyear)

Alt 1 - Base Year Des. Value

(3-year period centered on 2005)

80.0

86.0

81.0

84.0

85.0

86.0

80.0

76.0

82.0

Alt 2 - Base Year Des. Value

(Highest 3-year period including 2005)

85.0

89.0

86.0

88.0

89.0

87.0

81.0

90.0

Alt 1 - Future Year Projected Value

75.3

81.4

78.3

79.5

79.7

81.0

75.6

72.0

77.7

Alt 2 - Future Year Projected Value

80.0

84.3

83.2

83.3

83.5

83.8

82.2

76.7

85.2

Figure 64. Estimated Future Year Values (unmonitored grid cells)

Finally, EPA’s latest regional ozone modeling was considered as corroborative information.

This modeling was performed as part of the June 2007 proposal to revise the ozone standard

(EPA, 2007b). EPA applied the CMAQ model with 2001 meteorology to first estimate ozone

levels in 2020 based on the current standard and national rules in effect or proposed (i.e., the

baseline), and then to evaluate strategies for attaining a more stringent (70 ppb) primary

standard. Baseline (2020) ozone levels were predicted to be below the current standard in 481

of the 491 counties with ozone monitors. Of the 10 counties predicted to be above the

standard, there is one county in the LADCO region (i.e., Kenosha County, WI at 86 ppb). This

result is consistent with LADCO’s Base K modeling for 2018 (i.e., Kenosha County, WI at 86.7

ppb), which is not surprising given that EPA’s modeling and LADCO’s Base K modeling have a

similar base year (2001 v. 2002).

Analysis of Trends:

EPA’s modeling guidelines note that while air quality models are generally

the most appropriate tools for assessing the expected impacts of a change in emissions, it may

also be possible to extrapolate future trends based on measured historical trends of air quality

and emissions. To do so, USEPA’s guidance suggests that ambient trends should first be

normalized to account for year-to-year variations in meteorological conditions (EPA, 2002).

Meterologically-adjusted 4

high 8-hour ozone concentrations were derived using the air quality

– meteorological regression model developed by EPA (i.e., Cox method – see Section 2.1).

The historical trend in these met-adjusted ozone concentrations were extrapolated to estimate

future year ozone concentrations based on historical and projected trends in precursor

emissions. Both VOC and NOx emissions affect ozone concentrations. Given that observation-

based methods show that urban areas in the region are generally VOC-limited and rural areas

in the region are NOx-limited (see Section 2.1), urban VOC emissions and regional NOx

emissions are considered important. The trends in urban VOC and regional NOx emissions

were calculated to produce appropriate weighting factors.

The resulting 2009 and 2012 ozone values are provided in Figure 65, along with the primary

and alternative modeling ozone values for key sites in the Lake Michigan, Cleveland, and

Cincinnati areas. The results reflect a fairly wide scatter, but, on balance, the supplemental

information is supportive of the primary modeling results (i.e., sites in the Lake Michigan area

and Cleveland are expected to be close to the standard).

Sheboygan

100

2002

2005

2009

2012

Holland

100

2002

2005

2009

2012

Sycamore

100

2002

2005

2009

2012

Ashtabula

100

2002

2005

2009

2012

Figure 65. Estimates of Future Year Ozone Concentrations – Lake Michigan Area (Sheboygan and Holland), Cincinnati (Sycamore), and

Cleveland (Ashtabula)

Note: Primary (guideline) modeling values (Base K and Base M results) are represented by large red diamonds, additional modeling

values by small black circles, and trends-based values by small pink squares

Observational Models and Diagnostic Analyses:

The observation-based modeling (i.e.,

MAPPER) is presented in Section 3. The key findings from this modeling are that most urban

areas are VOC-limited and rural areas are NOx-limited.

The primary diagnostic analysis is source apportionment modeling with CAMx to provide more

quantitative information on source region (and source sector) impacts (Baker, 2007a).

Specifically, the model estimated the impact of 18 geographic source regions (which are

identified in Figure 66) and 6 source sectors (EGU point, non-EGU point, on-road, off-road,

area, and biogenic sources) at ozone monitoring sites in the region.

Figure 66. Source regions (left) and key monitoring sites (right) for ozone modeling analysis

Modeling results for 2009 (Base M) and 2012 (Base K) are provided in Appendix II for several

key monitoring sites. For each monitoring site, there are two graphs: one showing sector-level

contributions, and one showing source region and sector-level contributions in terms of

percentages. (Note, in the sector-level graph, the contributions from NOx emissions are shown

in blue, and from VOC emissions in green.)

The sector-level results (see, for example, Figure 67) show that on-road and nonroad NOx

emissions generally have the largest contributions at the key monitor locations (> 15% each).

EGU and non-EGU NOx emissions are also important contributors (> 10% each). The source

group contributions vary by receptor location due to emissions inventory differences.

Figure 67. Source-sector results for Holland (left) and Ashtabula (right) monitors – 2009 (Base M)

Kenosha County

Cook County

Indianapolis

Holland

St. Louis

Detroit

Cincinnati

Door County

Cleveland

The source region results (see, for example, Figure 68) show that while nearby areas generally

have the highest impacts (e.g., the northeastern IL/northwestern IN/southeastern WI

nonattainment area contributes 25-35% to high sites in the Lake Michigan area, and Cleveland

nonattainment counties contribute 20-25% to high sites in northeastern Ohio), there is an even

larger regional impact (i.e., contribution from other states).

Figure 68. Source-region results for Holland (left) and Ashtabula (right) monitors – 2009 (Base M)

Summary:

Air quality modeling and other supplemental analyses were performed to estimate

future year ozone concentrations. Based on this information, the following general conclusions

can be made:

•

Existing (“on the books”) controls are expected to produce significant

improvement in ozone air quality.

•

The choice of the base year affects the future year model projections. A key

difference between the base years of 2002 and 2005 is meteorology. As noted

above, 2002 was more ozone conducive than 2005. The choice of which base

year to use as the basis for the SIP is a policy decision (i.e., how much safeguard

to incorporate).

•

Most sites are expected to meet the current 8-hour standard by the applicable

attainment date, except, for sites in western Michigan and, possibly, in eastern

Wisconsin and northeastern Ohio.

•

Current monitoring data show significant nonattainment in these areas (e.g.,

peak design values on the order of 90 – 93 ppb). It is not clear whether sufficient

emission reductions will occur in the next couple of years to provide for

attainment.

•

Attainment by the applicable attainment date is dependent on actual future year

meteorology (e.g., if the weather conditions are consistent with [or less severe

than] 2005, then attainment is likely) and actual future year emissions (e.g., if the

emission reductions associated with the existing controls are achieved, then

attainment is likely). On the other hand, if either of these conditions is not met,

then attainment may be less likely.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

4.3 Weight-of-Evidence Determination for PM

2.5

The WOE determination for PM

2.5

consists of the primary modeling and other supplemental

analyses. A summary of this information is provided below.

Primary (Guideline) Modeling

: The results of the guideline modeling are presented in Section

4.1. Key findings from this modeling include:

•

Base M regional modeling shows attainment by 2009 at all sites, except Detroit,

Cleveland, and Granite City, and attainment at all sites by 2012, except for Detroit

and Granite City.

The regional modeling for PM

2.5

does not reflect any air quality benefit expected

from local controls. States are conducting local-scale analyses and will use these

results, in conjunction with the regional-scale modeling, to support their attainment

demonstrations for PM

2.5

•

Base K modeling results reflect generally higher future year values, and show more

sites in nonattainment in 2009 and 2012 compared to the Base M modeling. The

difference in the two modeling analyses is due mostly to lower base year design

values in Base M.

•

Base K and Base M modeling analyses are considered “SIP quality”, so the

attainment demonstration for PM

2.5

should reflect a weight-of-evidence approach,

with consideration of monitoring based information.

•

Base M modeling also shows that the new PM

2.5

24-hour standard will not be met at

many sites, even by 2018, with existing controls.

Additional Modeling:

EPA’s latest regional PM

2.5

modeling was considered as corroborative

information. This modeling was performed as part of the September 2006 revision to the PM

2.5

standard (USEPA, 2006). EPA applied the CMAQ model with 2001 meteorology to estimate

2.5

levels in 2015 and 2020 first with national rules in effect or proposed, and then with

additional controls to attain the current standard (15 ug/m

annual/65 ug/m

daily). Additional

analyses were performed to evaluate strategies for attaining more stringent standards in 2020

(15/35, and 14/35). Baseline (2015) PM

2.5

levels were predicted to be above the current

standard in four counties in the LADCO region: Madison County, IL at 15.2 ug/m

, Wayne

County, MI at 17.4, Cuyahoga County, OH at 15.4, and Scioto County, OH at 15.6. These

results are consistent with LADCO’s Base K modeling for 2012/2018, which is not surprising

given that EPA’s modeling and LADCO’s Base K modeling have a similar base year (2001 v.

2002).

Observational Models and Diagnostic Analyses:

The observation-based modeling (i.e.,

application of thermodynamic equilibrium models) is presented in Section 3. The key findings

from this modeling are that PM

2.5

mass is sensitive to reductions in sulfate, nitric acid, and

ammonia concentrations. Even though sulfate reductions cause more ammonia to be available

to form ammonium nitrate (PM-nitrate increases slightly when sulfate is reduced), this increase

is generally offset by the sulfate reductions, such that PM

2.5

mass decreases. Under conditions

with lower sulfate levels (i.e., proxy of future year conditions), PM

2.5

is more sensitive to

reductions in nitric acid compared to reductions in ammonia.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

The primary diagnostic analysis is source apportionment modeling with CAMx to provide more

quantitative information on source region (and source sector) impacts (Baker, 2007b).

Specifically, the model estimated the impact of 18 geographic source regions (which are

identified in Figure 69) and 6 source sectors (EGU point, non-EGU point, on-road, off-road,

area, and biogenic sources) at PM

2.5

monitoring sites in the region.

Figure 69. Source regions (left) and key monitoring sites (right) for PM

2.5

modeling analysis

Modeling results for 2012 (Base K) and 2018 (Base M) are provided in Appendix III for several

key monitoring sites. For each monitoring site, there are two graphs: one showing sector-level

contributions, and one showing source region and sector-level contributions in terms of absolute

modeled values.

The sector-level results (see, for example, Figure 70) show that EGU sulfate, non-EGU-sulfate,

and area organic carbon emissions generally have the largest contributions at the key monitor

locations (> 15% each). Ammonia emissions are also important contributors (> 10%). The

source group contributions vary by receptor location due to emissions inventory differences.

Figure 70. Source-sector results for Detroit (left) and Cleveland (right) monitors – 2018 (Base M)

Cook County

Steubenville

Granite City

Detroit

Cincinnati

Cleveland

The source region results (see, for example, Figure 71) show that while nearby areas generally

have the highest impacts (e.g., Detroit nonattainment counties contribute 40% to high sites in

southeastern Michigan, and Cleveland nonattainment counties contribute 35% to high sites in

northeastern Ohio), there is an even larger regional impact (i.e., contribution from other states).

Figure 71. Source-region results for Detroit (left) and Cleveland (right) monitors – 2018 (Base M)

Summary:

Air quality modeling and other supplemental analyses were performed to estimate

future year PM

2.5

concentrations. Based on this information, the following general conclusions

can be made:

•

Existing (“on the books”) controls are expected to produce significant

improvement in PM

2.5

air quality.

•

The choice of the base year affects the future year model projections. It is not

clear how much of this is attributable to differences in meteorology, because, as

noted in Section 3, PM

2.5

concentrations are not as strongly influenced by

meteorology as ozone.

•

Most sites are expected to meet the current PM

2.5

standard by the applicable

attainment date, except for sites in Detroit, Cleveland, and Granite City.

•

Current monitoring data show significant nonattainment in these areas (e.g.,

peak design values on the order of 16 – 17 ug/m

). It is not clear whether

sufficient emission reductions will occur in the next couple of years to provide for

attainment. States are conducting local-scale analyses for Detroit, Cleveland,

and Granite City, in particular, to identify appropriate additional local controls.

•

Attainment by the applicable attainment date is dependent (possibly) on actual

future year meteorology and (more likely) on actual future year emissions (e.g., if

the emission reductions associated with the “on the books” controls are

achieved, then attainment is likely). On the other hand, if either of these

conditions is not met (especially, with respect to emissions), then attainment may

be less likely.

Section 5. Reasonable Progress Assessment for Regional Haze

Air quality modeling and other information were used to assess the improvement in visibility that

would be provided by existing (“on the books”) controls and possible additional control

programs. In determining reasonable progress for regional haze, Section 169A of the Clean Air

Act and EPA’s visibility rule requires states to consider five factors:

•

costs of compliance

•

time necessary for compliance

•

energy and non-air quality environmental impacts of compliance

•

remaining useful life of any existing source subject to such requirements

•

uniform rate of visibility improvement needed to attain natural visibility conditions

by 2064

The uniform rate of visibility improvement requirement can be depicted graphically in the form of

a “glide path” (see Figure 72).

Figure 72. Visibility “glide paths” for northern Class I areas (units: deciviews)

5.1 Class I Areas Impacted

EPA’s visibility rule requires a state to “address regional haze in each mandatory Class I

Federal area located within the State and in each mandatory Class I Federal area located

outside the State which may be affected by emissions from within the State.” (40 CFR Part

51.308(d)) To meet this requirement, technical analyses conducted by the RPOs were

consulted to obtain information on areas of influence and culpability for Class I areas in the

eastern U.S. (MRPO, 2007). A summary of this information is provided in Table 1 (MRPO,

2007). The table shows that every LADCO State impacts multiple Class I areas in the eastern

U.S.

2004

2010

2020

2030

2040

2050

2060 2064

Voyageurs

Bounday Waters

Isle Royale

Seney

Voyageurs

Boundry Waters

20% best visibility days

20% worst visibility days

Baseline

Conditions

Baseline

Conditions

Natural

Conditions

Table 14. Draft List of Class I Areas Impacted by LADCO States

AREA NAME

81.401 Alabama.

Sipsey Wilderness Area

(1)

81.404 Arkansas.

Caney Creek Wilderness Area

(2), (4)

Upper Buffalo Wilderness Area

(1),(2),(4),(5)

(2), (4)

(2)

81.408 Georgia.

Cohotta Wilderness Area

Okefenokee Wilderness Area

Wolf Island Wilderness Area

81.411 Kentucky.

Mammoth Cave NP

(1), (2), (5)

(1), (2)

(1), (2), (5)

81.412 Louisiana.

Breton Wilderness Area

81.413 Maine.

Acadia National Park

(3)

Moosehorn Wilderness Area.

(3)

81.414 Michigan.

Isle Royale NP.

(1), (2)

Seney Wilderness Area

(1), (2)

81.415 Minnesota.

Boundary Waters Canoe Area Wilderness

(2)

(1), (2)

Voyageurs NP

(2)

(1), (2)

81.416 Missouri.

Hercules-Glades Wilderness Area

(2), (4), (5)

(2), (4)

(2)

Mingo Wilderness Area

(2), (4), (5)

(2)

(2), (4)

(2)

81.419 New Hampshire.

Great Gulf Wilderness Area

(3)

(1), (3)

Pres. Range-Dry River Wilderness Area.

81.42 New Jersey.

Brigantine Wilderness Area

(3)

(1), (3)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

81.422 North Carolina.

Great Smoky Mountains NP{1}

(1)

Joyce Kilmer-Slickrock Wilderness Area{2}

Linville Gorge Wilderness Area.

Shining Rock Wilderness Area.

Swanquarter Wilderness Area

81.426 South Carolina.

Cape Romain Wilderness

81.428 Tennessee.

Great Smoky Mountains NP{1}.

(1)

Joyce Kilmer-Slickrock Wilderness{2}

81.431 Vermont.

Lye Brook Wilderness

(2), (3)

(1), (2), (3)

81.433 Virginia.

James River Face Wilderness.

(2)

(2), (5)

Shenandoah NP

(2), (3)

(1), (2), (3)

(2), (3)

(1),(2),(3),(5)

81.435 West Virginia.

Dolly Sods/Otter Creek Wilderness.

(2), (3)

(1), (2), (3)

(1),(2),(3),(5)

Key

(1) MRPO Back Trajectory Analyses

(2) MRPO PSAT Modeling

(3) MANE-VU Contribution Assessment

(4) Missouri-Arkansas Contribution Assessment

(5) VISTAS Areas of Influence

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

5.2 Future Year Modeling Results

For regional haze, the calculation of future year conditions assumed:

•

baseline concentrations based on 2000-2004 IMPROVE data, with updated

(subsitituted) data for Mingo, Boundary Waters, Voyageurs, Isle Royale, and

Seney (see Section 2.3);

•

use of the new IMPROVE light extinction equation; and

•

use of EPA default values for natural conditions, based on the new IMPROVE

light extinction equation.

The uniform rate of visibility improvement values for the 2018 planning year were derived (for

the 20% worst visibility days) based on a straight line between baseline concentration value

(plotted in the year 2004 -- end year of the 5-year baseline period) and natural condition value

(plotted in the year 2064 -- date for achieving natural conditions). Plots of these “glide paths”

with the Base M modeling results are presented in Figure 73 for Class I areas in the eastern

U.S. A tabular summary of measured baseline and modeled future year deciview values for

these Class I areas are provided in Table 15 (2002 base year) and Table 16 (2005 base year)

The haze results show that several Class I areas in the eastern U.S. are expected to be greater

than (less improved than) the uniform rate of visibility improvement values (in 2018), including

those in northern Michigan and several in the northeastern U.S. Many other Class I areas in the

eastern U.S. are expected to be less than (more improved than) the uniform rate of visibility

improvement values (in 2018). As noted above, states should consider these results, along with

information on the other four factors, in setting reasonable progress goals.

An assessment of the five factors was performed for LADCO and the State of Minnesota by a

contractor (EC/R, 2007). Specifically, ECR examined reductions in SO2 and NOx emissions

from EGUs and industrial, commercial and institutional (ICI) boilers; NOx emissions from mobile

sources and reciprocating engines and turbines; and ammonia emissions from agricultural

operations. The impacts of “on the books” controls were also examined to provide a frame of

reference for assessing the impacts of the additional control measures.

The results of ECR’s analysis of the five factors are summarized below:

Factor 1 (Cost of Compliance): The average cost effectiveness values (in terms of $M

per ton) are provided in Table 16. For comparison, cost-effectiveness estimates

previously provided for “on the books” controls include:

CAIR SO2: $700 - $1,200, NOx: $1,400 – $2.600 ($/T)

BART SO2: $300 - $963, NOx: $248 - $1,770

MACT SO2: $1,500, NOx: $7,600

Most of the cost-effectiveness values for the additional controls are within the range of

cost-effectiveness values for “on the books” controls.

Model results reflect the grid cell where the IMPROVE monitor is located.

Voyageurs

Boundary Waters

Isle Royale

Seney

Mammoth Cave

Upper Buffalo

Figure 73. Visibility modeling results for Class I areas in eastern U.S.

Mingo

Shenandoah

Dolly Sods

Bringantine

Lye Brook

Acadia

Figure 73 (cont.) Visibility modeling results for Class I areas in eastern U.S

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Worst 20%

2018

2009

2012

2018

Site

Baseline

URP

OTB

EGU2

(5-state region)

EGU2

(12-state region)

BOWA1

19.86

17.70

19.05

19.01

18.94

18.40

17.72

VOYA2

19.48

17.56

19.14

19.19

19.18

18.94

18.38

SENE1

24.38

21.35

22.98

22.71

22.38

21.26

20.63

ISLE1

21.59

19.21

20.46

20.28

20.04

19.09

18.64

HEGL1

26.75

22.76

24.73

24.34

23.85

23.01

22.04

MING1

28.15

24.08

25.18

24.67

24.01

22.53

21.45

CACR1

26.36

22.55

24.01

23.55

22.99

22.43

21.57

UPBU1

26.27

22.47

24.02

23.58

23.06

22.31

21.38

MACA1

31.37

26.14

28.06

27.03

25.52

24.27

22.57

DOSO1

29.04

24.23

24.86

23.59

22.42

21.60

20.15

SHEN1

29.31

24.67

24.06

22.79

21.57

20.43

19.42

JARI1

29.12

24.48

24.81

23.79

22.42

21.59

20.88

BRIG1

29.01

24.68

25.87

25.25

24.39

23.91

23.45

LYBR1

24.45

21.16

21.80

21.32

20.69

20.18

19.79

Best 20%

2018

2009

2012

2018

Site

Baseline

URP

OTB

EGU2

(5-state region)

EGU2

(12-state region)

BOWA1

6.42

6.71

6.73

6.87

6.83

6.81

VOYA2

7.09

7.21

7.25

7.34

7.31

7.26

SENE1

7.14

7.19

7.23

7.06

6.91

ISLE1

6.75

6.57

6.51

6.47

6.20

6.06

HEGL1

12.84

12.61

12.62

12.61

12.43

12.02

MING1

14.46

13.96

13.93

13.94

13.74

13.33

CACR1

11.24

10.91

10.92

10.90

10.75

10.42

UPBU1

11.71

11.47

11.46

11.42

11.28

11.01

MACA1

16.51

16.06

15.91

15.54

15.18

14.75

DOSO1

12.28

11.72

11.45

11.19

10.93

10.67

SHEN1

10.93

9.73

9.53

9.17

9.05

8.90

JARI1

14.21

13.56

13.33

12.97

12.65

12.46

BRIG1

14.33

13.74

13.69

13.47

13.32

13.21

LYBR1

6.36

6.12

6.05

5.96

5.88

5.82

Table 15. Haze Results - Round 4 (Based on 2000-2004)

Worst 20%

2018

2009

2012

2018

Site

Baseline

URP

OTB

OTB+Will DO

BOWA1

19.86

17.94

18.45

18.33

17.94

17.92

VOYA2

19.48

17.75

18.20

18.07

17.63

17.66

SENE1

24.38

21.64

23.10

23.04

22.59

22.42

ISLE1

21.59

19.43

20.52

20.43

20.09

20.13

ISLE9

21.59

19.43

20.33

20.22

19.84

19.82

HEGL1

26.75

23.13

24.72

24.69

24.22

24.17

MING1

28.15

24.27

25.88

25.68

24.74

24.83

CACR1

26.36

22.91

23.39

23.29

22.44

22.40

UPBU1

26.27

22.82

23.34

23.27

22.59

22.55

MACA1

31.37

26.64

27.11

27.01

26.10

26.15

DOSO1

29.05

24.69

24.00

23.90

23.00

23.04

SHEN1

29.31

25.12

24.99

24.87

23.92

23.95

JARI1

29.12

24.91

25.17

25.01

24.06

24.12

BRIG1

29.01

25.05

25.79

25.72

25.21

25.22

LYBR1

24.45

21.48

22.04

21.86

21.14

ACAD1

22.89

20.45

21.72

21.49

Best 20%

2018

2009

2012

2018

Site

Baseline

Max

OTB

OTB+Will DO

BOWA1

6.42

6.21

6.19

6.14

6.12

VOYA2

7.09

6.86

6.83

6.75

6.76

SENE1

7.14

7.57

7.58

7.71

7.78

ISLE1

6.75

6.62

6.59

6.60

6.62

ISLE9

6.75

6.56

6.55

6.52

6.50

HEGL1

12.84

12.51

12.32

11.66

11.64

MING1

14.46

14.07

13.89

13.28

13.29

CACR1

11.24

10.88

10.85

10.52

UPBU1

11.71

11.13

11.08

10.73

10.74

MACA1

16.51

15.76

15.69

15.25

DOSO1

12.28

11.25

11.23

11.00

11.01

SHEN1

10.93

10.13

10.11

9.91

JARI1

14.21

13.38

13.14

BRIG1

14.33

14.15

14.08

13.92

LYBR1

6.37

6.25

6.23

6.14

6.15

ACAD1

8.78

8.86

8.82

Table 16. Haze Results - Round 5.1 (Based on 2000-2004)

101

Table 17. Estimated Cost Effectiveness for Potential Control Measures

Average Cost effectiveness ($/ton)

Emission category

Control strategy

Region

SO2

NOX

NH3

EGU

EGU1

3-State

1,540

2,037

9-State

1,743

1,782

EGU2

3-State

1,775

3,016

9-State

1,952

2,984

ICI boilers

ICI1

3-State

2,992

2,537

9-State

2,275

1,899

ICI Workgroup

3-State

2,731

3,814

9-State

2,743

2,311

Reciprocating engines

3-State

538

emitting 100 tons/year or

9-State

506

Reciprocating engines

and turbines

3-State

754

Turbines emitting 100

tons/year or more

9-State

754

3-State

1,286

Reciprocating engines

emitting 10 tons/year or more

9-State

1,023

3-State

800

Turbines emitting 10

tons/year or more

9-State

819

Agricultural sources

10% reduction

3-State

31 - 2,700

9-State

31 - 2,700

15% reduction

3-State

31 - 2,700

9-State

31 - 2,700

Mobile sources

Low-NOX Reflash

3-State

241

9-State

241

MCDI

3-State

10,697

9-State

2,408

Anti-Idling

3-State

(430) - 1,700

9-State

(430) - 1,700

Cetane Additive Program

3-State

4,119

9-State

4,119

Cement Plants

Process Modification

Michigan

Conversion to dry kiln

Michigan

9,848

LoTox™

Michigan

1,399

Glass Manufacturing

LNB

Wisconsin

1,041

Oxy-firing

Wisconsin

2,833

Electric boost

Wisconsin

3,426

SCR

Wisconsin

1,054

SNCR

Wisconsin

1,094

Lime Manufacturing

Mid-kiln firing

Wisconsin

688

LNB

Wisconsin

837

SNCR

Wisconsin

1,210

SCR

Wisconsin

5,037

FGD

Wisconsin

128 - 4,828

Oil Refinery

LNB

Wisconsin

3,288

SNCR

Wisconsin

4,260

SCR

Wisconsin

17,997

LNB+FGR

Wisconsin

4,768

ULNB

Wisconsin

2,242

FGD

Wisconsin

1,078

102

Factor 2 (Time Necessary for Compliance): All of the control measures can be

implemented by 2018. Thus, this factor can be easily addressed.

Factor 3 (Energy and Non-Air Quality Environmental Impacts): The energy and other

environmental impacts are believed to be manageable. For example, the increased

energy demand from add-on control equipment is less than 1% of the total electricity

and steam production in the region, and solid waste disposal and wastewater treatment

costs are less than 5% of the total operating costs of the pollution control equipment. It

should also be noted that the SO2 and NOx controls would have beneficial

environmental impacts (e.g., reduced acid deposition and nitrogen deposition).

Factor 4 (Remaining Useful Life): The additional control measures are intended to be

market-based strategies applied over a broad geographic region. It is not expected that

the control requirements will be applied to units that will be retired prior to the

amortization period for the control equipment. Thus, this factor can be easily addressed.

Factor 5 (Visibility Impacts): The estimated incremental improvement in 2018 visibility

levels for the additional measures is shown in Figure 74, along with the cost-

effectiveness expressed in $M per deciview improvement). These results show that

although EGU and ICI boiler controls have higher cost-per-deciview values (compared

to some of the other measures), their visibility impacts are larger.

103

Visibility Improvement (dv)

0.00

0.25

0.50

0.75

1.00

SO2 NOx NH3

Strategy 1 Strategy 2

EGUs

ICI Boilers

Rec.Eng.

Mobile

Cost Effectiveness ($M/dv)

1000

2000

3000

4000

5000

6000

SO2 NOx NH3

SO2NOx NH3

SO2 NOx NH3

SO2 NOxNH3

SO2 NOx NH3

Strategy 1

Strategy 2

EGUs

ICI Boilers

Rec.Eng.

Mobile

Figure 74. Results of ECR analysis of reasonable progress factors – visibility improvement (Factor

5) is on top, and cost effectiveness (Factor 1) is on bottom

104

5.3 Weight-of-Evidence Determination for Haze

The WOE determination for haze consists of the primary modeling and other supplemental

analyses. A summary of this information is provided below.

Primary (Guideline) Modeling

: The results of the guideline modeling are presented in Section

4.1. Key findings from this modeling include:

•

Base M modeling results show that the northern Minnesota Class I areas are close

to the glide path, whereas the northern Michigan Class I areas are above the glide

path in 2018. Other sites in the eastern U.S. are close to (or below) the glide path,

except for Mingo (MO), Brigantine (NJ), and Acadia (ME).

•

Base K modeling results show that the northern Minnesota and northern Michigan

Class I areas are above the glide path in 2018. Other sites in the eastern U.S. are

close to (or below) the glide path.

•

The difference in the two modeling analyses is due mostly to differences in future

year emission projections, especially for EGUs (e.g., use of IPM2.1.9 v. IPM3.0).

•

Base K and Base M modeling analyses are considered “SIP quality”, so the

attainment demonstration for haze should reflect a weight-of-evidence approach,

with consideration of monitoring based information.

Additional Modeling:

Two additional modeling analyses were considered: (1) the primary

modeling redone with different baseline values, and (2) modeling by the State of Minnesota

which looked at different receptor locations in the northern Class I areas (MPCA, 2008). Each

of these analyses is described below.

First, the primary modeling analysis (Base M) was revised using an alternative baseline value.

Specifically, the data for the period 2000-2005 were used to calculate the baseline, given that

the Base M modeling reflects a 2005 base year. The results of this alternative analysis (see

Table 18) are generally consistent with the primary modeling (see Table 16).

Second, Minnesota’s modeling reflects a 2002 base year and much of the data developed by

LADCO for its modeling. (Note, Minnesota conducted modeling for LADCO’s domain at 36 km,

and for a statewide domain at 12 km.) The purpose of the 12 km modeling was to address local

scale impacts on the northern Class I areas at several locations, not just the location of the

IMPROVE monitor. Results for the Boundary Waters on the 20% worst days range from 18.3 –

19.0 dv, with an average value of 18.7 dv, which is consistent with Minnesota’s 36 km modeling

results at the IMPROVE monitor. This variability in visibility levels should be kept in mind when

reviewing the values presented in Tables 15, 16, and 18, which reflect results at the IMPROVE

monitor locations.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Worst 20%

2009

2012

2018

Site

Baseline

URP

OTB

OTB+Will DO

BOWA1

20.10

18.12

18.63

18.51

18.12

18.09

VOYA2

19.62

17.86

18.27

18.15

17.70

17.72

SENE1

24.77

21.94

23.44

23.39

22.94

22.77

ISLE1

21.95

19.71

20.84

20.76

20.41

20.44

ISLE9

21.95

19.71

20.65

20.55

20.15

20.13

HEGL1

27.45

23.67

25.30

25.27

24.79

24.73

MING1

28.92

24.86

25.88

25.68

24.74

24.83

CACR1

27.05

23.44

23.88

23.78

22.92

22.86

UPBU1

26.97

23.36

23.92

23.85

23.14

23.09

MACA1

31.76

26.93

27.42

27.32

26.39

26.44

DOSO1

29.36

24.92

24.20

24.11

23.19

23.23

SHEN1

29.45

25.23

25.06

24.94

23.98

24.01

JARI1

29.40

25.13

25.32

25.17

24.22

24.28

BRIG1

29.12

25.14

25.84

25.77

25.26

LYBR1

24.71

21.69

22.22

22.06

21.36

ACAD1

22.91

20.47

21.72

21.49

Best 20%

2009

2012

2018

Site

Baseline

URP

OTB

OTB+Will DO

BOWA1

6.40

6.20

6.17

6.13

6.10

VOYA2

7.05

6.82

6.78

6.71

SENE1

7.20

7.60

7.61

7.73

7.80

ISLE1

6.80

6.67

6.64

6.65

6.66

ISLE9

6.80

6.62

6.61

6.57

6.55

HEGL1

13.04

12.71

12.51

11.85

11.82

MING1

14.68

14.07

13.89

13.28

13.29

CACR1

11.62

11.24

11.20

10.86

UPBU1

11.99

11.41

11.36

11.01

11.02

MACA1

16.64

15.88

15.82

15.37

15.38

DOSO1

12.24

11.21

11.19

10.96

10.97

SHEN1

10.85

10.04

10.02

9.82

9.83

JARI1

14.35

13.51

13.27

BRIG1

14.36

14.17

14.10

13.94

LYBR1

6.21

6.11

6.09

6.01

ACAD1

8.57

8.67

8.66

8.62

Table 18. Haze Results - Round 5.1 (Based on 2000-2005)

106

Observational Models and Diagnostic Analyses

: The observation-based modeling (i.e.,

application of thermodynamic equilibrium models) is presented in Section 3. The key findings

from this modeling are that PM

2.5

mass is sensitive to reductions in sulfate, nitric acid, and

ammonia concentrations. Even though sulfate reductions cause more ammonia to be available

to form ammonium nitrate (PM-nitrate increases slightly when sulfate is reduced), this increase

is generally offset by the sulfate reductions, such that PM

2.5

mass decreases and visibility

improves. Under conditions with lower sulfate levels (i.e., proxy of future year conditions), PM

2.5

is more sensitive to reductions in nitric acid compared to reductions in ammonia.

As discussed in Section 2, thermodynamic equilibrium modeling based on data collected at

Seney indicates that PM

2.5

there is most sensitive to reductions in sulfate, but also responsive to

reductions in nitric acid (Blanchard, 2004). An analysis using data from the Midwest ammonia

monitoring network for a site in Minnesota (i.e., Great River Bluffs, which is the closest ammonia

monitoring site to the northern Class I areas) suggested that reductions in sulfate, nitric acid,

and ammonia concentrations will lower PM

2.5

concentrations and improve visibility levels in the

northern Class I areas.

Trajectory analyses for the 20% worst visibility days for the four northern Class I areas are

provided in Figure 75. (Note, this figure is similar to Figure 34, but the trajectory results for each

Class I area are displayed separately here.) The orange areas are where the air is most likely

to come from, and the green areas are where the air is least likely to come from. Darker

shading represents higher frequency. As can be seen, bad air days are generally associated

with transport from regions located to the south, and good air days with transport from Canada.

107

Seney

Isle Royale

Boundary Waters

Voyageurs

Figure 75. Trajectory analysis results for northern Class I areas on 20% worst visibility days

The primary diagnostic analysis is source apportionment modeling with CAMx to provide more

quantitative information on source region (and source sector) impacts (Baker, 2007b).

Specifically, the CAMx model was applied to provide source contribution information.

Specifically, the model estimated the impact of 18 geographic source regions (which are

identified in Figure 76) and 6 source se ctors (EGU point, non-EGU point, on-road, off-road,

area, and ammonia sources) at visibility/haze monitoring sites in the eastern U.S.

108

Figure 76. Source regions (left) and key monitoring sites (right) for haze modeling analysis

Modeling results for 2018 (Base K and Base M) are provided in Appendix IV for several key

monitoring sites (Class I areas). For each monitoring site, there are two graphs: one showing

sector-level contributions, and one showing source region and sector-level contributions in

terms of absolute modeled values.

The sector-level results (see, for example, Figure 77) show that EGU sulfate, non-EGU-sulfate,

and ammonia emissions generally have the largest contributions at the key monitor locations.

The source group contributions vary by receptor location due to emissions inventory differences.

Figure 77. Source-sector results for Seney (left) and Boundary Waters (right) – 2018 (Base M)

The source region results (see, for example, Figure 78) show that emissions from a number of

nearby states contribute to regional haze levels.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

109

Figure 78. Source-region results for Seney (left) and Boundary Waters (right) – 2018 (Base M)

Table 19 provides a summary of the estimated state-level culpabilities based on the LADCO

back trajectory analyses and the PSAT analyses for 2018.

Summary:

Air quality modeling and other supplemental analyses were performed to estimate

future year visibility levels. Based on this information, the following general conclusions can be

made:

•

Existing (“on the books”) controls are expected to improve visibility levels in the

northern Class I areas.

•

Visibility levels in a few Class I areas in the eastern U.S. are expected to be

greater than (less improved than) the uniform rate of visibility improvement

values in 2018, including those in northern Michigan and some in the

northeastern U.S.

•

Visibility levels in many other Class I areas in the eastern U.S. are expected to

be less than (more improved than) the uniform rate of visibility improvement

values in 2018.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

110

Table 19. State Culpabilities Based on PSAT Modeling and Trajectory Analyses

Boundary Waters

Seney

LADCO -

Round 4

PSAT

LADCO -

Round 5

PSAT

MPCA-

PSAT

CENRAP -

PSAT

LADCO -

Traj. Analysis

LADCO -

Round 4

PSAT

LADCO -

Round 5

PSAT

CENRAP -

PSAT

LADCO -

Traj. Analysis

Michigan

3.4%

4.8%

3.0%

1.9%

0.7%

13.8%

18.1%

14.7%

Minnesota

30.5%

23.5%

28.0%

30.6%

37.6%

4.8%

1.6%

3.8%

Wisconsin

10.4%

10.9%

10.0%

6.4%

10.6%

12.6%

10.9%

8.4%

Illinois

5.2%

5.1%

6.0%

3.5%

2.7%

13.0%

14.3%

7.4%

Indiana

2.9%

3.9%

3.0%

1.8%

1.2%

9.6%

11.6%

2.2%

Iowa

7.6%

8.3%

8.0%

2.5%

7.4%

6.2%

3.8%

5.7%

Missouri

5.2%

3.4%

6.0%

2.1%

3.3%

6.5%

4.8%

3.2%

N. Dakota

5.7%

1.1%

6.0%

4.6%

5.9%

1.5%

0.1%

0.6%

Canada

1.9%

2.7%

3.0%

12.5%

15.1%

2.1%

1.2%

11.1%

CENRAP-

WRAP

10.9%

13.5%

4.2%

10.1%

13.1%

10.0%

7.0%

83.6%

77.2%

73.0%

70.2%

94.6%

83.3%

76.4%

64.1%

Voyageurs

Isle Royale

LADCO -

Round 4

PSAT

LADCO -

Round 5

PSAT

MPCA-

PSAT

CENRAP -

PSAT

LADCO -

Traj. Analysis

LADCO -

Round 4

PSAT

LADCO -

Round 5

PSAT

CENRAP -

PSAT

LADCO -

Traj. Analysis

Michigan

2.0%

4.9%

2.0%

1.0%

1.6%

12.7%

13.4%

Minnesota

35.0%

20.2%

31.0%

31.5%

36.9%

14.1%

9.5%

Wisconsin

6.3%

7.9%

6.0%

3.7%

9.7%

16.3%

14.7%

Illinois

3.0%

7.1%

3.0%

1.8%

1.2%

7.0%

8.7%

Indiana

1.6%

4.6%

2.0%

0.8%

5.6%

5.2%

Iowa

7.4%

7.1%

7.0%

2.4%

10.2%

6.9%

8.3%

Missouri

4.3%

4.0%

1.6%

0.3%

3.9%

4.6%

N. Dakota

10.3%

1.7%

13.0%

6.1%

7.1%

3.6%

0.3%

Canada

2.7%

3.3%

5.0%

17.2%

13.3%

2.2%

1.7%

CENRAP-

WRAP

10.2%

13.7%

6.1%

16.5%

12.5%

12.6%

82.7%

74.5%

73.0%

72.2%

96.8%

84.9%

79.0%

111

Section 6. Summary

To support the development of SIPs for ozone, PM

2.5

, and regional haze in the States of Illinois,

Indiana, Michigan, Ohio, and Wisconsin, technical analyses were conducted by LADCO, its

member states, and various contractors. The analyses include preparation of regional

emissions inventories and meteorological modeling data for two base years, evaluation and

application of regional chemical transport models, and review of ambient monitoring data.

Analyses of monitoring data were conducted to produce a conceptual model, which is a

qualitative summary of the physical, chemical, and meteorological processes that control the

formation and distribution of pollutants in a given region. Key findings of the analyses include:

Ozone

•

Current monitoring data show about 20 sites in violation of the 8-hour ozone

standard of 85 ppb. Historical ozone data show a steady downward trend over the

past 15 years, especially since 2001-2003, due likely to federal and state emission

control programs.

•

Ozone concentrations are strongly influenced by meteorological conditions, with

more high ozone days and higher ozone levels during summers with above normal

temperatures.

•

Inter- and intra-regional transport of ozone and ozone precursors affects many

portions of the five states, and is the principal cause of nonattainment in some areas

far from population or industrial centers

2.5

•

Current monitoring data show 30 sites in violation of the annual PM

2.5

standard of 15

ug/m

. Nonattainment sites are characterized by an elevated regional background

(about 12 – 14 ug/m

) and a significant local (urban) increment (about 2 – 3 ug/m

Historical PM

2.5

data show a slight downward trend since deployment of the PM

2.5

monitoring network in 1999.

•

2.5

concentrations are also influenced by meteorology, but the relationship is more

complex and less well understood compared to ozone.

•

On an annual average basis, PM

2.5

chemical composition consists of mostly sulfate,

nitrate, and organic carbon in similar proportions.

Haze

•

Current monitoring data show visibility levels in the Class I areas in northern

Michigan are on the order of 22 – 24 deciviews. The goal of EPA’s visibility program

is to achieve natural conditions, which is on the order of 12 deciviews for these

Class I areas, by the year 2064.

•

Visibility impairment is dominated by sulfate and nitrate.

Air quality models were applied to support the regional planning efforts. Two base years were

used in the modeling analyses: 2002 and 2005. EPA’s modeling guidance recommends using

112

2002 as the baseline inventory year, but also allows for use of an alternative baseline inventory

year, especially a more recent year. Initially, LADCO conducted modeling with a 2002 base

year (i.e., Base K modeling, which was completed in 2006). A decision was subsequently made

to conduct modeling with a 2005 base year (i.e., Base M, which was completed in 2007).

Statistical analyses showed that 2002 and 2005 both had above normal ozone-conducive

conditions, although 2002 was more severe compared to 2005. Examination of multiple base

years provides for a more complete technical assessment. Both sets of model runs are

discussed in this document.

Basecase modeling was conducted to evaluate model performance (i.e., assess the model's

ability to reproduce the observed concentrations). This exercise was intended to assess

whether, and to degree, confidence in the model is warranted (and to assess whether model

improvements are necessary). Model performance for ozone and PM

2.5

was generally

acceptable and can be characterized as follows:

Ozone

•

Good agreement between modeled and monitored concentration for higher

concentration levels (> 60 ppb) – i.e., bias within 30%

•

Regional modeled concentrations appear to be underestimated in the 2002 base

year, but show better agreement (with monitored data) in the 2005 base year due to

model and inventory improvements.

•

Day-to-day and hour-to-hour variation in and spatial patterns of modeled

concentrations are consistent with monitored data

•

Model accurately simulates the change in monitored ozone concentrations due to

reductions in precursor emissions.

2.5

•

Good agreement in the magnitude of fine particle mass, but some species are

overestimated and some are underestimated

•

Sulfates: good agreement in the 2002 base year, but underestimated in

the summer in the 2005 base year due probably to meteorological factors

•

Nitrates: slightly overestimated in the winter in the 2002 base year, but

good agreement in the 2005 base year as a result of model and inventory

improvements

•

Organic Carbon: grossly underestimated in the 2002 and 2005 base

years due likely to missing primary organic carbon emissions

•

Temporal variation and spatial patterns of modeled concentrations are consistent

with monitored data

Future year strategy modeling was conducted to determine whether existing (“on the books”)

controls would be sufficient to provide for attainment of the standards for ozone and PM

2.5

and if

not, then what additional emission reductions would be necessary for attainment. Traditionally,

attainment demonstrations involved a “bright line” test in which a single modeled value (based

on EPA guidance) was compared to the ambient standard. To provide a more robust

assessment of expected future year air quality, other information was considered. Furthermore,

according to EPA’s modeling guidance, if the future year modeled values are “close” to the

113

standard (i.e., 82 – 87 ppb for ozone and 14.5 – 15.5 ug/m

for PM

2.5

), then the results of the

primary modeling should be reviewed along with the supplemental information in a “weight of

evidence” (WOE) assessment of whether each area is likely to achieve timely attainment. Key

findings of the WOE determination include:

•

Existing controls are expected to produce significant improvement in ozone and

2.5

concentrations and visibility levels.

•

The choice of the base year affects the future year model projections. A key

difference between the base years of 2002 and 2005 is meteorology. 2002 was

more ozone conducive than 2005. The choice of which base year to use as the

basis for the SIP is a policy decision (i.e., how much safeguard to incorporate).

•

Most sites are expected to meet the current 8-hour standard by the applicable

attainment date, except for sites in western Michigan and, possibly, in eastern

Wisconsin and northeastern Ohio.

•

Most sites are expected to meet the current PM

2.5

standard by the applicable

attainment date, except for sites in Detroit, Cleveland, and Granite City.

The regional modeling for PM

2.5

does not reflect air quality benefits expected

from local controls. States are conducting local-scale analyses and will use

these results, in conjunction with the regional-scale modeling, to support their

attainment demonstrations for PM

2.5

•

These findings of residual nonattainment for ozone and PM

2.5

are supported by

current (2005 – 2007) monitoring data which show significant nonattainment in

the region (e.g., peak ozone design values on the order of 90 – 93 ppb, and peak

2.5

design values on the order of 16 - 17 ug/m

). It is unlikely that sufficient

emission reductions will occur in the next few of years to provide for attainment at

all sites.

•

Attainment at most sites by the applicable attainment date is dependent on actual

future year meteorology (e.g., if the weather conditions are consistent with [or

less severe than] 2005, then attainment is likely) and actual future year

emissions (e.g., if the emission reductions associated with the existing controls

are achieved, then attainment is likely). If either of these conditions is not met,

then attainment may be less likely.

•

The new PM

2.5

24-hour standard and the new lower ozone standard will not be

met at several sites, even by 2018, with existing controls.

•

Visibility levels in a few Class I areas in the eastern U.S. are expected to be

greater than (less improved than) the uniform rate of visibility improvement

values in 2018 based on existing controls, including those in northern Michigan

and some in the northeastern U.S. Visibility levels in many other Class I areas in

the eastern U.S. are expected to be less than (more improved than) the uniform

rate of visibility improvement values in 2018.

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

114

Section 7. References

Alpine Geophysics, 2007, “Emissions Inventory Assistance: 2005 Base Year Biogenic and

Other (non-LADCO) State Emissions, Preparation and Delivery of Non-MRPO Emission Files<

March 12, 2007.

AER, 2004, “Analysis of Recent Regional Haze Data”, Atmospheric and Environmental

Research, Inc., August 2004.

Baker, K., “Hot Spot Test (Draft), Lake Michigan Air Directors Consortium, March 23, 2005

Baker, K. 2007a, “Ozone Source Apportionment Results for Receptors in Non-Attainment

Counties in the Great Lakes Region”, Lake Michigan Air Directors Consortium, October 2007.

Baker, K., 2007b, “Source Apportionment Results for PM

2.5

and Regional Haze at Receptors in

the Great Lakes Region”, Lake Michigan Air Directors Consortium, November 2007.

Baker, K. and D. Kenski, 2007, “Diagnostic and Operational Evaluation of 2002 and 2005 8-

Hour Ozone to Support Model Attainment Demonstrations”, Lake Michigan Air Directors

Consortium, October 2007.

Blanchard, C.L., and S. Tanenbaum, 2003, “VOC and NO

Limitation of Ozone Formation at

Monitoring Sites in Illinois, Indiana, Michigan, Missouri, Ohio, and Wisconsin, 1998-2002”,

Report to LADCO, February 2003..

Blanchard, C.L., and S. Tanenbaum, 2004, “The Effects of Changes in Sulfate, Ammonia, and

Nitric Acid on Fine PM Composition at Monitoring Sites in Illinois, Indiana, Michigan, Missouri,

Ohio, and Wisconsin, 2000-2002”, Report to LADCO, February 2004.

Blanchard, C.L., and S. Tanenbaum, 2005a, “Weekday/Weekend Differences in Ambient

Concentrations of Primary and Secondary Air Pollutants in Atlanta, Baltimore, Chicago, Dallas-

Fort Worth, Denver, Houston, New York, Phoenix, Washington, and Surrounding Areas”,

Prepared for National Renewable Energy Laboratory, NREL Project ES04-1, July 30, 2005.

Blanchard, C.L., and S. Tanenbaum, 2005b, “Analysis of Data from the Midwest Ammonia

Monitoring Project”, Draft Final Technical Memorandum, March 31, 2005.

Breiman, L., J. Friedman, R. Olshen, and C. Stone, 1984, “Classification and Regression

Trees”, Chapman & Hall (1984).

Camalier, L., Cox, W., Dolwick, P., 2007. The effects of meteorology on ozone in urban areas

and their use in assessing ozone trends. Atmospheric Environment 41, 7127-7137

Cox, W.M, and S.-H. Chu, 1993, Meteorologically Adjusted Ozone Trends in Urban Areas: A

Probabililistic Approach, Atmospheric Environment 27B(4):425-434 (1993).

DeBell, L.J., K. Gebhart, J. Hand, W. Malm, M. Pitchford, B. Schichtel, and W. White, 2006,

“Spatial and Seasonal Patterns and Temporal Variability of Haze and its Constituents in the

United States, Report IV”, Cooperative Institute for Research in the Atmosphere, November

2006

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

115

DRI, 2005, “Source Apportionment Analysis of Air Quality Monitoring Data: Phase II”, Desert

Research Institute. March 2005

EC/R, Incorporated, 2004, “Fire Emissions Inventory Development for the Midwest Regional

Planning Organization”, Final Report, September 30, 2004.

EC/R, Incorporated, 2007, “Reasonable Progress for Class I Areas in the Northern Midwest –

Factor Analysis”, Draft Final Technical Memorandum, July 18, 2007.

E.H. Pechan, 2004, “LADCO Nonroad Emissions Inventory Project – Development of Local

Data for Construction and Agricultural Equipment”, Final Report, September 10, 2004

E.H. Pechan, 2005, “Development of Updated Growth and Control Factors for Lake Michigan

Air Directors Consortium (LADCO)”, Draft Report, December 29, 2005.

E. H. Pechan, 2007, “Development of 2005 Base Year Growth and Control Factors for Lake

Michigan Air Directors Consortium (LADCO)”, Final Report, September 2007.

Environ, 2004, “LADCO Nonroad Emissions Inventory Project for Locomotive, Commercial

Marine, and Recreational Marine Emission Sources, Final Report, December 2004.

Environ, 2007a, “Modeling Weekend and Weekday Ozone in Southeast Michigan”, Draft Final

Report, Prepared for National Renewable Energy Laboratory, July 30, 2007.

Environ, 2007b, “LADCO 2005 Locomotive Emissions”, Draft, February 2007.

Environ, 2007c, “LADCO 2005 Commercial Marine Emissions”, Draft, March 2, 2007.

Environ, 2008, “LADCO On-Road Emission Inventory Development Using CONCEPT MV”,

January 2008..

Hollander, M. and D. Wolfe, 1973, “Nonparametric Statistical Methods”, John Wiley & Sons,

New York (1973).

Hopke, P. K., 2005, Analyses of Midwest PM-Related Measurements, Report to LADCO, March

2005.

Kenski, D.M., 2004, Quantifying Transboundary Transport of PM

2.5

: A GIS Analysis, Proc.

AWMA 97

Annual Conf., Indianapolis, June 2004.

Kenski, D.M., 2007a, “CART Analysis for Ozone Trends and Meteorological Similarity”, May 10,

2007.

Kenski, D.M., 2007b, “Impact of Missing Data on Worst Days at Midwest Northern Class 1

Areas”, Donna Kenski, Midwest RPO, March 12, 2007 (revised 6/19/07).

Kenski, D.M., 2008a, “Updated Cox Trends Analysis”, February 2008.

Kenski, D.M. 2008b, “Updated CART Analysis for Ozone through 2007”, January 30, 2008.

116

LADCO, 2005, “Meteorological Modeling Performance Summary for Applications to

2.5

/Haze/Ozone Modeling Projects”, February 18, 2005.

LADCO, 2006a, “Base K/Round 4 Strategy Modeling” Emissions”, May 16, 2006.

LADCO, 2006b, “Base K/Round 4 Modeling: Summary”, August 31, 2006.

LADCO, 2007a, “Modeling Protocol: 2002 Basecase Technical Details”, October 17, 2007

LADCO, 2007b, “Modeling Protocol: 2005 Basecase Technical Details”, October 19, 2007

LADCO, 2007c, “An Evaluation of an Annual 2005 MM5 Simulation to Support Photochemical

and Emissions Modeling Applications”, January 10, 2007.

LADCO, 2007d, “Base M/Round 5 Modeling: Summary (DRAFT), October 10, 2007

LADCO, 2008a, “Base M Strategy Modeling: Emissions (Revised)”, February 27, 2008

LADCO, 2008b, “Critique: Modeled Attainment Probability (i.e., Bob Lopez’ analysis)”, January

10, 2008

Lopez, 2007, “Assessing the Site-Specific and Region-wide Probability of Attainment Level

Ozone Air Quality under varying seasonal conditions using the results of 2002 (Round 4) and

2005 (Round 5) Regional Modeling”, Draft 2, Bob Lopez, WDNR, December 12, 2007

Mansell, G, Z. Wang, R. Zhang, J. Fadel, T. Ramsey, H. Xin, Y. Liang, and J. Arogo, 2005, “An

Improved Process-Based Ammonia Emissions Model for Agricultural Sources – Emission

Estimates”

MPCA, 2008, “Technical Support Document for the Minnesota State Implementation Plan for

Regional Haze”, Draft, March 4, 2008.

MRPO, 2001, “Principles for Regional Planning”, March 9, 2001.

MRPO, 2007, “Draft List of Class I Areas Located Within (or Impacted by) Midwest RPO

States”, June 26, 2007.

MRPO, 2008, “Regional Haze in the Upper Midwest: Summary of Technical Information”,

Version 2.2, February 22, 2008.

Millstein, D.E., R.A. Harley, and S.V. Hering, 2007, “Weekly cycles in fine particulate matter”,

Atmospheric Environment (2007), doi:10.1016/j.atmosenv.2007.10.010.

Poirot, R. L.; Wishinski, P. R.; Hopke, P. K.; Polissar, A. V.; Comparative Application of Multiple

Receptor Methods To Identify Aerosol Sources in Northern Vermont, Environ. Sci. Technol.,

(4); 820-820 (2002).

Sheesley, R.J., J.J. Schauer, E. Bean, and D. Kenski, Trends in Secondary Organic Aerosol at

a Remote Site in Michigan’s Upper Peninsula, Env. Sci. and Tech.

(24); 6491-6500 (Oct

2004).

117

STI, 2004, “Data Processing and Analysis of Aloft Air Quality Data Collected in the Upper

Midwest”, Executive Summary, STI-903470-2568-ES3, Sonoma Technology, Inc, August 5,

2004.

STI, 2006a, “Integration of Results for the Upper Midwest Urban Organics Study”, Final Report,

STI-903520-2942-FR, Sonoma Technology, Inc. March 31, 2006.

STI, 2008, “Data Analysis and Source Apportionment of PM

2.5

in Selected Midwestern Cities”,

Final Report, STI-907018.03-3264-DRF, Sonoma Technology, Inc., February 2008.

EPA, 1999, Regional Haze Regulations, 64 FR 35714, July 1, 1999.

EPA, 2003, “Guidance for Estimating Natural Visibility Conditions Under the Regional Haze

Rule”, EPA-454/B-03-005, September 2003.

EPA, 2002, “Recommended Approach for Performing Mid-Course Review of SIPs to Meet the

1-Hour NAAQS for Ozone”, January 2002.

EPA, 2005, Final Rule to Implement the 8-Hour Ozone National Ambient Air Quality Standard,

70 FR 71612, November 29, 2005

EPA, 2006, “Regulatory Impact Analysis of the Revisions to the National Ambient Air Quality

Standards”, September 2006 (http://www.epa.gov/ttn/ecas/ria.html#ria2007)

EPA, 2007a, "Guidance for 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/B07-002, April

2007.

EPA, 2007b, “Regulatory Impact Analysis of the Proposed Revisions to the National Ambient Air

Quality Standards”, EPA-452/R-07-008, July 2007

(http://www.epa.gov/ttn/ecas/ria.html#ria2007

)

EPA, 2007c, Clean Air Fine Particle Implementation Rule, 72 FR 20586, April 25, 2007

Zhang, R, T. Ramsey, J. Fadel, J. Arogo, Z. Wang, G. Mansell, and H. Xin, 2005, “An Improved

Process-Based Ammonia Emissions Model for Agricultural Sources – Model Development”

Web Sites:

http://www.ladco.org/tech/emis/basem/canada/index.htm

http://www.ladco.org/tech/emis/basek/BaseK_Reports.htm

http://www.ladco.org/tech/emis/r5/round5_reports.htm

http://vista.cira.colostate.edu/views/

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

118

APPENDIX I

Ozone and PM

Modeling Results

Key Sites

2005 BY 2002 BY

'03

'04

'05

'06

'07

'03-'05 '04-'06 '05-'07

Average Average

RRF

Round 5

Lake Michigan Area

Chiwaukee

550590019

84.7

98.3

0.968

82.0

Chiwaukee

Racine

551010017

80.3

91.7

0.966

77.6 Racine

Milwaukee-Bayside

550190085

82.7

91.0

0.963

79.6 Milwaukee-Bayside

Harrington Beach

550890009

83.3

93.0

0.960

80.0 Harrington Beach

Manitowoc

550710007

85.0

87.0

0.957

81.3 Manitowoc

Sheboygan

551170006

88.0

97.0

0.959

84.4

Sheboygan

Kewaunee

550610002

82.7

89.3

0.954

78.9 Kewaunee

Door County

550290004

101

88.7

91.0

0.956

84.8

Door County

Hammond

180892008

77.7

88.3

0.971

75.4 Hammond

Whiting

180890030

79.3

0.971

77.0 Whiting

Michigan City

180910005

77.0

90.3

0.964

74.2 Michigan City

Ogden Dunes

181270020

78.3

86.3

0.967

75.7 Ogden Dunes

Holland

260050003

90.0

94.0

0.951

85.6

Holland

Jenison

261390005

82.0

86.0

0.950

77.9 Jenison

Muskegon

261210039

85.0

90.0

0.951

80.8 Muskegon

Indianapolis Area

Noblesville

189571001

101

82.7

93.7

0.944

78.0 Noblesville

Fortville

180590003

78.0

91.3

0.948

73.9 Fortville

Fort B. Harrison

180970050

78.7

90.0

0.951

74.8 Fort B. Harrison

Detroit Area

New Haven

260990009

102

86.0

92.3

0.962

82.7

New Haven

Warren

260991003

101

84.0

90.0

0.982

82.5

Warren

Port Huron

261470005

82.7

88.0

0.956

79.0 Port Huron

Cleveland Area

Ashtabula

390071001

89.0

95.7

0.954

84.9

Ashtabula

Geauga

390550004

79.3

99.0

0.954

75.7 Geauga

Eastlake

390850003

86.3

92.7

0.959

82.8

Eastlake

Akron

391530020

83.7

93.3

0.948

79.3

Cincinnati Area

Wilmington

390271002

82.3

94.3

0.945

77.8 Wilmington

Sycamore

390610006

84.7

90.3

0.965

81.7 Sycamore

Lebanon

391650007

87.7

87.0

0.954

83.6

Lebanon

Columbus Area

London

390970007

79.7

88.7

0.946

75.4 London

New Albany

390490029

86.3

93.0

0.954

82.4

New Albany

Franklin

290490028

80.3

86.0

0.958

77.0 Franklin

St. Louis Area

W. Alton (MO)

291831002

86.3

90.0

0.954

82.4

W. Alton (MO)

Orchard (MO)

291831004

87.0

90.0

0.958

83.3

Orchard (MO)

Sunset Hills (MO)

291890004

82.3

88.3

0.966

79.5 Sunset Hills (MO)

Arnold (MO)

290990012

82.3

84.7

0.956

78.7 Arnold (MO)

Margaretta (MO)

295100086

83.0

87.7

0.962

79.8 Margaretta (MO)

Maryland Heights (MO)

291890014

87.3

0.967

84.5

Maryland Heights (MO)

4th High 8-hour Value

Des. Values (truncated)

2008 - OTB

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Key Sites

2005 BY 2002 BY

'03

'04

'05

'06

'07

'03-'05 '04-'06 '05-'07

Average Average

RRF

Round 5 Round 4

RRF

Round 5

Lake Michigan Area

Chiwaukee

550590019

84.7

98.3

0.972

82.3

92.0

0.971

82.2

Chiwaukee

Racine

551010017

80.3

91.7

0.965

77.5

84.9

0.964

77.4 Racine

Milwaukee-Bayside

550190085

82.7

91.0

0.965

79.8

84.9

0.964

79.7 Milwaukee-Bayside

Harrington Beach

550890009

83.3

93.0

0.961

80.1

85.4

0.960

80.0 Harrington Beach

Manitowoc

550710007

85.0

87.0

0.951

80.8

78.9

0.949

80.7 Manitowoc

Sheboygan

551170006

88.0

97.0

0.955

84.0

88.9

0.953

83.9

Sheboygan

Kewaunee

550610002

82.7

89.3

0.945

78.1

81.0

0.943

78.0 Kewaunee

Door County

550290004

101

88.7

91.0

0.946

83.9

81.8

0.945

83.8

Door County

Hammond

180892008

77.7

88.3

0.971

75.4

86.6

0.970

75.3 Hammond

Whiting

180890030

79.3

0.971

77.0

0.970

77.0 Whiting

Michigan City

180910005

77.0

90.3

0.960

73.9

86.5

0.959

73.8 Michigan City

Ogden Dunes

181270020

78.3

86.3

0.965

75.6

82.8

0.964

75.5 Ogden Dunes

Holland

260050003

90.0

94.0

0.948

85.3

83.4

0.947

85.2

Holland

Jenison

261390005

82.0

86.0

0.940

77.1

77.6

0.939

77.0 Jenison

Muskegon

261210039

85.0

90.0

0.947

80.5

81.5

0.945

80.3 Muskegon

Indianapolis Area

Noblesville

189571001

101

82.7

93.7

0.945

78.1

83.7

0.946

78.2 Noblesville

Fortville

180590003

78.0

91.3

0.947

73.9

83.8

0.948

73.9 Fortville

Fort B. Harrison

180970050

78.7

90.0

0.955

75.1

83.7

0.956

75.2 Fort B. Harrison

Detroit Area

New Haven

260990009

102

86.0

92.3

0.947

81.4

85.3

0.947

81.4 New Haven

Warren

260991003

101

84.0

90.0

0.968

81.3

83.3

0.969

81.4 Warren

Port Huron

261470005

82.7

88.0

0.937

77.5

79.1

0.938

77.5 Port Huron

Cleveland Area

Ashtabula

390071001

89.0

95.7

0.937

83.4

82.7

0.941

83.7

Ashtabula

Geauga

390550004

79.3

99.0

0.942

74.7

88.8

0.945

75.0 Geauga

Eastlake

390850003

86.3

92.7

0.949

81.9

82.8

0.954

82.4

Eastlake

Akron

391530020

83.7

93.3

0.934

78.1

81.4

0.935

78.2

Cincinnati Area

Wilmington

390271002

82.3

94.3

0.941

77.5

83.5

0.942

77.6 Wilmington

Sycamore

390610006

84.7

90.3

0.967

81.9

84.7

0.968

82.0 Sycamore

Lebanon

391650007

87.7

87.0

0.947

83.0

79.0

0.948

83.1

Lebanon

Columbus Area

London

390970007

79.7

88.7

0.941

75.0

78.4

0.942

75.0 London

New Albany

390490029

86.3

93.0

0.947

81.8

82.6

0.948

81.8 New Albany

Franklin

290490028

80.3

86.0

0.945

75.9

76.5

0.948

76.2 Franklin

St. Louis Area

W. Alton (MO)

291831002

86.3

90.0

0.938

81.0

85.2

0.932

80.5 W. Alton (MO)

Orchard (MO)

291831004

87.0

90.0

0.942

82.0

82.2

0.939

81.7 Orchard (MO)

Sunset Hills (MO)

291890004

82.3

88.3

0.956

78.7

81.9

0.954

78.5 Sunset Hills (MO)

Arnold (MO)

290990012

82.3

84.7

0.938

77.2

77.4

0.937

77.1 Arnold (MO)

Margaretta (MO)

295100086

83.0

87.7

0.955

79.3

83.4

0.955

79.3 Margaretta (MO)

Maryland Heights (MO)

291890014

87.3

0.955

83.4

0.954

83.3

Maryland Heights (MO)

4th High 8-hour Value

Des. Values (truncated)

2009 - OTB

2009 - Will Do

Key Sites

2005 BY 2002 BY

'03

'04

'05

'06

'07

'03-'05 '04-'06 '05-'07

Average Average

RRF

Round 5 Round 4

RRF

Round 5

Lake Michigan Area

Chiwaukee

550590019

84.7

98.3

0.956

80.9

90.3

0.900

76.2 Chiwaukee

Racine

551010017

80.3

91.7

0.947

76.1

82.9

0.886

71.2 Racine

Milwaukee-Bayside

550190085

82.7

91.0

0.944

78.0

82.3

0.880

72.7 Milwaukee-Bayside

Harrington Beach

550890009

83.3

93.0

0.939

78.3

82.9

0.870

72.5 Harrington Beach

Manitowoc

550710007

85.0

87.0

0.925

78.6

76.3

0.853

72.5 Manitowoc

Sheboygan

551170006

88.0

97.0

0.930

81.8

86.4

0.857

75.4 Sheboygan

Kewaunee

550610002

82.7

89.3

0.918

75.9

79.1

0.845

69.9 Kewaunee

Door County

550290004

101

88.7

91.0

0.919

81.5

79.3

0.843

74.7 Door County

Hammond

180892008

77.7

88.3

0.960

74.6

86.3

0.922

71.6 Hammond

Whiting

180890030

79.3

0.960

76.2

0.922

73.1 Whiting

Michigan City

180910005

77.0

90.3

0.942

72.5

85.4

0.884

68.1 Michigan City

Ogden Dunes

181270020

78.3

86.3

0.951

74.5

82.0

0.904

70.8 Ogden Dunes

Holland

260050003

90.0

94.0

0.920

82.8

81.0

0.846

76.1 Holland

Jenison

261390005

82.0

86.0

0.909

74.5

75.5

0.838

68.7 Jenison

Muskegon

261210039

85.0

90.0

0.918

78.0

79.4

0.846

71.9 Muskegon

Indianapolis Area

Noblesville

189571001

101

82.7

93.7

0.914

75.6

82.0

0.831

68.7 Noblesville

Fortville

180590003

78.0

91.3

0.916

71.4

82.1

0.835

65.1 Fortville

Fort B. Harrison

180970050

78.7

90.0

0.931

73.2

82.4

0.879

69.1 Fort B. Harrison

Detroit Area

New Haven

260990009

102

86.0

92.3

0.932

80.2

83.5

0.885

76.1 New Haven

Warren

260991003

101

84.0

90.0

0.961

80.7

81.9

0.924

77.6 Warren

Port Huron

261470005

82.7

88.0

0.913

75.5

77.0

0.858

70.9 Port Huron

Cleveland Area

Ashtabula

390071001

89.0

95.7

0.910

81.0

80.2

0.844

75.1 Ashtabula

Geauga

390550004

79.3

99.0

0.916

72.7

86.2

0.848

67.3 Geauga

Eastlake

390850003

86.3

92.7

0.932

80.5

80.6

0.883

76.2 Eastlake

Akron

391530020

83.7

93.3

0.903

75.6

78.5

0.821

68.7 Akron

Cincinnati Area

Wilmington

390271002

82.3

94.3

0.910

74.9

81.1

0.830

68.3 Wilmington

Sycamore

390610006

84.7

90.3

0.948

80.3

82.9

0.881

74.6 Sycamore

Lebanon

391650007

87.7

87.0

0.921

80.7

77.0

0.846

74.2 Lebanon

Columbus Area

London

390970007

79.7

88.7

0.911

72.6

76.5

0.832

66.3 London

New Albany

390490029

86.3

93.0

0.922

79.6

80.2

0.845

73.0 New Albany

Franklin

290490028

80.3

86.0

0.923

74.1

74.7

0.859

69.0 Franklin

St. Louis Area

W. Alton (MO)

291831002

86.3

90.0

0.911

78.6

84.0

0.868

74.9 W. Alton (MO)

Orchard (MO)

291831004

87.0

90.0

0.919

80.0

80.4

0.876

76.2 Orchard (MO)

Sunset Hills (MO)

291890004

82.3

88.3

0.937

77.1

80.6

0.897

73.9 Sunset Hills (MO)

Arnold (MO)

290990012

82.3

84.7

0.918

75.6

75.8

0.874

72.0 Arnold (MO)

Margaretta (MO)

295100086

83.0

87.7

0.939

77.9

82.5

0.896

74.4 Margaretta (MO)

Maryland Heights (MO)

291890014

87.3

0.936

81.7

0.894

78.1 Maryland Heights (MO)

4th High 8-hour Value

Des. Values (truncated)

2012 - OTB

2018 - OTB

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

2005 BY

2002 BY

Key Site

County

Site ID

'03

'04

'05

'06

'07

'03 - '05 '04 - '06 '05 - '07

Average

w/ 2007

Average

Round 5

Round4

Key Site

Chicago - Washington HS Cook

170310022

15.6 14.2 16.9 13.2 15.7

15.6

14.8

15.3

15.2

15.9

14.1

14.8

Chicago - Washington HS

Chicago - Mayfair

Cook

170310052

15.9 15.3 17.0 14.5 15.5

16.1

15.6

15.7

15.8

17.1

14.4

15.8

Chicago - Mayfair

Chicago - Springfield

Cook

170310057

15.6 13.8 16.7 13.5 15.1

15.4

14.7

15.1

15.0

15.6

13.9

14.5

Chicago - Springfield

Chicago - Lawndale

Cook

170310076

14.8 14.2 16.6 13.5 14.3

15.2

14.8

14.9

15.6

13.8

14.5

Chicago - Lawndale

Blue Island

Cook

170312001

14.9 14.1 16.4 13.2 14.3

15.1

14.6

14.8

15.6

13.7

14.5

Blue Island

Summit

Cook

170313301

15.6 14.2 16.9 13.8 14.8

15.6

15.0

15.2

15.2

16.0

14.2

14.8

Summit

Cicero

Cook

170316005

16.8 15.2 16.3 14.3 14.8

16.1

15.3

15.1

15.5

16.4

14.4

15.3

Cicero

Granite City

Madison

171191007

17.5 15.4 18.2 16.3 15.1

17.0

16.6

16.5

16.7

17.3

15.1

16.0

Granite City

E. St. Louis

St. Clair

171630010

14.9 14.7 17.1 14.5 15.6

15.6

15.4

15.7

15.6

16.2

14.1

14.9

E. St. Louis

Jeffersonville

Clark

180190005

15.8 15.1 18.5 15.0 16.5

16.5

16.2

16.7

16.4

17.2

13.8

15.5

Jeffersonville

Jasper

Dubois

180372001

15.7 14.4 16.9 13.5 14.4

15.7

14.9

15.2

15.5

12.4

13.8

Jasper

Gary

Lake

180890031

16.8 13.3 14.5

16.8

15.1

14.9

15.6

13.0

Gary

Indy-Washington Park

Marion

180970078

15.5 14.3 16.4 14.1 15.8

15.4

14.9

15.4

15.3

16.2

12.8

14.5

Indy-Washington Park

Indy-W 18th Street

Marion

180970081

16.2 15.0 17.9 14.2 16.1

16.4

15.7

16.1

16.0

13.4

Indy-W 18th Street

Indy- Michigan Street

Marion

180970083

16.3 15.0 17.5 14.1 15.9

16.3

15.5

15.8

15.9

16.6

13.4

14.8

Indy- Michigan Street

Allen Park

Wayne

261630001

15.2 14.2 15.9 13.2 12.8

15.1

14.4

14.0

14.5

15.8

13.0

14.5

Allen Park

Southwest HS

Wayne

261630015

16.6 15.4 17.2 14.7 14.5

16.4

15.8

15.5

15.9

17.3

14.2

15.8

Southwest HS

Linwood

Wayne

261630016

15.8 13.7 16.0 13.0 13.9

15.2

14.2

14.3

14.6

15.5

13.1

14.1

Linwood

Dearborn

Wayne

261630033

19.2 16.8 18.6 16.1 16.9

18.2

17.2

17.5

19.3

15.8

17.7

Dearborn

Wyandotte

Wayne

261630036

16.3 13.7 16.4 12.9 13.4

15.5

14.3

14.2

14.7

16.6

13.1

15.1

Wyandotte

Middleton

Butler

390170003

17.2 14.1 19.0 14.1 15.4

16.8

15.7

16.2

16.2

16.5

13.5

14.2

Middleton

Fairfield

Butler

390170016

15.8 14.7 17.9 14.0 14.9

16.1

15.5

15.6

15.8

15.9

13.1

13.5

Fairfield

Cleveland-28th Street

Cuyahoga

390350027

15.4 15.6 17.3 13.0 14.5

16.1

15.3

14.9

15.4

16.5

13.5

14.4

Cleveland-28th Street

Cleveland-St. Tikhon

Cuyahoga

390350038

17.6 17.5 19.2 14.9 16.2

18.1

17.2

16.8

17.4

18.4

15.2

16.1

Cleveland-St. Tikhon

Cleveland-Broadway

Cuyahoga

390350045

16.4 15.3 19.3 14.1 15.3

17.0

16.2

16.5

16.7

14.4

14.6

Cleveland-Broadway

Cleveland-E14 & Orange

Cuyahoga

390350060

17.2 16.4 19.4 15.0 15.9

17.7

16.9

16.8

17.1

17.6

15.0

15.3

Cleveland-E14 & Orange

Newburg Hts - Harvard Ave Cuyahoga

390350065

15.6 15.2 18.6 13.1 15.8

16.5

15.6

15.8

16.0

16.2

14.0

14.1

Newburg Hts - Harvard Ave

Columbus - Fairgrounds

Franklin

390490024

16.4 15.0 16.4 13.6 14.6

15.9

15.0

14.9

15.3

16.5

12.9

14.6

Columbus - Fairgrounds

Columbus - Ann Street

Franklin

390490025

15.3 14.6 16.5 13.8 14.7

15.5

15.0

15.1

16.0

12.7

14.1

Columbus - Ann Street

Columbus - Maple Canyon Franklin

390490081

14.9 13.6 14.6 12.9 13.1

14.4

13.7

13.5

13.9

16.0

11.7

14.0

Columbus - Maple Canyon

Cincinnati - Seymour

Hamilton

390610014

17.0 15.9 19.8 15.5 16.5

17.6

17.1

17.3

17.3

17.7

14.5

15.5

Cincinnati - Seymour

Cincinnati - Taft Ave

Hamilton

390610040

15.5 14.6 17.5 13.6 15.1

15.9

15.2

15.4

15.5

15.7

12.8

13.6

Cincinnati - Taft Ave

Cincinnati - 8th Ave

Hamilton

390610042

16.7 16.0 19.1 14.9 15.9

17.3

16.7

16.6

16.9

17.3

14.0

14.6

Cincinnati - 8th Ave

Sharonville

Hamilton

390610043

15.7 14.9 16.9 14.5 14.8

15.8

15.4

15.6

16.0

12.9

13.6

Sharonville

Norwood

Hamilton

390617001

16.0 15.3 18.4 14.4 15.1

16.6

16.0

16.2

16.3

13.4

14.2

Norwood

St. Bernard

Hamilton

390618001

17.3 16.4 20.0 15.9 16.1

17.9

17.4

17.3

17.6

17.3

14.7

15.2

St. Bernard

Steubenville

Jefferson

390810016

17.7 15.9 16.4 13.8 16.2

16.7

15.4

15.5

15.8

17.7

12.8

16.3

Steubenville

Mingo Junction

Jefferson

390811001

17.3 16.2 18.1 14.6 15.6

17.2

16.3

16.1

16.5

17.5

13.5

15.5

Mingo Junction

Ironton

Lawrence

390870010

14.3 13.7 17.0 14.4 15.0

15.0

15.5

15.2

15.7

12.8

14.2

Ironton

Dayton

Montgomery

391130032

15.9 14.5 17.4 13.6 15.6

15.9

15.2

15.5

15.5

15.9

13.2

13.7

Dayton

New Boston

Scioto

391450013

14.7 13.0 16.2 14.3 14.0

14.6

14.5

14.8

14.7

17.1

12.1

15.4

New Boston

Canton - Dueber

Stark

391510017

16.8 15.6 17.8 14.6 15.9

16.7

16.0

16.1

16.3

17.3

14.0

15.0

Canton - Dueber

Canton - Market

Stark

391510020

15.0 14.1 16.6 11.9 14.4

15.2

14.2

14.3

14.6

15.7

12.6

13.6

Canton - Market

Akron - Brittain

Summit

391530017

15.4 15.0 16.4 13.5 14.4

15.6

15.0

14.8

15.1

16.4

13.0

14.4

Akron - Brittain

Akron - W. Exchange

Summit

391530023

14.2 13.9 15.7 12.8 13.7

14.6

14.1

14.3

15.6

12.3

13.6

Akron - W. Exchange

Annual Average Conc.

Design Values

2009 Modeling Results

2005 BY

2002 BY

Key Site

County

Site ID

'03

'04

'05

'06

'07

'03 - '05 '04 - '06 '05 - '07

Average

w/ 2007

Average

Round 5

Round4

Key Site

Chicago - Washington HS Cook

170310022

15.6 14.2 16.9 13.2 15.7

15.6

14.8

15.3

15.2

15.9

14.0

14.6

Chicago - Washington HS

Chicago - Mayfair

Cook

170310052

15.9 15.3 17.0 14.5 15.5

16.1

15.6

15.7

15.8

17.1

14.2

15.5

Chicago - Mayfair

Chicago - Springfield

Cook

170310057

15.6 13.8 16.7 13.5 15.1

15.4

14.7

15.1

15.0

15.6

13.8

14.3

Chicago - Springfield

Chicago - Lawndale

Cook

170310076

14.8 14.2 16.6 13.5 14.3

15.2

14.8

14.9

15.6

13.7

14.3

Chicago - Lawndale

Blue Island

Cook

170312001

14.9 14.1 16.4 13.2 14.3

15.1

14.6

14.8

15.6

13.6

14.3

Blue Island

Summit

Cook

170313301

15.6 14.2 16.9 13.8 14.8

15.6

15.0

15.2

15.2

16.0

14.0

14.6

Summit

Cicero

Cook

170316005

16.8 15.2 16.3 14.3 14.8

16.1

15.3

15.1

15.5

16.4

14.3

15.1

Cicero

Granite City

Madison

171191007

17.5 15.4 18.2 16.3 15.1

17.0

16.6

16.5

16.7

17.3

14.9

15.8

Granite City

E. St. Louis

St. Clair

171630010

14.9 14.7 17.1 14.5 15.6

15.6

15.4

15.7

15.6

16.2

13.9

14.7

E. St. Louis

Jeffersonville

Clark

180190005

15.8 15.1 18.5 15.0 16.5

16.5

16.2

16.7

16.4

17.2

13.7

15.0

Jeffersonville

Jasper

Dubois

180372001

15.7 14.4 16.9 13.5 14.4

15.7

14.9

15.2

15.5

12.2

13.5

Jasper

Gary

Lake

180890031

16.8 13.3 14.5

16.8

15.1

14.9

15.6

12.8

Gary

Indy-Washington Park

Marion

180970078

15.5 14.3 16.4 14.1 15.8

15.4

14.9

15.4

15.3

16.2

12.6

14.2

Indy-Washington Park

Indy-W 18th Street

Marion

180970081

16.2 15.0 17.9 14.2 16.1

16.4

15.7

16.1

16.0

13.2

Indy-W 18th Street

Indy- Michigan Street

Marion

180970083

16.3 15.0 17.5 14.1 15.9

16.3

15.5

15.8

15.9

16.6

13.1

14.9

Indy- Michigan Street

Allen Park

Wayne

261630001

15.2 14.2 15.9 13.2 12.8

15.1

14.4

14.0

14.5

15.8

12.8

14.1

Allen Park

Southwest HS

Wayne

261630015

16.6 15.4 17.2 14.7 14.5

16.4

15.8

15.5

15.9

17.3

13.9

15.3

Southwest HS

Linwood

Wayne

261630016

15.8 13.7 16.0 13.0 13.9

15.2

14.2

14.3

14.6

15.5

12.8

13.7

Linwood

Dearborn

Wayne

261630033

19.2 16.8 18.6 16.1 16.9

18.2

17.2

17.5

19.3

15.5

17.1

Dearborn

Wyandotte

Wayne

261630036

16.3 13.7 16.4 12.9 13.4

15.5

14.3

14.2

14.7

16.6

12.8

14.7

Wyandotte

Middleton

Butler

390170003

17.2 14.1 19.0 14.1 15.4

16.8

15.7

16.2

16.2

16.5

13.2

13.7

Middleton

Fairfield

Butler

390170016

15.8 14.7 17.9 14.0 14.9

16.1

15.5

15.6

15.8

15.9

12.9

Fairfield

Cleveland-28th Street

Cuyahoga

390350027

15.4 15.6 17.3 13.0 14.5

16.1

15.3

14.9

15.4

16.5

13.2

13.8

Cleveland-28th Street

Cleveland-St. Tikhon

Cuyahoga

390350038

17.6 17.5 19.2 14.9 16.2

18.1

17.2

16.8

17.4

18.4

14.8

15.4

Cleveland-St. Tikhon

Cleveland-Broadway

Cuyahoga

390350045

16.4 15.3 19.3 14.1 15.3

17.0

16.2

16.5

16.7

14.0

Cleveland-Broadway

Cleveland-E14 & Orange

Cuyahoga

390350060

17.2 16.4 19.4 15.0 15.9

17.7

16.9

16.8

17.1

17.6

14.6

14.7

Cleveland-E14 & Orange

Newburg Hts - Harvard Ave Cuyahoga

390350065

15.6 15.2 18.6 13.1 15.8

16.5

15.6

15.8

16.0

16.2

13.6

13.5

Newburg Hts - Harvard Ave

Columbus - Fairgrounds

Franklin

390490024

16.4 15.0 16.4 13.6 14.6

15.9

15.0

14.9

15.3

16.5

12.6

14.0

Columbus - Fairgrounds

Columbus - Ann Street

Franklin

390490025

15.3 14.6 16.5 13.8 14.7

15.5

15.0

15.1

16.0

12.4

13.5

Columbus - Ann Street

Columbus - Maple Canyon Franklin

390490081

14.9 13.6 14.6 12.9 13.1

14.4

13.7

13.5

13.9

16.0

11.4

13.4

Columbus - Maple Canyon

Cincinnati - Seymour

Hamilton

390610014

17.0 15.9 19.8 15.5 16.5

17.6

17.1

17.3

17.3

17.7

14.3

14.8

Cincinnati - Seymour

Cincinnati - Taft Ave

Hamilton

390610040

15.5 14.6 17.5 13.6 15.1

15.9

15.2

15.4

15.5

15.7

12.6

13.0

Cincinnati - Taft Ave

Cincinnati - 8th Ave

Hamilton

390610042

16.7 16.0 19.1 14.9 15.9

17.3

16.7

16.6

16.9

17.3

13.8

14.0

Cincinnati - 8th Ave

Sharonville

Hamilton

390610043

15.7 14.9 16.9 14.5 14.8

15.8

15.4

15.6

16.0

12.7

13.0

Sharonville

Norwood

Hamilton

390617001

16.0 15.3 18.4 14.4 15.1

16.6

16.0

16.2

16.3

13.2

13.6

Norwood

St. Bernard

Hamilton

390618001

17.3 16.4 20.0 15.9 16.1

17.9

17.4

17.3

17.6

17.3

14.4

14.6

St. Bernard

Steubenville

Jefferson

390810016

17.7 15.9 16.4 13.8 16.2

16.7

15.4

15.5

15.8

17.7

12.5

15.9

Steubenville

Mingo Junction

Jefferson

390811001

17.3 16.2 18.1 14.6 15.6

17.2

16.3

16.1

16.5

17.5

13.2

15.0

Mingo Junction

Ironton

Lawrence

390870010

14.3 13.7 17.0 14.4 15.0

15.0

15.5

15.2

15.7

12.5

13.7

Ironton

Dayton

Montgomery

391130032

15.9 14.5 17.4 13.6 15.6

15.9

15.2

15.5

15.5

15.9

12.9

13.2

Dayton

New Boston

Scioto

391450013

14.7 13.0 16.2 14.3 14.0

14.6

14.5

14.8

14.7

17.1

11.9

14.8

New Boston

Canton - Dueber

Stark

391510017

16.8 15.6 17.8 14.6 15.9

16.7

16.0

16.1

16.3

17.3

13.6

14.3

Canton - Dueber

Canton - Market

Stark

391510020

15.0 14.1 16.6 11.9 14.4

15.2

14.2

14.3

14.6

15.7

12.3

13.0

Canton - Market

Akron - Brittain

Summit

391530017

15.4 15.0 16.4 13.5 14.4

15.6

15.0

14.8

15.1

16.4

12.7

13.6

Akron - Brittain

Akron - W. Exchange

Summit

391530023

14.2 13.9 15.7 12.8 13.7

14.6

14.1

14.3

15.6

12.0

13.0

Akron - W. Exchange

Annual Average Conc.

Design Values

2012 Modeling Results

2005 BY

2002 BY

Key Site

County

Site ID

'03

'04

'05

'06

'07

'03 - '05 '04 - '06 '05 - '07

Average

w/ 2007

Average

Round 5

OTB

Round 5

Will Do

Round4

Key Site

Chicago - Washington HS Cook

170310022

15.6 14.2 16.9 13.2 15.7

15.6

14.8

15.3

15.2

15.9

13.9

13.8

14.4

Chicago - Washington HS

Chicago - Mayfair

Cook

170310052

15.9 15.3 17.0 14.5 15.5

16.1

15.6

15.7

15.8

17.1

13.9

13.8

15.0

Chicago - Mayfair

Chicago - Springfield

Cook

170310057

15.6 13.8 16.7 13.5 15.1

15.4

14.7

15.1

15.0

15.6

13.7

13.5

14.1

Chicago - Springfield

Chicago - Lawndale

Cook

170310076

14.8 14.2 16.6 13.5 14.3

15.2

14.8

14.9

15.6

13.6

13.4

14.1

Chicago - Lawndale

Blue Island

Cook

170312001

14.9 14.1 16.4 13.2 14.3

15.1

14.6

14.8

15.6

13.4

13.3

14.1

Blue Island

Summit

Cook

170313301

15.6 14.2 16.9 13.8 14.8

15.6

15.0

15.2

15.2

16.0

13.9

13.8

14.4

Summit

Cicero

Cook

170316005

16.8 15.2 16.3 14.3 14.8

16.1

15.3

15.1

15.5

16.4

14.2

14.0

14.9

Cicero

Granite City

Madison

171191007

17.5 15.4 18.2 16.3 15.1

17.0

16.6

16.5

16.7

17.3

14.3

14.2

15.5

Granite City

E. St. Louis

St. Clair

171630010

14.9 14.7 17.1 14.5 15.6

15.6

15.4

15.7

15.6

16.2

13.4

13.3

14.5

E. St. Louis

Jeffersonville

Clark

180190005

15.8 15.1 18.5 15.0 16.5

16.5

16.2

16.7

16.4

17.2

13.4

14.4

Jeffersonville

Jasper

Dubois

180372001

15.7 14.4 16.9 13.5 14.4

15.7

14.9

15.2

15.5

11.8

11.9

13.0

Jasper

Gary

Lake

180890031

16.8 13.3 14.5

16.8

15.1

14.9

15.6

12.4

Gary

Indy-Washington Park

Marion

180970078

15.5 14.3 16.4 14.1 15.8

15.4

14.9

15.4

15.3

16.2

12.0

12.1

13.7

Indy-Washington Park

Indy-W 18th Street

Marion

180970081

16.2 15.0 17.9 14.2 16.1

16.4

15.7

16.1

16.0

12.6

12.7

Indy-W 18th Street

Indy- Michigan Street

Marion

180970083

16.3 15.0 17.5 14.1 15.9

16.3

15.5

15.8

15.9

16.6

12.6

14.0

Indy- Michigan Street

Allen Park

Wayne

261630001

15.2 14.2 15.9 13.2 12.8

15.1

14.4

14.0

14.5

15.8

12.4

13.3

Allen Park

Southwest HS

Wayne

261630015

16.6 15.4 17.2 14.7 14.5

16.4

15.8

15.5

15.9

17.3

13.5

14.4

Southwest HS

Linwood

Wayne

261630016

15.8 13.7 16.0 13.0 13.9

15.2

14.2

14.3

14.6

15.5

12.5

13.0

Linwood

Dearborn

Wayne

261630033

19.2 16.8 18.6 16.1 16.9

18.2

17.2

17.5

19.3

15.1

16.1

Dearborn

Wyandotte

Wayne

261630036

16.3 13.7 16.4 12.9 13.4

15.5

14.3

14.2

14.7

16.6

12.5

13.9

Wyandotte

Middleton

Butler

390170003

17.2 14.1 19.0 14.1 15.4

16.8

15.7

16.2

16.2

16.5

12.8

13.1

Middleton

Fairfield

Butler

390170016

15.8 14.7 17.9 14.0 14.9

16.1

15.5

15.6

15.8

15.9

12.5

12.6

12.2

Fairfield

Cleveland-28th Street

Cuyahoga

390350027

15.4 15.6 17.3 13.0 14.5

16.1

15.3

14.9

15.4

16.5

12.7

12.9

Cleveland-28th Street

Cleveland-St. Tikhon

Cuyahoga

390350038

17.6 17.5 19.2 14.9 16.2

18.1

17.2

16.8

17.4

18.4

14.3

14.5

14.4

Cleveland-St. Tikhon

Cleveland-Broadway

Cuyahoga

390350045

16.4 15.3 19.3 14.1 15.3

17.0

16.2

16.5

16.7

13.5

13.7

13.1

Cleveland-Broadway

Cleveland-E14 & Orange

Cuyahoga

390350060

17.2 16.4 19.4 15.0 15.9

17.7

16.9

16.8

17.1

17.6

14.1

14.2

13.7

Cleveland-E14 & Orange

Newburg Hts - Harvard Ave Cuyahoga

390350065

15.6 15.2 18.6 13.1 15.8

16.5

15.6

15.8

16.0

16.2

13.1

13.3

12.6

Newburg Hts - Harvard Ave

Columbus - Fairgrounds

Franklin

390490024

16.4 15.0 16.4 13.6 14.6

15.9

15.0

14.9

15.3

16.5

12.0

12.1

13.0

Columbus - Fairgrounds

Columbus - Ann Street

Franklin

390490025

15.3 14.6 16.5 13.8 14.7

15.5

15.0

15.1

16.0

11.9

12.5

Columbus - Ann Street

Columbus - Maple Canyon Franklin

390490081

14.9 13.6 14.6 12.9 13.1

14.4

13.7

13.5

13.9

16.0

10.9

11.0

12.5

Columbus - Maple Canyon

Cincinnati - Seymour

Hamilton

390610014

17.0 15.9 19.8 15.5 16.5

17.6

17.1

17.3

17.3

17.7

13.8

13.9

14.0

Cincinnati - Seymour

Cincinnati - Taft Ave

Hamilton

390610040

15.5 14.6 17.5 13.6 15.1

15.9

15.2

15.4

15.5

15.7

12.2

12.3

Cincinnati - Taft Ave

Cincinnati - 8th Ave

Hamilton

390610042

16.7 16.0 19.1 14.9 15.9

17.3

16.7

16.6

16.9

17.3

13.4

13.2

Cincinnati - 8th Ave

Sharonville

Hamilton

390610043

15.7 14.9 16.9 14.5 14.8

15.8

15.4

15.6

16.0

12.3

12.4

12.2

Sharonville

Norwood

Hamilton

390617001

16.0 15.3 18.4 14.4 15.1

16.6

16.0

16.2

16.3

12.8

Norwood

St. Bernard

Hamilton

390618001

17.3 16.4 20.0 15.9 16.1

17.9

17.4

17.3

17.6

17.3

14.0

14.1

13.8

St. Bernard

Steubenville

Jefferson

390810016

17.7 15.9 16.4 13.8 16.2

16.7

15.4

15.5

15.8

17.7

12.7

16.2

Steubenville

Mingo Junction

Jefferson

390811001

17.3 16.2 18.1 14.6 15.6

17.2

16.3

16.1

16.5

17.5

13.4

15.3

Mingo Junction

Ironton

Lawrence

390870010

14.3 13.7 17.0 14.4 15.0

15.0

15.5

15.2

15.7

12.3

13.2

Ironton

Dayton

Montgomery

391130032

15.9 14.5 17.4 13.6 15.6

15.9

15.2

15.5

15.5

15.9

12.4

12.5

12.3

Dayton

New Boston

Scioto

391450013

14.7 13.0 16.2 14.3 14.0

14.6

14.5

14.8

14.7

17.1

11.6

14.2

New Boston

Canton - Dueber

Stark

391510017

16.8 15.6 17.8 14.6 15.9

16.7

16.0

16.1

16.3

17.3

13.3

13.6

Canton - Dueber

Canton - Market

Stark

391510020

15.0 14.1 16.6 11.9 14.4

15.2

14.2

14.3

14.6

15.7

11.9

12.0

12.2

Canton - Market

Akron - Brittain

Summit

391530017

15.4 15.0 16.4 13.5 14.4

15.6

15.0

14.8

15.1

16.4

12.3

12.9

Akron - Brittain

Akron - W. Exchange

Summit

391530023

14.2 13.9 15.7 12.8 13.7

14.6

14.1

14.3

15.6

11.5

11.6

12.2

Akron - W. Exchange

Annual Average Conc.

Design Values

2018 Modeling Results

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

24-Hour PM

2.5

Base Year

Key Site

County

Site ID

'03

'04

'05

'06

'07

'03-'05 '04-'06 '05-'07

Average

w/ 2007

2009

2012

2018

Key Site

Chicago - Washington HS

Cook

170310022

37.7 32.5 45.7 27.0 35.7

38.6

35.1

36.1

36.6

Chicago - Washington HS

Chicago - Mayfair

Cook

170310052

37.3 38.8 48.3 31.6 39.4

41.5

39.6

39.8

40.3

Chicago - Mayfair

Chicago - Springfield

Cook

170310057

36.4 33.1 46.5 27.7 38.9

38.7

35.8

37.7

37.4

Chicago - Springfield

Chicago - Lawndale

Cook

170310076

32.6 39.7 45.1 29.0 37.2

39.1

37.9

37.1

38.1

Chicago - Lawndale

McCook

Cook

170311016

43.0

McCook

Blue Island

Cook

170312001

39.6 38.5 43.8 28.1 35.1

40.6

36.8

35.7

37.7

Blue Island

Schiller Park

Cook

170313103

40.7 50.3 30.0 36.6

45.5

40.3

39.0

41.6

Schiller Park

Summit

Cook

170313301

38.4 42.4 49.1 27.4 36.7

43.3

39.6

37.7

40.2

Summit

Maywood

Cook

170316005

38.5 42.5 44.6 29.2 36.9

41.9

38.8

36.9

39.2

Maywood

Granite City

Madison

171191007

40.8 35.4 44.1 36.3 36.0

40.1

38.6

38.8

39.2

Granite City

E. St. Louis

St. Clair

171630010

32.6 30.2 39.6 29.2 33.1

34.1

33.0

34.0

33.7

E. St. Louis

Jeffersonville

Clark

180190005

28.4 45.5 35.9 43.3

37.0

36.6

41.6

38.4

Jeffersonville

Jasper

Dubois

180372001

39.5 30.0 41.2 31.6 39.5

36.9

34.3

37.4

36.2

Jasper

Gary - IITRI

Lake

180890022

39.0

Gary - IITRI

Gary - Burr School

Lake

180890026

39.0

Gary - Burr School

Gary

Lake

180890031

38.7 27.1 36.2

38.7

32.9

34.0

35.2

Gary

Indy-West Street

Marion

180970043

38.0

Indy-West Street

Indy-English Avenue

Marion

180970066

38.0

Indy-English Avenue

Indy-Washington Park

Marion

180970078

39.3 31.0 42.5 31.7 37.6

37.6

35.1

37.3

36.6

Indy-Washington Park

Indy-W 18th Street

Marion

180970081

36.2 31.9 45.7 34.8 38.4

37.9

37.5

39.6

38.3

Indy-W 18th Street

Indy- Michigan Street

Marion

180970083

36.7 31.3 40.3 33.5 37.2

36.1

35.0

37.0

36.0

Indy- Michigan Street

Luna Pier

Monroe

261150005

34.7 35.0 49.3 32.6 32.2

39.7

39.0

38.0

38.9

Luna Pier

Oak Park

Oakland

261250001

36.6 32.5 52.2 33.0 35.3

40.4

39.2

40.2

39.9

Oak Park

Port Huron

St. Clair

261470005

37.2 32.2 47.6 37.9 36.3

39.0

39.2

40.6

39.6

Port Huron

Ypsilanti

Washtenaw 261610008

38.8 31.5 52.1 31.3 34.5

40.8

38.3

39.3

39.5

Ypsilanti

Allen Park

Wayne

261630001

40.5 36.9 43.0 34.1 35.9

40.1

38.0

37.7

38.6

Allen Park

Southwest HS

Wayne

261630015

33.6 36.0 49.7 36.2 34.0

39.8

40.6

40.0

40.1

Southwest HS

Linwood

Wayne

261630016

46.2 38.3 51.8 36.9 34.8

45.4

42.3

41.2

43.0

Linwood

E 7 Mile

Wayne

261630019

37.1 35.0 52.3 36.2 33.0

41.5

41.2

40.5

41.0

E 7 Mile

Dearborn

Wayne

261630033

42.8 39.4 50.2 43.1 36.6

44.1

44.2

43.3

43.9

Dearborn

Wyandotte

Wayne

261630036

34.8 32.3 46.7 33.2 28.6

37.9

37.4

36.2

37.2

Wyandotte

Newberry

Wayne

261630038

36.8 57.5 28.6 33.4

39.1

39.8

42.7

Newberry

FIA

Wayne

261630039

43.9 32.4 34.8

37.0

39.7

FIA

Middleton

Butler

390170003

38.6 37.2 47.6 30.2 37.1

41.1

38.3

39.3

Middleton

Fairfield

Butler

390170016

34.8 32.2 43.4 35.2 34.5

36.8

36.9

37.7

37.1

Fairfield

Butler

390170017

34.6 34.3 44.9

37.9

39.6

40.8

Cleveland-28th Street

Cuyahoga

390350027

41.3 40.9 35.7 31.5 39.0

39.3

36.0

35.4

36.9

Cleveland-28th Street

Cleveland-St. Tikhon

Cuyahoga

390350038

47.3 42.5 51.2 36.1 39.7

44.9

47.0

42.3

44.2

Cleveland-St. Tikhon

Cleveland-Broadway

Cuyahoga

390350045

42.2 36.1 46.2 29.5 37.0

41.5

37.3

37.6

38.8

Cleveland-Broadway

Cleveland-GT Craig

Cuyahoga

390350060

45.5 42.2 49.5 31.0 38.7

45.7

40.9

39.7

42.1

Cleveland-GT Craig

Newburg Hts - Harvard Ave

Cuyahoga

390350065

39.1 36.1 47.9 27.8 39.1

41.0

37.3

38.3

38.9

Newburg Hts - Harvard Ave

Columbus - Fairgrounds

Franklin

390490024

39.2 35.1 45.0 34.0 34.2

39.8

38.0

37.7

38.5

Columbus - Fairgrounds

Columbus - Ann Street

Franklin

390490025

37.0 35.5 44.9 34.0 35.5

39.1

38.1

38.5

Columbus - Ann Street

Cincinnait

Hamilton

390610006

45.0 33.3 34.7

37.7

40.6

Cincinnait

Cincinnati - Seymour

Hamilton

390610014

37.8 42.0 38.5 35.2 38.1

39.4

38.6

37.3

38.4

Cincinnati - Seymour

Cincinnati - Taft Ave

Hamilton

390610040

31.9 30.5 45.8 32.8 34.7

36.1

36.4

37.8

36.7

Cincinnati - Taft Ave

Cincinnati - 8th Ave

Hamilton

390610042

33.8 31.9 44.4 34.5 35.9

36.7

36.9

38.3

37.3

Cincinnati - 8th Ave

Sharonville

Hamilton

390610043

37.3 31.4 39.9 34.9 34.0

36.2

35.4

36.3

36.0

Sharonville

Norwood

Hamilton

390617001

37.1 34.6 47.1 34.0 33.7

39.6

38.6

38.3

38.8

Norwood

St. Bernard

Hamilton

390618001

35.8 33.9 51.4 36.1 35.4

40.4

40.5

41.0

40.6

St. Bernard

Steubenville

Jefferson

390810016

39.6 43.8 43.8 32.1 43.5

42.4

39.9

39.8

40.7

Steubenville

Mingo Junction

Jefferson

390811001

40.9 51.5 44.2 32.9 35.4

45.5

42.9

37.5

42.0

Mingo Junction

Dayton

Montgomery 391130032

42.7 32.5 45.0 30.3 36.9

40.1

35.9

37.4

37.8

Dayton

Canton - Dueber

Stark

391510017

34.2 36.3 47.6 32.2 33.4

39.4

38.7

37.7

38.6

Canton - Dueber

Akron - Brittain

Summit

391530017

36.9 36.9 45.2 31.5 33.3

39.7

37.9

36.7

38.1

Akron - Brittain

Green Bay - Est High

Brown

550090005

33.5 32.3 41.5 36.9 37.1

35.8

36.9

38.5

37.1

Green Bay - Est High

Madison

Dane

550250047

32.0 31.9 40.1 33.4 44.3

34.7

35.1

39.3

36.4

Madison

Milwaukee-Health Center

Milwaukee

550790010

33.2 38.4 38.7 40.7 40.6

36.8

39.3

40.0

38.7

Milwaukee-Health Center

Milwaukee-SER Hdqs

Milwaukee

550790026

29.6 28.7 41.5 42.6 39.8

33.3

37.6

41.3

37.4

Milwaukee-SER Hdqs

Milwaukee-Virginia FS

Milwaukee

550790043

39.2 41.4 37.1 44.0

39.2

40.8

39.7

39.9

Milwaukee-Virginia FS

Milwaukee- Fire Dept Hdqs

Milwaukee

550790099

33.7 38.9 37.1 38.3 40.7

36.6

38.1

38.7

37.8

Milwaukee- Fire Dept Hdqs

Waukesha

551330027

29.1 38.4 41.1 28.2 33.8

36.2

35.9

34.4

35.5

Waukesha

98th Percentile (24-hour)

Design Values

Round 5 Modeling Results

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Site ID

State

County

Season

Species

Species Comp. of Ave.

FRM (fraction)

Species RRF

1703100521 IL

Cook

winter

so4

0.1772

0.9342

1703100521 IL

Cook

winter

no3

0.3099

1.0128

1703100521 IL

Cook

winter

ocm

0.2147

0.9942

1703100521 IL

Cook

winter

0.0372

0.888

1703100521 IL

Cook

winter

soil

0.0242

1.1674

1703100521 IL

Cook

winter

nh4

0.1421

0.97

1703100521 IL

Cook

winter

pbw

0.0947

0.9678

1703100521 IL

Cook

spring

so4

0.32

0.8018

1703100521 IL

Cook

spring

no3

0.0609

0.9385

1703100521 IL

Cook

spring

ocm

0.2742

1.0629

1703100521 IL

Cook

spring

0.0501

0.8712

1703100521 IL

Cook

spring

soil

0.0505

1.1796

1703100521 IL

Cook

spring

nh4

0.1203

0.8619

1703100521 IL

Cook

spring

pbw

0.0984

0.8492

1703100521 IL

Cook

summer

so4

0.3089

0.725

1703100521 IL

Cook

summer

no3

1.0124

1703100521 IL

Cook

summer

ocm

0.1599

1.069

1703100521 IL

Cook

summer

0.0351

0.8683

1703100521 IL

Cook

summer

soil

0.0318

1.204

1703100521 IL

Cook

summer

nh4

0.0932

0.7354

1703100521 IL

Cook

summer

pbw

0.094

0.7217

1703100521 IL

Cook

fall

so4

0.1872

0.9151

1703100521 IL

Cook

fall

no3

0.1628

0.9408

1703100521 IL

Cook

fall

ocm

0.2389

1.0091

1703100521 IL

Cook

fall

0.0403

0.8623

1703100521 IL

Cook

fall

soil

0.0284

1.1443

1703100521 IL

Cook

fall

nh4

0.1062

0.9247

1703100521 IL

Cook

fall

pbw

0.0614

0.9233

1711910071 IL

Madison

winter

so4

0.213

0.9195

1711910071 IL

Madison

winter

no3

0.2705

1.0306

1711910071 IL

Madison

winter

ocm

0.2093

0.9289

1711910071 IL

Madison

winter

0.0434

0.9083

1711910071 IL

Madison

winter

soil

0.0306

1.1782

1711910071 IL

Madison

winter

nh4

0.1528

0.9513

1711910071 IL

Madison

winter

pbw

0.0804

0.9243

1711910071 IL

Madison

spring

so4

0.3194

0.7717

1711910071 IL

Madison

spring

no3

0.0189

0.8611

1711910071 IL

Madison

spring

ocm

0.2455

1.1103

1711910071 IL

Madison

spring

0.0564

1.0046

1711910071 IL

Madison

spring

soil

0.0459

1.2252

1711910071 IL

Madison

spring

nh4

0.1121

0.7894

1711910071 IL

Madison

spring

pbw

0.1085

0.7783

1711910071 IL

Madison

summer

so4

0.313

0.705

1711910071 IL

Madison

summer

no3

0.884

1711910071 IL

Madison

summer

ocm

0.153

1.1546

1711910071 IL

Madison

summer

0.0345

1.0513

1711910071 IL

Madison

summer

soil

0.0302

1.2532

1711910071 IL

Madison

summer

nh4

0.102

0.7409

1711910071 IL

Madison

summer

pbw

0.1096

0.7133

1711910071 IL

Madison

fall

so4

0.2058

0.9037

1711910071 IL

Madison

fall

no3

0.1308

0.9426

1711910071 IL

Madison

fall

ocm

0.259

1.0233

1711910071 IL

Madison

fall

0.0563

0.9248

1711910071 IL

Madison

fall

soil

0.0549

1.1412

1711910071 IL

Madison

fall

nh4

0.1073

0.9185

1711910071 IL

Madison

fall

pbw

0.0655

0.918

PM2.5 RRFs by Species and Season (2009)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Site ID

State

County

Season

Species

Species Comp. of Ave.

FRM (fraction)

Species RRF

1803720011 IN

Dubois

winter

so4

0.2669

0.8833

1803720011 IN

Dubois

winter

no3

0.2548

0.9526

1803720011 IN

Dubois

winter

ocm

0.1747

0.9374

1803720011 IN

Dubois

winter

0.0313

0.9319

1803720011 IN

Dubois

winter

soil

0.0192

1.1349

1803720011 IN

Dubois

winter

nh4

0.1646

0.9069

1803720011 IN

Dubois

winter

pbw

0.0885

0.9006

1803720011 IN

Dubois

spring

so4

0.4141

0.6808

1803720011 IN

Dubois

spring

no3

0.0022

0.8106

1803720011 IN

Dubois

spring

ocm

0.178

0.9997

1803720011 IN

Dubois

spring

0.0324

0.9083

1803720011 IN

Dubois

spring

soil

0.0218

1.1284

1803720011 IN

Dubois

spring

nh4

0.1432

0.7075

1803720011 IN

Dubois

spring

pbw

0.1556

0.6916

1803720011 IN

Dubois

summer

so4

0.3687

0.644

1803720011 IN

Dubois

summer

no3

0.8029

1803720011 IN

Dubois

summer

ocm

0.1174

1.0136

1803720011 IN

Dubois

summer

0.0207

0.913

1803720011 IN

Dubois

summer

soil

0.0213

1.1988

1803720011 IN

Dubois

summer

nh4

0.1168

0.6789

1803720011 IN

Dubois

summer

pbw

0.1246

0.6613

1803720011 IN

Dubois

fall

so4

0.2964

0.8232

1803720011 IN

Dubois

fall

no3

0.138

0.8797

1803720011 IN

Dubois

fall

ocm

0.2116

0.9861

1803720011 IN

Dubois

fall

0.0437

0.9019

1803720011 IN

Dubois

fall

soil

0.03

1.1387

1803720011 IN

Dubois

fall

nh4

0.1449

0.8444

1803720011 IN

Dubois

fall

pbw

0.0941

0.8558

1809700811 IN

Marion

winter

so4

0.2358

0.9192

1809700811 IN

Marion

winter

no3

0.2729

0.9769

1809700811 IN

Marion

winter

ocm

0.1851

0.9546

1809700811 IN

Marion

winter

0.0385

0.8647

1809700811 IN

Marion

winter

soil

0.0239

1.0835

1809700811 IN

Marion

winter

nh4

0.1561

0.9446

1809700811 IN

Marion

winter

pbw

0.0877

0.944

1809700811 IN

Marion

spring

so4

0.3745

0.6868

1809700811 IN

Marion

spring

no3

0.0167

0.8082

1809700811 IN

Marion

spring

ocm

0.2034

0.9881

1809700811 IN

Marion

spring

0.0447

0.8547

1809700811 IN

Marion

spring

soil

0.0376

1.0625

1809700811 IN

Marion

spring

nh4

0.1313

0.7182

1809700811 IN

Marion

spring

pbw

0.1309

0.7056

1809700811 IN

Marion

summer

so4

0.3582

0.6529

1809700811 IN

Marion

summer

no3

0.8099

1809700811 IN

Marion

summer

ocm

0.1231

1.0043

1809700811 IN

Marion

summer

0.03

0.8444

1809700811 IN

Marion

summer

soil

0.0253

1.0918

1809700811 IN

Marion

summer

nh4

0.1114

0.6854

1809700811 IN

Marion

summer

pbw

0.1163

0.6674

1809700811 IN

Marion

fall

so4

0.2751

0.8538

1809700811 IN

Marion

fall

no3

0.149

0.9452

1809700811 IN

Marion

fall

ocm

0.223

0.9648

1809700811 IN

Marion

fall

0.0525

0.8412

1809700811 IN

Marion

fall

soil

0.0358

1.089

1809700811 IN

Marion

fall

nh4

0.1378

0.8905

1809700811 IN

Marion

fall

pbw

0.0865

0.8888

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Site ID

State

County

Season

Species

Species Comp. of Ave.

FRM (fraction)

Species RRF

2616300331 MI

Wayne

winter

so4

0.1587

0.9206

2616300331 MI

Wayne

winter

no3

0.2394

0.9813

2616300331 MI

Wayne

winter

ocm

0.3193

1.0781

2616300331 MI

Wayne

winter

0.0383

0.9279

2616300331 MI

Wayne

winter

soil

0.0541

1.0206

2616300331 MI

Wayne

winter

nh4

0.1188

0.9518

2616300331 MI

Wayne

winter

pbw

0.0714

0.9566

2616300331 MI

Wayne

spring

so4

0.3383

0.7398

2616300331 MI

Wayne

spring

no3

0.0259

0.8787

2616300331 MI

Wayne

spring

ocm

0.3543

1.0234

2616300331 MI

Wayne

spring

0.0504

0.8671

2616300331 MI

Wayne

spring

soil

0.0915

1.0153

2616300331 MI

Wayne

spring

nh4

0.1191

0.7818

2616300331 MI

Wayne

spring

pbw

0.1126

0.7619

2616300331 MI

Wayne

summer

so4

0.3311

0.6681

2616300331 MI

Wayne

summer

no3

0.8431

2616300331 MI

Wayne

summer

ocm

0.2297

1.0029

2616300331 MI

Wayne

summer

0.0362

0.8332

2616300331 MI

Wayne

summer

soil

0.061

1.0177

2616300331 MI

Wayne

summer

nh4

0.1027

0.6974

2616300331 MI

Wayne

summer

pbw

0.1073

0.6754

2616300331 MI

Wayne

fall

so4

0.1898

0.854

2616300331 MI

Wayne

fall

no3

0.1075

0.9367

2616300331 MI

Wayne

fall

ocm

0.3689

1.0607

2616300331 MI

Wayne

fall

0.0546

0.8862

2616300331 MI

Wayne

fall

soil

0.1676

1.0317

2616300331 MI

Wayne

fall

nh4

0.0866

0.8919

2616300331 MI

Wayne

fall

pbw

0.0553

0.8821

3903500381 OH

Cuyahoga

winter

so4

0.2117

0.8993

3903500381 OH

Cuyahoga

winter

no3

0.2665

0.9856

3903500381 OH

Cuyahoga

winter

ocm

0.2048

0.9716

3903500381 OH

Cuyahoga

winter

0.0413

0.8903

3903500381 OH

Cuyahoga

winter

soil

0.0465

1.0959

3903500381 OH

Cuyahoga

winter

nh4

0.1459

0.9416

3903500381 OH

Cuyahoga

winter

pbw

0.0832

0.9541

3903500381 OH

Cuyahoga

spring

so4

0.3334

0.7145

3903500381 OH

Cuyahoga

spring

no3

0.0374

0.8393

3903500381 OH

Cuyahoga

spring

ocm

0.2068

1.0899

3903500381 OH

Cuyahoga

spring

0.052

0.9362

3903500381 OH

Cuyahoga

spring

soil

0.0697

1.0601

3903500381 OH

Cuyahoga

spring

nh4

0.1256

0.7666

3903500381 OH

Cuyahoga

spring

pbw

0.115

0.7761

3903500381 OH

Cuyahoga

summer

so4

0.3241

0.6303

3903500381 OH

Cuyahoga

summer

no3

0.89

3903500381 OH

Cuyahoga

summer

ocm

0.1306

1.0998

3903500381 OH

Cuyahoga

summer

0.0419

0.9354

3903500381 OH

Cuyahoga

summer

soil

0.0583

1.0906

3903500381 OH

Cuyahoga

summer

nh4

0.1074

0.7038

3903500381 OH

Cuyahoga

summer

pbw

0.1183

0.6674

3903500381 OH

Cuyahoga

fall

so4

0.2055

0.8193

3903500381 OH

Cuyahoga

fall

no3

0.1275

0.9189

3903500381 OH

Cuyahoga

fall

ocm

0.2234

1.0245

3903500381 OH

Cuyahoga

fall

0.0499

0.8913

3903500381 OH

Cuyahoga

fall

soil

0.0675

1.0927

3903500381 OH

Cuyahoga

fall

nh4

0.1034

0.8615

3903500381 OH

Cuyahoga

fall

pbw

0.0637

0.8564

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Site ID

State

County

Season

Species

Species Comp. of Ave.

FRM (fraction)

Species RRF

3904900241 OH

Franklin

winter

so4

0.2555

0.8622

3904900241 OH

Franklin

winter

no3

0.2373

1.0002

3904900241 OH

Franklin

winter

ocm

0.2082

0.974

3904900241 OH

Franklin

winter

0.0375

0.8537

3904900241 OH

Franklin

winter

soil

0.0259

1.0844

3904900241 OH

Franklin

winter

nh4

0.1495

0.9261

3904900241 OH

Franklin

winter

pbw

0.0861

0.9274

3904900241 OH

Franklin

spring

so4

0.3754

0.6615

3904900241 OH

Franklin

spring

no3

0.0176

0.8436

3904900241 OH

Franklin

spring

ocm

0.2069

1.062

3904900241 OH

Franklin

spring

0.0405

0.8678

3904900241 OH

Franklin

spring

soil

0.0371

1.0551

3904900241 OH

Franklin

spring

nh4

0.1296

0.7212

3904900241 OH

Franklin

spring

pbw

0.128

0.6992

3904900241 OH

Franklin

summer

so4

0.3703

0.622

3904900241 OH

Franklin

summer

no3

0.9056

3904900241 OH

Franklin

summer

ocm

0.1343

1.0654

3904900241 OH

Franklin

summer

0.0311

0.8565

3904900241 OH

Franklin

summer

soil

0.0267

1.0667

3904900241 OH

Franklin

summer

nh4

0.1142

0.7021

3904900241 OH

Franklin

summer

pbw

0.1186

0.6614

3904900241 OH

Franklin

fall

so4

0.2692

0.8119

3904900241 OH

Franklin

fall

no3

0.1186

0.9099

3904900241 OH

Franklin

fall

ocm

0.2489

1.019

3904900241 OH

Franklin

fall

0.0533

0.8371

3904900241 OH

Franklin

fall

soil

0.0423

1.0924

3904900241 OH

Franklin

fall

nh4

0.1217

0.8539

3904900241 OH

Franklin

fall

pbw

0.0821

0.8519

3906100141 OH

Hamilton

winter

so4

0.2685

0.8104

3906100141 OH

Hamilton

winter

no3

0.2378

1.0886

3906100141 OH

Hamilton

winter

ocm

0.19

0.961

3906100141 OH

Hamilton

winter

0.035

0.8969

3906100141 OH

Hamilton

winter

soil

0.0229

1.4146

3906100141 OH

Hamilton

winter

nh4

0.1583

0.9077

3906100141 OH

Hamilton

winter

pbw

0.0874

0.8687

3906100141 OH

Hamilton

spring

so4

0.3583

0.6331

3906100141 OH

Hamilton

spring

no3

0.0025

1.0155

3906100141 OH

Hamilton

spring

ocm

0.1986

1.0798

3906100141 OH

Hamilton

spring

0.0466

0.9228

3906100141 OH

Hamilton

spring

soil

0.0289

1.3785

3906100141 OH

Hamilton

spring

nh4

0.1215

0.6968

3906100141 OH

Hamilton

spring

pbw

0.128

0.6307

3906100141 OH

Hamilton

summer

so4

0.3722

0.577

3906100141 OH

Hamilton

summer

no3

1.0923

3906100141 OH

Hamilton

summer

ocm

0.121

1.082

3906100141 OH

Hamilton

summer

0.0309

0.9099

3906100141 OH

Hamilton

summer

soil

0.0199

1.537

3906100141 OH

Hamilton

summer

nh4

0.1178

0.6441

3906100141 OH

Hamilton

summer

pbw

0.1261

0.5734

3906100141 OH

Hamilton

fall

so4

0.2608

0.7754

3906100141 OH

Hamilton

fall

no3

0.1184

0.9857

3906100141 OH

Hamilton

fall

ocm

0.213

1.0235

3906100141 OH

Hamilton

fall

0.0512

0.8876

3906100141 OH

Hamilton

fall

soil

0.0328

1.4007

3906100141 OH

Hamilton

fall

nh4

0.1254

0.846

3906100141 OH

Hamilton

fall

pbw

0.0828

0.8172

Site ID

State

County

Season

Species

Species Comp. of Ave.

FRM (fraction)

Species RRF

3908110011 OH

Jefferson

winter

so4

0.2367

0.8217

3908110011 OH

Jefferson

winter

no3

0.1709

1.0522

3908110011 OH

Jefferson

winter

ocm

0.3288

0.8819

3908110011 OH

Jefferson

winter

0.0435

0.9091

3908110011 OH

Jefferson

winter

soil

0.0272

0.4368

3908110011 OH

Jefferson

winter

nh4

0.1199

0.8904

3908110011 OH

Jefferson

winter

pbw

0.073

0.8583

3908110011 OH

Jefferson

spring

so4

0.3508

0.6666

3908110011 OH

Jefferson

spring

no3

0.0154

0.9156

3908110011 OH

Jefferson

spring

ocm

0.3078

0.9995

3908110011 OH

Jefferson

spring

0.0395

0.9853

3908110011 OH

Jefferson

spring

soil

0.0407

0.4844

3908110011 OH

Jefferson

spring

nh4

0.114

0.7054

3908110011 OH

Jefferson

spring

pbw

0.1095

0.6713

3908110011 OH

Jefferson

summer

so4

0.3779

0.6156

3908110011 OH

Jefferson

summer

no3

1.0837

3908110011 OH

Jefferson

summer

ocm

0.2098

1.0145

3908110011 OH

Jefferson

summer

0.0308

0.9689

3908110011 OH

Jefferson

summer

soil

0.0323

0.3632

3908110011 OH

Jefferson

summer

nh4

0.1065

0.6428

3908110011 OH

Jefferson

summer

pbw

0.1007

0.625

3908110011 OH

Jefferson

fall

so4

0.2315

0.7694

3908110011 OH

Jefferson

fall

no3

0.0702

1.0302

3908110011 OH

Jefferson

fall

ocm

0.372

0.9312

3908110011 OH

Jefferson

fall

0.051

0.9086

3908110011 OH

Jefferson

fall

soil

0.0344

0.4555

3908110011 OH

Jefferson

fall

nh4

0.0859

0.8284

3908110011 OH

Jefferson

fall

pbw

0.0629

0.7951

3911300321 OH

Montgomer winter

so4

0.2613

0.8598

3911300321 OH

Montgomer winter

no3

0.2407

1.029

3911300321 OH

Montgomer winter

ocm

0.1954

0.9442

3911300321 OH

Montgomer winter

0.036

0.8746

3911300321 OH

Montgomer winter

soil

0.0259

1.1295

3911300321 OH

Montgomer winter

nh4

0.1531

0.9304

3911300321 OH

Montgomer winter

pbw

0.0876

0.9205

3911300321 OH

Montgomer spring

so4

0.3659

0.6606

3911300321 OH

Montgomer spring

no3

0.0163

0.8639

3911300321 OH

Montgomer spring

ocm

0.1895

1.0976

3911300321 OH

Montgomer spring

0.0442

0.9417

3911300321 OH

Montgomer spring

soil

0.0253

1.0873

3911300321 OH

Montgomer spring

nh4

0.1313

0.7149

3911300321 OH

Montgomer spring

pbw

0.1326

0.6839

3911300321 OH

Montgomer summer

so4

0.375

0.6234

3911300321 OH

Montgomer summer

no3

0.9474

3911300321 OH

Montgomer summer

ocm

0.128

1.1047

3911300321 OH

Montgomer summer

0.029

0.9496

3911300321 OH

Montgomer summer

soil

0.0205

1.1299

3911300321 OH

Montgomer summer

nh4

0.1114

0.6931

3911300321 OH

Montgomer summer

pbw

0.1114

0.6482

3911300321 OH

Montgomer fall

so4

0.3062

0.8033

3911300321 OH

Montgomer fall

no3

0.1012

0.9634

3911300321 OH

Montgomer fall

ocm

0.2221

1.0158

3911300321 OH

Montgomer fall

0.0514

0.877

3911300321 OH

Montgomer fall

soil

0.028

1.1391

3911300321 OH

Montgomer fall

nh4

0.1352

0.8625

3911300321 OH

Montgomer fall

pbw

0.0982

0.8475

Site ID

State

County

Season

Species

Species Comp. of Ave.

FRM (fraction)

Species RRF

3915100171 OH

Stark

winter

so4

0.2362

0.8558

3915100171 OH

Stark

winter

no3

0.2234

1.0222

3915100171 OH

Stark

winter

ocm

0.2478

0.9255

3915100171 OH

Stark

winter

0.0414

0.8866

3915100171 OH

Stark

winter

soil

0.0334

1.099

3915100171 OH

Stark

winter

nh4

0.1376

0.925

3915100171 OH

Stark

winter

pbw

0.0802

0.9155

3915100171 OH

Stark

spring

so4

0.3581

0.6834

3915100171 OH

Stark

spring

no3

0.0236

0.855

3915100171 OH

Stark

spring

ocm

0.221

1.0892

3915100171 OH

Stark

spring

0.0501

1.0017

3915100171 OH

Stark

spring

soil

0.058

1.0528

3915100171 OH

Stark

spring

nh4

0.1288

0.7264

3915100171 OH

Stark

spring

pbw

0.1256

0.7009

3915100171 OH

Stark

summer

so4

0.3621

0.6277

3915100171 OH

Stark

summer

no3

0.8203

3915100171 OH

Stark

summer

ocm

0.1483

1.0984

3915100171 OH

Stark

summer

0.0403

1.016

3915100171 OH

Stark

summer

soil

0.037

1.0781

3915100171 OH

Stark

summer

nh4

0.1157

0.6739

3915100171 OH

Stark

summer

pbw

0.124

0.651

3915100171 OH

Stark

fall

so4

0.2293

0.8041

3915100171 OH

Stark

fall

no3

0.1262

0.9363

3915100171 OH

Stark

fall

ocm

0.2722

1.0226

3915100171 OH

Stark

fall

0.0545

0.9202

3915100171 OH

Stark

fall

soil

0.0461

1.0959

3915100171 OH

Stark

fall

nh4

0.1105

0.8549

3915100171 OH

Stark

fall

pbw

0.0706

0.8428

3915300171 OH

Summit

winter

so4

0.2511

0.8771

3915300171 OH

Summit

winter

no3

0.2376

1.0052

3915300171 OH

Summit

winter

ocm

0.2185

0.9429

3915300171 OH

Summit

winter

0.0334

0.8677

3915300171 OH

Summit

winter

soil

0.0255

1.0835

3915300171 OH

Summit

winter

nh4

0.1489

0.9374

3915300171 OH

Summit

winter

pbw

0.0851

0.945

3915300171 OH

Summit

spring

so4

0.387

0.7046

3915300171 OH

Summit

spring

no3

0.0072

0.8466

3915300171 OH

Summit

spring

ocm

0.1901

1.0967

3915300171 OH

Summit

spring

0.035

0.9482

3915300171 OH

Summit

spring

soil

0.0304

1.0524

3915300171 OH

Summit

spring

nh4

0.1294

0.7521

3915300171 OH

Summit

spring

pbw

0.1342

0.7384

3915300171 OH

Summit

summer

so4

0.3694

0.6378

3915300171 OH

Summit

summer

no3

0.8587

3915300171 OH

Summit

summer

ocm

0.1417

1.1077

3915300171 OH

Summit

summer

0.0332

0.9506

3915300171 OH

Summit

summer

soil

0.0198

1.0744

3915300171 OH

Summit

summer

nh4

0.1121

0.6961

3915300171 OH

Summit

summer

pbw

0.1146

0.6691

3915300171 OH

Summit

fall

so4

0.2443

0.8074

3915300171 OH

Summit

fall

no3

0.1175

0.9392

3915300171 OH

Summit

fall

ocm

0.2636

1.0252

3915300171 OH

Summit

fall

0.0623

0.8883

3915300171 OH

Summit

fall

soil

0.0494

1.086

3915300171 OH

Summit

fall

nh4

0.109

0.8622

3915300171 OH

Summit

fall

pbw

0.0723

0.8506

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

119

APPENDIX II

Ozone Source Apportionment Modeling Results

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

120

APPENDIX III

2.5

Source Apportionment Modeling Results

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Chicago (Cicero), Illinois

2005 (Round 5)

2012 (Round 4)

2018 (Round 5)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Clark County, Indiana

2005 (Round 5)

2012 (Round 4)

2018 (Round 5)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Dearborn, Michigan

2005 (Round 5)

2012 (Round 4)

2018 (Round 5)

Cincinnati, Ohio

2005 (Round 5)

2012 (Round 4)

2018 (Round 5)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Cleveland, Ohio

2005 (Round 5)

2012 (Round 4)

2018 (Round 5)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Steubenville, Ohio

2005 (Round 5)

2012 (Round 4)

2018 (Round 5)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

121

APPENDIX IV

Haze Source Apportionment Modeling Results

Boundary Waters, Minnesota

2005 (Round 5)

2018 (Round 4)

2018 (Round 5)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Voyageurs, Minnesota

2005 (Round 5)

2018 (Round 4)

2018 (Round 5)

Seney, Michigan

2005 (Round 5)

2018 (Round 4)

2018 (Round 5)

Isle Royale, Michigan

2005 (Round 5)

2018 (Round 4)

2018 (Round 5)

Shenandoah, Virginia

2005 (Round 5)

2018 (Round 4)

2018 (Round 5)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Mammoth Cave, Kentucky

2005 (Round 5)

2018 (Round 4)

2018 (Round 5)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A

Lye Brook, Vermont

2005 (Round 5)

2018 (Round 4)

2018 (Round 5)

Electronic Filing - Received, Clerk's Office, January 21, 2009

Appendix A