vvEPA
         Office of Transportation                   EPA420-D-06-004
United States    and Air Quality                      February 2006
Environmental Protection	
Agency
           Draft Regulatory Impact
           Analysis: Control of
           Hazardous Air Pollutants from
           Mobile Sources

-------
                                                                   EPA420-D-06-004
                                                                       February 2006
                     of                   Air
                         Assessment and Standards Division
                       Office of Transportation and Air Quality
                       U.S. Environmental Protection Agency
                                     NOTICE
   This Technical Report does not necessarily represent final EPA decisions or positions.
It is intended, to present technical analysis of issues using data that are currently available.
         The purpose in the release of such reports is to facilitate an exchange of
        technical information and to inform the public of technical developments.

-------
                                Table of Contents
Executive Summary 	ES-1



Chapter 1: Mobile Source Air Toxics Health Information	1-1

Chapter 2: Emission Inventories	2-1

Chapter 3: Air Quality and Resulting Health and Welfare Effects of Air Pollution from Mobile
         Sources	3-1

Chapter 4: Industry Characterization	4-1

Chapter 5: Vehicle Technological Feasibility	5-1

Chapter 6: Feasibility of Complying with a Benzene and Other Control Standards	6-1

Chapter 7: Gas Can Feasibility and Test Procedures	7-1

Chapter 8: Impact of New Requirements on Vehicle Costs	8-1

Chapter 9: Cost of Proposed Gasoline Benzene Standard and Other Control Options
         Considered	9-1

Chapter 10: Gas Can Costs	10-1

Chapter 11: Cost per Ton of Emissions Reduced	11-1

Chapter 12: Cost-Benefit Analysis	12-1

Chapter 13: Economic Impact Analysis	13-1

Chapter 14: Small-Business Flexibility Analysis	14-1
                                          in

-------
List of Acronyms
ug/kg-day          Micrograms per kilogram day
ug/m3              Microgram per cubic meter
AAM              Alliance of Automobile Manufacturers
ABT               Averaging, Banking, and Trading
ACS               American Cancer Society
AEO               Annual Energy Outlook (an EIA publication)
AGO               Atmospheric Gas Oil (a refinery stream)
ALTP              Absolute Level Trigger Point
AMS/EPA          Regulatory Model AERMOD
APHEA            Air Pollution and Health: A European Approach
API                American Petroleum Institute
AQMTSD          Air Quality Modeling Technical Support Document
ARB               (California) Air Resources Board
ASPEN            Assessment System for Population Exposure Nationwide
ASTM             American Society of Testing Materials
ATB               Atmospheric Tower Bottoms (a refinery stream)
ATSDR            Agency for Toxic Substances and Disease Registry
ATV               All-Terrain Vehicles
BBL               Barrel
BC                Black Carbon
BenMAP           Environmental Benefits Mapping and Analysis Program
BPCD              Barrels Per Calendar Day
BPD               Barrels Per Day
BTEX              Benzene, Toluene, Ethylbenzene, and Xylene isomers
BTX               Benzene, Toluene, and Xylene  isomers
BZ                Benzene
C                  Celsius
C6 or C6            A hydrocarbon molecule with a specified number of carbon atoms, in this case 6 carbons
CA                California
CAA               Clean Air Act
CAIR              Clean Air Interstate Rule
California EPA      California Environmental Protection Agency
CAMR             Clean Air Mercury Rule
CAMx             Comprehensive Air Quality Model with Extension
CAND             Clean Air Nonroad Diesel
CARD              California Air Resources Board
CASAC            Clean Air Science Advisory Committee
CAVR             Clean Air Visibility Rule
CB                Chronic Bronchitis
CD                Criteria Document
CDC               Center for Disease Control
CE10              Gasoline with 10 percent ethanol content
CEA               Cost-Effectiveness Analysis
CFEIS              Certification and Fuel Economy Information System
CFR               Code of Federal Regulations
CG                Conventional Gasoline
CHAD             Consolidated Human Activity Database
CI                 Compression Ignition
CUT               Chemical Industry Institute of Toxicology
CIMT              Carotid Intima-Media Thickness
CM 15              Gasoline with 15 percent methanol content
CMAI              Chemical Market Associates Incorporated
CMAQ             Community Multiscale Air Quality Model
                                                IV

-------
CNG               Compressed Natural Gas
CNS               Central Nervous System
CO                Carbon Monoxide
CO2               Carbon Dioxide
COI               Cost-of-Illness
COPD              Chronic Obstructive Pulmonary Disease
C-R               Concentration-Response
CRC E-55/E-59      Coordinating Research Council Emission Test Program for Heavy Duty Trucks
CUA               Cost-Utility Analysis
DCS               Distributed Control System
DEOG             Diesel Exhaust Organic Gases
DOE               Department of Energy
DPM               Diesel Paniculate
E&C               Engineering and Construction
E10                Gasoline Blend with Nominal 10 volume percent Ethanol
E200               Percent of Fuel Evaporated at 200 Degrees F (ASTM D 86)
E300               Percent of Fuel Evaporated at 300 Degrees F (ASTM D 86)
EAC               Early Action Compact
EC/OC             Elemental/Organic Carbon
ECM               Engine Control Module
EGU               Electrical Generating Utility
EIA               Economic Impact Analysis
EIA               Energy Information Administration (part of the U.S. Department of Energy)
EIM               Economic Impact Model
EMS-HAP          Emissions Modeling System for Hazardous Air Pollutants
EO                Executive Order
EPA               Environmental Protection Agency
EPAct              Energy Policy Act of 2005
ETBE              Ethyl Tertiary Butyl Ether
ETC               Electronic Throttle Control
ETS               Environmental Tobacco Smoke
EVOH             Ethylene vinyl alcohol
F                  Fahrenheit
FACES             Fresno Asthmatic Children's Environment Study
FBP               Feed Boiling Point (also Final Boiling Point)
FCC               Fluidized Catalytic Cracker
FCCU              Fluidized Catalytic Cracking Unit
PEL               Family Emission Level
FHWA             Federal Highway Administration
FOEB              Fuel Oil Equivalent Barrel
FRM               Final Rulemaking
FRTP              Fixed Reduction Trigger Point
FTC               Federal Trade Commission
FTP               Federal Test Procedure
g/gal/day           Grams per gallon per day
GDP               Gross Domestic Product
GIS                Geographic Information System
GM                General Motors
GPA               Geographic Phase-in Area
GVW              Gross Vehicle Weight
GVWR             Gross Vehicle Weight Rating
H2                 Hydrogen gas
HAD               Diesel Health Assessment Document
HAP               Hazardous Air Pollutant
HAPEM5           Hazardous Air Pollutant Exposure Model version 5
HC                Hydrocarbon

-------
HCO               Heavy Cycle Oil (a refinery stream)
HDN               Naphtha Hydrotreater (also Hydro-Denitrogenation Unit)
HDPE              High density polyethylene
HDS               Hydro-Desulfurization Unit
HOT               Hydrotreater
HEGO             (Heated) Exhaust Gas Oxygen
HHDDT            Heavy Heavy-Duty Diesel Truck
HI                 Hazard Index
HLDT              Heavy Light-Duty Truck
HQ                Hazard Quotient
HSR               Heavy Straight Run (a refinery stream)
HVGO             Heavy Vacuum Gas Oil (a refinery stream)
IBP                Initial Boiling Point
ICAO              International Civil Aviation Organization
ICD-9              International Classification of Diseases - Ninth Revision
ICI                Independent Commercial Importer
IFF                Institute Francais du Petrole
IMO               International Maritime Organization
IMPROVE          Interagency Monitoring of Protected Visual Environments
IRFA              Initial Regulatory Flexibility Analysis
IRIS               Integrated Risk Information System
ISC                Integrated Source Complex
ISCST              Industrial Source Complex Short Term
JAMA              Journal of the American Medical Association
K                  Thousand
KBBL              Thousand barrels
KWH              Kilowatt Hour
LB                Pound
LCO               Light Cycle Oil (a refinery stream)
LDGT              Light Duty Gasoline Truck
LDGV             Light Duty Gasoline Vehicle
LOT               Light-Duty Truck
LDV               Light-Duty Vehicle
LEV I              Low Emission Vehicle I
LEV II             Low Emission Vehicle II
LEV               Low Emission Vehicle
LLDT              Light Light-Duty Truck
LLE               Liquid-Liquid Extraction
LNS               Light Naphtha Splitter
LP                Linear Programming (a type of refinery model)
LPG               Liquefied Petroleum Gas
LRS               Lower Respiratory Symptom
LSR               Light Straight Run (a refinery stream)
MAP               (Engine) Mass Air Flow
MAP               (Engine) Manifold Absolute Pressure
MDPV             Medium-Duty Passenger Vehicle
mg/m3              Milligrams per cubic meter
MHDDT           Diesel-Fueled Medium Heavy-Duty Truck
MHDT             Medium Heavy-Duty Truck
MI                Myocardial Infarction
MILY              Morbidity Inclusive Life Years
MLE               Maximum Likelihood Estimate
MM               Million
MNCPES           Minnesota Children's Pesticide Exposure Study
MOBILE6.2        EPA's Highway Vehicle Emission Model
MON              Motor Octane Number
                                                VI

-------
MRAD
MRL
MSAT
MS ATI
MTBE
NAAQS
NAICS
NAS
NATA
NATTS
NCI
NCLAN
NEI
NEMS
NESCAUM
NESHAP
NHAPS
NHEXAS
NIOSH
NLEV
NMHC
NMIM
NMIM2005
NMMAPS
NMOG
NO2
NONROAD
NONROAD2005
NOX
NPRA
NPRM
NRC
NSTC
O&M
OEHHA
OGJ
OMB
OP
OSHA
OTAQ
PADD
PAH
PC
PC
PFC
PM
PM/NMHC
PM10
PM25
POM
PONA
PPM
PRTP
PSI
PSR
QALY
Minor Restricted Activity Days
Minimum Risk Level
Mobile Source Air Toxics
2001 Mobile Source Air Toxics Rule
Methyl Tertiary-Butyl Ether
National Ambient Air Quality Standards
North American Industrial Classification System
National Academy of Science
National Scale Air Toxics Assessment
National Air Toxics Trends Sites
National Cancer Institute
National Crop Loss Assessment Network
National Emissions Inventory
National Energy Modeling System
Northeast States for Coordinated Air Use Management
National Emissions Standards for Hazardous Air Pollutants
National Human Activity Pattern Survey
National Human EXposure Assessment Survey
National Institute for Occupational Safety and Health
National Low Emission Vehicle
Non-Methane Hydrocarbons
National Mobile Inventory Model (EPA software tool)
National Mobile Inventory Model Released in 2005
National Mortality, Morbidity and Air Pollution
Non-Methane Organic Gases
Nitrogen Dioxide
EPA's Non-road Engine Emission Model
EPA's Non-road Engine Emission Model Released in 2005
Oxides of Nitrogen
National Petroleum Refiners Association
Notice of Proposed Rulemaking
National Research Council
National Science and Technology Council
Operating and maintenance
Office of Environmental Health Hazard Assessment
Oil and Gas Journal
Office of Management and Budget
Original Production
U.S. Department of Labor Occupational Safety and Health Organization
Office of Transportation and Air Quality
Petroleum Administration for Defense District
Polycyclic Aromatic Hydrocarbon
Particle Count
Passenger car
Portable Fuel Containers
Paniculate Matter
Paniculate Matter to Non-Methane Hydrocarbon Ratio
Coarse Particle
Fine Particle
Polycyclic Organic Matter
Paraffin, Olefin, Naphthene, Aromatic
Parts Per Million
Percentage Reduction Trigger Point
Pounds per Square Inch
Power Systems Research
Quality-Adjusted Life Year
                                                vn

-------
R&D               Research and Development
RAMS             Regional Atmospheric Modeling System
RBOB             Reformulated Blendstock for Oxygenate Blending
REMSAD          Regional Modeling System for Aerosols and Deposition
RfC                Reference Concentration
RfD                Oral reference dose
RFG               Reformulated Gasoline
RFS                Renewable Fuels Standard
RIA                Regulatory Impact Analysis
RIOPA             Relationship Between Indoor, Outdoor and Personal Air
ROI                Return on Investment
RON               Research Octane Number
RSM               Response Surface Metamodel
RVP               Reid Vapor Pressure
SAB               Science Advisory Board
SAB-HES          Science Advisory Board - Health Effects Subcommittee
S AE               Society of Automotive Engineers
SBA               Small Business Administration
SBAR Panel, or 'the Panel'        Small Business Advocacy Review Panel
SBREFA           Small Business Regulatory Enforcement Fairness Act of 1996
SCF                Standard Cubic Foot
SECA             SOX Emission Control Area
SER               Small Entity Representative
SFTP               Supplemental Federal Test Procedure
SHED             Sealed Housing for Evaporative Determination
SI                 Spark Ignition
SPECIATE         EPA's repository of Total Organic Compound (TOC) & Paniculate Matter (PM) speciated
                   profiles
SUV               Sports-Utility Vehicle
SVM               Small Vehicle Manufacturer
SVOC             Semi-Volatile Organic Compound
SwRI               Southwest Research Institute
TDM               Travel Demand Model
TEACH            Toxic Exposure Assessment - Columbia/Harvard
TEAM             Total Exposure Assessment Methodology
THC               Total Hydrocarbon
TMP               2,2,4-Trimethyrpentane
TSP                Total Suspended Paniculate Matter
TWC               Three-Way Catalyst
UC                Unified Cycle Emission Test Procedure from ARE
UCL               Upper Confidence Limit
ULSD             Ultra-Low Sulfur Diesel
URE               Unit Risk Estimate
URS               Upper Respiratory Symptom
UV                Ultraviolet
UVb               Ultraviolet-b
VGO               Vacuum Gas Oil (a refinery stream)
VMT               Vehicle Miles Traveled
VOC               Volatile Organic Compound
VSL               Value of a Statistical Life
VTB               Vacuum Tower Bottoms (a refinery stream)
WLD               Work Loss Days
WTP               Willingness-to-Pay
                                               Vlll

-------
                               Executive Summary

       EPA is proposing new standards to reduce emissions of Mobile Source Air Toxics
(MSATs) including benzene and overall hydrocarbons from motor vehicles, motor vehicle fuels,
and portable gasoline containers (gas cans).  This Regulatory Impact Analysis provides technical,
economic, and environmental  analyses of the proposed new emission standards.  The anticipated
emission reductions will significantly reduce exposure to harmful pollutants and also provide
assistance to states and regions facing ozone and particulate air quality problems that are causing
a range of adverse health effects, especially in terms of respiratory impairment and related
illnesses.

       Chapter 1  reviews information related to the health effects of mobile source air toxics.
Chapter 2 provides emissions  inventory estimates, including estimates of anticipated emissions
reductions.  Chapter 3 presents air quality and resulting health and welfare effects associated
with air toxics, ozone, and particulate matter (PM). Chapter 4 contains an overview of the
affected refiners and manufacturers, including a description of the range of products involved
and their place in  the market.  Chapters 5 through 7 summarize the available information
supporting the specific standards we are proposing, providing a technical justification for the
feasibility of the standards for vehicles, fuels, and gas cans, respectively. Chapters 8 throughlO
present cost estimates of complying with the proposed standards or vehicles, fuels, and gas cans,
respectively.  Chapter 11 compares the costs and the emission reductions to generate an estimate
of the cost per ton of pollutant removed.  Chapters 12 and 13 describe the estimated  societal
costs and benefits of the proposed rulemaking.  Chapter  14 presents our Regulatory Flexibility
Analysis, as called for in the Regulatory Flexibility Act.

       The following paragraphs briefly describe the standards that we are proposing and the
estimated impacts.

Emissions Standards

Vehicles

       We  are proposing new standards for both exhaust and evaporative emissions from
passenger vehicles. The new exhaust emissions standards would significantly reduce non-
methane hydrocarbon (NMHC) emissions from passenger vehicles at cold temperatures.  These
hydrocarbons include many mobile source air toxics (including benzene), as well as VOC.

       The current NMHC standards are typically tested at 75° F, and recent research and
analysis indicates that these standards are not resulting in robust control of NMHC at lower
temperatures. (There is an  existing cold temperature standard, but it applies only to CO.)  We
believe that cold temperature NMHC control can be substantially improved using the same
technological approaches that are generally already being used in the Tier 2 vehicle fleet to meet
the stringent standards at 75° F. These cold-temperature NMHC controls would also result in
lower direct PM emissions at cold temperatures.
                                          ES-1

-------
       Accordingly, we are proposing that light-duty vehicles, light-duty trucks, and medium-
duty passenger vehicles would be subject to a new non-methane hydrocarbon (NMHC) exhaust
emissions standard at 20° F.  Vehicles at or below 6,000 pounds gross vehicle weight rating
(GVWR) would be subject to a sales-weighted fleet average NMHC level of 0.3 grams/mile.
Vehicles between 6,000 and 8,500 pounds GVWR and medium-duty passenger vehicles would
be subject to a sales-weighted fleet average NMHC level of 0.5 grams/mile. For lighter vehicles,
the standard would phase in between 2010 and 2013.  For heavier vehicles, the new standards
would phase in between 2012 and 2015. We are also proposing a credit program and other
provisions designed to provide flexibility to manufacturers, especially during the phase-in
periods. These provisions are designed to allow the earliest possible phase-in of standards and
help minimize costs and ease the transition to new standards

       We are also proposing a set of nominally more stringent evaporative emission standards
for all light-duty vehicles, light-duty trucks, and medium-duty passenger vehicles.  The proposed
standards are equivalent to California's Low Emission Vehicle II (LEV II) standards, and they
reflect the evaporative emissions levels that are already being achieved nationwide. The
standards we propose today would codify the approach that manufacturers are already taking for
50-state evaporative systems, and thus the standards would prevent backsliding in the future.  We
are proposing to implement the evaporative emission standards in 2009 for lighter vehicles and
in 2010 for the heavier vehicles.

Gasoline Fuel Standards

       We are proposing that beginning January 1, 2011, refiners and fuel importers would meet
an average gasoline benzene content standard of 0.62% by volume on all their gasoline, both
reformulated  and conventional (except for California, which is already covered by a similar
relatively stringent state program).

       This proposed fuel standard would result in air toxics emissions reductions that are
greater than required under all existing gasoline toxics programs.  As a result, EPA is proposing
that upon full implementation in 2011, the regulatory provisions for the benzene control program
would become the single regulatory mechanism used to implement the RFG and Anti-dumping
annual average toxics requirements. The current RFG and Anti-dumping annual average
provisions would be replaced by the proposed benzene control program. The MSAT2 benzene
control program would also replace the MS ATI requirements.  In addition, the program would
satisfy certain fuel MS AT conditions of the Energy Policy Act of 2005. In all of these ways, we
would significantly consolidate and simplify the existing national fuel-related MSAT regulatory
program.

       We are also proposing that refiners could generate benzene credits and use  or transfer
them as a part of a nationwide averaging, banking, and trading (ABT) program. From 2007-
2010 refiners could generate benzene credits by taking early steps to reduce gasoline benzene
levels. Beginning in 2011 and continuing indefinitely, refiners  could generate credits by
producing gasoline with benzene levels below the 0.62% average standard. Refiners could apply
the credits towards company compliance, "bank" the credits for later use, or transfer ("trade")
                                          ES-2

-------
them to other refiners nationwide (outside of California) under the proposed program. Under
this program, refiners could use credits to achieve compliance with the benzene content standard.

Portable Gasoline Container (Gas Can) Controls

       Portable gasoline containers, or gas cans, are consumer products used to refuel a wide
variety of gasoline-powered equipment, including lawn and garden equipment, recreational
equipment, and passenger vehicles that have run out of gas. We are proposing standards that
would reduce hydrocarbon emissions from evaporation, permeation, and spillage. These
standards would significantly reduce benzene and other toxics, as well as VOC more generally.
VOC  is an ozone precursor.

       We propose a performance-based standard of 0.3 grams per gallon per day of
hydrocarbons, based on the emissions from the can over a diurnal test cycle.  The standard would
apply to gas cans manufactured on or after January  1, 2009. We also propose test procedures
and a certification and compliance program, in order to ensure that gas cans would meet the
emission standard over a range of in-use conditions. The proposed standards would result in the
best available control technologies, such as durable permeation barriers, automatically closing
spouts, and cans that are well-sealed.

       California implemented an emissions control program for gas cans in 2001, and since
then, several other states have adopted the program. Last year,  California adopted a revised
program, which will take effect July 1, 2007.  The revised California program is very similar to
the program we are proposing. Although a few aspects of the program we are proposing are
different, we believe manufacturers would be able to meet both EPA and California requirements
with the same gas can designs.

Projected Impacts

       The following paragraphs and tables summarize the projected emission reductions and
costs associated with the emission standards.  See the detailed analysis later in this document for
further discussion of these estimates.

Emissions Reductions

Toxics

       Air toxic emissions from light-duty vehicles depend on both fuel benzene content and
vehicle hydrocarbon emission controls. Similarly, the air toxic  emissions from gas cans depend
on both fuel benzene content and the gas can emission controls. Tables 1 and 2 below
summarize the expected reductions in benzene and total MSAT emissions, respectively, from our
proposed vehicle, fuel, and gas can controls.  Although the proposal does not apply to nonroad
engines or the gasoline distribution industry, the fuels controls would reduce benzene emissions
from these sources as well due to lower benzene levels in gasoline.  In 2030, annual benzene
emissions from gasoline on-road mobile sources would be 44% lower as a result of this proposal.
Annual  benzene emissions from gasoline light-duty vehicles would be 45% lower in 2030  as a
                                          ES-3

-------
result of this proposal. Gasoline would have 37% lower benzene overall. Finally, this proposal
would reduce annual emissions of benzene from gas cans by 78% in 2030.

Table 1:  Estimated Reductions in Benzene Emissions from Proposed Control Measures by
                             Sector, 2020 and 2030 (tons)


                                         2020                        2030

 Fuels                         18,145                       20,272

 Vehicles                     28,105                       47,689

 Gas Cans                     1,567                        1,772

 Total                        45,241                       65,282


 Table 2: Estimated Reductions in MSAT Emissions from Proposed Control Measures by
                             Sector, 2020 and 2030 (tons)

Fuels
Vehicles
Gas Cans
Total
2020
18,145
181,509
24,158
221,081

20,272
308,887
27,342
351,894
2030




voc

      VOC emissions would be reduced by the hydrocarbon emission standards for both light-
duty vehicles and gas cans.  Annual VOC emission reductions from these sources would be 35%
lower in 2030 because of this proposal.

Table 3: Estimated Reductions in VOC Emissions from Light-Duty Gasoline Vehicles and
                            Gas Cans, 2020 and 2030 (tons)

Vehicles
Gas Cans
Total
2020
536,484
192,683
729,167

913,439
218,080
1,131,519
2030



                                        ES-4

-------
PM2.5

       We expect that only the proposed vehicle control would reduce emissions of direct PM2.5.
As shown in Table 4, we expect this control to reduce direct PM2.s emissions by about 20,000
tons in 2030.  In addition, the VOC reductions from the proposed vehicle and gas can standards
would also reduce secondary formation of PM25

  Table 4. Estimated National Reductions in Direct PMi.s Exhaust Emissions from Light-
                Duty Gasoline Vehicles and Trucks, 2020 and 2030 (tons)

PM2.5 Reductions from Proposed
Vehicle Standards (tons)
2020
11,803
2030
20,096
Costs

Fuels

       The refinery model estimates that the proposed benzene standard would cost 0.13 cents
per gallon, averaged over the entire U.S. gasoline pool. (When averaged only over those
refineries which are assumed to take steps to reduce their benzene levels, the average cost would
be 0.19 cents per gallon.) This per-gallon cost would result from an industry-wide investment in
capital equipment of $500 million to reduce gasoline benzene levels. This would amount to an
average of $5 million in capital investment in each refinery that adds such equipment.  The
aggregate costs for the fuel program for 2020 and 2030 are provided in Table 5. The increase in
costs is due to the projected increase in gasoline usage.

 Table 5. Estimated Aggregate Annual Cost for the Proposed Benzene Standard, 2020 and
                                         2030

                                          2020                         2030

 Fuels program                  $212,606,000                   $248,421,000
Vehicles

       We project that the average incremental costs associated with the new cold temperature
standards would be less than $1 per vehicle. We are not projecting changes to vehicle hardware
as a result of the proposed standard. Costs would be associated with vehicle R&D and
recalibration as well as facilities upgrades to handle additional development testing under cold
conditions. Also, we are not anticipating additional costs for the proposed new evaporative
                                         ES-5

-------
emissions standard. We expect that manufacturers will continue to produce 50-state evaporative
systems that meet LEV II standards. Therefore, harmonizing with California's LEV-II
evaporative emission standards would streamline certification and be an "anti-backsliding"
measure. It also would codify the approach manufacturers have already indicated they are taking
for 50-state evaporative systems.

       We also estimated annual aggregate costs associated with the new cold temperature
emissions standards. These costs are projected to increase with the phase-in of standards and
peak in 2014 at about $13.4 million per year, then decrease as the fixed costs are fully amortized.
As shown in Table 6, the costs would be fully amortized by 2020.

 Table 6. Estimated Aggregate Annual Cost for the Proposed Vehicle Standards, 2020 and
                                         2030

                                         2020                         2030

 Vehicles program              $0                           $0
Gas Cans

       Table 7 summarizes the projected near-term and long-term per unit average costs to meet
the new emission standards.  Long-term impacts on gas cans are expected to decrease as
manufacturers fully amortize their fixed costs. The table also shows our projections of average
fuel savings over the life of the gas can.

          Table 7 Estimated Average Gas Can Costs and Lifetime Fuel Savings
                                                       Cost

               Near-Term Costs              $2.69

               Long-Term  Costs              $1.52

               Fuel Savings (NPV)            $4.24

       We have also estimated aggregate costs and fuel savings which are projected to peak in
2013 at about $51 million and then drop to about $29 million once fixed costs are recovered.
The aggregate annual costs and fuel savings estimates for 2020  and 2030 are provided in Table
  Table 8.  Estimated Aggregate Annual Cost and Fuel Savings for the Proposed Gas Can
                               Standards, 2020 and 2030

                                          2020                         2030
                                         ES-6

-------
 Gas Can Costs                 $31,767,000                    $38,724,000

 Gas Can Fuel Savings           $98,861,000                    $111,210,000


Cost Per Ton

       We have calculated the cost per ton of HC, benzene, total MSATs, and PM emissions
reductions associated with the proposed fuel, vehicle, and gas can programs. We have calculated
the costs per ton using the net present value of the annualized costs of the program, including gas
can fuel savings, from 2009 through 2030 and the net present value of the annual emission
reductions through 2030. We have also calculated the cost per ton of emissions reduced in the
year 2020 and 2030 using the annual costs and emissions reductions in that year alone. This
number represents the long-term cost per ton of emissions reduced. For fuels, the cost per ton
estimates include  costs and emission reductions that will occur from all motor vehicles and
nonroad engines fueled with gasoline as well as gas cans and gasoline distribution.

       We have not attempted to apportion costs across these various pollutants for purposes of
the cost per ton calculations since there is no distinction in the technologies, or associated costs,
used to control the pollutants. Instead, we have calculated costs per ton by  assigning all costs to
each individual pollutant. If we apportioned costs among the pollutants, the costs per ton
presented here would be proportionally lowered depending on what portion of costs were
assigned to the various pollutants. The results of the analysis are provided in Tables 9 through
12.

       The cost per ton estimates for each individual program are presented separately in the
tables below, and are part of the justification for each of the programs. For informational
purposes, we also present the cost per ton for the three programs combined.
                                          ES-7

-------
   Table 9 HC Aggregate Cost per Ton and Long-Term Annual Cost Per Ton
                               ($2003)

Vehicles
Gas Cans
(without fuel
savings)
Gas Cans (with
fuel savings)
Combined (with
fuel savings)
Discounted
Lifetime
Cost per ton at 3%
$14
$230
$0
$0
Discounted
Lifetime
Cost per ton at 7%
$18
$250
$0
$0
Long-Term Cost
per Ton in 2020
$0
$160
$0
$0
Long-Term Cost
per Ton in 2030
$0
$180
$0
$0
Table 10 Benzene Aggregate Cost per Ton and Long-Term Annual Cost Per Ton
                               ($2003)

Fuels
Vehicles
Gas Cans
(without fuel
savings)
Gas Cans (with
fuel savings)
Combined (with
fuel savings)
Discounted
Lifetime
Cost per ton at 3%
$11,700
$260
$27,800
$0
$3,700
Discounted
Lifetime
Cost per ton at 7%
$11,900
$340
$30,900
$0
$4,000
Long-Term Cost
per Ton in 2020
$11,700
$0
$20,000
$0
$3,200
Long-Term Cost
per Ton in 2030
$12,300
$0
$21,600
$0
$2,700
                                ES-8

-------
      Table 11 MSAT Aggregate Cost per Ton and Long-Term Annual Cost Per Ton
                                        ($2003)

Fuels
Vehicles
Gas Cans
(without fuel
savings)
Gas Cans (with
fuel savings)
Combined (with
fuel savings)
Discounted
Lifetime
Cost per ton at 3%
$11,700
$40
$1,800
$0
$770
Discounted
Lifetime
Cost per ton at 7%
$11,900
$53
$2,000
$0
$850
Long-Term Cost
per Ton in 2020
$11,700
$0
$1,300
$0
$660
Long-Term Cost
per Ton in 2030
$12,300
$0
$1,400
$0
$500
    Table 12 Direct PM Aggregate Cost per Ton and Long-Term Annual Cost Per Ton
                                        ($2003)

Vehicles
Discounted
Lifetime
Cost per ton at 3%
$620
Discounted
Lifetime
Cost per ton at 7%
$820
Long-Term Cost
per Ton in 2020
$0
Long-Term Cost
per Ton in 2030
$0
Benefits

       This analysis projects significant benefits throughout the period from initial
implementation of the proposed standards through 2030.  When translating emission benefits to
health effects and monetized values, however, we only quantify the PM-related benefits
associated with the proposed cold temperature vehicle standards.  The reductions in PM from the
proposed cold temperature vehicle standards would result in significant reductions in premature
deaths and other serious human health effects, as well as other important public health and
welfare effects. Table 13 provides the estimated monetized benefits of the proposed cold
temperature vehicle standards for 2020 and 2030. We estimate that in 2030, the benefits we are
able to monetize are expected to be approximately $6.5 billion using a 3 percent discount rate
and $5.9 billion using a 7 percent discount rate, assuming a background PM threshold of 3 ug/m3
in the calculation of PM mortality. There are no compliance costs associated with the proposed
cold temperature vehicle program after 2019; vehicle compliance costs are primarily research
and development, and facility costs are expected to be recovered by manufacturers  over the first
ten years of the program beginning in 2010. Total costs of the entire MSAT proposal,  which
                                         ES-9

-------
include both the proposed gasoline container and vehicle fuel standards, are $205 million in 2030
(in 2003$, including fuel savings).

       EPA's consistent approach has been to model premature mortality associated with PM
exposure as a nonthreshold effect; that is, with harmful effects to exposed populations modeled
regardless of the absolute level of ambient PM concentrations.  This approach has been shaped
and supported by advice from EPA's technical peer review panel, the Science Advisory Board's
Health Effects Subcommittee (SAB-HES).  Note, however, that it is not certain whether there
exists a threshold below which there would be no benefit to further reductions in PM2.5.  We
consider the impact of a threshold in the PM-mortality concentration response function in
Section 12.6.1.1 oftheRIA.
Table 13 Estimated Monetized PM-Related Health Benefits of the Proposed Mobile Source
                    Air Toxics Standards: Cold Temperature Controls

Using a 3% discount rate
Using a 7% discount rate
Total Benefits3' b'c (billions 2003$)
2020
$3.4 + B
$3.1 + B
2030
$6.5 + B
$5.9 + B
   Benefits include avoided cases of mortality, chronic illness, and other morbidity health endpoints. PM-related
   mortality benefits estimated using an assumed PM threshold at background levels (3 ug/m3).  There is
   uncertainty about which threshold to use and this may impact the magnitude of the total benefits estimate. For a
   more detailed discussion of this issue, please refer to Section 12.6.1.1 of the PJA.
   For notational purposes, unqualified benefits are indicated with a "B" to represent the sum of additional
   monetary benefits and disbenefits. A detailed listing of unqualified health and welfare effects is provided in
   Table 13-2 of the RIA.
   Results reflect the use of two different discount rates: 3 and 7 percent, which are recommended by EPA's
   Guidelines for Preparing Economic Analyses and OMB Circular A-4. Results are rounded to three significant
   digits for ease of presentation and computation.
Economic Impact Analysis

       We prepared a draft Economic Impact Analysis (EIA) to estimate the economic impacts
of the proposed emission control program on the gas can, gasoline fuel, and light-duty vehicle
markets. We estimate the net social costs of the proposed program for 2020 and 2030 are
provided in Table 14 below.  These estimates reflects the estimated costs associated with the
gasoline, gas can, and vehicle controls and the expected fuel savings from better evaporative
controls on gas cans. The results of the economic impact modeling performed for the gasoline
fuel and gas can control programs suggest that the social costs of those two programs are
expected to be about $244.3 million in 2020 with consumers of these products expected to bear
about 60 percent of these costs.  We estimate fuel savings of about $72.8 million in 2020 that
will accrue to consumers.  There are no social costs associated with the vehicle program in 2020.

    Table 14 Net Social Costs Estimates for the Proposed Program (Millions of 2003$)
                                           ES-10

-------
                                          2020                          2030

 Net Social Costs                171.5                         205.2


Impact on Small Businesses

       We prepared a Regulatory Flexibility Analysis, which evaluates the potential impacts of
new standards and fuel controls on small entities. Before issuing our proposal, we analyzed the
potential impacts of this rule on small entities. As a part of this analysis, we interacted with
several small entities representing the various affected sectors and convened a Small Business
Advocacy Review Panel to gain feedback and advice from these representatives.  This feedback
was used to develop regulatory alternatives to address the impacts of the rule on small
businesses. Small entities raised general concerns related to potential difficulties and costs of
meeting the upcoming standards.

       The Panel  consisted of members from EPA, the Office of Management and Budget, and
the Small Business Administration's Office of Advocacy.  We either are proposing or requesting
comment on the Panel's recommendations. These provisions would reduce the burden on small
entities that would be subject to this rule's requirements. We have proposed provisions that give
small light-duty vehicle manufacturers, small gasoline refiners,  and small gas can manufacturers
several compliance options aimed specifically at reducing the burden on these small entities. In
general, for vehicles and fuels, the options proposed are similar to small entity provisions
adopted in prior rulemakings where EPA set vehicle and fuel standards.  The options proposed
for small gas can manufacturers are unique to this rulemaking since we are proposing gas can
standards for the first time. The small entity provisions for the three industry sectors would
reduce the burden on small entities that would be required to meet this proposed rule's
requirements.
                                         ES-11

-------
                                                             Chapter 1: Table of Contents

Chapter 1: Mobile Source Air Toxics Health Information	3
   1.1.   What Are MSATs?	3
      1.1.1.  Compounds Emitted by Mobile Sources and Identified in IRIS	3
      1.1.2.  Compounds Emitted by Mobile Sources and Included on Section 112(b) List of
            Hazardous Air Pollutants	6
      1.1.3.  Other Sources of Information on Compounds with Potential Serious Adverse
            Health Effects	7
      1.1.4.  Which Mobile Source Emissions Pose the Greatest Health Risk at Current Levels?8
      1.1.4.1.     Risk Drivers in 1999 National-Scale Air Toxics Assessment	8
      1.1.4.2.     1999 NATA Risk Drivers with Significant Mobile Source Contribution .... 10
   1.2.   Dose-Response and Agency Risk Assessment Practice	11
      1.2.1.  Cancer	11
      1.2.2.  Chronic Exposure and Noncancer Health Effects	13
      1.2.3.  Acute Exposure and Noncancer Health Effects	13
   1.3.   Summary of Air Toxic Health Effects	14
      1.3.1.  Benzene	14
      1.3.2.  1,3-Butadiene	17
      1.3.3.  Formaldehyde	17
      1.3.4.  Acetaldehyde	19
      1.3.5.  Acrolein	19
      1.3.6.  Naphthalene	20
      1.3.7.  2,2,4-Trimethylpentane	20
      1.3.8.  Ethylbenzene	21
      1.3.9.  n-Hexane	21
      1.3.10. Methyl Tertiary Butyl Ether (MTBE)	22
      1.3.11. Styrene	22
      1.3.12. Toluene	23
      1.3.13.Xylenes	24
      1.3.14. Polycyclic Organic Matter (POM)	24
      1.3.15. Diesel Particulate Matter (DPM) and Diesel Exhaust Organic Gases (DEOG)	25
   1.4.   Emerging Issues	26
      1.4.1.  Gasoline PM	26
      1.4.2.  Metals	28
                                          1-1

-------
1-2

-------
        Chapter 1: Mobile Source Air Toxics Health Information

1.1.     What Are MSATs?

       Section 202(1) refers to "hazardous air pollutants from motor vehicles and motor vehicle
fuels." We use the term "mobile source air toxics (MSATs)" to refer to compounds that are
emitted by mobile sources and have the potential for serious adverse health effects. There are a
variety of ways in which to identify compounds that have the potential for serious adverse health
effects.  For example, EPA's Integrated Risk Information System (IRIS) is EPA's database
containing information on human health effects that may result from exposure to various
chemicals in the environment.  In addition,  Clean Air Act section 112(b) contains a list of
hazardous air pollutants that EPA is required to control through regulatory standards; other
agencies or programs such as the Agency for Toxic Substances and Disease Registry and the
California EPA have developed health benchmark values for various compounds; and the
International Agency for Research on Cancer and the National Toxicology Program have
assembled evidence of substances that cause cancer in humans and issue judgments on the
strength of the evidence. Each source of information has its own strengths and limitations. For
example, there are inherent limitations on the number of compounds that have been investigated
sufficiently for EPA to conduct an IRIS assessment. There are some compounds that are not
listed in IRIS but are considered to be hazardous air pollutants under Clean Air Act section
112(b) and are regulated by the Agency (e.g., propionaldehyde,  2,2,4-trimethylpentane).

1.1.1.   Compounds Emitted by Mobile Sources and Identified in IRIS

       In its 2001 MS AT rule, EPA identified a list of 21 MSATs. We listed a compound as an
MSAT if it was emitted from mobile sources, and if the Agency had concluded in IRIS that the
compound posed a potential cancer hazard and/or if IRIS contained an inhalation reference
concentration or ingestion reference dose for the compound. Since 2001, EPA has conducted an
extensive review of the literature to produce a list of the compounds identified in the exhaust or
evaporative emissions from onroad and nonroad equipment, using baseline as well as alternative
fuels (e.g., biodiesel, compressed natural gas).1 This list, the Master List of Compounds Emitted
by Mobile Sources ("Master List"), currently includes approximately 1,000 compounds.  It is
available in the public docket for this rule and on the web (www.epa.gov/otaq/toxics.htm).  Table
1.1 .-1 lists those compounds from the Master List that currently meet those 2001 MSAT criteria,
based on the current IRIS.

       Table 1.1 .-1 identifies all of the compounds from the Master List that are present in IRIS
with (a) a cancer hazard identification of known, probable, or possible human carcinogens (under
the 1986 EPA cancer guidelines) or carcinogenic to humans, likely to be carcinogenic to humans,
or suggestive evidence of carcinogenic potential (under the 2005 EPA cancer guidelines); and/or
(b) an inhalation reference concentration or an ingestion reference dose. Although all these
compounds have been detected in emissions from mobile sources, many are emitted in trace
amounts and data are not adequate to develop an inventory.  Those compounds for which we
have developed an emissions inventory are  summarized in Table 2.2.-1. There are several
compounds for which IRIS assessments are underway and therefore are not included in Table
                                          1-3

-------
l.l.-l. These compounds are: cerium, copper, ethanol, ethyl tertiary butyl ether (ETBE),
platinum, propionaldehyde, and 2,2,4-trimethylpentane.

       The fact that a compound is listed in Table l.l.-l does not imply a risk to public health or
welfare at current levels, or that it is appropriate to adopt controls to limit the emissions of such a
compound from motor vehicles or their fuels. In conducting any such further evaluation,
pursuant to sections 202(a) or 21 l(c) of the Act, EPA would consider whether emissions of the
compound from motor vehicles cause or contribute to air pollution which may reasonably  be
anticipated to endanger public health or welfare.
                                           1-4

-------
Table l.l.-l. Compounds Emitted by Mobile Sources That Are Listed in IRIS*
1,1,1 ,2-Tetrafluoroethane
1,1,1 -Trichloroethane
1,1-Biphenyl
1 ,2-Dibromoethane
1 ,2-Dichlorobenzene
1,3-Butadiene
2,4-Dinitrophenol
2-Methylnaphthalene
2-Methylphenol
4-Methylphenol
Acenaphthene
Acetaldehyde
Acetone
Acetophenone
Acrolein (2-propenal)
Ammonia
Anthracene
Antimony
Arsenic, inorganic
Barium and compounds
B enz [a] anthracene
Benzaldehyde
Benzene
Cadmium
Carbon disulfide
Carbon tetrachloride
Chlorine
Chlorobenzene
Chloroform
Chromium III
Chromium VI
Chrysene
Crotonaldehyde
Cumene (isopropyl benzene)
Cyclohexane
Cyclohexanone
Di(2-ethylhexyl)phthalate
Dib enz [a,h] anthracene
Dibutyl phthalate
Di chl oromethane
Diesel PM and Diesel exhaust
organic gases
Di ethyl phthalate
Ethylbenzene
Ethylene glycol monobutyl
ether
Fluoranthene
Fluorene
Manganese
Mercury, elemental
Methanol
Methyl chloride
Methyl ethyl ketone
(MEK)
Methyl isobutyl ketone
(MIBK)
Methyl tert-butyl ether
(MTBE)
Molybdenum
Naphthalene
Nickel
Nitrate
N-Nitrosodiethylamine
N-Nitrosodimethylamine
N-Nitroso-di-n-
butylamine
N-Nitrosodi-N-
propylamine
N-Nitrosopyrrolidine
Pentachlorophenol
Phenol
Phosphorus
Phthalic anhydride
Pyrene
Selenium and compounds
Silver
                                 1-5

-------
Benzo[a]pyrene (BaP)
B enzo [b ]fluoranthene
Benzo[k]fluoranthene
Benzole acid
Beryllium and compounds
Boron (Boron and Borates
only)
Bromomethane
Butyl benzyl phthalate
Formaldehyde
Furfural
Hexachl orodib enzo-p-di oxin,
mixture (dioxin/furans)
n-Hexane
Hydrogen cyanide
Hydrogen sulfide
Indeno[ 1 ,2,3 -cd]pyrene
Lead and compounds
(inorganic)
Strontium
Styrene
Tetrachloroethylene
Toluene
Tri chl orofluoromethane
Vanadium
Xylenes
Zinc and compounds
   *Compounds listed in IRIS as known, probable, or possible human carcinogens and/or
   pollutants for which the Agency has calculated a reference concentration or reference dose.

1.1.2.   Compounds Emitted by Mobile Sources and Included on Section 112(b) List of
     Hazardous Air Pollutants

       Clean Air Act section 112(b) contains a list of hazardous air pollutants that EPA is
required to control through regulatory standards. As discussed above, there are some compounds
emitted by mobile sources that are not listed in IRIS but are considered to be hazardous air
pollutants under Clean Air Act section 112(b) and are regulated by the Agency  such as
propionaldehyde and 2,2,4-trimethylpentane. Compounds emitted by mobile sources that are
Clean Air Act section 112(b) hazardous air pollutants are listed in Table 1.1.-2. Although all
these compounds have been  detected in emissions from mobile sources, many are emitted in
trace amounts and data are not adequate to develop an inventory.  Those compounds for which
we have developed an emissions inventory are summarized in Table 2.2.-1.

   Table 1.1.-2. Compounds Emitted by Mobile Sources That Are Listed in CAA Section
1 , 1 ,2-Trichloroethane
1 ,2-Dibromoethane
1,3 -Butadiene
2,2,4-Trimethylpentane
2,3,7,8-Tetrachlorodibenzo-p-
dioxin
2,4-Dinitrophenol
Carbon disulfide
Carbon tetrachloride
Chlorine
Chlorobenzene
Chloroform
Chromium (III and VI)
Methyl ethyl ketone
Methyl tert-butyl ether
Methyl chloride
Naphthalene
Nickel compounds
N-Nitrosodimethylamine
                                          1-6

-------
2-Methylphenol (o-cresol)
4-Methylphenol (p-cresol)
Acetaldehyde
Acetophenone
Acrolein
Antimony compounds
Arsenic compounds
Benzene
Beryllium
Biphenyl
Bromomethane
Cadmium compounds
Cumene
Di(2-ethylhexyl)phthalate
(DEHP)
Dibutylphthalate
Dichloromethane
Ethyl benzene
Formaldehyde
Hexane
Hydrogen cyanide
("Cyanide compounds in
Section 112(b))
Lead compounds
Manganese
Mercury compounds
Methanol
Pentachlorophenol
Phenol
Phosphorus
Phthalic anhydride
Polycyclic organic matter*
Propionaldehyde
Selenium compounds
Styrene
Tetrachl oroethy 1 ene
Toluene
Xylenes (isomers and mixture)

*Includes organic compounds with more than one benzene ring, and which have a boiling point greater than or equal
to 100.5 C.

1.1.3.   Other Sources of Information on Compounds with Potential Serious Adverse
     Health Effects

       Additional sources of information are available to characterize the potential for cancer or
noncancer health effects from toxic air pollutants.  These include the Agency for Toxic
Substances and Disease Registry list of minimal risk levels (http://www.atsdr.cdc.gov/mrls.html),
California EPA list of Reference Exposure Levels,
(http://www.oehha.ca.gov/risk/ChemicalDB/index.asp), International Agency for Research on
Cancer lists of carcinogenic compounds (http://www.iarc.fr/ENG/Databases/index.php), the
National Toxicology Program list of carcinogenic compounds (http://ntp-server.niehs.nih.gov/),
and the U.S. EPA Emergency Planning and Community Right-to-Know Act list of extremely
hazardous substances (http://yosemite.epa.gov/oswer/ceppoehs.nsf/content/BackGround). EPA
relies on these sources of information, as appropriate, for certain types of analyses.2
                                           1-7

-------
1.1.4.   Which Mobile Source Emissions Pose the Greatest Health Risk at Current Levels?

       The 1999 National-Scale Air Toxics Assessment (NAT A) provides some perspective on
which mobile source emissions pose the greatest risk at current estimated ambient levels.A We
also conducted a national-scale assessment for future years, which is discussed more fully in
Chapters 2 and 3 of the RIA.  Our understanding of what emissions pose the greatest risk will
evolve over time, based on our understanding of the ambient levels and health effects associated
with the compounds.8

1.1.4.1.        Risk Drivers in 1999 National-Scale Air Toxics Assessment

       The 1999 NATA evaluates 177 hazardous air pollutants currently listed under CAA
section 112(b), as well as diesel PM. NATA is described in greater detail in Chapters 2 and 3 of
this RIA.  Additional information can also be obtained from the NATA website
(http://www.epa.gov/ttn/atw/natal999). Based on the assessment of inhalation exposures
associated with outdoor sources of these hazardous air pollutants, NATA has identified cancer
and noncancer risk drivers on a national and regional scale (Table 1.1.-3). A cancer risk driver
on a national scale is a hazardous air pollutant for which at least 25 million people are exposed to
risk greater than ten in one million.  Benzene is the only compound identified in the 1999 NATA
as a national cancer risk driver. A cancer risk driver on a regional scale is a hazardous air
pollutant for which at least one million people are exposed to risk greater than ten in one million
or at least 10,000 people are exposed to risk greater than 100 in one million.  Twelve compounds
(or groups of compounds in the case of POM) were identified as  regional cancer risk drivers.
The 1999 NATA concludes that diesel  particulate matter is  among the substances that pose the
greatest relative risk, although the cancer risk cannot be quantified.

       A noncancer risk driver at the national scale is a hazardous air pollutant for which at least
25 million people are exposed at a concentration greater than the inhalation reference
concentration.  The RfC is an estimate  (with uncertainty spanning perhaps an order of
magnitude) of a daily  exposure to the human population (including sensitive subgroups) that is
likely to be without appreciable risk of deleterious effects during a lifetime.  Acrolein is the only
compound identified in the 1999 NATA as a national noncancer risk driver. A noncancer risk
driver on a regional scale is defined as  a hazardous air pollutant for which at least 10,000 people
are exposed to an ambient concentration greater than the inhalation reference concentration.
Sixteen regional-scale noncancer risk drivers were identified in the 1999 NATA (see Table  1.1.-
3.).
       A It is, of course, not necessary for EPA to show that a compound is a national or regional risk driver to
show that its emission from motor vehicles may reasonably cause or contribute to endangerment of public health or
welfare.  A showing that motor vehicles contribute some non-trivial percentage of the inventory of a compound
known to be associated with adverse health effects would normally be sufficient.  Cf. Bluewater Network v. EPA.
370 F. 3d 1, 15 (D.C. Cir. 2004).
       B The discussion here considers risks other than those attributed to ambient levels of criteria pollutants.


                                            1-8

-------
Table 1.1.-3. National and Regional Cancer and Noncancer Risk Drivers in 1999 NATA
Cancer l
National drivers 2
Benzene
Regional drivers 3
Arsenic compounds
Benzidine
1,3-Butadiene
Cadmium compounds
Carbon tetrachloride
Chromium VI
Coke oven
Ethyl ene oxide
Hydrazine
Naphthalene
Perchl oroethy 1 ene
Polycyclic organic matter




Noncancer
National drivers 4
Acrolein
Regional drivers 5
Antimony
Arsenic compounds
1,3-Butadiene
Cadmium compounds
Chlorine
Chromium VI
Diesel PM
Formaldehyde
Hexamethylene 1-6-diisocyanate
Hydrazine
Hydrochloric acid
Maleic anhydride
Manganese compounds
Nickel compounds
2,4-Toluene diisocyanate
Triethylamine
         lrThe list of cancer risk drivers does not include diesel paniculate matter. However, the 1999 NATA
         concluded that it was one of the pollutants that posed the greatest relative cancer risk.
         2 At least 25 million people exposed to risk >10 in 1 million
         3 At least 1 million people exposed to risk >10 in 1  million or at least 10,000 people exposed to risk >100
         in 1 million
         4 At least 25 million people exposed to a hazard quotient > 1.0
         5 At least 10,000 people exposed to a hazard quotient > 1

       It should be noted that varying levels of confidence are associated with risk estimates for
individual pollutants, based on the quality of the data used to estimate emissions, ambient
concentrations and exposure.  For the pollutants included in NATA, EPA rated its confidence in
                                              1-9

-------
risk estimates, based on the quality of the data used for emissions, air quality, and exposure
modeling, as high, medium, or lower. EPA has a high level  of confidence in the data for
benzene, medium confidence in the data for formaldehyde, but lower confidence in data for 1,3-
butadiene and acrolein.
1.1.4.2.
1999 NATA Risk Drivers with Significant Mobile Source Contribution
       Among the national and regional-scale cancer and noncancer risk drivers identified in the
1999 NATA, seven compounds have significant contributions from mobile sources: benzene,
1,3-butadiene, formaldehyde, acrolein, polycyclic organic matter (POM), naphthalene, and diesel
particulate matter and diesel exhaust organic gases (Table 1.1.-4.). For example, mobile sources
contribute 68% of the national benzene inventory, with 49% from on-road sources and 19% from
nonroad sources.
           Table 1.1.-4. Mobile Source Contribution to 1999 NATA Risk Drivers
1999 NATA Risk Drivers
Benzene
1,3 -Butadiene
Formaldehyde
Acrolein
Polycyclic organic matter*
Naphthalene
Diesel PM and Diesel
exhaust organic gases
Percent
Contribution
from All
Mobile Sources
68%
58%
47%
25%
6%
27%
100%
Percent
Contribution
from On-road
Mobile Sources
49%
41%
27%
14%
3%
21%
38%
       *This POM inventory includes the 15 POM compounds: benzo[b]fluoranthene, benz[a]anthracene,
       indeno(l,2,3-c,d)pyrene, benzo[k]fluoranthene, chrysene, benzo[a]pyrene, dibenz(a,h)anthracene,
       anthracene, pyrene, benzo(g,h,i)perylene, fluoranthene, acenaphthylene, phenanthrene, fluorene, and
       acenaphthene.
                                           1-10

-------
1.2.     Dose-Response and Agency Risk Assessment Practice

       This section describes EPA's formal process for conducting risk assessment. The
EPA framework for assessing and managing risks reflects the risk assessment and risk
management paradigm set forth by the National Academy of Sciences in 19833 which
was incorporated into the 1986 EPA risk guidance4 and revised in 2005 in the EPA
Guidelines for Carcinogen Risk Assessment and Supplemental Guidance for Assessing
Susceptibility from Early-Life Exposure to Carcinogens.5 The paradigm divides the risk
assessment and management process into four general phases.  The first three phases
(exposure assessment, dose-response assessment, and risk characterization) comprise risk
assessment. The fourth phase, risk management,  involves evaluation of information
provided by the risk assessment to the environmental manager who makes a risk
management decision.

       An exposure assessment is the quantitative or qualitative evaluation of contact to
a specific pollutant and  includes such characteristics as intensity, frequency, and duration
of contact.  The numerical output of an exposure assessment may be either exposure or
dose, depending on the purpose of the evaluation  and available data.

       The dose-response assessment produces two sequential analyses. The first
analysis is the hazard identification, which identifies contaminants that are suspected to
pose health hazards, describes the specific forms of toxicity (e.g., neurotoxicity,
carcinogenicity, etc.) that they may cause, and evaluates the conditions under which these
forms of toxicity might  be expressed in exposed humans. The types of effects that are
relevant to a particular chemical (e.g., cancer, noncancer) are determined as part of the
hazard identification.

       The second analysis is the human  health dose-response assessment, which
generally describes the characterization of the relationship between the concentration,
exposure, or dose of a pollutant and the resultant health effects.  Dose-response
assessment methods generally consist of two parts. First is the evaluation of the
experimentally observed relationship between health effects and the concentration,
exposure and/or dose of a particular compound, and second is the extrapolation from the
observed range to lower doses and risks.

1.2.1.   Cancer

       The term 'cancer' is used to describe a group of related diseases that affect a
variety of organs and tissues. Cancer results from a combination of genetic damage and
nongenetic factors that favor the growth of damaged cells. The EPA document,
Guidelines for Carcinogen Risk Assessment6 (2005) provides guidance on hazard
identification for carcinogens.  The approach recognizes three broad categories of data:
(1)  human data (primarily, epidemiological); (2) results of long-term experimental animal
bioassays; and (3) supporting data, including a variety of short-term tests for genotoxicity
and other relevant properties. The 2005 Guidelines for hazard identification recommend
that an agent's human carcinogenic potential be described in a weight-of-evidence
                                      1-11

-------
narrative.  The narrative summarizes the full range of available evidence and describes
any conditions associated with conclusions about an agent's hazard potential (e.g.,
carcinogenic by some routes of exposure and not others). To provide additional clarity
and consistency in weight-of-evidence narratives, the Guidelines suggest a set of weight-
of-evidence descriptors to accompany the narratives. The five descriptors are:
Carcinogenic to Humans, Likely to be Carcinogenic to Humans, Suggestive Evidence of
Carcinogenic Potential, Inadequate Information to Assess Carcinogenic Potential, and
Not Likely to be Carcinogenic to Humans. These descriptors replace those based on the
EPA 1986 Risk Assessment Guidelines which classified a compound as Group A:
Carcinogenic to Humans, Group B: Probably Carcinogenic to Humans, Group C:
Possibly Carcinogenic to Humans, Group D: Not Classifiable as to Human
Carcinogenicity, or Group E: Evidence of Noncarcinogenicity for Humans.

       A quantitative assessment is performed depending on the weight-of-evidence and
the suitability of the available information regarding a relationship between the  dose of a
compound and the effect it causes (dose-response data). Dose-response models are used
to calculate unit risk estimates (URE). Inhalation cancer risks are quantified by EPA
using the unit risk, which represent the excess lifetime  cancer risk estimated to result
from continuous lifetime exposure to an agent at a concentration of 1 |ig/m3 in air. These
unit risks are typically upper-bound estimates, although where there are adequate
epidemiological data, the unit risk may be based on a maximum likelihood estimate
(MLE). Except for benzene and chromium, where risks are based on maximum
likelihood dose-response values, risks from mobile source air toxics should all be
considered upper-bound values.  This means they are plausible upper limits to risks. True
risks could be greater, but are likely to be lower, and could  be zero.  A discussion of the
confidence in a quantitative cancer risk estimate is provided in the IRIS file for  each
compound. The discussion of the confidence in the cancer  risk estimate includes an
assessment of the  source of the data (human or animal), uncertainties in dose estimates,
choice of the model used to fit the exposure and response data and how uncertainties and
potential confounders are handled.

       The 2005 Guidelines include Supplemental Guidance for Assessing Susceptibility
from Early-Life Exposure to Carcinogens.7 The Supplemental Guidance is part of EPA's
response to the recommendation of the National Research Council (1994) that "EPA
should assess risks to infants  and children whenever it appears that their risks might be
greater than those of adults."  For several potential carcinogens, there is some evidence of
higher cancer risks following early-life exposure. Accordingly, the Supplemental
Guidance describes the approaches that EPA could use in assessing cancer risks
following early-life exposures.  The 1999 NATA does not include default adjustments for
early life exposures recently recommended in the Supplemental Guidance. Incorporation
of such adjustments would lead to higher estimates of lifetime risk.
                                      1-12

-------
1.2.2.   Chronic Exposure and Noncancer Health Effects

       Noncancer effects resulting from chronic exposures include a wide range of
effects in all organ systems, e.g., respiratory, cardiovascular, immune, kidney.  Hazard
identification procedures for chronic noncancer effects are described in EPA guidelines.
The EPA has published guidelines for assessing several specific types of noncancer
effects, including mutagenicity,8 developmental toxicity,9 neurotoxicity10; and
reproductive toxicity.u For identification of hazards resulting from long-term (chronic)
exposures, available data on different health endpoints and target organs are ordered and
discussed, and the effects (and their attendant dose/exposure levels) are described.
Particular attention is given to effects that occur at relatively low doses or that may have
particular relevance to human populations. The inhalation reference concentration (RfC)
and oral reference dose (RfD) are the primary Agency consensus quantitative toxicity
values for use in chronic noncancer risk assessment. The RfC or RfD is defined as an
estimate, with uncertainty spanning perhaps an order of magnitude, of an inhalation
exposure/oral dose to the human population (including sensitive subgroups) that is likely
to be without appreciable risks of deleterious effects during  a lifetime. The RfC or RfD is
derived using 1) a thorough review of the health effects database for an individual
chemical and 2) the most sensitive and relevant endpoint and the principal study(ies)
demonstrating that endpoint. RfCs for inhalation are derived according to the Agency's
1994 guidance.12 A statement regarding the confidence in the RfC and/or RfD is
developed to reflect the confidence in the principal study or studies on which the RfC or
RfD are based and the confidence in the underlying database. Factors that affect the
confidence in the principal study include how well the study was designed, conducted
and reported.  Factors that affect the confidence in the database include an assessment of
the availability of information regarding identification of the critical effect, potentially
susceptible populations and exposure scenarios relevant to assessment of risk.  In 2002 an
EPA RfC/RfD Technical Panel prepared several recommendations for preparation of
noncancer reference values.13

1.2.3.   Acute Exposure and Noncancer Health Effects

       Noncancer health impacts resulting from acute (short-term) exposures have been
assessed for many compounds in the occupational setting. EPA currently does not have
acute exposures reference values in IRIS comparable to the RfC described above. EPA's
Office of Research and Development proposed an Acute Reference Exposure (ARE)
approach for evaluating short term exposure effects in 1998.14 In 2002 EPA completed a
review document which summarizes recommendations of the EPA RfC/RfD Technical
Panel for preparation of noncancer reference values including acute exposure values.15
In response to the EPA Science Advisory Board review of the Acute Reference Exposure
methodology and recommendations from EPA's RfC/RfD Technical Panel, ORD is
currently developing an advanced acute inhalation reference concentration (acute RfC)
methodology.  As part of this new methodology, acute inhalation assessments are being
developed.
                                       1-13

-------
1.3.     Summary of Air Toxic Health Effects

       From a public health perspective, it is important to assess the emission
contributions to atmospheric levels of various air toxics (including diesel PM and exhaust
organic gases) emitted by motor vehicle engines, including their physical properties,
sources of potential exposure, and health hazards. In this section, we describe the cancer
and noncancer health effects attributed to chronic exposure to various mobile source air
toxics as well as any acute exposure health effects, where data are available. We focus
here on the air toxics that are identified in the NATA as risk drivers (see Section 1.1) and
that account for a significant share of mobile sources emissions.  We also consider
compounds for which we expect emission reductions from today's proposed rule. We are
also including diesel paniculate matter and diesel exhaust organic gases in this discussion.
EPA has concluded that diesel exhaust ranks with the other substances that the national-
scale assessment suggests pose the greatest relative risk.

1.3.1.   Benzene

       Benzene is an aromatic hydrocarbon that is present as a gas in both exhaust and
evaporative emissions from mobile sources.  Inhalation is the major source of human
exposure to benzene in the occupational and non-occupational setting.

       The EPA's IRIS database lists benzene as  a known human carcinogen (causing
leukemia) by all routes of exposure.16 A number  of adverse noncancer health effects
including blood disorders and immunotoxicity, have also been associated with long-term
occupational exposure to benzene.

       Long-term inhalation occupational exposure to benzene has been shown to cause
cancers of the hematopoetic (blood cell) system in adults.  Among these are acute
                         j-i                                   I •71 O
nonlymphocytic leukemia,  and chronic lymphocytic leukemia.  '    A doubling of risk
for acute nonlymphocy tic leukemia and myelodysplastic syndrome was found at average
exposure levels under  10 ppm (32 mg/m3).19 EPA has not formally evaluated this study
as part of the IRIS review process.  Leukemias, lymphomas, and other tumor types have
been observed in experimental  animals exposed to benzene by inhalation or oral
administration.  Exposure to benzene and/or its metabolites has also been linked with
       c Leukemia is a blood disease in which the white blood cells are abnormal in type or number.
Leukemia may be divided into nonlymphocytic (granulocytic) leukemias and lymphocytic leukemias.
Nonlymphocytic leukemia generally involves the types of white blood cells (leukocytes) that are involved
in engulfing, killing, and digesting bacteria and other parasites (phagocytosis) as well as releasing
chemicals involved in allergic and immune responses. This type of leukemia may also involve
erythroblastic cell types (immature red blood cells). Lymphocytic leukemia involves the lymphocyte type
of white bloods cell that are responsible for the immune responses.  Both nonlymphocytic and lymphocytic
leukemia may, in turn, be separated into acute (rapid and fatal) and chronic (lingering, lasting) forms. For
example; in acute myeloid leukemia there is diminished production of normal red blood cells (erythrocytes),
granulocytes, and platelets (control clotting), which leads to death by anemia, infection, or hemorrhage.
These events can be rapid.  In chronic myeloid leukemia (CML) the leukemic cells retain the ability to
differentiate (i.e., be responsive to stimulatory factors) and perform function; later there is a loss of the
ability to respond.
                                         1-14

-------
chromosomal changes in humans and animals20' 21 and increased proliferation of mouse
                 fjrj ry^
bone marrow cells.  '

       The latest assessment by EPA places the excess risk of developing acute
nonlymphocytic leukemia from inhalation exposure to benzene at 2.2 x 10"6 to 7.8 x 10"6
per |ig/m3.  In other words, there is a risk of about two to eight excess leukemia cases in
one million people exposed to 1 |ig/m3 of benzene over a lifetime.24 This range of unit
risks are the MLEs calculated from different exposure assumptions and dose-response
models that are linear at low doses. At present, the true cancer risk from exposure to
benzene cannot be ascertained, even though dose-response data are used in the
quantitative cancer risk analysis, because of uncertainties in the low-dose exposure
scenarios and lack of clear understanding of the mode of action. A range of estimates of
risk is recommended, each having equal scientific plausibility. There are confidence
intervals associated with the MLE range that reflect random variation of the observed
data. For the upper end of the MLE range, the 5th and 95th percentile values are about a
factor of 5 lower and higher than the best fit value.  The upper end of the MLE range was
used in the 1999 NATA.

       It should be noted that not enough information is known to determine the slope of
the dose-response curve at environmental levels of exposure and to provide a sound
scientific basis to choose any particular extrapolation/exposure model to estimate human
cancer risk at low doses. EPA risk assessment guidelines suggest using an assumption of
linearity of dose response when (1) there is  an absence of sufficient information on
modes of action or (2) the mode of action information indicates that the dose-response
curve at low dose is or is expected to be linear.25  Since  the mode of action for benzene
carcinogenicity is unknown, the current cancer unit risk estimate assumes linearity of the
low-dose response. Data that were considered by EPA in its carcinogenic update
suggested that the dose-response relationship at doses below those examined  in the
studies reviewed in EPA's most recent benzene assessment may be supralinear. They
support the inference that cancer risks are as high, or higher than the estimates provided
in the existing EPA assessment.26 Data discussed in the EPA IRIS assessment suggest
that genetic abnormalities occur at low exposure in humans, and the formation of toxic
metabolites plateaus above 25 ppm (80,000 jig/m3).27 More recent data on benzene
adducts in humans, published after the most recent IRIS assessment, suggest that the
enzymes involved in benzene metabolism start to saturate at exposure levels as low as 1
ppm.28 Because there is a transition from linear to saturable metabolism below 1 ppm,
the assumption  of low-dose linearity extrapolated from much higher exposures could lead
to substantial underestimation of leukemia risks.  This is consistent with recent
epidemiological data which also suggest a supralinear exposure-response relationship and
which "[extend] evidence for hematopoietic cancer risks to levels substantially lower than
had previously been established".29'30 These data are from the largest cohort  study done
to date with individual worker exposure estimates. However, these data have not yet
been formally evaluated by EPA as part of the IRIS review process, and it is not clear
whether these data provide sufficient evidence to reject a linear dose-response curve. A
better understanding of the biological mechanism of benzene-induced leukemia is needed.
                                       1-15

-------
       Children may represent a subpopulation at increased risk from benzene exposure,
due to factors that could increase their susceptibility.  Children may have a higher unit
body weight exposure because of their heightened activity patterns which can increase
their exposures, as well as different ventilation tidal volumes and frequencies, factors that
influence uptake. This could entail a greater risk of leukemia and other toxic effects to
children if they are exposed to benzene at similar levels as adults.  There is limited
information from two studies regarding an increased risk to children whose parents have
                                      11 10
been occupationally exposed to benzene.  '   Data from animal studies have shown
benzene exposures result in damage to the hematopoietic (blood cell formation) system
during development.33'34'35  Also, key changes related to the development of childhood
leukemia occur in the developing fetus.36 Several studies have reported that genetic
changes related to eventual leukemia development occur before birth. For  example, there
is one study of genetic  changes in twins who developed T cell leukemia at 9 years of
age.37  An association between traffic volume, residential proximity to busy roads and
occurrence of childhood leukemia has also been identified in some studies, although
some studies show no association.  These studies are discussed later in Chapter 3.

       A number of adverse noncancer health effects, including blood disorders such as
preleukemia and aplastic anemia, have also been associated with long-term exposure to
benzene.38'39 People with long-term occupational exposure to benzene  have experienced
harmful effects on the blood-forming tissues,  especially in bone marrow. These  effects
can disrupt normal blood production and suppress the production of important blood
components, such as red and white blood cells and blood platelets, leading  to anemia (a
reduction in the number of red blood cells), leukopenia (a reduction in the number of
white blood cells), or thrombocytopenia (a reduction in the number of blood platelets,
thus reducing the ability of blood to clot).  Chronic inhalation exposure to benzene in
humans and animals results in pancytopenia,0 a condition characterized by decreased
numbers of circulating erythrocytes (red blood cells), leukocytes (white blood cells), and
thrombocytes (blood platelets).40'41 Individuals that develop pancytopenia and have
continued exposure to benzene may develop aplastic anemia, whereas others exhibit both
pancytopenia and bone marrow hyperplasia (excessive cell formation), a condition that
may indicate a preleukemic state.42' 43  The most sensitive noncancer effect observed in
humans, based on current data, is the depression of the absolute lymphocyte count in
blood.44' 45

       EPA's inhalation reference concentration (RfC) for benzene is 30 |ig/m3. The
overall confidence in this RfC is medium.  The RfC is based on suppressed  absolute
lymphocyte counts seen in humans under occupational exposure conditions.  Since
development of this RfC, there have appeared reports in the medical literature of
benzene's hematotoxic effects in humans that provide data suggesting a wide range of
       D Pancytopenia is the reduction in the number of all three major types of blood cells (erythrocytes,
or red blood cells, thrombocytes, or platelets, and leukocytes, or white blood cells). In adults, all three
major types of blood cells are produced in the bone marrow of the vertebra, sternum, ribs, and pelvis. The
bone marrow contains immature cells, known as multipotent myeloid stem cells, that later differentiate into
the various mature blood cells. Pancytopenia results from a reduction in the ability of the red bone marrow
to produce adequate numbers of these mature blood cells.
                                        1-16

-------
hematological endpoints that are triggered at occupational exposures of less than 5 ppm
(about 16 mg/m3)46 and, more significantly, at air levels of 1 ppm (about 3 mg/m3) or less
among genetically susceptible populations.47  These studies had large sample sizes and
extensive individual exposure monitoring. One recent study found benzene metabolites
in mouse liver and bone marrow at environmental doses, indicating that even
concentrations in urban air may elicit a biochemical response in rodents that indicates
toxicity.48 EPA has not formally evaluated these recent studies as part of the IRIS review
process to determine whether or not they will lead to a change in the current RfC. EPA
does not currently have an acute reference concentration for benzene. The Agency for
Toxic Substances and Disease Registry Minimal Risk Level for acute exposure to
benzene is 160 |ig/m3 for 1-14 days exposure.

1.3.2.   1,3-Butadiene

       1,3-butadiene is formed in engine exhaust by the incomplete combustion of fuel.
It is not present in engine evaporative emissions because it is not generally present in an
appreciable amount in vehicle fuels.

       EPA has characterized 1,3-butadiene as a leukemogen, carcinogenic to humans by
inhalation. 49' 50 The specific mechanisms of 1,3-butadiene-induced carcinogenesis are
unknown however, it is virtually certain that the carcinogenic effects are mediated by
genotoxic metabolites of 1,3-butadiene.  Animal data suggest that females may be more
sensitive than males for cancer effects; nevertheless, there are insufficient data from
which to draw any conclusions on potentially sensitive subpopulations.  The upper bound
cancer unit risk estimate is 0.08 per ppm or 3xlO"5 per |ig/m3 (based primarily on linear
modeling and extrapolation of human data).  In other words, it is estimated that
approximately 30 persons in one million exposed to 1 |ig/m3 of 1,3-butadiene
continuously for their lifetime would develop cancer as a result of this exposure. The
human incremental lifetime unit cancer risk estimate is based on extrapolation from
leukemias observed in an occupational epidemiologic study.51 This estimate includes a
two-fold adjustment to the epidemiologic-based unit cancer risk applied to reflect
evidence from the rodent bioassays suggesting that  the epidemiologic-based estimate
(from males) may underestimate total cancer risk from 1,3-butadiene exposure in the
general population, particularly for breast cancer in females. Confidence in the excess
cancer risk estimate of 0.08 per ppm is moderate.52

       1,3-Butadiene also causes a variety of reproductive and  developmental effects in
mice; no human data on these effects are available.  The most  sensitive effect was ovarian
atrophy observed in a lifetime bioassay of female mice.53  Based on this critical effect
and the benchmark concentration methodology, an RfC for chronic health effects was
calculated at 0.9 ppb (approximately 2 |ig/m3). Confidence in the inhalation RfC is
medium.

1.3.3.   Formaldehyde
                                       1-17

-------
       Formaldehyde is the most prevalent aldehyde in engine exhaust.  It is formed as a
result of incomplete fuel combustion in both gasoline and diesel engines, although
formaldehyde accounts for a smaller quantity of total exhaust hydrocarbons from
gasoline engines. Formaldehyde emissions can vary substantially by engine duty cycle,
emission control system and composition of fuel.  Formaldehyde is not a component of
evaporative emissions but it can be formed photochemically in the atmosphere.

       Since 1987, EPA has classified formaldehyde as a probable human carcinogen
based on evidence in humans and in rats, mice, hamsters, and monkeys.54   Recently
released research conducted by the National Cancer Institute (NCI) found an increased
risk of nasopharyngeal cancer among workers exposed to formaldehyde.55'56 A recent
National Institute of Occupational Safety and Health (NIOSH) study of garment workers
also found increased risk of death due to leukemia among workers exposed to
formaldehyde.57 In 2004, the working group of the  International Agency for Research on
Cancer concluded that formaldehyde is carcinogenic to humans (Group 1 classification),
on the basis of sufficient evidence in humans and  sufficient evidence in experimental
animals—a higher classification than previous IARC evaluations. In addition, the
National Institute of Environmental Health Sciences recently nominated formaldehyde
for reconsideration as a known human carcinogen under the National Toxicology
Program. Since 1981 it has been listed as a "reasonably anticipated human carcinogen."

       In the past 15 years there has been substantial research on the inhalation
dosimetry for formaldehyde in rodents and primates by the CUT Centers for Health
Research (formerly the Chemical Industry Institute of Toxicology), with a focus on use of
rodent data for refinement of the quantitative cancer dose-response assessment.58'59'60
CIIT's risk assessment of formaldehyde incorporated mechanistic and dosimetric
information on formaldehyde. The risk assessment analyzed carcinogenic risk from
inhaled formaldehyde using approaches that are consistent with EPA's draft guidelines
for carcinogenic risk assessment. In 2001, Environment Canada relied on this cancer
dose-response assessment in their assessment of formaldehyde.61  In 2004, EPA also
relied on this cancer unit risk estimate during the development of the plywood and
composite wood products national emissions  standards for hazardous air pollutants
(NESHAPs).62  In these rules, EPA concluded that the CUT work represented the best
available application of the available mechanistic  and dosimetric science on the dose-
response for portal of entry cancers due to formaldehyde exposures. EPA is reviewing
the recent work cited above from the NCI and NIOSH, as well as the analysis by the CUT
Centers for Health Research and other studies, as part of a reassessment of the human
hazard and dose-response associated with formaldehyde.

       Noncancer effects of formaldehyde have been observed in humans and several
animal species and include irritation to eye, nose and throat tissues in conjunction with
increased mucous secretions.
                                      1-18

-------
1.3.4.   Acetaldehyde

       Acetaldehyde is formed as a result of incomplete fuel combustion in both gasoline
and diesel engines, although acetaldehyde accounts for a smaller quantity of total exhaust
hydrocarbons from gasoline engines. Acetaldehyde emissions can vary substantially by
engine duty cycle, emission control system and composition of fuel.  Acetaldehyde is not
a component of evaporative emissions but it can be formed photochemically in the
atmosphere.

       Acetaldehyde is classified in EPA's IRIS database as a probable human
carcinogen and is considered moderately toxic by inhalation.63 Based on nasal tumors in
rodents, the upper confidence limit estimate of a lifetime extra cancer risk from
continuous acetaldehyde exposure is about 2.2xlO"6 per |ig/m3. In other words, it is
estimated that about 2 persons in one million exposed to 1 |ig/m3 acetaldehyde
continuously for their lifetime (70 years) would develop cancer as a result of their
exposure although the risk could be as low as zero.

       In short-term (4 week) rat studies, compound-related histopathological changes
were observed only in the respiratory system at various concentration levels of
exposure.64'65 Data from these studies showing degeneration of the olfactory epithelium
were found to be sufficient for EPA to develop an RfC for acetaldehyde of 9 |ig/m3.
Confidence in the principal study is medium and confidence in the database is low, due to
the lack of chronic data establishing a no observed adverse effect level and due to the
lack of reproductive and developmental toxicity data. Therefore, there is low confidence
in the RfC.66 The agency is currently conducting a reassessment of risk from inhalation
exposure to acetaldehyde.

       The primary acute effect of exposure to acetaldehyde vapors is irritation of the
eyes, skin, and respiratory  tract.67 Some asthmatics have been shown to be a sensitive
subpopulation to decrements  in functional expiratory volume (FEV1  test) and broncho-
constriction upon acetaldehyde inhalation.68

1.3.5.   Acrolein

       Acrolein is found in vehicle exhaust and is formed as a result of incomplete
combustion of both gasoline and diesel fuel. It is not a component of evaporative
emissions but it can be formed photochemically from 1,3-butadiene in the atmosphere.

       EPA determined in 2003 using the 1999 draft cancer guidelines that the human
carcinogenic potential of acrolein could not be determined because the available data
were inadequate. No information was available on the carcinogenic effects of acrolein in
humans and the animal data provided inadequate evidence of carcinogenicity.

       Acrolein is an extremely volatile organic compound which possesses considerable
water solubility.69 As such, it readily absorbs into  airway fluids in the respiratory tract
when inhaled. The toxicological data base demonstrating the highly irritating nature of
                                       1-19

-------
this vapor has been consistent, regardless of test species. Acrolein is intensely irritating
to humans when inhaled, with acute exposure resulting in upper respiratory tract irritation,
mucus hypersecretion and congestion.

       Lesions to the lungs and upper respiratory tract of rats, rabbits, and hamsters
exposed to acrolein formed the basis of the reference concentrations for inhalation (RfC)
developed in 2003.70 The Agency has  developed an RfC for acrolein of 0.02 |ig/m3 and
an RfD of 0.5 ug/kg-day.71 The overall confidence in the RfC assessment is judged to be
medium and the confidence in the RfD is medium to high.

       The Agency is currently in the process of conducting an assessment of acute
exposure effects for acrolein. The intense irritancy of this carbonyl has been
demonstrated during  controlled tests in human subjects who suffer intolerable eye and
nasal mucosal sensory reactions within minutes of exposure.72

1.3.6.   Naphthalene

       Naphthalene is found in small quantities in gasoline and diesel fuels.
Naphthalene emissions have been measured in larger quantities in both gasoline and
diesel exhaust and evaporative  emissions from mobile  sources.

       In 2004, EPA released an  external review draft (External Review Draft, IRIS
Reassessment of the Inhalation Carcinogenicity of Naphthalene, U. S. EPA.
http://www.epa.gov/iris) of a reassessment of the inhalation carcinogenicity of
naphthalene.73 The draft reassessment completed external peer review in 2004 by Oak
Ridge Institute for Science and Education.74 Based on external comments, additional
analyses are being considered.  California EPA has also released a new risk assessment
for naphthalene with  a cancer unit risk  estimate of 3xlO"5 per |ig/m3.75  The California
EPA value was used in the 1999 NATA and in the analyses done for this rule. In
addition, IARC has reevaluated naphthalene and re-classified it as Group 2B: possibly
carcinogenic to humans.76  Noncancer data on hyperplasia and metaplasia in nasal tissues
form the basis of the inhalation RfC of 3 |ig/m3.77 A low to medium confidence rating
was given to this RfC, in part because it cannot be said with certainty that this RfC will
be protective for hemolytic anemia and cataracts, the more well-known human effects
from naphthalene exposure.

1.3.7.   2,2,4-Trimethylpentane

       2,2,4-Trimethylpentane is a colorless liquid hydrocarbon also known as isooctane,
isobutyltrimethylmethane,  and  TMP. Automotive exhaust and automotive evaporative
emissions are important sources of 2,2,4-trimethylpentane in the atmosphere.

       EPA is in the  process of assembling a review draft of a reassessment of its 1991
2,2,4-TMP health effects assessment in EPA's IRIS database.  The earlier document
found little conclusive evidence of specific health effects associated with 2,2,4-TMP
                   •yo
exposures in humans  . Overall, there was "inadequate information to assess
                                       1-20

-------
carcinogenic potential," in accordance with EPA's Guidelines for Carcinogen Risk
Assessment (U.S. EPA, 1986), for 2,2,4-trimethylpentane. No chronic bioassay studies
were available that assessed the carcinogenic effects of 2,2,4-trimethylpentane in humans.

       Oral studies existed linking 2,2,4-TMP with male rat kidney toxicity and an
increase in alpha2U-globulin protein and hyaline droplet accumulation in the proximal
tubules of the kidneys79.  These effects were not seen in the female rat test subjects.
Inhalation studies in animals had been performed but none were adequate to calculate an
inhalation RfC for the compound.

1.3.8.   Ethylbenzene

       Ethylbenzene is present as in both gasoline and diesel exhaust and in evaporative
emissions from gasoline-powered vehicles.80 Limited information is available on the
carcinogenic effects of ethylbenzene in humans and  animals. Under the 1987 Cancer
Guidelines, EPA has classified ethylbenzene as a Group D carcinogen, meaning it is not
classifiable as to human carcinogenicity.  This classification is the result  of inadequate
data from animal bioassays and human studies.81

       Chronic (long-term) exposure to ethylbenzene by inhalation in humans may result
in effects on the blood, kidney and liver.  No information is available  on the
developmental or reproductive effects of ethylbenzene in humans, although animal
studies have reported developmental effects via inhalation. The data from these studies
were found to be sufficient for EPA to develop an RfC of 1x103 ug/m3 for ethylbenzene
exposure. Confidence in the RfC is considered low because higher study exposure levels
might have been more informative and no chronic studies or multi-generational
developmental studies were available at the time. Animal studies have reported effects on
the blood, liver, and kidneys  from ingestion exposure to ethylbenzene. The data from
these studies were found to be sufficient for EPA to  develop an RfD for  ethylbenzene
exposure of 100 ug/kg-day. Confidence in this RfD is considered low because rats of
only one sex were tested,  no  chronic studies were then available, and no other oral
toxicity data were found.  Ethylbenzene is currently  undergoing an IRIS update for both
cancer and noncancer effects, based on new data.

       Acute (short-term) exposure to ethylbenzene in humans results in noncancer
respiratory effects, such as throat irritation and chest constriction, irritation of the eyes,
and neurological effects such as dizziness.

1.3.9.   n-Hexane

       n-Hexane is a component of gasoline and is also found in exhaust and evaporative
emissions from motor vehicles.  Monitoring data indicate that n-hexane  occurs widely  in
the atmosphere.82

       Under the 2005 Guidelines for Carcinogen Risk Assessment, there is inadequate
information to assess the carcinogenic potential of n-hexane.83 Chronic exposure to n-
                                       1-21

-------
hexane in air is associated with polyneuropathy in humans, with numbness in the
extremities, muscular weakness, blurred vision, headache, and fatigue observed.
Neurotoxic effects have also been exhibited in rats.  Mild inflammatory and degenerative
lesions in the nasal cavity have been observed in rodents chronically exposed by
inhalation.  Limited information is available on the reproductive or developmental effects
of n-hexane; one study reported testicular damage in rats exposed to n-hexane through
inhalation.  Birth defects have not been observed in the offspring of rats chronically
exposed via inhalation in several studies.  The data from a study of peripheral neuropathy
was used to develop an RfC of 700 ug/m3 for n-hexane exposure.84 This RfC has been
given a confidence rating of medium due to medium confidence in the underlying study
and medium confidence in the database. The database lacks chronic exposure
information on the pure compound via any route of exposure, a multigenerational
developmental and reproductive toxicity study and a developmental neurotoxicity study.

       Acute inhalation exposure of humans to high levels of n-hexane causes mild
central nervous system (CNS) depression  and irritation of the skin and mucous
membranes. Nervous system effects include dizziness, giddiness, slight nausea, and
headache in humans.

1.3.10.  Methyl Tertiary Butyl Ether (MTBE)

       Methyl tert-buty\ ether (MTBE) has been used in the United States since the late-
1970's as an octane-enhancing agent in gasoline.

       In 1994, EPA's Office of Research and Development concluded that, under the
1986 EPA cancer risk assessment guidelines, inhalation cancer test results support
placing MTBE in Group C as a "possible human carcinogen."85  An Interagency
Assessment of Oxygenated Fuels similarly concluded that "While there are no studies on
the carcinogenicity of MTBE in humans, there is sufficient evidence to indicate that
MTBE is an animal carcinogen and to regard MTBE as having a human hazard potential.
However, estimates of human risk from MTBE contain large uncertainties in both human
exposure and cancer potency."86 The Agency is currently conducting a reassessment of
MTBE.

       By the inhalation route, MTBE has been found to cause increases in liver and
kidney weights and increased severity of spontaneous kidney lesions, as well as swelling
around the eyes and increased prostration  in laboratory rats87. These effects are cited as
the basis for EPA's current inhalation reference concentration (RfC) of 3  mg/m3 for
MTBE.  The RfC has a medium to high confidence  rating.

1.3.11.  Styrene

       Styrene is found in the exhaust from both gasoline- and diesel-powered engines.
Several epidemiologic studies suggest that there may be an association between styrene
exposure and an increased risk of leukemia and lymphoma. However, the evidence is
inconclusive due to confounding factors.  Animal studies have produced both negative
                                      1-22

-------
and positive results. EPA is currently assessing the potential of styrene to cause cancer.

       Chronic exposure of humans to styrene results in effects on the central nervous
system (CNS), such as headache, fatigue, weakness, depression, peripheral neuropathy,
minor effects on some kidney enzyme functions and on the blood. Human studies are
inconclusive on the reproductive and developmental effects of styrene. The data from
human studies looking at central nervous system effects was found to be sufficient for
EPA to develop an RfC of 1 mg/m3 for styrene exposure. The RfC is assigned an overall
confidence rating of medium. Data from animal oral exposure studies was found to be
sufficient for EPA to also develop an RfD of 200 ug/kg-day for styrene oral exposure.
The RfD is assigned an overall confidence rating of medium.

       Acute exposure to styrene results in mucous membrane and eye irritation, and
                                      88 SQ
central nervous system effects in humans.  '

1.3.12.  Toluene

       Toluene is found in evaporative as well as exhaust emissions from motor vehicles.
Under the 2005 Guidelines for Carcinogen Risk Assessment, there is inadequate
information to assess the carcinogenic potential of toluene because studies of humans
chronically exposed to toluene are inconclusive and animal studies have generally been
negative.  90

       The central nervous system (CNS) is the primary target for toluene toxicity in
both humans and animals for acute and chronic exposures. CNS dysfunction (which is
often reversible) and narcosis have been frequently  observed in humans acutely exposed
to low or moderate levels of toluene by inhalation; symptoms include fatigue, sleepiness,
headaches, and nausea. Central nervous system depression has been reported to occur in
chronic abusers exposed to high levels of toluene.  Symptoms include ataxia,  tremors,
cerebral atrophy, nystagmus (involuntary eye movements), and impaired speech, hearing,
and vision. Chronic inhalation exposure of humans to toluene also causes irritation of the
upper respiratory tract, eye irritation, dizziness, headaches, and difficulty with sleep.91

       Human studies have also reported developmental effects, such as CNS
dysfunction, attention deficits, and minor craniofacial and limb anomalies, in the children
of women who abused toluene during pregnancy. A substantial database examining the
effects of toluene in subchronic and chronic occupationally exposed  humans exists.  The
weight of evidence from these studies indicates neurological effects (i.e., impaired color
vision, impaired hearing, decreased performance in neurobehavioral  analysis, changes in
motor and sensory nerve conduction velocity,  headache, dizziness) as the most sensitive
endpoint.  The data from these human studies was found to be sufficient for EPA to
develop an RfC of 5 mg/m3 for toluene exposure. The overall confidence in this RfC is
high. Additional data from animal oral exposure studies was found to be sufficient for
EPA to also develop an RfD of 80 ug/kg-day for toluene oral exposure.92 The overall
confidence in the RfD is medium.
                                      1-23

-------
1.3.13.  Xylenes

       Mixed xylenes are blended into gasoline and are present in diesel fuels. Xylenes
are emitted in the exhaust emissions and evaporative emissions of both gasoline- and
diesel-powered engines.

       Inadequate information is available on the carcinogenic effects of mixed xylenes
in humans, and animal studies have been inconclusive. Under the 1999 Draft Revised
Guidelines for Carcinogen Risk Assessment, data are inadequate for an assessment of the
carcinogenic potential of xylenes.93

       Chronic inhalation exposure in humans to mixed xylenes results primarily in
central nervous system effects, such as headache, nausea, fatigue and also included eye
and nose irritation and sore throat.94  Animal studies have reported developmental effects,
such as an increased incidence of skeletal variations in fetuses, and fetal resorptions via
inhalation.  EPA developed an RfC of 100 ug/m3 for xylenes based on impaired motor
coordination in rats. The confidence rating assigned to the RfC for xylenes is medium.
Data from animal oral exposure studies, looking at decreased body weight and increased
mortality were found to be sufficient for EPA to develop an RfD of 200 ug/kg-day for
oral xylene exposure. The RfD was assigned an overall confidence rating of medium.95

       Acute inhalation exposure to mixed xylenes in humans results in irritation of the
nose and throat, gastrointestinal effects such as nausea, vomiting, and gastric irritation,
mild transient eye irritation, and neurological effects.

1.3.14.  Poly cyclic Organic Matter (POM)

       POM is a class of chemicals consisting of organic compounds having multiple
benzene rings and boiling points in excess of 100 degrees Celsius. POM is a byproduct
of the incomplete combustion of fossil fuels and, as such, is a component of diesel and
gasoline engine emissions. Many of the compounds included in the class of compounds
known as POM are classified by EPA as probable human carcinogens based on animal
data.  One of these compounds, naphthalene, is discussed separately in this section.

       Recent studies have found that maternal exposures to polyaromatic hydrocarbons
(PAHs), a subclass of POM, in a population of pregnant women were associated with
several adverse birth outcomes, including low birth weight and reduced length at birth.96
These studies are discussed later in Chapter 3.
                                       1-24

-------
1.3.15.  Diesel Particulate Matter (DPM) and Diesel Exhaust Organic Gases
     (DEOG)

       In EPA's Diesel Health Assessment Document (HAD),97 diesel exhaust was
classified as likely to be carcinogenic to humans by inhalation at environmental
exposures, in accordance with the revised draft 1996/1999 EPA cancer guidelines. A
number of other agencies (National Institute for Occupational Safety and Health, the
International Agency for Research on Cancer, the World Health Organization, California
EPA, and the U.S. Department of Health and Human Services) have made similar
classifications. EPA concluded in the Diesel HAD that it is not possible currently to
calculate a cancer unit risk for diesel exhaust due to a variety of factors that limit the
current studies, such as limited quantitative exposure histories in occupational groups
investigated for lung cancer.

       However, in the absence of a cancer unit risk, the EPA Diesel HAD  sought to
provide additional insight into the significance of the cancer hazard by estimating
possible ranges of risk that might be present in the population. The possible risk range
analysis was developed by  comparing a typical environmental exposure level for
highway diesel sources to a selected range of occupational exposure levels.  The
occupationally observed risks were then proportionally scaled according to the exposure
ratios to obtain an estimate of the possible environmental risk. A number of calculations
are needed to accomplish this, and these can be seen in the EPA Diesel HAD.  The
outcome was that environmental risks from diesel exhaust exposure could range from a
low of 10"4 to 10"5 to as high as 10"3, reflecting the range of occupational exposures that
could be associated with the relative and absolute risk levels observed in the occupational
studies. Because of uncertainties, the analysis acknowledged that the risks could be
lower than 10"4 or 10"5, and a zero risk from diesel exhaust exposure was not ruled out.

       The acute and chronic exposure-related effects  of diesel exhaust emissions are
also of concern to the Agency.  EPA derived an RfC from consideration of four well-
conducted chronic rat inhalation studies showing adverse pulmonary effects.98' "' 100' 101
The RfC is 5 |ig/m3  for diesel exhaust as measured by diesel PM. This RfC does not
consider allergenic effects such as those associated with asthma or immunologic effects.
There is growing evidence, discussed in the Diesel HAD, that diesel exhaust can
exacerbate these  effects, but the exposure-response data are presently lacking to derive an
RfC.

       The Diesel HAD also briefly summarizes health effects associated with ambient
PM and the EPA's annual National Ambient Air Quality Standard (NAAQS) of 15 |ig/m3.
There is a much more extensive body of human data showing a wide spectrum of adverse
health effects associated with exposure to ambient PM, of which diesel exhaust is an
important component.  The  RfC is not meant to say that 5 |ig/m3 provides adequate public
health protection for ambient PM2.5.  In fact, there may be benefits to reducing diesel PM
below 5 |ig/m3 since diesel PM is a major contributor to ambient PM2.5.
                                       1-25

-------
1.4.          Emerging Issues

       Beyond the specific areas of quantifiable risk discussed above in Chapter 1.1.2,
EPA is interested in emerging mobile source toxics issues that might require action in the
future. The emerging issues currently under investigation by EPA are gasoline PM and
metals.

1.4.1.   Gasoline PM

       Gasoline exhaust is a complex mixture that has not been evaluated in EPA's IRIS,
in contrast to diesel exhaust, which has been evaluated in IRIS. However, there is
evidence for the mutagenicity and cytotoxicity of gasoline exhaust and gasoline PM.
Seagrave et al. investigated the combined particulate and semivolatile organic fractions of
gasoline and diesel engine emissions.102 Their results demonstrate that emissions from
gasoline engines are mutagenic and can induce inflammation and have cytotoxic effects.
Gasoline exhaust is a ubiquitous source of parti culate matter, contributing to the health
effects observed for ambient PM which is discussed extensively in the EPA Particulate
Matter Criteria Document.103  The PM Criteria Document notes that the PM components
of gasoline and diesel engine exhaust are hypothesized, important contributors to the
observed increases in lung cancer incidence and mortality associated with ambient
PM2 5.104 Gasoline PM is also a component of near-roadway emissions that may be
contributing to the  health effects observed in people who live near roadways (see Chapter
3.1.3.1).

       EPA is working to improve the understanding of PM emissions from gasoline
engines, including the potential range of emissions and factors that influence emissions.
EPA led a cooperative test program that recently completed testing approximately  500
randomly procured vehicles in the Kansas City metropolitan area. The purpose of this
study was to determine the distribution of gasoline PM emissions from the in-use light-
duty fleet. Results from this study are expected to be available in 2006.  Some  source
apportionment studies show gasoline and diesel PM can result in larger contributions to
ambient PM than predicted by EPA emission inventories.105'106 These source
apportionment studies were one impetus behind the Kansas City  study.

       Another issue related to gasoline PM is the effect of mobile source on  ambient
PM, especially secondary PM. Ambient PM is composed of primary PM emitted directly
into the atmosphere and secondary PM that is formed from chemical reactions in the
atmosphere.  Sulfates and nitrates are major examples of inorganic secondary PM, both of
which have been well studied  and quantified.  Carbonaceous PM, from both primary PM
emissions and  secondary PM formed in the atmosphere, is a major source of PM,
especially in urban areas. Various studies show that carbonaceous PM specifically from
mobile sources is a major PM constituent in many urban areas over many portions of the
country (including  urban areas in the Northeast,  Southeast, Midwest, and
California/Washington portions of the United States). This information is included in
EPA reports and various source apportionment studies.107'108'109'110'111'112'113
                                       1-26

-------
       Primary carbonaceous mobile source emissions can be evaluated from emission
inventories. The ambient PM levels from these emissions and secondary PM formed in
the atmosphere from mobile sources can then be estimated by air quality modeling
studies using the CMAQ (Community Multi-scale Air Quality) model.  In addition to
primary carbonaceous (organic aerosol) emissions, some specific compounds contribute
to atmospheric PM loadings via formation of secondary organic aerosols (SOA).  These
compounds include monoterpenes and possibly isoprene and sesquiterpenes, as well as
anthropogenic aromatic hydrocarbons (and probably higher molecular weight non-
aromatic hydrocarbons).

       Also, there is strong indication that benzene forms SOA.  In other ongoing
research, EPA scientists are investigating SOA formation from benzene, which has been
recently detected for the first time in European smog chamber experiments.114

       Upon release into the atmosphere, these numerous compounds can react with free
radicals in the atmosphere to form SOA. While SOA formation from many reactive
hydrocarbons has been investigated in the laboratory, there is relatively little information
available on the chemical composition of SOA compounds from specific hydrocarbon
precursors. This lack of information is largely due to having few reliable methods for
measuring the polar, high molecular weight compounds that are thought to make up much
of ambient SOA. The absence of compositional data has largely prevented identifying
aromatically-derived SOA in ambient samples which, in turn, has prevented observation-
based measurements of the aromatic and other SOA contributions to ambient PM levels.

       Recently EPA has taken the first step in addressing these issues by developing a
tracer-based method for detecting SOA  precursors in ambient samples.  The method
consists of irradiating the SOA precursor of interest in a smog chamber in the presence of
NOx, collecting the SOA produced on filters, and then analyzing the  samples for highly
polar compounds using advanced organic chemistry methods. Employing this method,
candidate tracers have been identified for several hydrocarbon compounds which are
emitted in significant quantities and known to produce SOA in the atmosphere.  Some of
these compounds forming SOA that have been investigated in the current study are
toluene, a variety of monoterpenes, isoprene,  and /?-caryophyllene, the latter three of
which are emitted by vegetation. The tracers provide a means to identify the
hydrocarbon SOA precursors present in ambient PM2.5 samples and show promise for
estimating their contributions to the organic carbon concentrations.

       The results of a recent EPA field study, not yet published in the peer-reviewed
literature, suggest aromatic hydrocarbon emissions, including toluene and possibly
xylenes, contribute to SOA in Research Triangle Park, North Carolina, with initial
estimates as high as 0.7 ug/m3  during smog events in July/August.  The level of toluene-
derived SOA is the lowest in the November-February time frame (0.2 ug/m3) with
intermediate levels in the other months.  Currently, EPA is conducting similar analyses of
ambient PM2.5 samples in Cincinnati, OH, Northbrook, IL, Detroit, MI, Bondville, IL,
and St. Louis, MO, the results of which  will be available by the end of 2006. After
acceptance of the EPA field study results in the peer-reviewed literature, they will used to
                                      1-27

-------
assess whether current treatment of aromatic SOA in the EPA CMAQ model, which
along with most of the other state of the science air quality models, predict low levels of
aromatic SOA, need to be modified.

       One caveat regarding this work is that a large number of gaseous hydrocarbons
emitted into the atmosphere having the potential to form SOA have not yet been studied
in this way. It is possible that hydrocarbons which have not yet been studied produce
some of SOA species which are being used as tracers for other gaseous hydrocarbons.
This means that the present work could over-estimate the amount of SOA in the
atmosphere to the gaseous hydrocarbons studied to date.

       The issue of SOA formation from aromatic precursors is an important one to
which EPA and others are paying significant attention.  Due to the large contribution of
mobile source emissions to overall aromatic levels in the atmosphere, this issue is a
crucial one for assessing what further reductions are possible  in mobile source PM.

1.4.2.   Metals

       The emission of metals to the environment is receiving increasing attention.
Metals comprise a complex class of elements, some of which are toxic at very low
exposure levels. The chemical form in which a metal or metal compound is emitted often
determines the potential toxicity and ultimate fate of the element in the environment.
Research in recent years suggests that some metals (e.g., transition metals) play an
important role in the toxicity of ambient PM, and inhalation as well as ingestion of metals
is known to cause a diverse array of cancer and noncancer effects in mammals. Since
metals do not degrade in the environment, concerns arise regarding their accumulation in
plants, animals, soil and water.  The emission of metals from mobile sources is an
emerging area of interest since the emissions are in the breathing zone and are distributed
in a concentrated fashion in the roadway environment.

       Emission of metals from mobile sources occurs as the result of metallic impurities
in lubricating oil and fuel, catalyst wear, engine wear, brake wear, and tire wear.
Emission rates of most metals from mobile sources are quite low, presenting challenges
for many common measurement methods. In recent years, improvements in analytical
chemistry allow both the quantification of very low levels of metals in mobile source
exhaust as well as some characterization of the form of the metals emitted. Currently,
there are many gaps in our understanding of the quantity, chemical form and size
distribution of metals in exhaust or from tire and brake wear.  Application of state-of-the-
art measurement techniques to mobile source metal emissions is just beginning. For
example, EPA is currently conducting an emissions characterization program to
understand the emission rate and chemical form of mercury in motor vehicle exhaust and
the total mercury concentration in gasoline, diesel fuel, lubricating oil, and brake wear
emissions. This work will help us understand the potential sources of motor vehicle
mercury emissions, and the contribution of motor vehicles relative to other sources of
mercury emissions. This information is necessary for any future consideration of control
options.
                                       1-28

-------
       Metals can also be emitted from mobile sources as a result of their use as an
additive to gasoline and/or diesel fuel. As discussed in Chapter III.G of the preamble,
Clean Air Act section 211 provides EPA with the authority to require a fuel additive
manufacturer to collect necessary data to enable EPA to make a determination about the
potential for risk to public health.
                                       1-29

-------
References for Chapter 1
1 Rao, S.; Pollack, A.; Lindhjem, C. 2004 Expanding and updating the master list of
compounds emitted by mobile sources - Phase III Final Report. Environ International
Corporation.

2 www.epa.gov/ttn/atw/toxsource/summary.html.  Tables of dose-response values on this
website, used in EPA Office of Air Quality Planning and Standards risk assessments, are
available in Docket EPA-HQ-OAR-2005-0036.

3 National Academy of Sciences.  1983. Risk Assessment in the Federal Government:
Managing the Process. Committee on the Institutional Means for Assessment of Risks to
Public Health, National Research Council. Found in various EPA library collections
through http://www.epa.gov/natlibra/ols.htm by its OCLC catalog no.09374015.

4  EPA.  1986. Guidelines for carcinogen risk assessment. Federal Register 51:33992-
34003. September 24.

5 EPA. 2005. Guidelines for carcinogen risk assessment and Supplemental Guidance for
Assessing Susceptibility from Early-Life Exposure to Carcinogens. EPA/630/P-03/001F.

6  EPA.  1986 Guidelines for carcinogen risk assessment. Federal Register 51:33992-
34003. September 24.

7 U. S. EPA. 2005 Supplemental Guidance for Assessing Susceptibility from Early-Life
Exposure to Carcinogens. Report No. EPA/630/R-03/003F.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid= 116283

8  EPA. 1986. Guidelines for mutagenicity risk assessment. Federal Register 51:34006-
34012. Sept. 24.

9  EPA. 1991. Guidelines for developmental toxicity risk assessment. Federal Register
56:63798-63826.

10  EPA.  1998. Guidelines for neurotoxicity risk assessment. Federal Register 63:26926.
May 14.  http://www.epa.gov/ncea/nurotox.htm

11  EPA.  1996. Guidelines for reproductive toxicity risk assessment. EPA/630/R-96/009.
Federal Register 56274-56322, 31 October 1996.
http ://www. epa. gov/ORD/WebPub s/repro/

12  EPA. 1994. Methods for derivation of inhalation reference concentrations and
application of inhalation dosimetry.  Washington D.C. EPA/600/8-90/066F.

13 EPA (2002) A review of the reference dose and reference concentration processes.
EPA/630/P-02/002F.
                                      1-30

-------
14  EPA (1998) Methods for exposure-response analysis for acute inhalation exposure to
chemicals: development of the acute reference exposure.  Review draft.  Office of
Research and Development, Washington, D.C. EPA/600/R-98/051.

15 EPA (2002) A review of the reference dose and reference concentration processes.
EPA/630/P-02/002F.

16 EPA 2005 "Full IRIS Summary for Benzene (CASRN 71-43-2)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0276.htm.

17 U.S. EPA (1985) Environmental Protection Agency, Interim quantitative cancer unit
risk estimates due to inhalation of benzene, prepared by the Office of Health and
Environmental Assessment, Carcinogen Assessment Group, Washington, DC. for the
Office of Air Quality Planning and  Standards, Washington, DC.,  1985. Document no.
EPA600-X-85-022.

18 Clement Associates, Inc. (1991) Motor vehicle air toxics health information, for U.S.
EPA Office of Mobile Sources, Ann Arbor, MI, September 1991. Available in Docket
EPA-HQ-OAR-2005-0036.

19 Hayes, R. B., S.  N. Yin, M. S. Dosemici, et al.  (1997) Benzene and the dose-related
incidence of hematological neoplasms in China. J. Nat. Cancer Inst. 89:1065-1071.

20 International Agency for Research on Cancer (IARC) (1982) IARC monographs on the
evaluation of carcinogenic risk of chemicals to humans, Volume 29, Some industrial
chemicals and dyestuffs, International Agency for Research on Cancer, World Health
Organization, Lyon, France, p. 345-389.

21 U.S. EPA (1998) Environmental Protection Agency, Carcinogenic Effects of Benzene:
An Update, National Center for Environmental Assessment, Washington, DC. EPA600-
P-97-001F. http://www.epa.gov/ncepihom/Catalog/EPA600P97001F.html


22 Irons, R.D., W.S. Stillman, D.B. Colagiovanni, and V.A. Henry (1992) Synergistic
action of the benzene metabolite hydroquinone on myelopoietic stimulating activity of
granulocyte/macrophage colony-stimulating factor in vitro, Proc.  Natl. Acad. Sci.
89:3691-3695.

23 U.S. EPA (1998) Environmental Protection Agency, Carcinogenic  Effects of Benzene:
An Update, National Center for Environmental Assessment, Washington, DC. EPA600-
P-97-001F. http://www.epa.gov/ncepihom/Catalog/EPA600P97001F.html

24 U.S. EPA (1998) Environmental Protection Agency, Carcinogenic  Effects of Benzene:
                                      1-31

-------
An Update, National Center for Environmental Assessment, Washington, DC. EPA600-
P-97-001F. http://www.epa.gov/ncepihom/Catalog/EPA600P97001F.html

25 U. S. EPA (2005) Guidelines for Carcinogen Risk Assessment. Report No.
EPA/630/P-03/001F.  http://cfpub.epa.gov/ncea/raf/recordisplav.cfm?deid=l 16283.

26
27
  U.S. EPA (1998) Carcinogenic Effects of Benzene: An Update. EPA/600/P-97/001F.

  Rothman, N; Li, GL; Dosemeci, M; et al. (1996) Hematotoxicity among Chinese
workers heavily exposed to benzene. Am J Ind Med 29:236-246.

28 Rappaport, S.M.; Waidyanatha, S.; Qu, Q.; Shore, R.; Jin, X.; Cohen, B.; Chen, L.;
Melikian, A.; Li, G; Yin, S.; Yan, H.; Xu, B.; Mu, R.; Li, Y.; Zhang, X.; and Li, K.
(2002) Albumin adducts of benzene oxide and 1,4-benzoquinone as measures of human
benzene metabolism. Cancer Research.  62:1330-1337.

29 Hayes, R.B.; Yin, S.; Dosemeci, M.; Li, G.; Wacholder, S.; Travis, L.B.; Li, C.;
Rothman, N.; Hoover, R.N.; and Linet, M.S. (1997) Benzene and the dose-related
incidence of hematologic neoplasms in China. J. Nat. Cancer Inst. 89:1065-1071.

30 Hayes, R.B.; Songnian, Y.; Dosemeci, M.; and Linet, M. (2001) Benzene and
lymphohematopoietic malignancies in humans. Am J Indust Med, 40:117-126.

31 Shu, X.O,; Gao, Y.T.; Brinton, L.A.; et al. (1988) A population-based case-control
study of childhood leukemia in Shanghai. Cancer 62:635-644.

32 McKinney P.A.; Alexander, F.E.; Cartwright, R.A.; et al. (1991) Parental occupations
of children with leukemia in west Cumbria, north Humberside,  and Gateshead, Br Med J
302:681-686.

33 Keller, KA; Snyder, CA. (1986) Mice exposed in utero to low concentrations of
benzene exhibit enduring changes in their colony forming hematopoietic cells.
Toxicology 42:171-181.

34 Keller, KA; Snyder, CA. (1988) Mice exposed in utero to 20 ppm benzene exhibit
altered numbers of recognizable hematopoietic cells up to seven weeks after exposure.
Fundam Appl Toxicol 10:224-232.

35 Corti, M; Snyder, CA. (1996) Influences of gender, development, pregnancy and
ethanol consumption on the hematotoxicity of inhaled 10 ppm benzene. Arch Toxicol
70:209-217.

36 U. S. EPA.  (2002). Toxicological Review of Benzene (Noncancer Effects).  National
Center for Environmental Assessment, Washington, DC.  Report No. EPA/635/R-
02/00IF. http://www.epa.gov/iris/toxreviews/0276-trflj.pdf
                                      1-32

-------
37 Ford, AM; Pombo-de-Oliveira, MS; McCarthy, KP; MacLean, JM; Carrico, KC;
Vincent, RF; Greaves, M. (1997) Monoclonal origin of concordant T-cell malignancy in
identical twins. Blood 89:281-285.

38 Aksoy, M. (1989) Hematotoxicity and carcinogenicity of benzene.  Environ. Health
Perspect. 82: 193-197.

39 Goldstein, B.D. (1988) Benzene toxicity. Occupational medicine. State of the Art
Reviews. 3: 541-554.

40 Aksoy, M. (1991) Hematotoxicity, leukemogenicity and carcinogenicity of chronic
exposure to benzene. In: Arinc, E.; Schenkman, J.B.; Hodgson, E., Eds. Molecular
Aspects of  Monooxygenases and Bioactivation of Toxic Compounds. New York:
Plenum Press, pp. 415-434.

41 Goldstein, B.D. (1988) Benzene toxicity. Occupational medicine. State of the Art
Reviews. 3: 541-554.

42 Aksoy, M., S. Erdem, and G. Dincol. (1974) Leukemia in shoe-workers exposed
chronically to benzene.  Blood 44:837.

43 Aksoy, M. and K. Erdem. (1978) A follow-up study on the mortality and the
development of leukemia in 44 pancytopenic patients associated with long-term exposure
to benzene. Blood 52: 285-292.

44 Rothman, N., G.L. Li, M. Dosemeci, W.E. Bechtold, G.E.  Marti, Y.Z. Wang,  M. Linet,
L.Q. Xi, W. Lu, M.T. Smith, N. Titenko-Holland, L.P. Zhang, W. Blot, S.N. Yin, and
R.B. Hayes (1996) Hematotoxicity among Chinese workers heavily exposed to benzene.
Am. J. Ind.  Med. 29: 236-246.

45 EPA 2005 "Full IRIS Summary for Benzene (CASRN 71-43-2)" Environmental
Protection Agency,  Integrated Risk Information System (IRIS),  Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0276.htm.

46 Qu, Q., R. Shore, G.  Li, X. Jin, L.C. Chen, B. Cohen,  et al. (2002). Hematological
changes among Chinese workers with a broad range of benzene exposures. Am. J.
Industr. Med. 42: 275-285.

47 Lan, Qing, Zhang, L., Li, G., Vermeulen, R., et al. (2004). Hematotoxically in
Workers Exposed to Low Levels of Benzene.  Science 306: 1774-1776.

47 Turtletaub, K.W.  and Mani, C. (2003). Benzene metabolism in rodents at doses
relevant to human exposure from Urban Air. Research Reports  Health Effect Inst. Report
No.113.
                                      1-33

-------
49 U.S. EPA. (2002). Health Assessment of 1,3-Butadiene. Office of Research and
Development, National Center for Environmental Assessment, Washington Office,
Washington, DC. Report No. EPA600-P-98-001F at
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=54499

50 U.S. EPA (1998) A Science Advisory Board Report: Review of the Health Risk
Assessment of 1,3-Butadiene. EPA-SAB-EHC-98.

51 Delzell, E., Sathiakumar, N., Hovinga, M., Macaluso, M., Julian, J., Larson, R., Cole,
P. and Muir, D.C.F., 1996. A Follow-up Study of Synthetic Rubber Workers. Toxicology,
113, 182-189.

52 EPA 2005 "Full IRIS Summary for 1,3-butadiene (CASRN 106-99-0)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0139.htm.

53 Bevan, C.; Stadler, J.C.; Elliot, G.S.; et al. (1996) Subchronic toxicity of 4-
vinylcyclohexene in rats and mice by inhalation.  Fundam. Appl. Toxicol. 32:1-10.

54 U.S. EPA (1987) Assessment of Health Risks to Garment Workers and Certain Home
Residents from Exposure to Formaldehyde, Office of Pesticides and Toxic Substances,
April 1987. Found in various EPA library collections through
http://www.epa.gov/natlibra/ols.htm by its OCLC catalog no. 16989049.

55 Hauptmann, M..; Lubin, J. H.; Stewart, P. A.; Hayes, R. B.; Blair, A. 2003.  Mortality
from lymphohematopoetic malignancies among workers in formaldehyde industries.
Journal of the National Cancer Institute 95: 1615-1623.

56 Hauptmann, M..; Lubin, J. H.; Stewart, P. A.; Hayes, R. B.; Blair, A. 2004.  Mortality
from solid cancers among workers in formaldehyde industries. American Journal of
Epidemiology  159: 1117-1130.

57 Pinkerton, L. E. 2004. Mortality among a cohort of garment workers exposed to
formaldehyde: an update. Occup.  Environ. Med. 61: 193-200.

58 Conolly,  RB, JS Kimbell, D Janszen, PM Schlosser, D Kalisak, J Preston, and FJ
Miller. 2003.  Biologically motivated computational modeling of formaldehyde
carcinogenicity in the F344 rat.  Tox Sci 75: 432-447.

59 Conolly,  RB, JS Kimbell, D Janszen, PM Schlosser, D Kalisak, J Preston, and FJ
Miller. 2004. Human respiratory tract cancer risks of inhaled formaldehyde: Dose-
response predictions derived from biologically-motivated computational modeling of a
combined rodent and human dataset. Tox Sci 82: 279-296.

60 Chemical Industry Institute of Toxicology (CUT). 1999. Formaldehyde: Hazard
                                      1-34

-------
characterization and dose-response assessment for carcinogenicity by the route of
inhalation.  CUT, September 28, 1999. Research Triangle Park, NC.

61 Health Canada (2001) Priority Substances List Assessment Report. Formaldehyde.
Environment Canada, Health Canada, February 2001. The document may be accessed at
http://www.hc-sc.gc.ca/ewh-semt/pubs/contaminants/psl2-
Isp2/formaldehyde/index e.html .

62 U.S. EPA (2004) National Emission Standards for Hazardous Air Pollutants for
Plywood and Composite Wood Products Manufacture: Final Rule. (69 FR 45943,
7/30/04)

63 EPA 2005 "Full IRIS Summary for Acetaldehyde (CASRN 75-07-0)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0290.htm.

64 Appleman, L. M., R. A. Woutersen, V. J. Feron, R. N. Hooftman, and W. R. F. Notten.
(1986). Effects of the variable versus fixed exposure levels on the toxicity of
acetaldehyde in rats. J. Appl. Toxicol. 6: 331-336.

65 Appleman, L.M., R.A. Woutersen, and VJ. Feron. (1982). Inhalation toxicity of
acetaldehyde in rats. I. Acute and subacute studies.  Toxicology. 23: 293-297.

66 EPA 2005 "Full IRIS Summary for Acetaldehyde (CASRN 75-07-0)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0290.htm.

67 EPA 2005 "Full IRIS Summary for Acetaldehyde (CASRN 75-07-0)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0290.htm.

68 Myou, S.; Fujimura, M.; Nishi K.; Ohka, T.; and Matsuda, T. (1993) Aerosolized
acetaldehyde induces histamine-mediated bronchoconstriction  in asthmatics.  Am Rev
RespirDis 148(4 Pt 1): 940-3.

69 Agency for Toxic Substances and Disease Registry (ATSDR). Toxicological Profile
for Acrolein. Public Health Service, U.S. Department of Health and Human Services,
Atlanta, GA. 2005. Publication # PB/91/180307/AS at
http://www.atsdr.cdc. gov/toxprofiles/tp 124. html

70 EPA 2005 "Full IRIS Summary for Acrolein (CASRN 107-02-8)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
                                      1-35

-------
OH http://www.epa.gov/iris/subst/0364.httn.

71 EPA 2005 "Full IRIS Summary for Acrolein (CASRN 107-02-8)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0364.htm.

72 Sim VM, Pattle RE. Effect of possible smog irritants on human subjects JAMA165:
1980-2010, 1957.

73 U. S. EPA. (2004) External Review Draft, IRIS Reassessment of the Inhalation
Carcinogenicity of Naphthalene,  http://www.epa.gov/iris

74 Oak Ridge Institute for Science and Education.  (2004) External Peer Review for the
IRIS Reassessment of the Inhalation Carcinogenicity of Naphthalene.  August 2004.
http://cfpub2.epa.gov/ncea/cfm/recordisplay.cfm?deid=86019

75 California EPA (2004) Long Term Health Effects of Exposure to Naphthalene.  Office
of Environmental Health Hazard Assessment at
http: //www. oehha. ca. gov/air/toxi c_contaminants/draftnaphth. html

76 International Agency for Research on Cancer  (IARC) (2002) Monographs on the
Evaluation of the Carcinogenic Risk of Chemicals for Humans. Vol. 82. Lyon, France.

77 EPA 2005 "Full IRIS Summary for Naphthalene (CASRN 91-20-3)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0436.htm.

78 EPA 2005 "Full IRIS Summary for 2,2,4-Trimethylpentane (CASRN 540-84-1)"
Environmental Protection Agency, Integrated Risk Information System (IRIS), Office of
Health and Environmental Assessment, Environmental Criteria and Assessment Office,
Cincinnati, OH http://www.epa.gov/iris/subst/0614.htm.

79 Hazardous Substances Data Bank (HSDB) on Isooctane 2005. National Library of
Medicine Bethesda, MD found at http://toxnet.nlm.nih.gov/index.html

80 ATSDR (1999) Toxicological Profile for Ethylbenzene (update). USDHHS, PHS,
ATSDR. Publication PB/99/166647 at http://www.atsdr.cdc.gov/toxprofiles/tp 110.html.

81 EPA 2005 "Full IRIS Summary for Ethylbenzene (CASRN 100-41-4)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0051 .htm.
                                      1-36

-------
82ATSDR. 1999. Toxicological Profile for n-Hexane.  USDHHS, PHS, ATSDR.
Publication# PB/99/166688 at http://www.atsdr.cdc.gov/toxprofiles/tpll3.html.

83 EPA 2005 "Full IRIS Summary for n-Hexane (CASRN 110-54-3)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0486.htm.

84 EPA 2005 "Full IRIS Summary for n-Hexane (CASRN 110-54-3)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0486.htm.

85 EPA. 1994. Health risk perspectives on fuel oxygenates. Washington, DC: Office of
Research and Development; report no. EPA 600/R-94/217.

86 Interagency Oxygenated Fuels Assessment Steering Committee. 1997. Interagency
assessment of oxygenated fuels. Washington, DC: National Science and Technology
Council, Committee on Environment and Natural Resources and Office  of Science and
Technology Policy, http://www.epa.gov/otaq/fuels.html.

87 EPA 2005 "Full IRIS Summary for Methyl-tertiary-butyl Ether (MTBE, CASRN 1634-
04-4)" Environmental Protection Agency, Integrated Risk Information System (IRIS),
Office of Health and Environmental Assessment, Environmental Criteria and Assessment
Office, Cincinnati, OH http://www.epa.gov/iris/subst/0545.htm.

88 EPA 2005 "Full IRIS Summary for Styrene (CASRN 100-42-5)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0104.htm.

89 Agency for Toxic Substances Disease Registry (1992) Toxicological profile for
styrene. Atlanta: ATDSR.

90 EPA 2005 "Full IRIS Summary for Toluene (CASRN 108-88-3)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0118.htm.

91 EPA 2005 "Full IRIS Summary for Toluene (CASRN 108-88-3)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0118.htm.
92 EPA 2005 "Full IRIS Summary for Toluene (CASRN 108-88-3)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
                                      1-37

-------
OH http://www.epa.gov/iris/subst/0118.htm.
93 EPA 2005 "Full IRIS Summary for Xylenes (CASRN 1330-20-7)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0270.htm.
94 EPA Toxicological Review of Xylenes. January 2003. EPA 635/R-03/001.

95 EPA 2005 "Full IRIS Summary for Xylenes (CASRN 1330-20-7)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0270.htm.
96 Perera, F.P.; Rauh, V.; Tsai, W-Y.; et al. (2002) Effect of transplacental exposure to
environmental pollutants on birth outcomes in a multiethnic population. Environ Health
Perspect 111:201-205.

97 U.S. EPA (2002) Health Assessment Document for Diesel Engine Exhaust.
EPA/600/8-90/057F Office of Research and Development, Washington DC. This
document is available electronically at
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=29060 .

98 Ishinishi, N; Kuwabara, N; Takaki, Y; et al. (1988) Long-term inhalation experiments
on diesel exhaust. In: Diesel exhaust and health risks. Results of the HERP studies.
Ibaraki, Japan: Research Committee for HERP Studies; pp. 11-84.

99 Heinrich, U; Fuhst, R; Rittinghausen, S; et al. (1995) Chronic inhalation exposure of
Wistar rats and two different  strains of mice to diesel engine exhaust, carbon black, and
titanium dioxide. Inhal Toxicol 7:553-556.

100 Mauderly, JL; Jones, RK;  Griffith, WC; et al. (1987) Diesel exhaust is a pulmonary
carcinogen in rats exposed chronically by inhalation. Fundam Appl Toxicol 9:208-221.

101 Nikula, KJ; Snipes, MB; Barr, EB;  et al. (1995) Comparative pulmonary toxicities and
carcinogenicities of chronically inhaled diesel exhaust and carbon black in F344 rats.
Fundam Appl Toxicol 25:80-94.

102 Seagrave, J.; McDonald, J.D.; Gigliotti, A.P.; Nikula, K.J.; Seilkop, S.K.; Gurevich,
M. and Mauderly, J.L. (2002) Mutagenicity and in Vivo Toxicity of Combined
Particulate and Semivolatile Organic Fractions of Gasoline and Diesel Engine Emissions.
Toxicological Sciences 70:212-226.

103 U.S.  Environmental Protection Agency (2004) Air Quality Criteria for Particulate
Matter. Research Triangle Park, NC: National Center for Environmental Assessment -
RTF Office; Report No. EPA/600/P-99/002aF.
                                      1-38

-------
104 U.S. Environmental Protection Agency (2004) Air Quality Criteria for Particulate
Matter. Research Triangle Park, NC: National Center for Environmental Assessment -
RTF Office; Report No. EPA/600/P-99/002aF, p. 8-318.

105 Fujita, E.; Watson, M.J.; Chow, M.C.; et al. (1998) Northern Front Range Air Quality
Study, Volume C: Source apportionment and simulation methods and evaluation.
Prepared for Colorado State University, Cooperative Institute for Research in the
Atmosphere, by Desert Research Institute, Reno, NV. May be downloaded at
http://www.nfraqs.colostate.edu/nfraqs/index2.html  as "Final Report and Appendices
A,B and C"

106 Schauer, J.J.; Rogge, W.F.; Hildemann, L.M.; et al. (1996) Source apportionment of
airborne particulate matter using organic compounds as tracers. Atmos Environ
30(22):3837-3855.

107 EPA, 2001, "National Air Quality and Emission Trends Report, 1999," EPA 454/R-
01-004.

108 Lawson, Douglas R., Ralph E. Smith, 1998, "The Northern Front Range Air Quality
Study" Executive summary.

109 Schauer, James J., Wolfgang F. Rogge, Lynn M. Hildemann, Monica A. Mazurek,
Glen R. Cass, and Bernd Simoneit, 1996, "Source Apportionment  of Airborne Particulate
Matter Using Organic Compounds as Tracers," Atmospheric Environment, 30, 3837-
3855.

110 Zheng, Mei, Glen R. Cass, James J. Schauer, and Eric S. Edgerton, 2002, "Source
Apportionment of PM2.5 in the  Southeastern United States Using  Solvent-Extractable
Organic Compounds as Tracers," Environmental Science and Technology, 36, 2631-2371.

111 Hannigan, Michael P., William F. Busby, Jr., Glen R.  Cass, 2005, "Source
Contributions to the Mutagenicity of Urban Particulate Air Pollution," Journal of the Air
and Waste Management Association, 55, 399-410.

112 Kleeman, Michael J.., Glen R. Cass, 1999, "Identifying the Effect of Individual
Emission Sources on Particulate Air Quality Within a Photochemical Aerosol Processes
Trajectory Model," Atmospheric Environment, 33, 4597-4613.

113 Schauer, James J., Glen R. Cass, 2000, "Source Apportionment of Wintertime Gas-
Phase and Particle-Phase Air Pollutants Using Organic Compounds as Tracers,
Environmental Science and Technology, 1821-1832.

114 Martin-Reviego, M. and K. Wirtz, 2005. "Is benzene a precursor for secondary
organic aerosol?" Environmental Science and Technology, 39, 1045-1054
                                      1-39

-------
                              Chapter 2: Table of Contents

Chapter 2: Emission Inventories	3
   2.1  Criteria Pollutants	3
     2.1.1   Methods	3
        2.1.1.1 Highway Vehicles	3
        2.1.1.2  Portable Fuel Containers	8
     2.1.2  Emission Reductions of Proposed Controls	9
     2.1.3   Strengths and Limitations of Criteria Pollutant Inventories	18
   2.2  Air Toxics	19
     2.2.1  Emission Inventories Used in Air Quality Modeling	19
        2.2.1.1  Methods Used to Develop Air Toxics Inventories for Air Quality Modeling . 20
           2.2.1.1.1  Highway Vehicles	20
           2.2.1.1.2  Nonroad Equipment in the Nonroad Model	25
           2.2.1.1.3     Commercial Marine Vessels, Locomotives and Aircraft	25
           2.2.1.1.4  Stationary Sources	26
           2.2.1.1.5  Precursor Emissions	28
           2.2.1.1.6  Strengths and Limitations	29
        2.2.1.2 Trends in Air Toxic Emissions	32
           2.2.1.2.1  Emission Trends Without Proposed Controls	32
           2.2.1.2.2  Impact on Inventory of Proposed Fuel Benzene Control	45
     2.2.2   Emission Reductions from Proposed Controls	50
        2.2.2.1 Methodology Changes from Air Quality Inventories	50
           2.22.1.1 Highway Vehicles	50
           2.2.2.1.2  Nonroad Equipment	52
           2.2.2.1.3  Portable Fuel Containers	53
           2.2.2.1.4 Gasoline Distribution	55
        2.2.2.2 Estimated Reductions for Air Toxic Pollutants of Greatest Concern	56
           2.2.2.2.1  Fuel Benzene Standard	56
           2.2.2.2.2  Cold Temperature VOC Emission Control	65
           2.2.2.2.3  Portable Fuel Container Control	69
           2.2.2.2.4  Cumulative Reductions of Proposed Controls	72
   2.3  Potential Implications of New Emissions Data for Inventories	79
     2.3.1  Newer Technology Light Duty Vehicles	79
     2.3.2  Heavy-Duty Vehicles (CRC E-55/E-59)	80
     2.3.3  Small Spark Ignition Engines	81
     2.3.4  Nonroad CI engines	83
   2.4    Description of Current Mobile  Source Emissions Control Programs that Reduce
         MSATs	84
     2.4.1   Fuels Programs	84
        2.4.1.1RFG	84
        2.4.1.2 Anti-dumping	85
        2.4.1.32001 Mobile Source Air  Toxics Rule (MSATI)	85
        2.4.1.4 Gasoline Sulfur	86
        2.4.1.5 Gasoline Volatility	86
        2.4.1.6 Diesel Fuel	86
                                          2-1

-------
   2.4.1.7 Phase-Out of Lead in Gasoline	87
2.4.2  Highway Vehicle and Engine Programs	87
2.4.3  Nonroad Engine Programs	89
   2.4.3.1 Land-based Diesel Engines	90
   2.4.3.2 Small Land-Based SI Engines	90
   2.4.3.3 Large Land-Based SI engines	90
   2.4.3.4 Recreational Vehicles	91
   2.4.3.5 Marine engines	91
   2.4.3.6 Locomotives	92
   2.4.3.7 Aircraft	92
2.4.4  Voluntary Programs	93
                                    2-2

-------
                       Chapter 2: Emission Inventories

       This chapter describes the methods used to develop inventories for air quality modeling,
estimation of emission benefits and calculation of cost-effectiveness for this rule. The chapter
also presents and discusses these inventories.  MSAT inventories for air quality modeling were
developed well in advance of rule proposal, because of the lead time required to conduct air
quality, exposure, and risk analyses.  Thus, these inventories do not include revised estimates of
emissions at cold temperature in vehicles, emissions from portable fuel containers, or revisions
in the gasoline distribution inventory to reflect changes made for the 2002 National Emissions
Inventory.  Therefore, the chapter has separate sections discussing MSAT inventories used for
modeling, and revised inventories used to estimate emission benefits of the rule and cost-
effectiveness.

2.1  Criteria Pollutants

2.1.1   Methods

2.1.1.1 Highway Vehicles

       Highway vehicle hydrocarbon (HC) emission inventories were calculated by using
vehicle emission rates produced from the emission model MOBILE6.2 multiplied by vehicle
miles traveled (VMT) using the National Mobile Inventory Model (NMIM).1 MOBILE6.2 uses
emission factors obtained through the analysis of emissions data collected from vehicle emission
research2. The VMT used by NMIM was estimated for base years using historical data from the
Federal Highway Administration, allocated to counties using the methodology documented for
the National Emissions Inventory, and projected to future years using the Energy Information
Administration's National Energy Modeling System (NEMS) Transportation Model. This is the
same approach used in the Clean Air Interstate Air Quality (CAIR) rule.3

       Analysis of vehicle emission  certification data submitted to EPA as part of requirements
to comply with requirements  for cold temperature carbon monoxide  (CO) standards by vehicle
manufacturers, as  well as surveillance testing data from the California Air Resources Board,
indicated that MOBILE6.2 was substantially underestimating start emission at cold temperatures
for Tier 1 and later vehicles. This data was supplemented with test data collected by the Office of
Transportation and Air Quality (OTAQ) at Southwest Research Institute (SwRI)4 and was then
used to adjust the temperature and engine start emission factors in MOBILE6.2 to provide inputs
to NMIM which calculates county level national inventories.5

       EPA cold CO certification data was paired as 20 °F versus 75 °F tests per engine family
to calculate the additional hydrocarbon (HC) emissions due to lower temperature. The bag
emission data where available indicated that at 20 °F, as in the standard FTP at 75 °F, the
majority of HC emissions occur during vehicle start and that lower vehicle soak and start
temperatures result in higher  HC emissions. Table 2.1 .-1 indicates the trends found in the EPA
Cold CO program certification data.
                                          2-3

-------
       The state of California has a 50 °F emission standard requirement and that data, also
supplied by manufacturers, reflects the same trend over the smaller temperature difference
(Table 2.1.-2).

       The testing done by OTAQ at SwRI was performed on four Tier 2 vehicles to confirm the
effects seen in the certification data and to extend the range of soak temperature to 0 °F. A
summary of the hydrocarbon data is found in Table 2.1.-3.
Table 2.1.-1. FTP HC Data From Federal Certified Vehicles
(grams per mile)

Emission Standard
Tier 1
TLEV
LEV
ULEV
LEV2
2004 Tier 2
2005 Tier 2
2006 Tier 2

Sample Size
410
64
695
132
119
172
190
90
75°
Mean
0.1190
0.0804
0.0501
0.0335
0.0296
0.0406
0.0415
0.0408
Std. Dev.
0.0553
0.0286
0.0209
0.0214
0.0123
0.0169
0.0203
0.0239
20°
Mean
0.8630
0.6996
0.6402
0.4675
0.5035
0.5641
0.5651
0.5502
Std. Dev.
0.7269
0.2778
0.3723
0.2727
0.2549
0.3269
0.3247
0.3107
Table 2.1.-2. FTP HC Emissions Data from California Certified
Vehicles
(grams per mile)

Emission
Standard
LEV
ULEV
LEV2

Sample
Size
53
14
21
75°
Mean
0.0397
0.0162
0.0346
Std.
Dev.
0.0259
0.0043
0.0097
50°
Mean
0.0988
0.0403
0.0843
Std.
Dev.
0.0631
0.0176
0.0310

Ratio of
Averages
2.49
2.48
2.44
                                          2-4

-------
Table 2.1.-3. SwRI FTP (Bag 1 Only) Emissions from Four
Tier 2 Vehicles
Temperature in °F
Number of Observations
Average THC (gm/mile)
Standard deviation
Ratio to 75 °F
75
4
0.115
0.072
1
20
8
1.658
0.780
14.446
0
4
3.752
2.117
32.699
       MOBILE6.2 currently has engine start emission factors based on 75° emission test data
on 1981 and newer vehicles. These engine start emissions are the difference, in grams, between
the emissions from phase 1 of the FTP after a 12 hour engine soak and the emissions of the same
driving fully warm and without the engine start.  Temperature effects on HC emissions are
estimated using a multiplier that depends on ambient temperature. This process is described in
the MOBILE6.2 documentation6. The current engine start adjustments in MOBILE6.2 are not as
large for Tier 1  and later vehicles as what is indicated in the certification  and SwRI test data.  A
method of correcting the emission factors was developed using the test data. Those methods are
covered in detail in EPA technical report no. EPA420-D-06-001,  "Cold Temperature Effects on
Vehicle HC Emissions."

      Based on our analysis from Tier 1  and newer vehicles, it was decided that additive values
would be applied to 75 °F start emission factors based on temperature and vehicle technology
(i.e., Tier 1, NLEV, Tier 2, etc). Additive values can more closely approximate the additional
hydrocarbon emissions caused strictly by the start and warm-up of the engine and/or the exhaust
aftertreatment at the different temperatures than multiplicative values. These values were
obtained from subtracting the FTP emissions at 0, 20, and 50 °F from the FTP emissions at 75 °F
using the certification  and SwRI test data. For emissions at temperature points where data was
not available (i.e., 50 °F for Tier 2 vehicles), linear interpolation between the 20 and 75 °F test
data was used.  All of the difference in emissions is attributed to the increase in engine start
emissions.  The values used for inputs for start adjustments are found in Table 2.1 .-4.

      It is not  clear what impact this phenomenon has on HC emissions in malfunctioning or
deteriorated vehicles.  Emissions could go up proportionally to properly operating vehicles or
could go up at a lower rate.  Properly operating vehicles are very  clean due to their emissions
technology. Vehicle starts represent a period of operation where  the vehicle's emissions
equipment is not fully  operational and the oxidation of fuel to carbon dioxide and water is not
optimal.  This situation is similar to the conditions found in a deteriorated or improperly
maintained vehicle except that the condition is temporary in a normal vehicle.  While MOBILE
currently uses a multiplier to account for temperature effects, doing so in the case results in
extremely high  and unrealistic emission rates.  Therefore we have used the MOBILE6.2 estimate
of FTP emissions at 20 °F for model year 2005 high-emitting vehicles in calendar year 2005 as
the additive factor for  all Tier 2 high-emitting vehicles.  Those values are  found in Table 2.1.-5.
We are not changing high-emitting vehicle emission factors for Tier 1 and older vehicles.
                                           2-5

-------
Table 2.1.-4. Increase in Engine Start Hydrocarbon Emissions
Over the 75 °F Baseline at Low Temperatures
(grams per engine start after a 12 hour soak)

Index
1
2
3
4
5
6
7
8
9
10
11

Index
1
2
3
4
5
6
7
8
9
10
11
12

Description
Tier 0 (not used)
Intermediate Tier 1
Tierl
Tier 2 (not used)
Intermediate Transitional Low Emission Vehicle
Transitional Low Emission Vehicle
Intermediate Low Emission Vehicle
Low Emission Vehicle (LEV)
Transitional Ultra Low Emission Vehicle
Ultra Low Emission Vehicle (ULEV)
Zero Emission Vehicle (ZEV) (not used)

Tier 2 (All Cars & Trucks) By Model Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
"F
0
25.96
25.96
25.96
18.26
21.60
21.60
20.59
20.59
15.14
15.14
0.00

0
18.26
18.27
17.77
17.77
17.77
17.77
17.77
17.77
17.77
17.77
17.77
17.77
20
12.98
12.98
12.98
9.13
10.80
10.80
10.29
10.29
7.57
7.57
0.00

20
9.13
9.13
8.88
8.88
8.88
8.88
8.88
8.88
8.88
8.88
8.88
8.88
50
3.09
3.09
3.09
3.27
2.09
2.09
1.30
1.30
0.87
0.87
0.00

50
3.27
3.27
3.27
3.27
3.27
3.27
3.27
3.27
3.27
3.27
3.27
3.27
         Table 2.1.-5. Tier 2 High Emitter HC Adjustment
Based on 2005 Model Year MOBILE6.2 Results in Calendar Year 2005
Temperature °F
Engine start grams without
adjustment
Additional grams
0
63.335
50.522
20
41.360
28.547
50
21.821
9.008
75
12.813
N/A
                              2-6

-------
       The above tables and the new emission standard were used to determine the effects of the
proposed emission standard on start emission factors.  The predicted reductions were applied to
Tier 2 vehicles over the phase-in period of the standards.  Those values are found in Table 2.1 .-6.
No reductions beyond those found for normally-emitting Tier 2 vehicles are applied for Tier 2
high-emitting vehicles.

       With the appropriate HC start emission temperature adjustment factors, we can provide
the necessary emission factors required as inputs to NMIM to project pre-control and control
inventories for this rule.  No modification to any other components of NMIM is needed to
calculate these inventories.
Table 2.1.-6. Adjustments to Engine Start Hydrocarbon Emissions
Over the 75 °F Baseline at Low Temperatures
For MSAT Rule
(grams per engine start after a 12 hour soak)

Index
1
2
o
J
4
5
6
7
8
9
10
11
12


Index
1
2
o
J
4
5
6
7
8
9
10
11
12

Tier 2 Cars & Light Trucks <6,000 Ibs GVWR
By Model Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015


Tier 2 Light Trucks >6,000 Ibs GVWR By Model
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
"F
0
18.26
18.27
17.77
17.77
17.77
17.77
6.66
6.66
6.66
6.66
6.66
6.66

20
9.13
9.13
8.88
8.88
8.88
8.88
3.3
3.3
3.3
3.3
3.3
3.3

50
3.27
3.27
3.27
3.27
3.27
3.27
1.215
1.215
1.215
1.215
1.215
1.215

"F
0
18.26
18.27
17.77
17.77
17.77
17.77
17.77
17.77
11.0
11.0
11.0
11.0
20
9.13
9.13
8.88
8.88
8.88
8.88
8.88
8.88
5.5
5.5
5.5
5.5
50
3.27
3.27
3.27
3.27
3.27
3.27
3.27
3.27
2.025
2.025
2.025
2.025
Phase In
Fraction
0
0
0
0
0
0
0.25
0.50
0.75
1.00
1.00
1.00

Phase In
Fraction
0
0
0
0
0
0
0
0
0.25
0.50
0.75
1.00
                                           2-7

-------
2.1.1.2  Portable Fuel Containers

       In 1999, California's Air Resources Board (ARE) proposed a methodology to estimate
annual emissions from portable fuel containers (PFCs) within California. Their approach relied
on survey data to first estimate the number of PFCs, and then to combine those estimates with
results from testing PFCs to develop a statewide annual inventory.

       EPA has modified California's approach.  We first used our NONROAD emissions
model to estimate (for each month of the year and for each state) the quantity of gasoline
dispensed from PFCs that was used to fuel nonroad equipment.  Then using some of the
California survey data on the amount of gasoline in stored in each PFC, EPA estimated the
number of PFCs in use (each season) in each state. These estimated  counts of PFCs were similar
(but not identical) to the California estimates.  EPA also adjusted the California emission
estimates to account for daily temperature variations  and seasonal RVP variations.  EPA then
combined its state-by-state estimates of PFC usage with its adjusted emission rates to obtain
seasonal VOC inventory estimates for each state.7

       For each of the 50 states plus the District of Columbia, this EPA approach produced the
estimates for calendar year 1990 given in Table 2.1.-7. Assuming no changes (i.e.,  no  controls),
each of these estimates will increase by approximately 1.21 percent annually due to the increase
in gasoline consumption predicted by the NONROAD model.

       Six states (California, Delaware, Maine, Maryland, New York, and Pennsylvania) have
implemented controls on the design of PFCs that will reduce HC emissions.  The California
program began in 2001.  The other states started their programs in 2005.  Additionally, seven
other states plus the District of Columbia (Connecticut, Massachusetts, New Jersey, Rhode
Island, Texas, Vermont, Virginia, and Washington DC) are also planning to adopt the California
PFC program.

       Additionally, California has begun to adopt more stringent emission standards that will
require each PFC to emit (permeation plus evaporation) no more than 0.3 grams of VOC per day
for each gallon of capacity.  This requirement will be effective  July 1, 2007.  Assuming that
PFCs have a typical life of about five years on average, the  "new" versions of the PFCs should
replace virtually all of the earlier versions by 2013. As these state programs result in replacing
the existing PFCs with lower-emitting PFCs, the estimated national inventory of VOCs
associated with PFCs will drop  by about 20 percent.

       To estimate the VOC emissions from gas cans assuming the proposed rules  are
implemented, we made to following three changes to our inventory estimates:

1.      Since the proposed rule  makes it unlikely for  a newly designed gas can to be left in the
       "open" position, we altered the distribution of the cans (from  the California survey) to
       100 percent "closed." This change reduced the VOC emissions from both evaporation as
       well as spillage during transport. (Note, the 13 states plus the District of Columbia that
                                          2-8

-------
       are adopting the California gas can rules already had this change applied. So, this
       affected the VOC emissions from only gas cans in the other 37 states.)

2.      This proposed rule also produces changes (to the design of the individual gas cans) that
       are expected to reduce the spillage by 50 percent (when these gas cans are used to refuel
       individual pieces of equipment). Again, this emission reduction was already included in
       the base case for those states that are adopting the California rules. Therefore, only the
       gas cans in the remaining 37 states contributed to our estimated reductions of spillage.

3.      Finally, the proposed rule includes a maximum emission rate of 0.3 grams per gallon per
       day for the new gas cans. We used this emission standard to estimate the total
       permeation plus evaporative emissions from each newly designed gas can.  Only
       California has adopted (or plans to adopt) this requirement. Thus, the effect of this
       proposed national requirement applies to the remaining 49 states.

       The change in VOC emissions was then calculated by subtracting the emissions (on a
state-by-state basis) estimated using these preceding three changes from our base estimates.  The
national estimate was simply the sum of the 50 individual state (plus DC) estimates.

2.1.2  Emission Reductions of Proposed Controls

       Light-Duty Gasoline Vehicles — We are proposing a 20° F FTP  emission standard for
non-methane hydrocarbon (NMHC) emissions from spark ignition vehicles of 0.3 grams per mile
for light duty vehicles and trucks that weigh 6000 pounds or less and a  0.5 gram per mile
standard for vehicles that weigh more than 6000 pounds. The standard will be applied to a
manufacturer on a sales-weighted fleet-wide basis.  Furthermore, the standards will be phased  in
over a period of time following the schedule found in Table 2.1.-8.

       The resulting reductions were modeled based upon the above standard and the phase-in
period. This was  done as outlined  in Section 2.1.1.1 with an external data file provided as input
to MOBILE6.2 that altered MOBILE6.2 start emission factors for Tier  2 vehicles only.
MOBILE6.2 was  then used with NMEVI to generate county and nationwide inventories of the
control case.  When the standard is fully phased in we expect a 60 % reduction in start emissions
in gasoline fuel vehicles that have  a gross vehicle weight rating (GVWR) of less than or equal  to
6000 Ibs and a 30 % in gasoline-fueled vehicles that have a GVWR greater than 6000 Ibs.  The
impact on future nationwide VOC  inventories is found in Table 2.1.-9.
                                           2-9

-------
Table 2.1-7. PFC Emissions (Tons per Year) by Source Type (for 1990)

State
AL
AK
AZ
AR
CA
CO
CT
DE
DC
FL
GA
HI
ID
IL
IN
IA
KS
KY
LA
ME
MD
MA
Ml
MN
MS
MO
MT
NE
NV
NH
NJ
NM
NY
NC
ND
OH
OK
OR
PA
Rl
Refilling PFC at Pump
Vapor Displ
159.6
17.5
273.4
88.3
1,602.2
209.9
148.9
33.6
5.7
817.5
305.6
51.9
43.6
383.4
213.7
105.7
93.7
107.4
132.1
47.7
248.2
230.9
452.7
155.6
70.3
193.4
23.7
53.9
81.0
51.4
351.5
56.3
479.6
368.1
17.7
523.5
124.9
165.0
396.8
29.9
Spillage
13.2
1.4
23.5
7.0
136.0
17.0
12.8
3.0
0.5
72.2
29.2
3.9
4.6
39.7
20.7
9.5
9.2
10.2
11.0
4.1
21.5
20.1
33.4
14.8
6.5
18.0
2.3
5.6
7.8
4.2
31.0
5.0
45.6
28.6
1.8
42.1
10.0
13.3
39.1
3.2

Spillage
During
Transport
395.5
46.8
655.2
218.5
3,815.5
485.9
367.9
87.8
18.2
2,026.0
838.6
110.6
135.9
1,148.0
606.1
283.9
269.7
311.8
339.7
125.6
604.5
584.2
993.3
444.2
204.2
536.6
72.7
166.4
217.1
125.7
889.5
147.9
1,339.2
828.9
53.6
1,223.4
304.0
383.2
1,164.9
92.9
Refueling Equipment
Vapor
Disnl
159.6
17.5
273.4
88.3
1,602.2
209.9
148.9
33.6
5.7
817.5
305.6
51.9
43.6
383.4
213.7
105.7
93.7
107.4
132.1
47.7
248.2
230.9
452.7
155.6
70.3
193.4
23.7
53.9
81.0
51.4
351.5
56.3
479.6
368.1
17.7
523.5
124.9
165.0
396.8
29.9
Spillage
871.3
83.3
1,665.4
428.9
9,452.1
1,174.2
884.4
210.5
37.1
4,998.5
1,971.4
273.4
301.8
2,673.0
1,406.0
625.7
614.6
656.2
694.8
285.5
1,521.8
1,372.7
2,253.8
940.8
412.9
1,182.5
143.5
367.6
550.7
283.1
2,093.1
338.8
2,918.2
1,937.1
105.1
2,886.9
669.3
915.1
2,670.4
217.2

Permeation
Plus
Evaporation
3,572.3
548.0
2,910.4
2,467.9
21,553.8
3,025.9
2,230.0
450.8
176.1
10,172.5
4,107.6
972.6
663.6
4,385.3
2,981.2
1,876.5
1,620.4
2,233.4
3,697.3
979.6
2,950.2
3,390.3
10,004.8
2,657.3
1,852.0
3,161.3
511.9
786.8
709.2
939.0
5,136.2
1,019.5
7,196.1
6,327.8
355.5
8,553.9
3,094.2
2,601.9
6,988.9
367.6

Totals by
State
5,171.4
714.6
5,801.2
3,299.0
38,161.8
5,123.0
3,793.0
819.5
243.3
18,904.2
7,558.0
1,464.3
1,193.0
9,012.8
5,441.4
3,007.0
2,701.3
3,426.3
5,006.9
1,490.3
5,594.5
5,829.1
14,190.8
4,368.2
2,616.3
5,285.1
777.7
1,434.1
1,646.8
1,454.8
8,852.8
1,623.8
12,458.3
9,858.5
551.3
13,753.4
4,327.4
4,243.4
11,656.9
740.6
                              2-10

-------

State
SC
SD
TN
TX
UT
VT
VA
WA
WV
Wl
WY
50-State
Refilling PFC at Pump
Vapor Displ
161.1
18.8
181.5
743.1
63.6
21.8
295.4
245.8
51.8
190.2
14.5
10,903.6
Spillage
14.1
1.9
16.6
68.1
6.5
2.0
26.2
20.5
4.4
16.9
1.4
961.1
Refueling
Equipment
Spillage
During
Transport
407.3
59.4
496.5
1,968.7
192.0
60.7
752.9
595.4
141.9
505.2
44.3
27,887.6
Refueling Equipment
Vapor
Displ
161.1
18.8
181.5
743.1
63.6
21.8
295.4
245.8
51.8
190.2
14.5
10,903.6
Spillage
974.5
118.3
1,086.4
4,654.3
419.2
134.2
1,845.0
1,411.9
279.8
1,118.3
90.5
65,221.2

Permeation
Plus
Evaporation
2,519.7
359.8
3,789.5
11,008.5
941.6
380.2
4,211.6
3,627.0
1,502.9
3,547.8
269.1
171,387.4

Totals by
State
4,237.8
577.1
5,751.9
19,185.9
1,686.4
620.7
7,426.5
6,146.4
2,032.6
5,568.4
434.4
287,264.5
       Table 2.1.-8. Proposed Phase-in Schedule for 20°F Standard by Model Year
Vehicle GVWR
(Category)
< 6000 Ibs
(LDV/LLDT)
> 60001bs HLDT
(and MDPV)
2010

25%



2011

50%



2012

75%

25%

2013

100%

50%

2014



75%

2015



100%

 Table 2.1.-9. Impact on Nationwide VOC Emissions from Light Duty Vehicles and Trucks
           of a 20 °F FTP Emission Standard for Non-Methane Hydrocarbons.
Year
1999
2010
2015
2020
2030
Tons Without Standard
4,899,891
2,936,905
2,625,076
2,556,751
2,889,269
Tons With Standard
N. A.
2,790,971
2,305,203
2,020,267
1,985,830
Reduction
N.A.
145,934
319,874
536,484
913,439
      These benefits are primarily realized in regions of the country with colder winter
temperatures.  Table 2.1 .-10 shows the impacts on a State by State basis in year 2030.
                                        2-11

-------
       Test data show that the proposed controls on cold temperature hydrocarbon emissions
will have the ancillary benefit of reducing PM emissions as well. Emissions generated during
cold temperature starts tend to be elevated due to a combination of a cold catalyst and excess fuel
in the combustion chamber. These factors increase emissions of benzene and other
hydrocarbons, and at the same time allow for unburned or pyrolized fuel to be emitted.

       A number of source apportionment studies have indicated previously that emissions from
vehicles starting at cold temperatures contribute disproportionately to ambient PM2.5.  For
instance, the Northern Front Range Air Quality study conducted in the Denver, CO area during
the winter of 1997 estimated that, on average, 12% of ambient PM2 5 could be attributed to cold
start light-duty gasoline vehicle emissions.8

       At this point, the PM emission factors in MOBILE6.2 for PM from light-duty gasoline
vehicles are not sensitive to temperatures. However, as outlined above,  the emission factors for
hydrocarbons and gaseous toxics are temperature-dependent.

       In order to estimate the expected emission reductions in PM as a result of the cold
temperature standards in this proposal, we evaluated the relationship between PM and NMHC in
Tier 2 vehicles operating at different temperatures.  All emissions benefits of the cold
temperature standard are expected to affect only the cold temperature starting emissions. As
such, all analyses were restricted to Bag 1.  However, similar results were obtained when using
full weighted FTP results.

       First, data from the only extant testing program of Tier 2 vehicles at multiple
temperatures was obtained from Southwest Research Institute.9  Figure 2.1 .-1 shows the PM
emission factors as a function of temperature. Like NMHC, PM emission factors  increase
exponentially with lower temperatures through the  entire range of testing.
                                          2-12

-------
Table 2.1.-10. Impacts on State Light Duty Vehice and Truck VOC Emissions of
    20 °F FTP Emission Standard for Non-Methane Hydrocarbons in 2030.

AL
AK
AZ
AR
CA
CO
CT
DE
DC
FL
GA
HI
ID
IL
IN
IA
KS
KY
LA
ME
MD
MA
Ml
MN
MS
MO
MT
NE
NV
NH
NJ
NM
NY
NC
ND
OH
OK
OR
PA
Rl
SC
SD
TN
Reference Case
Tons
49,848
11,377
50,563
28,603
249,670
59,856
28,578
7,573
3,462
110,729
99,741
6,979
20,716
117,780
87,191
36,930
34,192
49,849
35,684
17,412
49,383
49,937
141,535
87,180
23,418
73,449
17,728
23,655
26,445
18,650
57,554
27,037
155,448
89,150
12,087
119,496
44,642
53,308
116,128
7,615
46,158
12,261
67,115
Control Case
Tons
38,155
6,130
38,008
21,104
178,119
38,363
17,443
4,883
2,329
100,275
75,155
6,820
13,068
73,217
57,078
23,614
22,590
33,028
28,657
10,288
31,758
30,477
88,464
52,242
17,721
49,197
10,506
15,038
18,852
1 1 ,440
36,810
19,911
97,923
64,947
7,041
77,175
32,578
34,494
74,186
4,729
33,346
7,441
47,317
Reduction in
Tons
1 1 ,692
5,247
12,556
7,499
71,552
21,493
11,135
2,690
1,133
10,454
24,586
158
7,648
44,563
30,113
13,315
1 1 ,602
16,821
7,026
7,124
17,625
19,460
53,072
34,938
5,697
24,252
7,222
8,617
7,593
7,210
20,744
7,126
57,525
24,202
5,045
42,321
12,064
18,814
41,942
2,886
12,812
4,820
19,799
Percent
Reduction
23
46
25
26
29
36
39
36
33
9
25
2
37
38
35
36
34
34
20
41
36
39
37
40
24
33
41
36
29
39
36
26
37
27
42
35
27
35
36
38
28
39
29
                                 2-13

-------

TX
UT
VT
VI
WA
WV
Wl
WY
Reference Case
Tons
176,753
28,151
11,451
79,427
72,891
16,139
77,447
10,900
Control Case
Tons
146,569
17,576
6,993
54,082
44,616
10,259
47,205
6,614
Reduction in
Tons
30,184
10,575
4,458
25,345
28,275
5,881
30,242
4,286
Percent
Reduction
17
38
39
32
39
36
39
39
         Figure 2.1.-1.  FTP Bag 1 PM Emissions vs. Temperature, Tier 2 Vehicles
                          i   i   i
                                       0  20 40  60
                                       i   i   i    i
                                                          J	L
                f
                CL
                 01
                 03
                 DO
-
-
—
-2 -
-4 -
-6 -
-R -
10 -
Vehicle 4
o
@
o

o
Vehicle 1

!°8§,
o
o
Vehicle 5
o
8


Vehicle 2
o
o
o
Vehicle 6
».
o o
o
o o

Vehicles
o
a
o
o

- -2
- -4
- -6
- -8
- -10
-
-
                         0  20  40 GO                0  20 40 60

                                     Temperature (F)

Figure 2.1.-2 illustrates the relationship between FTP Bag 1 NMHC and PM emission factors in
this test program.  Lower temperature tests are found to the upper right corner, corresponding to
elevated emissions of both NMHC and PM. The symbol used for each data point represents the
different vehicles in the test program.  As shown, there is a clear, linear association. Thus, we
concluded that estimated reductions in PM as  a result of the hydrocarbon emission controls in
this rule could be estimated by applying a PM to NMHC ratio to the estimated reduction in
NMHC.
                                          2-14

-------
     Figure 2.1.-2. FTP Bag 1 PM and FTP Bag 1 NMHC for Various Tier 2 Vehicles
               "0:1
               O)
               ra
               ffi
                    oo _

A
$ ^<
.»' ¥ *
v 4-Q v
V 0 A + V
V •%> o
V v
X A o
1 1 1 1
-3 -2 -1 0
A +




I I
1 2
                                   Bag1 NHMC-ln(g/mi)
                               Plot Icons are Vehicle-Specific

       In order to determine an appropriate PM/NMHC ratio for calculating PM reductions from
NMHC reductions during cold start conditions, we employed mixed models with random vehicle
terms.10 We fit several models to the data, treating the PM/NMHC ratio as a dependent variable.
In summary, the model fit to the data was:
       Y=n+T+b+e
Here,  Y is a matrix of dependent variables (emission factors);
       |l is the intercept term or "grand mean";
       b is the change in emission factor associated with discrete testing temperatures;
       T is the vehicle effect, normally distributed around zero;
       e is the random error term (normally distributed).
Tests in which temperature was treated as a continuous variable were also employed.
Overall, the b term was found to be significant only at 75° testing,  and this may have been due to
random measurement errors in the PM/NMHC ratio as a result of very low emissions at 75°. The
b term became insignificant when it was allowed to vary randomly by vehicle. In addition,
because the proposed standards apply only to cold starting conditions, the  effect on the ratio at
75° is not relevant to changes in overall emissions. Therefore, we used the mean PM/NMHC
ratio of 0.022 to calculate the expected ancillary reductions in PM.  The 95% confidence interval
for the mean was 0.020-0.024.
                                         2-15

-------
       Using this number, the expected reductions in PM from this rule are estimated to be
7,037 tons in 2015, 11,803 tons in 2020 and 20,096 tons in 2030.  These calculations provide
initial evidence that the potential public health impacts of this proposal are substantial.

       In a subsequent test program in which the feasibility of the NMHC standards in today's
proposal was demonstrated, the test vehicle exhibited substantial reductions in PM emission as
well.  These PM emission reductions at 20° F were of similar magnitude as those predicted by
the above calculation. However, in that test program, the average PM/NHMC ratio was slightly
smaller than in the SwRI test program. The vehicle tested in the feasibility program reflected a
unique control technology that requires careful coordination among the engine air-fuel ratio  and
secondary air injection timing and air volume to provide the maximum emission benefits.  The
feasibility program was a "proof of concept" type study that did not have the ability to fully
explore ideal control  coordination and sizing of the emission control system. PM reductions
would very likely have been even greater if this coordination was possible.  The six current
unmodified production vehicles tested in the SwRI test program are considered to be more
representative  of emission control technologies found throughout the fleet.

       Several factors are not accounted for in the emission reduction estimation procedures,
which adds uncertainty to the level of emission reductions reported here.  First, if manufacturers
employ control technologies that differ substantially from those in the SwRI test program, actual
emission reductions could differ from the estimates here. Second, actual PM reductions may be
affected by the extent to which different vehicle  or engine technologies penetrate into the vehicle
market (such as hybrid electric drivetrains and direct injection  gasoline engines).

      Portable Fuel  Containers — The portable  fuel container controls proposed in this rule will
also reduce emissions of hydrocarbons. As noted in Section 2.1.1.2, thirteen states plus the
District of Columbia  have adopted controls on PFCs independent of the controls proposed in this
RIA.  In Figure 2.1 .-3, we have graphed the estimated annual national VOC emissions (in tons)
associated with PFCs for the following three scenarios:

  —   a base scenario in which no PFC controls  are used illustrated with the dotted  (black) line,

  —   a scenario in which only those 13 states plus DC have implemented PFC controls
      illustrated with the solid (blue) line,  and

  —   a scenario in which the PFC controls proposed in this RIA are implemented nationwide
      illustrated with the dashed (red) line
                                          2-16

-------
                   Figure 2.1-3. Comparison of PFC Control Scenarios
         Annual Nationwide VOC Emissions (Tons) from PFCs by Calendar Year
   500,000
   400,000
   300,000
   200,000
   100,000
Base

State  Programs

Nationwide
          1990
                         2000
                                         2010
                                     Calendar Year
                                                        2020
                                                                        2030
As noted in Section 2.1.1.2, the estimates of the VOC inventory in the basic scenario are
increasing (annually) at a rate of about 1.21 percent. The scenario containing just the state
programs has the estimated VOC inventory increasing at an annual rate of about 1.33 percent
once all of the programs are phased in. Similarly, the scenario in which nationwide requirements
(of this RIA) are phased in exhibit an annual increase in the VOC inventory of about 1.44
percent after phase-in.

Table 2.1.-11 compares the estimated national (annual) inventory  of PFC-related VOC with the
proposed control program to a reference case scenario that includes only State level controls.

            Table 2.1.-11. Nationwide Annual Gas Can VOC Emissions (tons)
Calendar
Year
1999
2007
2010
2015
2020
2030
With NO EPA
PFC Controls
318,596
310,744
279,374
296,927
318,384
362,715
With EPA
PFC Controls
NA
NA
250,990
116,431
125,702
144,634
Reduction
NA
NA
28,384
180,496
192,683
218,080
                                          2-17

-------
2.1.3   Strengths and Limitations of Criteria Pollutant Inventories

       Light-Duty Gasoline Vehicles — Emission factors for hydrocarbons in the MOBILE
model are based on tens of thousands of tests under a wide variety of conditions, and account for
leaking fuel systems, aggressive driving, air conditioner use and a variety of other parameters.
These data  are supported by over 50 technical reports, and many of them received extensive
scientific peer review. The strengths and limitations of the MOBILE model have been evaluated
by the Coordinating Research Council and the National Research Council.11'12

       There are significant uncertainties in emission inventories resulting from the use of
national default data rather than local inputs, as well as "top-down" allocation schemes in
estimating toxic emissions. Examples include use of national default vehicle registration
distributions,  default average speed distributions, and use of county level population data to
allocate State or urban level VMT.

       Also, it should be noted that there are greater uncertainties in projection year estimates.
Estimates of emissions from advanced technology vehicles and engines that will comply with
planned future emission standards include assumptions regarding levels of emission deterioration
and performance under various conditions.  Also, vehicle miles traveled are estimated using
economic projections with similar inherent limitations.

       The revised estimates of cold start VOC emissions are based on a robust dataset at
temperatures of 20°F and above. At lower temperatures, however, data are more limited and the
magnitude  of cold temperature effects is not as certain.  Similarly, the estimate of PM reductions
from NMHC cold temperature controls are based on limited data, although PM shows a very
strong correlation with NMHC.  Future control strategies may also employ mechanisms that
result in different PM/NMHC ratios than found in existing vehicles.

       Portable Fuel Containers — To estimate PFC inventories we were able to build on  survey
and test data collected by the California Air Resources Board.  We also developed inventories
using a "bottom up" approach which provides flexibility and permits very detailed fine-tuning of
the various scenarios.  However, the inventory involved many assumptions, including refueling
activity and temperature effects. Spillage occurring when non-road equipment is refueled  is a
significant  source of VOC emissions.  We are assuming (from EPA's NONROAD model) that
spillage is a constant 17 grams for each refueling event.  We are also assuming that each
refueling event occurs when the fuel tank on that piece of equipment is empty.  However, if the
user "tops off the fuel tank prior to each use, then we are underestimating the total VOC
emissions.

       Another assumption relates to whether inactive PFCs are stored with fuel. For example,
we assumed that a residence that uses a PFC to only fuel a lawn mower (perhaps six months of
the year) will have that PFC empty the remainder of the year (i.e., no permeation or evaporative
emissions). However, if that PFC were to contain a small amount of gasoline for those non-
mowing months, then we are underestimating the total inventory.
                                          2-18

-------
       Uncertainty in the characterization of the population of PFCs (i.e., commercial versus
residential usage, open versus closed, metal versus plastic) is the major source of uncertainty in
our estimates of the inventory of VOCs from PFCs.  Our characterization of the population of
PFCs is based on surveys performed by the Air Resources Board (ARE) of California.  We used
the same distribution of open versus closed PFCs determined by ARE.  Since the rest of the
country might not be exactly like California (relative to PFCs), we performed a sensitivity
analysis to determine the effects of varying that distribution. We found that even relatively large
changes in that distribution produced changes in estimated total VOC of less than 13 percent.13
Other source of uncertainty include estimates of the frequency of refilling of containers,
estimates of effects of ambient temperature on vapor displacement and spillage estimates of
effects of RVP on vapor displacement, impacts of temperature of the fuel itself on emissions, and
estimates of the amount of spillage during refilling.

2.2  Air Toxics

2.2.1 Emission Inventories Used in Air Quality Modeling

       The data and methods employed to develop the county-level air toxics inventories used
for air quality, exposure and risk modeling to support this rule are discussed in detail in the EPA
Technical Report, "National Scale Modeling of Air Toxics for the Mobile Source Air Toxics
Rule; Technical Support Document," Report Number EPA-454/R-06-002. In addition, the
reference case emissions modeling (i.e., emissions modeling without proposed controls) has been
externally peer-reviewed in a journal article currently in press.14 All underlying data and
summary statistics are included in  the docket for this rule.  The following sections summarize the
methods  used to develop these inventories and present results. While air quality, exposure, and
risk modeling was done for years 1999, 2015, 2020,  and 2030  (with modeling for 1999 done as
the National Scale Air Toxics Assessment), reference case inventories were also developed for
2007 and 2010 in order to better assess emission trends over time. Inventories for 1990 and 1996
which are methodologically consistent with later year inventories are also discussed to put
emission trends for later years into perspective. Control case modeling was done for proposed
fuel benzene standards in 2015, 2020 and 2030. Inventories which included revised estimates of
cold temperature hydrocarbon and air toxics emissions and portable fuel container emissions
were not completed in time to be included in this modeling. For the reference case, we modeled
all air toxic compounds listed in section 112 of the Clean Air Act for which we had adequate
data to estimate emissions. Table 2.2.-1 lists the pollutants included in these inventories which
were used in subsequent modeling of air quality, exposure, and risk. For the control case, we
modeled  a smaller subset of pollutants as discussed below. Emission inventories included
stationary sources, highway vehicles, and nonroad equipment.
                                          2-19

-------
2.2.1.1  Methods Used to Develop Air Toxics Inventories for Air Quality Modeling

2.2.1.1.1 Highway Vehicles

       For modeling calendar year 1999, we used the 1999 National Emissions Inventory
(NEI), final version 3.15  This inventory was also used in the 1999 National-Scale Air
Toxics Assessment. This inventory estimated highway vehicle emissions using the
MOBILE6.2 emission factor model.16'17  The 1999 NEI includes vehicle refueling
emissions as part of the stationary source inventory; thus, in developing inventories for
air quality, exposure and risk modeling these emissions were treated as stationary
sources.

       Within the MOBILE6.2 model, six MSATs (benzene, formaldehyde,
acetaldehyde, 1,3 butadiene, acrolein, and methyl tertiary butyl ether [MTBE]) can be
calculated directly by including detailed fuel parameters within the MOBILE6.2 scenario
descriptions.  These fuel parameters are:  sulfur content, olefms content, aromatics
content, benzene content, E200 value, E300 value, oxygenate content by type, and
oxygenate sales fraction by type.A  Since these fuel parameters are area-specific, EPA
developed county-level inputs for each of these parameters by season.  Fuel parameters
were collected for winter and summer seasons using a number of different data sources.
These sources include the Alliance of Automobile Manufacturers, Northrop Grumman
Mission Systems (formerly TRW Petroleum Technologies),  and EPA reformulated
gasoline surveys. Documentation for the NEI describes the development of the fuel
parameter database used with MOBILE6.2 in detail. The fuel  parameter data through
1999 are posted at the following website:

ftp://ftp.epa.gov/EmisInventory/fmalnei99ver3/haps/datafiles/onroad/auxiliary/

       MOBILE6.2 also has a command (ADDITIONAL HAPS) which allows the user
to enter emission factors or air toxic ratios for additional  air toxic pollutants. Emission
factors for the other HAPs in Table 2.2.-1 were calculated by MOBILE6.2 through the
use of external data files specifying emission factors for these pollutants in one of three
ways: as fractions of volatile organic compounds (VOC), fractions of PM, or by
supplying the basic emission factors. The ratios used with this command must be
expressed as milligrams of HAP per gram of VOC or PM. Gaseous hydrocarbons were
estimated as fractions of VOC. Poly cyclic aromatics hydrocarbons (PAHs) were
calculated as fractions of PM,  although the data used to calculate mass ratios included
both gas and particle phase PAH emissions.
 ' E200 and E300, represent the percentage of vapor that gasoline produces at 200 and 300 °F, respectively.
                                      2-20

-------
        Table 2.2.-1. Air Toxics Included in Emission Inventories and Used for Air
                     Quality, Exposure, and Risk Modeling.
1,3 -Butadiene
2,2,4-Trimethylpentane
Acenaphthene
Acenaphthylene
Acetaldehyde
Acrolein
Anthracene
Benzene
Benz(a)anthracene
Benzo(a)pyrene
Benzo(b)fluoranthene
Benzo(g,h,i)perylene
Benzo(k)fluoranthene
Chromium
Chrysene
Dib enzo(a, h)anthracene
Ethyl Benzene
Fluoranthene
Fluorene
Formaldehyde
n-Hexane
Indeno(l,2,3,c,d)-pyrene
Manganese
Methyl tert-butyl ether (MTBE)
Naphthalene
Nickel
Phenanthrene
Propionaldehyde
Pyrene
Styrene
Toluene
Xylenes
       Metals were estimated using basic emission factors. Evaporative emissions (e.g.,
toluene, xylenes) can only be estimated as fractions of VOC. Because toxic to VOC
ratios for several gaseous HAPs vary between baseline gasoline and gasoline oxygenated
with MTBE or ethanol, separate ADDITIONAL HAPS input files were developed for: 1)
baseline gasoline; 2) gasoline oxygenated with 2% MTBE by weight (e.g., Federal
reformulated gasoline); 3) gasoline oxygenated with 2.7% MTBE by weight (e.g., winter
oxygenated gasoline); and 4) gasoline oxygenated with 3.5% ethanol by weight
(gasohol). The documentation for the NEI provides more information on the
development of HAP inventories using this command. ADDITIONAL HAPs inputs
(including PAHs) for the 1999 NEI, final version 3  can be obtained at the same link given
above for the final 1999 NEI fuel parameter files.

       Although fuel parameter data were prepared for only two seasons (summer and
winter), four seasonal scenarios were developed.  The months corresponding to each
season were selected to best coincide with seasonal fuel requirements.   Summer fuel
parameters were applied in the fall scenarios and winter fuel parameters were applied in
the spring scenarios.

       The number of MOBILE6.2 input files required to model all counties in a State
were determined based on unique combinations of control programs and fuel parameters.
                                      2-21

-------
For counties where there was more than one fuel type sold, such as reformulated
gasolines with MTBE and ethanol, two sets of MOBILE6.2 input files were developed,
and resulting emission factors were weighted by gasoline market shares to derive overall
county-level emission factors. The county level emission factors were multiplied by
VMT from the Highway Performance Monitoring System (HPMS), as described in the
documentation for the 1999 NEI.  It should also be noted that California provided its own
air toxic emissions estimates for 1999, which replaced those generated by EPA.

       To develop projection year inventories for highway vehicles, we used NMIM.18' 19
NMIM develops inventories using EPA's MOBILE6.2 emission factor model for
highway vehicles, EP A' s NONRO AD emissions inventory model  for nonroad
equipment, and model inputs stored in data files. Model inputs include  data such as
temperatures, fuel properties, vehicle registration distributions, inspection and
maintenance programs, vehicle miles traveled, and toxics inputs in the form of toxic to
volatile organic compound (VOC) ratios,  toxic to particulate matter (PM) ratios, or toxic
emission factors. The toxics inputs were developed from a variety of emissions testing
programs conducted by EPA, States, and industry over many years (see Section 2.2.1.1.6
for more information).  Details  on data sources can be found in the documentation for the
National Emissions Inventory.  Projection year fuel parameters were developed using
results of several refinery modeling analyses conducted to assess impacts of fuel control
programs on fuel properties.20' 21' 22

       The projection year fuel parameters were calculated by applying adjustment
factors to the base year parameters.23  In addition, NMIM uses monthly  rather than
seasonal fuel parameters, and parameters for spring and fall months are  estimated by
interpolating from summer and winter data. Documentation of the fuel  parameters used
in NMIM was compiled in 2003 (Eastern Research Group, 2003),  and subsequently, a
number of changes were made,  based  on comments from States. These changes are
documented  in the change log for NMEVI, dated May, 14, 2004. This change log is
included in the docket for this rule, along with the original documentation.  In general,
multiplicative adjustment factors were used to calculate future year gasoline parameters
(i.e., future year parameter = base year parameter x adjustment factor). However, additive
adjustment factors were used to calculate  future year parameters for E200, E300, and
oxygenate market shares (i.e., future year parameter = base year parameter +  adjustment
factor). The database used for this assessment assumes no Federal ban on MTBE, but
does include State bans. Also, it did not include the renewable fuels mandate in the
recent Energy Policy Act. Vehicle miles traveled used in this assessment were those
developed for the Clean Air Interstate Air Quality Rule (CAIR).24

       NMIM outputs for 1999, 2007, 2010, 2015 and 2020 were  used to develop ratios
of future year to 1999 air toxic inventories.  These were then applied to  1999  NEI
inventory estimates by SCC, county and HAP:
                         NMIM, 1999
                                      2-22

-------
where PF20xx is the projection factor for 2007, 2010, 2015, 2020, or 2030, E20xx is the
emissions for the corresponding year and £1999 is the 1999 emissions.  NMIM results
were provided for the following emission types - exhaust, non-refueling evaporative and
refueling evaporative.  ENMIM was computed as the sum of non-refueling evaporative and
exhaust emissions for pollutants with both an exhaust and evaporative emissions
component (benzene, 2,2,4-trimethylpentane, naphthalene, toluene, xylenes, n-hexane,
and ethylbenzene). Separate ratios were developed for each vehicle class, pollutant and
county combination. In addition, separate ratios were developed for vehicle refueling,
and these ratios were used to project refueling emissions in the stationary source
inventory.

       In cases where the 1999 NEI included aggregated or different categories other
than those in NMIM, we aggregated NMIM results prior to applying ratios. For example,
California reported heavy duty diesel vehicle (HDDV) emissions in the 1999 NEI as an
aggregated HDDV "total" vehicle type rather than the specific HDDV classes (e.g., Class
2B, Class 3, 4, and 5).  Thus, we aggregated NMIM HDDV results for California in order
to apply a projection ratio to the HDDV "total" emissions.  In the event that the NEI had
HAPs not covered by NMIM (resulting from a state or local agency inventory
submission), we developed ratios based on NMIM PM or VOC results.

       For years 2015, 2020, and 2030, inventories were developed that reflected the
impacts of the fuel benzene standard proposed in this rule.  These control case inventories
included the following pollutants: benzene, 1,3-butadiene, formaldehyde, acetaldehyde
and acrolein.  In MOBILE6.2, emissions of other pollutants are not affected by changes
in fuel benzene or aromatics levels.

       To develop these inventories, NMIM was rerun with revised gasoline fuel
parameter inputs for fuel benzene and aromatics levels. These inputs were revised based
on refinery modeling done for the rule. As part of the refinery modeling, average fuel
properties under the new standards were estimated for each Petroleum Administration for
Defense District (PADD). Average fuel benzene levels for conventional gasoline  and
reformulated gasoline in each PADD before and after implementation of the proposed
standards were used to develop multiplicative factors which were applied to the reference
case fuel benzene levels for each county in the NMIM database. These multiplicative
factors are summarized in Table 2.2.-2. Although California is part of PADD5, it was
treated separately, since California has its own reformulated gasoline program. The
refinery modeling also indicated that the reduction in fuel benzene levels would result in
small decreases in aromatics levels as well.25  Thus aromatics levels were adjusted using
the additive factors calculated as follows:

              Additive Factor = 0.77*(BZ(control) - BZ(ref))            (2)

              Where BZ = benzene

An Excel workbook, designated "fuel changes.xls", summarizes the control and reference
case fuel benzene and aromatics levels used for 2015, 2020, and 2030.  This file is
                                       2-23

-------
included in the docket for the rule. We also checked the control case fuel benzene levels
to make sure the nationwide average level was close to the proposed standard. We did
this by weighting county fuel benzene level by VMT as a surrogate for fuel sales.  The
resulting nationwide average level was a little under 0.63%, very close to the standard.
The refinery modeling methodology is discussed in Chapter 9 of the Regulatory Impact
Analysis.  Since the reduction in fuel benzene changes well below one percent of the
gasoline, the level of uncertainty in the impacts on other fuel parameters and emissions is
quite small.

       Once fuel parameters were developed for the control case, NMEVI was rerun with
the same data files used in original reference case runs. Output included total exhaust and
non-refueling evaporative emissions, exhaust emissions, non-refueling evaporative
emissions, and refueling evaporative emissions. Projection factors for each emissions
type, by gasoline vehicle class, county and pollutant, were calculated as follows:
              PF
              1 * 20XX
                       ENMIM
                            ControUOXX
(3)
                            Reference20XX
    Table 2.2.-2. Average Fuel Benzene Level (Volume Percent) by PADD with
 Implementation of Proposed Fuel Benzene Standard (CG - Conventional Gasoline;
                         RFC - Reformulated Gasoline)

Reference
Case

Control Case

Multiplicative
Factor


CG
RFG
CG
RFG
CG
RFG
PADD
1
0.91 %
0.59%
0.55%
0.54%
0.60
0.92
PADD
2
1.26%
0.80%
0.68%
0.71%
0.54
0.89
PADD
3
0.95%
0.57%
0.54%
0.55%
0.57
0.96
PADD
4
1.47%
1.05%
0.93%
0.62%
0.63
0.59
PADD
5
1.42%
0.65%
0.85%
0.60%
0.60
0.92
Calif.
0.62%
0.62%
0.61%
0.61%
0.98
0.98
Pp2oxx is the projection factor for 2015, 2020, or 2030, and ENMIM Controi2oxx is the NMIM
emissions for the control scenario. It includes exhaust and non-refueling evaporative
emissions, but not refueling emissions. ENMiMReference2oxx is the NMIM reference case
MSAT emissions, and includes exhaust and non-refueling evaporative emissions, but not
refueling emissions. Although vehicle refueling was estimated as part of the stationary
source inventory, changes in MOBILE6.2 vehicle refueling emissions with fuel benzene
                                      2-24

-------
control were used to adjust the reference case refueling inventory to obtain the control
case inventory.

2.2.1.1.2 Nonroad Equipment in the Nonroad Model

       Nonroad equipment in the NONROAD model includes such sources as
recreational, construction, industrial, lawn and garden, farm, light commercial, logging,
airport service, railway maintenance, recreational marine vessels. For modeling calendar
year 1999, we used the 1999 National Emissions Inventory (NEI), final version 3.  This
inventory used NONROAD2004, which was also used in the recent Clean Air Nonroad
Diesel Rule.26 As with highway vehicles, exhaust gaseous hydrocarbons were estimated
as fractions of VOC, PAHs were calculated as fractions of PM, and metals were
estimated using basic emission factors. Evaporative emissions were estimated as
fractions of VOC. The projection of the portion of the nonroad inventory included in the
NONROAD model followed a similar methodology as for the on-road.  Projection factors
were developed using the 1999 and future year NMIM runs and were applied to nonroad
categories in the 1999 NEI. Retrospective inventories for nonroad  equipment in 1990 and
1996 are available at the same link given for the 1990 and 1996 highway inventories and
are described in the documentation for the 1999 NEI.

       Changes in fuel benzene and aromatics levels are expected to result in similar
emission changes  for nonroad gasoline equipment as for gasoline highway vehicles.
However, NMIM  does not have the capability to model  impacts of these fuel changes on
nonroad equipment emissions. Thus, we assumed that changes in county level exhaust
emissions of nonroad gasoline equipment were proportional to changes in highway light
duty gasoline vehicle exhaust emissions, and changes in county level evaporative
emissions of nonroad gasoline equipment were proportional to changes in highway light
duty gasoline vehicle evaporative (refueling and non-refueling) emissions:

                  ,   ,           ELDOVexhaustNMIMControl20XX
       Pr nonroad exhaust20XX = —	              (4)
                                h,LDGVExhaustNMIMReference20XX

        jjj-,        ,           ^LDGVevaPNMIMControl20XX
       PF nonroad evap20XX =	                   (5)
2.2.1.1.3      Commercial Marine Vessels, Locomotives and Aircraft

       These source sectors will not be impacted by the fuel benzene standards being
proposed in this rule.

       Emissions for these source sectors in 1999 were obtained from the 1999 National
Emissions Inventory, Final Version 3. Gaseous air toxic and PAH emissions for turbine
engine aircraft were estimated by applying toxic to VOC ratios obtained from detailed
characterization of turbine engine emissions.  Since no emissions data were available for
                                      2-25

-------
piston engine aircraft, a speciation profile from a non-catalyst light-duty gasoline vehicle
was used as a surrogate.  Metal emissions were not estimated for aircraft. No speciated
emissions data were available for commercial marine vessels. For diesel marine vessels,
profiles from heavy-duty diesel highway vehicles were used; for steamships, a profile for
stationary and industrial  boilers was used. Locomotive air toxic emissions were
estimated using speciation data from a year 2000 study done by the California Air
Resources Board.27  More detailed information on methods used to develop air toxic
                                                                          rjO
inventories for these sectors can be found in the documentation for the 1999 NEI.   This
documentation also  describes methods used to develop inventories for 1990 and 1996.

       The following approaches were used to project emissions for these source
categories:

       Locomotives and commercial marine vessels - For gaseous HAPs, inventories
were developed by applying ratios of future year to 1999 national level 50 state VOC
inventory estimates  (from the recent Clean Air Nonroad Diesel rule) by SCC code. For
polycyclic aromatic  hydrocarbons, PM ratios were used. Metal inventory estimates were
projected to future years based on activity.  Locomotive activity was projected using fuel
consumption data from the Energy Information Administration, as discussed in the
Regulatory Impact Analysis for Clean Air Nonroad Diesel Rule. For commercial marine
vessels, projected equipment populations from 1998 Power Systems Research (PSR)  data
were used to develop factors.  The future year inventories do not account for potential
reductions of additional locomotive or commercial marine vessel emission controls
currently under consideration.

       Aircraft - To project emissions from aircraft and from aviation gas distribution
emissions, we developed and applied growth factors (in EMS-HAP) to 1999 emissions
based on landing and take off data.  The Federal Aviation Administration's Terminal
Area Forecast System provided landing and take off data for future years up to 2020,
associated with commercial aircraft, general aviation,  air taxi and military aircraft.29
These four categories map directly to the inventory categories for aircraft emissions.  The
landing and take off data were summed across airports to create growth factors at the
national level.   The general aviation growth factors were used for aviation gas
distribution emissions. After 2020, activity was assumed to increase at the same rate  as
the increase from 2015 to 2020.

2.2.1.1.4 Stationary Sources

       Stationary source estimates for 1990, 1996, and 1999 were obtained from the
National Emissions  Inventories for those years.30'31'32'33

       For nearly all stationary sources (point and non-point source inventories), we used
the Emissions Modeling System for Hazardous Air Pollutants (EMS-HAP), Version 3.0
to apply growth and control factors to the 1999 NEI, source type by source type.34 EMS-
HAP has the capability of projecting emissions to 2020. After 2020, stationary source
emissions were assumed to remain constant.
                                       2-26

-------
       The general methodology for projecting stationary source emissions using EMS-
HAP is as follows:

Future Year Emissions = Base Year Emissions * Growth Factor * (100% - % Reduction)/100

       The actual equations used by EMS-HAP also allow the application of a "new
source" reduction to a fraction of the emissions to allow for a different level of emission
reduction to be applied to a portion of the emissions. In addition, if the source is already
controlled, and the value of the overall control efficiency is provided in the emission
inventory, EMS-HAP adjusts the percent reduction (% Reduction) based on the overall
control efficiency value provided in the inventory.  The actual projection equations are
provided in Chapter 6 (PtGrowCntl) of the EMS-HAP User's Guide (U. S. EPA, 2004b,
pp. 6-15-6-17).

       Stationary source growth — EMS-HAP allows growth factors to be applied to the
inventory on either a national, state or county level basis, based on one of the following
inventory codes that describe the source: (1) MACT, which identifies an emission source
as a belonging to a particular regulatory category or subcategory; (2) Standard Industrial
Classification (SIC), which classifies establishments by their primary type of activity, as
defined by the U.S. Census Bureau; (3)  Source Category Code (SCC), which defines the
source using EPA's coding system for the NEI. The MACT and SCC code definitions
are contained in the code tables supplied with the NEI.  Note that even though the code is
called "MACT", it is also used for other regulations besides MACT such as section 129
rules.  The hierarchy built into EMS-HAP is to use a MACT-based growth factor first,
followed by an SIC-based and lastly, an SCC-based growth factor.  The most detailed
geographic level is used first (e.g., a state-specific growth factor replaces a national
growth factor).  EMS-HAP does not have the capability to apply growth factors to
specific point source facilities, nor can they be applied differently for the different
pollutants for a particular source category.

   For stationary sources, growth factors were developed using three primary sources of
information:

       Regional Economic Models, Inc. (REMI) Policy Insight® model, version 5.5;35'36
       Regional and National fuel-use forecast data from the Energy Information
       Administration,  U.S. Department of Energy, Annual Energy Outlook (AEO)37
       Rule development leads or economists who had obtained economic information in
       the process of rule development.

The first two sources of information were also used in projecting criteria pollutant
emissions for EPA's 2005 Clean Air Interstate Rule.38

       More details on  how these sources were used can be found in the EPA technical
report, "National Scale Modeling of Mobile  Source Air Toxic Emissions, Air Quality,
Exposure and Risk for the Mobile Source Air Toxics Rule," cited previously.
                                       2-27

-------
       For refueling emissions, which are related to mobile sources but inventoried as
stationary sources, we developed SCC-based growth factors based on changes in
refueling emissions predicted using MOBILE6.2.

       Stationary source reductions -- Emission reductions were applied to the grown
emissions to account for regulatory efforts which are expected to reduce HAPs from 1999
levels. The percent reductions we determined were primarily based on estimates of
national average reductions for specific HAPs or for groups of HAPs from a source
category or subcategory as a  result of regulatory efforts.  These efforts are primarily the
MACT and section 129 standards, mandated in Title III of the 1990 Clean Air Act
Amendments.  We determined percent reductions, and whether they apply to major only
or both major and area sources, for the various rules from rule preambles, fact sheets and
through the project leads (questionnaire and phone calls). A major source is defined as
any stationary  source or group of stationary sources located within a contiguous area and
under common control that has the potential to emit  considering controls, in  the
aggregate,  10 tons per year or more of any hazardous air pollutant or 25 tons per year or
more of any combination of hazardous air pollutants. For some rules, percent reductions
were provided for specific  HAPs or groups of HAPs (e.g., all metals, or all volatiles)
rather than a single number for all HAPs in the categories. After 2010, stationary source
emissions are based  only on economic growth.  They do not account for reductions from
ongoing toxics programs such as the urban air toxics program, residual risk standards and
area source program, which are expected to further reduce toxics.

       Impact of Fuel Benzene Controls - The fuel benzene controls in this rule will
reduce emission from vehicle refueling, and also emissions from gasoline distribution.
Gasoline distribution emissions include emissions at bulk terminals, bulk plants, and
service stations, and emissions during transport  by trucks, marine vessels, and rail.
Reductions in emissions from all these sources were assumed to be proportional to
reductions in vehicle refueling emissions.

2.2.1.1.5 Precursor Emissions

       In addition to the air toxics in Table 2.2.-1, emissions of a number of other
compounds were estimated because they are precursor emissions which are
atmospherically transformed  into air toxics. These pollutants are listed in Table 2.2.-3,
along with air toxic pollutants included in the inventory which can be transformed into
other air toxics. Precursor emissions in 1999 were estimated by applying speciation
profiles from SPECIATE to VOC estimates from version 2 of the 1999 NEI.39 For
mobile sources, precursor emissions were projected to future years using ratios of VOCs
for future years versus 1999.   Stationary source  precursor emissions were assumed to
remain at  1999 levels since the impact of growth and control is unknown.
                                       2-28

-------
2.2.1.1.6 Strengths and Limitations

       Highway Vehicles - Limitations in the VOC and PM emission estimates which
are the basis for calculating air toxic emissions are discussed in Section 2.1.3.
MOBILE6.2 toxic to VOC ratios for key toxics from gasoline vehicles, such as benzene,
1,3-butadiene, formaldehyde and acetaldehyde, are based on almost 900 vehicle tests on a
wide variety of fuels.  These data account for impacts of emissions control technology,
normal vs. high emitters, and impacts of a variety of fuel properties, including benzene.
level, aromatics levels, olefm level, sulfur level, RVP, E200, E300, and oxygenate
content.

       However, there are a number of significant uncertainties in our highway vehicle
air toxic inventories for air quality modeling. Among the uncertainties are:
   •   The Agency has limited emissions data for advanced technology highway
       vehicles, including hybrid and alternative technology vehicles. The toxic to VOC
       ratios in MOBILE6.2 are all based on Tier 0 and earlier vehicles. EPA has
       recently evaluated data on more recent technology vehicles and what might be the
       potential impacts of these data on inventories.  The result of this analysis is
       discussed in Section 2.3.1.
   •   MOBILE6.2 uses the same toxic to VOC ratios for cold starts and hot running
       operation even though these ratios for benzene and 1,3-butadiene are higher
       during cold starts than hot running.
   •   We have a limited understanding of the impact of off-cycle operation on highway
       vehicle air toxic emissions.
   •   Data are limited for certain sources and pollutants not significant to this rule. For
       heavy-duty highway vehicles (both gasoline and diesel  engines) the toxic to VOC
       ratios used in MOBILE6.2 to develop inventory estimates are based on very
       limited data.  Moreover, we lack data on how diesel fuel properties impact air
       toxic emissions, and we have very little data on mobile source metal emissions.

       There are also significant uncertainties resulting from the use  of national default
data rather than local inputs, as well as "top-down" allocation schemes in estimating toxic
emissions. Examples include use of national default vehicle registration distributions,
default average speed distributions, and use of county level population data to allocate
State or urban level VMT. A recent paper evaluated the impacts of these default inputs
and allocation schemes on local level inventories.40
                                       2-29

-------
Table 2.2.-S. Precursor Pollutants.
Pollutant
Acetaldehyde
1,3-Butadiene
1-Butene
1-2,3-Dimethyl butene
1-2-Ethylbutene
1-2-Methyl butene
1-3-Methyl butene
2-Butene
2-2-Methyl butene
1-Decene
Ethanol
Ethene
1-Heptene
2-Heptene
1-Hexene
2-Hexene
3-Hexene
Precursor for
Formaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert), Acrolein (reactive
and inert)
Formaldehyde (reactive and
inert), Propionaldehyde
(reactive and inert)
Formaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
Acetaldehyde (reactive and
inert)
Acetaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
Acetaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
Acetaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
Acetaldehyde (reactive and
inert)
Propionaldehyde (reactive
and inert)
Pollutant
Isoprene
MTBE
Methanol
1-Nonene
2-Nonene
1-Octene
2-Octene
1-Pentene
1-2,4,4-Trimethyl pentene
1-2-Methyl pentene
1-3-Methyl pentene
1-4-Methyl pentene
2-Pentene
2-3-Methyl pentene
2-4-Methyl pentene
Propene
2-Methylpropene
Precursor for
Formaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
Acetaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
Acetaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
Acetaldehyde (reactive and
inert),
Propionaldehyde (reactive
and inert)
Acetaldehyde (reactive and
inert)
Acetaldehyde (reactive and
inert)
Acetaldehyde (reactive),
Acetaldehyde (inert),
Formaldehyde (reactive and
inert)
Formaldehyde (reactive and
inert)
               2-30

-------
       Finally, as discussed in Section 2.1.3, there are greater uncertainties in projection
year estimates.

       NonroadEquipment - The toxic to VOC ratios in NMIM for lawn and garden
equipment, which makes the single largest contribution of any nonroad sector to the air
toxics inventory, is supported by a large amount of test data.   The VOC estimates for
uncontrolled engines in the NONROAD model are based on a large amount of in-use test
data and peer reviewed methodologies.  Estimates for controlled engines are based on
certification test data and emission standards. However, for a number of source
categories—in particular heavy-duty diesel engines and aircraft engines—the toxic to
VOC ratios used to develop inventory estimates are based on very limited data.  In
addition, the Agency has limited emissions data for nonroad equipment with emission
controls. The Agency has been doing test data to address some of the limitations. This
work is discussed in Sections 2.3.3 and 2.3.4. There are also significant uncertainties
associated with allocating nonroad equipment emissions from the national to the local
level. As with highway sources, future year inventories are more uncertain. Finally, the
relationship between fuel parameters and emission rates for gasoline nonroad equipment
is much more poorly understood than the relationship for highway gasoline vehicles. In
our modeling, we assumed that the impacts of fuel control on emissions from nonroad
equipment would be proportional to the impact on highway vehicle emissions, as
discussed above.

       Stationary Sources — For the 1999 NEI, there are a number of known or
suspected issues for stationary source emissions listed  on the emission inventory website
(U. S. EPA, 2004a).  The issues listed are generally limited to specific geographic areas
and are not expected to influence national-level results. Of these, it is expected that
issues related to acrolein are most likely to affect the results for assessment of noncancer
effects. Another uncertainty concerning the base year  inventory is the proper
identification of sources using the inventory codes. These codes are utilized for applying
growth and reduction factors.

       There are several uncertainties associated with  the growth and reduction
information.  The growth information is uncertain for a number of reasons.  For most
sources, activity growth is used as a surrogate for emissions growth, which  may not be
appropriate for some industry sectors. In addition the growth information available is
from economic models, is typically specific to broad industry categories, and is not
resolved geographically for all categories. The stationary source reductions are uncertain
because they are generally based on national-average reductions (although we have used
facility-specific reductions where available). We do not expect this uncertainty to have
an impact on national-level results.

       As previously mentioned, after 2010, stationary source emissions are based only
on economic growth.  They do not account for reductions from ongoing toxics programs
such as the urban air toxics program, residual risk standards and area source program,
which are expected to further reduce toxics. Furthermore, the 2030 stationary source
                                       2-31

-------
inventory estimates are equal to the 2020 estimates, because of additional uncertainties in
the available growth data past 2020 and the lack of knowledge of the effect of stationary
source control programs that far into the future.

2.2.1.2 Trends in Air Toxic Emissions

2.2.1.2.1  Emission Trends Without Proposed Controls

       In 1999, based on the National Emissions Inventory (NEI), mobile sources
accounted for 44% of total emissions of 188 hazardous air pollutants (see Figure 2.2.-1).
Diesel particulate matter is not included in this list of 188 pollutants.  Sixty-five percent
of the mobile source tons in this inventory were attributable to highway mobile sources,
and the remainder to nonroad sources. Furthermore, over 90% of mobile source air toxic
emissions are attributable to gasoline vehicles and equipment.

       Overall, air toxic emissions are projected to decrease from 5,030,000 tons in 1999
to 4,010,000 tons in 2020, as a result of existing and planned emission controls on major,
area, and mobile sources. In the absence of Clean Air Act emission controls currently in
place, EPA estimates air toxic emissions would total 11,590,000 tons in 2020 (Figure 2.2-
1). As indicated in Figure 2.2.-1, mobile source air toxic emissions will be reduced 60%
between 1999  and 2020 without the controls in this proposal, from 2.2 million to 880,000
tons.  This reduction will occur despite a projected 57% increase in vehicle miles
traveled,  and a 63% projected increase in nonroad activity (See Figures 2.2.-2 and 2.2.-3).
It should be noted, however, that EPA anticipates mobile source air toxic emissions will
begin to increase after 2020, from about 880,000 tons in 2020 to 920,000 tons in 2030.
Benzene  emissions from all sources decrease from about 347,000 tons in 1999 to 222,000
tons in 2020, and as is the case with total air toxic emissions, begin to increase slightly
between 2020  and 2030 (Figure 2.2.-4).

       None of the inventory trends data presented in this  section includes revised estimates of
emissions at cold temperature in vehicles, addition of emissions from portable fuel containers,
and revisions in the gasoline distribution inventory used to estimate emission benefits of the rule
and cost-effectiveness. These revisions are discussed in Section 2.2.2.
                                       2-32

-------
 Figure 2.2.-1.  Contribution of Source Categories to Air Toxic Emissions, 1990 to
2020 (not Including Diesel Particulate Matter).  Dashed Line Represents Projected
                      Emissions without Clean Air Act Controls.
I
o
"H-
CO
c
o
CO
CO
CO


LU
12

11

10

9

8

7

6

5

4

3

2

1
         U.S. Contributions of Source Categories to Total Emissions for all HAPs
           I   I Major
           I   I Area and Other
           I   I Fires - Prescribed and Wild
           I   I Non-Road Mobile
           I   I On-Road Mobile
7.X'
                2.69
                2.55
                1990
                        5.03
                        1.26
                        1.29
                        0.29

                        0.76
                        1.44
                        1999
                                4.09
                0.95
                                1.45
                0.29

                0.65
                                0.77
                                2007
                         3.9
                         0.9
                        1.51
0.29

0.58
                                        0.63
                                        2010
                                   3.S8
                                               .,-
                                                    0.97
                                    1.64
                                                    0.29
                                                    0.49
                                                    0.49

                                                    2015
                                           4.01
                                            1.06
                                            1.78
                                           0.29
                                           0.44
                                                      0.44

                                                       2020
                                           2-33

-------
Figure 2.2.-2.  Trend in Highway Vehicle Air Toxic Emissions Versus VMT, 1990 to
                                             2030.
                           Highway Vehicle Air Toxic Emissions vs. VMT
    1,000,000 --
     500,000 --
            1990    1996    1999   2007    2010    2015   2020    2030
                                                              2,000,000
                                                             -- 1,000,000
   Figure 2.2.-S.  Trend in Emissions of Nonroad Equipment Air Toxic Emissions
 (Excluding Commercial Marine Vessels, Locomotives and Aircraft) versus Activity,
                                        1990 to 2030.
                          Nonroad Equipment Air Toxic Emissions Versus Activity
       900,000


       800,000


       700,000

Emissions
(Tons)
       400,000 -


       300,000 -


       200,000 -


       100,000 -
              1990   1996    19:
                                                                  Activity (hp-hrs)
                                2007    2010   2015   2020   2030

                                   Year
Air Toxic Emissions (tons)
Activity fhp-hrt
                                             2-34

-------
      400000
      350000
      300000
      250000
  Tons 200000
      150000
      100000
      50000
                                    Figure 2.2.-4

                           Trend in Benzene Emissions - 1999 to 2030
• Nonroad
DOnroad
D Fires
• Area and Other
D Major	
               1999       2007      2010      2015
                                       Year
                                                      2020
                                                               2030
       Highway Vehicle Trends - Table 2.2.-3 summarizes nationwide emissions of
individual air toxics from highway vehicles from 1999 to 2030. Fifteen POM compounds
listed in Table 2.2.-1 (except for naphthalene) are grouped together as POM.  For mobile
sources, eighteen percent of the chromium was assumed to be the highly toxic hexavalent
form, based on combustion data from stationary  combustion turbines that burn diesel
fuel.41
                                        2-35

-------
 Table 2.2.-S.  Nationwide Emissions (Tons) of Individual Air Toxic Pollutants from
                               Highway Vehicles.
Pollutant
1,3-Butadiene
2,2,4-Trimethylpentane
Acetaldehyde
Acrolein
Benzene
Chromium VI
Ethyl Benzene
Formaldehyde
n-Hexane
MTBE
Manganese
Naphthalene
Nickel
POM
Propionaldehyde
Styrene
Toluene
Xylenes
1999
23623
166208
29928
3993
170355
4
69480
80677
65164
82570
16
3978
16
460
4209
13168
456344
267324
2007
10876
90621
17049
1974
95766
5
37951
40168
43107
33458
20
2490
19
256
2343
6570
242800
142123
2010
8807
73768
13909
1570
79550
5
30838
32240
35832
28026
22
2229
20
228
1953
5284
196528
115004
2015
6913
58013
11317
1242
63920
6
24165
26150
27727
21124
25
2007
23
208
1621
4200
154225
90182
2020
6468
51820
10721
1170
58109
6
21472
24879
23087
16117
28
1976
26
211
1553
3910
138365
80799
2030
6864
53786
11651
1263
60660
8
22229
27188
23292
15225
36
2255
32
243
1693
4132
143714
83948
       Table 2.2.-4 summarizes total tons of air toxic emissions from highway vehicles
by vehicle class in 1999, 2007, 2010, 2015, 2020, and 2030. Table 2.2.-5 provides the
percentage of total highway vehicle emissions associated with each vehicle class.  In
1999, 54% of air toxic emissions from highway vehicles were emitted by light duty
gasoline vehicles (LDGVs) and 37% by light duty trucks (LDGTs). EPA projects that in
2020, only 27% of highway vehicle HAP emissions will be from LDGVs and 63% will
be from LDGTs. More detailed summaries of emissions by individual pollutant, by State,
and for urban versus rural area can be found in Excel workbooks included in the docket
for this rule.
                                      2-36

-------
  Table 2.2.-4. Tons of Air Toxic Emissions from Highway Vehicle Classes, 1999 to
                  2030 (Not Including Diesel Particulate Matter).
Vehicle Type
HDDV
HDGV
LDDT
LDDV
LDGT1
LDGT2
LDGV
MC
Total Highway
Emissions (tons/yr)
1999
38,534
80,227
1,279
977
342,839
186,078
778,772
8,826
1,437,532
2007
26,923
35,096
766
139
239,534
139,447
317,021
8,691
767,617
2010
23,707
24,838
617
60
208,636
126,396
232,547
9,035
625,836
2015
20,570
17,342
552
34
177,486
114,204
153,050
9,854
493,092
2020
20,435
13,666
491
23
170,855
105,843
118,762
10,673
440,748
2030
23,336
12,023
402
22
179,122
102,085
128,305
12,957
458,252
HDDV: Heavy Duty Diesel Vehicles
HDGV: Heavy Duty Gasoline Vehicles
LDDT: Light Duty Diesel Tracks
LDDV: Light Duty Diesel Vehicles
LDGT1: Light Duty Gasoline Tracks 1
LDGT2: Light Duty Gasoline Tracks 2
LDGV: Light Duty Gasoline Vehicles
MC: Motorcycles
Table 2.2.-S. Percent Contribution of Vehicle Classes to Highway Vehicle Air Toxic
         Emissions, 1999 to 2020 (Not Including Diesel Particulate Matter).
Vehicle
LDGV
LDGT1 and 2
HDGV
HDDV
Other (motorcycles and
light duty diesel
vehicles and trucks)
1999
54%
37%
6%
3%
1%
2007
41%
49%
5%
4%
1%
2010
37%
53%
4%
4%
1%
2015
31%
59%
4%
4%
2%
2020
27%
63%
3%
5%
2%
2030
28%
61%
3%
5%
3%
       Tables 2.2.-6 through 2.2.-11 summarize total tons of emissions nationwide for
benzene, 1,3-butadiene, formaldehyde, acetaldehyde, naphthalene, and acrolein from
highway vehicles. About 87% of benzene emissions from gasoline vehicles were in
exhaust, with the remainder in evaporative and refueling emissions. Benzene emissions
from diesel vehicles were all exhaust. There are no evaporative emissions of 1,3-
butadiene, formaldehyde, acetaldehyde, and acrolein.
                                      2-37

-------
 Table 2.2.-6.  Tons of Benzene Emissions from Highway Vehicle Classes, 1999 to
                                   2030.
Vehicle Type
HDGV
HDDV
LDDT
LDDV
LDGT1
LDGT2
LDGV
MC
Total Highway
Emissions
(tons/yr)
1999
7967
2674
167
120
42433
20638
95591
764
170355
2007
4041
1872
100
17
30773
17701
40478
784
95766
2010
2970
1650
82
7
27498
16805
29722
817
79550
2015
2152
1434
74
4
23835
15694
19835
892
63920
2020
1760
1426
67
3
23346
14897
15643
967
58109
2030
1539
1628
57
3
24856
14505
16895
1177
60660
Table 2.2.-7. Tons of 1,3-Butadiene Emissions from Highway Vehicle Classes, 1999
                                  to 2030.
Vehicle Type
HDGV
HDDV
LDDT
LDDV
LDGT1
LDGT2
LDGV
MC
Total Highway
Emissions
(tons/yr)
1999
1507
1430
64
44
5132
3483
11743
220
23623
2007
483
995
38
6
3218
1919
3983
234
10876
2010
260
877
31
3
2801
1735
2855
244
8807
2015
130
760
29
1
2307
1524
1895
266
6913
2020
103
755
26
1
2291
1503
1500
288
6468
2030
84
859
23
1
2447
1486
1614
350
6864
                                    2-38

-------
Table 2.2.-S. Tons of Formaldehyde Emissions from Highway Vehicle Classes, 1999
                                  to 2030.
Vehicle Type
HDGV
HDDV
LDDT
LDDV
LDGT1
LDGT2
LDGV
MC
Total Highway
Emissions
(tons/yr)
1999
6648
19887
495
391
14907
9809
27957
582
80677
2007
2242
13921
297
56
8540
5264
9239
609
40168
2010
1309
12272
238
24
6787
4164
6811
635
32240
2015
741
10663
211
14
5572
3628
4628
693
26150
2020
599
10601
186
9
5516
3513
3705
751
24879
2030
498
12109
148
9
5975
3509
4028
912
27188
Table 2.2.-9.  Tons of Acetaldehyde Emissions from Highway Vehicle Classes, 1999
                                  to 2030.
Vehicle Type
HDGV
HDDV
LDDT
LDDV
LDGT1
LDGT2
LDGV
MC
Total Highway
Emissions
(tons/yr)
1999
1569
7568
200
164
5766
3433
11057
171
29928
2007
722
5310
120
24
3947
2411
4311
204
17049
2010
465
4682
96
10
3265
2023
3155
214
13909
2015
297
4071
84
6
2714
1789
2123
233
11317
2020
245
4049
73
4
2682
1726
1690
253
10721
2030
209
4633
57
4
2899
1710
1831
309
11651
                                    2-39

-------
 Table 2.2.-10. Tons of Acrolein Emissions from Highway Vehicle Classes, 1999 to
                                     2030.
Vehicle Type
HDGV
HDDV
LDDT
LDDV
LDGT1
LDGT2
LDGV
MC
Total Highway
Emissions
(tons/yr)
1999
714
807
24
16
661
357
1396
18
3993
2007
177
561
14
2
434
255
511
19
1974
2010
79
494
12
1
368
222
374
20
1570
2015
25
429
11
1
306
198
251
22
1242
2020
18
425
10
0
302
191
199
24
1170
2030
12
483
9
0
326
188
215
29
1263
 Table 2.2.-11.  Tons of Naphthalene Emissions from Highway Vehicle Classes, 1999
                                    to 2030.
Vehicle Type
HDGV
HDDV
LDDT
LDDV
LDGT1
LDGT2
LDGV
MC
Total Highway
Emissions
(tons/yr)
1999
752
172
6
7
766
491
1758
26
3978
2007
540
98
3
1
612
260
950
27
2490
2010
388
67
2
0
645
268
831
28
2229
2015
241
33
1
0
702
274
726
30
2007
2020
189
20
1
0
774
281
678
33
1976
2030
170
16
1
0
906
316
807
40
2255
      NonroadEquipment Trends — Table 2.2.-12 summarizes nationwide emissions of
individual air toxics from nonroad equipment, from 1999 to 2030. The lead emissions in
the table are from piston engine aircraft, which use leaded gasoline.  Table 2.2.-13
summarizes total tons of air toxic emissions from categories of nonroad equipment by
equipment type in 1999, 2007, 2010, 2015, 2020, and 2030. Table 2.2.-14 provides the
percentage of total nonroad equipment emissions associated with each equipment type.
Air toxic emissions from nonroad equipment are dominated by lawn and garden
equipment, recreational equipment, and pleasure craft, which collectively account for
almost 80% of nonroad HAP emissions in all years. More detailed summaries of
emissions by individual pollutant, by State, and for urban versus rural area can be found
in Excel workbooks included in the docket for this rule.
                                      2-40

-------
   Table 2.2.-12.  Nationwide Emissions of Individual Air Toxics from Nonroad
                       Equipment, from 1999 to 2030.

Pollutant
1,3-Butadiene
2,2,4-Trimethylpentane
Acetaldehyde
Acrolein
Benzene
Chromium VI
Ethyl Benzene
Formaldehyde
n-Hexane
Lead
MTBE
Manganese
Naphthalene
Nickel
POM
Propionaldehyde
Styrene
Toluene
Xylenes
Annual Total Nonroad Emissions tons)
1999
9718
94546
23479
3083
65360
4
42731
56254
28765
550
24338
5
1254
34
356
4735
4254
205186
193016
2007
7906
81056
19333
2655
54232
4
36719
45526
25230
551
10922
6
1236
36
320
3792
3604
192855
160347
2010
6799
71985
17390
2496
46951
4
32395
41214
22784
565
9569
6
1214
38
302
3358
3091
173428
140968
2015
6298
59516
15425
2360
42031
4
27587
36911
19872
587
8819
7
1258
39
290
2956
2735
143943
118662
2020
6237
51944
14516
2330
40444
4
25260
34979
18451
609
8664
7
1318
41
284
2765
2606
125562
107495
2030
6765
51957
14988
2505
43252
5
26660
36320
19464
654
9459
8
1465
45
300
2827
2802
127370
112660
Table 2.2.-13. Tons of Air Toxic Emissions from Nonroad Equipment Types, 1999
              to 2030 (Not Including Diesel Particulate Matter).
Equipment type
Agriculture
Aircraft
Airport Support
Commercial
Commercial Marine Vessel
Construction
Industrial
Lawn/Garden
Logging
Pleasure Craft
Railroad
Recreational
Underground Mining
Total Nonroad
Emissions (tons/yr)
1999
23,098
14,276
421
46,990
8,736
39,675
14,559
196,257
3,816
258,190
4,416
146,526
176
759,565
2007
15,954
14,315
311
33,732
9,557
25,138
7,456
115,652
2,325
172,930
4,143
244,129
155
647,754
2010
13,476
14,965
251
27,281
9,742
21,702
5,114
99,485
2,339
144,245
3,984
231,291
138
575,831
2015
10,546
16,081
206
29,004
10,213
17,937
3,157
101,535
2,394
122,057
3,896
171,593
112
490,454
2020
8,530
17,256
191
31,451
10,973
15,609
2,573
109,328
2,562
111,936
3,758
128,661
100
444625
2030
7,129
19,603
205
36,981
13,354
14,303
2,382
125,823
3,054
108,260
3,531
124,142
104
460,627
                                   2-41

-------
 Table 2.2.-14. Contribution of Equipment Types to Nonroad Air Toxic Emissions,
              1999 to 2020 (not Including Diesel Particulate Matter).
Equipment
Type
Lawn and
Garden
Pleasure Craft
Recreational
All Others
1999
26%
34%
19%
21%
2007
18%
27%
38%
17%
2010
17%
25%
40%
18%
2015
21%
25%
35%
19%
2020
25%
25%
29%
21%
2030
27%
24%
27%
22%
       Almost 90% of nonroad toxic emissions are from 2-stroke and 4-stroke gasoline
engines, with the remainder from diesel engines and turbine engine aircraft. Similarly,
almost 90% of benzene emissions from nonroad equipment are from gasoline engines,
and these emissions would be reduced by a fuel benzene standard.

       Tables 2.2.-15 through 2.2.-20 summarize total tons of emissions nationwide for
benzene, 1,3-butadiene, formaldehyde, acetaldehyde, naphthalene, and acrolein from
nonroad equipment types.

Table 2.2.-15. Tons of Benzene Emissions from Nonroad Equipment Types, 1999 to
                                     2030.
Equipment Type
Agriculture
Aircraft
Airport Support
Commercial
Commercial Marine
Vessel
Construction
Industrial
Lawn/Garden
Logging
Pleasure Craft
Railroad
Recreational
Underground Mining
Total Nonroad
Emissions
(tons/yr)
1999
2203
1102
44
6809
644
3601
1976
20451
267
20304
162
7781
15
65360
2007
1569
1114
33
5323
705
2310
986
14729
185
14177
150
12938
13
54232
2010
1323
1163
26
4206
719
1957
633
12112
180
12113
144
12365
12
46951
2015
1058
1247
21
4529
753
1639
368
12039
177
10507
140
9544
10
42031
2020
877
1335
20
4964
809
1450
291
12960
187
9787
134
7622
9
40444
2030
744
1511
22
5906
982
1348
258
14941
221
9598
125
7587
9
43252
                                     2-42

-------
Table 2.2.-16. Tons of 1,3-Butadiene Emissions from Nonroad Equipment Types,
                              1999 to 2030.
Equipment Type
Agriculture
Aircraft
Airport Support
Commercial
Commercial Marine
Vessel
Construction
Industrial
Lawn/Garden
Logging
Pleasure Craft
Railroad
Recreational
Underground Mining
Total Nonroad
Emissions
(tons/yr)
1999
243
824
7
1140
6
407
302
3423
44
2071
114
1136
1
9718
2007
176
821
5
892
6
259
143
2445
29
1423
107
1600
1
7906
2010
148
859
3
683
6
214
88
1933
29
1201
104
1530
1
6799
2015
120
924
3
738
6
182
50
1887
29
1018
102
1238
1
6298
2020
101
993
3
813
6
165
39
2030
31
928
99
1029
1
6237
2030
85
1131
3
972
7
156
33
2342
36
895
94
1009
1
6765
Table 2.2.-17. Tons of Formaldehyde Emissions from Nonroad Equipment Types,
                              1999 to 2030.
Equipment Type
Agriculture
Aircraft
Airport Support
Commercial
Commercial Marine
Vessel
Construction
Industrial
Lawn/Garden
Logging
Pleasure Craft
Railroad
Recreational
Underground Mining
Total Nonroad
Emissions
(tons/yr)
1999
9816
6549
139
3418
4715
12417
3046
6867
432
4136
1901
2731
87
56254
2007
6671
6505
105
2907
5153
8958
1790
4727
248
2848
1793
3743
77
45526
2010
5630
6809
90
2435
5252
7742
1404
3830
214
2447
1730
3562
68
41214
2015
4288
7333
71
2236
5499
5937
963
3678
167
2105
1690
2890
55
36911
2020
3363
7885
63
2131
5899
4779
832
3856
155
1932
1629
2404
50
34979
2030
2749
8990
65
2128
7152
4074
837
4371
163
1879
1529
2333
51
36320
                                  2-43

-------
 Table 2.2.-18. Tons of Acetaldehyde Emissions from Nonroad Equipment Types,
                               1999 to 2030.
Equipment Type
Agriculture
Aircraft
Airport Support
Commercial
Commercial Marine
Vessel
Construction
Industrial
Lawn/Garden
Logging
Pleasure Craft
Railroad
Recreational
Underground Mining
Total Nonroad
Emissions
(tons/yr)
1999
4493
2019
63
1400
2364
5723
1350
2478
176
1703
853
820
39
23479
2007
3058
2004
49
1270
2588
4138
857
1920
102
1179
805
1330
34
19333
2010
2581
2098
42
1071
2639
3578
676
1548
85
1002
776
1264
31
17390
2015
1966
2259
33
975
2768
2745
459
1480
62
854
758
1041
25
15425
2020
1542
2430
29
920
2974
2210
389
1546
55
782
731
886
22
14516
2030
1260
2770
30
906
3619
1883
381
1748
55
757
686
870
23
14988
Table 2.2.-19. Tons of Acrolein Emissions from Nonroad Equipment Types, 1999 to
                                   2030.
Equipment Type
Agriculture
Aircraft
Airport Support
Commercial
Commercial Marine
Vessel
Construction
Industrial
Lawn/Garden
Logging
Pleasure Craft
Railroad
Recreational
Underground Mining
Total Nonroad
Emissions
(tons/yr)
1999
285
968
6
156
98
392
119
388
16
316
131
206
2
3083
2007
194
960
4
127
109
280
71
252
9
212
124
312
2
2655
2010
164
1005
4
105
112
241
55
207
8
179
120
295
2
2496
2015
125
1083
3
99
118
186
38
201
7
152
117
228
1
2360
2020
98
1165
3
98
129
151
33
212
7
139
113
180
1
2330
2030
81
1329
3
102
161
130
34
241
8
134
107
176
1
2505
                                   2-44

-------
  Table 2.2.-20. Tons of Naphthalene Emissions from Nonroad Equipment Types,
                                  1999 to 2030.
Equipment Type
Agriculture
Aircraft
Airport Support
Commercial
Commercial Marine
Vessel
Construction
Industrial
Lawn/Garden
Logging
Pleasure Craft
Railroad
Recreational
Underground Mining
Total Nonroad
Emissions
(tons/yr)
1999
49
456
1
98
65
61
30
261
4
112
61
56
0
1254
2007
36
475
1
106
69
46
18
245
4
103
51
81
0
1236
2010
32
496
1
98
68
42
15
224
4
100
44
90
0
1214
2015
26
530
1
108
72
32
9
232
4
101
42
101
0
1258
2020
21
566
1
119
79
23
6
251
4
104
40
105
0
1318
2030
15
638
1
142
102
16
4
289
5
110
35
109
0
1465
       Diesel Paniculate Matter -- If diesel particulate matter emissions were added to
the mobile source total mass of air toxic emissions, mobile sources would account for
48% of a total 5,398,000 tons in 1999.  Table 2.2.-21 summarizes the trend in diesel
particulate matter between 1999 and 2030, by source category. As controls on highway
diesel engines and nonroad diesel engines phase in, diesel-powered locomotives and
commercial marine vessels increase from 11% of the inventory in 1999 to 27% in 2020.

     Table 2.2.-21. Percent Contribution of Mobile Source Categories to Diesel
   Particulate Matter Emissions, 1999 to 2020 in Tons Per Year (Percent of Total).
Source
Highway Vehicles
Commercial
Marine Vessels
Locomotives
Other Nonroad
Equipment
1999
144,000
(39%)
20,000
(5%)
21,000
(6%)
183,000
(50%)
2007
85,000
(33%)
19,000
(7%)
18,000
(7%)
134,000
(52%)
2010
63,000
(30%)
18,000
(8%)
15,000
(7%)
118,000
(55%)
2015
38,000
(25%)
17,000
(11%)
14,000
(9%)
83,000
(55%)
2020
30,000
(26%)
17,000
(15%)
14,000
(12%)
53,000
(46%)
2.2.1.2.2 Impact on Inventory of Proposed Fuel Benzene Control

       The fuel benzene control proposed in this rule would reduce benzene emissions
from highway gasoline vehicles, nonroad gasoline equipment, gasoline distribution and
portable fuel containers. The total benzene emissions reduced in the inventories used for
                                      2-45

-------
air quality modeling for these sectors are 12,800 tons, or 6 percent of the national
benzene inventory from all sources. It should be emphasized that the air quality,
exposure and risk modeling inventory underestimates the total emissions benefit since it
does not account for portable fuel container emissions and underestimates cold
temperature emissions for highway vehicles. For inventories which include these
emissions, see Section 2.2.2.2.

       Table 2.2.-22 summarizes the nationwide impact of the proposed benzene
standard on emissions of key air toxics from highway vehicles in 2015, 2020, and 2030.
Although EPA's MOBILE emissions model estimates very small increases in emissions
of 1,3-butadiene, formaldehyde, and acetaldehyde, the reductions in benzene emissions
are dramatic,  roughly 11 to  12%. Similar impacts are projected for nonroad equipment
(Tables 2.2.-23 and 2.2.-24). In addition, fuel benzene controls would reduce emissions
within the gasoline distribution sector, and during vehicle refueling. Table 2.2.-25
presents estimated reductions for these sources in 2015 and 2020, which total over 2000
tons per year.  These vehicle refueling and gasoline distribution reductions are also based
on inventory projections from the 1999 NEI, as discussed above.  However, subsequent
to the air quality, exposure and risk modeling for this rule, new emission estimates for
this sector were released as part of the 2002 NEI42.  These revisions are discussed in
Section 2.2.2, and were used in developing estimates of emission benefits for this rule.
More detailed summaries of emissions by individual pollutant, by State, and for urban
versus rural area can be found in Excel workbooks included in the docket for this rule.
                                       2-46

-------
Table 2.2.-22. Nationwide Impact of the Proposed Benzene Control on Emissions of Key Air Toxics from Highway Vehicles in
                                              2015, 2020, and 2030.

Pollutant
1,3-Butadiene
Acetaldehyde
Acrolein
Benzene
Formaldehyde
5 MSAT Total
Annual Emissions (tons) by Vehicle Ty
2015
Reference
Case
6913
11317
1242
63920
26150
109542
2015
Control
Case
6926
11336
1242
56596
26195
102295
2015
Reduction
-14
-19
0
7324
-45
7247
2020
Reference
Case
6468
10721
1170
58109
24879
101347
2020
Control
Case
6480
10738
1170
51711
24921
95020
2020
Reduction
-13
-17
0
6398
-41
6327
pe
2030
Reference
Case
6864
11651
1263
60660
27188
107626
2030
Control
Case
6877
11669
1263
54154
27231
101194
2030
Reduction
-13
-18
0
6506
-43
6433
                                                      2-47

-------
Table 2.2.-2S. Nationwide Impact of the Proposed Benzene Control on Emissions of Key Air Toxics from all Nonroad
                                    Equipment in 2015, 2020, and 2030.

Pollutant
1,3-Butadiene
Acetaldehyde
Acrolein
Benzene
Formaldehyde
5 MSAT Total
Annual Emissions (tons) by Vehicle Ty
2015
Reference
Case
6298
15425
2360
42031
36911
103025
2015
Control
Case
6310
15435
2360
37531
36940
98576
2015
Reduction
-12
-10
0
4500
-29
4449
2020
Reference
Case
6237
14516
2330
40444
34979
98505
2020
Control
Case
6249
14525
2330
36022
35007
94132
2020
Reduction
-12
-9
0
4422
-28
4373
pe
2030
Reference
Case
6765
14988
2505
43252
36320
103830
2030
Control
Case
6778
14998
2505
38489
36350
99120
2030
Reduction
-13
-10
0
4763
30
4710
                                                  2-48

-------
Table 2.2.-24. Nationwide Impact of the Proposed Benzene Control on Emissions of Key Air Toxics from Gasoline Nonroad
                                      Equipment in 2015, 2020, and 2030.
Annual Emissions (tons) for Gasoline Nonroad Equipment
Pollutant
1,3 -Butadiene
Acetaldehyde
Acrolein
Benzene
Formaldehyde
5 MSAT Total
2015
Reference
Case
5071
3663
632
37747
9423
56535
2015
Control
Case
5083
3672
632
33247
9452
52087
2015
Reduction
-12
-10
0
4500
-29
4448
2020
Reference
Case
4982
3558
591
36440
9103
54675
2020
Control
Case
4994
3567
591
32018
9131
50302
2020
Reduction
-12
-9
0
4422
-28
4373
2030
Reference
Case
5401
3807
625
39163
9740
58736
2030
Control
Case
5413
3817
625
34399
9770
54025
2030
Reduction
-13
-10
0
4763
-30
4711
                                                    2-49

-------
    Table 2.2.-25. Nationwide Impact of the Proposed Controls on Emissions of
 Benzene from Vehicle Refueling and Gasoline Distribution in 2015 and 2020 (2030
                        Assumed to be the Same as 2020).



Vehicle Refueling
Gasoline Distribution
2015
Reference
Case
724
5419
2015
Control
Case
459
3663

2015
Reduction
265
1756
2020
Reference
Case
720
5606
2020
Control
Case
459
3804

2020
Reduction
261
1802
2.2.2   Emission Reductions from Proposed Controls

       Section 2.2.2 describes revisions made to emission inventories subsequent to
development of MS AT inventories for air quality modeling.  These include revised estimates of
emissions at cold temperature in vehicles, addition of emissions from portable fuel containers,
and revisions in the gasoline distribution inventory to reflect changes made for the 2002 National
Emissions Inventory. The revised inventories were used to estimate  emission benefits of the rule
and cost-effectiveness.

2.2.2.1 Methodology Changes from Air Quality Inventories

2.2.2.1.1 Highway Vehicles

       Section 2.1.1.1 describes the changes made to hydrocarbon emission rates in MOBILE6.2
to reflect the higher measured emissions during cold  starts at cold temperature for Tier 1 and
later vehicles.  Since the algorithms used to calculate toxic to hydrocarbon emission ratios in
MOBILE6.2 do not vary with temperature, reductions in hydrocarbon emissions result in
proportional reductions in air toxic emissions.

       The assumption in MOBILE6.2 that reductions in air toxic emissions are proportional to
hydrocarbon emission reductions was based on testing done at temperatures ranging from -20 to
75 °F in EPA's Office of Research  and Development in the late 1980's.43'44 These studies found
that, overall, the composition of hydrocarbon emissions did not vary appreciably with
temperature, although fractions of formaldehyde increased somewhat with lower temperature in
port fuel injected vehicles.  The validity of the assumption was re-evaluated for later model
vehicles.

       EPA's Office of Research and Development recently tested several later model vehicles
as the same temperature ranges cited above.45'46'47  The results of the test program are
unpublished, but are included in the docket for the rule. Vehicles included in the test program
were a 1993 Chevrolet Cavalier, a  1987 and 1993 Ford Taurus, a 1996 Chrysler Concord, a 2001
Ford Focus, a 1993 Buick Regal, and a  2001 Dodge Intrepid. This test program found increasing
emissions of individual air toxics at lower temperatures. Benzene and 1,3-butadiene emissions
increased proportionally with hydrocarbon emissions, with a very strong correlation. However,
                                      2-50

-------
correlations were not as strong with aldehydes. Results from the 1993 Cavalier and 1993 Taurus
found a statistically significant correlation for acetaldehyde but not for formaldehyde, whereas
analysis of data from the other vehicles found a correlation for formaldehyde but not
acetaldehyde.

        A major vehicle manufacturer also recently tested two Tier 2 compliant vehicles at 75
and 20 °F.  Although the data are confidential, they show emission of air toxics increase at the
same rate as hydrocarbons, with a very high correlation.

       A third source of data is testing done by Southwest Research Institute for U. S. EPA,
Office of Transportation and Air Quality on four model year 2005 vehicles - a Ford F-150, a
Mazda 3, a Honda Odyssey and  a Chevrolet Equinox.48 The four vehicles were tested at 0, 20
and 75 °F.  Benzene and  1,3-butadiene correlated very strongly with hydrocarbon emissions,
with r-square values above 0.9.  Benzene accounted for about 3.6 percent of exhaust non-
methane hydrocarbon emissions at all temperatures, while 1,3-butadiene accounted for about
0.3%.  However, formaldehyde and acetaldehyde fractions appeared to decrease with decreasing
temperature. When data for the  largest vehicle, the Ford F-150, were removed, there seemed to a
stronger correlation between aldehyde emissions and non-methane hydrocarbons. This could be
because this larger engine is running richer during cold starts than the other vehicles, and not
enough oxygen is available for aldehyde formation.

       Recent EPA testing of a Chevrolet Trailblazer, with its engine recalibrated to meet the
proposed cold temperature standard, showed reductions in acetaldehyde and acrolein
proportional to the reduction in VOC. Formaldehyde was also reduced, but was not reduced  as
much as acetaldehyde and acrolein. Other air toxic compounds, including benzene, were not
included in this testing. Figure 2.2.-5 depicts the relationship between carbonyl compounds and
NMHC.

Figure 2.2.-S.  Regressions of Carbonyl Emissions Versus NMHC for Chevrolet Trailblazer
                    Recalibrated to Meet Cold Temperature Standard.
Acetaldehyde vs. NMHC
7 _,
c
•S 5
•= 4 .

0)
09
•1
n -
y - 0 OOOSx + 1 66

-—^2****^
+±****r*^
t**^



0 1000 2000 3000 4000 5000 6000
NMHC
                                      2-51

-------
c.
—• 4 _
-§
"3)
EQ
C
• u— 7F nclv+ R °1 7 ...

^ • R^= 0.0082
•



0 1000 2000 3000 4000 5000 6000
NMHC (mg/mi)
          Given available data, we have concluded it is reasonable to retain the assumption that
ratios of toxic emissions to hydrocarbon emission do not vary with temperature. However, as
more data become available, this assumption should be reevaluated, particularly for aldehydes.

2.2.2.1.2 Nonroad Equipment

No changes were made from the inventory estimates of nonroad equipment that were
developed for air quality modeling.  In estimating the emission reductions from proposed
controls, no changes were made from the inventory estimates, with and without the
proposed fuel benzene control, developed for air quality modeling. It should be noted,
however, that EPA recently released newer versions of NONROAD and NMIM,
NONROAD2005 and NMEVI2005 that include a number of significant revisions. 49'50
Most importantly, there are new evaporative categories for tank permeation, hose
permeation, hot soak, and running loss emissions.  If these revisions were included in the
estimation of emission reductions from the proposed fuel benzene control, the estimated
reductions would be larger.
                                      2-52

-------
2.2.2.1.3 Portable Fuel Containers

Any MSATs contained in the liquid gasoline will be present as a component of the VOCs
associated with the PFCs. Specifically, the VOC emissions (estimated in Sections 2.1.1.2
and 2.1.3) will contain the following eight MSATs:

       benzene
       MTBE
       n-hexane
       toluene
       xylenes
       ethylbenzene
       naphthalene
       2,2,4-trimethylpentane

We estimated only nationwide emission totals for all MSATs except benzene, where
State level totals were estimated.

       For all compounds except benzene and MTBE, the fraction of total PFC
emissions that is composed of each of those HAPs is assumed to be directly proportional
to the ratio of each of those HAPs in total evaporative emissions from light-duty gasoline
vehicles. These ratios were obtained from the database of toxic to VOC ratios in the
NMEVI model, discussed in previous sections. NMEVI has ratios that vary by fuel type
(conventional or baseline gasoline, ethanol oxygenated gasoline, and MTBE oxygenated
gasoline). Based on the sales of the various gasoline blends, we generated the ratios
given in Table 2.2.-26.
Table 2.2.-26. Ratios of Pollutants to Total Evaporative VOC Emissions.
Pollutant Name
Naphthalene
Ethyl Benzene
Toluene
n-Hexane
2,2,4-
Trimethylpentane
Xylenes
Baseline
0.0004
0.0077
0.0413
0.0234
0.0158
0.0223
10%
Gasohol
0.0004
0.0045
0.0195
0.0096
0.0158
0.0119
MTBE
Gasoline
0.0004
0.0063
0.0276
0.0087
0.0158
0.0188
Weighted
Ratios
0.0004
0.0067
0.0337
0.0175
0.0158
0.0192
       In this table, the weighted ratios are based on the estimate that the nationwide
distribution of gasoline is 58.3 percent baseline, 23.5 percent gasohol (i.e., E10), and the
                                      2-53

-------
remainder (18.2 percent) oxygenated with MTBE. This estimate is based on 2003 sales
data for ethanol oxygenated gasoline compiled by the Federal Highway Administration,
estimates of reformulated gasoline sales from the Energy Information Administration,
and estimates of the amount of MTBE oxygenated gasoline sold as part of the Federal
Reformulated Gasoline Program from Federal Reformulated Gasoline surveys.51'52

       Because of the localized use of MTBE in gasoline, we used a different approach to
estimate nationwide emissions of this pollutant.  The nationwide quantity of MTBE emitted by
PFCs by permeation or evaporation was estimated based on the ratio of nationwide MTBE
refueling emissions in the 2002 NEI to total VOC refueling emissions. The resulting ratio was
0.024.  Since several States have eliminated the use of MTBE in reformulated gasoline, and
further reductions in the use of MTBE are anticipated in the future, this approach likely
overestimates MTBE emissions from PFCs in future years.

       Another approach was used to estimate emissions of benzene with and without
PFC control, and also with and without the fuel benzene standard proposed in this rule.
We assumed that the fraction of PFC emissions that is benzene is proportional to the
benzene fraction in refueling  emissions. First, we divided State-level benzene refueling
emissions by State-level VOC refueling emissions estimated by NMIM, for both
reference and control case scenarios. The resultant ratios were multiplied by VOC
emissions from evaporation, vapor displacement, and spillage.

       A separate ratio was used for permeation emissions since recent research suggests
that the ratio of benzene from permeation is higher than for evaporation, vapor
displacement or spillage. Thus, we also needed  to split the "permeation plus
evaporation" estimates in Table 2.1 .-1.  Analyses (referenced in Section  2.1.1.2) suggest
that the permeation emissions account for 33.87  percent of the combined permeation plus
evaporation for the sealed PFCs. As noted, a recent study53 suggests that the ratio of
benzene from permeation to total VOC  from permeation is about 1.7727 times higher
than the ratio associated with evaporation.  Thus, we multiplied the benzene refueling
ratios for each state by 1.7727 to obtain the benzene to VOC ratios for permeation.

       It should be noted that because the PFC inventories for air toxics include
emissions spillage while refueling nonroad equipment, and because estimates of nonroad
equipment evaporative emissions in NONROAD also include this source of emissions,
there is some double counting of overall air toxics emissions and emission benefits from
fuel benzene control (This is not an issue for estimates of VOCs).  However, the spillage
component of evaporative emissions in NONROAD is significantly smaller than the
estimates in the PFC inventory, and the double counting accounts for well under 1% of
the total emission benefits of fuel benzene control.
                                      2-54

-------
2.2.2.1.4 Gasoline Distribution

       Subsequent to the development of the gasoline distribution inventories used in the
modeling of air quality, exposure, and risk from mobile source air toxics, EPA improved its
methodology for estimating gasoline distribution emissions in the 2002 National Emissions
Inventory (NEI). The key changes were:

1) Vehicle refueling emissions were estimated as part of the highway vehicle inventory using
   NMTM.  Details of how the modeling was done can be found in the documentation for the
   mobile source 2002 NEI.54 The previous methodology is described in the nonpoint 1999 NEI
   documentation.55  IN this older method, national VOC emissions were calculated using fuel
   sales data and estimates of emissions per fuel volume in areas with and without Stage 2 vapor
   recovery systems. Air toxic emissions were estimated from VOC by applying speciation
   profiles for different fuel types, such as baseline gasoline, MTBE oxygenated gasoline, and
   ethanol oxygenated gasoline.  Total emissions for each combination of vapor recovery
   system and fuel type were allocated to individual counties using vehicle miles traveled.

2) For all other source categories in the gasoline distribution sector, EPA used an improved set
   of methods. These improvements include:  (a) for source categories where activity-based
   emission factors were available (all except bulk terminals and pipelines), EPA established
   methods that maintain mass balance for storage and transfer activities, such that there is
   agreement with the activity estimates used for each of the different distribution sectors; (b)
   EPA developed criteria pollutant and air toxic emission estimates using the same activity
   data and a consistent set of speciation profiles; and (c) EPA accounted for local differences in
   fuel properties for downstream emissions (e.g. bulk plants, transit, unloading, storage, Stage
   1 evaporative losses).  More details on these improvements can be found in a technical
   memorandum on the website for the 2002 NEI.56

       The results of these changes were a significant increase in the air toxic inventory
estimates for vehicle refueling and a small increase nationwide for other sources of gasoline
distribution emissions. County-level estimates for some gasoline distribution sources changed
considerably since local differences in fuel properties were accounted for. Table 2.2.-27
compares benzene estimates in the 1999 NEI, final version 3, and the final 2002 NEI.

Table 2.2.-17. Vehicle Refueling and Gasoline Distribution Benzene Emissions (Tons), 1999
                                     and 2002 NEI.

Vehicle Refueling
Gasoline Distribution
1999 NEI
1558
4978
2002 NEI
2129
5119
% Difference
+36
+3
In order to develop better estimates of the emission benefits of the proposed fuel benzene control
in this rule, EPA developed updated air toxic inventories for vehicle refueling and gasoline
distribution to reflect the changes made in the 2002 NEI. The changes were made as follows:
                                       2-55

-------
1)  Vehicle refueling emissions were estimated using NMIM projections. Refueling emissions
    were estimated for reference case inventories in 1999, 2007, 2010, 2015, 2020 and 2030.
    Control case inventories were estimated for 2015, 2020 and 2030.

2)  For other gasoline distribution emissions, for each air toxic pollutant, EPA estimated a
    national-scale adjustment factor as follows:

    Adjustment factor = 2002 NEI national emissions/2002 national emissions estimated from
    interpolation of the 1999 NEI and the 2007 projection for air quality, exposure and risk
    modeling

3)  EPA developed new county level reference case inventories for these pollutants by applying
    these adjustment factors to county-level gasoline distribution emissions. Revised inventories
    were developed for years 1999, 2007, 2010, 2015, 2020, and 2030.

4)  EPA developed new control case inventories  for gasoline distribution, for benzene, for years
    2015, 2020, and 2030.  These revised county  level inventories were estimated by applying
    the following ratios:

       emissions original control scenario/emissions original reference case

2.2.2.2 Estimated Reductions for Air Toxic Pollutants of Greatest Concern

2.2.2.2.1 Fuel Benzene Standard

       Highway Gasoline Vehicles - The proposed fuel benzene standard will reduce emissions
from light-duty gasoline vehicles and trucks, motorcycles, and heavy-duty gasoline trucks.
Tables 2.2-28, 2.2.-29, and 2.2.-30 present nationwide  benzene emissions for these vehicle
classes with and without the proposed fuel standard in  2015, 2020, and 2030.  Total benzene
emissions from these vehicle classes were  178,000 tons in 1999. Since impacts of fuel benzene
control on emissions of other MSATs are negligible (see Section 2.2.1.2), they are not presented
here, although they are available in the docket for the rule.

   Table 2.2.-2S.  Impact of Fuel Benzene Control on Benzene Emissions from Highway
                                 Vehicle Classes,  2015.
Vehicle Class
LDGV
LDGT1
LDGT2
MC
HDGV
Reference Case
Tons
39,485
41,796
20,074
728
1,715
Control Case Tons
35,253
37,296
17,834
626
1,503
Reduction
4,232
4,500
2,240
102
212
                                      2-56

-------
   Table 2.2.-29. Impact of Fuel Benzene Control on Benzene Emissions from Highway
                                 Vehicle Classes, 2020.
Vehicle Class
LDGV
LDGT1
LDGT2
MC
HDGV
Reference Case
Tons
37,635
47,352
21,083
787
1,399
Control Case Tons
33,730
42,391
18,822
677
1234
Reduction
3,905
4,961
2,261
110
165
   Table 2.2.-30. Impact of Fuel Benzene Control on Benzene Emissions from Highway
                                 Vehicle Classes, 2030.
Vehicle Class
LDGV
LDGT1
LDGT2
MC
HDGV
Reference Case
Tons
44,871
56,290
23,737
947
1,213
Control Case Tons
40,271
50,520
21,245
816
1067
Reduction
4,600
5,770
2,492
131
146
       Reductions from the proposed fuel benzene control vary significantly across the U.S.,
depending on the average level of benzene in gasoline sold, as discussed in Section 2.2.1.2 on air
quality modeling inventories. Table 2.2.-31  summarizes impacts of fuel benzene control on the
benzene emission inventory for gasoline vehicles in each State in 2030.
                                      2-57

-------
Table 2.2.-31. Impacts of Fuel Control on Gasoline Vehicle Benzene by State in 2030.
State
ALABAMA
ALASKA
ARIZONA
ARKANSAS
CALIFORNIA
COLORADO
CONNECTICUT
DELAWARE
DISTRICT OF COLUMBIA
FLORIDA
GEORGIA
HAWAII
IDAHO
ILLINOIS
INDIANA
IOWA
KANSAS
KENTUCKY
LOUISIANA
MAINE
MARYLAND
MASSACHUSETTS
MICHIGAN
MINNESOTA
MISSISSIPPI
MISSOURI
MONTANA
NEBRASKA
NEVADA
NEW HAMPSHIRE
NEW JERSEY
NEW MEXICO
NEW YORK
NORTH CAROLINA
NORTH DAKOTA
OHIO
OKLAHOMA
OREGON
PENNSYLVANIA
RHODE ISLAND
SOUTH CAROLINA
SOUTH DAKOTA
TENNESSEE
TEXAS
UTAH
VERMONT
VIRGINIA
WASHINGTON
WEST VIRGINIA
WISCONSIN
WYOMING
2030 Reference
Case Tons
2183
1270
1936
1275
9115
2870
1023
281
122
4220
4210
194
1224
4744
3895
1704
1833
2351
1543
765
1860
1874
6974
4129
1000
3439
1057
1195
1086
797
2068
1402
6601
3738
656
5263
1942
3190
5023
271
2034
619
2896
6544
1473
541
3061
4450
792
3657
671
2030 Control
Case Tons
1961
879
1783
1137
8489
2503
1009
277
120
3754
3821
193
1039
4359
3426
1471
1548
2083
1364
731
1809
1849
6030
3480
890
3018
904
1022
1034
769
2041
1169
6236
3363
553
4597
1740
2684
4685
268
1837
534
2612
5949
1276
500
2891
3709
700
3253
571
2030 Tons
Reduced
222
390
153
138
625
367
13
4
2
466
389
1
185
385
469
233
285
268
179
34
51
25
944
649
110
421
153
174
51
27
27
234
365
375
103
666
202
507
338
3
197
85
284
595
197
41
170
741
92
404
100
% Change
10
31
8
11
7
13
1
1
1
11
9
0
15
8
12
14
16
11
12
4
3
1
14
16
11
12
14
15
5
3
1
17
6
10
16
13
10
16
7
1
10
14
10
9
13
8
6
17
12
11
15
                                 2-58

-------
Gasoline Nonroad Equipment - Table 2.2.-24 summarizes the nationwide impact of the proposed
fuel benzene control on benzene emissions from gasoline nonroad equipment. As with highway
gasoline vehicles, emission benefits vary across the U. S. As can be seen in Table 2.2.-32, these
benefits vary from 1 to 32% by State in 2030.

       Portable Fuel Containers -Table 2.2.-33 summarizes MSAT emissions from PFCs with
no fuel benzene or Federal PFC control (but including State control programs). The proposed
fuel benzene control will reduce benzene emissions from portable fuel containers. Table 2.2.-34
summarizes the nationwide impact of fuel benzene control on PFC benzene emissions. Again,
emission benefits vary across the U. S., as seen in Table 2.2.-35.

       Gasoline Distribution - Table 2.2.-36 presents revised national reference case  inventory
estimates for gasoline distribution. Vehicle refueling emissions are included in the highway
vehicle inventory. Table 2.2.-37 presents the benzene inventory from gasoline distribution (not
including refueling) in 2015 and 2020 with and without the proposed fuel benzene control.
Table 2.2.-38 presents the inventory for 2020 at the State level with and without proposed fuel
benzene control.  More detailed inventory estimates by county are available in the docket for the
rule.
                                      2-59

-------
Table 2.2.-S2. Gasoline Nonroad Equipment Emission Reductions (Tons) from Proposed
                           Fuel Control by State, 2030.
State
ALABAMA
ALASKA
ARIZONA
ARKANSAS
CALIFORNIA
COLORADO
CONNECTICUT
DELAWARE
DISTRICT OF COLUMBIA
FLORIDA
GEORGIA
HAWAII
IDAHO
ILLINOIS
INDIANA
IOWA
KANSAS
KENTUCKY
LOUISIANA
MAINE
MARYLAND
MASSACHUSETTS
MICHIGAN
MINNESOTA
MISSISSIPPI
MISSOURI
MONTANA
NEBRASKA
NEVADA
NEW HAMPSHIRE
NEW JERSEY
NEW MEXICO
NEW YORK
NORTH CAROLINA
NORTH DAKOTA
OHIO
OKLAHOMA
OREGON
PENNSYLVANIA
RHODE ISLAND
SOUTH CAROLINA
SOUTH DAKOTA
TENNESSEE
TEXAS
UTAH
VERMONT
VIRGINIA
WASHINGTON
WEST VIRGINIA
WISCONSIN
WYOMING
2030 Reference
Case
707
160
631
443
3018
578
468
132
20
3085
1136
102
279
1371
763
452
364
514
782
288
739
756
1829
1055
450
856
153
248
242
237
1118
203
2050
1187
128
1542
486
589
1496
111
641
124
793
3378
350
113
839
869
249
934
104
2030 Control
Case
605
109
561
375
2705
493
459
130
19
2662
987
102
229
1266
653
375
298
441
650
269
707
742
1521
873
377
734
128
204
221
221
1097
164
1920
1023
105
1298
411
483
1368
109
549
103
682
2978
296
101
780
708
212
807
86
2030
Reductions
102
51
70
68
314
85
8
3
0
424
148
1
49
104
110
77
66
73
131
19
32
14
308
182
73
122
25
43
20
16
21
40
129
164
24
244
76
107
128
2
91
21
111
400
54
12
58
161
37
128
18
% Change
14
32
11
15
10
15
2
2
2
14
13
1
18
8
14
17
18
14
17
6
4
2
17
17
16
14
17
17
8
7
2
19
6
14
18
16
16
18
9
2
14
17
14
12
15
11
7
19
15
14
17
                                  2-60

-------
          Table 2.2.-3S. MSAT Emissions (Tons) from Uncontrolled PFCs.
Pollutant
2,2,4-Trimethylpentane
Benzene
Ethylbenzene
n-Hexane
MTBE
Naphthalene
Toluene
Xylenes
1999
5023
2229
2132
5570
7646
127
10,731
6,123
2007
4899
2254
2080
5432
7458
124
10,467
5,972
2010
4405
2118
1870
4884
6705
112
9,410
5,369
2015
4682
2262
1987
5191
7126
119
10,002
5,707
2020
5020
2423
2131
5566
7641
127
10,724
6,119
2030
5719
2757
2428
6341
8705
145
12,218
6,971
Table 2.2.-S4. Reduction in Benzene PFC Emissions (Tons) with Proposed Fuel Control
                        (No Control on PFC Emissions).
Year
1999
2015
2020
2030
Reference Case
2229
2262
2423
2757
Control Case
N. A.
1359
1456
1657
Reduction
N.A.
903
967
1100
                                  2-61

-------
Table 2.2.-3S. Reduction in Benzene PFC Emissions (Tons) with Proposed Fuel Control in
                   2030 by State (No Control on PFC Emissions).
State
AK
AL
AR
AZ
CA
CO
CT
DC
DE
FL
GA
HI
IA
ID
IL
IN
KS
KY
LA
MA
MD
ME
Ml
MN
MO
MS
MT
NC
ND
NE
NH
NJ
NM
NV
NY
OH
OK
OR
PA
Reference Case
Tons
42
65
47
45
92
83
11
1
3
224
84
8
42
28
98
68
46
50
68
18
20
5
219
67
80
35
18
115
10
23
12
26
31
13
47
167
51
92
45
Control Case
Tons
8
15
9
15
92
24
8
0
2
74
27
2
12
9
37
22
14
13
14
13
15
4
48
20
24
8
5
30
3
8
3
19
9
5
36
46
11
25
34
Reduction
34
50
38
29
0
59
3
0
1
150
57
6
30
19
61
45
33
36
54
5
5
2
171
47
56
27
13
86
7
15
9
7
22
8
11
121
40
66
11
% Change
81
77
80
65
1
71
27
31
24
67
67
75
72
68
62
67
70
73
80
26
23
31
78
70
70
77
73
74
72
66
74
26
72
61
23
73
78
72
25
                                   2-62

-------
State
Rl
SC
SD
UT
VA
VT
WA
Wl
WV
WY
Reference Case
Tons
2
49
9
34
30
3
135
78
33
10
Control Case
Tons
2
14
3
11
22
2
40
21
7
3
Reduction
0
35
6
23
7
1
95
57
26
7
% Change
15
71
70
68
25
25
70
73
79
71
 Table 2.2.-S6. Emissions of Mobile Source Air Toxics from Gasoline Distribution in tons
                         (2030 assumed to be same as 2020).
Pollutant
2,2,4-trimethylpentane
Benzene
Ethyl Benzene
n-Hexane
MTBE
Naphthalene
Toluene
Xylenes
1999
5,473
5,502
1,444
10,700
16,934
427
10,693
6,452
2007
5,646
5,695
1,547
10,925
17,346
446
11,121
6,859
2010
5,825
5,863
1,622
11,174
17,879
460
11,473
7,137
2015
5,981
5,999
1,710
11,309
18,113
471
11,771
7,449
2020
6,174
6,207
1,824
11,607
18,543
489
12,219
7,871
Table 2.2.-S7. Nationwide Impact of the Proposed Controls on Emissions of Benzene from
                       Gasoline Distribution in 2015 and 2020.



Tons of
Benzene
2015
Reference
Case

5,999
2015
Control
Case

4,054

2015
Reduction

1,945
2020
Reference
Case

6,207
2020
Control
Case

4,210

2020
Reduction

1,997
                                    2-63

-------
Table 2.2.-3S. Reduction in Gasoline Distribution Emissions of Benzene (Tons) with Proposed
                          Fuel Benzene Control by State, 2020.
State
ALABAMA
ALASKA
ARIZONA
ARKANSAS
CALIFORNIA
COLORADO
CONNECTICUT
DELAWARE
DISTRICT OF COLUMBIA
FLORIDA
GEORGIA
HAWAII
IDAHO
ILLINOIS
INDIANA
IOWA
KANSAS
KENTUCKY
LOUISIANA
MAINE
MARYLAND
MASSACHUSETTS
MICHIGAN
MINNESOTA
MISSISSIPPI
MISSOURI
MONTANA
NEBRASKA
NEVADA
NEW HAMPSHIRE
NEW JERSEY
NEW MEXICO
NEW YORK
NORTH CAROLINA
NORTH DAKOTA
OHIO
OKLAHOMA
OREGON
PENNSYLVANIA
RHODE ISLAND
SOUTH CAROLINA
SOUTH DAKOTA
TENNESSEE
TEXAS
UTAH
VERMONT
VIRGINIA
WASHINGTON
WEST VIRGINIA
WISCONSIN
WYOMING
Reference
Case
89
9
81
42
246
66
48
8
9
236
107
8
56
241
97
94
158
119
284
55
71
67
208
127
105
66
30
26
18
10
78
77
819
99
23
208
151
137
194
12
50
15
119
935
63
4
111
79
151
67
31
Control Case
51
5
48
24
242
42
44
8
8
143
65
5
35
172
56
51
86
70
162
42
60
61
112
68
60
42
19
14
11
8
72
44
707
60
12
112
82
82
124
11
30
8
64
666
40
2
80
47
91
42
20
Reduction
39
3
32
18
4
24
4
1
1
94
42
3
20
69
41
43
73
49
123
13
10
6
96
58
45
24
11
12
7
2
7
33
112
39
11
96
70
55
70
1
20
7
55
269
23
1
32
32
60
24
11
%
Change
43
40
40
43
2
37
8
8
8
40
40
40
37
29
42
46
46
41
43
24
15
8
46
46
43
36
37
46
40
17
8
43
14
40
46
46
46
40
36
8
40
46
46
29
37
40
28
40
40
37
37
                                         2-64

-------
2.2.2.2.2 Cold Temperature VOC Emission Control

       Reductions in MSATs are proportional to reduced NMHC start emissions from
vehicles subject to this rule. The magnitude of the reductions in these vehicles on a given
gasoline is based entirely on the number and duration of events between engine off and
engine on (vehicle soak) and the ambient conditions during them. The emissions reduced
are those created in the engine start following the vehicle soak. These parameters are
currently modeled by vehicle class and vehicle age in MOBILE6.2.57'58'59'60
MOBILE6.2 also provides the necessary information to adjust MSAT emission factors to
account for geographic and seasonal effects on in-use fuels.

       When all the affected vehicle classes meet the new emission standard we expect a
60% reduction of benzene and 1,3 butadiene from gasoline-fueled highway vehicles with
GVWR < 6000 Ibs and 30% from gasoline fueled highway vehicles with GVWR > 6000
Ibs. Effects on the trends in the inventories for the affected MSATs are shown in Table
2.2.-39 through Table 2.2.-44.
   Table 2.2.-40. Reference Case, Light Duty Gasoline Vehicles and Trucks, 1999
                                MSAT Inventory.
Pollutant
1,3 -Butadiene
2,2,4-Trimethylpentane
Acetaldehyde
Acrolein
Benzene
Ethyl Benzene
Formaldehyde
n-Hexane
MTBE
Styrene
Toluene
Xylenes
Total MSATS
Emissions in Tons
20,868
170,366
21,035
2,234
171,154
67,091
54,104
55,360
51,457
13,070
453,141
255,940
1,341,572
                                      2-65

-------
Table 2.2.-41. Reference and Vehicle Control Case, Light-Duty Gasoline Vehicles
                    and Trucks, 2010 MSAT Inventories.
Pollutant
1,3 -Butadiene
2,2,4-Trimethylpentane
Acetaldehyde
Acrolein
Benzene
Ethyl Benzene
Formaldehyde
n-Hexane
MTBE
Styrene
Toluene
Xylenes
Total MSATs
Reference Case Tons
in Calendar Year
2010
10,091
96,626
12,218
1,191
104,779
38,003
25,180
34,639
26,271
7,096
253,844
143,177
756,352
Vehicle Control Case
Tons in Calendar
Year 20 10
9,347
90,312
11,215
1,104
96,980
35,567
23,110
33,415
25,931
6,533
236,623
133,474
706,745
Reduction
in Tons
744
6,314
1003
87
7,799
2,436
2,070
1,223
340
563
17,221
9,703
49,607
Percent
Reduction
7
7
8
7
7
6
8
4
1
8
7
7
7
  Table 2.2.-42. Reference and Vehicle Control Case, Light-Duty Vehicles, 2015
                             MSAT Inventories.
Pollutant
1,3 -Butadiene
2,2,4-Trimethylpentane
Acetaldehyde
Acrolein
Benzene
Ethyl Benzene
Formaldehyde
n-Hexane
MTBE
Styrene
Toluene
Xylenes
Total MSATs
Reference Case Tons
in Calendar Year
2015
9,585
90.361
11,901
1,140
101,355
35,418
24,201
29,589
20,319
6,901
239,097
134,834
707,877
Vehicle Control Case
Tons in Calendar
Year 20 15
7,964
76,521
9,695
948
84,496
30,079
19,753
26,911
19,594
5667
201,351
113,568
599,492
Reduction
in Tons
1,621
13,840
2,206
192
16,859
5,339
4,448
2,679
725
1234
37,746
21,266
108,385
Percent
Reduction
17
15
19
17
17
15
18
9
4
18
16
16
15
                                   2-66

-------
   Table 2.2.-4S. Reference and Vehicle Control Case, Light-Duty Vehicles, 2020
                               MSAT Inventories.
Pollutant
1,3 -Butadiene
2,2,4-Trimethylpentane
Acetaldehyde
Acrolein
Benzene
Ethyl Benzene
Formaldehyde
n-Hexane
MTBE
Styrene
Toluene
Xylenes
Total MSATs
Reference Case Tons
in Calendar Year
2020
10,189
92,586
12,703
1,204
106,071
36,175
25,661
27,287
16,056
7,364
246,984
139,250
724,840
Vehicle Control Case
Tons in Calendar
Year 2020
7,470
69,374
9,006
882
77,966
27,213
18,323
22,801
14,909
5,292
183,618
103,549
543,332
Reduction
in Tons
2,719
23,212
3,697
322
28,105
8,962
7,338
4,486
1,147
2,072
63,366
35,701
181,508
Percent
Reduction
27
25
29
27
27
25
29
16
7
28
26
26
25
   Table 2.2.-44. Reference and Vehicle Control Case, Light-Duty Vehicles, 2030
                               MSAT Inventories.
Pollutant
1,3 -Butadiene
2,2,4-Trimethylpentane
Acetaldehyde
Acrolein
Benzene
Ethyl Benzene
Formaldehyde
n-Hexane
MTBE
Styrene
Toluene
Xylenes
Total MSATs
Reference Case Tons
in Calendar Year
2030
12,067
107,911
15,165
1,422
124,898
42,092
30,486
29,958
15,670
8,760
289,066
162,961
844,366
Vehicle Control Case
Tons in Calendar
Year 2030
7,379
68,389
8,938
875
77,208
26,807
18,218
22,322
13,793
5,228
180,996
102,072
535,479
Reduction
in Tons
4,688
39,522
6,227
547
47,690
15,285
12,268
7,636
1,877
3,532
108,070
60,889
308,887
Percent
Reduction
39
37
41
39
38
36
40
25
12
40
37
37
37
       State level reductions in calendar year 2030 benzene inventories are reported in
Table 2.2.-45. Reductions are higher in States with cold winter temperatures, such as
Alaska, where the reduction is 50%, and lowest in States with no winter or mild winters,
such as Hawaii and Florida, where reductions are 4% and 14%, respectively.
                                      2-67

-------
Table 2.2.-4S. 2030 Light-Duty Gasoline Vehicle Benzene Reference and Vehicle
                          Control Cases by State.
State
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
DC
Delaware
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
2030 Reference Case
Benzene Tons in
Calendar 2030
2128
1260
1886
1252
8984
2817
1010
120
275
4081
4117
188
1208
4674
3837
1682
1809
2315
1509
754
1829
1838
6885
4086
980
3385
1047
1180
1053
788
2030
1363
6520
3660
650
5177
1906
3131
4947
267
2030 Control Case
Benzene Tons in
Calendar 2030
1495
639
1241
854
5436
1645
535
69
155
3512
2807
181
707
2652
2315
1000
1110
1409
1124
413
1041
975
4043
2277
674
2083
586
702
664
454
1118
926
3721
2393
359
3029
1295
1864
2852
143
Reduction
in Tons
633
620
646
398
3548
1172
475
51
120
569
1309
7
501
2022
1522
682
698
907
385
340
788
863
2842
1809
307
1302
461
479
390
334
912
437
2799
1268
291
2148
611
1268
2095
124
Percent
Reduction
30%
49%
34%
32%
39%
42%
47%
42%
44%
14%
32%
4%
41%
43%
40%
41%
39%
39%
26%
45%
43%
47%
41%
44%
31%
38%
44%
41%
37%
42%
45%
32%
43%
35%
45%
41%
32%
40%
42%
46%
                                   2-68

-------
State
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
2030 Benzene Totals
2030 Reference
Case Benzene Tons
in Calendar 2030
1999
612
2843
6373
1442
536
3021
4383
784
3612
663
124898
2030 Control Case
Benzene Tons in
Calendar 2030
1328
348
1807
4918
822
305
1863
2431
454
2056
379
77208
Reduction
in Tons
671
263
1036
1455
621
231
1158
1952
330
1556
285
47670
Percent
Reductions
34%
43%
36%
23%
43%
43%
38%
45%
42%
43%
43%
38%
2.2.2.2.3 Portable Fuel Container Control

       The effect of PFC control on nationwide MS AT emissions are reported in Tables
2.2.-46 through 2.2.-49.  Table 2.2.-50 reports reductions in benzene with PFC control by
State in 2030.  Similar patterns are expected for other MSATs, although State level
inventories were not developed.

 Table 2.2.-46. Estimated Reductions in MSAT Emissions from PFC Control, 2010.
Pollutant
Benzene
Naphthalene
Ethyl Benzene
Toluene
n-Hexane
2,2,4-Trimethylpentane
Xylenes
MTBE
Total
Reference
Case
2118
112
1870
9410
4884
4405
5369
6705
34873
Control
Case
1885
100
1680
8454
4388
3957
4824
6024
31312
Reduction
in Tons
233
11
190
956
496
448
546
681
3561
Percent
Reduction
11
10
10
10
10
10
10
10
10
                                      2-69

-------
Table 2.2.-47. Estimated Reductions in MSAT Emissions from PFC Control, 2015.
Pollutant
Benzene
Naphthalene
Ethyl Benzene
Toluene
n-Hexane
2,2,4-Trimethylpentane
Xylenes
MTBE
Total
Reference
Case
2262
119
1987
10002
5191
4682
5707
7126
37075
Control
Case
794
47
779
3922
2035
1836
2238
2794
14445
Reduction
in Tons
1468
72
1208
6080
3155
2846
3469
4332
22630
Percent
Reduction
65
61
61
61
61
61
61
61
61
Table 2.2.-4S. Estimated Reductions in MSAT Emissions from PFC Control, 2020.
Pollutant
Benzene
Naphthalene
Ethyl Benzene
Toluene
n-Hexane
2,2,4-Trimethylpentane
Xylenes
MTBE
Total
Reference
Case
2423
127
2131
10724
5566
5020
6119
7641
39752
Control
Case
856
50
841
4234
2197
1982
2416
3017
15594
Reduction
in Tons
1567
77
1290
6490
3368
3038
3703
4624
24157
Percent
Reduction
65
61
61
61
61
61
61
61
61
Table 2.2.-49. Estimated Reductions in MSAT Emissions from PFC Control, 2030.
Pollutant
Benzene
Naphthalene
Ethyl Benzene
Toluene
n-Hexane
2,2,4-Trimethylpentane
Xylenes
MTBE
Total
Reference
Case
2757
145
2428
12218
6341
5719
6971
8705
45283
Control
Case
985
58
968
4872
2528
2280
2780
3471
17942
Reduction
in Tons
1772
87
1460
7346
3812
3438
4191
5234
27341
Percent
Reduction
64
60
60
60
60
60
60
60
60
                                   2-70

-------
Table 2.2.-50. Reductions in Benzene Emissions (Tons) with PFC Control by State,
                                   2030.
State
ALABAMA
ALASKA
ARIZONA
ARKANSAS
CALIFORNIA
COLORADO
CONNECTICUT
DELAWARE
DISTRICT OF COLUMBIA
FLORIDA
GEORGIA
HAWAII
IDAHO
ILLINOIS
INDIANA
IOWA
KANSAS
KENTUCKY
LOUISIANA
MAINE
MARYLAND
MASSACHUSETTS
MICHIGAN
MINNESOTA
MISSISSIPPI
MISSOURI
MONTANA
NEBRASKA
NEVADA
NEW HAMPSHIRE
NEW JERSEY
NEW MEXICO
NEW YORK
NORTH CAROLINA
NORTH DAKOTA
OHIO
OKLAHOMA
OREGON
PENNSYLVANIA
RHODE ISLAND
SOUTH CAROLINA
SOUTH DAKOTA
TENNESSEE
TEXAS
UTAH
VERMONT
VIRGINIA
WASHINGTON
WEST VIRGINIA
WISCONSIN
WYOMING
Reference
Case
65
42
45
47
92
83
11
3
1
224
84
8
28
98
67
42
46
50
68
5
20
18
219
67
35
80
18
23
13
12
26
31
47
115
10
167
51
92
45
2
49
9
69
104
34
3
30
135
33
78
10
Control Case
39
25
27
27
55
52
10
3
1
134
50
8
18
64
37
23
25
29
39
4
16
20
118
36
20
46
11
13
10
9
24
18
32
69
6
90
28
55
29
2
29
5
42
65
21
2
21
81
20
46
7
Reduction
26
17
18
20
37
31
1
0
0
90
34
0
10
34
30
19
21
21
29
1
4
-2
101
31
15
34
7
10
3
3
2
13
15
46
4
77
23
37
16
0
20
4
27
39
13
1
9
54
13
32
3
%
Change
40
40
40
43
40
37
9
0
0
40
40
0
36
35
45
45
46
42
43
20
20
-11
46
46
43
43
39
43
23
25
8
42
32
40
40
46
45
40
36
0
41
44
39
38
38
33
30
40
39
41
30
                                   2-71

-------
2.2.2.2.4 Cumulative Reductions of Proposed Controls

       Air toxic emissions from light-duty vehicles depend on both fuel benzene content
and vehicle hydrocarbon emission controls.  Similarly, the air toxic emissions from gas
cans depend on both fuel benzene content and the gas can emission controls. Tables 2.2.-
51 and 2.2.-52 summarize the expected reductions in benzene and MSAT emissions,
respectively, from the combined effects of our proposed vehicle, fuel, and gas can
controls.

       Table 2.2.-53  summarizes the cumulative benzene emission reductions from these
controls on highway gasoline vehicles, nonroad gasoline vehicles, gas cans, and gasoline
distribution at the State level in 2030.

       Table 2.2.-54  presents the impact of proposed controls on total benzene emissions
from mobile sources and portable fuel containers, and the impacts on total benzene
emissions from all sources.  Table 2.2.-55 presents the cumulative impact of proposed
controls on total emissions of mobile source air toxics from mobile source and portable
fuel containers, as well  as the impact on total emissions of mobile source air toxics from
both mobile and stationary sources. As discussed previously, the fuel benzene control
reduces stationary source emissions of benzene  associated with gasoline distribution.
                                       2-72

-------
Table 2.2-51.  Estimated Reductions in Benzene Emissions from All Proposed Control Measures by Sector, 2015 to 2030.
Benzene
Gasoline On-
road Mobile
Sources
Gasoline
Nonroad
Mobile Sources
Gas Cans
Gasoline
Distribution
Total
1999

178,465
58,710
2,229
5,502
244,905
2015
Without
Rule
(tons)
103,798
37,747
2,262
5,999
149,806
With Rule
(tons)
77,155
33,247
492
4,054
114,948
Reductions
(tons)
26,643
4,500
1,770
1,945
34,858
2020
Without
Rule
(tons)
108,256
36,440
2,423
6,207
153,326
With
Rule
(tons)
71,326
32,018
531
4,210
108,085
Reductions
(tons)
36,930
4,422
1,892
1,997
45,241
2030
Without
Rule
(tons)
127,058
39,162
2,757
6,207
175,184
With Rule
(tons)
70,682
34,400
610
4,210
109,902
Reductions
(tons)
56,376
4,762
2,147
1,997
65,282
                                                    2-73

-------
Table 2.2.-S2. Estimated Reductions in MSAT Emissions from All Proposed Control Measures by Sector, 2015 to 2030.
MSAT
Gasoline On-
road Mobile
Sources
Gasoline
Nonroad
Mobile
Sources
Gas Cans
Gasoline
Distribution
Total
1999

1,415,502
673,922
39,581
50,625
2,179,630
2015
Without
Rule
(tons)
731,283
432,953
37,076
62,804
1,264,116
With
Rule
(tons)
613,227
428,506
14,143
60,859
1,116,735
Reductions
(tons)
118,056
4,447
22,933
1,945
147,381
2020
Without
Rule
(tons)
745,769
390,468
39,751
64,933
1,240,921
With
Rule
(tons)
555,541
386,095
15,268
62,936
1,019,840
Reductions
(tons)
190,228
4,373
24,483
1,997
221,081
2030
Without
Rule
(tons)
865,767
405,119
45,284
64,933
1,381,103
With
Rule
(tons)
548,298
400,408
17,567
62,936
1,029,209
Reductions
(tons)
317,469
4,711
27,717
1,997
351,894
                                                    2-74

-------
Table 2.2.-5S. Cumulative Benzene Emission Reductions From All Proposed Controls at the State level in 2030.


ALABAMA
ALASKA
ARIZONA
ARKANSAS
CALIFORNIA
COLORADO
CONNECTICUT
DELAWARE
DISTRICT OF
COLUMBIA
FLORIDA
GEORGIA
HAWAII
IDAHO
ILLINOIS
INDIANA
IOWA
KANSAS
KENTUCKY
LOUISIANA
MAINE
MARYLAND
MASSACHUSETTS
MICHIGAN
MINNESOTA
MISSISSIPPI
Gasoline Highway
Vehicles
Tons
Reduced
801
821
756
500
3979
1402
482
122
51
989
1599
8
614
2261
1826
829
881
1081
527
359
820
877
3428
2187
388
%
37
65
39
39
44
39
47
44
42
23
38
4
50
48
47
49
48
46
34
47
44
47
49
53
39
Nonroad
Gasoline
Engines
Tons Reduced
102
51
70
68
314
85
8
3
0
424
148
1
49
104
110
77
66
73
131
19
32
14
308
182
73

%
14
32
11
15
10
15
2
2
2
14
13
1
18
8
14
17
18
14
17
6
4
2
17
17
16
Gas Cans
Tons
Reduced
56
37
35
42
37
68
4
1
0
180
68
6
23
74
55
36
39
42
60
3
7
6
193
56
30

%
86
89
79
89
40
81
33
30
37
80
80
75
80
75
82
85
84
85
88
51
37
31
88
84
87
Gasoline
Distribution
Tons Reduced
39
3
32
18
4
24
4
1
1
94
42
3
20
69
41
43
73
49
123
13
10
6
96
58
45

%
43
40
40
43
2
37
8
8
8
40
40
40
37
29
42
46
46
41
43
24
15
8
46
46
43
Total
Tons
Reduced
998
912
893
628
4334
1579
498
127
52
1687
1857
18
706
2508
2032
985
1059
1245
841
394
869
903
4025
2483
536

%
34
64
34
36
35
37
35
30
43
21
34
6
47
40
43
45
47
42
33
36
32
35
43
47
36
                                                        2-75

-------


MISSOURI
MONTANA
NEBRASKA
NEVADA
NEW HAMPSHIRE
NEW JERSEY
NEW MEXICO
NEW YORK
NORTH
CAROLINA
NORTH DAKOTA
OHIO
OKLAHOMA
OREGON
PENNSYLVANIA
RHODE ISLAND
SOUTH
CAROLINA
SOUTH DAKOTA
TENNESSEE
TEXAS
UTAH
VERMONT
VIRGINIA
WASHINGTON
WEST VIRGINIA
WISCONSIN
WYOMING
Gasoline Highway
Vehicles
Tons
Reduced
1582
549
587
428
350
928
605
3018
1536
350
2567
763
1587
2302
126
812
314
1233
1963
741
286
1271
2386
386
1802
344
%
46
52
49
39
44
45
43
46
41
53
49
39
50
46
46
40
51
43
30
50
47
42
54
49
49
51
Nonroad
Gasoline
Engines
Tons Reduced
122
25
43
20
16
21
40
129
164
24
244
76
107
128
2
91
21
111
400
54
12
58
161
37
128
18

%
14
17
17
8
7
2
19
6
14
18
16
16
18
9
2
14
17
14
12
15
11
7
19
15
14
17
Gas Cans
Tons
Reduced
66
15
97
9
10
8
26
22
97
9
142
45
77
24
1
40
7
59
55
27
1
14
111
29
66
9

%
83
83
84
85
80
32
84
47
84
85
85
88
83
52
22
83
84
85
53
80
55
47
82
88
85
82
Gasoline
Distribution
Tons Reduced
24
11
12
7
2
7
33
112
39
11
96
70
55
70
1
20
7
55
269
23
1
32
32
60
24
11

%
36
37
46
40
17
8
43
14
40
46
46
46
40
36
8
40
46
46
29
37
40
28
40
40
37
37
Total
Tons
Reduced
1794
600
739
464
378
964
704
3281
1836
394
3049
954
1826
2524
130
963
349
1458
2687
845
300
1375
2690
512
2020
382

%
41
50
50
34
37
31
43
36
36
49
43
38
47
39
35
36
47
39
25
45
42
35
50
47
44
49
2-76

-------
Table 2.2.-S4. Impact of proposed controls on total benzene emissions from mobile sources, and the impacts on total benzene
                                            emissions from all sources.

2015
Fuel Benzene Control
Vehicle Control
Fuel, Vehicle and RFC Control
2020
Fuel Benzene Control
Vehicle Control
Fuel, Vehicle and RFC Control
2030
Fuel Benzene Control
Vehicle Control
Fuel, Vehicle and RFC Control
Mobile Source and
RFC Tons Reduced

16687
16858
32912

16790
28104
43245

20997
47688
65281
Mobile Source and
RFC Tons

149602
149602
149602

152618
152618
152618

174753
174753
174753
% of Mobile
Source and
RFC Tons
Reduced

11
11
22

11
18
28

12
27
37
Total Tons
Reduced

18632
16858
34857

18787
28104
45242

20997
47688
65281
Total Mobile
and Stationary
Tons

269787
269787
269787

276295
276295
276295

298430
298430
298430
% of Mobile
and
Stationary
Tons
Reduced

7
6
13

7
10
16

7
16
22
                                                      2-77

-------
   Table 2.2.-5S.  Cumulative impact of proposed controls on total emissions of mobile source air toxics from mobile source and
portable fuel containers, as well as the impact on total emissions of mobile source air toxics from both mobile and stationary sources.



2015
2020
2030
Mobile
Source and
RFC Tons
Reduced
145436
240032
373658


Mobile Source
and RFC Tons
1260205
1229591
1369867
% of Mobile
and RFC
Tons
Reduced
12
20
27


Total Tons
Reduced
147381
242029
375655
Total Mobile
and
Stationary
Tons
4164490
4362301
4502577

% of Mobile and
Stationary Tons
Reduced
4
6
8
                                                          2-78

-------
2.3  Potential Implications of New Emissions Data for Inventories

2.3.1  Newer Technology Light Duty Vehicles

       MOBILE6.2 explicitly estimates emissions for the following air toxic compounds:
benzene, 1,3-butadiene, formaldehyde, acetaldehyde, MTBE, and acrolein.61'62
MOBILE6.2 estimates air toxics emission factors by multiplying an air toxic to VOC
(volatile organic compound) ratio by MOBILE6.2 VOC. For light-duty gasoline vehicles
and trucks, the product for exhaust emissions is then multiplied by an off-cycle
adjustment factor, which accounts for the difference in toxic fractions between Federal
Test Procedure (FTP) and Unified Cycle (UC) operation.

       Toxic to VOC ratios vary by technology group, vehicle type, whether a vehicle is
a normal or high emitter (same definition as MOBILE6.2), and fuel characteristics.
Evaporative toxic/VOC ratios do not vary among gasoline vehicle classes. Since toxic
emission rates  are a product of toxic/VOC emission ratios and VOC emission rates,
anything that reduces VOC will also result in toxic emission reductions. Toxic/VOC
ratios for individual technology group/vehicle type/emitter class combinations are
determined using a series of algorithms which calculate the ratios based on fuel
parameter inputs. These algorithms were derived from tests on 1990 model year
technology vehicles and form the basis of the Complex Model for Reformulated
Gasoline. MOBILE6.2 assumes that the same ratios are applicable to all post-1990
technology vehicles, including advanced technology low emission vehicles (LEVs)
meeting Tier 2 standards.63

       Eastern Research Group, under contract to EPA, recently compared exhaust
emissions data from newer technology vehicles to see if the toxic to VOC fractions
estimated from these data were statistically different from ratios predicted by
MOBILE6.2.  To make these comparisons,  we used data collected by EPA Office of
Research and Development/National Exposure Research Laboratory on 23 1998-2003
vehicles, the California Air Resources Board (46 vehicles) and Southwest Research
Institute (3 vehicles). The contractor report and the data used are available in the docket
for this rule.64  The data from EPA's Office  of Research and Development have been
published.65

       The conclusions from t-test comparisons were as follows:

    1) When the off-cycle adjustment for benzene is factored out of the model results,
       MOBILE6.2 predicts  statistically higher toxic fractions than one gets from the
       California Air Resources Board and  Southwest Research Institute data, although
       for the large California dataset, the difference is only 10%. The fractions from the
       EPA Office of Research and Development data are higher than predicted by
       MOBILE6.2, but the difference is not statistically significant.

   2)  MOBILE6.2 is over-predicting toxic fractions for 1,3-butadiene.
                                      2-79

-------
   3)  The available data do not support a conclusion that MOBILE6.2 underestimates
       or overestimates fractions for MTBE, formaldehyde, acetaldehyde or acrolein.

       There is a significant amount of scatter in the available test data, which makes it
difficult to draw strong conclusions from the statistical comparisons. Also data are very
limited for high emitters and off-cycle operation, which make a large contribution to total
emissions. Nonetheless, at this point it appears that MOBILE6.2 toxic to VOC fractions
for benzene, MTBE, formaldehyde, acetaldehyde, and acrolein are reasonably accurate
for newer technology vehicles, but that fractions used for 1,3-butadiene are
overestimating emissions for this pollutant.

       The recent Energy Policy Act passed by Congress requires EPA to develop a new
fuel effects model that reflects a 2007 fleet.  The collection of a large amount of data and
substantial analytical work is needed to meet this requirement, and to update the
algorithms used in the  current Complex Model and MOBILE6.2. Initial work is
underway in a collaborative test program between EPA and members of the Alliance of
Automobile Manufacturers to examine emissions of both regulated pollutants and air
toxics from Tier 2 compliant vehicles.  The current program focuses on changes in fuel
sulfur, vapor pressure,  and benzene levels, and will provide data for the air toxics
rulemaking process as  well as inform the design of a more comprehensive program
covering a wider range of fuel properties and vehicle certification levels.

2.3.2  Heavy-Duty Vehicles (CRC E-55/E-59)

       The primary  objective of the E-55/59 research program was to quantify gaseous
and PM emissions from primarily in-use heavy-duty diesel trucks in California's South
Coast Air Basin, in support of emissions inventory development.66 A second program
objective was to quantify the influence of tampering and mal-maintenance on emissions
from these vehicles.  The program was conducted in four Phases (denoted as 1, 1.5, 2 and
3). The Phase 1 test fleet consisted of 25 heavy heavy-duty diesel trucks (HHDDT),
selected to match a distribution of model years (MY) and to reflect engines in common
use in California. In  Phase 1.5 an additional twelve HHDDT were studied, with a
thirteenth truck tested at idle alone. The Phase 2 test fleet consisted often HHDDT and
nine medium heavy-duty trucks (MHDT), which included seven diesel-fueled medium
heavy-duty trucks (MHDDT) and two  gasoline-fueled medium heavy-duty trucks
(MHDGT). Phase 3 gathered data from nine HHDDT, eight MHDDT, and two MHDGT.
The Phase 2 and 3 data added post-2002 MY HHDDT (at 2.5 g/bhp-hr NOX standard) to
the program.

       Sampling for chemical speciation was performed on thirteen HHDDT in Phase 1
and on five HHDDT and one MHDDT in Phase 2.  However, only three of the thirteen
Phase 1 trucks  had their exhaust samples analyzed  for air toxic emissions, and the
remaining samples were being archived. Toxics species were measured from five
HHDDT and one MHDDT (medium HDDTs) in the Phase 2 test fleet.  PM data were
acquired in Phases 1.5, 2 and 3. Exhaust data were acquired for methane  and VOC.
Semi-volatile organic compounds and  PM soluble fractions were captured and analyzed,
                                      2-80

-------
along with carbonyls and nitrosamines. Ions and elemental/organic carbon (EC/OC) split
were determined from quartz filters. The ion and metal analyses varied widely between
trucks.
       These data will be incorporated into EPA's MS AT inventories, and will help
address limitations discussed in Sections 2.1.4 and 2.2.1.1.5.
2.3.3  Small Spark Ignition Engines

       The National Mobile Inventory Model (NMIM) calculates air toxic emissions for
small Spark Ignition (SI) engines by multiplying compound-specific fractions with
volatile organic carbon (VOC) or particulate matter (PM) emission outputs from EPA's
NONROAD model.  These fractions were used in the 1999 National Air Toxics
Assessment (NAT A).  These data were all obtained from a small number of uncontrolled
engines.67'68'69'70'71   In fiscal year 2004 EPA tested a mixture of in-use and new pre-
control and Phase 1 small hand held SI trimmers, chain saws and a leaf blower72.  In the
same time period EPA performed engine tests on Phase I residential four-stroke lawn
mowers.  The emission data from both programs may impact future versions of NMIM
and the inventories it calculates.

       EPA tested four pre-control, nine Phase 1, two California-certified, and eight
Phase 2 handheld engines. Five  of the Phase 2 engines were new. All tests were fueled
by either of two summer grades of gasoline. One was a gasoline ethanol blend meant to
represent a reformulated gasoline and the other a conventional gasoline. All but one of
the engines were two-cycle designs. However, the four-cycle engine was designed to
operate on a typical two-cycle fuel lubricating oil mixture.  All the test engines require
that lubricating oil be mixed and consumed with the fuel. The program therefore used
two different types of lubricating oil, one a mineral-based product and the other a "low
smoke" synthetic. Both oils were commercially available. The testing was done over the
Composite Two Mode (C2M) duty cycle. Table 2.3.-1 compares the emission factors
used in NONROAD and the fractions used in NMIM with those based on the testing.

       NONROAD  and NMEVI have not been adjusted to use the new data, but some
increase in projected benzene inventories is likely once this occurs. In all  but one engine
and fuel combination the benzene/VOC fraction is greater than that currently used in
NMIM. It is significant that two-cycle engines have a large proportion of their fuel being
emitted in an unburned state. A reduction in fuel benzene content will have a significant
effect on benzene emissions from them.

       The other MSAT fractions are found in Table 2.3.-2. Some of the  measured
values are more consistent with NMIM values, but some are not (e.g., xylenes).

       The second EPA test program involved six new Phase 2 four cycle lawn mower
engines.  These data are unpublished.  The engines were tested after 20 hours of
operation. The testing was done using the certification test procedure on certification
gasoline. Formaldehyde and acetaldehyde were the only MSATs measured in the test
program. A comparison  of NMIM fractions and measured fractions are in Table 2.3.-.3.
                                      2-81

-------
      The measured values are similar to the values used in NMIM. Incorporation of
the new test data would not result in a dramatic change in inventories from these engines
and use types.
    Table 2.3.-1. Comparison between NONROAD Outputs and NMIM MSAT
 Fractions and Averaged Test Data for PM, VOC and Benzene from EPA Testing of
   18 Handheld SI Engines Aggregated by Use, Engine Class, Emission Standard
                      (Phase), Catalyst, and Engine Cycle



Type



Class



Condition



Phase


Catalyst
Equipped


Engine
Cycle

NONROAD
PM10EF
(g/bhp)
Average
Tested
PM2.5
(g/bhp)

NONROAD
HCEF
(g/bhp)
Average
Tested
THC
(g/bhp)

NMIM
Benzene
Fraction
Average
Tested
Benzene
Fraction
BLOWER
CHAIN
SAW
CHAIN
SAW
CHAIN
SAW
CHAIN
SAW
CHAIN
SAW
STRING
TRIMMER
STRING
TRIMMER
STRING
TRIMMER
STRING
TRIMMER
STRING
TRIMMER
STRING
TRIMMER
V
IV
IV
IV
IV
V
III
III
IV
IV
IV
IV
New
New
Used
Used
Used
Used
Used
Used
New
New
Used
Used
2
2
0
1
2
1
0
1
2
2
0
1
YES
YES
NO
NO
NO
NO
NO
NO
YES
NO
NO
NO
2
2
2
2
2
2
2
2
2
4
2
2
7.70
7.70
9.24
9.93
9.93
9.75
9.24
9.55
7.70
0.06
9.24
9.93
0.028
0.228
3.072
2.051
1.483
1.330
4.915
7.519
0.641
0.231
3.093
3.856
40.15
26.87
313.20
231.84
42.66
152.00
313.20
272.79
26.87
25.83
313.20
231.84
24.842
30.254
185.976
110.567
98.066
80.026
265.205
243.167
31.581
12.791
221.354
154.140
0.024
0.080
0.080
0.080
0.080
0.080
0.011
0.011
0.011
0.011
0.011
0.011
0.038
0.022
0.016
0.014
0.014
0.016
0.019
0.013
0.028
N.A.
0.015
0.017
                                    2-82

-------
       Table 2.S.-2. NMIM MSAT Fractions versus Fractions from EPA Testing of
18 Handheld SI Engines
Type Standard Fuel
BLOWER
SAW
SAW
SAW
SAW
SAW
TRIMMER
TRIMMER
TRIMMER
TRIMMER
TRIMMER
TRIMMER
TRIMMER
TRIMMER
Ph2
Phi
Phi
Ph2
Ph2


Phi
Phi
Phi
Phi
Ph2
Ph2
CG
CG
CG
RFG
CG
RFG
CG
RFG
CG
CG
RFG
RFG
CG
RFG
Formaldehyde
NMTM
0.0068
0.0068
0.0068
0.0068
0.0068
0.0029
0.0029

0.0029

0.0029
0.0029
0.0029
Tested
0.0050
0.0042
0.0053
0.0052
0.0056
0.0072
0.0077

0.0039

0.0045
0.0050
0.0080
Acetaldehyde
NMTM
0.0013
0.0013
0.0013
0.0013
0.0013
0.0006
0.0006

0.0006

0.0006
0.0006
0.0006
Tested
0.0011
0.0009
0.0046
0.0011
0.0055
0.0016
0.0066

0.0009

0.0046
0.0010
0.0073
Acrolein
NMTM
0.0004
0.0004
0.0004
0.0004
0.0004
0.0003
0.0003

0.0003

0.0003
0.0003
0.0003
Tested
0.0003
0.0003
0.0004
0.0004
0.0004
0.0006
0.0006

0.0003

0.0003
0.0003
0.0005
Propionaldehyde
NMTM
0.0001
0.0003
0.0004
0.0004
0.0004
0.0004
0.0004

0.0003

0.0004
0.0006
0.0009
Tested
Toluene
NMTM
0.0978
0.0002 0.0598
O.OOOq 0.0598
0.0002| 0.0598
0.0002
0.0002
0.0002
0.0002

0.0002

0.0002
0.0002
0.0002
0.0598
0.0598
0.0978
0.0890
0.0978
0.0978
0.0890
0.0890
0.0978
0.0890
Tested
0.0979
0.0998
0.1064
0.1105
0.1065
0.0955
0.1049
0.0891
0.1093
0.1000
0.1096
0.0906
0.1303
0.1014
2,2,4-
Trimethypentane
NMTM
0.0372
0.0372
0.0372
0.0372
0.0372
0.0372
0.0372
0.0372
0.0372
0.0372
0.0372
0.0372
0.0372
0.0372
Tested
0.0122
0.0490
0.0487
0.0280
0.0409
0.0252
0.0437
0.0242
0.0432
0.0497
0.0249
0.0279
0.0559
0.0326
Xylene
NMIM
0.1075
0.0931
0.0931
0.0931
0.0931
0.0931
0.1075
0.0978
0.1075
0.1075
0.0978
0.0978
0.1075
0.0978
Tested
0.0224
0.0166
0.0151
0.0231
0.0177
0.0228
0.0174
0.0232
0.0204
0.0163
0.0299
0.0238
0.0205
0.0235
         Table 2.3.-S. Comparison of NMIM Acetaldehyde and Formaldehyde to
        VOC fractions with Measured Fractions from OTAQ Test Program
MSAT
Acetaldehyde
Formaldehyde
NMIM Fraction
0.00440
0.01256
Average Measured Fraction
0.00396
0.01541
2.3.4  Nonroad CI engines

       The Agency conducted three separate emission test programs measuring exhaust
emissions from fifteen nonroad diesel engines and in-use pieces of nonroad diesel
equipment73'74'75. The engines tested derived from construction, utility and agricultural
equipment applications for the most part and ranged from seven horsepower (hp) up
through 850 hp (425 hp, as tested). The test fuels used varied by sulfur concentration
from nonroad-grade diesel fuels at 2500 and 3300 ppm sulfur to a nominal "D-2" diesel
at 350 ppm sulfur and, lastly, to an ultra-low sulfur diesel, measured at less than 10 ppm
sulfur. Test engines were run over both steady-state and transient duty cycles.  Several of
the transient cycles were application-specific, having been based on rubber-tire loader or
excavator operations, for example. Criteria pollutants in the exhaust emissions were
quantified for each test engine as well as sulfate, ammonia, N2O and a range of Ci  - Cn
compounds (aldehydes, ketones, alcohols, etc.).  Emissions of several additional air toxic
compounds were identified in two of the three programs.  These emission species
included benzene, toluene, ethylbenzene, xylenes, polyaromatic hydrocarbons (PAH),
nitrated-PAHs and several metals. Emission results were  summarized in both grams/hour
and grams/brake-horsepower/hour.

        With the emission data, EPA will address differences between Tier 1  and
                                      2-83

-------
unregulated NR diesel emissions, the impact of diesel fuel sulfur level on engine
emissions, whether any adjustments to default modeling TAFs (transient adjustment
factors) used in the NONROAD emissions model are warranted by the new data, and the
necessity of creating category- and power-specific TAFs for NONROAD. The data will
also be used to update NMIM inventories for toxic air compounds.

2.4    Description of Current  Mobile Source Emissions Control
Programs that Reduce MSATs

       As described above, existing  mobile source control programs will reduce MSAT
emissions (not including diesel PM)  by 60% between 1999 and 2020. Diesel PM from
mobile sources will be reduced by 70% between 1999 and 2020. The mobile source
programs include controls on fuels, highway vehicles, and nonroad equipment.  These
programs are also reducing hydrocarbons and PM more generally, as well as oxides of
nitrogen. The sections immediately below provide general descriptions  of these
programs, as well as voluntary programs to reduce mobile source emissions, such as the
National Clean Diesel Campaign and Best Workplaces for Commuters.

2.4.1   Fuels Programs

       Several federal fuel programs reduce MSAT emissions.  Some of these programs
directly control air toxics, such as the reformulated gasoline (RFG) program's benzene
content limit and required reduction in total toxics emissions,  and the anti-backsliding
requirements of the anti-dumping and current MSAT programs, which require that
gasoline cannot  get dirtier with respect to toxics emissions. Others, such as the gasoline
sulfur program,  control toxics indirectly by reducing hydrocarbon and related toxics
emissions.

2.4.1.1 RFG

       The RFG program contains two direct toxics  control requirements. The first is a
fuel benzene standard, requiring RFG to average no greater than 0.95 volume percent
benzene annually (on a refinery or importer basis). The RFG  benzene requirement
includes a per-gallon cap on fuel benzene level of 1.3 volume percent. In 1990, when the
Clean Air Act was amended to require reformulated gasoline,  fuel benzene averaged 1.60
volume percent. For a variety of reasons, including other regulations, chemical product
prices and refining efficiencies, most refiners and importers have achieved significantly
greater reductions in benzene than required by the program. In 2003, RFG benzene
content averaged 0.62 percent. The RFG benzene requirement includes  a per-gallon cap
on fuel benzene  level of 1.3 volume percent.

       The second RFG toxics control requires that RFG achieve a specific level of
toxics emissions reduction. The requirement has increased in stringency since the RFG
program began in 1995, when the requirement was that RFG annually achieve a 16.5%
reduction in total (exhaust plus evaporative)  air toxics emissions. Currently, a 21.5%
reduction is required. These reductions are determined using  the Complex Model.  As
                                     2-84

-------
mentioned above, for a variety of reasons most regulated parties have overcomplied with
the required toxics emissions reductions. During the 1998-2000 timeframe, RFG
achieved, on average, a 27.5% reduction in toxics emissions.

2.4.1.2 Anti-dumping

       The anti-dumping regulations were intended to prevent the dumping of "dirty"
gasoline components, which were removed to produce RFG, into conventional gasoline
(CG).  Since the dumping of "dirty" gasoline components, for example, benzene or
benzene-containing blending streams, would show up as increases in toxics emissions,
the anti-dumping regulations require that a refiner's or importer's CG be no more
polluting with respect to toxics emissions than the refiner's or importer's 1990 gasoline.
The anti-dumping program considers only exhaust toxics emissions and does not include
evaporative emissions.8  Refiners and importers have either a unique individual anti-
dumping baseline or they have the statutory anti-dumping baseline if they did not fulfill
the minimum requirements for developing a unique individual baseline.  In 1990, average
exhaust toxics emissions (as estimated by the Complex Model)  were 104.5 mg/milec; in
2004, CG exhaust toxics emissions averaged 90.7 mg/mile.  Although CG has no benzene
limit, benzene levels have declined significantly from the 1990  level of 1.6 volume
percent to 1.1 volume percent for CG in 2004.

2.4.1.3 2001 Mobile Source Air Toxics Rule (MSAT1)

       As discussed above, both RFG and CG have, on average, exceeded their
respective toxics control requirements. In 2001, EPA issued a mobile source air toxics
rule (MS ATI, for the purposes of this second proposal), as discussed in section ID.  The
intent of MS ATI is to prevent refiners and importers from backsliding from the toxics
performance that was being achieved by RFG and CG. In order to lock in superior levels
of control, the rule requires that the annual average toxics performance of gasoline must
be at least as clean as the average performance of the gasoline produced or imported
during the three-year period 1998-2000. The period 1998-2000 is called the baseline
period. Toxics performance is determined separately for RFG and CG, in the same
manner as the toxics determinations required by the RFG76 and  anti-dumping  rules.

       Like the anti-dumping provisions, MSAT1 utilizes an individual baseline against
which compliance is determined. The average 1998-2000 toxics performance level, or
baseline, is determined separately for each refinery and importer.0 To establish a unique
individual MS ATI baseline, EPA requires each refiner and importer to submit
documentation supporting the determination of the baseline. Most refiners and many
importers in business during the baseline period had sufficient data to establish an
     BSee RFG rule for why evaporative emissions are not included in the anti-dumping toxics
determination.
     cPhase II
     DExcept for those who comply with the anti-dumping requirements for conventional gasoline on an
aggregate basis, in which case the MS ATI requirements for conventional gasoline must be met on the same
aggregate basis (40 CFR Part 80, Subpart E).
                                      2-85

-------
individual baseline.  An MS ATI baseline volume is associated with each unique
individual baseline value. The MS ATI baseline volume reflects the average annual
volume of such gasoline produced or imported during the baseline period. Refiners and
importers who did not have sufficient refinery production or imports during 1998-2000 to
establish a unique individual MSAT1 baseline must use the default baseline provided in
the rule.

       The MS ATI program began with the annual averaging period beginning January
1, 2002.  Since then, the toxics performance for RFG has improved from a baseline
period average of 27.5% reduction to 29.5% reduction in 2003.  Likewise, CG toxics
emissions have decreased from an average of 95 mg/mile during 1998-2000 to 90.7
mg/milein2003.

2.4.1.4 Gasoline Sulfur

       EPA's gasoline sulfur program77 requires, beginning in 2006, that sulfur levels in
gasoline can be no higher in any one batch than 80 ppm, and must average 30 ppm
annually. When fully effective, gasoline will have 90 percent less sulfur than before the
program. Reduced sulfur levels are necessary to ensure that vehicle emission control
systems are not impaired. These systems effectively reduce non-methane organic gas
(NMOG) emissions, of which some are air toxics.  With lower sulfur levels, emission
control technologies can work longer and more efficiently.  Both new and older vehicles
benefit from reduced gasoline sulfur levels.

2.4.1.5 Gasoline Volatility

       A fuel's volatility defines its evaporation characteristics. A gasoline's volatility is
commonly referred to as its Reid vapor pressure, or RVP.  Gasoline summertime RVP
ranges from about 6-9 psi, and wintertime RVP ranges from about 9-14 psi, when
additional vapor is required for starting in cold temperatures.  Gasoline vapors contain a
subset of the liquid gasoline components, and thus can contain toxics compounds such as
benzene. EPA has controlled summertime gasoline RVP since 1989 primarily as a VOC
and ozone precursor control, which also results in some toxics pollutant reductions.

2.4.1.6 Diesel Fuel

       In early 2001, EPA issued rules requiring that diesel fuel for use in highway
vehicles contain no more than 15 ppm sulfur beginning June 1, 2006. 78 This program
contains averaging, banking and trading provisions, as well as other compliance
flexibilities. In June 2004, EPA issued rules governing the sulfur content of diesel fuel
used in nonroad diesel engines.79  In the nonroad rule, sulfur levels are limited to a
maximum of 500 ppm sulfur beginning in 2007 (current levels are approximately 3000
ppm).  In 2010, nonroad diesel sulfur levels must not exceed 15 ppm.

       EPA's diesel fuel requirements are part of a comprehensive program to combine
engine and fuel controls to achieve the greatest emission reductions.  The diesel fuel
                                      2-86

-------
provisions enable the use of advanced emission-control technologies on diesel vehicles
and engines. The diesel fuel requirements will also provide immediate public health
benefits by reducing PM emissions from current diesel vehicles and engines.

2.4.1.7 Phase-Out of Lead in Gasoline

       One of the first programs to control toxic emissions from motor vehicles was the
removal of lead from gasoline. Beginning in the mid-1970s, unleaded gasoline was
phased in to replace leaded gasoline. The phase-out of leaded gasoline was completed
January 1, 1996, when lead was banned from motor vehicle gasoline. The removal of
lead from gasoline has essentially eliminated on-highway mobile source emissions of this
highly toxic substance.

2.4.2   Highway Vehicle and Engine Programs

       The 1990 Clean Air Act Amendments set specific emission standards for
hydrocarbons and for PM.  Air toxics are present in both of these pollutant categories. As
vehicle manufacturers develop technologies to comply with the hydrocarbon (HC) and
particulate standards (e.g., more efficient catalytic converters), air toxics are reduced as
well.  Since 1990, we have developed a number of programs to address exhaust and
evaporative hydrocarbon emissions and PM emissions.  Table 2.4-1 shows current mobile
source programs for highway vehicles.

       Two of our recent initiatives to control  emissions from motor vehicles and their
fuels are the Tier 2 control program for light-duty vehicles and the 2007 heavy-duty
engine rule. Together these two initiatives define a set of comprehensive standards for
light-duty and heavy-duty motor vehicles and their fuels. In both of these initiatives, we
treat vehicles and fuels  as a system. The Tier 2 control program establishes stringent
tailpipe and evaporative emission standards for light-duty vehicles and a reduction in
sulfur levels in gasoline fuel beginning in 2004.80  The 2007 heavy-duty engine rule
establishes stringent exhaust emission standards for new heavy-duty engines and vehicles
for the 2007 model year as well as reductions in diesel fuel sulfur levels starting in
2006.81  Both of these programs will provide substantial emissions reductions through the
application of advanced technologies. We expect 90% reductions in PM from new diesel
engines compared to engines under current standards.

       Some of the key earlier programs controlling highway vehicle and engine
emissions are the Tier 1 and NLEV standards for light-duty vehicles and trucks;
enhanced evaporative emissions standards; the supplemental federal test procedures
(SFTP); urban bus  standards; and heavy-duty diesel and gasoline standards for the
2004/2005 time frame.
                                      2-87

-------
   Table 2.4-1.  Current On-Highway Engine and Vehicle Programs Providing
                   Significant Additional MSAT Reductions.
Category
Light-duty cars and
trucks



Heavy-duty trucks
Urban Buses
Highway
motorcycles
Rule & FRM Date
Tier 2 (including low sulfur fuel),
February 10, 2000
NLEV (National Low-Emitting
Vehicle)
Enhanced Evaporative Emissions
SFTP (Supplemental FTP)
Procedures
2004 Heavy-duty Rule
October 6, 2000
2007 Heavy-duty Rule (including
low sulfur fuel), January 18, 2001
HD Diesel Retrofit
December 2003
Implementation
Schedule
2004 - 2009
1999-2003

2001 (start)
2004 - 2007
2007-2010
1994 - 1998
2006-2010
voc
Standards*
/
/
/
/
/

/
PM
Standards
/
/


/
/

* Standards in various forms including HC, NMHC, NMOG, and NOx+NMHC

-------
             Table 2.4-2 Current Nonroad Engine/Vehicle Programs.
Category
Land-based diesel
Locomotives
Vlarine



^arge spark-
ignition engines
Small spark-
ignition engines

Aircraft
(NOx Std in 2005;
Smoke Std in 1982)
lecreational
vehicles
Rule & FRM Date
Tier 2, October 23, 1998
Tier 3, October 23, 1998
Tier 4 (w/ low sulfur fuel)
June 29, 2004
Tier 0, Tier 1, Tier 2
April 16, 1998
Spark-ignition Gasoline
Engine Standards,
October 4, 1996
Diesel Engines, less than 50hp
Recreational Diesel,
November 8, 2002
Commercial Diesel,
February 28, 2003
Tier 1 Standards
Tier 2 Standards
November 8, 2002
Phase 1 Standards,
Handheld Phase 2 Standards,
April 25, 2000
Non-handheld Phase 2
Standards, March 30, 1999

November 8,2002
Implementation
Schedule
2001-2006
2006-2008
2008-2014
2002 - 2005
1998-2006
1999-2005
Starting 2006/2009
Starting 2004/2007
2004 - 2007
2007 - 20XX
1997-2007
2002 - 2007
2001 -2007
No current/recent
standards for VOC or
PM
2006-2012
VOC
Standards*
/
/
/
/
/

/
/
/
/
/


/
PM
Standards
/
/
/
/

/
/
/






* Standards in various forms including HC, NMHC, NMOG, and NOx+NMHC

2.4.3   Nonroad Engine Programs

       There are various categories of nonroad engines, including  land-based diesel
engines (e.g., farm and construction equipment), small land-based spark-ignition (SI)
engines (e.g., lawn and garden equipment, string trimmers), large land-based SI engines
(e.g., forklifts, airport ground service equipment), marine engines (including diesel and
SI, propulsion and auxiliary, commercial and recreational), locomotives, aircraft, and
recreational vehicles (off-road motorcycles, "all terrain" vehicles and snowmobiles).
                                      2-89

-------
Table 2.4-2 shows current mobile source programs for nonroad engines. Brief summaries
of our current and anticipated programs for these nonroad categories follow. As with
highway vehicles, the VOC standards we have established for nonroad engines will also
significantly reduce VOC-based toxics from nonroad engines. In addition, the standards
for diesel engines (in combination with the stringent sulfur controls on nonroad diesel
fuel) will significantly reduce diesel PM and exhaust organic gases, which are mobile
source air toxics.

       In addition to the engine-based emission control programs described below, fuel
controls will also reduce emissions of air toxics from nonroad engines. For example,
restrictions on gasoline formulation (the removal of lead, limits on gasoline volatility and
RFG) are projected to reduce nonroad MSAT emissions because most gasoline-fueled
nonroad vehicles are  fueled with the same gasoline used in on-highway vehicles. An
exception to this is lead in aviation gasoline. Aviation gasoline, used in general (as
opposed to commercial) aviation, is a high octane fuel used in a relatively small number
of aircraft (those with piston engines).  Such aircraft are generally used for personal
transportation, sightseeing, crop dusting, and similar activities.

2.4.3.1 Land-based Diesel Engines

       We recently finalized stringent new emissions standards for land-based nonroad
diesel engines, used in agricultural and construction equipment as well as many other
applications (although the standards do not apply to locomotive, mining equipment and
marine engines).82 These standards are similar in stringency to the 2007 highway diesel
engine standards, and are likewise enabled by stringent controls on sulfur levels is diesel
fuel, as explained earlier in section 2.4.1.6. The new engine standards, starting in 2008,
will reduce PM from  new 2008 nonroad diesel engines by about 95 percent compared to
engines under today's standards. The fuels controls are scheduled to begin in mid-2007.

2.4.3.2 Small Land-Based SI Engines

       Small land-based spark-ignition (Small SI) engines at or below 25 hp are used
primarily in lawn and garden equipment such as lawn mowers, string trimmers, chain
saws, lawn and garden tractors, and other similar equipment.  Our Phase 1 emission
controls for this category of engines took effect beginning in  1997 and are projected to
result in a roughly 32 percent reduction in VOC emissions for new engines, on average,
versus pre-controlled engines.83  We also have Phase 2 regulations for these engines
which, when fully phased-in, are projected to result in additional combined HC and NOx
reductions beyond the Phase 1 levels of 60 percent for new non-handheld engines and of
70 percent for new handheld engines.84 We are currently developing a proposal for Phase
3 standards that would further reduce HC emissions from  Small SI engines.

2.4.3.3 Large Land-Based SI engines

       Since the MSAT1 rule was published, we have also finalized emissions standards
                                                      oc
for SI engines above  25 hp used in commercial applications.    Such engines are used in a
                                       2-90

-------
variety of industrial equipment such as forklifts, airport ground service equipment,
generators and compressors.  The Tier 1 standards went into effect in 2004 and the Tier 2
standards will start in 2007, providing additional emissions reductions. These standards
will provide about a 90 percent reduction in HC emissions on average for new engines
versus Tier 1 controlled engines.

2.4.3.4 Recreational Vehicles

       Standards for recreational vehicles, including snowmobiles, off-road motorcycles
and "all terrain" vehicles, will begin in 2006.  These standards will require significant
reductions in HC emissions from new engines, ranging from 50 to 86 percent compared
to pre-controlled engines.86

2.4.3.5 Marine engines

       Marine engines cover a very wide range of products, from 10-horsepower
outboard engines to 100,000-horsepower engines on oceangoing vessels.  We have active
emission-control programs to address the need for emission controls for every kind of
marine engine. For gasoline-fueled engines, we adopted an initial tier of standards with a
phase-in schedule that is complete in the 2006 model year. These standards, which apply
to outboard and personal-watercraft engines, have led to a major shift to four-stroke
engines and advanced-technology two-stroke  engines for an estimated 75 percent
reduction in hydrocarbon emissions from uncontrolled levels.87 We are developing a
proposal to adopt new, more stringent standards for these engines that would reduce
emissions from these engines by an additional 60 percent or more from the previous tier.

       Another kind of gasoline-fueled marine engine, referred to as stern drive and
inboard engines, uses an automotive-type engine. These engines have uncontrolled
emission rates that are well below the current standards that apply to outboard and
personal-watercraft engines.  These engines are not currently subject to emission
standards, but we intend to include new emission standards for these engines in an
upcoming gasoline marine engine proposal.88 These new standards would likely be based
on the application of catalyst technology to substantially reduce hydrocarbon and NOx
emissions.

       We  have adopted emission standards for marine diesel  engines in four separate
rulemakings. All of these standards are based on in-engine controls and do not require
aftertreatment. First, we adopted two tiers of standards for marine engines below 50
horsepower that apply equally to land-based and marine engines.  These standards were
phased in from 1999 to 2005.  Second, we adopted emission standards for commercial
marine diesel engines with per-cylinder engine displacement up to 30 liters.  These
standards are comparable to the standards for land-based nonroad diesel engines that
apply in the same time frame,  with several adjustments to test procedures and compliance
provisions appropriate for marine engines.89  The emission standards generally apply in
2007 for locomotive-size engines and in 2004 for smaller engines. Third, the emission
standards adopted for recreational marine diesel engines are very similar to the
                                       2-91

-------
comparable commercial engines, with implementation scheduled two years after the
commercial standards take effect.  All the emission standards in these three rulemakings
targeted reductions in NOx and PM emissions. Finally, we adopted standards to control
NOx emissions at levels consistent with the requirements from the International Maritime
Organization (IMO), but we adopted these as EPA standards under the Clean Air Act to
make them mandatory for all engines with per-cylinder displacement above 2.5 liters
installed on U.S.-flag vessels starting in the 2004 model year.  We are in the process of
reviewing the emission standards for all sizes of marine diesel engines  and expect to
propose new requirements in the near future.

       EPA is also investigating the possibility of designating U.S. coastal areas as SOx
Emission Control Areas (SECAs) under the IMO. Such a designation would trigger a
requirement for any vessel entering such an area to use reduced-sulfur fuel or operate
exhaust scrubbers to prevent SOx emissions.

2.4.3.6 Locomotives

       Our regulations for locomotive engines consist of three tiers of standards,
applicable depending on the date a locomotive or a particular engine was originally
manufactured.90 The first set of standards (Tier 0) applies to locomotives and their
locomotive engines originally manufactured from 1973 through 2001, starting from the
time the engine was manufactured or later at "remanufacture."E  The second set of
standards (Tier 1) applies to locomotives and their engines manufactured from 2002
through 2004 and again at engine manufacture or rebuild. The third  set of standards (Tier
2) applies to locomotive engines manufactured in 2005 and later. The Tier 0 and Tier 1
regulations were primarily intended to reduce NOx emissions. The Tier 2 regulations are
projected to result in 50 percent reductions in VOC and diesel PM as compared to
unregulated engine emission levels, as well as additional NOx reductions beyond the Tier
0 and Tier 1 regulations. We are currently developing a new tier of more stringent
emissions standards for locomotive engines.

2.4.3.7 Aircraft

       A variety of emission regulations have been applied to commercial gas turbine
aircraft engines, beginning with limits on smoke and fuel venting in  1974. In 1984, limits
were placed on the amount of unburned HC that gas turbine engines  can emit per landing
and takeoff cycle. In 1997, we adopted standards that were equivalent  to the existing
International Civil Aviation Organization (ICAO) NOx and CO emission standards for
gas turbine engines.  In 2005, we tightened the NOx emission  standards to levels that are
equivalent to the ICAO standards that became effective in 2004. These actions have
resulted in minimal  emissions reductions, and have largely served to prevent increases in
aircraft emissions. We continue to explore ways to reduce emissions from aircraft
throughout the nation.
      ' "Remanufacture" is an engine rebuild "to new" during four-to-eight year long maintenance cycles.


                                       2-92

-------
2.4.4   Voluntary Programs

       In addition to the fuel and engine control programs described above, we are
actively promoting several voluntary programs to reduce emissions from mobile sources,
such as the National Clean Diesel Campaign, anti-idling measures, and Best Workplaces
for Commuters. While the stringent emissions standards described above apply to new
highway and nonroad diesel engines, it is also important to reduce emissions from the
existing fleet of about 11 million diesel engines. EPA has launched a comprehensive
initiative called the National Clean Diesel Campaign, one component of which is to
promote the reduction of emissions in the existing fleet of engines through a variety of
cost-effective and innovative strategies. The goal  of the Campaign is to reduce emissions
from the 11 million existing engines by 2014. Emission reduction strategies include
switching to cleaner fuels, retrofitting engines through the addition of emission control
devices, and engine replacement. For example, installing a diesel particulate filter
achieves diesel particulate matter reductions of approximately 90 percent (when
combined with the use of ultra low sulfur diesel fuel). The Energy Policy Act of 2005
includes grant authorizations and other incentives to help facilitate voluntary clean diesel
actions nationwide.

       The National Clean Diesel Campaign is focused on leveraging local, state, and
federal resources to retrofit or replace diesel engines, adopt best practices, and track and
report results.  The Campaign targets five key sectors:  school buses, ports, construction,
freight, and agriculture.

       Reducing vehicle idling provides important environmental benefits. As a part of
their daily routine, truck drivers often keep their vehicles at idle during stops to provide
power, heat and air conditioning. EPA's SmartWay Transport Partnership is helping the
freight industry to adopt innovative idle reduction technologies and take advantage of
proven systems that provide drivers with basic necessities without using the engine. To
date, there are 50 stationary anti-idling projects, and mobile technology has been installed
on nearly 20,000 trucks.  The SmartWay Transport Partnership also works with the
freight industry to reduce fuel use (with a concomitant reduction in emissions)  by
promoting a wide range of new technologies such as advanced aerodynamics, single-wide
tires, weight reduction speed control and intermodal shipping.

       Daily commuting represents another significant source of emissions from motor
vehicles. EPA's Best Workplaces for CommutersSM program is working with employers
across the country to reverse the trend of longer, single-occupancy vehicle commuting.
OTAQ has created a national list of the Best Workplaces for Commuters to formally
recognize employers that offer superior commuter benefits such as free transit passes,
subsidized vanpools/carpools, and flexi-place, or work-from-home, programs. More than
1,300 employers representing 2.8 million U.S. workers have been designated Best
Workplaces for Commuters.

       Much of the growth in the Best Workplaces for Commuters program has been
through metro area-wide campaigns. Since 2002, EPA has worked with coalitions in 14
                                       2-93

-------
major metropolitan areas to increase the penetration of commuter benefits in the
marketplace and the visibility of the companies that have received the BWC designation.
Another significant path by which the program has grown is through Commuter Districts
including corporate and industrial business parks, shopping malls, business improvement
districts and downtown commercial areas.  To date EPA has granted the Best
Workplaces for Commuters "District" designation to twenty locations across the country
including downtown Denver, Houston, Minneapolis and Tampa.
                                      2-94

-------
References for Chapter 2
1 http ://www. epa. gov/otaq/nonrdtndl .htm#model

2 http://www.epa.gov/otaq/models/mobile6/m6tech.htm

3 Mullen, M., Neumann, J. Technical Memorandum: Documentation of 2003 VMT
Projection Methodology, Prepared by E. H. Pechan and Associates and Industrial
Economics, Inc. for James DeMocker, Office of Air and Radiation, Office of Policy
Analysis and Review, U. S. EPA, Contract No. 68-W-02-048, WA B-41, March 2004.
Included as Appendix G: Clean Air Interstate Rule: Emission Inventory Technical
Support Document, http://www.epa.gov/cair/technical.html.  This document is available
in Docket EPA-HQ-OAR-2003-0053.

4 "VOC/PM Cold Temperature Characterization and Interior Climate Control
Emission/Fuel Economy Impact", Alan P. Stanard , Southwest Research Institute Project
No. 03.11382.01, October 2005

5 Brzezinski, D.; Fernandez; A. (2005) Cold Temperature Effects on Vehicle HC
Emissions.  U. S. EPA, Assessment and Standards Division, National Vehicle and Fuel
Emissions Laboratory, Ann Arbor, MI, Report No. EPA420-D-06-001.

6 http://www.epa.gov/otaq/models/mobile6/m6tech.htm

7 Landman, L. C. (2006) Estimating Emissions Associated with Portable Fuel Containers
(PFCs). U. S. EPA, Assessment and Standards Division, National Vehicle and Fuel
Emissions Laboratory, Ann Arbor, MI, Report No. EPA420-D-06-003. This document is
available in Docket EPA-HQ-OAR-2003-0053.

8 Lawson, D.R.; Smith, R.E. (1998) The Northern Front Range Air Quality Study. A
report to the Governor and General Assembly.  Colorado State University. [Online at
http://www.nfraqs.colostate.edu]

9 Stanard, A.P. (2005) VOC/PM cold temperature characterization and interior climate
control emissions/fuel economy impact. Draft final report volume I.  EPA contract 68-C-
06-018. Work assignment No. 0-4. Southwest Research Institute project no.
03.11382.04.

10 Bailey, C.R. (2005) Cold-temperature exhaust particulate matter emissions.
Memorandum to Docket EPA-HQ-OAR-2005-0036.

11 Pollack, A. K.; Lindhjem, C.; Stoeckenius, T. E.; Tran, C.; Mansell, G.; Jiminez, M.;
Wilson, G.; Coulter-Burke, S. 2004. CRC Project E-64: Evaluation  of the U. S. EPA
MOBILE6 Highway Vehicle Emission Factor Model. Prepared by Environ Corp. for
                                     2-95

-------
Coordinating Research Council and U. S. EPA. March, 2004.
http://www.epa.gov/otaq/m6. htm#m60

12 National Research Council. 2000. Modeling Mobile Source Emissions. National
Academy Press; Washington, D.C. http://www.nap.edu/books/0309070880/html/

13 Ibid.

14 Strum, M., Cook, R., Pope, A., Palma, T., Shedd, S., Mason, R., Michaels, H.,
Thurman, J., Ensley, D. 2005.  Projection of Hazardous Air Pollutant Emissions to
Future Years. Science of the Total Environment, in press.  This document is available in
Docket EPA-HQ-OAR-2005-0036.

15 U. S. EPA. 2004.  1999 National Emissions Inventory, Final Version 3.
http://www.epa.gov/ttn/chief/net/1999inventory.html

16 U. S. EPA. MOBILE6 Vehicle Emissions Modeling Software.
http://www.epa.gov/otaq/m6.htm.

17 Cook, R.; Glover, E. "Technical Description of the Toxics Module for MOBILE6.2
and Guidance on Its Use for Emission Inventory Preparation", U.S. EPA, Office of
Transportation and Air Quality, Ann Arbor, MI, November 2002; Report No., EPA420-
R-02-011.  http://www.epa.gov/otaq/m6.htm This document is available in Docket EPA-
HQ-OAR-2005-0036.

18 Michaels, H., Brzezinski, D., Cook, R..  EPA's National Mobile Inventory Model
(NMIM), A Consolidated Emissions Modeling System for MOBILE6 and NONROAD.
U. S. Environmental Protection Agency, Office of Transportation and Air Quality,
Assessment and Standards Division, Ann Arbor, MI, March 2005; Report No.  EPA-420-
R-05-003.  http://www.epa.gov/otaq/nmim.htm. This document is available in Docket
EPA-HQ-OAR-2005-0036.

19 Cook, R., Glover, E., Michaels, H., Brzezinski, D. 2004.  Modeling of Mobile Source
Air Toxic Emissions Using EPA's National Mobile Inventory Model.  Proceedings, 2004
Emission Inventory Conference, Clearwater Beach, FL.
http://www.epa.gov/ttn/chief/conference/eil3/index.html. This document is available in
Docket EPA-HQ-OAR-2005-0036.

20 MathPro.  1998.  Costs of Alternative Sulfur Content Standards for Gasoline in PADD
IV. Final Report. Prepared for the National Petrochemical and Refiners Association.
December 30.  This document is available in Docket EPA-HQ-OAR-2005-0036.

21 Mathpro, 1999. Costs of meeting 40 ppm Sulfur Content  Standard for Gasoline in
PADDs 1-3, Via MOBIL and CD  Tech Desulfurization Processes. Final Report.

                                      2-96

-------
Prepared for the American Petroleum Institute. February 26.  This document is available
in Docket EPA-HQ-OAR-2005-0036.

22MathPro. 1999. Analysis of California Phase 3 RFG Standards.  Prepared for the
California Energy Commission. December 7. This document is available in Docket
EPA-HQ-OAR-2005-0036.

23 Eastern Research Group. National Mobile Inventory Model (NMIM) Base and Future
Year County Database Documentation and Quality Assurance Procedures.  Prepared by
Eastern Research Group, Inc., for U.S. Environmental Protection Agency, Office of
Transportation and Air Quality, Ann Arbor, MI, 2003.  This document is  available in
Docket EPA-HQ-OAR-2005-0036.

24 Mullen, M., Neumann, J. Technical Memorandum: Documentation of 2003 VMT
Projection Methodology, Prepared by E. H. Pechan and Associates and Industrial
Economics, Inc. for James DeMocker, Office of Air and Radiation, Office of Policy
Analysis and Review, U. S. EPA, Contract No. 68-W-02-048, WA B-41,  March 2004.
Included as Appendix G: Clean Air Interstate Rule: Emission Inventory Technical
Support Document, http://www.epa.gov/cair/technical.html.  This document is available
in Docket EPA-HQ-OAR-2003-0053.

25 Wyborny, Lester; Memorandum to the Docket; Effect of Benzene Control on Gasoline
Quality, February 22, 2006.

26 U. S. Environmental Protection Agency. Regulatory Impact Analysis:  Clean Air
Nonroad Diesel Rule. Office of Transportation and Air Quality, Ann Arbor, MI, 2004d,
Report No.  EPA420-R-04-007. http://www.epa.gov/nonroad-diesel/2004fr.htmfeia

27 Fritz, S. 2000. Diesel Fuel Effects on Locomotive Exhaust Emissions. Prepared by
Southwest Research for the California Air Resources Board.  This document is available
in Docket EPA-HQ-OAR-2003-0053.

28 Eastern Research Group and E. H. Pechan and Associates.  2005.  Documentation for
Aircraft, Commercial Marine Vessels, Locomotive and Other Nonroad Components of
the National Emissions Inventory. Prepared for U. S. EPA, Office of Air Quality
Planning and Standards, http://www.epa.gov/ttn/chief/net/1999inventory.html. This
document is available in Docket EPA-HQ-OAR-2005-0036.

29 Federal Aviation Administration, 2004. Terminal Area Forecast System.
http://www.apo.data.faa.gov/faatafall.HTM

30 U. S. EPA.  2000.  National  Air Pollutant Emission Trends, 1990--1998. Office of
Air Quality Planning and Standards, Research Triangle Park, NC. Report No. EPA-
454/R-00-002. http://www.epa.gov/ttnchiel/trends/trends98/trends98.pdf.
                                     2-97

-------
31 Driver, L., Pope, A., Billings, R., Wilson, D. The 1996 National Toxics Inventory and
its role in evaluating the EPA's progress in reducing hazardous air pollutants in ambient
air. Proceedings, 92nd Annual Meeting of the Air & Waste Management Association, St.
Louis, MO, June 1999, paper 91-501.

32 ERG, Inc. 2003. Documentation for the Final 1999 Point Source National Emissions
Inventory for Hazardous Air Pollutants (Version 3).  Prepared for U. S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, July 2003.
http://www.epa.gov/ttn/chief/net/1999inventory.html.  This document is available in
Docket EPA-HQ-OAR-2005-0036.

33 ERG, Inc. 2003.  Documentation for the Final  1999 Nonpoint Source National
Emissions Inventory for Hazardous Air Pollutants (Version 3). Prepared for U. S.
Environmental Protection Agency, Office of Air Quality Planning and Standards, August
2003. http://www.epa.gov/ttn/chief/net/1999inventory.html.  This document is available
in Docket EPA-HQ-OAR-2005-0036.

34 U. S. Environmental Protection Agency. User's Guide for the Emissions Modeling
System for Hazardous Air Pollutants (EMS-HAP, Version 3.0), Office of Air Quality
Planning and Standards, Research Triangle Park, NC, 2004b,  Report No. EPA-454/B-00-
007.  http://www.epa.gov/scram001/tt22.htmtfaspen.  This document is available in
Docket EPA-HQ-OAR-2005-0036.

35 Regional Economic Models, Inc.  2004. REMI Policy Insight,  http://www.remi.com.

36 Fan W, Treyz F, Treyz G.  An evolutionary new economic geography model. J
Regional Sci 2000 40: 671- 696.

37 Energy Information Adminstration. Annual Energy Outlook 2005 with Projections to
2025. 2005, Report No. DOE/EIA-0383.  http://www.eia.doe.gov/oiaf/aeo/index.html.

38 U. S. Environmental Protection Agency. Clean Air Interstate Rule: Emissions
Inventory Technical Support Document, 2005b.  Available from:
http://www.epa.gov/cair/technical.html. This document is available in Docket EPA-HQ-
OAR-2003-0053.

39 U. S. Environmental Protection Agency. SPECIATE, Version 3.2. Available at:
http://www.epa.gov/ttn/chief/software/speciate/index.html

40 Cook, R., Beidler, A., Touma, J., Strum, M.  2005. Preparing Highway Emissions
Inventories for Urban Scale Modeling: A Case Study in Philadelphia.  Submitted to
Transportation Research Part D: Transport and Environment. This document is
available in Docket EPA-HQ-OAR-2005-0036.
                                      2-98

-------
41 Taylor, M. Memorandum: Revised HAP Emission Factors for Stationary Combustion
Turbines, Prepared by Alpha-Gamma Technologies, Inc for Sims Roy, EPA OAQPS
ESD Combustion Group. August, 2003.Docket ID: OAR-2003-0189.
http://docket.epa.gov/edkpub/do/EDKStaffItemDetailView?objectId=090007d480271237

42 See http://www.epa.gov/ttn/chief/net/index.html for status and availability of NEI2002
inventory and supporting documents.

43 Stump, F., Tejada, S., Ray, W., Dropkin, D.,  Black, F.,  Snow, R., Crews, W., Siudak,
P., Davis, C., Baker L., Perry, N. 1989. The Influence of Ambient Temperature on
Tailpipe Emissions from 1984 to 1987 Model Year Light-Duty  Gasoline Vehicles.
Atmospheric Environment 23: 307-320.

44 Stump, F., Tejeda, S., Ray, W., Dropkin, D.,  Black, F.,  Snow, R., Crews, W., Siudak,
P., Davis C., Carter, P. 1990. The Influence of Ambient Temperature on Tailpipe
Emissions from 1985-1987 Model Year Light-Duty Gasoline Vehicles — II.  Atmospheric
Environment 24A: 2105-2112.

45 Stump, F., Tejeda, S., Dropkin, D. Loomis, C. Unpublished.  Characterization of
Emissions from Malfunctioning Vehicles Fueled with Oxygenated Gasoline-MTBE Fuel
- Part I. U. S. EPA, National Exposure Research Laboratory, Office of Research and
Development.  This document is available  in Docket EPA-HQ-OAR-2005-0036.

46 Stump, F., Tejeda, S., Dropkin, D. Loomis, C., Park, C. Unpublished. Characterization
of Emissions from Malfunctioning Vehicles Fueled with Oxygenated Gasoline-Ethanol
(E10) Fuel - Part II.  U. S. EPA, National Exposure Research Laboratory, Office of
Research and Development.  Unpublished. This document is available in Docket EPA-
HQ-OAR-2005-0036.

47 Stump, F., Tejeda, S., Dropkin, D. Loomis, C., Park, C. Unpublished.  Characterization
of Emissions from Malfunctioning Vehicles Fueled with Oxygenated Gasoline-Ethanol
(E10) Fuel - Part III. U. S. EPA, National Exposure Research Laboratory, Office of
Research and Development. This document is available in Docket EPA-HQ-OAR-2005-
0036.

48 Stanard, Alan P. 2005. VOC/PM Cold Temperature Characterization and Interior
Climate Control Emissions/Fuel Economy Impact. Prepared for U. S. EPA, Office of
Transportation  and Air Quality, Ann Arbor, MI, by Southwest Research Institute.
Southwest Research Project No. 03.11382.04. This document is available in Docket
EPA-HQ-OAR-2005-0036.

49 http://www.epa.gov/otaq/nonrdmdl.htm#model

50 http://www.epa.gov/otaq/nmim.htm

                                     2-99

-------
51 Federal Highway Adminstration. 2003. Highway Statistics.
http://www.fhwa.dot.gov/policy/ohim/hs03/

52 Energy Information Adminstrati on. 2003. Petroleum Marketing Annual.
http://www.eia.doe.gov/pub/oil_gas/petroleum/data_publications/
petrol eum_marketing_annual/historical/2003/pma_2003.html

53 Haskew, H. M.; Liberty, T. F.; McClement, D. 2004. Fuel Permeation from
Automotive Systems. Prepared for the Coordinating Research Council by Harold
Haskew and Associates and Automotive Testing Laboratories, Inc. September 2004.
CRC Project No. E-65. http://www.crcao.com.  This document is available in Docket
EPA-HQ-OAR-2005-0036.

54 U. S. EPA. 2005.  Documentation for the Final 2002 Mobile National Emissions
Inventory. Prepared by U. S. EPA, Assessment and Standards Division and E. H. Pechan
and Associates for U. S. EPA, Office of Air Quality Planning and Standards.  September
2005.

55 ERG, Inc. 2003.  Documentation for the for the Final 1999 Nonpoint Area Source
National Emission Inventory for Hazardous Air Pollutants (Version 3). Prepared for U.
S. EPA, Office of Air Quality Planning and Standards, August 26, 2003.
http://www.epa.gov/ttn/chief/net/1999inventory.html.  This document is available in
Docket EPA-HQ-OAR-2005-0036.

56 Driver, L.  2005.  Memorandum to State/Local Agencies that submitted 2002 gasoline
distribution emissions requesting review/comments on EPA plans for final 2002 NEI.
ftp://ftp. epa.gov/EmisInventory /draftnei2002/nonpoint/gasoline_marketing/gasdistributio
ninstructions.pdf. This document is available in Docket EPA-HQ-OAR-2005-0036.

57 Glover, E.; Brzezinski, D.  2001.  Soak Length Activity Factors for Start Emissions. U.
S. EPA, Office of Transportation and Air Quality, Ann Arbor, MI.  Report No. EPA420-
R-01-011. http://www.epa.gov/otaq/models/mobile6/r01011 .pdf This document is
available in Docket EPA-HQ-OAR-2005-0036.

58 Enns, P.; Brzezinski, D.  2001. Comparison of Start Emissions in theLA92 and ST01
Test Cycles. U. S. EPA, Office of Transportation and Air Quality, Ann Arbor, MI.
Report No. EPA420-R-01 -025.  http://www.epa.gov/otaq/models/mobile6/rO 1025.pdf
This document is available in Docket EPA-HQ-OAR-2005-0036.

59 Glover, E; Carey, P.  2001. Determination of Start Emissions as  a Function  of
Mileage and Soak Time for 1981-1993 Model Year Light-Duty Vehicles. U. S.  EPA,
Office  of Transportation and Air Quality, Ann Arbor, MI.  Report No. EPA420-R-01-
058. http://www.epa.gov/otaq/models/mobile6/r01058.pdf  This document is available
in Docket EPA-HQ-OAR-2005-0036.

                                     2-100

-------
60 Glover, E.; Brzezinski, D. 2001. Exhaust Emission Temperature
Correction Factors for MOBILE6: Adjustments for Engine Start and Running LA4
Emissions for Gasoline Vehicles. U. S. EPA, Office of Transportation and Air Quality,
Ann Arbor, MI. Report No. EPA420-R-01-029.
http://www.epa.gov/otaq/models/mobile6/r01029.pdf. This document is available in
Docket EPA-HQ-OAR-2005-0036.

61 Cook, R.; Glover, E.  "Technical Description of the Toxics Module for MOBILE6.2
and Guidance on Its Use for Emission Inventory Preparation", U.S. EPA, Office of
Transportation and Air Quality, Ann Arbor, MI, November 2002; Report No., EPA420-
R-02-011. http://www.epa.gov/otaq/m6.htm.  This document is available in Docket
EPA-HQ-OAR-2005-0036.

62 MOBILE6 Vehicle Emissions Modeling Software, http://www.epa.gov/otaq/m6.htm.

63EPA. 1993. Final Regulatory Impact Analysis for Reformulated Gasoline. December
13, 1993. http://www.epa.gov/otaq/regs/fuels/rfg/.  This document is available in Docket
EPA-HQ-OAR-2005-0036.

64 Weatherby, M. F., Fincher, S., DeFries, T. H., Kishan, S. 2005.  Comparison of Toxic
to Hydrocarbon Ratios from In-Use New Technology Vehicles to Those Used in
MOBILE6.2. Prepared by Eastern Research Group, Austin, TX, for Rich Cook, Office of
Transportation and Air Quality, U. S. EPA. ERG Report No. ERG No.:
0136.04.001.001. This document is available in Docket EPA-HQ-OAR-2005-0036.

65 Baldauf, R. W., P. Gabele, W. Crews, R. Snow, and R. Cook.  2005.  Criteria and Air
Toxic Emissions from In-Use Automobiles in the National Low-Emission Vehicle
Program. Journal of the Air and Waste Management Association, 55:1263 -1268.

66 Clark, N., Gautam, M., Wayne, W., Lyons, D., & Thompson, G., "California Heavy
Heavy-Duty Diesel Truck Emissions Characterization for Program E-55/59: (Draft) Final
Report" (Short Title: "E-55/59 All Phases") Nov. 11, 2005 West Virginia University
Research Corporation, Morgantown, WV. EPA document # 420-D-06-005.

67 Censullo, A. 1991.  Development of Species Profiles for Selected Organic Emission
Sources.  California Polytechnic State University. San Luis Obispo, CA. This document
is available in Docket EPA-HQ-OAR-2005-0036.

68 Gabele, Peter. 1997. Exhaust emissions from four-stroke lawn mower engines.  J.  Air
& Waste Manage. Assoc. 47:945-950.

69 Hare, C. T.; Carroll, J. N. 1993. Speciation of Organic Emissions to Study Fuel
Dependence of Small Engine Exhaust Photochemical Reactivity. Report for Advisory
Committee  on Research, Southwest Research Institute, July 1993. This document is
available in Docket EPA-HQ-OAR-2005-0036.
                                    2-101

-------
70 Hare, C.T.; White, JJ.  1991. Toward the Environmentally Friendly Small Engine:
Fuel, Lubricant, and Emission Measurement Issues.  SAE Paper No. 9 11222. This
document is available in Docket EPA-HQ-OAR-2005-0036.

71 Carroll, J. N. 1991. Emission Tests of In-use Small Utility Engines:  Task III Report,
Nonroad Source Emission Factors Improvement. Prepared for U. S. EPA by Southwest
Research Institute. Report No. SwRI 3426-0006.  This document is available in Docket
EPA-HQ-OAR-2005-0036.

72 "Characterization of Emissions from Small Hand-Held, In-Use 2- Cycle Engines",
Richard Snow and William Crews, BKI, Inc , EPA/600/X-4/191, October 2004. This
document is available in Docket EPA-HQ-OAR-2005-0036.

73 Starr, M. February, 2004 "Air Toxic Emissions from In-Use Nonroad Diesel
Equipment; Final Report" (SwRI-08.05004.04). Prepared for U. S. EPA, Office of
Transportation and Air Quality, Ann Arbor, MI by Southwest Research Institute. (EPA
document # 420-R-04-019). This document is available in Docket EPA-HQ-OAR-2005-
0036.

74 Starr, M. February, 2004 "Nonroad Duty Cycle Testing for Toxic Emissions; Final
Report" (SwRI-08.05004.05).  Prepared for U. S. EPA, Office of Transportation and Air
Quality, Ann Arbor, MI by Southwest Research Institute. (EPA document # 420-R-04-
018).  This document is available in Docket EPA-HQ-OAR-2005-0036.

75 Starr, M. May,  2003 "Transient and Steady-State Emissions Testing of Ten Different
Nonroad Diesel Engines Using Four Fuels; Final Report" (SwRI-08.03316). Prepared for
California Air Resources Board, El Monte, CA by Southwest Research Institute. (EPA
document # 420-R-03-901). This document is available in Docket EPA-HQ-OAR-2005-
0036.
76
  40CFRPart80, SubpartD.
77 65 FR 6822 (February 10, 2000)

78 66 FR 5002 (January 18, 2001) http://www.epa.gov/otaq/diesel.html

79
  69 FR 38958 (June 29, 2004)
80 65 FR 6697, February 10, 2000.

81 66 FR 5001, January 18, 2001.

82 69 FR 38958, June 29, 2004.
                                    2-102

-------
83 60 FR 34582, July 3, 1995.
84
  64 FR 15208, March 30, 1999 and 65 FR 24267, April 25, 2000.
85 675 FR 68241, November 8, 2002.




86 67 FR 68241, November 8, 2002.




87 61 FR 52088, October 4, 1996.




88 67 FR 53050, August 14, 2002.




89 64 FR 73300, December 29, 1999 and 68 FR 9746, February 28, 2003.
90
  63 FR 18978, April 16, 1998.
                                   2-103

-------
                            Chapter 3: Table of Contents

Chapter 3: Air Quality and Resulting Health and Welfare Effects of Air Pollution from Mobile
      Sources	3
 3.1    Air Quality and Exposure Measurements	3
    3.1.1   Ambient Monitoring	3
    3.1.2   Population-Based (Representative) Exposure Measurements	7
    3.1.3   Elevated Concentrations and Exposures in Mobile Source-Impacted Areas	10
       3.1.3.1 Concentrations Near Major Roadways	11
         3.1.3.1.1     Particulate Matter	11
         3.1.3.1.2     Gaseous Air Toxics	13
       3.1.3.2 Exposures Near Major Roadways	14
         3.1.3.2.1     In Vehicles	14
         3.1.3.2.2     InHomesand Schools	17
         3.1.3.2.3     Pedestrians and Bicyclists	19
         3.1.3.2.4     Measurement Uncertainties	20
       3.1.3.3 Exposure and Concentrations in Homes with Attached Garages	21
       3.1.3.4 Exposure and Concentrations in Parking Garages	24
       3.1.3.5 Exposure and Concentrations at Service Stations	26
       3.1.3.6 Occupational Exposure	28
 3.2    Modeled Air Quality, Exposures, and Risks for Air Toxics	29
    3.2.1   National-Scale Modeled Air Quality, Exposure, and Risk for Air Toxics	29
       3.2.1.1 Air Quality Modeling	31
         3.2.1.1.1     Methods	31
         3.2.1.1.2     Air Quality Trends for Air Toxics	32
         3.2.1.1.3     Distributions of Air Toxic Concentrations across the U. S	35
         3.2.1.1.4     Impacts of Proposed Fuel Benzene Controls on Ambient
                Concentrations	36
       3.2.1.2 Exposure and Risk Modeling	39
         3.2.1.2.1     Methods	39
         3.2.1.2.2     Exposure and Risk Trends for Air Toxics	43
         3.2.1.2.3     Distributions of Air Toxics Risk across  the U. S	49
         3.2.1.2.4     Impacts of Proposed Fuel Benzene Controls on Average Inhalation
                Cancer Risk	57
       3.2.1.3 Impacts of Near Roadway Microenvironment on Modeled Exposures to
             Benzene	61
         3.2.1.3.1     Assessment Methods	61
         3.2.1.3.2     Results	63
       3.2.1.4 Strengths and Limitations	65
       3.2.1.5. Perspective on Cancer Cases	68
    3.2.2   Local-Scale Modeling	71
 3.3 Ozone	75
    3.3.1   Science of Ozone Formation	75
    3.3.2   Health Effects of Ozone	77
    3.3.3   Current 8-Hour Ozone Levels	80

                                         3-1

-------
     3.3.4  Projected 8-Hour Ozone Levels	82
        3.3.4.1 CAIR Ozone Air Quality Modeling	82
        3.3.4.2 Ozone Response Surface Metamodel Methodology	87
        3.3.4.3 Ozone Response Surface Metamodel Results	90
     3.3.5  Environmental Effects of Ozone Pollution	91
        3.3.5.1 Impacts on Vegetation	91
   3.4ParticulateMatter	93
     3.4.1  Science of PM Formation	93
     3.4.2  Health Effects of Paniculate Matter	94
        3.4.2.1 Short-Term Exposure Mortality and Morbidity Studies	94
        3.4.2.2 Long-Term Exposure Mortality and Morbidity Studies	95
        3.4.2.3 Roadway-Related Pollution Exposure	95
     3.4.3  Current and Projected PM Levels	96
        3.4.3.1 Current PM2.5 Levels	96
        3.4.3.2 Current PMio Levels	97
        3.4.3.3ProjectedPM2.5Levels	97
           3.4.3.3.1    PM Modeling Methodology	97
           3.4.3.3.2   Areas at Risk of Future PM2.5 Violations	98
     3.4.4  Environmental Effects of PM Pollution	100
        3.4.4.1 Visibility Degradation	100
           3.4.4.1.1    Current Visibility Impairment	101
           3.4.4.1.2   Current Visibility Impairment at Mandatory Class I Federal Areas.. 101
           3.4.4.1.3    Future Visibility Impairment	102
           3.4.4.1.4   Future Visibility Impairment at Mandatory Class I Federal Areas.... 102
        3.4.4.2 Atmospheric Deposition	103
           3.4.4.2.1    Heavy Metals	103
           3.4.4.2.2   Polycyclic Organic Matter	104
        3.4.4.3 Materials Damage and Soiling	105
   3.5     Health and Welfare Impacts of Near-Roadway  Exposure	105
     3.5.1  Mortality	107
     3.5.2  Non-Allergic Respiratory Symptoms	108
     3.5.3  Development of Allergic Disease and Asthma	109
     3.5.4  Cardiovascular Effects	110
     3.5.5  Birth Outcomes	Ill
     3.5.6  Childhood Cancer	Ill
     3.5.7  Summary  of Near-Roadway Health Studies	113
Appendix 3 A: 8-Hour Ozone Nonattainment	115
Appendix 3B: PM Nonattainment	131
Appendix 3C: Visibility Tables	134
                                          5-2

-------
 Chapter 3: Air Quality and Resulting Health and Welfare Effects of
                      Air Pollution from Mobile Sources

3.1    Air Quality and Exposure Measurements

3.1.1   Ambient Monitoring

       Ambient air toxics data is useful for identifying pollutants of greatest concern, areas of
unhealthy ambient air toxics concentrations, and air toxics trends; evaluating and improving
models; and assessing the effectiveness of air toxics reduction strategies. Ambient air toxics data
though has limitations for use in risk assessments.  While EPA, states, tribes, and local air
regulatory agencies collect monitoring data for a number of toxic air pollutants, both the
chemicals monitored and the geographic coverage of the monitors vary from state to state.1 In
recent years, the US EPA and states have initiated more extensive monitoring of air toxics to
assist in air pollution management through measurement and mitigation.2 EPA is working with
its regulatory partners to build upon the existing monitoring sites to create a national monitoring
network for a number of toxic air pollutants.  The goal is to ensure that those compounds that
pose the greatest risk are measured.  EPA also recently published a draft National Air Toxics
Monitoring Strategy to advance this goal.3

       The available monitoring data help air pollution control agencies track trends in toxic air
pollutants in various locations around the country. EPA conducted a pilot city monitoring
project in 2001 that included sampling in four urban areas and six small city/rural areas (see
Figure 3.1-1). This program helped answer several important national network design questions
(e.g., sampling and analysis precision, sources of variability, and minimal detection levels).

               Figure 3.1-1. Map of Ten Cities in Monitoring Pilot Project


                     • Seattle

                   u                               &
                                                      Detroit        •
                                                       •      Providence
                            Grand Junction
                               •                        •
                                                      Charleston
                   San Jacinto     Rio Rancho
                                                                San Juan
                                                      Tampa*       |
                                          5-3

-------
       Building on the pilot program, the US EPA and states established a national air toxics
monitoring program beginning with a 10-city pilot program, which now consists of 22 national
air toxics trends sites (NATTS), and numerous community-scale monitoring studies.4 To guide
development of the monitoring program, a qualitative data analysis project was begun in 2001
and the first phase was completed in 2004.  The analysis showed that typical urban concentration
ranges for most VOCs are approximately an order of magnitude (or more) higher than the
background concentrations. Because air toxics concentrations vary spatially, other monitoring
networks are needed to provide additional, especially rural, concentrations.  Extrapolation for
most air toxics beyond the urban scale is not recommended without a network of rural
measurements capable of capturing gradients between urban and rural areas. For the latest
information on national air toxics monitoring, see www.epa.gov/ttn/amtic/airtxfil.html.

       Figure 3.1-2 shows measurements of benzene taken from 95 urban monitoring sites
around the country.  These urban areas generally have higher levels of benzene than other areas
of the country.  Measurements taken at these sites show, on average, a 47% drop in benzene
levels from 1994 to 2000. During this period, EPA phased in new (so-called "tier 1") car
emission standards; required many cities to begin using cleaner-burning gasoline; and set
standards that required significant reductions in benzene and other pollutants emitted from oil
refineries and chemical processes.

Figure 3.1-2. Ambient Benzene, Annual Average Urban Concentrations, Nationwide,  1994-
                                         2000
                           of sites hove concentrations below this line
                           90% of sites have concentrations below this line
                                     96       97       96       99
                                  1994-00: 47% decrease
       Following is a summary of analyses recently performed on ambient measurements of air
toxics to identify pollutants and geographic areas of concern and to evaluate trends. Use of
monitoring data to evaluate and improve models is discussed in Section 3.2.

       New York State has a systematic program in place that has been measuring air toxics
since the 1990s.5  The network of monitors is located throughout urban, industrial, residential
and rural locations.  The New York State Department of Environmental Conservation recently
examined the spatial and temporal characteristics of benzene by analyzing five of the 32 total
network sites across the state (see Table 3.1-1).  Spatial trends show a wide range of annual
                                          3-4

-------
average benzene concentrations, with the lowest value at a rural site and the highest at an
industrial site.  The recent 3-year period of 2001-2003 was also compared with the longer 1990-
2003 period. The 3-year period exhibits a decrease in mean concentration compared to the entire
period, indicating that benzene concentrations are decreasing over New York State throughout
this period. The mean annual rate of change in the period 1990 to 2003 was determined using
linear regression of the concentration data. The analysis indicated that ambient concentration
levels of benzene  decreased by as much as 60% during 1990 to 2003. These decreases occurred
in ozone nonattainment areas that had reformulated gasoline (RFG) requirements as well as in
the rest of the state.  The downward trend can be attributed to regulatory measures aimed at
reducing toxic emissions from industrial sources, replacement of older higher emitting vehicles
with vehicles meeting more stringent EPA standards for hydrocarbon emissions, as well as the
adoption of RFG in 1995 and 1999 for the  1-hour ozone nonattainment areas in New York State.
Since trends were observed for sites that were not part of the RFG program, decreases may also
be attributed to the improvement in vehicle emissions technology and the state-wide adoption of
the California Low Emission Vehicle program.

       The downward trend in benzene concentrations reported for New York is consistent with
other reported  changes in ambient levels of benzene. In California, the Air Resources Board
(ARB) maintains an Almanac of Emissions and Air Quality.6  The Almanac summarizes
statewide emissions, statewide annual average concentrations (calculated as a mean of monthly
means), and statewide average health risks for selected air toxics.  Currently there are data
available for ten air toxics in California, including benzene. The ARB network consists of 18 air
quality monitoring stations. The data collected, analyzed, and reported reflect a spatial average;
therefore, ambient concentrations for individual locations may  be higher or lower.  Estimates
show that approximately 84% of the benzene emitted in California comes from motor vehicles,
including evaporative leakage and unburned fuel exhaust.  The predominant sources of total
benzene emissions in the atmosphere are gasoline fugitive emissions and gasoline motor vehicle
exhaust. Approximately 49% of the statewide benzene emissions can be attributed to on-road
motor vehicles, with an additional 35% attributed to other mobile sources such as recreational
boats, off-road recreational vehicles, and lawn and garden equipment.  Currently, the benzene
content of gasoline is less than 1%.  Some of the benzene in the fuel is emitted from vehicles  as
unburned fuel. Benzene is also formed as a partial combustion product of larger aromatic fuel
components. Industry-related stationary sources contribute 15% and area-wide sources
contribute 1%  of the statewide benzene emissions. The primary stationary sources of reported
benzene emissions are crude petroleum and natural gas mining, petroleum refining, and electric
generation. The primary area-wide sources include residential combustion of various types such
as cooking and water heating. The primary natural sources are petroleum seeps that form where
oil or natural gas emerge from subsurface sources to the ground or water surface.  The statewide
benzene levels have shown generally steady improvement since 1990.  To examine the trend  in
benzene while minimizing the influences of weather on the trend, the statewide average benzene
concentration for  1990-1992 was compared to that for 2001-2003. The result was a 72%
decrease in benzene concentration.  These downward trends for benzene and other air toxics are
a result of many control measures implemented to reduce emissions.
                                          5-5

-------
Table 3.1-1. Site Descriptions of the Monitoring Stations Along with Mean Benzene Concentration from 1990-2003 and 2001-
                                                       2003

Site Character
Location Area
2000 Population
(thousands)
Annual Vehicle
Miles Traveled
(million miles)
Period 1990-2003
Mean Concentration
(ug/m3)
Period 200 1-2003
Mean Concentration
(ug/m3)
Lackawanna
Industrial
Buffalo
950
8250
5.09
2.26
Eastern District
High School
Urban
Brooklyn
2465
4246
2.85
2.05
Troy
Small Urban
Hudson Valley
153
1413
2.31
1.68
Niagara Falls
Urban Industrial
Niagara
220
1546
1.80
1.08
Whiteface
Mountain Base
Lodge
Rural
Essex
39
577
0.86
0.54
                                                        3-6

-------
       Another recent evaluation of hazardous air pollutant (HAP) trends was conducted for
selected metropolitan areas.7  Researchers retrieved historical concentration and emissions data
from the US EPA for Boston, New York City, Philadelphia, Tampa Bay, Detroit, Dallas, St.
Louis, Denver, Los Angeles, and Seattle, chosen for each of EPA's ten regions.  Annual and
seasonal trends were generated to evaluate reductions in HAP emissions and ambient
concentrations during the time period 1990-2003. Several air toxics were targeted, including
benzene. To evaluate the trends, average concentrations from 1990-1994 were compared to
2002-2003 (these time periods were chosen due to availability of data).  The results showed that
over 85% of the metropolitan area-HAP combinations decreased in their HAP concentrations,
while less than 15% realized an increase.  For example, Table 3.1-2 shows that benzene
concentrations decreased in seven of the ten metropolitan areas (range 19 to 79%).

       Each of these  analyses consistently illustrates the significant reductions in national annual
average concentrations of benzene and other air toxics. The air pollution management efforts of
the US EPA and states have been effective.  Additional reductions are expected with the
implementation of additional regulatory measures such as this one.

3.1.2   Population-Based (Representative) Exposure Measurements

       In addition to  measurements of outdoor and microenvironmental concentrations, an
important component of understanding human exposure to air toxics is the body of studies that
employ survey techniques to assess representative populations' exposures. Typically, these
studies are designed to represent a discrete geographic area.  The  personal exposure
concentration summaries from these studies are shown in Table 3.1-3.

       The National Human EXposure Assessment Survey (NHEXAS) was a series of
population-based exposure studies. The states in EPA Region 5 were the focus of one NHEXAS
study, which was conducted in mid-1990.8 Nearly 400 personal and indoor air samples were
obtained from both smokers and non-smokers, along  with a smaller number of outdoor air
samples in residential areas. Measurements took place over 6 days per subject.  Overall, average
personal exposure to benzene was 7.52  ug/m3, with indoor air concentrations averaging 7.21
ug/m3.  Outdoor air concentrations averaged 3.61 ug/m3. Personal air concentrations were
significantly associated with indoor air concentrations, as well as  blood concentrations.

       The results of the NHEXAS study in Arizona, another study area, indicate that median
indoor concentrations were 1.3 ug/m3 during the mid-1990's, while outdoor concentrations were
1.0 ug/m3.9  Furthermore, reported results from the Arizona study indicate that fuel-related
VOCs are elevated in homes with attached garages.

       In another study based on a random population-based sample of an urban population, 37
non-smoking residents of South Baltimore, MD were equipped with passive monitors to assess
3-day average personal exposure to VOCs, in addition to indoor and outdoor air.10 Monitoring
took place in 2000 and 2001.  Modeled air quality data from the ASPEN dispersion model,
employed in EPA's National Air Toxics Assessment  for 1996, was also obtained for the study
                                          5-7

-------
Table 3.1-2. Benzene Emission (Tons Per Year) and Concentration (ug/m3) Comparison
Metropolitan
Area
Boston
New York City
Philadelphia
Tampa Bay
Detroit
Dallas
St. Louis
Denver
Los Angeles
Seattle
1990
Emissions
6262
16653
5961
3103
6480
7933
4358
2800
19762
5844
2002
Emissions
2229
7512
2577
2408
4388
2832
2304
1913
4168
4315
% Change
in
Emissions
-64.4
-54.9
-56.8
-22.4
-32.3
-64.3
-47.1
-31.7
-78.9
-26.2
1990-1994
Average
Concentration
3.93
3.24
3.60
NA
4.19
1.21
5.16
NA
8.97
NA
2002-2003
Average
Concentration
0.81
1.35
1.26
NA
3.40
0.78
1.43
2.75
2.34
1.39
% Change
in
Concentration
-79.5
-58.5
-64.9
NA
-18.7
-35.8
-72.3
NA
-73.9
NA

-------
area.  Overall, outdoor modeled concentrations of benzene and other fuel-related VOCs
corresponded well with measured data in the area. Average personal exposure to benzene was
4.06 ug/m3, while 95th percentile values were 7.30 ug/m3. For indoors, the respective values
were 3.70 and 8.34 ug/m3, while for outdoors the values were 1.84 and 3.14 ug/m3. Overall, the
study provides evidence that modeling outdoor benzene concentrations using ASPEN, as is done
in this rule, provides adequate representation of outdoor values. However, indoor and personal
exposures are also influenced by other sources, as is described in the section on attached garages.

       While not a population-based study, the recently-completed Relationship Between
Indoor, Outdoor and Personal Air (RIOPA) study provides a depiction of indoor, outdoor, and
personal concentrations of benzene and other toxics in three regions with differing source
mixtures.11 100 non-smoking homes in each of Los Angeles, CA, Houston, TX, and Elizabeth,
NJ were selected for sampling in areas representing locations dominated by emissions from
mobile sources, stationary sources, and a mixture of sources, respectively.  In the adult sample,
average personal exposures to benzene were 3.64 ug/m3, with a 95th percentile of 10.7 ug/m3.
Respective statistics for indoor air were 3.50 ug/m3 and 10.0 ug/m3, while outdoor statistics were
2.15 and 5.16 ug/m3.  In further EPA-funded analysis of the data from Elizabeth, NJ,
concentrations of benzene, toluene, ethylbenzene, and xylene isomers were found to be
associated with proximity to both major roadways and gas stations, as was PM2.5, EC, and
several PAHs.12'13 Section 3.1.3 provides more detail on concentrations and exposures in these
types of mobile-source impacted areas.

       Few studies have systematically addressed exposures among representative samples of
children.  Several have been done in Minnesota, with  others in New York, Los Angeles, and
Baltimore areas.

       For the Minnesota Children's Pesticide Exposure Study  (MNCPES), conducted in urban
and rural areas in the vicinity of Minneapolis-St. Paul, MN,14 all monitoring used the same 6-day
monitoring duration as used in the Region 5 NHEXAS study. In the first phase of the study, a
statistically representative sample of 284 homes with children underwent air monitoring for
VOCs. Low-income and minority homes were over sampled to ensure representation.  Indoor
benzene concentrations averaged 4.6 ug/m3, with the data skewed right.  The 95th  percentile
concentration was 12.7 ug/m3.  Homes with attached garages had significantly higher
concentrations of benzene indoors. In the  second phase of the study, a subset of 100 children
underwent intensive monitoring of personal, indoor, and outdoor air as well as activity tracking
via diary.  Overall personal exposures were 4.8 ug/m3, with a 95th  percentile of 9.1 ug/m3.
Indoor concentrations in the intensive period averaged 3.9 ug/m3 and outdoor averaged 3.3
ug/m3. Regression analysis indicated that  personal exposures generally were higher than the
time-weighted average of indoor and outdoor air. Furthermore, living in a home with an
attached garage was associated with elevated personal exposures to both benzene and toluene.
                                           5-9

-------
       In another study, students recruited from an inner-city school in Minneapolis, MN
participated in an exposure study called SHIELD.15  Students were recruited using stratified
random sampling, with a total of 153 children participating between two seasons. Home and
personal samples were collected and averaged over two continuous days of sampling using
passive methods.  School measurements took place during school hours only, over the course of
5 days, and outdoor measurements were set up to run continuously outside the school through
each week sampled (Monday through Friday).  The study reported median,  10th, and 90th
percentile concentrations. In personal samples, median benzene concentrations were 1.5 ug/m3
in spring and 2.1 ug/m3 in winter.16

       The TEACH exposure study tracked inner-city high school students' exposures in New
York, NY and Los Angeles, CA.  In the New York City study, 42 students underwent personal,
indoor home, and outdoor home air quality monitoring during two seasons.17  Average winter
benzene personal concentrations were 4.70 ug/m3, while indoor and outdoor concentrations
averaged 5.97 and 2.55 ug/m3. Summer values were 3.09, 1.75,  and  1.31 ug/m3, respectively.
The authors noted that VOC concentrations within the city tracked traffic patterns.  Generally,
indoor concentrations in Los Angeles were of similar magnitude, while personal exposures were
not reported as of the time of this proposal.  There was no substantial evidence for indoor sources
of benzene.18

       Overall, these studies show that personal and indoor concentrations of benzene and other
VOCs are significantly higher than found outdoors.  Some of the factors leading to these elevated
concentrations are a result of motor vehicle impacts such as exhaust and evaporative emissions in
attached garages, exposures during on-road commutes and exposures during vehicle re-fueling.
These and other factors are discussed in more detail in Section 3.1.3. This suggests that risk
reductions from the controls in this proposal will be greater than can  currently be estimated using
national-scale modeling tools.

3.1.3   Elevated Concentrations and Exposures in Mobile Source-Impacted Areas

       Air quality measurements near roads often identify elevated concentrations of air toxic
pollutants at these locations.  The concentrations of air toxic pollutants near heavily trafficked
roads, as well as the pollutant composition and characteristics, differ from those measured distant
from heavily trafficked roads. Thus, exposures for populations residing, working, or going to
school near major roads are likely different than for other populations.  Following is an overview
of concentrations of air toxics and exposure to air toxics in areas experiencing elevated pollutant
concentrations due to the impacts of mobile source emissions.
                                          3-10

-------
        Table 3.1-3. Personal Exposure to Benzene from Population-Based Studies"
Location
EPA Region 5
Baltimore, MD
NJ, TX, CA
Minneapolis -
St. Paul, MN
Minneapolis,
MN
New York, NY
Year(s)
1995-
1996
2000-
2001
1999-
2001
1997
2000
1999
Includes
Smokers
Yes
No
No
Yese
Yese
No
Average
(ug/m3)
7.21
4.06
3.64
4.8
2.1 Winter
1.5 Spring
4. 7 Winter
3. 1 Summer
"Upper
Bound"
(ns/m3)
13.71b
7.30C
10.7C
9.1
6.5 Winter"
4.2 Springb
11.4Winterd
7. 0 Summerd
Reference
Clayton et al.
(1999)
Payne-Sturges et
al. (2004)
Weisel et al.
(2005)
Adgate et al.
(2004a)
Adgate et al.
(2004b)
Kinney et al.
(2002)
   a Children's studies in italics
   b 90th percentile
   c 95th percentile
   d Mean +2 standard deviations
   e Smoking in homes
3.1.3.1 Concentrations Near Major Roadways

3.1.3.1.1      Parti culate Matter

       Mobile sources influence temporal and spatial patterns of criteria pollutants, air toxics,
and PM concentrations within urban areas. Motor vehicle emissions may lead to elevated
concentrations of pollutants near major roads. Since motor vehicle emissions generally occur
within the breathing zone, near-road populations may be exposed to "fresh" primary emissions as
well as combustion pollutants "aged" in the atmosphere. For paniculate matter, these fresh
versus aged emissions can result in the presence of varying particle sizes near roadways,
including ultrafine, fine, and coarse particle modes.

       The range of particle sizes of concern is quite broad and is divided into smaller
categories. Defining different size categories is useful since particles of different sizes behave
differently in the atmosphere and in the human respiratory system.  Table 3.1-4 lists the four
terms for categorizing particles of different sizes as defined by the US EPA.19
                                          3-11

-------
         Table 3.1-4. Descriptions and Particle Sizes of Each Category of Particles
Description
Supercoarse
Coarse (or Thoracic Coarse Mode)
Fine (or Accumulation Mode)
Ultrafme (or Nuclei Mode)a
Particle Size, dp (urn)
dp>10
2.5 < dp < 10
0.1 
-------
emission rates, atmospheric chemistry, and meteorology.  Re-entrained road dust, as well as
brake and tire wear will also contribute to increased concentrations of coarse PM.

       Meteorological factors can affect exposures to motor vehicle emissions near the road.
Researchers have noted that particle number concentrations changed significantly with changing
wind conditions, such as wind speed, near a road.33  Wind direction also affects traffic-related air
pollution mass concentrations inside and outside of schools near motorways.34'35 Diurnal
variations in mixing layer height will also influence both near-road and regional air pollutant
concentrations.  Decreases in the height of the mixing layer (due to morning inversions, stable
atmosphere, etc.) will lead to increased pollutant concentrations at both local and regional scales.

3.1.3.1.2     Gaseous Air Toxics

       Concentrations of mobile source air toxics have been estimated by a number of different
sources such as the NATA National-Scale Assessment, local-scale modeling assessments, and
from air quality monitoring in locations in immediate proximity to busy roadways. Each
approach offers a different level of representation of the concentrations of air toxics near
roadways.

       The NATA National-Scale Assessment estimates average concentrations within a census
tract, but it does not differentiate between locations near roadways and those further away.
Local-scale modeling can better characterize distributions of concentrations, as  observed in
assessments done in Houston, TX and Portland, OR. The Houston study calculated the average
benzene concentration to be 2.29 ug/m3,36 using the same emissions inventory as used in the
1996 NATA National-Scale Assessment but with more refined allocation of highway vehicle
emissions. In this study, spatially defined inventories placed vehicle emissions  at the location of
actual roadway links, thus characterizing with greater resolution the spatial distribution of
ambient benzene concentrations. As a result, there was better agreement with monitor data (2.97
ug/m3), than what was obtained by gridding emissions (2.09 ug/m3). The Portland study
modeled concentrations of air toxics at the center of every census block group in the Portland,
OR metropolitan area.37  A subsequent analysis determined average 1,3-butadiene, benzene, and
diesel PM concentrations at several distances from major  roadways (0-50, 50-200, 200-400, and
> 400 m). For benzene, the resulting average concentrations were 1.29,  0.64, 0.40, and 0.12
ug/m3, respectively, illustrating the steep concentration gradient around roadways. The overall
mean benzene concentration modeled in Portland was 0.21 ug/m3.

       Air quality monitoring is another means of evaluating pollutant concentrations at
locations near sources such as roadways. Several studies  have found that concentrations of
benzene and other mobile source air toxics are significantly elevated near busy roads compared
to "urban background" concentrations measured  at a fixed site 38>39>40>41'42>43 por example,
measurements near a tollbooth in Baltimore observed mean benzene concentrations to vary by
time of day from 3 to 22.3  ug/m3 depending on traffic volume, vehicle type, and meteorology.44
In comparison with ambient levels, Maryland's Department of Environment reported the range
of benzene annual averages measured at seven different monitoring sites in 2000 between 0.27-
0.71 ug/m3.45 Another study measured the average benzene concentration in a relatively high
traffic density (~ 16000 automobiles/day) sampling area at 9.6 ug/m3 and in rural  areas with
                                          3-13

-------
hardly any traffic (< 50 automobiles/day) at 1.3 ug/m3.46 The concentration of benzene, along
with several other air toxics (toluene and the isomeric xylenes), in the urban area far exceeded
those in the rural area.

       Ambient VOC concentrations were measured around residences in Elizabeth, NJ, as part
of the Relationship among Indoor, Outdoor, and Personal Air (RIOPA) study. Data from that
study was analyzed to assess the influence of proximity of known ambient emission sources on
residences.47 The ambient concentrations of benzene, toluene, ethylbenzene, and xylene isomers
(BTEX) were found to be inversely associated with: distances from the sampler to interstate
highways and major urban roads; distance from the sampler to gasoline stations; atmospheric
stability; temperature; and wind speed. The data indicate that BTEX concentrations around
homes within 200 m of roadways and gas stations are 1.5 to 4 times higher than urban
background  levels.

       According to Gaussian dispersion theory, pollutants emitted along roadways will show
highest concentrations nearest a road, and concentrations exponentially decrease with increasing
distance downwind. These near-road pollutant gradients have been confirmed by measurements
of both criteria pollutants and air toxics.48'49'50'51'52 Researchers have demonstrated exponential
decreases in CO, as well as PM number, and black carbon (as measured by an aethalometer),
concentration with increasing downwind distance from a freeway in Los Angeles.53'54 These
pollutants reached background levels approximately 300 m downwind of the freeway.

3.1.3.2 Exposures Near Major Roadways

       The modeling assessments and air quality monitoring studies discussed above have
increased our understanding of ambient concentrations of mobile source air toxics and potential
population exposures. Results from the following exposure studies reveal that populations
spending time near major roadways likely experience elevated personal exposures to motor
vehicle related pollutants. In addition, these populations may experience exposures to differing
physical and chemical compositions of certain air toxic pollutants depending on the amount of
time spent in close proximity to motor vehicle emissions. Following is a detailed discussion on
exposed populations near major roadways.

3.1.3.2.1     In Vehicles

       Several studies suggest that people may experience  significant exposures while driving in
vehicles. A recent in-vehicle monitoring study was conducted by EPA and consisted of in-
vehicle air sampling throughout work shifts within ten police patrol cars used by the North
Carolina State Highway Patrol (smoking not permitted inside the vehicles).55 Troopers operated
their vehicles in typical patterns, including highway and city driving and refueling. In-vehicle
benzene concentrations averaged 12.8 ug/m3, while concentrations measured at an "ambient" site
located outside a nearby state environmental office  averaged 0.32 ug/m3.  The study also found
that the benzene concentrations were closely associated with other fuel-related VOCs measured.

       The American Petroleum Institute funded a  screening study of "high-end" exposure
microenvironments as required by section 21 l(b) of the Clean Air Act.56 The study included
                                          3-14

-------
vehicle chase measurements and measurements in several vehicle-related microenvironments in
several cities for benzene and other air toxics. In-vehicle microenvironments (average
concentrations in parentheses) included the vehicle cabin tested on congested freeways (17.5
ug/m3), in parking garages above-ground (155 ug/m3) and below-ground (61.7 ug/m3), in urban
street canyons (7.54 ug/m3), and during refueling (46.0 ug/m3). It should be noted that sample
sizes in this screening study were small, usually with only one to two samples per
microenvironment.

      In 1998, the California Air Resources Board published an extensive study of
concentrations of in-vehicle air toxics in Los Angeles and Sacramento, CA.57 The data set is
large and included a variety of sampling conditions.  On urban freeways, in-vehicle benzene
concentrations ranged from 3 to  15 ug/m3 in Sacramento and  10 to 22 ug/m3 in Los Angeles. In
comparison, ambient benzene concentrations ranged from 1 to 3 ug/m3 in Sacramento and 3 to 7
ug/m3 in Los Angeles.

       Studies have also been conducted in diesel buses, such as the one recently conducted of
LA school buses.58'59 In the study, five conventional diesel buses, one diesel bus equipped with a
catalytic particle filter, and one natural gas bus were monitored for benzene, among other
pollutants. These buses were driven on a series of real school bus routes in and around Los
Angeles,  CA.  Average benzene  concentrations in the buses were 9.5 ug/m3, compared with 1.6
ug/m3 at a background urban fixed site in west Los Angeles. Type of bus, traffic congestion
levels, and encounters with other diesel vehicles contributed to high exposure variability between
runs.

      The same researchers additionally determined the relative importance of school bus-
related microenvironments to children's pollutant exposure.60 Real-time concentrations of black
carbon (BC), particle-bound PAH,  nitrogen  dioxide (NC^), particle counts (0.3-0.5 um size
range), and PM2.5 mass were measured inside school buses during long commutes, at bus stops
along the routes, at bus loading and unloading zones, and at nearby urban background sites.
Across all the pollutants, mean concentrations during bus commutes were higher than in any
other microenvironment. Mean exposures in bus commutes were 50 to 200 times  more than for
loading and unloading zones at the school, and 20 to 40 times more than for bus stops along the
route, depending on the pollutant.  The in-cabin exposures were dominated by the effect of
surrounding traffic when windows were open and by the bus'  own exhaust when the windows
were closed. The mean pollutant concentrations in the three school bus commute-related
environments and background air are presented in the Table 3.1-5.
                                          3-15

-------
   Table 3.1-5.  Mean Concentrations of Black Carbon (BC), Particle Bound PAH, NO2,
  Particle Count (PC), and PMi.s in Three School Bus Commute Microenvironments and
                                    Background Air

BC (ug/m3)
Particle Bound
-PAH (ug/m3)
N02 (ppb)
PC
(count/cm3)
PM2.5 (ug/m3)
Mean Concentrations
Background
2 ±0.1
0.027 ±0.0015
49 ± 1.0
83±3.1
20 ±2.4
(Un)Loading
Zone
2 ±0.3
0.015 ±0.0003
35 ±0.2
Not collected
Not collected
Bus Stops
4 ±0.4
0.044 ±0.0045
54± 1.9
62 ± 1.8
25b
Bus
Commutes3
3-19(8)
0.064-0.400
(0.134)
34-110(73)
77-236 (130)
21-62 (43)
     aRanges are associated with different bus types and window positions. Values in
     parenthesis are the mean for all runs.
     bNot enough data to establish a confidence interval.

       In another recent study of commuter buses, concentrations of benzene and other VOCs
were measured in buses on several routes in Detroit, MI.61  The average in-bus concentration of
benzene was 4.5 ug/m3, while the average concentrations at three fixed sites taken during the
study period ranged from 0.9-2.0 ug/m3.  In this study, daily bus/ambient concentration ratios
were reported, and ranged from 2.8-3.3 on the three reported study days.  The in-bus
concentrations were found to be most influenced by local traffic sources.  A number of other
studies similarly observe that passenger car commuters are exposed to elevated pollutant
concentrations while driving on busy roads.62'63'64'65'66'67

       Older studies that examine in-vehicle concentrations in older model year vehicles are
difficult to apply for regulatory analyses, due to the relatively rapid changes in vehicle emission
controls over the last 15 years. In general, these studies indicate that concentrations in vehicles
are significantly higher than ambient concentrations.68'69'70 The average benzene measurements
of these older in-vehicle studies (Raleigh, NC and CA South  Coast Air Basin) are in Table 3.1-6
along with the more recent studies for comparison.

       Overall, these studies show that concentrations experienced by commuters and other
roadway users are substantially higher than ambient air measured in typical urban air.  As a
result, the time a person spends in a vehicle will significantly affect their overall exposure.
                                          3-16

-------
  Table 3.1-6. Benzene Concentrations (ug/m3) Measured in Vehicles and in Ambient Air
Study

Raleigh, NC (1 989) a
CA South Coast Air Basin (1989) b
Boston, MA (1991) c
Los Angeles, C A (1998)
Sacramento, CA (1998)
Detroit, MI (2000) d
API Gasoline Screening (2002)
LA, CA School Buses (2003)
NC State Highway Patrol (2003)
In- Vehicle
Mean
11.6
42.5
17.0
10-22
3-15
4.5
17.5
9.5
12.8
Max
42.8
267.1
64.0
—
—
10.8
—
—
43.1
Ambient Air
Mean
1.9
9.3-16.9
—
3-7
1-3
0.9-2.0
—
1.6
0.32
Max
8.5
—
—
—
—
—
—
—
1.92
3.1.3.2.2
a A one-hour measurement was taken for each experimental trip.
b The estimated sampling time period was 1.5 hours/round-trip. n=191.
c In-vehicle measurement includes both interstate and urban driving, n=40.
d Measurements taken from interiors of urban buses.

   In Homes and Schools
       The proximity of schools to major roads may result in elevated exposures for children
due to potentially increased concentrations indoors and increased exposures during outdoor
activities. Here we discuss international studies in addition to the limited number of US studies,
because while fleets and fuels outside the U.S. can be much different, the spatial distribution of
concentrations is relevant.

       There are many sources of indoor air pollution in any home or school. These include
indoor sources and outdoor sources, such as vehicle exhaust.  Outdoor air enters and leaves a
house by infiltration, natural ventilation, and mechanical ventilation. In infiltration, outdoor air
flows into the house through openings, joints, and cracks in walls, floors, and ceilings, and
around windows and doors. In natural ventilation, air moves through opened windows and doors.
Air movement associated with infiltration and natural ventilation is caused by air temperature
differences between indoors and outdoors and by wind. Finally, there are a number of
mechanical ventilation devices, from outdoor-vented fans that intermittently remove air from a
single room, such as bathrooms and kitchen, to air handling systems that use fans and duct work
to continuously remove indoor air and distribute filtered and conditioned outdoor air to strategic
points throughout the house.  The majority of what is outdoors can therefore get indoors.  A
review of the literature determined that approximately 100% of gaseous compounds, such as
benzene, and 80% of diesel PM can penetrate indoors.71'72

       In the Fresno Asthmatic Children's Environment Study (FACES), traffic-related
pollutants were measured on selected days from July 2002 to February 2003 at a central site, and
inside and outside of homes and outdoors at schools of asthmatic children.73 Preliminary data
indicate that PAH concentrations are higher at elementary schools located near primary roads
than at elementary schools distant from primary roads (or located near primary roads with
                                          3-17

-------
limited access). PAH concentrations also appear to increase with increase in annual average
daily traffic on nearest major collector. Remaining results regarding the variance in traffic
pollutant concentrations at schools in relation to proximity to roadways and traffic density will
be available in 2006.

       The East Bay Children's Respiratory Health Study studied traffic-related air pollution
outside of schools near busy roads in the  San Francisco Bay Area in 2001,74 Concentrations of
the traffic pollutants PMi0, PM2 5, black carbon, total NOX, and NO2 were measured at ten school
sites in neighborhoods that spanned a busy traffic corridor during the spring and fall seasons.
The school sites were selected to represent a range of locations upwind and downwind of major
roads. Differences were observed in concentrations between schools nearby (< 300 m) versus
those more distant (or upwind) from major roads.  Investigators found spatial variability in
exposure to black carbon, NOx, NO, and (to a lesser extent) NO2, due specifically to roads with
heavy traffic within a relatively small geographic area.

       An exposure assessment of PMi0  from a major highway interchange in East Los Angeles
found that children in nearby schools were exposed to elevated pollutant levels.75 Each of the
four chosen schools was located within 500 m of a major limited-access highway, and three of
them were within 150 m of the roadway.  Using a computer model to calculate dispersion
analysis, researchers predicted that average 24-hour (assuming 10-hour school-based exposure
duration to account for time in class and at after-school programs) particle concentrations, which
were dominated by road dust, would be 10.45, 14.58, 5.78, and 8.27 ug/m3, respectively, for the
four schools studied. These results indicate a trend for increased emissions at school locations in
closer proximity to the traffic source, with the exception of one school which was 25 m farther.
These values reflect the increase in concentration over ambient exposure, not the total ambient
exposure.
                                                                                      2
       A study to assess children's exposure to traffic-related air pollution while attending
schools near roadways was performed in the Netherlands.76  Investigators measured PM2.5, NO
and benzene inside and outside of 24 schools located within 400 m of roadways.  The indoor
average benzene concentration was 3.2 ug/m3, with a range of 0.6-8.1 ug/m3.  The outdoor
average benzene concentration was 2.2 ug/m3, with a range of 0.3-5.0 ug/m3.  Overall results
indicate that indoor pollutant concentrations are significantly correlated with traffic density and
composition, percentage of time downwind, and distance from major roadways.

       In another study performed in the Netherlands, investigators measured indoor
concentrations of black smoke, PMio, and NO2 in twelve schools between the periods of May
and August 1995.77 The schools were located at varying distances from the motorway (35-645
m). Results indicate that black smoke and NO2 concentrations inside the schools were
significantly correlated with truck and/or car traffic intensity as well as percentage of time
downwind from the motorway and distance of the school from the motorway. PMio
concentrations measured in classrooms during school hours were highly variable and much
higher than those measured outdoors, but they did not correlate with any of the distance or traffic
parameters.
                                          3-18

-------
       The TEACH study (Toxic Exposure Assessment - Columbia/Harvard) measured the
concentrations of VOCs, PM2 5, black carbon, and metals outside the homes of high school
students in New York City.78 The study was conducted during winter and summer of 1999  on 46
students and in their homes. Average winter (and summer) indoor concentrations exceeded
outdoor concentrations by a factor of 2.3 (1.3). In addition, analyses of spatial and temporal
patterns of MTBE concentrations, used as a tracer for motor vehicle pollution, were consistent
with traffic patterns.

       The RIOPA study was conducted in three cities (Los Angeles, CA, Houston, TX, and
Elizabeth, NJ) during four seasons.79'80 The study examined 100 non-smoking homes sited  in
high-emissions environments, including residential areas near freeways, service stations,
petroleum industrial estates, and mixed sources. The cities involved were selected to represent
different sources: Los Angeles (mobile source dominated), Houston (stationary source
dominated), and Elizabeth, NJ (mixture of sources).  Of the poly cyclic aromatic hydrocarbons
(PAHs) analyzed, the presence of 5-7 ring PAHs indoors was attributed to outdoor sources
which, in Los Angeles and Elizabeth, NJ, could be attributed to mobile sources.

       Average benzene concentrations were determined in a recent evaluation of the exposure
of urban inhabitants to atmospheric benzene in Athens, Greece.81  Home and personal levels of
50 non-smokers in six monitoring campaigns varied between 6.0-13.4 and 13.1-24.6 ug/m3,
respectively.  Urban levels varied between 15.4 and 27.9 ug/m3 with an annual mean of 20.4
ug/m3. The highest values were observed during the first two sampling periods in fall and
winter, when wind speed was low. The low summer values were attributed to decreased vehicle
traffic. Among home factors, only proximity to busy roads was determined to be an important
influence on indoor benzene levels.

       Children are exposed to elevated levels of air toxics not only in their homes, classrooms,
and outside on school grounds, but also during their commute to school. See above discussion of
in-vehicle (school bus and passenger car) concentrations of air toxics for one method of
commuting. The discussion below also presents potential  exposures to children from another
commuting method.

3.1.3.2.3       Pedestrians and Bicyclists

       Researchers have noted that pedestrians and cyclists along major roads experience
elevated exposures to motor vehicle related pollutants. Although commuting near roadways
leads to higher levels of exposure to traffic pollutants, the  general consensus is that exposure
levels of those commuting by walking or biking is lower than for those who travel by car or bus,
(see discussion on in-vehicle exposure in previous section above). For example, investigators
found that personal measurements of exposure to PMio concentrations were 16% higher inside
the car than for the walker on the same route, but noted that a walker may have a larger overall
exposure due to an increase in journey time.82  Similarly, researchers found that traffic-related
pollutant exposure concentrations of car drivers were higher than for cyclists.83 Cyclists are
typically on the border of the road or on dedicated bike paths and therefore further away from the
vehicle emissions and are less delayed by traffic jams.  However, after accounting for cyclists'
higher ventilation, the uptake of CO, benzene, toluene, and xylenes by cyclists sometimes
                                          3-19

-------
approached that of car drivers, and for NO2 it was significantly higher.

       In the early 1990's, researchers studied the in-vehicle concentrations of a large number of
compounds associated with motor vehicle use and the exposure to VOCs of a pedestrian on an
urban sidewalk (50 m from roadways) in Raleigh, NC.84 The mean concentration of benzene in
the six pedestrian sidewalk samples was 6.8 ug/m3. This concentration was lower than the in-
vehicle measurement (11.6 ug/m3), but hi
urban roadways 100-300 m from streets.
vehicle measurement (11.6 ug/m3), but higher than the fixed-site measurement (1.9 ug/m3) on
       The same researchers studied the exposure of commuters in Boston to VOCs during car
                                        85
driving, subway travel, walking, and biking.  For pedestrians, mean time-weighted
concentrations of benzene, toluene, and xylenes of 10.6, 19.8, and 16.7 ug/m3, respectively, were
reported. For cyclists, the time-weighted concentrations were similar to those of pedestrians, at
9.2, 16.3, and 13.0 ug/m3, respectively. In-vehicle exposure concentrations were higher as
discussed above.

       Numerous other studies which were conducted in Europe and Asia yield similar results.
A survey of CO concentration was conducted for various transport modes along heavy traffic
routes in Athens, Greece.86 Results showed that mean CO levels for trips of 30 min were 21.4
ppm for private car,  10.4 ppm for bus, and 11.5  ppm for pedestrians. In Northampton, UK
during the winter 1999, personal measurements  of exposure to PMio, PM2.5, and PMi were made
during walking and in-car journeys on two suburban routes.87 In-car measurements were highest
(43.16, 15.54, and 7.03 ug/m3 for PMio, PM2.5, and PMI, respectively) followed by walking
(38.18, 15.06, and 7.14 ug/m3, respectively).  Background levels were only available for PMio
(26.55  ug/m3), but were significantly lower than the walking exposure levels.  Researchers found
similar results for CO exposure levels of schoolchildren commuters.88 So although personal
exposures are greater for in-vehicle commutes, pedestrians and bicyclists in proximity to heavy
traffic are exposed to elevated pollutant levels relative to background.

3.1.3.2.4      Measurement Uncertainties

       A number of uncertainties limit our ability to fully describe the impacts of motor vehicle
emissions.  As described above, most people in the U.S. experience  some level of exposure to
emissions from motor vehicles. Thus, proper characterization of the level of these exposures is
critical. However, the exposure assessment techniques used may not adequately represent the
populations' true exposures to motor vehicle emissions.

       Air quality measurements are expensive and therefore are limited. The high costs of
measurement techniques affect the quantity of samples that can be collected and quantity of
compounds that can be identified.  As a result, measurements may only occur at central
monitoring sites, rather than in microenvironments impacted by motor vehicle emissions. Air
quality monitoring at these central sites often do not represent actual exposures, especially for
populations living near roads.

       Air quality samples are often integrated and therefore lack time resolution.  This can
result in difficulty in determining source contributions. Additionally, some compounds are hard
                                          3-20

-------
to measure accurately. For example, 1,3-butadiene is very reactive in the ambient atmosphere
and has a short atmospheric lifetime, estimated to be only two hours.89 Thus, this compound can
easily break down before samples are analyzed. Also, a vapor pressure of 3.3 atm at 25°C makes
it a very volatile compound.  Secondary reactions are a confounding factor in air quality
measurements and can add additional uncertainty to measured ambient concentrations.

       Results from emissions studies suggest that simple methods of estimating the
contribution of motor vehicle exhaust to exposure likely do not capture the substantial variability
in the chemical  and physical characteristics of motor vehicle exhaust. Comprehensive
assessments of exposure will be a critical  factor in identifying which compounds are impacting
the near-road environment.

3.1.3.3 Exposure and Concentrations in Homes with Attached Garages

       Residential indoor air quality is a major determinant of personal exposure, with  most
people spending the majority of their time indoors at home. According to the National Human
Activity Pattern Survey, nationally, people spend an average of 16.68 hours per day indoors in a
residence.90  The large fraction of time spent in this microenvironment implies that sources that
impact indoor air are likely to have a substantial effect on personal exposure.

       Indoor air quality is in large part determined by ventilation of indoor spaces. Natural
ventilation occurs as a result of two factors: wind-induced pressure and the "stack effect."  The
latter occurs when hot air rises in a home, causing a pressure drop in the lower part of the home,
which then creates airflow into the home from higher-pressure locations outside the home.
Natural ventilation can also be influenced by opening of windows and doors.  Mechanical
ventilation employs  fans and sometimes ductwork to manage ventilation  within a home.

       Air can be drawn into a home from either outdoors, or in a home with an attached garage,
from the garage. Air from the garage can have higher concentrations of VOCs and other
pollutants as a result of the storage of vehicles, other engines and equipment, fuel (gasoline in
gas cans), solvents, or cleaning products.  As a result, homes with a greater fraction of airflow
from the garage are more susceptible to air quality decrements from in-garage emissions.

       Several studies have examined homes with attached garages to determine the fraction of
residential air intake from the garage. One study from Minnesota examined homes constructed
in 1994, 1998, and 2000.91 Homes built in 1994 had 17.4% of airflow originating in the garage.
Homes built in 1998 and 2000 had 10.5% and 9.4% of airflow from the garage, respectively. In
another study conducted in Ottawa, Ontario, an average of 13% of home  air intake came from
the garage.92  That study also found that the house-garage interface area was as leaky as the rest
of the building envelope. In another study from Washington, D.C., the house-garage interface
was found to be 2.5 times as permeable as the rest of the house.93 This discrepancy may indicate
that homes built in colder climates are built more tightly than homes in warmer regions as a
result of weather-sealing. However, there is no evidence that in regions with cold weather,
colder temperatures  lead to elevated indoor concentrations of VOCs.94
                                          3-21

-------
       Several studies have examined the influence of attached garages on indoor air and
personal exposure. In the 1980's researchers identified attached garages as a major source of
benzene and other VOCs in residences. The Total Exposure Assessment Methodology (TEAM)
was a series of large, probability-based samples of people who underwent study of the air inside
and outside their homes and in their personal breathing zones.  The study took place in the
1980's, and found that a large fraction of an average nonsmoker's benzene exposure originated
from sources in attached garages.95  Work done as part of the TEAM study also identified stored
gasoline as an important source of elevated benzene levels indoors.96 This stored gasoline can be
found primarily in gas cans as well as the fuel tanks of lawn and garden equipment, such as lawn
mowers and string trimmers. Gas can emissions, however,  are significantly higher than
evaporative emissions from lawn and garden equipment, because the fuel tanks are much smaller
than gas cans, typically 0.3 to 0.4 gallons.  Emissions are also higher from gas cans because
vents and spouts are left open.

       These early studies have highlighted the role of evaporative emissions within the garage
as contributors to indoor air pollution.  Since then, major changes have affected emissions from
vehicles, including additional controls on evaporative emissions, on- board diagnostics, and state
inspection and maintenance programs addressing evaporative emission controls.  Several
researchers have subsequently conducted air measurements in homes and in attached garages to
evaluate the effects on indoor air.

       Garage concentrations of benzene and other VOCs are generally much higher than either
indoor or outdoor air, and constitute one of the highest-concentration microenvironments to
which a person might typically be exposed. The garage also supplies contaminated air to the
home to which it is attached, and emits the rest. One recent study from Michigan found average
garage benzene concentrations of 36.6 ug/m3, with a standard deviation of 38.5 ug/m3, compared
to mean and standard deviation concentrations of 0.4 ug/m3 and 0.12 ug/m3 in ambient air.97  In
Alaska, where  fuel benzene levels tend to be very high and  homes built very airtight, garage
concentrations have been measured at even higher levels. One study measured average garage
benzene concentrations of 101 ug/m3, with a standard deviation of 38 ug/m3.98

       Other studies have studied the effect of garages or the sources within them on indoor air
quality. Most prominently, a group of Canadian investigators conducted source apportionment
of indoor non-methane hydrocarbons (NMHC) in 16 Ontario homes in the late 1990's." They
also assembled source profiles from hot soak and cold start emissions, which they used to
conduct source apportionment of total indoor air NMHC. All emissions samples and house
testing were conducted using the same 1993 model year vehicle. Overall, while the vehicle was
hot-soaking in  the garage over a four hour sampling period, between 9 and 71% of the NMHC
inside the house could be attributable to that vehicle's emissions.  Similarly, in the two hours
following a cold start event, between 13 and 85% of indoor NMHC could be attributed to the
vehicle cold start. Prior to the hot soak testing, average indoor benzene concentrations were 3.77
ug/m3, while during the hot soak, concentrations averaged 13.4 ug/m3.  In the garage,
concentrations averaged 121 ug/m3 during the cold start.  Prior to a cold start, indoor benzene
concentrations averaged 6.98 ug/m3, while for the two hours following cold start, concentrations
averaged 25.9 ug/m3. In the garage, concentrations averaged 422 ug/m3 over the two hours
following cold start.
                                          3-22

-------
       The study also conducted real-time monitoring of CO and total hydrocarbons (THC)
within the house and garage. Overall, concentrations of CO and THC were relatively constant
during hot-soaks, but following a cold start, indoor concentrations of CO and THC tended to rise
sharply, and fall over the next two hours. This study provides direct evidence that a high fraction
of indoor NMHC (or VOCs) are directly attributable to emission events occurring in the garage.

       Other studies have examined the influence of attached garages by comparing homes with
and without attached garages.  In another study from Alaska, 137 Anchorage homes underwent
indoor air quality monitoring for benzene and other VOCs.100 Homes with attached garages had
significantly higher concentrations of indoor benzene compared to homes without attached
garages (70.8 ug/m3 vs. 8.6 ug/m3). In addition, elevated benzene indoors was also associated
with the presence of a vehicle in the garage, fuel being opened in the  garage,  and the use of
forced-air heaters.

       In another Alaska study,  concentrations of benzene and toluene in indoor air were found
to be not significantly associated with their urinary biomarkers, but indoor concentrations were
associated with the number of gasoline-powered engines stored in the garage.101 In a recent
follow-up to the study, ventilation patterns in two homes were evaluated using perfluorocarbon
tracers and a multi-zone indoor air quality model.102 In the study, average garage concentrations
were consistently elevated relative to the home.  Furthermore, the study calculated the "virtual"
source strengths for benzene and toluene within the garage, and the garage was the only major
source of benzene within the home. Median garage source strengths for benzene ranged from
14-126mg/h.

       Several population-based surveys have also found evidence of the influence of attached
garages.  The National Human Exposure Assessment Survey (NHEXAS) Phase I pilot study in
Arizona was a representative exposure survey of the population.  It found that in non-smoking
homes with attached garages, distribution of toluene concentrations indoors was shifted
significantly higher in homes with attached garages.103 Homes with attached garages had median
toluene levels of 24 ug/m3, while homes without garages had median concentrations of 5 ug/m3.
The NHEXAS study in EPA Region 5 states was of similar design, but covering the states of the
upper Midwest. Using multivariate statistics, investigators found that VOCs  including benzene
were associated with the storage of gasoline-powered equipment in an attached garage.104

       In one study from New Jersey, investigators evaluated the indoor air effects of a vehicle
fueled with "M85" - an 85% methanol, 15% gasoline blend - parking in the garage of a single
home.105 Testing was undertaken with both normally-functioning and malfunctioning
evaporative emissions controls, as well as with the HVAC system on and off. Garage benzene
concentrations exceeded indoor concentration by approximately  10-fold. Furthermore, the room
adjacent to the garage had substantially higher concentrations than a room on the opposite side of
the house.  This study provides evidence that the garage is a major source of benzene inside the
house.

       EPA undertook an investigation of the effect of attached garages on indoor air under
various scenarios.106  This study  was undertaken to evaluate the magnitude of exposure
                                          3-23

-------
underestimation using the national-scale exposure modeling techniques discussed above. Using
a mass balance model, steady-state concentrations of benzene were calculated as a function of
the concentration of air in the garage, the concentration of outdoor air, and the fraction of house
air intake from a garage. Data were obtained from studies discussed above.  Overall, using in-
garage concentration data from Michigan, average indoor concentrations increase by
approximately 4.2 ug/m3, relative to concentrations estimated without an attached garage term.
Using data from Alaska, average indoor concentrations increase by 11.6 ug/m3,  and using New
Jersey data, by 9.2 ug/m3.  As noted above, the National Human Activity Pattern Survey
(NHAPS) estimates that the average person spends 16.68  hours per day indoors in a residence.
Taking that into account, overall modeled exposures would be expected to increase by at least
2.9 ug/m3, using the Michigan data.  These calculations imply that predicted exposures would
more than double if attached garages were treated systematically in a national exposure model.

       Proposed reductions in fuel benzene content, new  standards for cold temperature exhaust
emissions during vehicle starts, and reduced emissions from gas cans are all expected to
significantly reduce this major source of exposure.

3.1.3.4 Exposure and Concentrations in Parking Garages

       Relatively limited air quality data for parking garages is available in the  literature. The
following are results of air quality studies performed in parking garages, all of which indicate
that air toxics and criteria pollutants measured in these environments are significantly higher
than found in outdoor air.

       In November 1990, a study of microenvironments, partially funded by the US EPA,
evaluated the potential range in concentrations of selected air toxics.107 Ten parking garages,
along with gasoline stations and office buildings, were randomly chosen for sampling since they
were among the least studied of the potentially important  exposure microenvironments. The
principal air contaminants monitored were benzene, formaldehyde, and CO.  Additional
compounds included toluene, xylenes, 1,2-dichloroethane, chloroform, carbon tetrachloride,
perchloroethylene, 1,1,1-trichloroethane, 1,3-butadiene, and trichloroethylene.  The majority of
the compounds measured were significantly higher inside the garage compared to the ambient
sample. For example, the median 5-minute concentration of benzene was 67.1  ug/m3 in the
parking garage and 12.8 ug/m3 in ambient air.  CO was 11000 ppb in the parking garage and
2000 ppb in ambient air. The researchers identified elevated levels of selected air toxics in
parking garages and pointed out the potential contribution from cold starts at the end of the work
day.

       A more recent 2002 study was funded by The American Petroleum Institute to screen
"high-end" exposure microenvironments as required by section 21 l(b) of the Clean Air Act.108
The study included measurements at underground parking garages and surface parking lots in
several cities. Air toxics quantified included hydrocarbons (HCs), carbonyl compounds, BTEX,
total VOC, and CO. When sampling at parking lot exits, spikes in pollutant concentrations were
observed when vehicles accelerated out of the parking lot, while presumably prior to full catalyst
warm-up.  In underground garages, the levels of BTEX and other compounds of interest varied
                                          3-24

-------
with traffic level and reached concentrations that were significantly higher than ambient levels
outside the garage.

       A comparative study of indoor air quality in Hong Kong showed that the levels of CO,
NOx, and nonmethane hydrocarbons (NMHC) detected in a local park garage were the highest
among 13  other indoor sampling locations.109 The study did not specify the type or size of the
chosen parking garage, but indicated that it was located in an urban commercial area. High
indoor/outdoor ratios indicated that the air quality was mainly affected by indoor sources,
namely the vehicle exhaust. They also concluded that the pollution generated might cause health
hazards to the users and workers using such an environment.

       Another assessment of the air quality in indoor park garages was performed in Hong
Kong in August through December 2000.no Air samples were collected in two different garages
(an enclosed and semi-enclosed parking garage)  as well as outdoors (within 10m of each
parking garage) and analyzed for one hundred different C3-C12 VOCs.  Other compounds
measured included CO, CO2, PMi0,  and PM2.5. The CO levels in the enclosed garage were more
than in the semi-enclosed garage, and double the levels of the outdoor air.  The PMio and PM2.5
concentrations were also found to be higher in the parking garage environments than outdoors.
High mass fractions of aliphatic and aromatic compounds detected in the enclosed garage
showed that fuel evaporation and motor vehicular exhaust were the major contributors to the
VOCs. The total concentrations of NMHC in the enclosed and semi-enclosed garages ranged
from 580 to 4610 ug/m3 and 43.1 to  175 ug/m3, respectively. The mean concentration of NMHC
measured in the enclosed garage (1910 ug/m3) was about 17 times higher than in the semi-
enclosed garage (94.6  ug/m3),  and 3 times higher than measured at the outdoor sites. Not only
was the level of VOCs higher in the enclosed garage, but also the abundance of species
identified.  The most abundant species in similar ranking order for both garages was toluene, 2-
methylbutane,  w/p-xylenes, w-pentane, 2-methylpentane, w-hexane, and w-butane.  Other major
gasoline components such as benzene, xylenes, and C4-C7 saturated HCs were also very high in
the enclosed garage. The difference between the two sites could be associated with the
ventilation and location, since the occupancy rates and fleet mixes were similar. They also noted
that the absence of sunlight in the enclosed garage would result in a slower or negligible
photochemical depletion rate of unsaturated hydrocarbons, and consequently an increased
abundance of the species observed.

       In another study of multi-level  parking garages in an Athens urban area, CO levels were
characterized in autumn 1999.111 Samples were  collected at the exit sites (ramp where the flow
of vehicles was concentrated), the indoor site (first underground level where the majority of cars
parked), and immediately outside of each garage. Results indicate that CO levels varied
significantly over site, time, and day of measurement.  The peak 1-hour value at the indoor sites
ranged from 22.9 to 109.3 ppm.  At the indoor site, levels snowed little variation and remained
high over time. The peak 1-hour value at the exit sites ranged from 8.9 to 57.3 ppm. At the exit
sites, 15-minute maximum concentrations were 5-15 times higher than the maximum recorded
CO level immediately outside the garage. CO levels on Saturday were much lower than  a typical
weekday due to the reduced traffic, and weekday values were highest during the afternoon
sampling times (12:00-16:00 hour) corresponding with peak traffic volumes.
                                         3-25

-------
       In Mumbai, India, ambient levels of benzene were determined during different seasons at
several different locations, including two parking areas.112 Parameters of the parking areas at
Liberty Cinema and Natraj Cinema were not specified, but 24-hour geometric means of benzene
measured 117.4 and 74.2 ug/m3 during the summer, 94.5 and 75.4 ug/m3 during the monsoon,
and 148.0 and 703.0 ug/m3 during the winter seasons, respectively. These values were
considerably higher in comparison to less heavily trafficked residential locations.  The mean
benzene concentrations of four different residential locations ranged from 4.7 to 32.9 ug/m3, 1.9
to 33.5 ug/m3, and 4.7 to 18.8 ug/m3, respectively, for the summer, monsoon, and winter
seasons.  The high concentrations in parking areas were attributed to cold start-up emissions of
engines.

       A study in the UK of twelve underground parking garages identified high pollutant levels
of NOx, CO, CO2, BTEX, and PM.113 The parking garages selected covered a cross-section of
sizes (1 to 8 decks), ventilation system (natural and mechanical), designs (50 to 690 spaces), and
usages (business, shopping,  and/or residential). Monitoring sites were located inside and at the
exit of the parking garage.  The highest 15-minute average CO levels were measured at the exit
of parking garages, but a number of the parking garages had CO levels consistently higher inside
than at their exit.  The NO2 measurements  showed similar trends.  Weekday benzene
concentration measurements averaged over one hour inside the parking garage and at the exit
ranged from 60 to 870 ug/m3 and 10 to 350 ug/m3, respectively.

       In Madrid, Spain, atmospheric pollution produced by vehicles in  parking garages was
studied.114  Two parking garages of different design were chosen for measurements of PMio,
lead, 12 PAHs, and CO. In both garages, CO, NO, TSP, and lead concentrations directly
correlated with vehicle traffic flow into and out of the garage. Also, higher values were observed
on the weekdays than during the weekend, for CO, NO, PAHs, and TSP  in both garages. For
example, in one garage, the average daily TSP concentrations were 78-122 ug/m3 on the
weekdays versus 39 ug/m3 on the weekend, which was similar to outdoor city average
measurement (50 ug/m3).  The researchers conclude that maximum concentrations for NO were
observed during maximum parking garage exits and therefore due to vehicle cold-starts. They
also conclude that the mechanical ventilation used in both garages was not sufficient to disperse
the pollutants emitted by the vehicles.

3.1.3.5 Exposure and Concentrations at  Service Stations

       Although there is relatively limited air quality data for service stations available currently
in the literature, the general  consensus is that exposures to air toxics at service stations
significantly exceed ambient background levels. The studies below measure personal exposures
and concentrations during refueling either inside or outside of vehicles throughout the United
States.  Several studies conducted outside of the United States chronicle  similar results but are
not presented here due to differences in fuels and control technologies.

       The Total Exposure Assessment Methodology (TEAM)  Study was planned in 1979 and
completed in  1985.115  The goal of this study was to develop methods to  measure individual total
exposure (through air, food and water) and resulting body burden to toxic and carcinogenic
chemicals, and then to apply these methods with a probability-based sampling framework to
                                          3-26

-------
estimate the exposures and body burdens of urban populations in several U.S. cities. The study
measured personal exposures of 600 people to a number of air toxics. The subjects were selected
to represent residents of cities in New Jersey, North Carolina, North Dakota, and California.
One of the major findings was that pumping gas as well as exposure to auto exhaust was a
specific and major source of benzene exposure. People who filled their tanks with gasoline had
twice as much benzene in their breath as people who did not. Estimated concentrations at the
breathing zone could then exceed  1000 ug/m3 (100 times the ambient level), based on the median
breath benzene value measured (n=67) for those who had worked at or been in a service station
during the past 24 hours. But since then, implementation of fuel controls and Stage II vapor
recovery have changed emission and concentration levels as discussed  in Section 3.1.1.

       In March 1990, another study randomly sampled 100 self-service filling stations
throughout Southern California along with samples at 10 parking garages and 10 offices nearby
those garages.116 The study took five-minute samples of 13 motor vehicle air pollutants (CO,
formaldehyde, and VOCs) in each microenvironment and in the ambient environment.  The
median benzene concentration measured at the  service stations was 28.8 ug/m3 with the
maximum reported value of 323 ug/m3. The median benzene concentration in ambient air was
significantly lower at 12.8 ug/m3.

       A 1993 National Institute for Occupational Safety and Health (NIOSH) study assessed
benzene and MTBE  concentrations and service station attendant exposures at service stations
with and without Stage II vapor recovery in Cincinnati, Phoenix, and Los Angeles.117 The mean
(and maximum) benzene exposure measurements were 96 (927), 160 (1662), and 192 (607)
ug/m3, respectively.  The study found that Stage II vapor recovery did not significantly reduce
exposure to benzene during refueling. However, the efficiency of Stage II vapor recovery has
improved over the years. Northeast States for Coordinated Air Use Management (NESCAUM)
has suggested that Stage II vapor recovery systems are greater than 90% effective at capturing
MTBE and benzene  vapors during refueling.118 These systems would therefore be expected to
reduce exposure beyond that shown in the NIOSH exposure assessment.

       In March 1996 to July 1997, concentrations of MTBE, benzene, and toluene were
determined inside automobile cabins during fueling.119 Air samples were collected at service
stations in New Jersey, and the  mean benzene in-cabin concentration was 54.3 ug/m3 (n=46).
The background concentration at the pump island measured 9.6 ug/m3 (n=36).  The highest in-
cabin concentrations for all three pollutants occurred in a car that had a malfunctioning vapor
recovery system and in a series of cars sampled on an unusually warm, calm winter day when the
fuel volatility was high, the evaporation maximal, and the wind dispersion minimal. The in-
cabin concentrations were also typically higher when the car window was opened during the
entire fueling process.

       Most recently, as discussed in the section on in-vehicle and parking garage exposure and
concentrations, a screening study of "high-end" exposure microenvironments was performed by
the American Petroleum Institute.56  The study  included several vehicle-related
microenvironments in Houston and Atlanta during summer 2002.  Among the various
microenvironments examined, the highest short-term concentrations occurred during refueling.
The in-vehicle average concentration of benzene measured during  refueling was 46.0 ug/m3.
                                         3-27

-------
3.1.3.6 Occupational Exposure

       Occupational settings can be considered a microenvironment in which exposure to
benzene and other air toxics can occur. Occupational exposures to benzene from mobile sources
or fuels can be several orders of magnitude greater than typical exposures in the non-
occupationally exposed population. Several key occupational groups are discussed below.

       Occupations that involve fuel distribution,  storage, and tank remediation lead to elevated
exposure to mobile-source related air toxics.  Researchers published a review of benzene and
total hydrocarbon exposures in the downstream petroleum industry, including exposure data
from the past two decades among workers in the following categories: refinery, pipeline, marine,
rail, bulk terminals, tank truck drivers, service stations, underground storage tanks, tank cleaning,
and site remediation.120 The studies reviewed indicate that benzene exposure can range from <1
to more than 10 mg/m3, which is approximately three orders of magnitude higher than typical
non-occupational exposures (although there are occurrences of high benzene exposures in non-
occupational settings as well). This review is relevant because of the potential for mobile source
benzene emission reductions to reduce their exposures as well.  This statement is echoed by
researchers in the occupational literature.121

       Handheld and non-handheld equipment operators are also exposed to elevated
concentrations of fuel-related air toxics. Several studies were conducted in work categories
employing small engine equipment, such as lawn and garden workers, workers in
construction/demolition, and others. Many of these occupations require the use of personal
protective equipment to prevent high exposures to carbon monoxide or other species. At present,
there are no representative samples of exposures among these categories. Non-occupational
exposures from these equipment types may also be important contributors to overall  exposure.
EPA recently conducted a study of occupational exposures among lawn and garden workers
using riding tractors, walk-behind lawn mowers, string trimmers, and chainsaws.122 Results
demonstrated that equipment operators can experience highly variable exposures, with short-
term personal concentrations of CO and PM2.5 ranging over two orders of magnitude. Air toxics
data will be available later this year. This  study illustrated the role of operator's activity in
affecting exposure levels to fuel-related air toxics.

       Another study provides some insight into the possible range of benzene exposures in
workers who operate gasoline-powered engines, particularly those with 2-stroke engine cycles.123
A study of snowmobile rider exposures in  Sweden found benzene concentrations ranging from
under 10  ug/m3 to 2.5 mg/m3, a range  of at least two orders of magnitude. Exposures measured
on riders  on the back of the vehicle ranged from 0.7-0.8 mg/m3.  These measurements illustrate
the potential for relatively high exposures when operating 2-stroke equipment, as used in this
study.  Yellowstone National Park commissioned a study in 2002 to examine occupational
exposures of park employees to benzene, other VOCs, PMio, and CO.124 Work shift benzene
concentrations at a snowmobile entry  gate 176.7 ug/m3, while snowmobile-bound mobile patrol
officers' exposure concentrations averaged 137.20 ug/m3. The highest observed work  shift
concentration in the study was 514.1 ug/m3.  At major sites of tourist interest where
snowmobiles parked, such as the Old Faithful geyser, concentrations averaged 41.3 to 48.8
                                          3-28

-------
ug/m3. 15-minute "peak" samples of workers' personal air ranged from 46.8 ug/m3 to 842.8
ug/m3. This study provides an indication of the variability of occupational benzene exposure
concentrations with time, and highlights the potential for elevated work shift exposures over
several hours.

       A preliminary report published by the Northeast States for Coordinated Air Use
Management further illustrates the occupational impact of nonroad heavy-duty diesel
equipment.125  In-cabin and work site perimeter measurements were collected for diesel
equipment emissions from the agricultural, construction (building and roadway), and lumber
industries in the Northeast. Initial results indicate that PM2 5 concentrations  were 1-16 times
greater than the average ambient concentrations in each monitoring area. In-cabin exposures to
PM2.5 for operators ranged from 2 ug/m3 to over 660 ug/m3.  Additionally, measured
concentrations of acetaldehyde, benzene, and formaldehyde were found to be significantly
elevated, although concentrations were not presented.

       In addition, some occupations require that workers spend considerable time in vehicles,
which increases the time they spend in a higher-concentration microenvironment. In-vehicle
concentrations are discussed in  Section 3.1.3.2.1 above.

3.2    Modeled Air Quality, Exposures, and Risks for Air Toxics

3.2.1   National-Scale Modeled Air Quality, Exposure, and Risk for Air Toxics

       EPA assesses human health impacts from  outdoor, inhalation, chronic exposures to air
toxics in the National-Scale Air Toxics Assessment. It assesses lifetime risks assuming
continuous exposure to levels of air toxics estimated for a particular point in time.  The most
recent National-Scale Air Toxics Assessment was done for the year 1999.126  It has four steps:

       1) Compiling a national  emissions inventory of air toxics emissions from outdoor
       sources. The 1999 National Emissions Inventory is the underlying basis for the
       emissions information in the 1999 assessment.
       2) Estimating ambient concentrations based on emissions as input to  an air dispersion
       model (the Assessment System for Population Exposure Nationwide, or ASPEN
       model).127
       3) Estimating population exposures based on a screening-level inhalation exposure model
       (Hazardous Air Pollutant Exposure Model, version 5, or HAPEM5) and the estimated
                                         _                                      1 oo
       ambient concentrations (from the ASPEN model) as input to the exposure model.
       4) Characterizing 1999 potential public health risks due to inhalation of air toxics. This
       includes cancer and noncancer effects, using available information on air toxics health
       effects, current EPA risk assessment and risk characterization guidelines, and estimated
       population exposures.129

       For this rule, we have conducted air quality, exposure and risk modeling for the years
2015, 2020, and 2030, using the same tools and methods as the 1999 National-Scale Air Toxics
Assessment. Thus our results are comparable to the 1999 Assessment, other than in the few
situations in which risk values were re-computed  resulting from stationary source inventory
                                          3-29

-------
errors which were determined to impact a tract or county-level risk estimate. For the reference
case, which includes all control programs currently planned by EPA in regulations, we modeled
all the pollutants in Table 2.2-1. These pollutants
    •   Are on EPA's list of hazardous air pollutants in Section 112 of the Clean Air Act
    •   Are emitted by mobile sources
    •   Are included in the National Emissions Inventory
    •   Are included in the 1999 NATA
Note that the modeling did not include diesel PM and diesel exhaust organic gases. For the fuel
benzene control case, we modeled the following pollutants: benzene, 1,3-butadiene,
formaldehyde, acetaldehyde, and acrolein.  This modeling work is discussed in more detail in an
EPA technical report, "National Scale Modeling of Air Toxics for the Mobile Source Air Toxics
Rule; Technical Support Document,"  Report Number EPA-454/R-06-002. EPA has previously
done future year projections of the mobile source contribution to air toxics concentrations,
exposure, and risk for selected air toxics,130'131'132'133 but has never done a comprehensive
assessment that includes projections for all mobile source air toxics, as well as the stationary
source contribution for those pollutants.

       The National-Scale Air Toxics Assessment modeling framework has a number of
limitations which prevent its use as the sole basis for setting regulatory standards.  Even so, this
modeling framework is very useful in identifying air toxic pollutants and sources of greatest
concern, setting regulatory priorities, and informing the decision making process. Among the
significant limitations of the framework is that it cannot be used to identify "hot spots," such as
areas in immediate proximity to major roads, where the air concentration,  exposure and/or risk
might be significantly higher within a census tract or county. This limitation may result in
underestimates of exposure due to the design of ASPEN.  In addition, this kind of modeling
assessment cannot address the kinds of questions an epidemiology study might allow, such as the
relationship between asthma or cancer risk, and proximity of residences to point sources,
roadways and other  sources of air toxics emissions.  The framework also does not account for
risk from sources of air toxics originating indoors, such as stoves or out-gassing from building
materials or evaporative  benzene emissions from cars in attached garages. The ASPEN model
performs well for some pollutants, but has also been shown to systematically underestimate
pollutant concentrations  relative to measured levels for certain pollutants such as metals. The
cancer unit risk estimates for many pollutants are "upper bound," meaning they probably lead to
overestimates of risk. It should be noted, however, that the unit risk estimate for benzene is a
maximum likelihood estimate, which  is a best scientific estimate.  The above limitations are
discussed in detail in Section 3.2.1.4.

       Another tool which has been used by EPA to assess distributions of concentrations of air
toxics at the national scale is the Community Multiscale Air Quality Model (CMAQ).134 In
general, predicted concentrations of air toxics from CMAQ were within a factor of 2  of measured
values, with a tendency to underpredict measured ambient concentrations.135 CMAQ
underpredicts monitored benzene levels more than ASPEN, although it better calculates the
contribution of transport, and more accurately model the effect of benzene decay. CMAQ has
sophisticated photochemistry, but does not yet have the spatial resolution of dispersion models
such as ASPEN, and thus accounts for less of the total variability in levels of air toxics with
localized effects, such as benzene.136  Finally, CMAQ is requires more computational resources,
                                          3-30

-------
which makes it more difficult to use for evaluating trends in a large number of air toxics over
many years or impacts of control scenarios.

       Details of the methods used, and presentation of key results are discussed in the
following sections. Results do not account for significant sources of inhalation exposure, such as
benzene emissions from sources in attached garages (such as vehicles, snowblowers,
lawnmowers and gas cans). Furthermore, the modeling underestimates the contribution of
hydrocarbon and particulate matter emissions at cold temperatures, based on results of recent test
data discussed in Chapter 2.

3.2.1.1 Air Quality Modeling

3.2.1.1.1       Methods

       Prior to performing air quality modeling on the projected emissions, the emissions from
the stationary and mobile inventories (discussed in Chapter 2) are processed in the Emissions
Modeling System for Hazardous Air Pollutants (EMS-HAP) Version 3 to create the emissions
                     _                                     1 Q*7
input files used by ASPEN to calculate air quality concentrations.   In addition to projecting
stationary and area source emissions to future years for some source categories, EMS-HAP
spatially allocates emissions inventoried at the county level to the census tract level, and
temporally allocates them to eight three-hour time periods throughout the day. Once the
emissions were processed, they were input into ASPEN to calculate air quality concentrations.
In addition to the emissions, ASPEN uses meteorological parameters, and census tract centroid
locations for concentration calculations. ASPEN estimates do not account for day-of-week or
seasonal variations in emissions. The ASPEN model takes into account important determinants
of pollutant concentrations, such as: rate of release, location of release, the height from which the
pollutants are released, wind speeds and directions from the meteorological  stations nearest to
the release, breakdown of the pollutants in the atmosphere after being released (i.e., reactive
decay), settling of pollutants out of the atmosphere (i.e., deposition), and transformation of one
pollutant into another.  The model first estimates concentrations at receptors arranged in rings
around emission sources up to 50 km away.  The model then interpolates concentrations to
census tract centroids.  For 1999 NAT A, meteorological conditions in 1999 and 2000 census
tract data were used.

       In using ASPEN to estimate concentrations for emissions projected to years 2015, 2020,
and 2030, the same meteorology and census tract locations were used, as for the 1999 NAT A.
Details of how ASPEN processed emissions  data are provided in the technical document,
"National-Scale Modeling of Mobile Source Air Toxic Emissions, Air Quality, Exposure and
Risk for the Mobile Source Air Toxics Rule."  ASPEN only accounts for sources within a 50
kilometer radius of each source when calculating ambient concentrations. Thus, the contribution
to ambient levels of air toxics from sources further away than 50 kilometers, as well as the
contribution of uninventoried sources is addressed through the addition of a "background"
term.138 Mobile source pollutants which include a background component are 1,3-butadiene,
acetaldehyde, benzene, formaldehyde, and xylenes. Each of the three projection years used the
same 1999-based background. However, background levels are likely to change with emissions.
                                          3-31

-------
Thus, a sensitivity analysis was done to evaluate the potential impact of not changing the
background concentration (see Section 3.2.1.4).

       We estimated the contributions to ambient concentrations for the following source
sectors: major, area and other, onroad gasoline, onroad diesel, nonroad gasoline, remaining
nonroad (diesel and compressed natural gas), and background.A

3.2.1.1.2      Air Quality Trends for Air Toxics

       Table 3.2-1 summarizes nationwide mean census tract ambient concentrations of mobile
source air toxics in 1999 and projection years for the following source sectors: major sources,
area and other sources, highway vehicles, nonroad sources, and background. The behavior of
benzene is typical of the projected trends.  Over 90% of the mobile source contribution to
ambient benzene levels is attributable to gasoline vehicles and engines. Figure 3.2-1 depicts the
trend in nationwide average census tract concentrations of benzene over this time  period. The
mobile source contribution to ambient benzene concentrations is projected to decrease 60% by
2015, with a decrease in ambient benzene concentration from all sources of over 30%. Summary
tables providing data by State, and for reformulated and non-reformulated gasoline areas, can be
found in the docket for this rule.  Due to greater population and vehicle activity, the average
ambient benzene concentration in 1999 is much higher for counties in reformulated gasoline
areas than nonreformulated gasoline areas - about  1.9 jig/m3 versus 1.1 |ig/m3. However the
percent reduction in average ambient concentration is similar regardless of fuel type - 29% for
non-reformulated gasoline counties versus 35% for reformulated gasoline counties.
A Major and "area and other" are stationary source emission sectors. Major sources, as defined by the Clean Air
Act, are those stationary facilities that emit or have the potential to emit 10 tons of any one toxic air pollutant or 25
tons of more than one toxic air pollutant per year.  Area and other sources include sources that generally have
smaller emissions on an individual basis than "major sources" and are often too small or ubiquitous in nature to be
inventoried as individual sources. "Area sources" include facilities that have air toxics emissions below the major
source threshold as defined in the air toxics sections of the Clean Air Act and thus emit less than 10 tons of a single
toxic air pollutant or less than 25 tons of multiple toxic air pollutants in any one year. Area sources include smaller
facilities, such as dry cleaners. "Other sources" include sources such as wildfires and prescribed burnings that may
be more appropriately addressed by other programs rather than through regulations developed under certain air
toxics provisions (section 112 or 129) in the Clean Air Act. For example, wildfires and prescribed burning are being
addressed through the burning policy agreed to by the Interim Federal Wildland Policy.  "Background" includes
emissions from transport and uninventoried sources.


                                             3-32

-------
O
f>
o
«s

•o
 a
 «
  Cv
o
«s
o
 o
H
 u

 3
 O

 o

 a
 u
 =
 o
U
•*j
 =
—
3
•T
E
o
ISI
C
O
S
i
to
15
3
c
s
o


''E
o
1
e concent
8
*n
cu
O)







total (including
background)
onroad

o
c
o

to
08
(0
0)
S3
0
'(3*
total (including
background)
nonroad
o
c
o

to
o
08

0
CD

D)
•D
c >-v
O ^c
™ c
D)
Pollutant
CXI
cp
o
CO
CD
CO
cp
CXI
CO
o
CO
CD
CXI
cp
CD
CD
LSD
*~
CXI
cp
CD
LSD
O
CXI
ro
cp
CD
CD
cxi
o
CD
CD
CXI
cp
CD
CD
CO
LSD
CXI
cp
1 —
CD
CD
CXI
LSD
CXI
cp
CD
CD
O
CXI
ro
cp
CO
CD
CD
CD
*~
CXI
o
o

1,3-Butadiene
o
i —
LSD
r— .
CXI
ro
CXI
cp
O
CO
r- •
CO
CD
cp
CD
CD
CXI
CXI
cp
CD
ro
CXI
cp
CO
CD
CD
o
o
CO
CO
ro
CXI
CO
o
CD
CD
CXI
o
o
LSD
CD
CXI
cp
o
CO
CD
o
ro
CXI
cp
o
CO
CXI
o
o
o
o
o
o
o
2,2,4-Trimethylpentane
o
CO
CO
LSD
CD
CD
O
O
*~
cp
1 —
CO
CO
CXI
CXI
cp
1 —
CD
O
LSD
CXI
cp
O
CD
CD
CD
cxi
O
O
LU
CD
CXI
O
CO
CD
O
CXI
1 —
LSD
CD
CD
CXI
cp
1 —
O
CD
LSD
CXI
cp
CD
CD
CXI
O
ro
CXI

Acetaldehyde
CXI
cp
CO
CD
CXI
cp
CD
CO
CD
r- •
*~
cp
CD
CXI
CD
*~
CXI
cp
r- •
CD
cxi
3.5272E-03
9
CD
LSD
CD
2.2393E-02
CXI
cp
ro
LSD
CD
CO
LSD
CXI
cp
CXI
o
ro
CD
CXI
ro
cp
CXI
ro
O
O
O
O
O
O
O
Acrolein
o
o
ro
CD
CXI
cp
O
ro
LSD
O
CD
cp
LSD
CXI
CXI
O
CD
LSD
r— .
CO
CXI
cp
CXI
LSD
CD
LSD
O
O
+
I_U
CD
ro
O
CD
CD
LSD
O
r- •
LSD
CD
CD
O
1 —
ro
ro
CD
*~
CXI
cp
CD
CO
ro
CXI
CXI
O
CO
CO
ro
CD

Benzene
ro
cp
LU
CO
LSD
cp
CXI
O
CO
CD
cp
ro
ro
*~
cp
CD
CD
ro
cp
CD
ro
O
ro
cp
LU
CXI
ro
LSD
cp
LU
CD
LSD
CD
CD
CD
LSD
cp
CD
ro
CD
cp
CD
CD
CXI
LSD
^
o
ro
O
CXI
CXI
CO
O
o
o
o
o
o
o
Chromium III
o
cn
o
LD
LD
o
o
ro
cn
*~
o
CXI
r~
ro
O
CO
o
CXI
1 —
cxi
o
CD
LD
ro
9
CD
o
LD
o
CO
O
cn
ro
LD
o
o
1 —
o
CXI
o
o
CO
cn
<—
o
r~
o
<—
o
o
o
o
o
o
o
Chromium VI
o
o
LD
LD
cxi
CXI
o
LD
CO
LD
"*"
CD
cn
i —
CO
o
CXI
CD
LU
CXI
CXI
CXI
o
ro
LD
CXI
CD
LD
ro
ro
cn
o
o
CXI
CXI
CD
CXI
CXI
O
cn
i —
LD
CO
CXI
O
O
LD
ro
CO
<—
O
O
O
O
O
O
O
Ethyl Benzene
o
o
cn
CXI
CXI
CXI
CD
ro
CXI
r~
^~
CD
LD
CD
o
<—
CXI
CD
ro
ro
CXI
CO
cn
4.9752E-02
o
o
r~
cn
cn
LD
2.4205E-01
o
i —
CD
1 —
CD
^
CXI
cp
LSD
LSD
CD
r- •
CO
CXI
cp
CD
CD
CD
ro
o
CXI
CD

Formaldehyde
o
LSD
LSD
CXI
CXI
cp
ro
CXI
CD
ro
cp
O
ro
O
*~
O
LU
CXI
ro
LSD
CXI
cp
CO
CD
LSD
O
1 —
CXI
CXI
CO
CXI
cp
LU
CO
CO
LSD
O
CD
LSD
CD
LSD
CXI
O
LSD
^
CXI
cp
CO
CO
CO
CD
CD
O
O
O
O
O
O
O
H exane
ro
cp
ro
CO
CXI
ro
CD
LSD
cp
CXI
O
CD
*~
cp
ro
ro
CD
CD
*~
ro
cp
O
CO
CD
cxi
3.2250E-03
5.0455E-03
LSD
cp
r- •
CD
LSD
cp
O
CD
O
*~
ro
cp
LSD
CXI
CXI
ro
cp
CXI
CD
o
r- •
CXI
O
O
O
O
O
O
O
Manqanes e
o
CXI
CD
CD
CD
CXI
cp
CD
CD
CD
CD
CXI
cp
r- •
CD
CXI
CD
CD
CXI
cp
CD
r- •
CD
CO
LSD
CXI
cp
CXI
9
I_U
CXI
CD
LSD
CD
r-^
CXI
cp
LSD
LSD
CD
CO
O
CXI
O
CD
CXI
cp
CD
ro
CD
CXI
cp
CD
CXI
ro
*~
o
o
o
o
o
O
O
CQ
CXI
cp
LU
O
CD
cp
LSD
CO
LSD
^
cp
LSD
CD
CD
1 —
!"-
CXI
cp
LU
CD
CXI
LSD
LSD
CD
cp
LU
CD
CO
CD
ro
CXI
cp
LSD
CD
O
cp
CO
O
CXI
cp
CXI
*~
CXI
cp
CD
LSD
^
CD
cp
LSD
*
O
O
o
o
o
o
o
Naphthalene
ro
cp
O
ro
LSD
cxi
cp
CD
CXI
CXI
*~
cp
LSD
LSD
LSD
CXI
*~
cp
LSD
CD
CD
cp
O
CD
CD
CO
CO
2.3837E-03
cp
CD
CD
r- •
o
LSD
cp
CD
CD
CXI
CO
ro
cp
LSD
CD
*~
cp
CXI
CD
r- •
i —
O
O
O
O
O
O
O

CXI
cp
CO
LSD
CD
ro
cxi
cp
LSD
O
O
1 —
cp
LSD
CXI
CD
CD
CD
CXI
cp
LU
CO
CD
CD
CO
ro
cp
LU
o
CD
1 —
ro
CXI
cp
8
CD
ro
cxi
cp
CXI
CD
CD
CO
cp
LSD
*~
CXI
cp
CD
CD
*~
cp
CXI
CD
"*"
o
o
o
o
o
o
o
o
D_
o
CXI
LD
CXI
CXI
O
cn
CD
CXI
o
LD
LD
o
LD
CD
CXI
O
CXI
cn
ro
cxi
9.3087E-03
9
r~
o
cxi
CXI
o
CXI
o
LD
O
O
LD
CXI
CD
CD
<—
CXI
o
1 —
O
CXI
CXI
o
o
o
<—
o
o
o
o
o
o
o
Propi on aldehyde
CXI
o
LU
CO
CD
CXI
CD
O
CD
O
"*"
CD
O
CXI
CO
CXI
CD
CXI
cn
CO
CXI
o
LU
o
CXI
o
o
ro
CXI
CD
CD
LD
cn
CXI
ro
o
LU
LD
CD
LD
CO
CD
CXI
CD
CD
ro
i —
CD
CXI
CXI
O
o
LD
<—
CXI
o
CXI
LD
CXI
O
O
O
o
o
o
o
I
C/l
o
o
CD
o
CXI
CD
r- •
*~
cp
CD
LSD
LSD
1.0459E+00
1.4319E-01
o
o
LU
cn
ro
cn
cn
cxi
3.0633E-01
8
cn
o
CO

o
i —
ro
O
CO
o
ro
cn
CXI
o
CXI
o
o
o
o
o
o
o
Toluene
1.5538E+00
o
ro
O
cn
cn
*~
o
CD
CXI
CXI
ro
o
o
cn
cn
i —
i —
CXI
o
^.
o
CXI
CXI
CO
o
o
+
LU
o
o
CXI
CXI
cxi
o
^.
CO
LD
CD
ro
8
o
o

o
CXI
^J-
co
LSD
CXI
cp
CD
CO
CD
CD
O
O
O
O
r— .

ifi
1
X

-------
o
fO
o
«s

•a
 e
 «
  Cv
o
«s
o
a\
a\
a\
 o
H
 u

 S
 O
—

S
 o
 o

 e



 •_

"a

 u
 a
 o
U

"a
 
-------
           Figure 3.2-1. Nationwide Average Benzene Concentration, 1999-2030
        1.60
        1.40
        0.00
                  1999
                   2015
2020
2030
                                        Year
3.2.1.1.3
Distributions of Air Toxic Concentrations across the U. S.
       Table 3.2-2 gives the distribution of census tract concentrations, summed across all
source sectors and background, for mobile source air toxics across the nation in 2020.
Distributions for other years are similar. Summary tables providing distributions for other years,
as well as distributions by State and for reformulated and non-reformulated gasoline areas, can
be found in the docket for this rule.  From this table, it can be seen that 95th percentile of average
census tract concentrations for mobile-source dominated pollutants such as benzene and 1,3-
butadiene are typically two to five times higher than the median of census tract concentrations,
even though mobile source emissions are widely dispersed. For pollutants with large major
source contributions (e.g., manganese), the 95th percentile of census tract averages can be much
higher than the median. In addition, average census tract concentrations can span one to several
orders of magnitude.  Thus, there is considerable variation in average concentrations across the
U.S.

       Figure 3.2-2 depicts the geographic distribution of county median concentrations of
benzene in 2020.  Relatively high levels are seen in the Northeast, Southern California, Florida,
parts of Texas, and the Great Lakes Region, where there is  high population density and thus high
vehicle and nonroad equipment activity. Relatively high levels are also seen in the Pacific
Northwest, parts of Alaska, and the upper Great Lakes region. Analysis of fuel survey data
indicate higher than average fuel benzene levels in these areas.  Higher benzene levels in Idaho
are not due to fuel benzene levels, but are primarily due to wildfire emission estimates, which
were determined to be an error in the 1999 National Emissions Inventory and the subsequent
projections.
                                           3-35

-------
 Table 3.2-2. National Distribution of Census Tract Concentrations for Mobile Source Air
                                     Toxics in 2020

Pollutant
1 ,3-Butadiene
2,2,4-Trimethylpentane
Acetaldehyde
Acrolein
Benzene
Chromium III
Chromium VI
Ethyl Benzene
Formaldehyde
Hexane
MTBE
Manganese
Naphthalene
Nickel
POM
Propionaldehyde
Styrene
Toluene
Xylenes
2020 concentration (ng m'3) distribution
5th
percentile
2.24E-03
2.43E-02
5.33E-01
4.41 E-03
3.25E-01
7.72E-06
2.45E-06
1 .50E-02
3.99E-01
2.75E-02
2.16E-03
1 .54E-05
2.98E-03
1 .41 E-05
1 .72E-03
9.97E-03
2.06E-03
1.17E-01
2.44E-01
10th
percentile
4.29E-03
4.30E-02
5.61 E-01
7.44E-03
3.87E-01
1 .99E-05
6.56E-06
2.73E-02
5.16E-01
5.18E-02
5.19E-03
4.66E-05
6.06E-03
3.91 E-05
2.95E-03
1 .68E-02
3.95E-03
2.13E-01
3.04E-01
25th
percentile
2.70E-02
1.05E-01
6.51 E-01
1 .64E-02
5.67E-01
8.12E-05
2.97E-05
7.64E-02
7.80E-01
1 .68E-01
1 .61 E-02
2.06E-04
1 .91 E-02
1 .69E-04
5.70E-03
3.73E-02
9.58E-03
5.64E-01
5.44E-01
Median
7.65E-02
2.32E-01
8.31 E-01
3.44E-02
7.97E-01
3.14E-04
1.16E-04
1.78E-01
1.10E+00
4.29E-01
4. 96 E-02
8.72E-04
4. 63 E-02
6.80E-04
1.19E-02
8.19E-02
2. 11 E-02
1 .28E+00
1 .04E+00
75th
percentile
1.16E-01
4.02E-01
1 .07E+00
7. 11 E-02
1 .06E+00
1 .03E-03
3.30E-04
3.09E-01
1 .44E+00
8.21 E-01
1 .95E-01
3.56E-03
9.19E-02
2. 11 E-03
2.06E-02
1 .50E-01
4.45E-02
2.29E+00
1.74E+00
90th
percentile
1.71 E-01
6.39E-01
1 .46E+00
1.55E-01
1 .48E+00
2.66E-03
9.80E-04
5.00E-01
2.03E+00
1.65E+00
5.32E-01
1 .53 E-02
1.81 E-01
5.13E-03
3.59E-02
2.65E-01
9.45E-02
4.11E+00
3.15E+00
95th
percentile
2.85E-01
8.15E-01
1.78E+00
2.49E-01
1.84E+00
5.06E-03
1 .63E-03
7.17E-01
2.53E+00
2.81 E+00
7.62E-01
2.26E-02
2.80E-01
8.75E-03
5.76E-02
3.53E-01
1 .62E-01
5.88E+00
4.90E+00
       Similar benzene median county concentration maps for 1999, 2015, and 2030 can be
found in the docket for this rule, along with maps for other mobile source air toxics and tables of
concentration distributions.

3.2.1.1.4     Impacts of Proposed Fuel Benzene Controls on Ambient Concentrations

       The fuel benzene standard proposed in this rule will substantially reduce ambient
concentrations of benzene across the United States.  Table 3.2-3 shows that in 2015, 2020, and
2030, the highway vehicle portion of ambient concentrations will be reduced on average 8 to 9%
across the U.S., the nonroad equipment contribution will be reduced about 7%, and the area
source contribution about 4%.  The reduction for area sources is due to the impacts of fuel
benzene control on gasoline distribution emissions.  Reductions in non- Federal reformulated
gasoline areas (i.e., conventional gasoline areas) are even larger.  It should be noted that the
estimated total reductions are probably significantly underestimated, since we could not account
for the impacts of controls on background levels, which includes transport of emissions from
these sources.  The fuel benzene control proposed does not significantly affect ambient
concentrations of other air toxics.  Figure 3.2-3 presents the distribution of percent reductions in
median  ambient benzene concentrations for U. S. counties with the proposed fuel control in
2020. Summary tables providing data by State, as well as maps of benzene concentrations with
fuel controls and percent reductions with controls, can be found in the docket for the rule.
                                          3-36

-------
Similar data are also available for 1,3-butadiene, formaldehyde and acetaldehyde, even though
concentrations were not significantly affected.

    Figure 3.2-2.  Geographic Distribution of County Median Concentrations (ug/m3) of
                                      Benzene in 2020
                                                                             0.063 - 0.338

                                                                             0.339 - 0.567

                                                                             0.568 - 0.880

                                                                             0.881 - 1.316

                                                                             1.317-2.114

                                                                             2.115-4.929
                                            3-37

-------
Table 3.2-3.  Contributions of Source Sectors to Nationwide Average Census Tract Concentrations of Benzene, with and
                         without Proposed Fuel Benzene Standard, 2015, 2020, and 2030


Reference
Control
% Difference

2015 annual average concentrations (
major
0.02
0.02
-1

area&
other
0.19
0.18
-4

highway
vehicles
0.23
0.20
-9

nonroad
0.09
0.08
-7

ngm")
total
(including
background)
0.91
0.88
-4

Average Nationwide Difference in Ambient Benzene Concentration - Non RFC Areas

Reference
Control
% Difference


0.01
0.01
-1


0.15
0.15
-4


0.18
0.15
-14


0.06
0.06
-12


0.77
0.73
-5

Average Nationwide Difference in Ambient Benzene Concentration - RFC Areas

Reference
Control
% Difference

0.02
0.02
-1

0.25
0.24
-3

0.31
0.30
-4

0.14
0.13
-3

1.18
1.15
-2
2020 annual average concentrations (jig m_j
major
0.02
0.02
-1



0.01
0.01
-1



0.03
0.03
-1
area&
other
0.20
0.19
-4



0.16
0.15
-4



0.26
0.25
-3
highway
vehicles
0.20
0.18
-9



0.16
0.14
-13



0.28
0.27
-4
nonroad
0.09
0.09
-7



0.07
0.06
-12



0.14
0.14
-3
total
(including
background)
0.90
0.87
-4



0.76
0.72
-5



1.16
1.14
-2
2030 annual average concentrations (jig m_j
major
0.02
0.02
-1



0.01
0.01
-1



0.03
0.03
-1
areaS
other
0.20
0.19
-4



0.16
0.15
-4



0.26
0.25
-3
highway
vehicles
0.21
0.20
-8



0.16
0.14
-13



0.31
0.29
-4
nonroad
0.10
0.10
-7



0.07
0.06
-13



0.16
0.15
-3
total
(including
background)
0.92
0.89
-4



0.77
0.73
-5



1.20
1.18
-2
                                                   J-J

-------
       Figure 3.2-3. Distribution of Percent Reductions in Median Ambient Benzene
         Concentrations, 2020, for U. S. Counties with the Proposed Fuel Control
                                                                    Percent difference
                                                                         -13.547%--6.407%

                                                                         -6.406%--4.241%

                                                                         -4.240% - -2.863%

                                                                         -2.862%--1.821%

                                                                         -1.820%--0.956%

                                                                         -0.955% - -0
3.2.1.2 Exposure and Risk Modeling

3.2.1.2.1     Methods

       The HAPEM5 exposure model used in this assessment is the most recent version in a
series of models that the EPA has used to model population exposures and risks at the urban and
national scale in a number of assessments.139'140'141  HAPEM5 is designed to assess average
long-term inhalation exposures of the general population, or a specific sub-population, over
spatial scales ranging from urban to national. HAPEM5 uses the general approach of tracking
representatives of specified demographic groups as they move among indoor and outdoor
microenvironments and among geographic locations. The estimated pollutant concentrations in
each microenvironment visited are combined into a time weighted average concentration, which
is assigned to members of the demographic group.

       HAPEM5 uses four primary sources of information: population data from the US Census,
population activity data, air quality data, and microenvironmental data.  The population data used
is obtained from the US census.  Two kinds of activity data are used:  activity pattern data and
commuting pattern data. The activity pattern data quantify the amount of time individuals spend
in a variety of microenvironments and come from EPA's Consolidated Human Activity Database
(CHAD).142  The commuting data contained in the HAPEM5 default file were derived from a
special 1990 US Census study that specifies the number of residents of each tract that work in
that tract and every other US  Census tract. The air quality data come from ASPEN (after
                                          3-39

-------
background has been added). The microenvironmental data consist of factors that estimate air
toxic concentrations in specific microenvironments, based on penetration of outdoor air into the
microenvironment, proximity of the microenvironment to the emission source, and emission
sources within the microenvironment.  These factors vary among pollutants.143

       HAPEM5 has a number of technical improvements over the previous version of HAPEM.
These improvements, along with other details of the model, are described in the HAPEM5 User's
Guide.144  The projection year HAPEM runs used year 2000 census data and 1990 commuting
pattern data. Average lifetime exposure for an individual in a census tract was calculated from
data for individual demographic groups using a post-processing routine. We estimated the
contributions to ambient concentrations for the following source sectors: major, area and other,
onroad gasoline, onroad diesel, nonroad gasoline, remaining nonroad (diesel and compressed
natural gas), and background.

       Once HAPEM runs were completed, cancer risk and non-cancer risk calculations were
made for each of the mobile source air toxic pollutants. Table 3.2-4 lists the pollutants with their
respective Unit Risk Estimates (UREs) for cancer calculations and reference concentrations
(RfCs) for non-cancer calculations.  These are the same values used in the 1999 National-Scale
Air Toxics Assessment, and more detailed information on how dose-response values were
selected is provided at the website for that assessment. Also listed are the cancer weight of
evidence classifications and target organ system(s) for non-cancer calculations. The weight of
evidence classifications provided in this table were developed under EPA's 1986 risk assessment
guidelines where:

A = Known human carcinogen
Bl = Probable human carcinogen, based on incomplete human  data
B2 = Probable human carcinogen, based on adequate animal data
C = Possible human carcinogen

Dose-response values were selected using the following hierarchy:

       1)  EPA IRIS assessments.
       2)  Agency for Toxic Substances and Disease Registry (ATSDR) minimum risk levels
       (MRLs)  for non-cancer effects - used as RfC.
       3)  California Office of Environmental Health Hazard Assessment (OEHHA) values.
                                         3-40

-------
    Table 3.2-4. Dose Response Values Use in Risk Modeling (Concentrations in ug/m3)
HAP
1,3 -Butadiene
2,2,4-
Trimethylpentane
Acetaldehyde
Acrolein
Benzene
Chromium III
Chromium VI
Ethyl Benzene
Formaldehyde
Hexane
Manganese
MTBE
Naphthalene
Nickel
Propionaldehyde
POM1
POM2
POMS
POM4
POMS
POM6
POM7
POMS
Styrene
Toluene
Xylenes
Carcinogen
Class
A
N/A
B2

A
N/A
A

B



C
A
N/A
B2
B2
B2
B2
B2
B2
B2
B2



URE
(per ng/m3)
S.OxlO'5
N/A
2.2xlO'6
0
7.8xl(r6*
N/A
1.2xl(r2
0
5.5xlO'9
N/A
N/A
N/A
3.4xlO'5
1.6xl(r4
N/A
5.5xlO'5
5.5xlO'5
.OxlO'1
.OxlO'2
.OxlO'3
.OxlO'4
.OxlO'5
2.0xlO'4
N/A
N/A
N/A
Source
IRIS

IRIS

IRIS

IRIS

CUT



CAL
EPA/OAQ
PS

OAQPS
OAQPS
OAQPS
OAQPS
OAQPS
OAQPS
OAQPS
OAQPS



Organ
Systems
Reproductive
N/A
Respiratory
Respiratory
Immune
N/A
Respiratory
Developmental
Respiratory
Respiratory,
Neurological
Neurological
Liver, Kidney,
Ocular
Respiratory
Respiratory,
Immune
N/A








Neurological
Respiratory,
Neurological
Neurological
RfC(mg/
m3)
2.0xlO'3
N/A
9.0xlO'3
2.0xlO'5
S.OxlO'2
N/A
l.OxlO'4
1.0
9.8xlO'3
2.0X10'1
S.OxlO'5
3.0
S.OxlO'3
6.5xlO'5
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
1.0
4.0X10'1
l.OxlO'1
Source


IRIS
IRIS
IRIS

IRIS
IRIS
ATSDR
IRIS
IRIS
IRIS
IRIS
CAL









IRIS
IRIS
IRIS
   "represents upper end of a range of MLE values

There are a number of exceptions to this hierarchy:

       1)  Formaldehyde — EPA no longer considers the formaldehyde URE reported in IRIS,
       which is based on a 1987 study, to represent the best available science in the peer-
       reviewed literature. Accordingly, the 1999 risk estimates for formaldehyde are based on a
       dose-response value developed by the CUT Centers for Health Research (formerly the
       Chemical Industry Institute of Toxicology) and published in 1999. This issue is
       discussed in Chapter 1 of the RIA.
       2)  Nickel — The IRIS URE for nickel inhalation shown in Table 3.2.-4 was derived from
       evidence of the carcinogenic effects of insoluble nickel compounds in crystalline form.
       Soluble nickel species, and insoluble species in amorphous form, do not appear to
       produce genotoxic effects by the same toxic mode of action as insoluble crystalline
       nickel. Nickel speciation information for some of the largest nickel-emitting sources
       (including oil combustion, coal combustion, and others) suggests that at least 35% of
                                         3-41

-------
       total nickel emissions may be soluble compounds. The remaining insoluble nickel
       emissions are not well-characterized, however. Consistent with this limited information,
       this analysis has conservatively assumed that 65% of emitted nickel is insoluble, and that
       all insoluble nickel is crystalline. On this basis, the nickel URE (based on nickel
       subsulfide,  and representative of pure insoluble crystalline nickel) was adjusted to reflect
       an assumption that 65% of the total mass of nickel may be carcinogenic. The ATSDR
       MRL in Table 2 was not adjusted, however, because the noncancer effects of nickel are
       not thought to be limited to the crystalline, insoluble form.
       3) POM — POM was divided into eight toxicity categories to cover the range of unit
       risks of the individual POM species and POM groups contained in the 1999 NEI.  The
       unit risks for those eight categories were based on the midpoint of the range of unit risks
       defining the toxicity category.  More details on the development of these unit risks can be
       found on the website for the 1999 National  Scale Assessment and in Appendix H of the
       2001 EPA draft report to the Science Advisory Board on the  1996 National-Scale
       Assessment.145

       Cancer risk estimates (the product of unit risk estimates and exposure levels) for various
pollutants were assumed to be additive, since there was no evidence  of non-additive interactions
for any of the pollutants.  Most of the estimates are based on the statistical upper confidence limit
(UCL) of the fitted dose-response curve, but the estimates for hexavalent chromium, nickel, and
benzene are based  on the  statistical best fit ("maximum likelihood estimate," or MLE). Except
for benzene and chromium, where risks are based on maximum likelihood dose-response values,
risks from mobile source  air toxics should all be considered upper-bound values. True risks
could be greater, but are likely to be lower, and could be zero.

       To express  chronic noncancer hazards, we used the RfC as part of a calculation called the
hazard quotient (HQ), which is the ratio between the concentration to which a person is exposed
and the RfC. A value of the HQ less than one indicates that the exposure is lower than the RfC
and that no adverse health effects would be expected. A value of the  HQ greater than one
indicates that the exposure is higher than the RfC. However, because many RfCs incorporate
protective assumptions in the face of uncertainty, an HQ greater than one does not necessarily
suggest a likelihood of adverse effects. Furthermore, the HQ cannot be translated to a probability
that adverse effects will occur and is not likely to be proportional to risk. A HQ greater than one
can best be described as indicating that a potential exists for adverse  health effects. However
one should evaluate the weight of evidence supporting the RfC value for a particular chemical
before determining potential risks. Following the approach used in the 1999 National-Scale
Assessment, combined noncancer hazards were calculated using the  hazard index (HI), defined
as the sum of hazard quotients for individual air toxics compounds that affect the same organ or
organ system. The  HI is only an approximation of the combined effect, because some of the
substances may affect the target organs in different (i.e., non-additive) ways. As with the HQ, a
value of the HI below 1.0 will likely not result in adverse effects over a lifetime of exposure.
However, a value of the HI greater than 1.0 does not necessarily suggest a likelihood of adverse
effects. Furthermore, the HI cannot be translated to a probability that adverse effects will occur
and is not likely to be proportional to risk. A HI greater than one can be best described as
indicating that a potential may exist for adverse health effects.
                                          3-42

-------
3.2.1.2.2      Exposure and Risk Trends for Air Toxics

       Table 3.2-5 summarizes nationwide average census tract exposure concentrations of
mobile source air toxics in 1999, 2015, 2020, and 2030. It should be noted that all the other non-
inventoried sources, as well as the contribution from transport, contribute to background levels.
Overall, exposure concentrations tend to be less than ambient concentrations because penetration
rates to indoor microenvironments are typically less than one.  However, highway vehicles make
a larger contribution to overall average population exposures than they do to ambient levels.
This is largely because of elevated exposures experienced inside vehicles.

       Table 3.2-6 summarizes national average population cancer risk across census tracts for
these years by pollutant, as well as total cancer risk across pollutants.  The total cancer risk from
mobile source air toxics (including the stationary source contribution) was about 23 in a million
in 1999.  This compares to an overall nationwide average population cancer risk from all air
toxics in the 1999 National-Scale Assessment of 48 in a million.   About twenty-two percent of
this risk in the 1999 National Scale Assessment is attributable to benzene.

       In all projection years, benzene emissions are by far the largest contributor to cancer risk
from mobile sources (see Figure 3.2-4).  Furthermore, about 90%  of the mobile source risk from
all air toxics is due to gasoline vehicles and engines, and about 95% of the benzene risk from
mobile sources is from gasoline vehicles and engines. Other significant contributors to cancer
risk from mobile source air toxics include 1,3-butadiene, acetaldehyde, naphthalene, and
hexavalent chromium.

       Despite significant reductions in risk from mobile source air toxics, average inhalation
cancer risks for these pollutants, accounting for both mobile and stationary source contributions,
remain well above 10 in 1,000,000 (Figure 3.2-5). In addition, average risk from exposure to
benzene remains above 5 in 1,000,000.
                                          3-43

-------
Table 3.2-5.  Mean Population Exposure Concentrations of Mobile Source Air Toxics in 1999, 2015, 2020, and 2030

Pollutant
1,3-Butadiene
2,2,4-Trimethylpentane
Acetaldehyde
Acrolein
Benzene
Chromium III
Chromium VI
Ethyl Benzene
Formaldehyde
Hexane
MTBE
Manganese
Naphthalene
Nickel
POM
Propionaldehyde
Styrene
Toluene
Xylenes

background
(Mgm^)
3.82E-02
O.OOE+00
3.97E-01
O.OOE+00
2.98E-01
O.OOE+00
O.OOE+00
O.OOE+00
5.85E-01
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
1.27E-01
1999 average concentrations (ng m"3)
major
1.61E-03
1.80E-02
2.51 E-02
2.72E-03
1 .88E-02
3.28E-04
4.39E-05
1 .55E-02
3.47E-02
5.68E-02
9.45E-03
1.07E-03
3.94E-03
3.12E-04
2.89E-03
8.52E-03
2. 11 E-02
1.72E-01
8.53E-02
area & other
1 .72E-02
2.67E-02
4.64E-02
2.48E-02
1.42E-01
1 .86E-04
8.19E-05
7.52E-02
7.43E-02
3.89E-01
5.38E-02
9.12E-04
4.04E-02
5.94E-04
1 .03E-02
1 .95E-02
1 .20E-02
6.97E-01
5.11E-01
on road
6.53E-02
7.57E-01
7.60E-01
6.72E-02
7.58E-01
6.10E-05
1.36E-05
3.08E-01
5.79E-01
3.03E-01
7.03E-01
7.17E-05
1.73E-02
5.48E-05
1.28E-03
1.87E-01
3.24E-02
1 .98E+00
1.18E+00
n on road
1 .51 E-02
1.12E-01
1.41E-01
1 .95E-02
1 .33E-01
2.67E-05
5.89E-06
7.13E-02
2.14E-01
4.64E-02
7.95E-02
7.60E-06
4.01 E-03
4.76E-05
5.69E-04
3.50E-02
6.32E-03
2.74E-01
3.30E-01
total (including
background)
1 .37E-01
9.14E-01
1.37E+00
1.14E-01
1.35E+00
6.01 E-04
1 .45E-04
4.70E-01
1.49E+00
7.95E-01
8.45E-01
2.06E-03
6.57E-02
1 .01 E-03
1 .50E-02
2.50E-01
7.18E-02
3.13E+00
2.23E+00
2015 annual average concentrations (fig m"3)
major
1 .78E-03
9.01 E-03
2.54E-02
3.01 E-03
1 .35E-02
4.17E-04
5.58E-05
1 .05E-02
4.32E-02
5.04E-02
9.57E-03
1 .28E-03
3.46E-03
3.58E-04
2.28E-03
7.87E-03
2.48E-02
1 .22E-01
7.04E-02
area & other
1.73E-02
2.71 E-02
4.86E-02
2.23E-02
1.64E-01
2.52E-04
1.12E-04
9.77E-02
8.43E-02
4.62E-01
4.99E-02
1.20E-03
4.90E-02
6.68E-04
1.17E-02
2.02E-02
1.61 E-02
9.12E-01
6.82E-01
on road
1 .84E-02
2.56E-01
2.86E-01
1 .86E-02
2.71 E-01
9.27E-05
2.07E-05
1 .03E-01
1 .76E-01
1 .22E-01
1.14E-01
1.14E-04
9.44E-03
8.36E-05
6.16E-04
7.36E-02
9.92E-03
6.46E-01
3.82E-01
nonroad
9.27E-03
6.04E-02
9.77E-02
1.54E-02
8.28E-02
2.84E-05
6.25E-06
4.10E-02
1.49E-01
2.82E-02
2.59E-02
8.84E-06
4.00E-03
5.19E-05
5.02E-04
2.32E-02
3.67E-03
1.60E-01
1.80E-01
total (including
background)
8.58E-02
3.52E-01
8.58E-01
5.94E-02
8.33E-01
7.90E-04
1.95E-04
2.52E-01
1 .04E+00
6.63E-01
2.00E-01
2.60E-03
6.59E-02
1.16E-03
1.51 E-02
1.25E-01
5.45E-02
1 .84E+00
1 .44E+00
                                                  5-44

-------
Table 3.2-5 (cont'd). Mean Population Exposure Concentrations of Mobile Source Air Toxics in 1999, 2015, 2020, and 2030

Pollutant
1,3-Butadiene
2,2,4-Trimethylpentane
Acetaldehyde
Acrolein
Benzene
Chromium III
Chromium VI
Ethyl Benzene
Formaldehyde
Hexane
MTBE
Manganese
Naphthalene
Nickel
POM
Propionaldehyde
Styrene
Toluene
Xylenes

background
(Hgm3)
3.82E-02
O.OOE+00
3.97E-01
O.OOE+00
2.98E-01
O.OOE+00
O.OOE+00
O.OOE+00
5.85E-01
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
1.27E-01
2020 annual average concentrations (|j,g in 3)
major
1.92E-03
9.70E-03
2.65E-02
3.38E-03
1.48E-02
4.66E-04
6.36E-05
1.17E-02
4.90E-02
5.54E-02
1.07E-02
1.42E-03
3.88E-03
3.89E-04
2.53E-03
7.91E-03
2.83E-02
1.36E-01
7.95E-02
area &
other
1.73E-02
2.86E-02
4.97E-02
2.16E-02
1.71E-01
2.84E-04
1.26E-04
1.07E-01
8.86E-02
4.98E-01
5.19E-02
1.32E-03
5.22E-02
7.34E-04
1.20E-02
2.07E-02
1.78E-02
l.OOE+00
7.51E-01
onroad
1.70E-02
2.27E-01
2.49E-01
1.71E-02
2.43 E-01
1.04E-04
2.33E-05
9.05E-02
1.62E-01
1.02E-01
8.78E-02
1.29E-04
9.38E-03
9.38E-05
6.34E-04
6.26E-02
9.15E-03
5.75E-01
3.39E-01
nonroad
9.80E-03
5.93E-02
9.80E-02
1.62E-02
8.55E-02
2.90E-05
6.38E-06
4.16E-02
1.50E-01
2.86E-02
2.67E-02
9.47E-06
4.25E-03
5.43E-05
5.07E-04
2.32E-02
3.84E-03
1.60E-01
1.82E-01
total
(including
background)
8.50E-02
3.24E-01
8.24E-01
5.83E-02
8.16E-01
8.84E-04
2.20E-04
2.51E-01
1.04E+00
6. 83 E-01
1.77E-01
2.88E-03
6.97E-02
1.27E-03
1.57E-02
1.14E-01
5.91E-02
1.87E+00
1.48E+00
2030 annual average concentrations (jig m3)
major
1.91E-03
9.69E-03
2.65E-02
3.38E-03
1.48E-02
4.66E-04
6.36E-05
1.17E-02
4.90E-02
5.54E-02
1.07E-02
1.42E-03
3.88E-03
3.89E-04
2.53E-03
7.91E-03
2.83E-02
1.36E-01
7.94E-02
area &
other
1.73E-02
2.86E-02
4.97E-02
2.16E-02
1.71E-01
2.84E-04
1.26E-04
1.07E-01
8.86E-02
4.98E-01
5.19E-02
1.32E-03
5.22E-02
7.34E-04
1.20E-02
2.07E-02
1.78E-02
l.OOE+00
7.51E-01
onroad
1.82E-02
2.38E-01
2.60E-01
1.85E-02
2.57E-01
1.30E-04
2.90E-05
9.49E-02
1.75E-01
1.04E-01
8.40E-02
1.63E-04
1.09E-02
1.17E-04
7.41E-04
6.46E-02
9.81E-03
6.06E-01
3.56E-01
nonroad
1.12E-02
6.45E-02
1.07E-01
1.83E-02
9.61E-02
3.03E-05
6.67E-06
4.65E-02
1.65E-01
3.20E-02
3.02E-02
1.08E-05
4.84E-03
5.96E-05
5.64E-04
2.52E-02
4.36E-03
1.77E-01
2.03E-01
total
(including
background)
8.76E-02
3.41E-01
8.44E-01
6.18E-02
8.40E-01
9.11E-04
2.26E-04
2.60E-01
1.07E+00
6.89E-01
1.77E-01
2.91E-03
7.18E-02
1.30E-03
1.59E-02
1.18E-01
6.03E-02
1.92E+00
1.52E+00
                                                      5-45

-------
Table 3.2-6.  National Average Cancer Risk Across Census Tracts for 1999, 2015, 2020, and 2030 by Pollutant

Pollutant
Total Risk: All MSATs
POM
Nickel
Naphthalene
Formaldehyde
Chromium VI
Benzene
Acetaldehyde
1,3-Butadiene

background
4.35E-06
O.OOE+00
O.OOE+00
O.OOE+00
3.22E-09
O.OOE+00
2.32E-06
8.74E-07
1.15E-06
1999 average cancer risk
major
1.14E-06
1.75E-07
4.99E-08
1.34E-07
1.91E-10
5.26E-07
1.47E-07
5.51E-08
4.82E-08
area &
other
5.19E-06
1.01E-06
9.50E-08
1.38E-06
4.08E-10
9.83E-07
1.11E-06
1.02E-07
5.15E-07
onroad
1.04E-05
8.42E-08
8.78E-09
5.87E-07
3.18E-09
1.63E-07
5.91E-06
1.67E-06
1.96E-06
nonroad
2.05E-06
3.70E-08
7.61E-09
1.36E-07
1.18E-09
7.07E-08
1.04E-06
3.11E-07
4.52E-07
total
(including
background)
2.31E-05
1.31E-06
1.61E-07
2.23E-06
8.18E-09
1.74E-06
1.05E-05
3.01E-06
4.12E-06
2015 annual average cancer risk
major
1.20E-06
1.40E-07
5.73E-08
1.18E-07
2.37E-10
6.70E-07
1.05E-07
5.59E-08
5.33E-08
area &
other
6.19E-06
1.17E-06
1.07E-07
1.67E-06
4.64E-10
1.34E-06
1.28E-06
1.07E-07
5.20E-07
onroad
3.92E-06
3.97E-08
1.34E-08
3.21E-07
9.68E-10
2.48E-07
2.12E-06
6.30E-07
5.52E-07
nonroad
1.39E-06
3.32E-08
8.30E-09
1.36E-07
8.22E-10
7.50E-08
6.46E-07
2.15E-07
2.78E-07
total
(including
background)
1.71E-05
1.38E-06
1.86E-07
2.24E-06
5.74E-09
2.34E-06
6.49E-06
1.89E-06
2.57E-06

Pollutant
Total Risk: All MSATs
POM
Nickel
Naphthalene
Formaldehyde
Chromium VI
Benzene
Acetaldehyde
1,3-Butadiene

background
4.35E-06
O.OOE+00
O.OOE+00
O.OOE+00
3.22E-09
O.OOE+00
2.32E-06
8.74E-07
1.15E-06
2020 annual average cancer risk
major
1.34E-06
1.54E-07
6.22E-08
1.32E-07
2.70E-10
7.63E-07
1.15E-07
5.84E-08
5.75E-08
area &
other
6.57E-06
1.20E-06
1.18E-07
1.77E-06
4.87E-10
1.52E-06
1.33E-06
1.09E-07
5.19E-07
onroad
3.61E-06
4.07E-08
1.50E-08
3.19E-07
8.90E-10
2.79E-07
1.90E-06
5.48E-07
5.09E-07
nonroad
1.44E-06
3.37E-08
8.69E-09
1.44E-07
8.27E-10
7.66E-08
6.67E-07
2.16E-07
2.94E-07
total
(including
background)
1.74E-05
1.43E-06
2.03E-07
2.37E-06
5.73 E-09
2.64E-06
6.36E-06
1.81E-06
2.55E-06
2030 annual average cancer risk
major
1.34E-06
1.54E-07
6.22E-08
1.32E-07
2.70E-10
7.63E-07
1.15E-07
5.84E-08
5.74E-08
area &
other
6.57E-06
1.20E-06
1.18E-07
1.77E-06
4.87E-10
1.52E-06
1.33E-06
1.09E-07
5.18E-07
onroad
3.91E-06
4.75E-08
1.88E-08
3.70E-07
9.65E-10
3.49E-07
2.00E-06
5.73E-07
5.47E-07
nonroad
1.61E-06
3.76E-08
9.53E-09
1.65E-07
9.06E-10
8.00E-08
7.49E-07
2.36E-07
3.35E-07
total
(including
background)
1.78E-05
1.44E-06
2.08E-07
2.44E-06
5.88E-09
2.71E-06
6.55E-06
1.86E-06
2.63E-06
                                               5-46

-------
Figure 3.2-4. Contributions to Inhalation Cancer Risk from Air Toxics Emitted by Mobile
       Sources, 2020 (Not Including Diesel PM and Diesel Exhaust Organic Gases)
                 Naphthalene
                     9%
                Nickel
                 0%
                    Formaldehyde
                        0%
                  Chromium VI
                      7%
                                     1,3-Butadiene
                                        16%
                                                            Acetaldehyde
                                                                15%
   Figure 3.2-5. Average Nationwide Cancer Risk from Emissions of Mobile Source Air
 Toxics from both Mobile and Stationary Sources across Census Tracts, 1999 to 2030 (Not
                Including Diesel PM and Diesel Exhaust Organic Gases)
               2.5E-05
            a:

            8
            c
            TO
            O
            O
               2.0E-05 -
1.5E-05 -
            ~  1 .OE-05
            o
            D>
            ra
               5.0E-06
              O.OE-+00
                         1999
                     2015
2020
2030
                                         Year
                                         3-47

-------
       It should also be noted that because of population growth projected to occur in the United
States, the number of Americans above cancer risk benchmarks is will increase.  Figure 3.2-6
depicts the U. S. population at various risk benchmarks for mobile source air toxics in 1999,
2015, 2020, and 2030, using population projections from EPA's BenMAP model, a tool the EPA
uses to estimate benefits of air pollution control strategies, and average census tract exposures.
(BenMAP was recently used for EPA's Clean Air Interstate Air Quality Rule (CAIR).146 ).
These statistics do not include populations in Alaska and Hawaii.  More details on the
methodology used to project the U. S. population above various cancer risk benchmarks are
provided in the document "National-Scale Modeling of Mobile Source Air Toxic Emissions, Air
Quality, Exposure and Risk for the Mobile Source Air Toxics Rule." From this figure it can be
seen that, based on average census tract risks, the vast majority of the population experiences
risks between one in a million (IxlO"6) and one in ten thousand (IxlO"4). However, the number
of people experiencing risks above one in a hundred thousand (IxlO"5) increases from 214
million in 1999 to 240 million in 2030.
  Figure 3.2-6. U. S. Population at Various Cancer Risk Benchmarks due to Exposure to
                         Mobile Source Air Toxics, 1999 - 2030
        o
        re
        o
        o_
                      DRisk<1E-06
                      D1E-06<=Risk<1E-05
                      • 1E-05<=Risk<1E-04
                      DRisk>1E-04
                      1999
2020     2030
                                   Year
       Tables 3.2-7 and 3.2-8 summarize national average population hazard quotient for
chronic non-cancer effects across census tracts for these years by pollutant, as well as the
respiratory hazard index across pollutants. The respiratory system is the only target organ
system where the hazard index exceeds one.  Although the average respiratory hazard index for
mobile source air toxics decreases by almost 50% between 1999 and 2030 (Figure 3.2-7), it is
still over 3 in 2030, indicating a potential for adverse health effects. In addition, about 95% of
this non-cancer risk is attributable to acrolein in all projection years. It should be noted that the
                                          3-48

-------
confidence in the RfC for acrolein is medium. About 25% of primary acrolein emissions are
from mobile sources, and about 70% of ambient concentrations of acrolein (and about 75% of
exposure) is attributable to mobile sources.  The mobile source contribution to concentrations
and exposure is largely attributable to the contribution from mobile source 1,3-butadiene, which
is transformed to acrolein in the atmosphere. Moreover, projected growth in the U. S. population
will increase the number of Americans with a respiratory hazard index for mobile source air
toxics above one, from 250 million in 1999 to 273 million in  2030 (Figure 3.2-8).

       Detailed summary tables presenting cancer risk, hazard quotients and hazard indices by
State, and for reformulated and non-reformulated (i.e., conventional) gasoline areas, can be
found in the docket for this rule, along with statistics on number of individuals above various
cancer and non-cancer benchmarks, by source sector.

3.2.1.2.3      Distributions of Air Toxics Risk across the U.  S.

       Table 3.2-9 gives the distribution of nationwide average cancer risks for mobile source
air toxics in 2020.  Summary tables providing distributions for other years, as well as
distributions by State and for reformulated and non-reformulated gasoline areas, can be found in
the docket for this rule.  Risk distributions are broader than the distributions of ambient
concentrations in Table 3.2-2.  For instance, while the 95th percentile benzene concentration is
about twice the median value, the 95th percentile cancer risk is roughly six times the median risk.
A key reason for this is the variability in activity patterns, concentrations among
microenvironments, and commuting patterns. Figures 3.2-9 through 3.2-12 depict the
geographic distributions of median county cancer risks in 2020 for all mobile source air toxics,
benzene, acetaldehyde and 1,3-butadiene. These geographic  distributions closely track
distributions of ambient concentrations, with the highest risks in major population centers of the
country where mobile source activity is the greatest. Relatively high benzene risks are also seen
in areas of the country where fuel benzene levels are higher, such as the Pacific Northwest, parts
of Alaska, and the upper Great Lakes region, since higher fuel benzene levels lead to higher
benzene emissions and higher exposures.

       Table 3.2-10 gives the distribution of nationwide average census tract hazard quotients
for acrolein, and hazard indices for the respiratory target system  in 2020. Patterns for other years
are similar.  The average respiratory hazard index at the 95th  percentile is almost 40 times  that at
the 5th percentile, and about six times the median.  Thus, some populations are experiencing
much higher hazard indices than others.  Figure 3.2-13 depicts the geographic distribution  of
median county respiratory hazard indices in 2020.  The high hazard indices in Idaho are the
result of high inventory estimates for wildfires and reflect a known error in the Idaho inventory
for this source.
                                           3-49

-------
Table 3.2-7. National Average Population Hazard Quotient for Chronic Non-Cancer Effects across Census Tracts

Pollutant
1,3-Butadiene
Acetaldehyde
Acrolein
Benzene
Chromium VI
Ethyl
Benzene
Formaldehyde
Hexane
MTBE
Manganese
Naphthalene
Nickel
Styrene
Toluene
Xylenes

Target System
Reproductive
Respiratory
Respiratory
Immunological
Respiratory
Developmental
Respiratory
Neurological,
Respiratory
Liver, Kidney,
Ocular
Neurological
Respiratory
Respiratory,
Immunological
Neurological
Respiratory,
Neurological
Neurological

background
1.91E-02
4.41E-02
O.OOE+00
9.93E-03
O.OOE+00
O.OOE+00
5.97E-02
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
1.27E-03
1999 average Hazard Quotient
major
8.04E-04
2.78E-03
1.36E-01
6.27E-04
4.39E-04
1.55E-05
3.55E-03
2.84E-04
3.15E-06
2.14E-02
1.31E-03
4.79E-03
2.11E-05
4.29E-04
8.53E-04
area &
other
8.58E-03
5.15E-03
1.24E+00
4.72E-03
8.19E-04
7.52E-05
7.58E-03
1.94E-03
1.79E-05
1.82E-02
1.35E-02
9.14E-03
1.20E-05
1.74E-03
5.11E-03
onroad
3.27E-02
8.44E-02
3.36E+00
2.53E-02
1.36E-04
3.08E-04
5.91E-02
1.52E-03
2.34E-04
1.43E-03
5.76E-03
8.44E-04
3.24E-05
4.96E-03
1.18E-02
Nonroad
7.54E-03
1.57E-02
9.77E-01
4.42E-03
5.89E-05
7.13E-05
2.19E-02
2.32E-04
2.65E-05
1.52E-04
1.34E-03
7.32E-04
6.32E-06
6.85E-04
3.30E-03
total
(including
background)
6.87E-02
1.52E-01
5.72E+00
4.50E-02
1.45E-03
4.70E-04
1.52E-01
3.98E-03
2.82E-04
4.13E-02
2.19E-02
1.55E-02
7.18E-05
7.81E-03
2.23E-02
2015 average Hazard Quotient
major
8.88E-04
2.82E-03
1.50E-01
4.49E-04
5.58E-04
1.05E-05
4.41E-03
2.52E-04
3.19E-06
2.55E-02
1.15E-03
5.50E-03
2.48E-05
3.05E-04
7.04E-04
area &
other
8.66E-03
5.40E-03
1.12E+00
5.47E-03
1.12E-03
9.77E-05
8.60E-03
2.31E-03
1.66E-05
2.40E-02
1.63E-02
1.03E-02
1.61E-05
2.28E-03
6.82E-03
onroad
9.19E-03
3.18E-02
9.32E-01
9.04E-03
2.07E-04
1.03E-04
1.80E-02
6.11E-04
3.80E-05
2.28E-03
3.15E-03
1.29E-03
9.92E-06
1.61E-03
3.82E-03
nonroad
4.63E-03
1.09E-02
7.71E-01
2.76E-03
6.25E-05
4.10E-05
1.52E-02
1.41E-04
8.62E-06
1.77E-04
1.33E-03
7.98E-04
3.67E-06
4.00E-04
1.80E-03
total
(including
background)
4.29E-02
9.54E-02
2.97E+00
2.78E-02
1.95E-03
2.52E-04
1.07E-01
3.31E-03
6.65E-05
5.20E-02
2.20E-02
1.79E-02
5.45E-05
4.60E-03
1.44E-02
                                                 5-50

-------
Table 3.2-7 (cont'd). National Average Population Hazard Quotient for Chronic Non-Cancer Effects across Census Tracts

Pollutant
1,3-Butadiene
Acetaldehyde
Acrolein
Benzene
Chromium VI
Ethyl Benzene
Formaldehyde
Hexane
MTBE
Manganese
Naphthalene
Nickel
Styrene
Toluene
Xylenes

Target System
Reproductive
Respiratory
Respiratory
Immunological
Respiratory
Developmental
Respiratory
Neurological,
Respiratory
Liver, Kidney,
Ocular
Neurological
Respiratory
Respiratory,
Immunological
Neurological
Respiratory,
Neurological
Neurological

background
1.91E-02
4.41E-02
O.OOE+00
9.93E-03
O.OOE+00
O.OOE+00
5.97E-02
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
1.27E-03
2020 average Hazard Quotient
major
9.58E-04
2.95E-03
1.69E-01
4.93E-04
6.36E-04
1.17E-05
5.00E-03
2.77E-04
3.58E-06
2.84E-02
1.29E-03
5.98E-03
2.83E-05
3.40E-04
7.95E-04
area&
other
8.64E-03
5.52E-03
1.08E+00
5.70E-03
1.26E-03
1.07E-04
9.04E-03
2.49E-03
1.73E-05
2.63E-02
1.74E-02
1.13E-02
1.78E-05
2.50E-03
7.51E-03
onroad
8.48E-03
2.77E-02
8.57E-01
8.12E-03
2.33E-04
9.05E-05
1.65E-02
5.08E-04
2.93E-05
2.58E-03
3.13E-03
1.44E-03
9.15E-06
1.44E-03
3.39E-03
Nonroad
4.90E-03
1.09E-02
8.09E-01
2.85E-03
6.38E-05
4.16E-05
1.53E-02
1.43E-04
8.91E-06
1.89E-04
1.42E-03
8.36E-04
3.84E-06
4.00E-04
1.82E-03
total
(including
background)
4.25E-02
9.15E-02
2.91E+00
2.72E-02
2.20E-03
2.51E-04
1.06E-01
3.42E-03
5.91E-05
5.75E-02
2.32E-02
1.96E-02
5.91E-05
4.68E-03
1.48E-02
2030 average Hazard Quotient
major
9.57E-04
2.95E-03
1.69E-01
4.93E-04
6.36E-04
1.17E-05
5.00E-03
2.77E-04
3.58E-06
2.84E-02
1.29E-03
5.98E-03
2.83E-05
3.40E-04
7.94E-04
area&
other
8.64E-03
5.52E-03
1.08E+00
5.69E-03
1.26E-03
1.07E-04
9.04E-03
2.49E-03
1.73E-05
2.63E-02
1.74E-02
1.13E-02
1.78E-05
2.50E-03
7.51E-03
onroad
9.12E-03
2.89E-02
9.25E-01
8.57E-03
2.90E-04
9.49E-05
1.79E-02
5.20E-04
2.80E-05
3.26E-03
3.63E-03
1.80E-03
9.81E-06
1.52E-03
3.56E-03
nonroad
5.59E-03
1.19E-02
9.16E-01
3.20E-03
6.67E-05
4.65E-05
1.68E-02
1.60E-04
1.01E-05
2.15E-04
1.61E-03
9.17E-04
4.36E-06
4.42E-04
2.03E-03
total
(including
background)
4.38E-02
9.38E-02
3.09E+00
2.80E-02
2.26E-03
2.60E-04
1.09E-01
3.44E-03
5.89E-05
5.82E-02
2.39E-02
2.00E-02
6.03E-05
4.80E-03
1.52E-02
                                                     5-51

-------
  Table 3.2-8. National Respiratory Hazard Index for Chronic Non-Cancer Effects across
                                   Census Tracts
Respiratory System Average Hazard Index
Year
1999
2015
2020
2030
background
1.04E-01
1.04E-01
1.04E-01
1.04E-01
major
1.49E-01
1.65E-01
1.85E-01
1.85E-01
area & other
1.28E+00
1.16E+00
1.13E+00
1.13E+00
onroad
3.52E+00
9.88E-01
9.08E-01
9.79E-01
nonroad
1.02E+00
7.99E-01
8.38E-01
9.48E-01
total (including
background)
6.07E+00
3.22E+00
3.17E+00
3.35E+00
Figure 3.2-7. Average Respiratory Hazard Index for U.S. Population (Aggregate of Hazard
                          Quotients for Individual Pollutants)
        	1     \
     •5
     a
     nj
     I
I
              1999
                            2015
                                          2020
                                                        2030
                                   Year
                                        3-52

-------
  Figure 3.2-8. U. S. Population at Various Non-Cancer Hazard Benchmarks due to

                Exposure to Mobile Source Air Toxics, 1999 - 2030
          o


          1
          S
          a.
          o
          a.
                     1999
2015   2020


    Year
2030
Table 3.2-9.  Distribution of Average Census Tract Cancer Risks for Mobile Source Air

                                 Toxics in 2020

Pollutant
Total Risk: All HAPs
POM
Nickel
Naphthalene
Formaldehyde
Chromium VI
Benzene
Acetaldehyde
1,3-Butadiene
2020 risk distribution
5th percentile
3.59E-06
7.48E-08
9.32E-07
2.08E-06
1.65E-08
1.75E-09
1.02E-07
1.27E-09
9.64E-08
10th
percentile
4.61 E-06
1.40E-07
9.98E-07
2.54E-06
4.33E-08
2.29E-09
2.09E-07
3.38E-09
1.71E-07
25th
percentile
8.04E-06
7.38E-07
1.20E-06
3.87E-06
1 .82E-07
3.53E-09
6.20E-07
1.35E-08
3.32E-07
Median
1.34E-05
1.99E-06
1.60E-06
5.61 E-06
6.73E-07
5.12E-09
1 .44E-06
5.04E-08
6.81 E-07
75th
percentile
2.02E-05
3.05E-06
2.13E-06
7.63E-06
1 .83E-06
6.80E-09
2.79E-06
1 .53E-07
1.15E-06
90th percentile
3.34E-05
4.48E-06
2.94E-06
1.07E-05
5.32E-06
9.63E-09
5.38E-06
3.62E-07
1.93E-06
95th
percentile
4.39E-05
7.47E-06
3.64E-06
1.35E-05
8.58E-06
1.23E-08
8.47E-06
6.15E-07
3. 11 E-06
                                     3-53

-------
Figure 3.2-9. 2020 County Median Cancer Risk for All Mobile Source Air Toxics
         Figure 3.2-10. 2020 County Median Cancer Risk for Benzene
                                   3-54

-------
Figure 3.2-11.  2020 County Median Cancer Risk for Acetaldehyde
Figure 3.2-12.  2020 County Median Cancer Risk for 1,3-Butadiene
                            3-55

-------
 Table 3.2-10. Distribution of Average Census Tract Hazard Quotients/Hazard Indices for
      Mobile Source Air Toxics (from both Mobile and Stationary Sources) in 2020

Pollutant
Acrolein
Respiratory System
2020 average Hazard Quotient or Hazard Index
5th percent! le
2.13E-01
3.06E-01
10th
percentile
3.65E-01
4.79E-01
25th
percentile
8.08E-01
9.61 E-01
Median
1 .69E+00
1.91E+00
75th
percentile
3.36E+00
3.67E+00

90th percentile
6.99E+00
7.39E+00

95th
percentile
1.11E+01
1.17E+01
Figure 3.2-13.  2020 County Median Non-Cancer Hazard Index Respiratory Mobile Source
                                    Air Toxics
                                       3-56

-------
3.2.1.2.4      Impacts of Proposed Fuel Benzene Controls on Average Inhalation Cancer Risk

       The fuel benzene standard proposed in this rule will substantially reduce inhalation
cancer risk from exposure to benzene emitted by mobile sources across the United States. Table
3.2-11 shows that in 2015, 2020, and 2030, the highway vehicle contribution to benzene cancer
risk will be reduced on average 8 to 9 percent across the U.S., the nonroad equipment
contribution will be reduced about 7 percent, and the area source contribution about 4 percent.
Reductions in conventional gasoline areas (i.e., areas not subject to reformulated gasoline) are
almost 13 percent. In States with high fuel benzene levels, such as Minnesota and Washington,
the risk reduction exceeds 17 percent (Table 3.2-12). Figure 3.2-14 depicts the impact on the
mobile source contribution to nationwide average population cancer risk from benzene in 2020.
Figure 3.2-15 presents the distribution of percent reductions in average benzene cancer risk for
U. S. counties with the proposed control in 2020. Patterns are similar for other years. Summary
tables providing exposure and risk data by  State, as well as maps of benzene cancer risks with
fuel controls and percent reductions with controls, can be found in the docket for the rule.
Similar data are also available for 1,3-butadiene, formaldehyde and acetaldehyde, even though
cancer risks were not significantly affected. Data are also available for noncancer risks, which
are also not significantly affected.

       It should be noted that the estimated total  reductions are significant underestimates, since
we could not account for further reductions in emissions from transport, i.e., background sources.
In Section 3.2.1.4, we provide a quantitative estimate of the expected reductions in background
concentrations in future years.
                                           3-57

-------
Table 3.2-11. Contributions of Source Sectors to Nationwide Average Benzene Cancer Risk, with and without Proposed Fuel
                                     Benzene Standard, 2015, 2020, and 2030


Reference
Control
% Difference

2015 Average Risks
major
1.05E-07
1 .04E-07
0.8

area & other
1.28E-06
1 .23E-06
3.6

highway
vehicles
2.12E-06
1.92E-06
9.0

nonroad
6.46E-07
6.00E-07
7.0

Average Nationwide Difference in Risk - Non RFC Areas

Reference
Control
% Difference


7.24E-08
7.19E-08
0.8


1.05E-06
1 .OOE-06
4.2


1.66E-06
1 .44E-06
13.6

Average Nationwide Difference in Risk - RFC Areas

Reference
Control
% Difference

1.64E-07
1 .63E-07
0.7

1.70E-06
1 .65E-06
3.0

2.93E-06
2.80E-06
4.4

4.57E-07
4.02E-07
12.2



9.84E-07
9.57E-07
2.7
total (including
background)
6.49E-06
6.21 E-06
4.3



5.40E-06
5.08E-06
6.0



8.46E-06
8.25E-06
2.4
2020 Average Risks
major
1.15E-07
1.14E-07
0.8



8.06E-08
7.99E-08
0.8



1.78E-07
1.77E-07
0.8
area & other
1.33E-06
1 .28E-06
3.7



1.08E-06
1 .04E-06
4.2



1.79E-06
1.73E-06
3.1
highway
vehicles
1.90E-06
1.73E-06
8.7



1.49E-06
1 .30E-06
13.1



2.63E-06
2.52E-06
4.1
nonroad
6.67E-07
6.19E-07
7.1



4.74E-07
4.15E-07
12.4



1.01 E-06
9.85E-07
2.8
total (including
background)
6.36E-06
6.10E-06
4.1



5.29E-06
5.00E-06
5.6



8.28E-06
8.09E-06
2.3
2030 Average Risks
major
1.15E-07
1.14E-07
0.8



8.05E-08
7.99E-08
0.8



1.78E-07
1.76E-07
0.8
area & other
1.33E-06
1 .28E-06
3.7



1.08E-06
1 .03E-06
4.2



1.79E-06
1.73E-06
3.1
highway
vehicles
2.00E-06
1.84E-06
8.3



1.54E-06
1.34E-06
12.9



2.84E-06
2.73E-06
3.9
nonroad
7.49E-07
6.95E-07
7.2



5.35E-07
4.68E-07
12.5



1.13E-06
1.10E-06
2.8
total (including
background)
6.55E-06
6.28E-06
4.1



5.40E-06
5.09E-06
5.7



8.62E-06
8.42E-06
2.3
                                                        5-58

-------
Table 3.2-12. States with Highest Reductions in Average Benzene Cancer Risk Resulting
                          from Mobile Source Emissions, 2020
State
Alaska
Washington
Minnesota
New Mexico
Oregon
Average Risk -
Reference Case
1.22xlO'6
3.21xlO'6
2.60xlO'6
1.45xlO'6
2.97xlO'6
Average Risk -
0.62% Benzene
Standard
8.36xlO'7
2.64xlO'6
2.14xlO'6
1.19xlO'6
2.47xlO'6
Percent Difference
-31%
-18%
-18%
-18%
-17%
  Figure 3.2-14. Contribution to Nationwide Average Population Cancer Risk in 2020
                    Resulting from Proposed Fuel Benzene Controls
      4.0E-06
      3.5E-06
      3.0E-06
      2.5E-06
 Average
 Cancer
 Risk  2.0E-06
      1.5E-06
      1 .OE-06
      5.0E-07
     O.OE+00
• Without Proposed Control
DO.62% Benzene Standard
                  U.S.
                               Non-RFG Areas
                                                RFG Areas
                                          3-59

-------
 Figure 3.2-15.  Distribution of Percent Reductions in Median Benzene Cancer Risk, 2020,
                       for U.S. Counties with the Proposed Control
                                                                  Percent Difference

                                                                  ^^ -12.397% --7.180%

                                                                    | -7.179%--5.034%

                                                                  ^^ -5.033%--3.530%

                                                                      -3.529% - -2.332%

                                                                      -2.331%--1.252%

                                                                      -1.251%-0.002%
As a result of the proposed fuel benzene control, the number of people above the 1 in 100,000
cancer risk level due to exposure to all mobile source air toxics from all sources will decrease by
over 3 million in 2020 and by about 3.5 million in 2030. The number of people above the 1 in
100,000 increased cancer risk level from exposure to benzene from all sources decreases by over
4 million in 2020 and 5 million in 2030 (Table 3.2-13).

   Table 3.2-13.  Decrease in Number of People with Inhalation Exposure above the 1 in
    100,000 Cancer Risk Level due to Inhalation Exposure from Ambient Sources, with
                             Proposed Fuel Benzene Control
Year
2015
2020
2030
Benzene
4,976,000
4,150,000
5,253,000
All Mobile Source Air
Toxics
3,226,000
3,077,000
3,477,000
The proposed standard will have little impact on the number of people above various respiratory
hazard index levels, since this potential non-cancer risk is dominated by exposure to acrolein.
Population statistics on number of individuals above various cancer and non-cancer benchmarks,
by source sector, with fuel benzene control are available in the docket for this rule.
                                          3-60

-------
3.2.1.3 Impacts of Near Roadway Microenvironment on Modeled Exposures to Benzene

3.2.1.3.1      Assessment Methods

       In HAPEM5, if only a single outdoor concentration is provided for each census tract, as
is typical, this concentration is assumed to uniformly apply to the entire census tract. EPA has
recently developed a new version of the model, HAPEM6, which refines the model to account
for the spatial variability of outdoor concentrations within a tract due to higher outdoor
concentrations of onroad mobile source pollutants at locations near major roadways.8 The new
version of HAPEM more accurately reflects the average and range of exposure concentrations
within each census tract by accounting for some of the spatial variability in the outdoor
concentrations within the tract, and by extension some of the spatial variability in indoor
concentrations within the tract. At this time, HAPEM6 only accounts for near-roadway effects
for benzene.

       The new version of HAPEM was developed using the following three steps.

1) Estimating the fraction of the population living near major roadways in each census tract by
demographic group.

       First, the "zone of influence" of transportation facilities needed to be determined - that is,
the width of the area around major roads within which concentrations of benzene are elevated.
Second, population data of sufficient geographic specificity was needed.  Using geographic
information systems, we conducted a study of the populations  in three states, Colorado,  Georgia,
and New York147  In Colorado, 22% live within 75 meters of a major road, while an additional
33% live between 75 and 200 meters of major roads. In Georgia, the respective percentages are
17% living within 75 meters and an additional 24% living between 75 and 200 meters. In New
York, the percentages  are 31% and 36%.

       This was done  by  overlaying  extracts from the ESRI StreetMap US roadway geographic
database on a geographic database of US Census blocks.

2) Estimating the increase near major roadways of air toxic pollutant concentrations from
onroad motor vehicle emissions relative to concentrations at other outdoor locations.

       In this step, data on spatial gradients of pollutants near roads from several sources were
analyzed.148'149'150'151'152'153'154 Data were analyzed for their suitability to estimate
concentration distributions within 75 meters of a major roadway, or between 75 meters and 200
meters from such a road.  All the data sources analyzed were from monitoring studies, except for
one, which was a modeling study in Portland using the CALPUFF dispersion model to estimate
concentrations at receptors located at census block centroids (Cohen et al., 2005). The
monitoring data were consistent with the spatial gradients characterized using CALPUFF, but
had limitations which precluded their use in quantifying concentration distributions.  Among the
limitations were that measured concentrations did not span various distances near the road
B The term "major roadway" will be used to describe a "Limited Access Highway", "Highway", "Major Road" or
"Ramp", as defined by the Census Feature Class Codes (CFCC).
                                          3-61

-------
needed to develop a model, monitors were all downwind, or measurements were taken at limited
times of year, making it difficult to extrapolate to annual averages. Modeling output from
Portland included receptor locations at many distances from major roadways and calculated
annual averages of benzene concentration in ambient air at those receptors.  Thus the Portland
modeling data was used to develop concentration ratios via regression analysis.

       One way of comparing the concentrations for near-road and other locations is to examine
the distribution of ratios between concentrations at multiple distances from a major road. Figure
3.2-16 presents a distribution of the concentration ratios between locations "near" a major
roadway (within 75 meters) and locations "far" from a roadway (>200 meters distant). Also
shown is a distribution of concentration ratios between locations at "intermediate" locations
(between 75 meters and 200 meters) and those "far" from a roadway. These data were derived
from the Portland modeling.

     Figure 3.2-16. Distribution of Ratios of Near Roadway to Remote Concentrations

      1.0
I* 0.8
!5
ro
.a
8 0.6
Q_
0
    O
      0.4
    E
   O 0.2 +
      0.0
            75 to 200 meters:
            median = 1.6
                                      0 to 75 meters:
                                      median = 2.5
                                        4
                                       Ratio
                                                                    8
3) Modification of the HAP EM model

       HAPEM6 models exposure for a simulated, demographically representative population
within each census tract.  For each simulated individual, HAPEM6 randomly selects for each
home tract indoor microenvironment whether it is within Dl (75) meters of a major roadway,
from Dl to D2 (75 to 200) meters from a major roadway, or greater than D2 meters from a major
roadway, according to the database developed in the first step described above.

       If the simulated person is a commuter, HAPEM6 randomly selects for each work tract
indoor microenvironment whether it is within Dl meters of a major roadway,  from Dl to D2
meters from a major roadway,  or greater than D2 meters from a major roadway, according to the
fractions of the populations living near major roadways in Step  1.
                                         3-62

-------
       For each microenvironment, HAPEM6 selects a near-roadway ratio multiplier
distribution for the proper distance according to the probabilities specified in the Step 2; and
selects a ratio multiplier from that distribution.

       HAPEM6 calculates the ambient concentration for locations more than D2 meters from a
major roadway according to the equation:

Afar*Cfar + ADi_D2*CDi-D2 = CASPEN          or

Afar*Cfar + ADl-D2*RDl-D2*Cfar = CASPEN     Or

Cfar = CASPEN / (Afar+ADi_D2*RDi_D2)

Where:
CASPEN is the ASPEN concentration prediction for the tract
CDi-D2 is the ambient concentration in the area between Dl and D2 meters from a major roadway
(i.e., ASPEN concentration estimate x mean of the ratio multiplier distribution)
Cfar is ambient concentration in the area more than D2 meters from a major roadway
Aoi-D2 is the fraction of the tract area that is between Dl and D2 meters from a major roadway,
Afar is the fraction of the tract area that is more than D2 meters from a major roadway, and
Roi-D2 is the near roadway ratio multiplier selected for the Dimeters to D2 meters distance range.

       The implicit assumption for this step is that the ASPEN estimate for the average census
tract concentration represents the spatial average over the tract excluding the area within Dl
meters of a major roadway. This is a reasonable assumption given the way that the ASPEN
concentration estimate is generated.

       The ASPEN estimate for the census tract average concentration is an aggregate of the
contributions from all sources within 50 km of the  tract. For sources located outside of the tract
the concentration contribution is estimated at the geographic centroid of the tract and assumed to
be uniform throughout the tract. For sources within the tract, which we expect to be the
dominant contributors, the concentration contribution is calculated as a weighted average of the
concentrations at all the modeling receptors that fall within the tract.

       HAPEM6 calculates the "ambient" concentrations in at different distances from major
roads by applying the relevant ambient concentration ratio.  If located within 75 meters of a
major road, the concentration ratio for that area is applied to the Cfar concentration, shown above.
Indoor microenvironmental concentrations are calculated based on this ambient concentration.
Likewise if located between 75 and 200 meters of a major road, the concentration ratio for that
area is used to calculate ambient and indoor concentrations at that point.  If located more that 200
meters away from a major road, the Cfar concentration is used for ambient air and for calculating
indoor microenvironmental concentrations.

3.2.1.3.2      Results

       The revised model was  run for three geographic areas representing different parts of the

                                          3-63

-------
country.  For these initial runs, benzene was the only pollutant modeled. ASPEN output for
calendar year 1999 were used as inputs.  We studied the states of Colorado, Georgia, and New
York. These areas are intended to represent different geographies, development patterns, and
housing densities.

       Within a given census tract, the HAPEM model predicts 30 lifetime exposure
concentrations depicting the variation in potential individual exposures within the tract.  Such
variation can result from differences in human activity patterns and as is the case for the
simulation with HAPEM6, proximity of populations to roadways. Table 3.2-14 depicts the
results of a comparison between HAPEM5 (does not include near roadway residents) and
HAPEM6 (includes near roadway residents).  The Table shows the distribution of individual
exposure concentrations both within a given tract as well as in tracts across the state. When
applied to each of these states, the greatest change in exposures resulting from the use of
HAPEM6 occurred for the individuals at the upper end of the exposure distribution within each
tract.  Further, this effect was most pronounced at the tracts with the highest exposures within a
state.  In summary, the models show that including the effects of residence locations can result in
exposures to some individuals that are up to 50% higher than those predicted by HAPEM5 (as
was applied in the 1999 NAT A).

    Table 3.2-14. Comparison Predicted Exposure Concentrations from HAPEM5 and
                 HAPEM6 Results for Georgia, Colorado, and New York
Modeled
State
Georgia
Colorado
New
York
Model
Version
HAPEM6
HAPEM5
HAPEM6
HAPEM5
HAPEM6
HAPEM5
Percentile of
Tracts
Across State
5% tract
50% tract
95% tract
5% tract
50% tract
95% tract
5% tract
50% tract
95% tract
5% tract
50% tract
95% tract
5% tract
50% tract
95% tract
5% tract
50% tract
95% tract
Percentile of Exposure within Census Tract (ug/m3)
1%
0.54
0.83
1.29
0.63
0.91
1.65
0.49
0.71
0.66
0.56
0.76
0.89
0.97
1.37
2.10
1.19
1.57
2.54
5%
0.57
0.87
1.34
0.66
0.94
1.70
0.52
0.74
0.71
0.58
0.78
0.93
1.02
1.43
2.28
1.22
1.62
2.65
10%
0.60
0.89
1.39
0.68
0.96
1.75
0.53
0.76
0.76
0.59
0.81
0.96
1.07
1.49
2.39
1.26
1.66
2.71
25%
0.64
0.95
1.53
0.71
1.01
1.84
0.58
0.81
0.84
0.63
0.85
1.01
1.16
1.59
2.63
1.32
1.74
2.85
50%
0.73
1.06
1.86
0.77
1.08
2.01
0.67
0.92
0.99
0.69
0.92
1.09
1.39
1.85
3.00
1.42
1.86
3.00
75%
0.92
1.28
2.45
0.85
1.19
2.19
0.80
1.08
1.24
0.77
1.02
1.17
1.76
2.27
3.53
1.57
2.05
3.18
90%
1.15
1.57
3.11
0.95
1.32
2.40
1.01
1.30
1.50
0.87
1.13
1.24
2.20
2.79
4.10
1.77
2.29
3.39
95%
1.31
1.76
3.59
1.00
1.40
2.59
1.10
1.42
1.70
0.92
1.20
1.29
2.52
3.16
4.49
1.95
2.50
3.56
99%
1.55
2.07
4.43
1.11
1.57
2.76
1.27
1.61
2.18
1.04
1.34
1.40
3.09
3.82
5.05
2.31
2.92
3.79
       The results indicate that by accounting for within-tract variability in concentrations,
HAPEM6 substantially increases overall variability in exposure to benzene. Demonstrating
                                          3-64

-------
these trends, the results of this modeling exercise for the state of New York are shown in Figure
3.2-17.  In the graph, the horizontal axis shows percentiles of exposure within census tracts,
while the range of each bar represents the 50th, 5th, and 95th percentiles of exposure concentration
across census tracts within the state.

       Overall, these study results indicate that proximity to major roads can significantly
increase personal exposure for populations living near major roads.  These models will be
extended to a national  scale for the final rulemaking.

  Figure 3.2-17. Changes in Predicted Benzene Exposure Patterns between HAPEM5 (no
 near-roadway adjustment) and HAPEM6 (with near-roadway adjustment) for New York
                                   25%       50%       75%       90%
                                 Percent Exposure within Census Tract
                                                                       95%
                                                                                99%
3.2.1.4 Strengths and Limitations

       Air quality, exposure, and risk were assessed using the best available suite of tools for
national-scale analysis of air toxics. In addition, the modeling done to support this rule was
consistent with NATA for 1999, making direct comparisons of results possible. The first NAT A,
done for calendar year 1996, was reviewed by EPA's Science Advisory Board, and the analyses
done for 1999 incorporate several changes in response to comments made in this peer review.
Among the improvements were:
                                          3-65

-------
   •   Improved emission inventory with detailed characterization of source categories within
       the onroad and nonroad source sectors and more speciated data for some pollutant groups
       (POM) within particular source categories.
   •   Speciation of chromium to hexavalent form based on emission sources rather than a
       single number applied across all sources
   •   Improved surrogates for spatial allocation in EMS-HAP.
   •   Improved estimation of "background" concentrations for many pollutants.  These
       background levels were previously uniform across the country. Now, for many
       pollutants, background levels are based on recent monitor data and spatially vary
       depending on county population density.155
   •   Improved version of HAPEM, which includes more recent census data, commuting
       algorithms and better characterization of exposure distributions through improvements in
       modeling long-term activity patterns and variability in concentration levels in
       microenvironments.

   The SAB expressed their belief that due to the limitations inherent in the analysis, the 1996
NAT A should not be used to support regulatory action. However, the use of the improved
analyses does provide useful  insight on the nature of the mobile source air toxics problem and
the possible public health improvements associated with this rule.

       In addition to the strengths listed above, there are limitations due to uncertainty.  The
inventory uncertainties discussed in Chapter 2. There are a number of additional significant
uncertainties associated with the air quality, exposure and risk modeling.  These uncertainties
result from a number of parameters including: development of county-level estimates from
broader geographic data (i.e., state, regional or national), surrogates used to allocate emissions to
census tracts, parameters used to characterize photochemical processes, long range transport,
terrain effects, deposition rates, human activity pattern parameters, assumptions about
relationships between ambient levels in different microenvironments, and dose-response
parameters.  The modeling also has certain key limitations: results are most accurate for large
geographic areas and cannot be used to identify "hot spots," such as the near road
microenvironment, exposure modeling does not fully reflect variation among individuals, non-
inhalation exposure pathways and indoor sources are not accounted for; and for some pollutants,
the ASPEN dispersion model may underestimate concentrations. Also, the 1999 NATA does not
include default adjustments for early life exposures recently recommended in the Supplemental
Guidance for Assessing Susceptibility from Early-Life Exposure to Carcinogens.156
Incorporation of such adjustments would lead to higher estimates of lifetime risk. EPA will
determine as part of the IRIS assessment process which substances meet the criteria for making
adjustments, and future assessments will reflect them.

       As part of the 1999 NATA, EPA compared ASPEN-modeled  concentrations with
available, but geographically limited, ambient air quality monitoring data for 1999. For each
monitor-pollutant combination, EPA compared the annual average concentration estimated by
the ASPEN model at the exact geographical coordinates of the monitor location with the annual
average monitored value to get a point-to-point comparison between the model and monitor
concentrations. The agreement between model and monitor values for benzene was very good,
with a median model to monitor ratio of 0.95, and 74% of sites within a factor of 2. Agreement

                                          3-66

-------
for acetaldehyde was almost as good as benzene, but data suggest that ASPEN could be
underpredicting for other mobile source air toxics (see Table 3.2-15).

More detailed discussion of modeling limitations and uncertainties can be found on the 1999
NAT A website.

   Table 3.2-15. Agreement of 1999 Model and Monitors by Pollutant on a Point-to-Point
   Basis Pollutants listed were Monitored in at least 30 Sites and in a Broad Geographical
                                 Area (Several States)
Pollutant
Acetaldehyde
Benzene
Formaldehyde
Chromium
Manganese
Nickel
No. of
Sites
68
115
68
42
34
40
Median of
Ratios
0.92
0.95
0.64
0.29
0.4
0.53
Within
Factor of 2
74%
72%
60%
26%
44%
48%
Within
30%
44%
43%
28%
5%
15%
18%
Underestimated
56%
52%
76%
95%
91%
75%
       In addition to the limitations and uncertainties associated with modeling the 1999 base
year, there are additional ones in the projection year modeling. For instance, the modeling is not
accounting for impacts of demographic shifts that are likely to occur in the future. A key
limitation is using 1999 "background" levels to account for mid-range to long-range transport.
However, since background is related to emissions far away from receptors, these levels should
decrease as those emissions decrease. We performed a sensitivity analysis for benzene,
formaldehyde, acetaldehyde and 1,3-butadiene to evaluate the potential bias introduced by this
assumption.  We used background estimates scaled by the change in the inventory for a future
year relative to 1999. The scaling factors applied to the background level for an individual
county were based on emissions for counties within 300 kilometers of that county's centroid.
Our analysis indicated that using a scaled background reduced benzene concentrations about
15% on average across the U. S in 2015, 2020, and 2030. Table 3.2-16 compares national
average total concentrations using 1999 versus scaled backgrounds.  More details are provided in
the technical document previously referenced.
  Table 3.2-16. National Average Total Concentrations (All Sources and Background) for
    2015, 2020, and 2030 using both the 1999 Background and the Scaled Backgrounds
HAP
1,3 -Butadiene
Acetaldehyde
Benzene
Formaldehyde
Xylenes
Total Concentrations (ug ni 3) using 1999
Background
2015
9.81X10'2
9.66X10"1
9.13X10'1
1.22
1.55
2020
9.77xlO"2
9.36X10"1
9.02X10"1
1.22
1.61
2030
l.OOxlO"1
9.56X10"1
9.24X10"1
1.25
1.65
Total Concentrations (jig m 3) using Scaled
Concentrations
2015
7.57xlO"2
7.77X10"1
7.57X10"1
9.56XKT1
1.50
2020
7.50xlO"2
7.47X10"1
7.40X10"1
9.68XKT1
1.56
2030
7.86xlO"2
7.78X10"1
7.71X10'1
1.01
1.60
                                         3-67

-------
       The largest impacts were in the Midwest as can be seen in Figure 3.2-18, which depicts
ratios of the ASPEN-modeled ambient benzene concentrations with an adjusted background
versus the 1999 background in 2020. Data tables with results of the sensitivity comparison by U.
S. County, along with maps of pollutant concentrations with and without an adjusted background
can be found in the docket for the rule.

       While accounting for impacts of emission reductions on  background levels would reduce
estimated population risks, it would increase estimated reductions in risk of control strategies in
a given year, since background levels would be reduced.  Also, if the modeling accounted for:
(1) near road impacts; (2) impacts of emissions from vehicles, equipment and fuels in attached
garages; (3) increased risks from early lifetime exposures; and (4) properly estimated cold start
emissions, estimated risks and risk reductions  from fuel benzene control would be larger.

      Figure 3.2-18. Ratios of Benzene Concentrations with and without an Adjusted
                                   Background, 2020
3.2.1.5. Perspective on Cancer Cases

       We have not quantified the cancer-related health benefits of expected MSAT reductions
in terms of avoided cancer cases or dollars.  The EPA Science Advisory Board (SAB)
specifically commented in their review of the 1996 National Air Toxics Assessment (NATA)
that these tools were not yet ready for use in a national-scale benefits analysis, because they did
not consider the full distribution of exposure and risk, or address sub-chronic health effects.157
While EPA has since improved many of these tools, there remain critical limitations for
estimating cancer incidence. For the MSATs of greatest concern, for example, we are currently
unable to estimate cessation lag, which is the time between reduction in exposure and decline in
                                         3-68

-------
risk to "steady state level."158 We have also not resolved the analytical challenges associated
with quantifying partial lifetime probabilities of cancer for different age groups or estimating
changes in survival rates over time. Indeed, some of these issues are likely to remain highly
uncertain for the foreseeable future.

       We can, however, present some perspective on how average individual risks could
translate into cumulative excess cancer cases across the U.S. population over a lifetime,
assuming continuous exposure at a given level for 70 years.  The following equation is used to
develop this estimate.

 Excess Cancer Cases = (Average Individual Cancer Risk] x (Population)

       To estimate annual incidence, this would be divided by 70. However, without knowing
when within a lifetime cancer is more likely to occur, and without accounting for time-varying
exposure, any estimate of incidence for a given calendar year is highly uncertain. We also note
that a proper calculation would entail the use of a life table of incidence rates within discrete age
ranges and a dose-response formulation expressing rate ratios as a function of benzene inhalation
exposure concentration.  Also, a proper calculation would require characterization of the full
distribution of exposure and risks. However, the modeling in this chapter estimates average
nationwide risk based on average census tract risks; thus, the full distribution has not been
characterized.  Also,  since census tracts vary in population, the average risk is not a population-
weighted average.

       In 2030, the cumulative excess average individual cancer risk from outdoor emissions  of
mobile source air toxics is estimated at 1.7xlO"5.  If the entire U. S. population (projected to be
about 364 million)159 were exposed to this level of risk over a 70-year lifetime, it would result in
about 6300 cancer cases, which translates into 90 annual cancer cases.

       In its review of the 1996 NAT A, SAB recommended that if cancer cases were calculated
for benefits assessment, a "best estimate" of risk (rather than an upper bound), should be used.
We believe that the maximum likelihood unit risk range for benzene represents a best estimate.
In our analyses, we have used the upper end of this range, as did the  1999 NATA. Following  is  a
discussion related to benzene specifically, including a discussion of the potential implications  of
the limitations of our national-scale modeling, which were noted  in Section 3.2.1.4.  As
discussed in Chapter 1, the current unit risk estimate for benzene may underestimate risk from
acute nonlymphocytic leukemia, because some recent epidemiology data, including key studies
published after the most recent IRIS assessment, suggest a supralinear rather than linear dose-
response at low doses. However, the studies published after the most recent IRIS assessment
have not yet been formally evaluated by EPA as part of the IRIS review process, and it is not
clear whether these data provide sufficient evidence to reject a linear dose-response curve. A
better understanding of the biological mechanism of benzene-induced leukemia is needed.

       In 2030, the national average inhalation individual cancer risk from outdoor mobile and
stationary sources of benzene, in the absence of the proposed benzene standard, is estimated at
approximately 6.6xlO"6, based on the modeling done for this rule. If the entire U. S. population
                                          3-69

-------
were exposed to that level of risk over a 70-year lifetime, it would result in approximately 34
excess cancer cases per year (Equation 1).

(1) Excess Cancer Cases at 2030 Exposure Level =
(Average Individual Cancer Risk] x (2030 Population)
  = 6.6xlO~6 x3.64x!08 =2402
  Annual Cancer Cases = 2402/70 = 34

However, this estimate does not include the higher estimates of benzene emissions from light-
duty vehicles at cold temperatures, higher gasoline distribution emissions, or portable fuel
container emissions developed for this rule. These revisions increase the total benzene inventory
from about 228,000 tons to 298,000 tons.  Assuming risks increase proportionally to the change
in the inventory, the estimated number of excess cases would be approximately 44 per year,
assuming continuous exposure to 2030 levels (equation 2).

(2) Adj. Excess Cancer Cases in 2030 =
      ,     , _      _        Adjusted Inventory
 Unadjusted Cancer Cases x
                           Unadjusted Inventory
  „ A   298,000 tons   A A
= 34x	= 44
       228,000 tons
       As discussed in Section 3.1.3.3, EPA's estimate of risk due to exposure to benzene could
increase significantly if the influence of attached garages were included. When the exposures for
people with attached garages are averaged across the population, average individual exposures to
benzene could increase by roughly 1 to 3 |ig/m3.160 This could result in about another 40 to 120
excess cancer cases (equation 3).

(3) Attached Garage Excess Cancer Cases =
(Average Exposure} x (Benzene URE) x (Population)
= (l-3 jug/m3)x (?.8xl 0~6 /jUg/m3)  x (3.64 x 108)= 2839 - 8518
Annual Cancer Cases = 41-122

Thus, including attached garages would increase the number of benzene-related excess cancer
cases to somewhere between 85 and 166 annually, assuming continuous exposure to 2030 levels.
This estimate would still not include higher exposure levels from near-road impacts,
occupational exposures, vapor emissions from leaking underground storage tanks,  or other
accidental releases into the environment. Any population risk characterization that does not
account for these factors underestimates the excess cancer related to benzene.

       The controls proposed in this rule reduce nationwide benzene emissions from all sources
in 2030 by about 22%, from 298,000 tons to 233,000 tons. That would reduce excess leukemia
cases due to benzene exposure, not including attached  garage exposures, by 10, from about 44
per year to 34 per year, assuming cancer risk decreases proportionally to emissions. This
                                         3-70

-------
assumption likely understates the distribution of benzene exposure reductions, as populations
with significant fractions of their daily activity on or near roadways are more likely to inhale
vehicle-related pollutants than a person living downwind.161'162  A roughly 40% reduction in fuel
benzene will reduce attached garage exposures by about 40% as well, reducing excess cancer
cases from this source of exposure by another 16 to 49 excess cancer cases.  Thus, this rule
would prevent roughly 26 to 59 benzene-related excess cancer cases annually, assuming
continuous lifetime exposure to 2030 levels, given the assumptions of population size and
lifetime above, and not including excess leukemia from near-road exposure, occupational
exposure, or from leaking underground storage tanks.  Emission reductions in 2030 would
reduce cancer cases not just in 2030, but also well beyond this period. There would also be
further unquantified reductions in incidence due to the other air toxics reductions.

Such estimates should be interpreted with extreme caution since they could imply an artificial
sense of precision. Serious limitations include:
    •  Geographically heterogeneous percentage emissions reductions  do not translate directly
       into changes in ambient levels, exposure, and risk.
    •  High and low end exposures are not fully characterized.
    •  The U.S. population would have experienced higher average exposures  in previous years,
       but this is not accounted for.
    •  Cessation lag between reduction in exposure and reduction in risk is not accounted for.
    •  Differences in risk among various age groups are not known, and the age structure of the
       U.S. population is expected to change over time.
    •  The current unit risk estimate may underestimate risk from acute nonlymphocytic
       leukemia, because recent epidemiology data, discussed in Chapter 1 of the Regulatory
       Impact Analysis, suggest a supralinear rather than linear dose-response  at low doses. As
       noted earlier, these data have not yet been formally evaluated by EPA as part of the IRIS
       review process.

3.2.2  Local-Scale Modeling

       Modeling at the national or regional scale, such the modeling done for the NATA
National-Scale Assessment described in Section 3.2.1, is designed to identify and prioritize air
toxics, emission source types and locations which are  of greatest potential concern in terms of
contributing to population risk. Such assessments also help elucidate patterns of exposure and
risk across broad geographic areas,  and can help characterize trends in air toxics risk and
potential impacts of controls at a broad geographic scale, as demonstrated above. However,
more localized assessments are needed to characterize and compare risks at local levels, and
identify potential "hotspots."

       National or regional-scale assessments typically rely on  a "top down" approach to
estimate emissions. Under a "top down" approach, emissions are  estimated at the county level,
typically starting from more aggregated information (e.g., state or national level) on activity.
Spatial surrogates are then used to allocate emissions to grid cells  or census tracts for modeling.
Use of more local data can greatly improve the characterization of the magnitude and distribution
of air toxic emissions. Air quality modeling can also be conducted with better spatial resolution
than is computationally feasible in a regional or national-scale assessment. As  a result, spatial

                                          3-71

-------
gradients of air toxic concentrations and locations where the highest risks are likely to occur can
be more accurately identified.

       Local-scale modeling is typically done using steady-state plume dispersion models, such
as the Integrated Source Complex (ISC) Model, the newly promulgated AERMOD (AMS/EPA
Regulatory Model), or non-steady-state puff models such as CALPUFF. These models have a
limited ability to simulate chemical reactions in the atmosphere.  As discussed in Section 3.2.1,
grid-based models, such as CMAQ, which better simulate chemical processes, do not yet have
the spatial resolution of dispersion models.  Significant advances are being made, however, in
combining features of grid-based models and plume/puff models. These advances are described
in a recent paper.163 A case study of diesel exhaust particulate matter in Wilmington, CA was
recently conducting employing some of these advances.164  The researchers combined Gaussian
and regional photochemical grid models. They found that local data, when modeled, provided a
much more refined picture of the magnitude and distribution of possible community "hot spots"
than more traditional, regional data, which rely on more default assumptions. An evaluation of
the approach determined that spatial allocation and emission rates contribute most to uncertainty
in model results, and this uncertainty could be substantially reduced through the collection and
integration of site specific information about the location of emission sources, and the activity
and emission rates of key sources affecting model concentrations.  They conclude that for
neighborhood assessments, incorporating site-specific data can lead to improvement in modeled
estimates of concentrations, especially where site-specific data are lacking in regulatory
databases.

       The Wilmington study discussed above also allocated motor vehicle emissions to
individual road "links," rather than using spatial surrogates to allocate county level vehicle
emissions to grid cells.  In using spatial surrogates to allocate emissions, high local
concentrations may not be captured for environments near major roadways, which are often
clustered in urban centers.  One local-scale assessment done in the Minneapolis-St. Paul area of
Minnesota, using such an inventory with the ISC model, found that the model tended to
overpredict at low monitored benzene concentrations and underpredict at high monitored
concentrations.165 Local-scale modeling using activity data for individual road links can better
characterize distributions of concentrations, and differentiate between  locations near roadways
and those further away, as observed in the following studies.

       As discussed in Section 3.1.3.2, local-scale modeling in Houston assigned emissions to
individual road links.166 Researchers at US EPA developed a methodology which utilized a
Geographic Information System (GIS) to allocate benzene emissions in Houston to major road
segments in an urban area and model the segments as elongated area sources. The Industrial
Source Complex Short Term (ISCST) dispersion model used both gridded and link-based
emissions to evaluate the effect of improved spatial allocation of emissions on ambient modeled
benzene concentrations. Allocating onroad mobile emissions to road segments improved the
agreement between modeled concentrations when compared with monitor observations, and also
resulted in higher estimated concentrations in the urban center where the density of
neighborhood streets is greater and the largest amount of traffic found. The calculated annual
average benzene model concentrations at monitor sites are compared to the observed annual
average concentrations in Figure 3.2-18. Most of the gridded model emissions show lower

                                          3-72

-------
benzene concentrations than both the link-based and observed monitor concentrations.
Allocating the onroad mobile emissions to road segments resulted in an increase in the average
benzene concentration, resulting in values that more closely match concentrations reported by
monitors.

    Figure 3.2-19. Model to Monitor Comparisons of Houston Benzene Concentrations
6.0-,
5.5-
KJV 5.0-
.5 4.5-
|L 4.0-
1 3.0-
1 2.5-
g 2.0-
.= 1.5-
1.0-
0.5-
o n-
ISCST3 eridded D ISCST3 lM;-ba?ed . Monitor
^
««"""»•"=•
];!4in"!"^
!^ni^!«

A * A
D
a
D



4 5 6
A/OM?rarj;

- :f'Hm'
C^!|™1*t'
iH^i.Trf


o




3

       Recent air quality modeling in Portland, OR using the CALPUFF dispersion model
assigned emissions to specific roadway links.167  The resulting data were used to develop a
regression model to approximate the CALPUFF predicted concentrations, determine the impacts
of roadway proximity on ambient concentration of three hazardous air pollutants (1,3-butadiene,
benzene, and diesel PM), and to estimate the zone of influence around roadways. Concentrations
were modeled at several distances from major roadways (0-50, 5-200, 200-400, and > 400
meters). For benzene, the resulting average concentrations were 1.29, 0.64, 0.40, and 0.12
ug/m3, respectively, illustrating the steep concentration gradient along roadways. There was a
zone of influence between 200 and 400 meters, with concentrations falling to urban background
levels beyond this distance.  The overall mean  motor vehicle benzene concentration modeled in
Portland was about 0.21 ug/m3, with concentrations increasing to 1.29 ug/m3 at model receptor
sites within 50 meters of a road. The results indicate that in order to capture localized impacts of
hazardous air pollutants in a dispersion model, there is a need to include individual roadway
links.

A recent review of local-scale modeling studies concluded that:168
       1) Significant variations in air toxic concentrations occurred across the cities, with
       highest concentrations occurring near the highest emitting sources, illustrating the need
       for modeling on a local scale.
       2) Increasing the receptor density near high emission sources  changes the location of
       maximum concentrations, illustrating the concentration gradients that can occur near high
       emission sources and the importance of receptor placement and density for model
                                          3-73

-------
       performance.
       3) Allocating on-road mobile emissions to road segments improved the agreement
       between modeled concentrations when compared with the observations, and also resulted
       in higher estimated concentrations in the urban center.
       4) It is important to refine the national emissions inventory for input into local air quality
       model applications.

       In another US EPA study, researchers provide a comparison of "top down" and "bottom
up" approaches to developing a motor vehicle emissions inventory for one urban area,
Philadelphia, in calendar year 1999.169  Under the "top down" approach, emissions were
estimated at the county level, typically  starting from more aggregated information.  Data on
vehicle miles traveled (VMT) in the metropolitan statistical area were allocated to counties using
population information.  Default national model inputs (e.g. fleet characteristics, vehicle speeds)
rather than local  data were also used. The "bottom up" approach utilizes vehicle activity data
from a travel demand model (TDM), and this "bottom up"  approach estimates emission rates
using more local input data to better estimate levels and spatial distribution of onroad motor
vehicle emissions.  TDM data can include information on the spatial distribution of vehicle
activity, speeds along those roads (which can have a large impact on emissions), and the
distribution of the VMT  among vehicle classes for different speed ranges.  These data can be
used to more accurately estimate the magnitude of toxic emissions at the local scale and where
they occur. Both the spatial distribution of emissions and the total county emissions in
Philadelphia differed significantly between the top-down and the bottom-up methodologies as
shown in Table 3.2-17.
                                          3-74

-------
  Table 3.2-17.  Comparison of Annual 1999 Benzene Emissions from Two Approaches in
                               Philadelphia Area Counties
County
Camden
Delaware
Gloucester
Montgomery
Philadelphia
Total
Local (TDM)
Based
165
162
110
333
255
1,025
National
(NEI)
210
160
104
209
467
1,150
Percent
Difference
-27%
1%
6%
59%
-45%
-12%
       In the case of Philadelphia, using local registration distribution data resulted in
significantly lower air toxics emission factors and resultant emissions.

       Local-scale modeling could also be improved by using local data on nonroad equipment
activity for lawn and garden, recreational,  construction and other sectors. EPA's county-level
inventories used in NATA and other modeling are developed using activity allocated from the
national or state level using surrogates.

       The use of more spatially refined emission inventories, in conjunction with other refined
air quality modeling techniques, improve the performance of air quality models.  They also
enable better characterization of the magnitude and distribution of air toxic emissions, exposure
and risk in urban areas, including risks associated with locations heavily impacted by mobile
sources.

       In conclusion, local scale modeling studies indicated higher concentrations of air toxics
than predicted by National scale analysis, particularly in near-source microenvironments such as
near roads. Thus, National scale analyses  such as 1999 NATA are likely underestimating high
end exposures and risks.

3.3 Ozone

       In this section we review the health and welfare effects of ozone. We also describe the
air quality monitoring and modeling data which indicate that people in many areas across the
country continue to be exposed to high levels of ambient ozone and will continue to be into the
future. Emissions of volatile organic compounds (VOCs) from the gas cans subject to this
proposed rule have been shown to contribute to these ozone concentrations.  Information on air
quality was gathered from a variety of sources, including monitored ozone concentrations, air
quality modeling forecasts conducted for this rulemaking, and other state and local air quality
information.

3.3.1   Science of Ozone Formation

       Ground-level ozone, the main ingredient in smog,  is formed by the reaction of VOCs and

                                          3-75

-------
nitrogen oxides (NOX) in the atmosphere in the presence of heat and sunlight. These pollutants,
often referred to as ozone precursors, are emitted by many types of pollution sources such as
highway and nonroad motor vehicles, gas cans, power plants, chemical plants, refineries, makers
of consumer and commercial products, industrial facilities, and smaller "area" sources.  VOCs
can also be emitted by natural sources such as vegetation.

       The science of ozone formation, transport, and accumulation is complex.170 Ground-level
ozone is produced and destroyed in a cyclical set of chemical reactions, many of which are
sensitive to temperature and sunlight. When ambient temperatures and sunlight levels remain
high for several days and the air is relatively stagnant, ozone and its precursors can build up and
result in more ozone than typically would occur on a single high-temperature day. Further
complicating matters, ozone also can be transported into an area from pollution sources found
hundreds of miles upwind, resulting in elevated ozone levels even  in areas with low VOC or NOX
emissions.  As a result, spatial and temporal differences in VOC and NOX emissions and weather
patterns contribute to daily, seasonal, and yearly differences in ozone concentrations across
different locations.

       The highest levels of ozone are produced when both VOC and NOX emissions are present
in significant quantities on clear summer days. Relatively small amounts of NOX enable ozone to
form rapidly when VOC levels are relatively high, but ozone production is quickly limited by
removal of the NOX.  Under these conditions NOX reductions are highly effective in reducing
ozone while VOC reductions have little effect. Such conditions are called "NOX -limited".
Because the contribution of VOC emissions from  biogenic (natural) sources to local ambient
ozone concentrations can be significant, even some areas where man-made VOC emissions  are
relatively low can be NOX -limited.

       When NOX levels are relatively high and VOC levels relatively low, NOX forms inorganic
nitrates (i.e., particles) but relatively little ozone.  Such conditions  are called "VOC-limited."
Under these conditions, VOC reductions are effective in reducing ozone, but NOX reductions can
actually increase local ozone under certain circumstances. Even in VOC-limited urban areas,
NOX reductions are not expected to increase ozone levels if the NOX reductions are sufficiently
large.

       Rural areas are usually NOX -limited, due to the relatively large amounts of biogenic
VOC emissions in many rural areas. Urban areas  can be either VOC- or NOX -limited, or a
mixture of both, in which ozone levels exhibit moderate sensitivity to changes in either pollutant.

       Ozone concentrations in an area also can be lowered by the reaction of nitric oxide with
ozone,  forming nitrogen dioxide (NO2); as the air  moves downwind and the cycle continues, the
NO2 forms additional ozone. The  importance of this reaction depends, in part, on the relative
concentrations of NOX, VOC, and ozone, all of which change with time and location.

       The Clean Air Act (CAA) requires EPA to set National Ambient Air Quality Standards
(NAAQS) for wide-spread pollutants from diverse sources considered harmful to public health
and the environment. The CAA established two types of NAAQS: primary standards to protect
public health, secondary standards to protect public welfare. The primary and secondary ozone

                                          3-76

-------
NAAQS are identical.  The 8-hour ozone standard is met when the 3-year average of the annual
4th highest daily maximum 8-hour ozone concentration is less than or equal to 0.08 ppm. (62 FR
38855, July 18, 1997)

3.3.2  Health Effects of Ozone

       Exposure to ambient ozone contributes to a wide range of adverse health effects. We are
relying on the data and conclusions in the 1996 ozone criteria document (CD) and ozone staff
paper, which reflect EPA's analysis of policy-relevant science from the ozone CD, regarding the
health effects associated with ozone.171'172  In August 2005, the EPA released the second external
review draft of a new ozone  CD which is scheduled to be released in final form in February
2006.173  The new ozone criteria document summarizes the findings of the 1996 ozone CD and
critically assesses relevant new scientific information that has emerged in the past decade.  In all,
the new epidemiological studies that have become available since the 1996 ozone CD continue to
demonstrate the harmful effects of ozone on public health, and the need to attain and maintain
the ozone NAAQS.

       Ozone-related health effects include lung function decrements, respiratory symptoms,
aggravation of asthma, increased hospital and emergency room visits, increased medication
usage, inflammation of lung tissues, as well as a variety of other respiratory effects. People who
are particularly at risk for high ozone exposures include healthy children and adults who are
active outdoors. Susceptible subgroups include children, people with respiratory disease, such as
asthma, and people with unusual sensitivity to ozone.174'175'176'177'178

       Based on a large number of scientific studies, EPA has identified several key health
effects associated with human exposure to levels of ozone found today in many areas of the
country.  Short-term (1 to 3 hours) and prolonged exposures (6 to 8 hours) to higher ambient
ozone concentrations have been linked to lung function decrements, respiratory symptoms,
increased hospital admissions and emergency room visits for respiratory
problems.179'180'181'182'183'184'185'186'187'188'189'190'191'192 Repeated exposure to ozone can make people
more susceptible to lung inflammation and can aggravate preexisting respiratory diseases, such
as asthma.193'194'195'196'197'198'199'200'201  Repeated exposure to ozone can also cause inflammation of
the lung, impairment of lung defense mechanisms, and possibly irreversible changes in lung
structure.202'203'204'205'206'207

       Children and adults who are outdoors and active during the summer months, such as
construction workers and other outdoor workers, are among those most at risk of elevated ozone
exposures.208'209'210'211'212 Specifically, children and outdoor workers are most at risk from ozone
exposure because they  typically are active outside, working, playing and exercising, during the
summer when ozone levels are highest.213'214'215'216  These individuals, as well as people with
respiratory illnesses such as  asthma, especially asthmatic children, can experience reduced lung
function and increased respiratory symptoms, such as pain on deep inspiration and cough, when
exposed to relatively low ozone levels during prolonged periods of moderate exertion.217 For
example, summer camp studies in the Eastern United States and  Southeastern Canada have
reported significant reductions in lung function in children who are active
outdoors.218'219'220'221'222'223'224'225'226 Further,  children are more at risk of experiencing health

                                           3-77

-------
effects from ozone exposure than adults because their respiratory systems are still developing.

       There has been new research that suggests additional serious health effects beyond those
that had been known when the 1996 ozone CD was published.  Since then, over 1,700 new
ozone-related health and welfare studies have been published in peer-reviewed journals.227
Many of these studies have investigated the impact of ozone exposure on such health effects as
changes in lung structure and biochemistry, inflammation of the lungs, exacerbation and
causation of asthma, respiratory illness-related school absence, hospital and emergency room
visits for asthma and other respiratory causes,  and premature mortality.  EPA is currently in the
process of evaluating these  and other studies as part of the ongoing review of the criteria
document and NAAQS for ozone. Key new health information falls into four general areas:
development of new-onset asthma, hospital admissions for young children, school absence rate,
and premature mortality. Examples of new studies in these areas are briefly discussed below.

       Aggravation of existing asthma resulting from short-term ambient ozone exposure was
reported prior to the 1997 ozone NAAQS revision and has been observed in studies published
since then 228>229>230>231>232 More recent studies  now suggest the potential for a relationship
between long-term ambient ozone concentrations and the incidence of new-onset asthma.  In
particular, such a relationship in adult males (but not in females) was reported by McDonnell et
al. (1999).233 Subsequently, McConnell et al. (2002) reported that incidence of new diagnoses of
asthma in  children is associated with heavy exercise in communities with high ambient ozone
concentrations (i.e., mean 8-hour concentration of 59.6 ppb or greater of ozone).234  This
relationship was documented in children who played 3 or more sports and thus spent more time
outdoors.  It was not documented for those children who played one or two sports.0 The larger
effect of high activity  sports than low activity sports and an independent effect of time spent
outdoors also in the higher ozone communities strengthens the inference that exposure to ozone
may modify the effect of sports on the development of asthma in some children.

       Previous studies have shown relationships between ozone and hospital admissions in the
general population. More recently there have been studies that report the effects of ozone on
unscheduled respiratory hospital admissions of children.235'236'237'238 A study in Toronto reported
a significant relationship between 1-hour maximum ozone concentrations and respiratory
hospital admissions in children under the age of two.239 Given the relative vulnerability of
children in this age category, there is particular concern about these findings from the literature
on ozone and hospital admissions.

       Increased rates of illness-related school absenteeism have been associated with 1-hour
daily maximum240 and 8-hour average ozone concentrations.241'242  In a study by Chen and
colleagues (2000), daily school absenteeism was examined in 27,793 students (kindergarten to
sixth grade) from 57 elementary schools in Washoe County, NV over a two-year period.243 In
models adjusting for PMio and CO, ambient ozone levels were found to be associated with
school absenteeism. Ozone-related school absences were also examined in a study of 1,933
fourth grade students from 12 southern California communities participating in the Children's
Health Study.244 Due to the comprehensive characterization of health outcomes, this study is
c In communities with mean 8-hour ozone concentration of 59.6 ppb, the relative risk of developing asthma in
children playing three or more sports was 3.3 (95% CI 1.9 - 5.8) compared with children playing no sports.
                                          3-78

-------
valuable in assessing the effect of ozoneon illness-related school absenteeism in children. The
study spanned the months of January through June 1996, which captured a wide range of
exposures while staying mostly below the highest levels observed in the summer season.  Larger
ozone effects were seen for respiratory causes than for nonrespiratory causes. Park et al. (2002)
examined the association between air pollution and school absenteeism in 1,264 students, first to
sixth grade, attending school in Seoul, Korea.245  The study period extended from March 1996 to
December 1999, with 8-hour average ozone concentrations ranging from 3.13 ppb to 69.15 ppb
(mean 22.86 ppb). Same day ozone concentrations were positively associated with illness-related
absences,  but inversely associated with non-illness-related absences. These studies reported that
ambient ozone concentrations, on the same day as well as accumulated over two to four weeks,
are associated with school absenteeism, particularly illness-related absences.

       The air pollutant most clearly associated with premature mortality is PM, with many
studies reporting such an association. However, recent  studies have reported statistically
significant associations between ozone exposure and premature mortality. Key findings are
available from a multi-city time-series study that reports associations between ozone and
mortality based on analyses using data from the 90 U.S. cities in the original National Mortality,
Morbidity and Air Pollution (NMMAPS) study246'247 and from 95 U.S. cities in an extension to
the NMMAPS analyses248, and further analyses using a subset of 19 U.S. cities and focusing on
cause-specific mortality associations249.  An additional study used case-crossover design and data
from  14 U.S. cities, to further investigate the influence of adjustment for weather variables in  the
ozone-mortality relationship.250 Finally, results are available from a European study, Air
Pollution and Health: a European Approach (APHEA), an analysis using data from 23 cities and
4 cities.251'252

       In addition, several meta-analyses have been conducted on the relationship between Os
and mortality. These analyses reported fairly consistent and positive combined effect estimates
for an increase in mortality for a  standardized change in 63. Three recent meta-analyses
evaluated potential sources of heterogeneity in  ozone-mortality associations.253'254'255 Common
findings were observed across all three analyses, in that all reported that effect estimates were
larger in warm season analyses, reanalysis of results using default GAM criteria did not change
the effect estimates, and there was no strong evidence of confounding by PM.  Bell et al. (2005)
and Ito et  al. (2005) both provided suggestive evidence of publication bias, but ozone-mortality
associations remained after accounting for that potential bias.  These studies "provide strong
evidence that ozone is associated with mortality." This discussion is drawn from the second draft
of the ozone criteria document. EPA is in the process of finalizing the ozone criteria document
and the  discussion in the final rule will reflect the final ozone criteria document.

       There is a substantial amount of recent experimental evidence that links ozone exposure
with respiratory effects in laboratory animals and humans. These include structural changes in
the bronchiolar-alveolar transition (centriacinar) region of the lung, biochemical evidence of
acute cellular/tissue injury, inflammation, increased frequency and severity of experimental
bacterial infection, and temporary reductions in mechanical lung function. The data linking
ozone exposure with respiratory effects have been observed with exposure to ozone at ambient or
near-ambient concentrations. Thus, many of the reported epidemiologic associations of ambient
ozone with respiratory health effects have considerable biological credibility. Accordingly, the

                                          3-79

-------
new epidemiologic studies of ambient ozone discussed here are best considered in combination
with information on ambient ozone concentration and exposure, and toxicological effects of
ozone in animals and humans.  This discussion is drawn from the second draft of the ozone
criteria document. EPA is in the process of finalizing the ozone criteria document and the
discussion in the final rule will reflect the final ozone criteria document.

3.3.3  Current 8-Hour Ozone Levels

       The proposed gas can emission reductions would assist 8-hour ozone nonattainment areas
in reaching the standard by each area's respective attainment date and assist 8-hour ozone
maintenance areas in maintaining the 8-hour ozone standard in the future.  In this section and the
next section we present information on current and model-projected future 8-hour ozone levels.

       A nonattainment area is defined in the CAA as an area that is violating a NAAQS or is
contributing to a nearby area that is violating the NAAQS. EPA designated nonattainment areas
for the 8-hour ozone NAAQS in June 2004.  The final rule on Air Quality Designations and
Classifications for the 8-hour Ozone NAAQS (69 FR 23858, April 30, 2004) lays out the factors
that EPA considered in making the 8-hour ozone nonattainment designations, including 2001-
2003 measured data, air quality in adjacent areas, and other factors.0

       According to EPA's designations, as of September 29, 2005, approximately 159 million
people live in the 126 areas that are currently designated as nonattainment for either failing to
meet the 8-hour ozone NAAQS or for contributing to poor air quality in a nearby area.  There are
474 full or partial counties that make up the 126 8-hour ozone nonattainment areas.  Figure 3.3-1
illustrates the widespread nature of these problems.  Shown in this figure are counties designated
as nonattainment for the 8-hour ozone NAAQS, also depicted are PM2 5 nonattainment areas and
the Mandatory Class I Federal Areas.  The 8-hour ozone nonattainment areas, nonattainment
counties and populations are listed in Appendix 3 A to this RIA.
D An ozone design value is the concentration that determines whether a monitoring site meets the NAAQS for
ozone. Because of the way they are defined, design values are determined based on three consecutive-year
monitoring periods.  For example, an 8-hour design value is the fourth highest daily maximum 8-hour average ozone
concentration measured over a three-year period at a given monitor.  The full details of these determinations
(including accounting for missing values and other complexities) are given in Appendices H and I of 40 CFR Part
50.  Due to the precision with which the standards are expressed (0.08 parts per million (ppm) for the 8-hour), a
violation of the 8-hour standard is defined as a design value greater than or equal to 0.085 ppm or 85 parts per
billion (ppb). For a county, the design value is the highest design value from among all the monitors with valid
design values within that county. If a county does not contain an ozone monitor, it does not have a design value.
Thus, our analysis may underestimate the number of counties with design values above the level of NAAQS.

                                            3-80

-------
   Figure 3.3.-1. 8-Hour Ozone and PM2.s Nonattainment Areas and Mandatory Class I
                                       Federal Areas
     ^B Both PM 2.5 and Ozone NA Counties
     |   | 8-hr Ozone NA Counties Only
     |   | PM 2.5 NA Counties Only
       H Federal Class 1 Areas
    8-hr Ozone Nonattainment Counties (474) from 4D CFR Part B1 as of 3/5/05
(Some of these counties are in Ozone Early Action CompactAreas with future effective dates)
    PM2.5 Nonattainment Counties (209) from April 5. 2005 Effective Date (incl DC)
          Nonattainment areas include whole and partial counties
       Counties designated as 8-hour ozone nonattainment were classified, on the basis of their
one-hour ozone design value, as Subpart 1 or Subpart 2 (69 FR 23951, April 30, 2004).  Areas
classified as Subpart 2 were then further classified, on the basis of their 8-hour ozone design
value, as marginal, moderate, serious, severe or extreme.  The maximum attainment date
assigned to an ozone nonattainment area is based on the area's classification.

       Table 3A-1 presents the 8-hour ozone nonattainment areas, their 8-hour design values,
their category or classification and their maximum attainment date.  States with 8-hour ozone
nonattainment areas are required to take action to bring those areas into compliance in the future.
Based on the final rule designating and classifying 8-hour ozone nonattainment areas, most 8-
hour ozone nonattainment areas will be required to attain the 8-hour ozone NAAQS in the 2007
to 2014 time frame and then be required to maintain the 8-hour ozone NAAQS thereafter.E The
gas can emission standards being proposed in this action would become effective  in 2009. Thus,
the expected ozone precursor emission inventory reductions from the standards proposed in this
 ' The Los Angeles Southcoast Air Basin 8-hour ozone nonattainment area will have to attain before June 15, 2021.
                                             3-81

-------
action would be useful to States in attaining and/or maintaining the 8-hour ozone NAAQS.

3.3.4   Projected 8-Hour Ozone Levels

       Recent air quality modeling predicts that without additional local, regional or national
controls there will continue to be a need for reductions in 8-hour ozone concentrations in some
areas in the future. In the following sections we describe recent ozone air quality modeling from
the CAIR analysis as well as results of the ozone response surface metamodel (RSM) analysis
we completed to assess the potential ozone impacts resulting from the VOC emissions controls
proposed for gas cans.

3.3.4.1 CAIR Ozone Air Quality Modeling

       Recently ozone air quality analyses were performed for the Clean Air Interstate Rule
(CAIR), which was promulgated by EPA in 2005.  The Comprehensive Air Quality Model with
Extension (CAMx) was used as the tool for simulating base and future year concentrations of
ozone in support of the CAIR ozone air quality assessment.  The CAIR analysis included all final
federal rules up to and including CAIR controls. Details on the air quality modeling are
provided in the Air Quality Modeling Technical Support Document for the Final Clean Air
Interstate Rule, included in the docket for this proposed rule.256

       Air quality modeling performed for CAIR indicates that in the absence of additional
controls, counties with projected 8-hour ozone concentrations greater than or equal to 85 ppb are
likely to persist in the future.  The CAIR analysis provided estimates of future ozone levels
across the country. For example, in 2010, in the absence of controls beyond those relied on for
the CAIR modeling, we project that 24 million people would live in 37 Eastern counties with 8-
hour ozone concentrations at and above 85 ppb, see Table 3.3-l.F  Table 3.3-1 also lists the 148
Eastern counties, where 61 million people are projected to live, with 2010 projected design
values that do not violate the 8-hour ozone NAAQS but are within ten percent of it, in the
absence of emission reductions beyond those considered in the CAIR modeling. These are
counties that are not projected to violate the standard, but to be close to it. The proposed rule
may help ensure that these counties continue to maintain their attainment status and the emission
reductions from this proposed rule would be included by the states in their baseline inventory
modeling for their ozone maintenance plans.
                                          3-82

-------
Table 3.3-1. Eastern Counties with 2010 projected 8-hour Ozone Concentrations
                 Above and within 10% of the 8-hour Ozone Standard
State
Arkansas
Connecticut
Connecticut
Connecticut
Connecticut
Connecticut
Connecticut
D.C.
Delaware
Delaware
Delaware
Georgia
Georgia
Georgia
Georgia
Georgia
Georgia
Georgia
Georgia
Georgia
Illinois
Illinois
Illinois
Illinois
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Kentucky
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
2010 Projected
8-hour Ozone
County Concentration (ppb)a 2000 popb
Crittenden Co
Fairfield Co
Hartford Co
Middlesex Co
New Haven Co
New London Co
Tolland Co
Washington Co
Kent Co
New Castle Co
Sussex Co
Bibb Co
Cobb Co
Coweta Co
De Kalb Co
Douglas Co
Fayette Co
Fulton Co
Henry Co
Rockdale Co
Cook Co
Jersey Co
Lake Co
McHenry Co
Boone Co
Clark Co
Hamilton Co
Hancock Co
La Porte Co
Lake Co
Madison Co
Marion Co
Porter Co
Shelby Co
St Joseph Co
Campbell Co
Bossier Parish
East Baton Rouge Parish
Iberville Parish
Jefferson Parish
Livingston Parish
West Baton Rouge Parish
80.8
92.2
80.1
90.6
91.3
83.4
82.7
85.0
78.7
84.7
80.3
80.0
79.4
76.6
81.9
78.7
76.7
85.1
80.3
80.4
81.8
77.0
76.8
76.6
78.1
78.4
81.7
80.4
81.8
82.8
78.6
79.6
81.1
81.6
77.8
81.5
77.0
80.6
79.4
78.6
77.8
78.8
50,866
882,567
857,183
155,071
824,008
259,088
136,364
572,058
126,697
500,264
156,638
153,887
607,750
89,215
665,864
92,174
91,263
816,005
119,341
70,111
5,376,739
21,668
644,356
260,077
46,107
96,472
182,740
55,391
110,106
484,563
133,358
860,453
146,798
43,445
265,559
88,616
98,310
412,852
33,320
455,466
91,814
21,601
2010 pop0
52,889
891,694
859,080
164,202
829,181
267,199
142,988
554,474
139,376
534,631
181,962
158,291
744,488
111,522
698,335
114,380
117,580
855,826
153,957
87,977
5,363,464
22,905
731,690
307,400
54,035
107,096
230,565
65,282
111,566
489,220
137,710
879,932
165,350
46,565
275,031
92,109
110,838
465,411
33,089
493,359
124,895
22,672
                                       3-83

-------
State
Maine
Maine
Maryland
Maryland
Maryland
Maryland
Maryland
Maryland
Maryland
Maryland
Maryland
Maryland
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Missouri
Missouri
Missouri
Missouri
Missouri
New Hampshire
New Jersey
New Jersey
New Jersey
New Jersey
New Jersey
New Jersey
New Jersey
2010 Projected
8-hour Ozone
County Concentration (ppb)a 2000 popb
Hancock Co
York Co
Anne Arundel Co
Baltimore Co
Carroll Co
Cecil Co
Charles Co
Frederick Co
Harford Co
Kent Co
Montgomery Co
Prince Georges Co
Barnstable Co
Bristol Co
Essex Co
Hampden Co
Hampshire Co
Middlesex Co
Suffolk Co
Allegan Co
Benzie Co
Berrien Co
Cass Co
Genesee Co
Macomb Co
Mason Co
Muskegon Co
Oakland Co
Ottawa Co
St Clair Co
Washtenaw Co
Wayne Co
Clay Co
Jefferson Co
St Charles Co
St Louis City
St Louis Co
Hillsborough Co
Atlantic Co
Bergen Co
Camden Co
Cumberland Co
Gloucester Co
Hudson Co
Hunterdon Co
80.5
80.2
88.6
83.7
80.0
89.5
78.7
78.1
92.8
85.8
79.3
84.2
83.6
83.0
81.7
80.2
78.0
79.1
78.1
82.1
77.9
78.1
78.2
76.7
85.4
78.9
82.0
80.7
76.6
80.6
81.0
84.7
76.5
76.7
80.5
79.4
80.5
76.6
80.4
86.0
91.6
84.4
91.3
84.3
88.6
51,791
186,742
489,656
754,292
150,897
85,951
120,546
195,277
218,590
19,197
873,341
801,515
222,230
534,678
723,419
456,228
152,251
1,465,396
689,807
105,665
15,998
162,453
51,104
436,141
788,149
28,274
170,200
1,194,155
238,314
164,235
322,895
2,061,161
184,006
198,099
283,883
348,188
1,016,315
380,841
252,552
884,118
508,932
146,438
254,673
608,975
121,989
2010 pop0
53,886
201,082
543,785
792,284
179,918
96,574
145,763
234,304
268,207
20,233
940,126
842,221
249,495
558,460
747,556
452,718
158,130
1,486,428
674,179
121,415
17,849
164,727
53,544
441,196
838,353
30,667
175,901
1,299,592
277,400
178,391
344,398
1,964,209
213,643
230,539
341,686
324,156
1,024,964
412,071
269,754
898,450
509,912
149,595
278,612
607,256
139,641
3-84

-------
State
New Jersey
New Jersey
New Jersey
New Jersey
New Jersey
New Jersey
New York
New York
New York
New York
New York
New York
New York
New York
New York
New York
New York
New York
New York
New York
North Carolina
North Carolina
North Carolina
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Oklahoma
Pennsylvania
Pennsylvania
2010 Projected
8-hour Ozone
County Concentration (ppb)a 2000 popb
Mercer Co
Middlesex Co
Monmouth Co
Morris Co
Ocean Co
Passaic Co
Bronx Co
Chautauqua Co
Dutchess Co
Erie Co
Essex Co
Jefferson Co
Monroe Co
Niagara Co
Orange Co
Putnam Co
Queens Co
Richmond Co
Suffolk Co
Westchester Co
Mecklenburg Co
Rowan Co
Wake Co
Allen Co
Ashtabula Co
Butler Co
Clermont Co
Clinton Co
Cuyahoga Co
Delaware Co
Franklin Co
Geauga Co
Hamilton Co
Knox Co
Lake Co
Lorain Co
Lucas Co
Medina Co
Portage Co
Summit Co
Trumbull Co
Warren Co
Wood Co
Tulsa Co
Allegheny Co
Armstrong Co
95.2
92.1
86.4
85.5
100.3
79.7
79.7
81.8
81.0
86.9
77.6
80.5
76.9
82.3
77.1
82.3
78.3
87.1
90.8
84.7
81.4
80.1
77.2
76.8
83.5
78.0
78.0
81.4
77.3
77.3
81.9
86.6
78.6
76.5
82.2
78.5
80.0
76.5
79.8
82.4
79.7
80.0
77.4
79.2
81.9
79.7
350,761
750,162
615,301
470,212
510,916
489,049
1,332,649
139,750
280,150
950,265
38,851
111,738
735,343
219,846
341,367
95,745
2,229,379
443,728
1,419,369
923,459
695,453
130,340
627,846
108,473
102,728
332,806
177,977
40,543
1,393,977
109,989
1,068,977
90,895
845,302
54,500
227,511
284,664
455,053
151,095
152,061
542,898
225,116
158,383
121,065
563,299
1,281,665
72,392
2010 pop0
359,912
805,537
670,971
500,033
572,364
495,610
1,298,206
139,909
291,098
953,085
39,545
113,075
745,350
220,407
371,434
107,967
2,239,026
488,728
1,472,127
944,535
814,088
143,729
787,707
106,900
104,850
384,410
205,365
47,137
1,348,313
136,125
1,142,894
102,083
843,226
59,435
237,161
292,040
447,302
173,985
162,685
552,567
226,157
186,219
129,124
610,536
1,259,040
72,829
3-85

-------
State
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Rhode Island
Rhode Island
Rhode Island
South Carolina
Tennessee
Tennessee
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
2010 Projected
8-hour Ozone
County Concentration (ppb)a 2000 popb
Beaver Co
Berks Co
Bucks Co
Cambria Co
Chester Co
Dauphin Co
Delaware Co
Erie Co
Franklin Co
Lancaster Co
Lehigh Co
Mercer Co
Montgomery Co
Northampton Co
Philadelphia Co
Washington Co
Westmoreland Co
York Co
Kent Co
Providence Co
Washington Co
Richland Co
Sevier Co
Shelby Co
Brazoria Co
Collin Co
Dallas Co
Denton Co
Galveston Co
Gregg Co
Harris Co
Jefferson Co
Johnson Co
Montgomery Co
Tarrant Co
Alexandria City
Arlington Co
Charles City Co
Fairfax Co
Hampton City
Hanover Co
Henrico Co
Loudoun Co
Suffolk City
79.6
81.7
94.3
76.9
85.4
80.8
84.0
79.1
80.2
83.6
82.1
78.1
87.6
81.8
89.9
77.3
76.7
79.4
86.2
81.2
84.2
76.9
76.5
76.7
84.1
82.5
82.2
86.8
84.6
79.1
97.4
85.0
78.2
81.2
87.2
80.9
86.0
77.7
85.4
78.7
80.9
78.2
78.6
77.5
181,412
373,637
597,635
152,598
433,501
251,798
550,863
280,843
129,313
470,657
312,090
120,293
750,097
267,066
1,517,549
202,897
369,993
381,750
167,090
621,602
123,546
320,677
71,170
897,471
241,767
491,675
2,218,899
432,976
250,158
111,379
3,400,577
252,051
126,811
293,768
1,446,219
128,283
189,453
6,926
969,749
146,437
86,320
262,300
169,599
63,677
2010 pop0
183,693
388,194
648,796
146,811
478,460
265,019
543,169
284,835
135,088
513,684
323,215
122,546
772,849
279,797
1,420,803
205,153
372,941
404,807
174,126
621,355
137,756
349,826
96,097
958,501
281,960
677,868
2,382,657
554,033
283,963
121,241
3,770,129
260,847
157,545
413,048
1,710,920
130,422
193,370
7,382
1,085,483
153,246
98,586
294,174
214,469
69,003
3-86

-------
2010 Projected
8-hour Ozone
State County Concentration (ppb)a 2000 popb
Wisconsin Door Co
Wisconsin Kenosha Co
Wisconsin Kewaunee Co
Wisconsin Manitowoc Co
Wisconsin Milwaukee Co
Wisconsin Ozaukee Co
Wisconsin Racine Co
Wisconsin Sheboygan Co
Number of Violating Counties
Population of Violating Counties
Number of Counties within 10%
Population of Counties within 10%
82.1
91.0
79.9
80.0
82.1
85.8
83.9
87.7
37
148

27,961
149,577
20,187
82,887
940,164
82,317
188,831
112,646
22,724,010
58,453,962
2010 pop0
30,508
166,359
20,538
83,516
922,943
95,549
199,178
118,866
24,264,574
61,409,062
a) Bolded concentrations indicate levels above the 8-hour ozone standard.
b) Populations are based on 2000 census data.
c) Populations are based on 2000 census projections.

3.3.4.2 Ozone Response Surface Metamodel Methodology

       We performed ozone air quality modeling simulations for the Eastern United States using
the ozone RSM.  The ozone RSM is a screening-level air quality modeling tool that allows users
to  quickly assess the estimated air quality changes over the modeling domain. The ozone RSM
is a model of a full-scale air quality model and is based on statistical relationships between
model inputs and outputs obtained from the full-scale air quality model.  In other words, the
ozone RSM uses statistical techniques to relate a response variable to a set of factors that are of
interest, e.g., emissions of precursor pollutants from particular sources and locations. The
following section describes the modeling methodology, including the development of the multi-
dimensional experimental design for control strategies and implementation and verification of
the RSM technique. Additional detail is available in the Air Quality Modeling Technical
Support Document (AQMTSD) for this proposal.257

       The foundation for  the ozone response surface metamodeling analyses was the CAMx
modeling done in support of the final Clean Air Interstate Rule (CAIR).  The CAIR modeling is
fully described in the CAIR Air Quality Modeling Technical Support Document, but a brief
description is provided below.258  The modeling procedures used in the CAIR analysis (e.g.,
domain, episodes, meteorology) have been used for several EPA rulemaking analyses over the
past five years and are well-established at this point.

       The ozone RSM uses the 2015 controlled CAIR emissions inventory as its baseline.259
This inventory does not include the gas can emissions that are being controlled in this proposed
action. The uncontrolled and controlled gas can emissions have been incorporated into the base
and control runs  of the ozone RSM (see Section 2.1 for more detail about the gas can emissions
inventory). The inventory  also does not include the higher estimates of cold temperature
emissions for gasoline vehicles developed for this rule; however, these emissions are not likely
to  have a significant impact on ozone formation. Because the base years of our air quality
                                          3-87

-------
modeling analysis are 2020 and 2030, we extrapolate the model from 2015 to 2020 and 2030.
Additional detail on how the model was extrapolated to reflect gas can emissions and various
projection years is included in the AQMTSD for this proposal.260

       The modeling simulations that comprised the metamodeling were conducted using
CAMx version 3.10. It should be noted that because the ozone RSM is built from CAMx air
quality model runs, it therefore has the same strengths and limitations of the underlying model
and its inputs.  CAMx is a non-proprietary computer model that simulates the formation and fate
of photochemical oxidants including ozone for given input sets of meteorological conditions and
emissions.  The gridded meteorological data for three historical episodes were developed using
the Regional Atmospheric Modeling System (RAMS),  version 3b.261  In all, 30 episode days
were modeled using frequently-occurring, ozone-conducive, meteorological conditions from the
summer of 1995. Emissions estimates were developed for the evaluation year (1995) as well as a
future year (2015).

       The CAMx model applications were performed for a domain covering all, or portions of,
37 States (and the District of Columbia) in the Eastern U.S., as shown in Figure 3.3-2. The
domain has nested horizontal grids of 36 km and 12 km.  However, the output data from the
metamodeling is  provided at a 12 km resolution (i.e., cells from the outer 36 km cells populate
the nine finer scale cells, as appropriate).  Although the domain of the ozone RSM is the 37
Eastern states, the gas can controls are a nationwide program.  Section 2.1.3 describes the
nationwide inventory reductions that could be achieved by the proposed gas can controls.
Section 2.1.1.2 also details the states that  have their own gas can control programs and how the
controls proposed here impact states which already have gas can control programs.

-------
      Figure 3.3-2. Map of the CAMx Domain used for MSAT Ozone Metamodeling
       The ozone RSM used for assessing the impacts of proposed gas can emission reductions
was developed broadly to look at various control strategies with respect to attaining the 8-hour
ozone NAAQS. The experimental design for the ozone RSM covered three key areas: type of
precursor emission (NOx or VOC), emission source type (i.e., onroad vehicles, nonroad vehicles,
area sources, electrical generating utility (EGU) sources, and non-utility point sources), and
location in or out of a 2015 model-projected residual ozone nonattainment area. This resulted in
a set of 14 emissions factors. Since some of the spillage emissions associated with gas cans are
currently included in the NONROAD emissions model, for the purposes of the ozone RSM we
have included gas can emissions as part of the nonroad factor in our air quality modeling.

       The 14 emission factors were randomly varied and used as inputs to CAMx.  The
experimental design for these 14 factors was developed using a Maximin Latin Hypercube
method. Based on a rule of thumb of 10 runs per factor, we developed an overall design with
140 runs (a base case plus 139 control runs).  The range of emissions reductions considered
within the metamodel ranged from 0 to 120 percent of the 2015 CAIR emissions.  This
experimental design resulted in a set of CAMx simulations that serve as the inputs to the ozone
response surface metamodel. Because the metamodeling was going to be used to assess the
impacts of the proposed gas can standards, the experimental design also included oversampling
                                         3-89

-------
in the range of 0 to 10 percent control for the nonroad VOC sector, as well as CAMx runs that
only included VOC controls.

       To develop a response surface approximation to CAMx, we used a multidimensional
kriging approach, implemented through the MIXED procedure in SAS.  We modeled the
predicted changes in ozone in each CAMx grid cell as a function of the weighted average of the
modeled responses in the experimental design.  A response-surface was then fit for the ozone
design value metric. Validation was performed and is summarized in the AQMTSD. The
validation exercises indicated that the ozone RSM replicates CAMx response to emissions
changes very well for most emissions combinations and in most locations.

       The assessment of proposed gas can controls conducted for this analysis involved
adjusting the nonroad mobile source VOC  emissions both in and out of ozone nonattainment
areas and looking at the impact on the 8-hour ozone design value metric. We created an input or
adjustment factor for the nonroad mobile source VOC emission factor by adding future year gas
can emission estimates to the projected CAIR emission inventory and then relating the future
year emissions estimate to 2015. For this assessment the future years modeled are 2020 and
2030.

3.3.4.3 Ozone Response Surface Metamodel Results

       This section summarizes the results of our modeling of ozone air quality impacts in the
future with and without the proposed reductions in gas can emissions. Based upon our previous
CAIR air quality modeling, we anticipate that without emission reductions beyond those already
required under promulgated regulations and approved SIPs, ozone nonattainment will likely
persist into the future.

       The inventories that underlie the  ozone modeling conducted for this rulemaking included
emission reductions from all current or committed federal, state, and local controls, including the
recent CAIR. There was no attempt to examine the prospects of areas attaining or maintaining
the 8-hour ozone standard with possible  additional future controls  (i.e., controls beyond current
or committed federal, State,  and local controls).

       According to the ozone response surface metamodel (RSM), the proposed gas can
controls are projected to result in a very  small population-weighted net improvement in future
ozone. The net improvement is generally so small as to be rendered insignificant when
presenting design values.  The model changes are smaller than the precision with which the
ozone standard is expressed (0.08 parts per million (ppm)) and to which 8-hour ozone data is
reported.0 Nonetheless, there are some areas where the ozone improvement is more significant.
These areas include Chicago, Milwaukee, Detroit and New York City. It is also important to
note that the ozone RSM results indicate that the counties which are projected to experience the
greatest improvement in ozone design values are generally also those that are projected to have
the highest ozone design values.  Those counties that are projected to experience an extremely
small increase in ozone design values generally have design values that are lower, below 70 ppb.
The results from the metamodeling projections indicate a net overall  improvement in future 8-
 Appendix I of 40 CFR Part 50.
                                         3-90

-------
hour ozone design values due to the proposed gas can controls, when weighted by population.
The AQMTSD, contained in the docket for this proposal, includes additional detail on the ozone
RSM results.

3.3.5   Environmental Effects of Ozone Pollution

       There are a number of public welfare effects associated with the presence of ozone in the
           r\/:rj
ambient air.   In this section we discuss the impact of ozone on plants, including trees,
agronomic crops and urban ornamentals.

3.3.5.1 Impacts  on Vegetation

       The Air Quality Criteria Document for Ozone and related Photochemical Oxidants notes
that "ozone affects vegetation throughout the United States, impairing crops, native vegetation,
and ecosystems more than any other air pollutant."263 Like carbon dioxide (CO2) and other
gaseous substances, ozone enters plant tissues primarily through apertures (stomata) in leaves in
a process called "uptake." To a lesser extent, ozone can also diffuse directly through surface
layers to the plant's interior.264 Once ozone, a highly reactive substance, reaches the interior of
plant cells, it inhibits or damages essential  cellular components and functions, including enzyme
activities,  lipids, and cellular membranes, disrupting the plant's osmotic (i.e., water) balance and
                        O^S **}£*£*
energy utilization patterns.   '    This damage is commonly manifested as visible foliar injury
such as chlorotic or necrotic  spots, increased leaf senescence (accelerated leaf aging) and/or
reduced photosynthesis. All these effects reduce a plant's capacity to form carbohydrates, which
are the primary form of energy used by plants.267 With fewer resources available, the plant
reallocates existing resources away from root growth and storage, above ground growth or yield,
and reproductive processes, toward leaf repair and maintenance.  Studies have shown that plants
stressed in these ways may exhibit a general loss of vigor, which can lead to secondary impacts
that modify plants' responses to other environmental factors.  Specifically, plants may become
more sensitive to other air pollutants, more susceptible to disease, insect attack, harsh weather
(e.g., drought, frost) and other environmental stresses (e.g., increasing CO2 concentrations).
Furthermore, there is considerable evidence that ozone can interfere with the formation of
mycorrhiza, essential symbiotic fungi associated with the roots of most terrestrial plants, by
reducing the amount of carbon available for transfer from the host to the symbiont.268

       Not all plants, however, are equally sensitive to ozone. Much of the variation in
sensitivity between individual plants or whole species is related to the plant's ability to regulate
the extent of gas exchange via leaf stomata (e.g., avoidance of Os uptake through closure of
stomata).269'270'271 Other resistance mechanisms may involve the intercellular production of
detoxifying substances. Several biochemical substances capable of detoxifying ozone have been
reported to occur in plants including the antioxidants ascorbate and glutathione. After injuries
have occurred, plants may be capable of repairing the damage to a limited extent.272 Because of
the differing sensitivities among plants to ozone, ozone pollution can also exert a selective
pressure that leads to changes in plant community composition. Given the range of plant
sensitivities and the fact that numerous other environmental factors modify plant uptake and
response to ozone, it is not possible to identify threshold values above which ozone is toxic  for
all plants.  However, in general, the science suggests that ozone concentrations of 100  ppb or

                                           3-91

-------
greater can be phytotoxic to a large number of plant species, and can produce acute foliar injury
responses, crop yield loss and reduced biomass production. Ozone concentrations below 100 ppb
(50 to 99 ppb) can produce these effects in more sensitive plant species, and have the potential
over a longer duration of creating chronic stress on vegetation that can lead to effects of concern
associated with reduced carbohydrate production and decreased plant vigor. The next few
paragraphs present additional information on ozone damage to trees, ecosystems, agronomic
crops and urban ornamentals.

       Ozone also has been shown conclusively to cause discernible injury to forest trees.273'274
In terms of forest productivity and ecosystem diversity, ozone may be the pollutant with the
greatest potential for regional-scale forest impacts.275  Studies have demonstrated repeatedly that
ozone concentrations commonly observed in polluted areas can have substantial impacts on plant
function.276'277

       Because plants are at the center of the food web in many ecosystems, changes to the plant
community can affect associated organisms and ecosystems (including the suitability of habitats
that support threatened or endangered species and below ground organisms living in the root
zone). Ozone damages at the community and ecosystem level vary widely depending upon
numerous factors, including concentration and temporal variation of tropospheric ozone, species
composition,  soil properties and climatic factors.278  In most instances, responses to chronic or
recurrent exposure are subtle and not observable for many years. These injuries can cause stand-
level forest decline in sensitive ecosystems.279'280'281 It is not yet possible to predict ecosystem
responses to ozone with much certainty; however, considerable knowledge of potential
ecosystem responses has been acquired through long-term observations in highly damaged
forests in the  United States.

       Laboratory and field experiments have also shown reductions  in yields for agronomic
crops exposed to ozone, including vegetables (e.g., lettuce) and field crops (e.g., cotton and
wheat). The most extensive field experiments, conducted under the National Crop Loss
Assessment Network (NCLAN) examined 15 species and numerous cultivars. The NCLAN
results show that "several economically important crop species are sensitive to ozone levels
typical of those found in the Unites States."282 In addition, economic studies have shown a
relationship between observed ozone levels and crop yields.283'284'285

       Urban ornamentals represent an additional vegetation category likely to  experience some
degree of negative effects associated with exposure to ambient ozone levels and likely to impact
large economic sectors. In the absence of adequate exposure-response functions and economic
damage functions for the potential range of effects relevant to these types of vegetation, no direct
quantitative analysis has been conducted. It is estimated that more than $20 billion (1990
dollars) are spent annually on landscaping using ornamentals, both by private property
owners/tenants  and by governmental units responsible for public areas.286 This  is therefore a
potentially costly environmental effect.  However, methods are not available to  allow for
plausible estimates of the percentage of these expenditures that may be related to impacts
associated with ozone exposure.
                                          3-92

-------
3.4 Particulate Matter

       In this section we review the health and welfare effects of paniculate matter (PM). We
also describe air quality monitoring and modeling data that indicate many areas across the
country continue to be exposed to levels of ambient PM above the NAAQS. Emissions of PM
and VOC from the vehicles subject to this proposed rule contribute to these PM concentrations.
Information on air quality was gathered from a variety of sources, including monitored PM
concentrations, air quality modeling done for recent EPA rulemakings and other state and local
air quality information.

3.4.1   Science of PM Formation

       Particulate matter (PM) represents a broad class of chemically and physically diverse
substances. It can be principally characterized as discrete particles that exist in the condensed
(liquid or solid) phase spanning several orders of magnitude in size. PM is described based on
its size fractions. PMio refers to particles with an aerodynamic diameter less than or equal to a
nominal 10 micrometers. PM2 5 refers to fine particles, those particles with an aerodynamic
diameter less than or equal to a nominal 2.5 micrometers.  Coarse fraction particles refer to those
particles with an aerodynamic diameter less than or equal to a nominal 10 micrometers.
Inhalable coarse fraction particles refer to those particles with an aerodynamic diameter greater
than 2.5 micrometers, but less than or equal to a nominal 10 micrometers.  Ultrafine PM refers to
particles with diameters of less  than 100 nanometers (0.1 micrometers). Larger particles (greater
than 10 micrometers) tend to be removed by the respiratory clearance mechanisms whereas
smaller particles (PMio) are deposited deeper in the lungs.  Ambient fine particles are a complex
mixture including sulfates, nitrates, chlorides, ammonium compounds, organic carbon, elemental
carbon, geological material, and metals. Fine particles can remain in the atmosphere for days to
weeks and travel through the atmosphere hundreds to thousands of kilometers, while coarse
particles generally tend to deposit to the earth within minutes to hours and within tens of
kilometers from the emission source.

       The vehicles that would be covered by the proposed standards contribute to ambient PM
levels through primary (direct) and secondary (indirect) PM.  Primary PM is directly emitted into
the air, and secondary PM forms in the atmosphere from gases emitted by fuel combustion and
other sources. Along with primary PM, the vehicles controlled in this action emit VOC, which
react in the atmosphere to form secondary PM2.5, namely organic carbonaceous PM2.5. The gas
cans that would be covered by the proposed standards also  emit VOC which contribute to
secondary PM25. Both types of directly and indirectly formed particles from vehicles and gas
cans are found principally in  the fine fraction.
       EPA has National Ambient Air Quality Standards (NAAQS) for both PM2.5 and
The PM NAAQS consist of a short-term (24-hour) and a long-term (annual) standard.  The short-
term PM2 5 NAAQS is set at a level of 65 //g/m3 based on the 98th percentile concentration
averaged over three years. The long-term PM2 5 NAAQS specifies an expected annual arithmetic
mean not to exceed 15 |_ig/m3 averaged over three years.  The short-term (24-hour) PMio
NAAQS is set at a  level of 150 //g/m3 not to be exceeded more than once per year.  The long-
term PMio NAAQS specifies an expected annual arithmetic mean not to exceed 50 |_ig/m3.

                                          3-93

-------
       EPA has recently proposed to amend the PM NAAQS.H The proposal includes lowering
the level of the primary 24-hour fine particle standard from the current level of 65 micrograms
per cubic meter (|ig/m3) to 35 |ig/m3, retaining the level of the annual fine standard at 15|ig/m3,
and setting a new primary 24-hour standard for certain inhalable coarse particles (the indicator is
qualified so  as to include any ambient mix of PMi0-2.5 that is dominated by sources typically
found in urban environments, such as resuspended dust from high-density traffic on paved roads
and PM generated by industrial and construction sources, and excludes any ambient mix of
PMio-2.5 dominated by rural windblown dust and soils and PM generated by agricultural and
mining sources) at 70|ig/m3. The Agency is also requesting comment on various other standards
for fine and  inhalable coarse PM (71 FR 2620, Jan. 17, 2006).

3.4.2   Health Effects of Particulate Matter

       As stated in the EPA Air Quality Criteria Document for PM (PM Criteria Document),
available scientific findings "demonstrate well that human health outcomes are associated with
ambient PM."287 We are relying primarily on the data and conclusions in the PM Criteria
Document and PM staff paper, which reflects EPA's analysis of policy-relevant science from the
PM Criteria Document, regarding the health effects associated with particulate matter.288 We
also present additional recent studies published after the cut-off date for the PM criteria
document.  Taken together this information supports the conclusion that PM-related emissions
from the gasoline vehicles and gas cans being controlled in this action are associated with
adverse health effects. Information on PM-related mortality is presented first, followed by
information  on PM-related morbidity and near-roadway PM exposure.

3.4.2.1 Short-Term Exposure Mortality and Morbidity Studies

       As discussed in the PM Criteria Document (CD), short-term exposure to PM2.5 is
associated with premature mortality from cardiopulmonary diseases (CD, p. 8-305),
hospitalization and emergency department visits for cardiopulmonary diseases (CD, p. 9-93),
increased respiratory symptoms (CD, p. 9-46), decreased lung function (CD Table 8-34) and
physiological changes or biomarkers for cardiac changes (CD, Section 8.3.1.3.4).  In addition,
the CD describes a limited body of new evidence from epidemiologic studies for potential
relationships between short-term exposure to PM and health endpoints such as low birth weight,
preterm birth, and neonatal and infant mortality. (CD, Section 8.3.4).

       Among the studies of effects from short-term exposure to PM2.5, several studies
specifically address the contribution of mobile sources to short-term PM2.5 effects on daily
mortality. These studies indicate that there are statistically significant associations between
mortality and PM related to mobile source emissions (CD, p.8-85).  The analyses incorporate
source apportionment tools into daily mortality studies and are briefly mentioned here. Analyses
incorporating source apportionment by factor analysis with daily time-series studies of daily
death indicated a relationship between mobile source PMi.s and mortality.289'290  Another recent
study in 14 U.S. cities examined the effect of PMio exposures on daily hospital admissions for
H US EPA, National Ambient Air Quality Standards for Particulate Matter (71 FR 2620, Jan. 17, 2006). This
document is also available on the web at: http://www.epa.gov/air/particlepollution/actions.html.
                                          3-94

-------
cardiovascular disease. They found that the effect of PMio was significantly greater in areas with
a larger proportion of PMio coming from motor vehicles, indicating that PMio from these sources
may have a greater effect on the toxicity of ambient PMio when compared with other sources.291
These studies provide evidence that PM-related emissions, specifically from mobile sources, are
associated with adverse health effects.

3.4.2.2 Long-Term Exposure Mortality and Morbidity Studies

       Short-term exposure studies provide one way of examining the effect of short-term
variations in air quality on morbidity and mortality.  However, they do not allow for an
evaluation of the effect of long-term exposure to air pollution on human mortality and
morbidity.292 Longitudinal cohort studies allow for analysis of such effects.

       Long-term exposure to elevated ambient PM2 5 is associated with mortality from
cardiopulmonary diseases and lung cancer (CD, p. 8-307), and effects on the respiratory system
such as decreased lung function or the development of chronic respiratory disease (CD, pp. 8-
313, 8-314).  Of specific importance to this proposal, the PM Criteria Document also notes that
the PM components of gasoline and diesel engine exhaust are likely to be major contributors to
the observed increases in lung cancer mortality associated with ambient PM2.5 (CD, p. 8-318).

       The PM Criteria Document emphasizes the results of two long-term studies, the Six
Cities and American Cancer Society (ACS) prospective cohort studies, based on several factors -
the large study population in the ACS study, the large air quality data set in the  Six Cities study,
the generally representative study population used in the Six Cities study, and the fact that these
studies have undergone extensive reanalysis (CD, p. 8-306).293'294'295 One analysis of a subset of
the ACS  cohort data, which was published after the PM criteria document was finalized, found a
larger  association than had previously been reported between long-term PM2.5 exposure and
mortality in the Los Angeles area using a new exposure estimation method that accounted for
variations in concentration within the city.296 These studies provide  strong evidence of a link
between  PM2.5 and mortality, including all-cause, cardiorespiratory,  and lung cancer mortality
(CD, p. 8-307).

       As discussed in the PM Criteria Document, the newer morbidity studies that combine the
features of cross-sectional and cohort studies provide the best evidence for chronic exposure
effects. Long-term studies evaluating the effect of ambient  PM on children's development have
shown some evidence indicating effects of PM2.5 and/or PMio on reduced lung function growth
(CD, Section 8.3.3.2.3). In another recent publication, investigators in southern California
reported  the results of a cross-sectional  study of outdoor PM2.5 and measures of atherosclerosis in
the Los Angeles basin.297 The study found significant associations between ambient residential
PM2.5 and carotid intima-media thickness (CIMT), an indicator of subclinical atherosclerosis, an
underlying factor in cardiovascular disease.

3.4.2.3 Roadway-Related Pollution Exposure

       A recent body of studies reinforces the findings of these PM morbidity and mortality
effects by looking at traffic-related exposures, PM measured along roadways, or time spent in

                                           3-95

-------
traffic and adverse health effects. While many of these studies did not measure PM specifically,
they include potential exhaust exposures which include mobile source PM because they employ
indices such as roadway proximity or traffic volumes.  One study with specific relevance to
PM2.5 health effects is a study that was done in North Carolina looking at concentrations of PM2.5
inside police cars and corresponding physiological changes in the police personnel driving the
cars.  The authors report significant elevations in markers of cardiac risk associated with
concentrations of PM2.5 inside police cars on North Carolina state highways.298 Additional
information on near-roadway health effects is included in Section 3.5 of this RIA.

3.4.3   Current and Projected PM Levels

        The emission reductions from this proposed rule would assist PM nonattainment areas in
reaching the standard by each area's respective attainment date and assist PM maintenance areas
in maintaining the PM standards in the future.  In this section we present information on current
and future attainment of the PM standards.

3.4.3.1 Current PM2.s Levels

        A nonattainment area is defined in the Clean Air Act (CAA) as an area that is violating
an ambient standard or is contributing to a nearby area that is violating the standard. EPA has
recently designated nonattainment areas for the 1997 PM2.5 NAAQS by calculating air quality
design values (using 2001-2003 or 2002-2004 measurements) and considering other factors.1
The Air Quality Designations and Classifications for the Fine Particles (PM2 5) NAAQS rule lays
out the factors that EPA considered in making the nonattainment designations (70 FR 943, Jan.
5, 2005).  According to EPA's recent designations, approximately 88 million people live in the
39 PM2 5 areas designated as nonattainment for either failing to meet the 1997 PM2 5 NAAQS or
for contributing to poor air quality in a nearby area. There are 208 full or partial counties that
make up the PM2.5 nonattainment areas, as shown in Figure 3.3-1.  The PM2.5 nonattainment
counties, areas and populations, as of September 2005, are listed in Appendix 3B to this RIA.

        States with PM2 5 nonattainment areas will be required to take action to bring those areas
into compliance in the future. Most PM2.5 nonattainment areas will be required to attain the 1997
PM2.5 NAAQS in the 2009 to 2014 time frame and then be required to maintain the PM2.5
NAAQS thereafter/  The emission standards being proposed in this action would become
effective between 2009 and 2015. The expected PM2 5 and PM2 5 precursor inventory reductions
1 The full details involved in calculating a PM2 5 design value are given in Appendix N of 40 CFR Part 50.
1 While the final implementation process for bringing the nation's air into attainment with the PM2 5 NAAQS is still
being completed in a separate rulemaking action, the basic framework is well defined by the statute. The EPA
finalized PM25 attainment and nonattainment areas in April 2005. Following designation, Section 172(b) of the
Clean Air Act allows states up to 3 years to submit a revision to their state implementation plan (SIP) that provides
for the attainment of the PM25 standard. Based on this provision, states could submit these SIPs until April 2008.
Section 172(a)(2) of the Clean Air Act requires that these SIP revisions demonstrate that the nonattainment areas
will attain the PM2 5 standard as expeditiously as practicable but no later than 5 years from the date that the area was
designated nonattainment. However, based on the severity of the air quality problem and the availability and
feasibility of control measures, the Administrator may extend the attainment date "for a period of no greater than 10
years from the date of designation as nonattainment."  Based on section 172(a) provisions in the Act, we expect that
areas will need to attain the PM25 NAAQS in the 2010 (based on 2007 - 2009 air quality data) to 2015 (based on
2012 to 2014 air quality data) time frame, and then be required to maintain the NAAQS thereafter.
                                            3-96

-------
from the standards proposed in this action would be useful to states in attaining or maintaining
the 1997PM2.5NAAQS.

3.4.3.2  Current PMio Levels

       EPA designated PMio nonattainment areas in 1990.K As of September 2005,
approximately 29 million people live in the 55 areas that are designated as PMio nonattainment,
for either failing to meet the PMio NAAQS or for contributing to poor air quality in a nearby
area.  There are 54 full or partial counties that make up the PMio nonattainment areas. The PMio
nonattainment areas and populations are listed in Appendix 3B to this RIA.

       The attainment date for the initial moderate PMio nonattainment areas, designated by law
on November 15, 1990, was December 31,  1994.  Several additional moderate PMio
nonattainment areas were designated in January of 1994, and the attainment date for these areas
was December 31, 2000. The initial serious PMio nonattainment areas had an attainment date set
by the Act of December 31, 2001.  The Act provides that EPA may grant extensions of the
serious area attainment dates of up to 5 years, provided that the area requesting the extension
meets the requirements of section 188(e) of the Act. Four serious PMio nonattainment areas
(Phoenix, Arizona; Clark County (Las Vegas), NV; Coachella Valley, and South Coast (Los
Angeles), CA) have received extensions of the December 31, 2001  attainment date and thus have
new attainment dates of December 31, 2006.  We expect that most PMio nonattainment areas
will attain the PMio standard in the 2006 time frame, depending on an area's classification and
other factors, and then be required to maintain the PMio NAAQS thereafter.  The projected
reductions in emissions from the proposed controls would be useful to states to maintain the
PMi0NAAQS.L

3.4.3.3 Projected PM2.5 Levels

       Recent air quality modeling predicts that without additional controls there will continue
to be a need for reductions in PM concentrations in the future.  In the following sections  we
describe the recent PM air quality modeling and results of the modeling.

3.4.3.3.1      PM Modeling Methodology

       Recently PM air quality analyses were performed for the Clean Air Interstate Rule
(CAIR), which was promulgated by EPA in 2005.  The Community Multiscale Air Quality
(CMAQ) model was used as the tool for simulating base and future year concentrations of PM,
visibility and deposition in support of the CAIR air quality assessment.  The  CAIR analysis
K A PM10 design value is the concentration that determines whether a monitoring site meets the NAAQS for PM10.
The full details involved in calculating a PM10 design value are given in Appendices H and I of 40 CFR Part 50.
L As mentioned above, the EPA has recently proposed to amend the PM NAAQS, by establishing a new indicator
for certain inhalable coarse particles, and a new primary 24-hour standard for coarse particles described by that
indicator. EPA also proposed to revoke the current 24-hour PM10 standard in all areas of the country except in those
areas with a population of at least 100,000 people and which contain at least one monitor violating the 24-hour PM10
standard, based on the most recent 3 years of air quality data. In addition, EPA proposed to revoke upon
promulgation of this rule the current annual PM10 standard if EPA finalizes the proposed primary standard for PM10.
2 5 (71FR 2620, Jan. 17,2006).
                                           3-97

-------
included all final federal rules up to and including CAIR controls.  Details on the air quality
modeling are provided in the Air Quality Modeling Technical Support Document for the Final
Clean Air Interstate Rule, included in the docket for this proposed rule.299
3.4.3.3.2
Areas at Risk of Future PM2.5 Violations
       Air quality modeling performed for CAIR indicates that in the absence of additional
local, regional or national controls, counties with annual average PM2.5 levels above 15 |_ig/m3
are likely to persist in the future.  The CAIR analysis provided estimates of future PM2.5 levels
across the country.  For example, in 2010 based on emission controls currently adopted or
expected to be in placeM, we project that 19 million people will live in 28 Eastern counties with
average PM2s levels at and above 15 |-ig/m3, see Table 3.4-1. The proposed rule would assist
these counties in attaining the PM2 sNAAQS.  Table 3.4-1 also lists the 56 Eastern counties,
where 24 million people are projected to live, with 2010 projected design values that do not
violate the annual PM2.5 NAAQS but are within ten percent of it.  These are counties that are not
projected to violate the standard, but to be close to it.  The proposed rule may help ensure that
these counties continue to maintain their attainment status and the emission reductions from  this
proposed rule would be included by the states in their baseline inventory modeling for their
PM2.5 maintenance plans.

      Table 3.4-1.  Eastern Counties with 2010 Projected Annual PM2.s Design Values
                   Above and within 10% of the Annual PMi.5 Standard
State
Alabama
Alabama
Alabama
Alabama
Alabama
Alabama
Alabama
Delaware
District of
Columbia
Georgia
Georgia
Georgia
Georgia
Georgia
Georgia
Georgia
Georgia
Georgia
County
DeKalb Co
Jefferson Co
Montgomery Co
Morgan Co
Russell Co
Shelby Co
Talladega Co
New Castle Co
District of Columbia

Bibb Co
Chatham Co
Clarke Co
Clayton Co
Cobb Co
DeKalb Co
Dougherty Co
Floyd Co
Fulton Co
2010 Projected
PM25 Concentration
(//g/m3)3'"
13.97
17.46
14.10
14.11
15.15
13.83
14.00
14.84
13.68

15.17
14.02
14.96
16.29
15.35
15.70
13.85
15.87
16.98
2000 pop0
64,452
662,046
223,509
111,064
49,756
143,293
80,321
500,264
572,058

153,887
232,047
101,488
236,516
607,750
665,864
96,065
90,565
816,005
2010popd
70,826
667,602
240,104
121,931
52,706
202,915
84,163
534,631
554,474

158,291
242,134
106,187
265,407
744,488
698,335
99,323
95,238
855,826
M
  Counties forecast to remain in nonattainment may need to adopt additional local or regional controls to attain the
standards by dates set pursuant to the Clean Air Act. The emissions reductions associated with this proposed rule
would help these areas attain the PM standards by their statutory date.
                                            3-98

-------
State
Georgia
Georgia
Georgia
Georgia
Georgia
Georgia
Georgia
Illinois
Illinois
Illinois
Illinois
Illinois
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Kentucky
Kentucky
Kentucky
Kentucky
Maryland
Michigan
Michigan
Michigan
Michigan
Mississippi
Missouri
New Jersey
New York
New York
North Carolina
North Carolina
North Carolina
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
County
Gwinnett Co
Hall Co
Muscogee Co
Richmond Co
Walker Co
Washington Co
Wilkinson Co
Cook Co
DuPage Co
Madison Co
St. Clair Co
Will Co
Clark Co
Dubois Co
Elkhart Co
Lake Co
Marion Co
Vanderburgh Co
Bullitt Co
Fayette Co
Jefferson Co
Kenton Co
Baltimore city
Kalamazoo Co
Monroe Co
Oakland Co
Wayne Co
Jones Co
St. Louis City
Union Co
Bronx Co
New York Co
Catawba Co
Davidson Co
Mecklenburg Co
Butler Co
Cuyahoga Co
Franklin Co
Hamilton Co
Jefferson Co
Lawrence Co
Lucas Co
Montgomery Co
Scioto Co
Stark Co
Summit Co
2010 Projected
PM25 Concentration
(M9/m3)a'b
14.02
14.28
14.57
14.64
14.22
14.22
15.22
16.88
13.81
15.96
15.54
14.30
15.15
14.37
13.93
16.48
15.54
14.26
13.67
14.17
15.44
13.72
14.88
13.66
14.03
13.70
18.23
14.21
14.40
13.60
13.62
14.95
14.07
14.36
13.92
15.03
17.11
15.13
16.61
15.64
14.11
13.76
13.83
15.98
15.08
14.69
2000 pop0
588,447
139,276
186,290
199,774
61,053
21,176
10,220
5,376,739
904,160
258,940
256,081
502,265
96,472
39,674
182,790
484,563
860,453
171,922
61,236
260,511
693,603
151,464
651,153
238,602
145,945
1,194,155
2,061,161
64,958
348,188
522,540
1,332,649
1,537,194
141,685
147,246
695,453
332,806
1,393,977
1,068,977
845,302
73,894
62,319
455,053
559,061
79,195
378,097
542,898
2010popd
743,813
156,939
193,867
207,977
64,140
22,220
10,958
5,363,464
1,019,575
267,328
253,343
588,797
107,096
41,394
195,982
489,220
879,932
175,307
71,957
292,752
704,891
160,582
616,324
251,616
153,348
1,299,592
1,964,209
68,915
324,156
523,568
1,298,206
1,539,917
155,349
164,790
814,088
384,410
1,348,313
1,142,894
843,226
70,731
63,014
447,302
552,901
80,248
382,563
552,567
3-99

-------

State County

Pennsylvania Allegheny Co
Pennsylvania Beaver Co
Pennsylvania Berks Co
Pennsylvania Delaware Co
Pennsylvania Lancaster Co
Pennsylvania Philadelphia Co
Pennsylvania York Co
Tennessee Davidson Co
Tennessee Hamilton Co
Tennessee Knox Co
Tennessee McMinn Co
Tennessee Roane Co
Tennessee Shelby Co
Tennessee Sullivan Co
Texas Harris Co
West Virginia Brooke Co
West Virginia Cabell Co
West Virginia Hancock Co
West Virginia Kanawha Co
West Virginia Wood Co
Number of Violating Counties
Population of Violating Counties
Number of Counties within 10%
Population of Counties within 10%
2010 Projected
PM25 Concentration
(M9/m3)a'b
18.01
13.61
13.56
13.94
14.09
14.98
14.20
14.26
15.57
16.16
13.64
13.58
13.77
14.01
13.78
14.42
15.08
14.89
15.27
14.14
28

56


2000 pop0

1,281,665
181,412
373,637
550,863
470,657
1,517,549
381,750
569,890
307,895
382,031
49,015
51,910
897,471
153,048
3,400,577
25,447
96,784
32,667
200,072
87,986

19,488,510

23,310,383

2010popd

1,259,040
183,693
388,194
543,169
513,684
1,420,803
404,807
589,133
327,337
426,545
51,072
54,744
958,501
159,873
3,770,129
24,753
93,421
31,374
197,381
87,994

19,750,033

24,583,976
a) Bolded concentrations indicate levels above the annual PM2 5 standard.
b) Concentrations are calculated for counties with Federal Reference Method PM2 5 monitoring data.
c) Populations are based on 2000 census data.
d) Populations are based on 2000 census projections.

3.4.4  Environmental Effects of PM Pollution

       In this section we discuss public welfare effects of PM and its precursors including
visibility impairment, atmospheric deposition, and materials damage and soiling.

3.4.4.1 Visibility Degradation

       Visibility is important because it directly affects people's enjoyment of daily activities in
all parts of the country. Individuals value good visibility for the well-being it provides them
directly, both in where they live and work, and in  places where they enjoy recreational
opportunities. Visibility is also highly valued in significant natural areas such as national parks
and wilderness areas, because of the special emphasis given to protecting these lands now and
for future generations.

       Fine particles are the major cause of reduced visibility in parts of the United States. To
address the welfare effects of PM on visibility, EPA set secondary PM2 5 standards in 1997
                                            3-100

-------
which would work in conjunction with the establishment of a regional haze program.  The
secondary (welfare-based) PM2 5 NAAQS was established as equal to the suite of primary
(health-based) NAAQS (62 FR 38652, July 18, 1997). Furthermore, Section 169 of the Act
provides additional authorities to remedy existing visibility impairment and prevent future
visibility impairment in the 156 national parks, forests and wilderness areas labeled as
Mandatory Class I Federal Areas (62 FR 38652, July 18, 1997)N In July 1999 the regional haze
rule (64 FR 35714) was put in place to protect the visibility in Mandatory Class I Federal Areas.
A list of the Mandatory Class I Federal Areas is included in Appendix 3C.°

       Data showing PM2 5 nonattainment areas and visibility levels above background at the
Mandatory Class I Federal Areas demonstrate that unacceptable visibility impairment is
experienced throughout the U.S., in multi-state regions, urban areas, and remote mandatory
Federal class I areas. The PM and PM precursor emissions from the vehicles and gas cans
subject to this proposed rule contribute to these visibility effects.

3.4.4.1.1      Current Visibility Impairment

       The need for reductions in the levels of PM2 5 is widespread.  Currently, high ambient
PM2 5 levels are measured throughout the country.  Fine particles may remain suspended for days
or weeks and travel hundreds to thousands of kilometers, and thus fine particles emitted or
created in one county may contribute to ambient concentrations in a neighboring region.300

       As mentioned above the secondary PM2 5 standards were set as equal to the suite of
primary PM2.5 standards. Recently designated PM2.5 nonattainment areas indicate that almost 90
million people live in 208 counties that are in nonattainment for the 1997 PM2 5 NAAQS, see
Appendix 3.4-A. Thus, at least these populations (plus others who travel to these areas) would
likely be experiencing visibility impairment.

3.4.4.1.2      Current Visibility Impairment at Mandatory Class I Federal Areas

       Detailed information about current and historical visibility conditions in Mandatory Class
I Federal Areas is summarized in the EPA Report to Congress and the 2002 EPA Trends
Report.301'302  The conclusions draw upon the Interagency Monitoring of Protected Visual
Environments (IMPROVE) network data.  One of the objectives of the IMPROVE monitoring
network program is to provide regional haze monitoring representing all Mandatory Class I
Federal Areas where practical.  The National Park Service report also describes the state of
national park visibility conditions and discusses the need for improvement.303

       The regional  haze rule requires states to establish goals for each affected Mandatory
Class I Federal Area to improve visibility on the haziest days (20% most impaired days) and
N These areas are defined in Section 162 of the Act as those national parks exceeding 6,000 acres, wilderness areas
and memorial parks exceeding 5,000 acres, and all international parks which were in existence on August 7, 1977.
0 As mentioned above, the EPA has recently proposed to amend the PM NAAQS (71 FR 2620, Jan. 17, 2006). The
proposal would set the secondary NAAQS equal to the primary standards for both PM25 and PM10-25 EPA also is
taking comment on whether to set a separate PM2 5 standard, designed to address visibility (principally in urban
areas), on potential levels for that standard within a range of 20 to 30 ug/m3, and on averaging times for the standard
within a range of four to eight daylight hours.
                                          3-101

-------
ensure no degradation occurs on the cleanest days (20% least impaired days). Although there
have been general trends toward improved visibility, progress is still needed on the haziest days.
Specifically, as discussed in the 2002 EPA Trends Report, without the effects of pollution a
natural visual range in the United States is approximately 75 to 150 km in the East and 200 to
300 km in the West. In 2001, the mean visual range for the worst days was 29 km in the East
and 98 km in the West.304

3.4.4.1.3       Future Visibility Impairment

       Recent modeling for the CAIR was used to project PM2 5 levels in the U.S. in 2010. The
results suggest that PM2.5 levels above the 1997 NAAQS will persist in the future. We predicted
that in 2010, there will be 28 Eastern counties with a population of 19 million where annual
PM2.5 levels are above 15 //g/m3, see Table 3.4-1.  Thus, in the future, a percentage of the
population may continue to experience visibility impairment in areas where they live, work and
recreate.

       The PM and PM precursor emissions from the vehicles and gas cans subject to the
proposed controls contribute to visibility impairment. These emissions occur in and around areas
with PM2 5 levels above the annual 1997 PM2 5 NAAQS. Thus, the emissions from these sources
contribute to the current and anticipated visibility impairment and the proposed emission
reductions may help improve future visibility impairment.

3.4.4.1.4       Future Visibility Impairment at Mandatory Class I Federal Areas

       Achieving the annual 1997 PM2 5 NAAQS will help improve visibility across the country,
but it will not be sufficient to meet the statutory goal of no manmade impairment in the
Mandatory Class I Federal Areas (64 FR 35714, July 1, 1999 and 62 FR 38652, July 18,  1997).
In setting the NAAQS, EPA discussed how the NAAQS in combination with the regional haze
program, is deemed to improve visibility consistent with the goals  of the Act. In the East, there
are and will continue to be areas with PM2 5 concentrations above 15 |_ig/m3 and where light
extinction is significantly above natural background. Thus, large areas of the Eastern United
States have air pollution that is causing and will continue to cause visibility impairment.  In the
West, scenic vistas are especially important to public welfare. Although the annual 1997 PM2.5
NAAQS  is met in most areas outside of California, virtually the entire West is in close proximity
to a scenic Mandatory Class I Federal Area protected by  169 A and 169B of the CAA.

       Recent modeling for the Clean Air Interstate Rule (CAIR) was also used to project
visibility conditions in mandatory Federal class I areas across the country in 2015. The results
for the mandatory Federal class I areas suggest that these areas are predicted to continue to have
visibility impairment above background on the 20% worst days in the future.

       The overall goal of the regional haze program is to prevent future visibility impairment
and remedy existing visibility impairment in Mandatory Class I Federal  Areas. As shown by the
future visibility estimates in Appendix 3C it is projected that there will continue to be Mandatory
Class I Federal Areas with visibility levels above background in 2015.305 Additional emission
reductions will be needed from the broad set of sources that contribute, including the vehicles

                                         3-102

-------
and gas cans subject to this proposed rule. The reductions proposed in this action are a part of
the overall strategy to achieve the visibility goals of the Act and the regional haze program.

3.4.4.2 Atmospheric Deposition

       Wet and dry deposition of ambient particulate matter delivers a complex mixture of
metals (e.g., mercury, zinc, lead, nickel, aluminum, cadmium), organic compounds (e.g., POM,
dioxins, furans) and inorganic compounds (e.g., nitrate, sulfate) to terrestrial and aquatic
ecosystems. The chemical form of the compounds deposited is impacted by a variety of factors
including ambient conditions (e.g., temperature, humidity, oxidant levels) and the sources of the
material. Chemical and physical transformations of the parti culate compounds occur in the
atmosphere as well as the media onto which they deposit. These transformations in turn
influence the fate, bioavailability and potential toxicity of these compounds. Atmospheric
deposition has been identified as a key component of the environmental and human health hazard
posed by several pollutants including mercury, dioxin  and PCBs.306

       Adverse impacts on water quality can occur when atmospheric contaminants deposit to
the water surface or when material deposited on the land enters a waterbody through runoff.
Potential impacts of atmospheric deposition to waterbodies include adverse effects  to human
health and welfare through ingestion of contaminated fish, ingestion of contaminated water,
damage to the marine ecology, and limited recreational uses. Several  studies have been
conducted in U.S. coastal waters and in the Great Lakes Region in which the role of ambient PM
deposition and runoff is investigated.307'308'309'310'311

       Adverse impacts on soil chemistry and plant life have been observed for areas heavily
impacted by atmospheric deposition of metals and acid species, resulting in forest decline and
damage to forest productivity.  Potential impacts also include adverse effects to human health
through ingestion of contaminated vegetation or livestock (as in the case for dioxin deposition),
reduction in crop yield, and limited use of land due to  contamination.

       In the following subsections, atmospheric deposition of heavy metals and particulate
organic material  is discussed.

3.4.4.2.1      Heavy Metals

       Heavy metals, including cadmium, copper, lead, chromium, mercury, nickel and zinc,
have the greatest potential for influencing forest growth (CD, p. 4-87).312 Investigation of trace
metals near roadways and industrial facilities indicate  that a substantial burden of heavy metals
can accumulate on vegetative  surfaces. Copper, zinc,  and nickel have been documented to cause
direct toxicity to vegetation under field conditions (CD, p. 4-75).  Little research has been
conducted on the effects associated with mixtures of contaminants found in ambient PM.  While
metals typically exhibit low solubility, limiting their bioavailability and direct toxicity, chemical
transformations of metal compounds occur in the environment, particularly in the presence of
acidic or other oxidizing species. These chemical changes influence the mobility and toxicity of
metals in the environment. Once taken up into plant tissue, a metal compound can undergo
chemical changes, accumulate and be passed along to herbivores or can re-enter the soil and

                                          3-103

-------
further cycle in the environment.

       Although there has been no direct evidence of a physiological association between tree
injury and heavy metal exposures, heavy metals have been implicated because of similarities
between metal deposition patterns and forest decline (CD, p. 4-76).313 Contamination of plant
leaves by heavy metals can lead to elevated soil levels. Trace metals absorbed into the plant
frequently bind to the leaf tissue, and then are lost when the leaf drops (CD, p. 4-75). As the
fallen leaves decompose, the heavy metals are transferred into the soil.314'315

       The environmental sources and cycling of mercury are currently of particular concern due
to the bioaccumulation and biomagnification of this metal in aquatic ecosystems and the potent
toxic nature of mercury in the forms in which is it ingested by people and other animals. Mercury
is unusual compared with other metals in that it largely partitions into the gas phase (in elemental
form), and therefore has a longer residence time in the atmosphere than a metal found
predominantly in the particle phase. This property enables mercury to travel far from the primary
source before being deposited and accumulating in the aquatic ecosystem. The major source of
mercury in the Great Lakes is from atmospheric deposition, accounting for approximately eighty
percent of the mercury in  Lake Michigan.316'317  Over fifty percent of the mercury in the
Chesapeake Bay has been attributed to atmospheric deposition.318 Overall, the National Science
and Technology Council (NSTC, 1999) identifies atmospheric deposition as the primary source
of mercury to aquatic systems.  Thirty-seven states have issued health advisories for the
consumption offish contaminated by mercury; however, most of these advisories are issued in
areas without a mercury point source.

       Elevated levels of zinc and lead have been identified in streambed sediments, and these
elevated levels have been correlated with population density and motor vehicle use.319'320  Zinc
and nickel have also been identified in urban water and soils.  In addition, platinum, palladium,
and rhodium, metals found in the catalysts of modern motor vehicles, have been measured at
                             101
elevated levels along roadsides.   Plant uptake of platinum has been observed at these locations.

3.4.4.2.2      Poly cyclic Organic Matter

       Poly cyclic organic matter (POM) is a byproduct of incomplete combustion and consists
of organic compounds with more than one benzene ring and a boiling point greater than or equal
to 100 degrees centigrade.322 Poly cyclic aromatic hydrocarbons (PAHs) are a class  of POM that
contain compounds which are known or suspected carcinogens.

       Major sources of PAHs  include mobile sources.  PAHs in the environment may be
present as a gas or adsorbed onto airborne parti culate matter. Since the majority of PAHs are
adsorbed onto particles less than 1.0 [im in diameter, long range transport is possible.  However,
studies have shown that PAH compounds adsorbed onto diesel exhaust particulate and exposed
to ozone have half lives of 0.5 to 1.0 hours.323

       Since PAHs are insoluble, the compounds generally are particle reactive and accumulate
in sediments. Atmospheric deposition of particles is believed to be the major source of PAHs to
the sediments of Lake Michigan.324'325 Analyses of PAH deposition to Chesapeake and

                                         3-104

-------
Galveston Bay indicate that dry deposition and gas exchange from the atmosphere to the surface
water predominate.326'327  Sediment concentrations of PAHs are high enough in some segments
of Tampa Bay to pose an environmental health threat.  EPA funded a study to better characterize
the sources and loading rates for PAHs into Tampa Bay.328 PAHs that enter a waterbody through
gas exchange likely partition into organic rich particles and be biologically recycled, while dry
deposition of aerosols containing PAHs tends to be more resistant to biological recycling.329
Thus, dry deposition is likely the main pathway for PAH concentrations in sediments while
gas/water exchange at the surface may lead to PAH distribution into the food web, leading to
increased health risk concerns.

       Trends in PAH deposition levels are difficult to discern because of highly variable
ambient air concentrations, lack of consistency in monitoring methods, and the significant
influence of local sources on deposition levels.330 Van Metre et al. (2000) noted PAH
concentrations in urban reservoir sediments have increased by 200-300% over the last forty years
and correlates with increases in automobile use.331

       Cousins et al.  (1999) estimates that greater than ninety percent of semi-volatile organic
compound (SVOC) emissions in the United Kingdom deposit on soil.332 An analysis of
polycyclic aromatic hydrocarbon (PAH) concentrations near a Czechoslovakian roadway
indicated that concentrations were thirty times greater than background.333

3.4.4.3 Materials Damage and Soiling

       The deposition of airborne particles can also reduce the aesthetic appeal of buildings and
culturally important articles through soiling, and  can contribute directly (or in conjunction with
other pollutants) to structural damage by means of corrosion or erosion.334 Particles affect
materials principally by promoting and accelerating the corrosion of metals, by degrading paints,
and by deteriorating building materials such as concrete and limestone. Particles contribute to
these effects because  of their electrolytic, hygroscopic, and acidic properties, and their ability to
sorb  corrosive gases (principally sulfur dioxide).  The rate of metal corrosion depends on a
number of factors, including the deposition rate and nature of the pollutant; the influence of the
metal protective corrosion film; the amount of moisture present; variability in the
electrochemical reactions; the presence and concentration of other surface electrolytes; and the
orientation of the metal surface.

3.5    Health and Welfare Impacts of Near-Roadway Exposure

       Over the years there have been a large number of studies that have examined associations
between  living near major roads and different adverse health endpoints. These studies generally
examine  people living near heavily-trafficked roadways, typically within several hundred meters,
where fresh emissions from motor vehicles are not yet fully diluted with background air.

       As discussed in Chapter 3.1.3, many studies have measured elevated concentrations of
pollutants emitted directly by motor vehicles near large roadways, as compared to overall urban
background levels. These elevated concentrations generally occur within approximately 200
meters of the road,  although the distance may vary depending on traffic and environmental

                                         3-105

-------
conditions. Pollutants measured with elevated concentrations include benzene, polycyclic
aromatic hydrocarbons, carbon monoxide, nitrogen dioxide, black carbon, and coarse, fine, and
ultrafine particles. In addition, resuspended road dust, and wear particles from tire and brake use
also show concentration increases in proximity of major roadways.

       The concentration changes that occur near major roadways are not fully captured in our
current air quality models used to assess the public health impacts of the proposed standards.
The studies discussed in this section address exposures and health effects that are at least
partially captured by our modeling, but there are additional exposures and health effects
associated with pollutants that are not explicitly quantified.

       At this point, there exists no exposure metric specific to "traffic," although as noted
above, a wide variety of gaseous, particulate, and semi-volatile species are elevated near
roadways. As a result, the exposure metrics employed generally indicate the presence and/or
intensity of a mixture of air pollutants for exposure assessment. Many of the health studies
discussed below employ non-specific exposure metrics, including traffic on roads nearest home
or school, distance to the nearest road, measured outdoor nitrogen dioxide concentrations, air
quality dispersion modeling of specific traffic-generated chemicals, and exposure assignment
based on land use.  These exposure metrics represent the mixture of traffic-generated pollutants,
rather than individual pollutants. Accordingly, such results are not directly comparable with
community epidemiology studies that employ ambient measurements of particulate matter or
ozone over a fixed time period, or to toxicological  studies employing a single pollutant to
evaluate responses in humans or animals.

       A wide range of health effects are reported in the literature related to near roadway and
in-vehicle exposures. This is not unexpected, given the chemical and physical complexity of the
mixture to which people are exposed in this environment.  These effects overlap with those
identified in our discussion of the effects of PM and ozone. The near-roadway health studies
provide stronger evidence for some health endpoints than others.  Epidemiologic evidence of
adverse responses to traffic-related pollution is strongest for non-allergic respiratory symptoms,
and several well-conducted epidemiologic studies have shown associations with cardiovascular
effects, premature adult mortality, and adverse birth outcomes, including low birth weight and
size. Traffic-related pollutants have been repeatedly associated with increased prevalence of
asthma-related respiratory symptoms in children, although epidemiologic evidence remains
inconclusive for a hypothesized link between traffic and the development of allergies and new
onset asthma.

       For childhood cancer, in particular childhood leukemia, epidemiologic studies have
shown less ability to detect the risks predicted from toxicological studies. Several small  studies
report positive associations, though such effects have not been observed in two larger studies. As
described above in Chapter 1.3, benzene and 1,3-butadiene are both known human leukemogens
in adults from occupational exposures. As previously mentioned, epidemiologic studies  have
shown an increased  risk of leukemia among children whose parents have been occupationally
exposed to benzene. While epidemiologic studies of near-roadway exposures have not always
shown a statistically significant association with childhood leukemias, the results are consistent
with the risks predicted from the studies at higher exposure levels. As a whole the toxicology

                                          3-106

-------
and epidemiology are consistent with a potentially serious children's health concern and
additional research is needed.

       Significant scientific uncertainties remain in research on health effects near roads,
including the exposures of greatest concern, the importance of chronic versus acute exposures,
the role of fuel type (e.g. diesel or gasoline) and composition (e.g., percent aromatics), and
relevant traffic patterns.  Furthermore, in these studies, it is often difficult to understand the role
of co-stressors including noise and socioeconomic status (e.g., access to health care, nutritional
status), and the role of differential susceptibility.

3.5.1   Mortality

       In addition to studies that have documented the relationship between PM and premature
mortality, a few recent studies have investigated the relationship between premature mortality
and broader indicators of transportation emissions, such as residence near traffic.  The extent to
which these studies are detecting any additional effects not accounted for in the PM-premature
mortality relationship is unclear.

       Living near major roads has been investigated in both long-term and short-term mortality
studies. Long-term studies track subjects over time and investigate the mortality rates among
groups with different levels of exposure to ambient pollutants.  Short term studies employ daily
variation in ambient concentrations to estimate the daily deaths attributable to air pollution.

       A total of two cohort studies have examined premature mortality in relation to residence
near traffic, while one other has examined stroke mortality.  In addition, one study accounted for
the effect of residence along a major road on associations with daily deaths in a time-series
study. These studies constitute all of the studies examining mortality with reference to proximity
to traffic.

       Premature mortality in adults in association with living near high-traffic roadways has
been studied in two recent cohort studies for all-cause  and cardiopulmonary mortality from the
Netherlands and from Ontario, Canada.335'336'337 Canadian vehicles and emission standards
largely mirror the U.S. vehicle fleet. Both studies defined living near a major road as having a
residence within 100 meters of a highway or within 50 meters of a major urban roadway.  In the
first study, involving approximately 5,000 people over 55 years  old living throughout the
Netherlands, residence near major roadways was associated with a 41% increase in the mortality
rate from all causes and a 95% increase in the cardiopulmonary  mortality rate.338  The second
study involved over 5,200 subjects aged 40 years or more, all living in the Hamilton, Ontario
area.  This study examined total mortality, finding a statistically significant 18% increase
associated with living near a major roadway. No difference in response was found among those
with pre-existing respiratory illness. The study also calculated "rate advancement periods,"
which describe the effect of an exposure in terms of the time period by which exposed persons
reach prematurely the same disease risk as unexposed persons reach later on.  The rate
advancement period for total mortality was 2.5 years.  The rate advancement periods were also
calculated for other risk factors for mortality, including chronic  pulmonary disease excluding
asthma (3.4 years), chronic ischemic heart disease (3.1 years), and diabetes mellitus (4.4 years).

                                          3-107

-------
A subsequent follow-up study found elevated mortality rates from circulatory causes in the
Canadian study population. Despite differences in the vehicle fleets of Europe and Canada,
whose emission standards largely mirror those of the U.S., the results of these studies are similar.

       One cohort study conducted in the United Kingdom examined cardiocerebral (stroke)
mortality in relation to living near traffic.339 Those living in census areas near roadways had
significantly higher stroke mortality rates.  In a study involving nearly 190,000 stroke deaths in
1990-1992, Maheswaran and Elliott (2002) examined stroke mortality rates in census districts
throughout England and Wales. Census districts closest to major roads showed significant
increases in stroke mortality rates for men and women.  Compared to those living in census
districts whose center was greater than 1000 m from a main road, men and women living in
census regions with centers less than 200 m away had stroke mortality rates 7% and 4% higher,
respectively.

       One study from the Netherlands used time-series analysis to evaluate  the change in the
magnitude of the association between daily concentrations of black smoke, an air metric related
to black carbon, and daily deaths, for populations living along roads with at least 10,000 vehicles
per day.340  Compared with the population living elsewhere, the traffic-exposed population had
significantly higher associations between black smoke and daily mortality.

       Although the  studies of mortality have employed different study designs  and metrics of
exposure, they provide evidence for increased mortality rates in proximity of heavy traffic. In
evaluating the generalizability of these study results,  questions remain regarding differences in
housing stock, residential ventilation, vehicle type  and fuel differences, personal activity
patterns, and the appropriate exposure metric.  Furthermore, in the cohort studies, although
controls for income level were incorporated based  on postal code or census area, it is possible
that other unmeasured covariates explain the associations with traffic.

3.5.2   Non-Allergic Respiratory Symptoms

       Our analysis of the benefits associated with reduced exposure to PM2.5 includes chronic
bronchitis, hospital admissions for respiratory causes, emergency room visits for asthma,  acute
bronchitis, upper and lower respiratory symptoms and exacerbation of asthma. In addition,
studies in Europe, Asia and North America have found  increased risk of respiratory symptoms
such as wheeze, cough, chronic phlegm production, and dyspnea (shortness of breath) in children
and adults with increased proximity to roadways and/or associated with local traffic density.
Most of these studies were cross-sectional and relied solely on questionnaire  assessments of
health outcomes, in combination with simple exposure indicators. There are  a large number of
studies available, but for the sake of brevity, only studies conducted in the United States are
discussed here. European studies reach similar conclusions, as summarized in a recent review of
                      1A1
the European literature.    The discussion below covers all studies conducted in the United
States. EPA has not formally evaluated the extent  to which these studies may be documenting
health effects that are already included in the benefits analysis associated with PM.

       Most recently, a study from Cincinnati, OH examined the prevalence  of wheezing in a
group of infants less than one year of age.342  Infants  with at least one atopic parent qualified for

                                          3-108

-------
enrollment.  The study compared infants living near stop-and-go truck traffic with others living
near smoothly-flowing truck traffic, and others further from traffic. Infants with wheeze were
significantly more likely to live near stop-and-go traffic than either those living near smoothly-
flowing traffic or those living away from traffic. Truck volume was not associated with wheeze.

       A respiratory health study in the east San Francisco Bay area looked at a series of
community schools upwind and downwind of major roads along a major transportation corridor,
where ambient air quality was monitored.343 Over 1,100 children in grades three through five
attending the schools were assessed for respiratory symptoms and physician's diagnosis of
asthma. Overall, concentrations of traffic-related air pollutants measured at each school were
associated with increased prevalence of bronchitis symptoms and physician confirmed asthma,
both within the last 12 months.

       A case-control study in Erie County, NY compared home proximity to traffic among
children admitted into local hospitals for asthma with those admitted for non-respiratory
conditions.344 Overall, children hospitalized for asthma were more likely to live within 200
meters of roads above the 90th percentile of daily vehicle miles traveled, and to have trucks and
trailers passing within 200 meters of their residences. However, hospitalization for asthma was
not associated with residential distance from major state routes.

       A study in San Diego County, C A compared the residential location of asthmatic children
with children having a non-respiratory diagnosis within the state Medicaid system.345  Traffic
volumes on streets nearby the home were not associated with the prevalence of asthma.
However, among asthmatic children, high street volumes on the nearest street were  associated
with an increased annual frequency of medical visits for asthma.

       In the only U.S. study examining adult respiratory symptoms, Massachusetts veterans
were evaluated for traffic-health relationships.346  In the study, living within 50 m of a major
roadway was associated with increased reporting of persistent wheeze.  This trend held only for
roads with at least 10,000 vehicles per day. Patients experiencing chronic phlegm were also
more likely to live within 50 meters of roads with at least 10,000 vehicles per day. However,
chronic cough was not associated with living near traffic.

       The studies described above employ different exposure metrics and health endpoints,
making evaluation difficult. However, numerous other studies from around the world also
provide evidence for increased prevalence of respiratory symptoms among people living near
major roads. For a detailed listing, refer to the docket of this proposal.  Taken together, these
studies provide substantial evidence that respiratory symptoms may be associated with living
near major roadways, particularly in children, upon whom the preponderance of studies have
focused.

3.5.3  Development of Allergic Disease and Asthma

       A significant number of studies have examined  evidence of a role of traffic-generated
pollution in the development (e.g. new onset) of atopic  illnesses (i.e.,  hypersensitivity to
allergens), such as asthma, allergic rhinitis, and dermatitis. A critical review of evidence,

                                          3-109

-------
primarily generated in European studies, was recently published.347  Overall, the review
concluded that there is some limited evidence of an association between traffic-generated
pollutants and asthma incidence. Toxicological evidence provides some evidence that particles
from diesel engine exhaust may serve as adjuvants to IgE-mediated immune responses.  EPA's
Health Assessment Document for Diesel Engine Exhaust addresses many of the toxicological
studies on diesel exhaust.  However, in community epidemiology studies, the evidence remains
tentative.  The potential for these effects is not taken into account in the benefits analysis for PM
because EPA's various scientific advisors have argued that the literature is not strong enough to
support a causal  association.

3.5.4   Cardiovascular Effects

       Cardiovascular effects are currently seen as a potentially important mechanism whereby
PM2.5 may be leading to premature mortality. Non-fatal acute myocardial infarction and
cardiovascular hospital admissions are also PM-related cardiovascular effects.  The studies
described in Section 3.5.1  found higher relative risks for cardiopulmonary causes of death.

       Several additional studies have provided suggestive evidence that exposure to fresh
emissions from traffic predispose people to adverse cardiovascular events. Studies have focused
on both short-term variations in exposure, as well as long-term residential history. As discussed
in the summary section below, there are stressors in the roadway environment in addition to
ambient air pollutants (e.g., noise, anxiety) that also have an impact on cardiovascular activity.
The potential role of these co-stressors has not been adequately investigated.

       A study from Augsburg, Germany interviewed survivors of myocardial infarction (MI)
shortly after they had recovered to examine ambient pollution and activities that might
predispose someone to having a heart attack.348 Survivors of MI were nearly three times as
likely to be in a car, in transit, or on a bicycle in the hour prior to the event as they were to be in
traffic at other times. Ambient air pollutants measured in the hour prior to MI at a central site in
the city were not associated with the risk of MI.

       A study of healthy young North Carolina state patrolmen conducted by EPA's Office of
Research and Development monitored in-vehicle concentrations of PM2.5, VOCs, and metals.349
In-vehicle PM2 5 concentrations were associated with altered heart rate variability, an indicator of
cardiac stress. In-vehicle concentrations were also associated with increased concentrations of
factors in the blood associated with long-term cardiac risk, such as C-reactive protein, an
indicator of inflammation. This study provides information on possible mechanisms by which
cardiac stress could be induced by exposures to traffic-generated air pollution.

       Heart rate variability has also been measured in a study of elderly residents of the Boston
area.350 In the study, ambient PM2.5 was associated with changes consistent with reduced
autonomic control of the heart.  Black carbon, often a more reliable index of traffic-related
pollution, was also associated with these changes.  In a related study, ST-segment depression, a
cardiographic indicator of cardiac ischemia or inflammation, was associated with black carbon
levels as well.351  These studies further document a hypothesized mechanism associated with
                                          3-110

-------
motor vehicle emissions, but do not necessarily suggest effects independent of those identified in
our discussion of PM health effects.

3.5.5   Birth Outcomes

       Studies examining birth outcomes in populations living near major traffic sources have
generally found evidence of low birth weight, preterm birth, reduced head circumference and
heart defects among children of mothers living in close proximity to heavy traffic.  Our
discussion of PM health effects also quantitatively accounts for premature mortality effects in
infants and qualitatively accounts for low birth weight.

       One measure of exposure to traffic-generated pollution is "distance-weighted traffic
density," where traffic volume is treated as a measure that "disperses" along a Gaussian bell-
shaped curve evenly on both sides of a roadway. This approach captures some of the patterns of
dispersion from line sources, but does not account for micrometeorology. One study from Los
Angeles County, California employed this metric in a study of birth outcomes for births from
1994 to 1996.  The study  showed associations between distance-weighted traffic volume near
women's residences during pregnancy and premature birth and low birth weight in their
babies.352 The elevated risks occurred primarily for mothers whose third trimesters fell during
fall or winter months.

       The same researchers had conducted an earlier study of births occurring between 1989
and 1993. In that study, consisting of over 125,000  births, exposures to ambient carbon
monoxide (CO), an indicator of traffic pollution, during the third trimester were significantly
associated with increased risk of low birth weight.353 In another study, preterm birth was
associated with ambient PMi0 and CO.354 These authors have also reported in a separate study
on the increase in cardiac ventricular septal defects with increasing CO exposure during the
second month of pregnancy.355 The role of socioeconomic status and factors associated with it
should be investigated in future study design.

       Although the exposure metrics employed in these studies are based on surrogate
approaches to exposure estimation, other researchers have shown associations between New
York mothers' measured personal  exposure to polycyclic aromatic hydrocarbons (PAHs) during
pregnancy and an increased risk of low birth weight and size.356

       Overall, although the number of studies examining perinatal exposures is relatively small,
there is some evidence that exposure to traffic-related pollutants may be associated with adverse
birth outcomes, including low birth weight and  preterm birth.  However, given the variety of
exposure metrics employed and the relatively limited geographic extent of studies,  the
generalization of the conclusions requires a better understanding of relevant sources, pollutants,
susceptibility, and local factors.

3.5.6   Childhood Cancer

       Several MSATs are associated with cancer in adult populations. However, children have
physical and biochemical differences that may affect their susceptibility to and metabolism of

                                         3-111

-------
MSATs.  Particularly in the first year or two after birth, infants' liver enzyme profiles undergo
rapid change. As such, children may respond to MSATs in different ways from adults.  Some
evidence exists that children may face different cancer risks from adults as a result of exposure to
certain MSATs and other components of motor vehicle exhaust.  EPA recently recommended
default adjustments to cancer risk estimates to account for early life exposures in the
Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to
Carcinogens.357

       Furthermore, evidence from human and animal studies suggests that increases in
childhood leukemia may be associated with in utero exposures to benzene and maternal and
paternal exposure prior to conception. Furthermore, there is some evidence that key changes
related to the development of childhood leukemia occur in the developing fetus.358

       In the last 15 years, several studies have evaluated the association between maternal or
childhood residence near busy roads and the risk of cancer in children.  Most studies to date have
been ecological in nature, with several employing individual-level exposure estimates within
cohort designs.  The studies employed widely varying exposure metrics, including modeled air
quality, proximity to sources, and distance-weighted traffic volumes. Positive studies tend to
have used small population sizes, although one recent positive study used a large population.
Due to differences in ages studied, study design, exposure metrics, and study location (e.g.
Europe vs. U.S.), a systematic comparison between studies is difficult.  A description of several
key studies from this literature follows.

       One early study from Colorado showed significant elevated risk of childhood leukemia in
children under age  15 associated with living near roads with higher traffic volumes.  The
strongest associations were with roads with at least 10,000 vehicles per day.359 The study was
reanalyzed using an approach to combine traffic volume with residential distance from major
roads to assess "distance-weighted traffic volume."360 The study found that the significant,
monotonically increasing risks associated with increased distance-weighted traffic volume.

       NOi has been used as an indicator of traffic emissions in some studies; however, it is
important to note that NOi is not implicated as causing  cancer. For instance, a study used a
dispersion model of NC>2 from traffic to  conduct a case-control study of childhood cancer in
Sweden.361 The  study found that in the highest-exposed group, risk of any cancer was
significantly elevated. Risks in the most-exposed group were also elevated for leukemia and
central nervous system tumors, but were not statistically significant.

       These earlier studies were based on relatively small populations of children with cancer.
In response, subsequent studies focused on either replicating the earlier studies or studying larger
groups of children. A study in Los Angeles, California applied the same distance-weighted
traffic volume approach as the earlier Colorado study, but found no elevation in risk in a larger
group of children.362  A large study of nearly 2,000 Danish children with cancer found no
association between modeled concentrations of benzene and NC>2 at home and the risk of
leukemia, central nervous system tumors, or total cancers.363 However, the study did find a
dose-dependent relationship between Hodgkin's disease and modeled air pollution from traffic.
                                          3-112

-------
       Several large studies were conducted in California using a statewide registry of cancer.
These studies employed study sizes of several thousand subjects. In one cross-sectional study,
the potency-weighted sum of concentrations of 25 air toxics modeled using EPA's ASPEN
model was not associated with mobile source emissions, but increased rates of childhood
leukemia were found when accounting for all sources of air toxics together, and for point sources
separately.364 Another study from the same researchers found that roadway density and traffic
density within 500 meters of children's homes was not associated with risk of cancer.365

       Most recently,  a novel approach to assessing childhood leukemia in relation to early life
exposures was employed in the United Kingdom. The study examined all children dying of
cancer between 1955 and 1980, consisting of over 22,000 cases.  Birth and death addresses of
children with cancer who moved before death were compared with regard to proximity to nearby
sources and emissions of specific chemicals.366 An excess of births near sources, relative to
deaths, was used to indicate sources in early life associated with greatest cancer. Greater risks
were associated with birth addresses within 300 meters of high emissions of benzene, 1,3-
butadiene, NOX, PMi0, dioxins, and benzo[a]pyrene.  In addition, births within 1.0 km of bus
stations, hospitals, freight terminals,  railways, and oil installations were associated with elevated
risk. Overall, locations with the highest emissions of 1,3-butadiene and carbon monoxide
showed the greatest risk.

       In summary, the lack of consistency in results between large studies and the multiplicity
of study designs makes it difficult to draw firm conclusions.  Epidemiologic methods for
detection of childhood cancer risks may lack sufficient power to detect risks with precision.
However, given the well-established carcinogenicity  of benzene and 1,3-butadiene in the
toxicological  and occupational epidemiologic literature, and data suggesting exposure to benzene
prior to conception and in utero can lead to increased risk of childhood leukemia, the potential
for public health concern is present.  The standards proposed in this rule will reduce such
exposures.

3.5.7   Summary of Near-Roadway Health Studies

       Taken together, the available studies of health effects in residents near major  roadways
suggest a possible public health concern. These studies' exposure metrics are reflective of a
complex mixture from traffic, and the proposed standards will reduce a broad range of pollutants
present in higher concentrations near roadways. The extent to which these health effects are
attributable to PM versus other components of the complex mixture is unclear.

3.5.8   Size and Characteristics of Populations Living near Major Roads

       In assessing the public health implications of near-roadway health concerns, some
understanding of the population living near major roads is required. Those living near major
roadways are a subpopulation of the total population  included in quantitative analysis, and to the
extent that there may be additional exposures and health effects not captured in analyses for the
total population, we enumerate the size and characteristics of the subpopulation. As noted earlier,
we conducted a study in three states, Colorado, Georgia, and New York.  Geographic
information systems were used in the analysis.  In Colorado,  22% live within 75 meters of a

                                          3-113

-------
major road, while an additional 33% live between 75 and 200 meters of major roads.  In Georgia,
the respective percentages are 17% living within 75 meters and an additional 24% living between
75 and 200 meters.  In New York, the percentages are 31% and 36%.

       To date, the only source of national data on populations living in close proximity to major
transportation sources is the American Housing Survey, conducted by the U.S. Census Bureau.367
This study characterizes the properties and neighborhood characteristics of housing units
throughout the U.S.  According to the Census Bureau's summaries of the 2003 survey, among
approximately 120,777,000 housing units in the nation, 15,182,000 were within 300 feet of a "4-
or-more-lane highway, railroad, or airport." This constitutes 12.6% of total U.S. housing units.
A simple assumption that the U.S. population is uniformly distributed among all types of housing
leads to the conclusion that approximately 37.4 million people live in what might be considered a
"mobile source hot spot."

       According to the American Housing Survey's summary tables, occupied housing units in
central cities are 35% more likely to be close to major transportation sources than housing units
in suburban areas.368 Furthermore, nationally, housing units that are renter-occupied are 2.3
times more likely to be close to major transportation sources, compared to housing units that are
owner-occupied. In the 2003 American Housing Survey, median household income for owner-
occupied units was $52,803, while only $26,983 for renter-occupied units. These statistics imply
that those houses sited near major transportation sources are likely to be lower in income than
houses not located near major transportation sources.

       A few population-based epidemiology studies have also examined whether discrete
groups of people live close to major roadways.  In one study of veterans living in southeastern
Massachusetts, 23% lived within  50 meters of a "major road," 33% lived within 100 meters, and
51% within 200 meters.369 In examining traffic volumes, 13% lived within 50 meters of a road
with annual average daily traffic of 10,000 vehicles or more, while other distances were not
analyzed.

       In another study using 150 meters as a definition of "near" a road, 2.3%  of California
public schools were found to be near a road with more than 50,000 vehicles per day, while 7.2%
were near roads with between 25,000 and 49,999 vehicles per day.370  This corresponded to 2.6%
and 9.8% of total enrollment, respectively. In that study, traffic exposure increased, the fractions
of school populations comprised of black and Hispanic students also increased,  as did the
fraction of children in government-subsidized meal programs.

       Another study in California defined the issue differently, examining the  child population
living in census block groups and traffic density.371  The study found that approximately 3% of
the state child population resided  in the highest traffic density census tracts. Furthermore, block
groups with lower income were more likely to have high traffic density.  Children of color were
more likely than white children to live in high traffic density areas.

       In summary, a substantial  fraction of the U.S. population lives within approximately 200
meters of major roads.
                                          3-114

-------
         Appendix 3A: 8-Hour Ozone Nonattainment
Table 3A-1. 8-Hour Ozone Nonattainment Areas and Populations (Data is Current
through April 2005 and Population Numbers are from 2000 Census Data)
8-hour Ozone
Nonattainment Areas
Albany-Schenectady-
Troy, NY
Allegan Co, Ml
Allentown-Bethlehem-
Easton, PA
Altoona, PA
Amador and Calaveras
Cos (Central Mtn), CA
Atlanta, GA
Baltimore, MD
Baton Rouge, LA
Beaumont-Port Arthur,
TX
Benton Harbor, Ml
Benzie Co, Ml
Berkeley and Jefferson
Counties, WV
Birmingham, AL
Boston-Lawrence-
Worcester (E. MA), MA
Boston-Manchester-
Portsmouth(SE),NH
Buffalo-Niagara Falls,
NY
Canton-Massillon, OH
Cass Co, Ml
Charleston, WV
Charlotte-Gastonia-Rock
Hill, NC-SC
Chattanooga, TN-GA
Chicago-Gary-Lake
County, IL-IN
Chico, CA
Cincinnati-Hamilton, OH-
KY-IN
Clarksville-Hopkinsville,
TN-KY
Clearfield and Indiana
Cos, PA
Cleveland-Akron-Lorain,
OH
Columbia, SC
Columbus, OH
2001-2003
8-hr Design Category /
Population Value (ppb) Classification30 d
923,778
105,665
637,958
129,144
75,654
4,228,492
2,512,431
636,214
385,090
162,453
15,998
118,095
805,340
5,534,130
696,713
1,170,111
378,098
51,104
251,662
1,476,564

372,264
8,757,808
203,171
1,891,518
207,033
172,987
2,945,831
494,518
1,541,930
87
97
91
85
91
91
103
86
91
91
88
86
87
95
95
99
90
93
86
100

88
101
89
96
85
90
103
89
95
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Marginal
Moderate
Marginal
Marginal
Subpart 1
Subpart 1
Subpart 1 EAC
Subpart 1
Moderate
Moderate
Subpart 1
Subpart 1
Marginal
Subpart 1
Moderate

Subpart 1 EAC
Moderate
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Moderate
Subpart 1 EAC
Subpart 1
Maximum
Attainment
Date"
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2007
Jun. 15,2010
Jun. 15,2007
Jun. 15,2007
Jun. 15,2009
Jun. 15,2009
Dec. 31,2007
Jun. 15,2009
Jun. 15,2010
Jun. 15,2010
Jun. 15,2009
Jun. 15,2009
Jun. 15,2007
Jun. 15,2009
Jun. 15,2010

Dec. 31,2007
Jun. 15,2010
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2010
Dec. 31,2007
Jun. 15,2009
                              3-115

-------
8-hour Ozone
Nonattainment Areas
Dallas-Fort Worth, TX
Dayton-Springfield, OH
Denver-Boulder-
Greeley-Ft Collins-Love.,
CO
Detroit-Ann Arbor, Ml
Door Co, Wl
Erie, PA
Essex Co (Whiteface
Mtn), NY
Evansville, IN
Fayetteville, NC
Flint, Ml
Fort Wayne, IN
Franklin Co, PA
Frederick Co, VA
Fredericksburg, VA
Grand Rapids, Ml
Greater Connecticut, CT
Greene Co, IN
Greene Co, PA
Greensboro-Winston
Salem-High Point, NC
Greenville-Spartanburg-
Anderson, SC
Hancock, Knox, Lincoln
& Waldo Cos, ME
Harrisburg-Lebanon-
Carlisle, PA
Haywood and Swain
Cos (Great Smoky NP),
NC
Hickory-Morganton-
Lenoir, NC
Houston-Galveston-
Brazoria, TX
Huntington-Ashland,
WV-KY
Huron Co, Ml
Imperial Co, CA
Indianapolis, IN
Jackson Co, IN
Jamestown, NY
Jefferson Co, NY
Johnson City-Kingsport-
Bristol, TN
Johnstown, PA
Kalamazoo-Battle Creek,
Ml
2001-2003
8-hr Design Category /
Population Value (ppb) Classification30 d
5,030,828
950,558

2,811,580

4,932,383
27,961
280,843
1,000
224,305
302,963
524,045
331,849
129,313
82,794
202,120
812,649
1,543,919
33,157
40,672
1,285,879
799,147
92,476
629,401

288

309,512
4,669,571

189,439
36,079
142,361
1,607,486
41,335
139,750
111,738
206,611
152,598
452,851
100
90

87

97
94
92
91
85
87
90
88
93
85
99
89
95
88
89
93
87
94
88

85

88
102

91
87
87
96
85
94
97
86
87
86
Moderate
Subpart 1

Subpart 1 EAC

Marginal
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1 EAC
Subpart 1
Subpart 1
Subpart 1
Subpart 1 EAC
Moderate
Subpart 1
Moderate
Subpart 1
Subpart 1
Marginal EAC
Subpart 1 EAC
Subpart 1
Subpart 1

Subpart 1

Subpart 1 EAC
Moderate

Subpart 1
Subpart 1
Marginal
Subpart 1
Subpart 1
Subpart 1
Moderate
Subpart 1 EAC
Subpart 1
Subpart 1
Maximum
Attainment
Date"
Jun. 15,2010
Jun. 15,2009

Dec. 31,2007

Jun. 15,2007
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Dec. 31,2007
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Dec. 31 , 2007
Jun. 15,2010
Jun. 15,2009
Jun. 15,2010
Jun. 15,2009
Jun. 15,2009
Dec. 31,2007
Dec. 31,2007
Jun. 15,2009
Jun. 15,2009

Jun. 15,2009

Dec. 31,2007
Jun. 15,2010

Jun. 15,2009
Jun. 15,2009
Jun. 15,2007
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2010
Dec. 31,2007
Jun. 15,2009
Jun. 15,2009
3-116

-------
8-hour Ozone
Nonattainment Areas
Kent and Queen Anne's
Cos, MD
Kern Co (Eastern Kern),
CA
Kewaunee Co, Wl
Knoxville, TN
La Porte, IN
Lancaster, PA
Lansing-East Lansing,
Ml
Las Vegas, NV
Lima, OH
Los Angeles South
Coast Air Basin, CA
Los Angeles-San
Bernardino Cos(W
Mojave),CA
Louisville, KY-IN
Macon, GA
Madison and Page Cos
(Shenandoah NP), VA
ManitowocCo, Wl
Mariposa and Tuolumne
Cos (Southern Mtn),CA
Mason Co, Ml
Memphis, TN-AR
Milwaukee-Racine, Wl
Muncie, IN
Murray Co
(Chattahoochee Nat
Forest), GA
Muskegon, Ml
Nashville, TN
Nevada Co. (Western
Part), CA
New York-N. New
Jersey-Long Island, NY-
NJ-CT
Norfolk-Virginia Beach-
Newport News (HR),VA
Parkersburg-Marietta,
WV-OH
Philadelphia-Wilmin-
Atlantic Ci,PA-NJ-MD-
DE
Phoenix-Mesa, AZ
Pittsburgh-Beaver
Valley, PA
Portland, ME
2001-2003
8-hr Design Category /
Population Value (ppb) Classification3 cd
59,760
99,251
20,187
713,755
110,106
470,658
447,728
1,348,864
108,473
14,593,587

656,408

968,313
153,937
2
82,887
71,631
28,274
948,338
1,839,149
118,769

1,000

170,200
1,097,810
77,735

19,634,122

1,542,144
151,237

7,333,475

3,086,045
2,431,087
456,508
95
98
93
92
93
92
86
86
89
131

106

92
86
87
90
91
89
92
101
88

85

95
86
98

102

90
87

106

87
94
91
Marginal
Subpart 1
Subpart 1
Subpart 1
Marginal
Marginal
Subpart 1
Subpart 1
Subpart 1
Severe 17

Moderate

Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Marginal
Moderate
Subpart 1

Subpart 1

Marginal
Subpart 1 EAC
Subpart 1

Moderate

Marginal
Subpart 1

Moderate

Subpart 1
Subpart 1
Marginal
Maximum
Attainment
Date"
Jun. 15,2007
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2007
Jun. 15,2007
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2021

Jun. 15,2010

Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2007
Jun. 15,2010
Jun. 15,2009

Jun. 15,2009

Jun. 15,2007
Dec. 31,2007
Jun. 15,2009

Jun. 15,2010

Jun. 15,2007
Jun. 15,2009

Jun. 15,2010

Jun. 15,2009
Jun. 15,2009
Jun. 15,2007
3-117

-------

8-hour Ozone
Nonattainment Areas
Poughkeepsie, NY
Providence (All Rl), Rl
Raleigh-Durham-Chapel
Hill, NC
Reading, PA
Richmond-Petersburg,
VA
Riverside Co, (Coachella
Valley), CA
Roanoke, VA
Rochester, NY
Rocky Mount, NC
Sacramento Metro, CA
San Antonio, TX
San Diego, CA
San Francisco Bay Area,
CA
San Joaquin Valley, CA
Scranton-Wilkes-Barre,
PA
Sheboygan, Wl
South Bend-Elkhart, IN
Springfield (Western
MA), MA
St Louis, MO-IL
State College, PA
Steubenville-Weirton,
OH-WV
Sutter Co (Sutter
Buttes), CA
Terre Haute, IN
Tioga Co, PA
Toledo, OH
Ventura Co, CA
Washington Co
(Hagerstown), MD
Washington, DC-MD-VA
Wheeling, WV-OH
York, PA
Youngstown-Warren-
Sharon, OH-PA
Total


Population
717,262
1,048,319
1,244,053
373,638
919,277
324,750

235,932
1,098,201
143,026
1,978,348
1,559,975
2,813,431
6,541,828
3,191,367
699,312
112,646
448,350
814,967

2,504,603
135,758
132,008
1
105,848
41,373
576,119
753,197
131,923
4,452,498
153,172
473,043
715,039
159,271,919
2001-2003
8-hr Design
Value (ppb)
94
95
94
91
94
108

85
88
89
107
89
93
86
115
86
100
93
94

92
88
86
88
87
86
93
95
86
99
87
89
95


Category /
Classification3'0'"
Moderate
Moderate
Subpart 1
Subpart 1
Marginal
Serious

Subpart 1 EAC
Subpart 1
Subpart 1
Serious
Subpart 1 EAC
Subpart 1
Marginal
Serious
Subpart 1
Moderate
Subpart 1
Moderate

Moderate
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Moderate
Subpart 1 EAC
Moderate
Subpart 1
Subpart 1
Subpart 1

Maximum
Attainment
Date"
Jun. 15,2010
Jun. 15,2010
Jun. 15,2009
Jun. 15,2009
Jun. 15,2007
Jun. 15,2013

Dec. 31 , 2007
Jun. 15,2009
Jun. 15,2009
Jun. 15,2013
Dec. 31,2007
Jun. 15,2009
Jun. 15,2007
Jun. 15,2013
Jun. 15,2009
Jun. 15,2010
Jun. 15,2009
Jun. 15,2010

Jun. 15,2010
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009
Jun. 15,2010
Dec. 31,2007
Jun. 15,2010
Jun. 15,2009
Jun. 15,2009
Jun. 15,2009

a) Under the CAA these nonattainment areas are further classified as subpart 1 or subpart 2 (subpart 2 is
further classified as marginal, moderate, serious, severe or extreme) based on their design values. An Early
Action Compact (EAC) area is one that has entered into a compact with the EPA and have agreed to reduce
ground level ozone pollution earlier than the CAA would require in exchange the EPA will defer the
effective date of the nonattainment designation.  The severe designation is denoted as severe-15 or severe-
17 based on the maximum attainment date associated with the classification.
                                             3-118

-------
b) The nonattainment areas covered under subpart 1 will be required to attain the standard no later than 5
years after designation and, in limited circumstances; they may apply for an additional extension of up to 5
years (e.g., 2009 to 2014). The areas classified under subpart 2 have attainment dates ranging from up to 3
years for marginal areas (2007) to up to 20 years for extreme areas (2024).

c) Boston-Manchester-Portsmouth (SE), NH has the same classification as Boston-Lawrence- Worcester
(E. MA), MA.

d) Fredericksburg, VA has the same classification as Washington, DC-MD-VA.
                                             3-119

-------
Table 3A-2. 8-Hour Ozone Nonattainment Counties and Populations (Data is
Current through April 13, 2005 and Population Numbers are from 2000 Census
Data)
State
AL
AL
AZ
AZ
AR
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CO
CO
8-hour Ozone Nonattainment
County
Jefferson Co
Shelby Co
Maricopa Co
Final Co
Crittenden Co
Alameda Co
AmadorCo
Butte Co
Calaveras Co
Contra Costa Co
El Dorado Co
Fresno Co
Imperial Co
Kern Co
Kings Co
Los Angeles Co
Madera Co
Marin Co
Mariposa Co
Merced Co
Napa Co
Nevada Co
Orange Co
Placer Co
Riverside Co
Sacramento Co
San Bernardino Co
San Diego Co
San Francisco Co
San Joaquin Co
San Mateo Co
Santa Clara Co
Solano Co
Sonoma Co
Stanislaus Co
Sutter Co
Tulare Co
Tuolumne Co
Ventura Co
Yolo Co
Adams Co
Arapahoe Co
Whole (W) or Partial
(P) County?
W
W
P
P
W
W
W
W
W
W
P
W
W
P
W
P
W
W
W
W
W
P
W
P
P
W
P
P
W
W
W
W
P
P
W
P
W
W
P
W
W
W
Population
662,047
143,293
3,054,504
31,541
50,866
1,443,741
35,100
203,171
40,554
948,816
124,164
799,407
142,361
649,471
129,461
9,519,338
123,109
247,289
17,130
210,554
124,279
77,735
2,846,289
239,978
1,519,609
1,223,499
1,689,509
2,813,431
776,733
563,598
707,161
1,682,585
394,542
413,716
446,997
25,014
368,021
54,501
753,197
168,660
348,618
487,967
                                  3-120

-------
State
CO
CO
CO
CO
CO
CO
CO
CT
CT
CT
CT
CT
CT
CT
CT
DE
DE
DE
DC
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
IL
IL
IL
8-hour Ozone Nonattainment
County
Boulder Co
Broomfield Co
Denver Co
Douglas Co
Jefferson Co
Larimer Co
Weld Co
Fairfield Co
Hartford Co
Litchfield Co
Middlesex Co
New Haven Co
New London Co
Tolland Co
Windham Co
Kent Co
New Castle Co
Sussex Co
Entire District
Barrow Co
Bartow Co
Bibb Co
Carroll Co
Catoosa Co
Cherokee Co
Clayton Co
Cobb Co
Coweta Co
De Kalb Co
Douglas Co
Fayette Co
Forsyth Co
Fulton Co
Gwinnett Co
Hall Co
Henry Co
Monroe Co
Murray Co
Newton Co
Paulding Co
Rockdale Co
Spalding Co
Walton Co
Cook Co
Du Page Co
Grundy Co
Whole (W) or Partial
(P) County?
W
W
W
W
W
P
P
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
P
P
W
W
W
W
W
W
W
P
Population
269,814
38,272
554,636
175,766
525,507
239,000
172,000
882,567
857,183
182,193
155,071
824,008
259,088
136,364
109,091
126,697
500,265
156,638
572,059
46,144
76,019
153,887
87,268
53,282
141,903
236,517
607,751
89,215
665,865
92,174
91,263
98,407
816,006
588,448
139,277
119,341
50
1,000
62,001
81,678
70,111
58,417
60,687
5,376,741
904,161
6,309
3-121

-------
State
IL
IL
IL
IL
IL
IL
IL
IL
IL
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
KY
KY
KY
KY
KY
KY
KY
KY
LA
LA
LA
LA
LA
8-hour Ozone Nonattainment
County
Jersey Co
Kane Co
Kendall Co
Lake Co
Me Henry Co
Madison Co
Monroe Co
StClairCo
Will Co
Allen Co
Boone Co
Clark Co
Dearborn Co
Delaware Co
Elkhart Co
Floyd Co
Greene Co
Hamilton Co
Hancock Co
Hendricks Co
Jackson Co
Johnson Co
Lake Co
La Porte Co
Madison Co
Marion Co
Morgan Co
Porter Co
St Joseph Co
Shelby Co
Vanderburgh Co
Vigo Co
Warrick Co
Boone Co
Boyd Co
Bullitt Co
Campbell Co
Christian Co
Jefferson Co
Kenton Co
Oldham Co
Ascension Par
East Baton Rouge Par
Iberville Par
Livingston Par
West Baton Rouge Par
Whole (W) or Partial
(P) County?
W
W
P
W
W
W
W
W
W
W
W
W
P
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
Population
21,668
404,119
28,417
644,356
260,077
258,941
27,619
256,082
502,266
331,849
46,107
96,472
10,434
118,769
182,791
70,823
33,157
182,740
55,391
104,093
41,335
115,209
484,564
110,106
133,358
860,454
66,689
146,798
265,559
43,445
171,922
105,848
52,383
85,991
49,752
61,236
88,616
72,265
693,604
151,464
46,178
76,627
412,852
33,320
91,814
21,601
3-122

-------
State
ME
ME
ME
ME
ME
ME
ME
ME
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
8-hour Ozone Nonattainment
County
Androscoggin Co
Cumberland Co
Hancock Co
Knox Co
Lincoln Co
Sagadahoc Co
Waldo Co
York Co
Anne Arundel Co
Baltimore Co
Calvert Co
Carroll Co
Cecil Co
Charles Co
Frederick Co
Harford Co
Howard Co
Kent Co
Montgomery Co
Prince George's Co
Queen Anne's Co
Washington Co
Baltimore (City)
Barnstable Co
Berkshire Co
Bristol Co
Dukes Co
Essex Co
Franklin Co
Hampden Co
Hampshire Co
Middlesex Co
Nantucket Co
Norfolk Co
Plymouth Co
Suffolk Co
Worcester Co
Allegan Co
Benzie Co
Berrien Co
Calhoun Co
Cass Co
Clinton Co
Eaton Co
Genesee Co
Huron Co
Whole (W) or Partial
(P) County?
P
P
P
P
P
W
P
P
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
Population
3,390
252,907
29,805
33,563
28,504
35,214
604
164,997
489,656
754,292
74,563
150,897
85,951
120,546
195,277
218,590
247,842
19,197
873,341
801,515
40,563
131,923
651,154
222,230
134,953
534,678
14,987
723,419
71,535
456,228
152,251
1,465,396
9,520
650,308
472,822
689,807
750,963
105,665
15,998
162,453
137,985
51,104
64,753
103,655
436,141
36,079
3-123

-------
State
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
MO
MO
MO
MO
MO
NV
NH
NH
NH
NH
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
8-hour Ozone Nonattainment
County
Ingham Co
Kalamazoo Co
Kent Co
LapeerCo
Lenawee Co
Livingston Co
Macomb Co
Mason Co
Monroe Co
Muskegon Co
Oakland Co
Ottawa Co
StClairCo
Van Buren Co
Washtenaw Co
Wayne Co
Franklin Co
Jefferson Co
St Charles Co
St Louis Co
St Louis
Clark Co
Hillsborough Co
Merrimack Co
Rockingham Co
Strafford Co
Atlantic Co
Bergen Co
Burlington Co
Camden Co
Cape May Co
Cumberland Co
Essex Co
Gloucester Co
Hudson Co
Hunterdon Co
Mercer Co
Middlesex Co
Monmouth Co
Morris Co
Ocean Co
Passaic Co
Salem Co
Somerset Co
Sussex Co
Union Co
Whole (W) or Partial
(P) County?
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
P
P
P
P
P
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
Population
279,320
238,603
574,335
87,904
98,890
156,951
788,149
28,274
145,945
170,200
1,194,156
238,314
164,235
76,263
322,895
2,061,162
93,807
198,099
283,883
1,016,315
348,189
1,348,864
336,518
11,721
266,340
82,134
252,552
884,118
423,394
508,932
102,326
146,438
793,633
254,673
608,975
121,989
350,761
750,162
615,301
470,212
510,916
489,049
64,285
297,490
144,166
522,541
3-124

-------
State
NJ
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
8-hour Ozone Nonattainment
County
Warren Co
Albany Co
Bronx Co
Chautauqua Co
Dutchess Co
Erie Co
Essex Co
Genesee Co
Greene Co
Jefferson Co
Kings Co
Livingston Co
Monroe Co
Montgomery Co
Nassau Co
New York Co
Niagara Co
Ontario Co
Orange Co
Orleans Co
Putnam Co
Queens Co
RensselaerCo
Richmond Co
Rockland Co
Saratoga Co
Schenectady Co
Schoharie Co
Suffolk Co
Wayne Co
Westchester Co
Alamance Co
Alexander Co
Burke Co
Cabarrus Co
Caldwell Co
Caswell Co
Catawba Co
Chatham Co
Cumberland Co
Davidson Co
Davie Co
Durham Co
Edgecombe Co
Forsyth Co
Franklin Co
Whole (W) or Partial
(P) County?
W
W
W
W
W
W
P
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
P
W
P
W
W
P
W
W
W
W
W
W
W
Population
102,437
294,565
1,332,650
139,750
280,150
950,265
1,000
60,370
48,195
111,738
2,465,326
64,328
735,343
49,708
1,334,544
1,537,195
219,846
100,224
341,367
44,171
95,745
2,229,379
152,538
443,728
286,753
200,635
146,555
31,582
1,419,369
93,765
923,459
130,800
33,603
69,970
131,063
64,254
23,501
141,685
21,320
302,963
147,246
34,835
223,314
55,606
306,067
47,260
3-125

-------
State
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
8-hour Ozone Nonattainment
County
Gaston Co
Granville Co
Guilford Co
Haywood Co
Iredell Co
Johnston Co
Lincoln Co
Mecklenburg Co
Nash Co
Orange Co
Person Co
Randolph Co
Rockingham Co
Rowan Co
Swain Co
Union Co
Wake Co
Allen Co
Ashtabula Co
Belmont Co
Butler Co
Clark Co
Clermont Co
Clinton Co
Columbiana Co
Cuyahoga Co
Delaware Co
Fairfield Co
Franklin Co
Geauga Co
Greene Co
Hamilton Co
Jefferson Co
Knox Co
Lake Co
Licking Co
Lorain Co
Lucas Co
Madison Co
Mahoning Co
Medina Co
Miami Co
Montgomery Co
Portage Co
Stark Co
Summit Co
Whole (W) or Partial
(P) County?
W
W
W
P
P
W
W
W
W
W
W
W
W
W
P
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
Population
190,365
48,498
421,048
28
39,885
121,965
63,780
695,454
87,420
118,227
35,623
130,454
91,928
130,340
260
123,677
627,846
108,473
102,728
70,226
332,807
144,742
177,977
40,543
112,075
1,393,978
109,989
122,759
1,068,978
90,895
147,886
845,303
73,894
54,500
227,511
145,491
284,664
455,054
40,213
257,555
151,095
98,868
559,062
152,061
378,098
542,899
3-126

-------
State
OH
OH
OH
OH
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
Rl
Rl
Rl
Rl
Rl
8-hour Ozone Nonattainment
County
Trumbull Co
Warren Co
Washington Co
Wood Co
Adams Co
Allegheny Co
Armstrong Co
Beaver Co
Berks Co
Blair Co
Bucks Co
Butler Co
Cambria Co
Carbon Co
Centre Co
Chester Co
Clear-field Co
Cumberland Co
Dauphin Co
Delaware Co
Erie Co
Fayette Co
Franklin Co
Greene Co
Indiana Co
Lackawanna Co
Lancaster Co
Lebanon Co
Lehigh Co
Luzerne Co
Mercer Co
Monroe Co
Montgomery Co
Northampton Co
Perry Co
Philadelphia Co
Tioga Co
Washington Co
Westmoreland Co
Wyoming Co
York Co
Bristol Co
Kent Co
Newport Co
Providence Co
Washington Co
Whole (W) or Partial
(P) County?
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
Population
225,116
158,383
63,251
121,065
91,292
1,281,666
72,392
181,412
373,638
129,144
597,635
174,083
152,598
58,802
135,758
433,501
83,382
213,674
251,798
550,864
280,843
148,644
129,313
40,672
89,605
213,295
470,658
120,327
312,090
319,250
120,293
138,687
750,097
267,066
43,602
1,517,550
41,373
202,897
369,993
28,080
381,751
50,648
167,090
85,433
621,602
123,546
3-127

-------
State
SC
SC
SC
SC
SC
SC
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
8-hour Ozone Nonattainment
County
Anderson Co
Greenville Co
Lexington Co
Richland Co
Spartanburg Co
York Co
Anderson Co
Blount Co
Cocke Co
Davidson Co
Hamilton Co
Hawkins Co
Jefferson Co
Knox Co
Loudon Co
Meigs Co
Montgomery Co
Rutherford Co
Sevier Co
Shelby Co
Sullivan Co
SumnerCo
Williamson Co
Wilson Co
Bexar Co
Brazoria Co
Chambers Co
Collin Co
Comal Co
Dallas Co
Denton Co
Ellis Co
Fort Bend Co
Galveston Co
Guadalupe Co
Hardin Co
Harris Co
Jefferson Co
Johnson Co
Kaufman Co
Liberty Co
Montgomery Co
Orange Co
Parker Co
Rockwall Co
Tarrant Co
Whole (W) or Partial
(P) County?
W
W
P
P
W
P
W
W
P
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
Population
165,740
379,616
181,265
313,253
253,791
102,000
71,330
105,823
20
569,891
307,896
53,563
44,294
382,032
39,086
11,086
134,768
182,023
71,170
897,472
153,048
130,449
126,638
88,809
1,392,931
241,767
26,031
491,675
78,021
2,218,899
432,976
111,360
354,452
250,158
89,023
48,073
3,400,578
252,051
126,811
71,313
70,154
293,768
84,966
88,495
43,080
1,446,219
3-128

-------
State
TX
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
WV
WV
WV
8-hour Ozone Nonattainment
County
Waller Co
Arlington Co
Botetourt Co
Charles City Co
Chesterfield Co
Fairfax Co
Frederick Co
Gloucester Co
Hanover Co
Henrico Co
Isle Of Wight Co
James City Co
Loudoun Co
Madison Co
Page Co
Prince George Co
Prince William Co
Roanoke Co
Spotsylvania Co
Stafford Co
York Co
Alexandria
Chesapeake
Colonial Heights
Fairfax
Falls Church
Fredericksburg
Hampton
Hopewell
Manassas
Manassas Park
Newport News
Norfolk
Petersburg
Poquoson
Portsmouth
Richmond
Roanoke
Salem
Suffolk
Virginia Beach
Williamsburg
Winchester
Berkeley Co
Brooke Co
Cabell Co
Whole (W) or Partial
(P) County?
W
W
W
W
W
W
W
W
W
W
W
W
W
P
P
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
Population
32,663
189,453
30,496
6,926
259,903
969,749
59,209
34,780
86,320
262,300
29,728
48,102
169,599
1
1
33,047
280,813
85,778
90,395
92,446
56,297
128,283
199,184
16,897
21,498
10,377
19,279
146,437
22,354
35,135
10,290
180,150
234,403
33,740
1 1 ,566
100,565
197,790
94,911
24,747
63,677
425,257
11,998
23,585
75,905
25,447
96,784
3-129

-------
State
WV
WV
WV
WV
WV
WV
WV
WV
Wl
Wl
Wl
Wl
Wl
Wl
Wl
Wl
Wl
Wl

8-hour Ozone Nonattainment
County
Hancock Co
Jefferson Co
Kanawha Co
Marshall Co
Ohio Co
Putnam Co
Wayne Co
Wood Co
Door Co
Kenosha Co
Kewaunee Co
Manitowoc Co
Milwaukee Co
Ozaukee Co
Racine Co
Sheboygan Co
Washington Co
Waukesha Co

Whole (W) or Partial
(P) County?
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
Total
Population
32,667
42,190
200,073
35,519
47,427
51,589
42,903
87,986
27,961
149,577
20,187
82,887
940,164
82,317
188,831
112,646
117,493
360,767
159,271,919
3-130

-------
                  Appendix 3B: PM Nonattainment
Table 3B-1. PMi.s Nonattainment Areas and Populations (data is current through
September 2005 and the population numbers are from 2000 census data)
PM2s Nonattainment Area
Atlanta, GA
Baltimore, MD
Birmingham, AL
Canton-Massillon, OH
Charleston, WV
Chattanooga, AL-TN-GA
Chicago-Gary-Lake County, IL-IN
Cincinnati-Hamilton, OH-KY-IN
Cleveland-Akron-Lorain, OH
Columbus, OH
Dayton-Springfield, OH
Detroit-Ann Arbor, Ml
Evansville, IN
Greensboro-Winston Salem-High Point, NC
Harrisburg-Lebanon-Carlisle, PA
Hickory, NC
Huntington-Ashland, WV-KY-OH
Indianapolis, IN
Johnstown, PA
Knoxville, TN
Lancaster, PA
Libby, MT
Liberty-Clairton, PA
Los Angeles-South Coast Air Basin, CA
Louisville, KY-IN
Macon, GA
Martinsburg, WV-Hagerstown, MD
NewYork-N. New Jersey-Long Island, NY-NJ-CT
Parkersburg-Marietta, WV-OH
Philadelphia-Wilmington, PA-NJ-DE
Pittsburgh-Beaver Valley, PA
Reading, PA
Rome, GA
San Joaquin Valley, CA
St. Louis, MO-IL
Steubenville-Weirton, OH-WV
Washington, DC-MD-VA
Wheeling, WV-OH
York, PA
Total
Population
4,231,750
2,512,431
807,612
378,098
251,662
423,809
8,757,808
1,850,975
2,775,447
1,448,503
851,690
4,833,493
277,402
568,294
585,799
141,685
340,776
1,329,185
164,431
599,008
470,658
2,626
21,600
14,593,587
938,905
154,837
207,828
19,802,587
152,912
5,536,911
2,195,054
373,638
90,565
3,191,367
2,486,562
132,008
4,377,935
153,172
381,751
88,394,361
                                  3-131

-------
Table 3B-2.  PMi0 Nonattainment Areas and Populations (data is current through
September 29, 2005 and the population numbers are from 2000 census data)
 PM10 Nonattainment Areas Listed Alphabetically
Classification   Number     2000     EPA     State
               of     Population   Region
             Counties   (thousands)
              NAA
Ajo (Pima County), AZ
Anthony, NM
BonnerCo (Sandpoint), ID
Butte, MT
Clark Co, NV
Coachella Valley, CA
Columbia Falls, MT
Coso Junction, CA
Douglas (Cochise County), AZ
Eagle River, AK
El Paso Co, TX
Eugene-Springfield, OR
Flathead County; Whitefish and vicinity, MT
Fort Hall Reservation, ID
Hayden/Miami, AZ
Imperial Valley, CA
Juneau, AK
Kalispell, MT
LaGrande, OR
Lake Co, OR
Lamar, CO
Lame Deer, MT
Lane Co, OR
Libby, MT
Los Angeles South Coast Air Basin, CA
Lyons Twsp., IL
Medford-Ashland, OR
Missoula, MT
Mono Basin, CA
Mun. of Guaynabo, PR
New Haven Co, CT
New York Co, NY
Nogales, AZ
Ogden, UT
Owens Valley, CA
Paul Spur, AZ
Phoenix, AZ
Pinehurst, ID
Poison, MT
Portneuf Valley, ID
Rillito, AZ
Ronan, MT
Sacramento Co, CA
Moderate
Moderate
Moderate
Moderate
Serious
Serious
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Serious
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Serious
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Serious
Moderate
Serious
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
1
1
1
1
1
1
1
1
1
4
1
1
1
1
1
1
1
1
1
1
1
2
1
1
2
1
1
1
8
3
37
35
1,376
182
4
7
16
195
564
179
5
1
4
120
14
15
12
3
9
1
3
3
14,594
109
78
52
0
92
124
1,537
25
77
7
1
3,112
2
4
66
1
3
1,223
9
6
10
8
9
9
8
9
9
10
6
10
8
10
9
9
10
8
10
10
8
8
10
8
9
5
10
8
9
2
1
2
9
8
9
9
9
10
8
10
9
8
9
AZ
NM
ID
MT
NV
CA
MT
CA
AZ
AK
TX
OR
MT
ID
AZ
CA
AK
MT
OR
OR
CO
MT
OR
MT
CA
IL
OR
MT
CA
PR
CT
NY
AZ
UT
CA
AZ
AZ
ID
MT
ID
AZ
MT
CA
                                        3-132

-------
Salt Lake Co, UT
San Bernardino Co, CA
San Joaquin Valley, CA
Sanders County (part);Thompson Falls and
vicinity, MT
Sheridan, WY
Shoshone Co, ID
Southeast Chicago, IL
Trona, CA
Utah Co, UT
Washoe Co, NV
Weirton, WV
Yuma, AZ
Moderate
Moderate
Serious
Moderate

Moderate
Moderate
Moderate
Moderate
Moderate
Serious
Moderate
Moderate
1
1
7
1

1
1
1
1
1
1
2
1
898
199
3,080
1

16
10
3
4
369
339
15
82
8
9
9
8

8
10
5
9
8
9
3
9
UT
CA
CA
MT

WY
ID
IL
CA
UT
NV
WV
AZ
55 Total Areas
54
28,918
                                          3-133

-------
                     Appendix 3C: Visibility Tables
Table 3C-1. List of 156 Mandatory Class I Federal Areas Where Visibility is an Important
Value (As Listed in 40 CFR 81)*
 State
Area Name
Acreage
Federal
 Land
Manager
Alabama
Alaska



Arizona











Arkansas

California



















Sipsey Wilderness Area
Bering Sea Wilderness Area
Denali NP (formerly Mt. McKinley NP)
Simeonof Wilderness Area
Tuxedni Wilderness Area
Chiricahua National Monument Wilderness
Area
Chiricahua Wilderness Area
Galiuro Wilderness Area
Grand Canyon NP
Mazatzal Wilderness Area
Mount Baldy Wilderness Area
Petrified Forest NP
Pine Mountain Wilderness Area
Saguaro Wilderness Area
Sierra Ancha Wilderness Area
Superstition Wilderness Area
Sycamore Canyon Wilderness Area
Caney Creek Wilderness Area
Upper Buffalo Wilderness Area
Agua Tibia Wilderness Area
Caribou Wilderness Area
Cucamonga Wilderness Area
Desolation Wilderness Area
Dome Land Wilderness Area
Emigrant Wilderness Area
Hoover Wilderness Area
John Muir Wilderness Area
Joshua Tree Wilderness Area

Kaiser Wilderness Area
Kings Canyon NP
Lassen Volcanic NP
Lava Beds Wilderness Area
Marble Mountain Wilderness Area
Minarets Wilderness Area
Mokelumme Wilderness Area
Pinnacles Wilderness Area
Point Reyes Wilderness Area
Redwood NP
12,646
41,113
1,949,493
25,141
6,402
9,440
18,000
52,717
1,176,913
205,137
6,975
93,493
20,061
71,400
20,850
124,117
47,757
4,344
9,912
15,934
19,080
9,022
63,469
62,206
104,311
47,916
484,673
429,690
36,300
22,500
459,994
105,800
28,640
213,743
109,484
50,400
12,952
25,370
27,792
USDA-FS
USDI-FWS
USDI-NPS
USDI-FWS
USDI-FWS
USDI-NPS
USDA-FS
USDA-FS
USDI-NPS
USDA-FS
USDA-FS
USDI-NPS
USDA-FS
USDI-FS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDI-NPS
USDI-BLM
USDA-FS
USDI-NPS
USDI-NPS
USDI-NPS
USDA-FS
USDA-FS
USDA-FS
USDI-NPS
USDI-NPS
USDI-NPS
                                    3-134

-------
State
Area Name
Acreage
Federal
  Land
Manager












Colorado











Florida


Georgia


Hawaii

Idaho




Kentucky
Louisiana
Maine



Michigan

San Gabriel Wilderness Area
San Gorgonio Wilderness Area

San Jacinto Wilderness Area
San Rafael Wilderness Area
Sequoia NP
South Warner Wilderness Area
Thousand Lakes Wilderness Area
Ventana Wilderness Area
Yolla Bolly-Middle Eel Wilderness Area

Yosemite NP
Black Canyon of the Gunnison Wilderness
Area
Eagles Nest Wilderness Area
Flat Tops Wilderness Area
Great Sand Dunes Wilderness Area
La Garita Wilderness Area
Maroon Bells-Snowmass Wilderness Area
Mesa Verde NP
Mount Zirkel Wilderness Area
Rawah Wilderness Area
Rocky Mountain NP
Weminuche Wilderness Area
West Elk Wilderness Area
Chassahowitzka Wilderness Area
Everglades NP
St. Marks Wilderness Area
Cohotta Wilderness Area
Okefenokee Wilderness Area
Wolf Island Wilderness Area
Haleakala NP
Hawaii Volcanoes NP
Craters of the Moon Wilderness Area3
Hells Canyon Wilderness Area
Sawtooth Wilderness Area
Selway-Bitterroot Wilderness Areab
Yellowstone NPC
Mammoth Cave NP
Breton Wilderness Area
Acadia National Park
Moosehorn Wilderness Area
Edmunds Unit
Baring Unit
Isle Royale NP
Seney Wilderness Area
36,137
56,722
37,980
20,564
142,722
386,642
68,507
15,695
95,152
111,841
42,000
759,172
11,180
133,910
235,230
33,450
48,486
71,060
51,488
72,472
26,674
263,138
400,907
61,412
23,360
1,397,429
17,745
33,776
343,850
5,126
27,208
217,029
43,243
83,800
216,383
988,770
31,488
51,303
5,000+
37,503
7,501
2,706
4,680
542,428
25,150
USDA-FS
USDA-FS
USDI-BLM
USDA-FS
USDA-FS
USDI-NS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDI-BLM
USDI-NPS
USDI-NPS
USDA-FS
USDA-FS
USDI-NPS
USDA-FS
USDA-FS
USDI-NPS
USDA-FS
USDA-FS
USDI-NPS
USDA-FS
USDA-FS
USDI-FWS
USDI-NPS
USDI-FWS
USDA-FS
USDI-FWS
USDI-FWS
USDI-NPS
USDI-NPS
USDI-NPS
USDA-FS
USDA-FS
USDA-FS
USDI-NPS
USDI-NPS
USDI-FWS
USDI-NPS
USDI-FWS
USDI-FWS
USDI-FWS
USDI-NPS
USDI-FWS
                                      3-135

-------
State
Area Name
Acreage
Federal
  Land
Manager
Minnesota

Missouri

Montana











Nevada
New Hampshire

New Jersey
New Mexico








North Carolina




North Dakota

Oklahoma
Oregon






Boundary Waters Canoe Area Wilderness
Area
Voyageurs NP
Hercules-Glades Wilderness Area
Mingo Wilderness Area
Anaconda-Pintlar Wilderness Area
Bob Marshall Wilderness Area
Cabinet Mountains Wilderness Area
Gates of the Mtn Wilderness Area
Glacier NP
Medicine Lake Wilderness Area
Mission Mountain Wilderness Area
Red Rock Lakes Wilderness Area
Scapegoat Wilderness Area
Selway-Bitterroot Wilderness Aread
U. L. Bend Wilderness Area
Yellowstone NPe
Jarbidge Wilderness Area
Great Gulf Wilderness Area
Presidential Range-Dry River Wilderness
Area
Brigantine Wilderness Area
Bandelier Wilderness Area
Bosque del Apache Wilderness Area
Carlsbad Caverns NP
Gila Wilderness Area
Pecos Wilderness Area
Salt Creek Wilderness Area
San Pedro Parks Wilderness Area
Wheeler Peak Wilderness Area
White Mountain Wilderness Area
Great Smoky Mountains NPf
Joyce Kilmer-Slickrock Wilderness Area9
Linville Gorge Wilderness Area
Shining Rock Wilderness Area
Swanquarter Wilderness Area
Lostwood Wilderness
Theodore Roosevelt NP
Wichita Mountains Wilderness
Crater Lake NP
Diamond Peak Wilderness
Eagle Cap Wilderness
Gearhart Mountain Wilderness
Hells Canyon Wilderness3

Kalmiopsis Wilderness
747,840
114,964
12,315
8,000
157,803
950,000
94,272
28,562
1,012,599
1 1 ,366
73,877
32,350
239,295
251,930
20,890
167,624
64,667
5,552
20,000
6,603
23,267
80,850
46,435
433,690
167,416
8,500
41,132
6,027
31,171
273,551
10,201
7,575
13,350
9,000
5,557
69,675
8,900
160,290
36,637
293,476
18,709
108,900
22,700
76,900
USDA-FS
USDI-NPS
USDA-FS
USDI-FWS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDI-NPS
USDI-FWS
USDA-FS
USDI-FWS
USDA-FS
USDA-FS
USDI-FWS
USDI-NPS
USDA-FS
USDA-FS
USDA-FS
USDI-FWS
USDI-NPS
USDI-FWS
USDI-NPS
USDA-FS
USDA-FS
USDI-FWS
USDA-FS
USDA-FS
USDA-FS
USDI-NPS
USDA-FS
USDA-FS
USDA-FS
USDI-FWS
USDI-FWS
USDI-NPS
USDI-FWS
USDA-NPS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDI-BLM
USDA-FS
                                      3-136

-------
 State
Area Name
Acreage
Federal
  Land
Manager






South Carolina
South Dakota

Tennessee

Texas

Utah




Vermont
Virgin Islands
Virginia

Washington







West Virginia

Wyoming






New Brunswick,
Canada
Mountain Lakes Wilderness
Mount Hood Wilderness
Mount Jefferson Wilderness
Mount Washington Wilderness
Strawberry Mountain Wilderness
Three Sisters Wilderness
Cape Remain Wilderness
Badlands Wilderness
Wind Cave NP
Great Smoky Mountains NPf
Joyce Kilmer-Slickrock Wilderness9
Big Bend NP
Guadalupe Mountains NP
Arches NP
Bryce Canyon NP
Canyonlands NP
Capitol Reef NP
Zion NP
Lye Brook Wilderness
Virgin Islands NP
James River Face Wilderness
Shenandoah NP
Alpine Lakes Wilderness
Glacier Peak Wilderness
Goat Rocks Wilderness
Mount Adams Wilderness
Mount Rainer NP
North Cascades NP
Olympic NP
Pasayten Wilderness
Dolly Sods Wilderness
Otter Creek Wilderness
Bridger Wilderness
Fitzpatrick Wilderness
Grand Teton NP
North Absaroka Wilderness
Teton Wilderness
Washakie Wilderness
Yellowstone NPh
Roosevelt Campobello International Park
23,071
14,160
100,208
46,116
33,003
199,902
28,000
64,250
28,060
241,207
3,832
708,118
76,292
65,098
35,832
337,570
221 ,896
142,462
12,430
12,295
8,703
190,535
303,508
464,258
82,680
32,356
235,239
503,277
892,578
505,524
10,215
20,000
392,160
191,103
305,504
351,104
557,311
686,584
2,020,625
2,721
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDI-FWS
USDI-NPS
USDI-NPS
USDI-NPS
USDA-FS
USDI-NPS
USDI-NPS
USDI-NPS
USDI-NPS
USDI-NPS
USDI-NPS
USDI-NPS
USDA-FS
USDI-NPS
USDA-FS
USDI-NPS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDI-NPS
USDI-NPS
USDI-NPS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDA-FS
USDI-NPS
USDA-FS
USDA-FS
USDA-FS
USDI-NPS
i
* U.S. EPA (2001) Visibility in Mandatory Federal Class I Areas (1994-1998): A Report to Congress.
EPA-452/R-01-008. This document is available in Docket EPA-HQ-OAR-2005-0036.
                                          3-137

-------
a) Hells Canyon Wilderness Area, 192,700 acres overall, of which 108,900 acres are in Oregon and
83,800 acres are in Idaho.

b) Selway Bitterroot Wilderness Area, 1,240,700 acres overall, of which 988,700 acres are in Idaho and
251,930 acres are in Montana.

c) Yellowstone National Park, 2,219,737 acres overall, of which 2,020,625 acres are in Wyoming,
167,624 acres are in Montana, and 31,488 acres are in Idaho

d) Selway-Bitterroot Wilderness Area, 1,240,700 acres overall, of which 988,770 acres are in Idaho and
251,930 acres are in Montana.

e) Yellowstone National Park, 2,219,737 acres overall, of which 2,020,625 acres are in Wyoming,
167,624 acres are in Montana, and 31,488 acres are in Idaho.

f) Great Smoky Mountains National Park, 514,758 acres overall, of which 273,551 acres are in North
Carolina, and 241,207 acres are in Tennessee.

g) Joyce Kilmer-Slickrock Wilderness Area, 14,033 acres overall, of which 10,201 acres are in North
Carolina, and 3,832 acres are in Tennessee.

h) Yellowstone National Park, 2,219,737 acres overall, of which 2,020,625 acres are in Wyoming,
167,624 acres are in Montana, and 31,488 acres are in Idaho.

i) Chairman, RCIP Commission.

Abbreviations Used in Table:
USDA-FS: U.S. Department of Agriculture, U.S. Forest Service
USDI-BLM: U.S. Department of Interior, Bureau of Land Management
USDI-FWS: U.S. Department of Interior, Fish and Wildlife Service
USDI-NPS: U.S. Department of Interior, National  Park Service
                                            3-138

-------
Table 3C-2. Current (1998-2002) Visibility, Projected (2015) Visibility, and Natural
Background Levels for the 20% Worst Days at 116 IMPROVE Sites
Class I Area Name"
Acadia
Agua Tibia
Alpine Lakes
Anaconda - Pintler
Arches
Badlands
Bandelier
Big Bend
Black Canyon of the Gunnison
Bob Marshall
Boundary Waters Canoe Area
Bridger
Brigantine
Bryce Canyon
Cabinet Mountains
Caney Creek
Canyonlands
Cape Remain
Caribou
Carlsbad Caverns
Chassahowitzka
Chiricahua NM
Chiricahua W
Craters of the Moon
Desolation
Dolly Sods
Dome Land
Eagle Cap
Eagles Nest
Emigrant
Everglades
Fitzpatrick
Flat Tops
Galiuro
Gates of the Mountains
Gila
Glacier
Glacier Peak
Grand Teton
Great Gulf
Great Sand Dunes
Great Smoky Mountains
Guadalupe Mountains
Hells Canyon
Isle Royale
State
ME
CA
WA
MT
UT
SD
NM
TX
CO
MT
MN
WY
NJ
UT
MT
AR
UT
SC
CA
NM
FL
AZ
AZ
ID
CA
WV
CA
OR
CO
CA
FL
WY
CO
AZ
MT
NM
MT
WA
WY
NH
CO
TN
TX
OR
MI
1998-2002 Baseline
Visibility
(deciviews)b
22.7
23.2
18.0
12.3
12.0
17.3
13.2
18.4
11.6
14.2
20.0
11.5
27.6
12.0
13.8
25.9
12.0
25.9
14.8
17.6
25.7
13.9
13.9
14.7
12.9
27.6
20.3
19.6
11.3
17.6
20.3
11.5
11.3
13.9
11.2
13.5
19.5
14.0
12.1
23.2
13.1
29.5
17.6
18.1
21.1
2015 CAIR Control
Case Visibility0
(deciviews)
21.0
23.2
17.4
12.2
12.1
16.8
13.2
18.3
11.4
14.0
19.0
11.3
25.4
11.9
13.4
24.1
12.0
23.9
14.6
17.9
23.0
13.9
13.9
14.7
12.8
23.9
19.9
19.0
11.4
17.4
19.2
11.3
11.4
14.1
10.8
13.5
19.1
13.8
12.0
21.2
13.0
26.1
17.5
18.0
20.1
Natural
Background
(deciviews)
11.5
7.2
7.9
7.3
7.0
7.3
7.0
6.9
7.1
7.4
11.2
7.1
11.3
7.0
7.4
11.3
7.0
11.4
7.3
7.0
11.5
6.9
6.9
7.1
7.1
11.3
7.1
7.3
7.1
7.1
11.2
7.1
7.1
6.9
7.2
7.0
7.6
7.8
7.1
11.3
7.1
11.4
7.0
7.3
11.2
                                       3-139

-------
Class I Area Name"
James River Face
Jarbidge
Joshua Tree
Joyce Kilmer - Slickrock
Kalmiopsis
Kings Canyon
La Garita
Lassen Volcanic
Lava Beds
Linville Gorge
Lostwood
Lye Brook
Mammoth Cave
Marble Mountain
Maroon Bells - Snowmass
Mazatzal
Medicine Lake
Mesa Verde
Mingo
Mission Mountains
Mokelumne
Moosehorn
Mount Hood
Mount Jefferson
Mount Rainier
Mount Washington
Mount Zirkel
North Cascades
Okefenokee
Otter Creek
Pasayten
Petrified Forest
Pine Mountain
Presidential Range - Dry
Rawah
Red Rock Lakes
Redwood
Rocky Mountain
Roosevelt Campobello
Salt Creek
San Gorgonio
San Jacinto
San Pedro Parks
Sawtooth
Scapegoat
Selway - Bitterroot
Seney
Sequoia
Shenandoah
State
VA
NV
CA
NC
OR
CA
CO
CA
CA
NC
ND
VT
KY
CA
CO
AZ
MT
CO
MO
MT
CA
ME
OR
OR
WA
OR
CO
WA
GA
WV
WA
AZ
AZ
NH
CO
WY
CA
CO
ME
NM
CA
CA
NM
ID
MT
MT
MI
CA
VA
1998-2002 Baseline
Visibility
(deciviews)b
28.5
12.6
19.5
29.5
14.8
23.5
11.6
14.8
16.6
27.9
19.6
23.9
30.2
17.1
11.3
13.1
17.7
12.8
27.5
14.2
12.9
21.4
14.0
15.7
18.9
15.7
11.7
14.0
26.4
27.6
14.7
13.5
13.1
23.2
11.7
12.1
16.5
14.1
21.4
17.7
21.5
21.5
11.4
13.6
14.2
12.3
23.8
23.5
27.6
2015 CAIR Control
Case Visibility0
(deciviews)
25.1
12.8
20.3
26.1
14.4
24.1
11.5
14.6
16.5
24.6
18.7
21.1
27.0
16.8
11.3
13.5
17.1
12.8
25.9
14.0
12.8
20.3
13.7
15.2
19.4
15.2
11.8
14.0
24.7
24.0
14.5
13.8
13.4
20.9
11.7
12.1
16.5
14.1
20.1
17.3
22.1
21.4
11.4
13.5
14.1
12.1
22.6
24.1
23.4
Natural
Background
(deciviews)
11.2
7.1
7.1
11.5
7.7
7.1
7.1
7.3
7.5
11.4
7.3
11.3
11.5
7.7
7.1
6.9
7.3
7.1
11.3
7.4
7.1
11.4
7.8
7.8
7.9
7.9
7.1
7.8
11.5
11.3
7.8
7.0
6.9
11.3
7.1
7.1
7.8
7.1
11.4
7.0
7.1
7.1
7.0
7.2
7.3
7.3
11.4
7.1
11.3
3-140

-------
Class I Area Name"
Sierra Ancha
Sipsey
South Warner
Strawberry Mountain
Superstition
Swanquarter
Sycamore Canyon
Teton
Theodore Roosevelt
Thousand Lakes
Three Sisters
UL Bend
Upper Buffalo
Voyageurs
Weminuche
West Elk
Wind Cave
Wolf Island
Yellowstone
YollaBolly- Middle Eel
Yosemite
Zion
State
AZ
AL
CA
OR
AZ
NC
AZ
WY
ND
CA
OR
MT
AR
MN
CO
CO
SD
GA
WY
CA
CA
UT
1998-2002 Baseline
Visibility
(deciviews)b
13.4
28.7
16.6
19.6
14.7
24.6
16.1
12.1
17.6
14.8
15.7
14.7
25.5
18.4
11.6
11.3
16.0
26.4
12.1
17.1
17.6
13.5
2015 CAIR Control
Case Visibility0
(deciviews)
13.7
26.1
16.5
19.2
15.0
21.9
16.6
12.1
16.8
14.6
15.2
14.1
24.3
17.6
11.4
11.3
15.4
24.9
12.1
16.9
17.4
13.3
Natural
Background
(deciviews)
6.9
11.4
7.3
7.5
6.9
11.2
7.0
7.1
7.3
7.3
7.9
7.2
11.3
11.1
7.1
7.1
7.2
11.4
7.1
7.4
7.1
7.0
a) 116 IMPROVE sites represent 155 of the 156 Mandatory Class I Federal Areas.  One isolated Mandatory Class I
Federal Area (Bering Sea, an uninhabited and infrequently visited island 200 miles from the coast of Alaska), was
considered to be so remote from electrical power and people that it would be impractical to collect routine aerosol
samples.  U.S. EPA (2003) guidance for Tracking Progress Under the Regional Haze Rule. EPA-454/B-03-004.
This document is available in Docket EPA-HQ-OAR-2005-0036.

b) The deciview metric describes perceived visual changes in a linear fashion over its entire range, analogous to the
decibel scale for sound. A deciview of 0 represents pristine conditions. The higher the deciview value, the worse the
visibility, and an improvement in visibility is a decrease in deciview value.

c) The 2015 modeling projections are based on the Clear Air Interstate Rule analyses (EPA, 2005).
                                                 3-141

-------
References for Chapter 3

1 U. S. EPA (2003) National Air Quality and Trends Report, 2003 Special Studies Edition.
Office of Air Quality Planning and Standards, Research Triangle Park, NC. Publication No. EPA
454/R-03-005. http://www.epa.gov/air/airtrends/aqtrnd03/ This document is available in Docket
EPA-HQ-OAR-2005-0036.
2 U. S. EPA (2004) Air Toxics Website, http://www.epa.gov/ttn/atw/stprogs.html
3 U. S. EPA (2004) National Air Toxics Monitoring Strategy, Draft.  Office of Air Quality
Planning and Standards, Research Triangle Park, NC. January 2004.
http://www.epa.gov/ttn/amtic/files/ambient/airtox/atstratl04.pdf
4 Kenski, D;  Koerber, M.; Hafner, H. et al. (2005) Lessons learned from air toxics data. A
national perspective. Environ Manage. June 2005:  19-22.
5 Aleksic, N.; Boynton, G.; Sistla, G.; Perry, J. (2005) Concentrations and trends of benzene in
ambient air over New York State during 1990-2003. Atmos Environ 39: 7894-7905.
6 California Air Resources Board (2005) The California Almanac of Emissions and Air Quality -
2005 Edition, http://www.arb.ca.gov/aqd/almanac/almanac05/almanac2005all.pdf
7 Oommen, R.; Hauser, J.; Dayton, D; Brooks, G. (2005) Evaluating HAP trends: A look at
emissions, concentrations, and regulation analyses for selected metropolitan statistical areas.
Presentation at the 14th International Emissions Inventory Conference: Transforming Emission
Inventories Meeting Future Challenges Today. April 12-14, 2005.
8 Clayton, C.A.; Pellizzari, E.D.; Whitmore, R.W.; et al. (1999) National Human Exposure
Assessment Survey (NHEXAS): distributions and associations of lead, arsenic, and volatile
organic compounds in EPA Region 5. J Exposure Analysis Environ Epidemiol 9: 381-392.
9 Gordon, S.M.; Callahan, P.J.; Nishioka, M.G.;  et al. (1999) Residential environmental
measurements in the National Human Exposure Assessment Survey (NHEXAS) pilot study in
Arizona: preliminary results for pesticides and VOCs. J Exposure Anal Environ Epidemiol 9:
456-470.
10 Payne-Sturges, D.C.; Burke, T.A.; Beysse, P.; et al. (2004) Personal exposure meets risk
assessment:  a comparison of measured and modeled exposures and risk in an urban community.
Environ Health Perspect 112: 589-598.
11 Weisel, C.P.; Zhang, J.; Turpin, B.J.; et al. (2005) Relationships of Indoor, Outdoor, and
Personal Air (RIOPA). Part I. Collection methods and descriptive analyses. Res Rep Health
Effects Inst 130.
12 Kwon, J. (2005) Development of a RIOPA database and evaluation of the effect of proximity
on the potential residential exposure to VOCs from ambient sources. Rutgers, the State
University of New Jersey and University  of Medicine and Dentistry of New Jersey. PhD
dissertation.
13 Weisel, C.P. (2004) Assessment of the  contribution to personal exposures of air toxics from
mobile sources. Final report to U.S. Environmental Protection Agency, Office of Transportation
and Air Quality. This document is available in Docket EPA-HQ-OAR-2005-0036.

                                         3-142

-------
14 Adgate, J.L.; Eberly, L.E.; Stroebel, C.; et al. (2004) Personal, indoor, and outdoor VOC
exposures in a probability sample of children. J Exposure Analysis Environ Epidemiol 14:  S4-
S13.
15 Adgate, J.L.; Church, T.R.; Ryan, A.D.; et al. (2004) Outdoor, indoor, and personal exposures
to VOCs in children.  Environ Health Perspect 112:  1386-1392.
16 Sexton, K.; Adgate, J.L.; Church, T.R.; et al. (2005) Children's exposure to volatile organic
compounds as determined by longitudinal measurements in blood. Environ Health Persepct 113:
342-349.
17 Kinney, P.L.; Chillrud, S.N.; Ramstrom, S.; et al. (2002) Exposures to multiple air toxics in
New York City. Environ Health Perspect 110 (suppl 4): 539-546.
18 Sax, S.N.; Bennett, D.H.;  Chillrud, S.N.; et al. (2004) Differences in source emission rates of
volatile organic compounds in innter-city residences of New York City and Los Angeles. J
Exposure Analysis Environ Epidemiol 14: S95-S109.
19 http://www.epa.gov/eogaptil/module3/category/category.htm This document is available in
Docket EPA-HQ-OAR-2005-0036.
20Kittleson, D.; Watts, W.; Johnson, J. (2002) Diesel aerosol sampling methodology.
Coordinating Research Council report E-43. http://www.crcao.com This document is available in
Docket EPA-HQ-OAR-2005-0036.
21 Zhang, K.M.; Wexler,  A.S.; Zhu, Y.F.; et al. (2004) Evolution of particle number distribution
near roadways.  Part II: the'Road-to-Ambient'process.  Atmos Environ 38: 6655-6665.
22
  Zhu, Y.; Hinds, W.C.; Kim, S.; et al. (2002) Study of ultrafine particles near a major highway
with heavy-duty diesel traffic. Atmos Environ 36: 4323-4335.
23
  Zhu, Y.; Hinds, W.C.; Kim, S.; Sioutas, C. (2002) Concentration and size distribution of
ultrafine particles near a major highway. J Air & Waste Manage Assoc 52: 1032-1042.
24
  Kittelson, D.B.; Watts, W.F.; Johnson, J.P. (2004) Nanoparticle emissions on Minnesota
highways. Atmos Environ 38: 9-19.
25 Reponen, T.; Grinshpun, S.A.; Trakumas, S.; et al. (2003) Concentration gradient patterns of
aerosol particles near interstate highways in the Greater Cincinnati airshed. J Environ Monit 5:
557-562.
26 Bunn, H.J.; Dinsdale, D.; Smith, T.; Grigg, J. (2001) Ultrafine particles in alveolar
macrophages from normal children. Thorax 56: 932-934.
27
  Kittelson, D.B.; Watts, W.F.; Johnson, J.P. (2004) Nanoparticle emissions on Minnesota
highways. Atmos Environ 38: 9-19.
28
  Kittelson, D.B.; Watts, W.F.; Johnson, J.P. (2004) Nanoparticle emissions on Minnesota
highways. Atmos Environ 38: 9-19.
29 Sanders, P.G.; Xu, N; Dalka, T.M.; Maricq, M.M. (2003) Airborne brake wear debris: size
distributions, composition, and a comparison of dynamometer and vehicle tests. Environ Sci
Technol 37:  4060-4069.

                                         3-143

-------
30 Kamens, R.M.; Jang, M.; Lee, S.; et al. (2003) Secondary organic aerosol formation: some
new and exciting insights. American Geophysical Union 5: 02915.
31
  Kupiainen, K.J.; Tervahattu, H; Raisanen, M.; et al. (2005) Size and composition of airborne
particles form pavement wear, tires, and traction sanding. Environ Sci Technol 39: 699-706.
32 Sanders, P.O.; Xu, N; Dalka, T.M.; Maricq, M.M. (2003) Airborne brake wear debris: size
distributions, composition, and a comparison of dynamometer and vehicle tests. Environ Sci
Technol 37:  4060-4069.
33 Hitchins, J.; Morawska, L.; Wolff, R.; Gilbert, D. (2000) Concentrations of submicrometre
particles from vehicle emissions near a major road. Atmos Environ 34: 51-59.
34 Janssen, N.A.H.; van Vliet, P.H.N.; Aarts, F.; et al. (2001) Assessment of exposure to traffic
related air pollution of children attending schools near motorways.  Atmos Environ 35:  3875-
3884.
35 Roorda-Knape M.C.; Janssen N.A.H.; De Hartog J.J.; et al. (1998) Air pollution from traffic in
city districts near major motorways. Atmos Environ 32: 1921-1930.
36 Kinnee, E.J.; Touma, J.S.; Mason, R.; Thurman, J.; Beidler, A.; Bailey, C.; Cook, R.  (2004)
Allocation of onroad mobile emissions to road segments for air toxics modeling in an urban area.
Transportation Res. Part D 9: 139-150.
37 Cohen, J.; Cook, R.; Bailey, C.R.; Carr, E. (2005) Relationship between motor vehicle
emissions of hazardous pollutants, roadway proximity, and ambient concentrations in Portland,
Oregon. Environ Modelling & Software 20: 7-12.
38 Skov, H.; Hansen, A.B.; Lorenzen, G.; et al. (2001) Benzene exposure and the effect  of traffic
pollution in Copenhagen, Denmark. Atmos Environ 35: 2463-2471.
39 Jo, W.; Kim, K.; Park, K.; et al. (2003) Comparison of outdoor and indoor mobile source-
related volatile organic compounds between low- and high-floor apartments. Environ Res 92:
166-171.
40 Fischer, P.H.; Joek, G.;  van Reeuwijk, H.; et al. (2000) Traffic-related differences in  outdoor
and indoor concentrations of particle and volatile organic compounds in Amsterdam. Atmos
Environ 34:  3713-3722.
41 Ilgen, E.; Karfich, N.; Levsen, K.; et al. (2001) Aromatic hydrocarbons in the atmospheric
environment: part I. Indoor versus outdoor sources, the influence of traffic.  Atmos Environ 35:
1235-1252.
42 Rodes, C.; Sheldon,  L.;  Whitaker, D.; et al. (1998) Measuring concentrations of selected air
pollutants inside California vehicles. Final report to California Air Resources Board. Contract
No. 95-339.
43 Sapkota, A.; Buckley, T.J. (2003) The mobile source effect on curbside 1,3-butadiene,
benzene, and particle-bound polycyclic aromatic hydrocarbons assessed at a tollbooth.  J Air
Waste Manage Assoc 53:  740-748.
                                         3-144

-------
44 Sapkota, A.; Buckley, TJ. (2003) The mobile source effect on curbside 1,3-butadiene,
benzene, and particle-bound polycyclic aromatic hydrocarbons assessed at a tollbooth. J Air
Waste Manage Assoc 53: 740-748.
45 http://www.mde.state.md.us/Programs/AirPrograms/airData/dataReport.asp This document is
available in Docket EPA-HQ-OAR-2005-0036.
46 Ilgen, E.; Karfich, N.; Levsen, K.; et al. (2001) Aromatic hydrocarbons in the atmospheric
environment: part I.  Indoor versus outdoor sources, the influence of traffic.  Atmos Environ 35:
1235-1252.
47 Kwon, J. (2005) Development of a RIOPA database and evaluation of the effect of proximity
on the potential residential exposure to VOCs from ambient sources. Rutgers, the State
University of New Jersey and University of Medicine and Dentistry of New Jersey. PhD
dissertation. This document is available in Docket EPA-HQ-OAR-2005-0036.
48 Hoek G.; Meliefste K.; Cyrys J.; et al. (2002) Spatial variability of fine particle concentrations
in three European areas. Atmos. Environ. 36: 4077-4088.
49 Etyemezian V.; Kuhns H.; Gillies J.; et al. (2003) Vehicle-based road dust emission
measurement (III): effect of speed, traffic volume, location, and season on PM10 road dust
emissions in the Treasure Valley, ID. Atmos. Environ. 37: 4583-4593.
50 Harrison R.M.; Tilling R.; Romero M.S.C.; et al. (2003) A study of trace metals and
polycyclic aromatic hydrocarbons in the roadside environment. Atmos. Environ. 37: 2391-2402.
51
  Zhang K.M.; Wexler A.S.; Zhu Y.F.; et al. (2004) Evolution of particle number distribution
near roadways. Part II: the 'Road-to-Ambient' process. Atmos. Environ. 38: 6655-6665.
52 Zhu Y.F.; Hinds W.C.;  Shen S.; Sioutas C. (2004) Seasonal trends of concentration and size
distribution of ultrafine particles near major highways in Los Angeles. Aerosol Sci. Technol. 38:
5-13.
53 Zhu, Y.; Hinds, W.C.; Kim, S.; et al. (2002) Study of ultrafine particles near a major highway
with heavy-duty diesel traffic. Atmos Environ 36: 4323-4335.
54 Zhu, Y.; Hinds, W.C.; Kim, S.; Sioutas, C. (2002) Concentration and size distribution of
ultrafine particles near a major highway. J Air & Waste Manage Assoc 52:  1032-1042.
55 Riediker, M.; Williams, R.; Devlin, R.; et al. (2003) Exposure to particulate matter, volatile
organic compounds, and other air pollutants inside patrol cars.  Environ Sci Technol 37: 2084-
2093.
56 Zielinska, B.; Fujita, E.M.; Sagebiel, J.C.; et al. (2002) Interim data report for Section 211(B)
Tier 2 high end exposure screening study of baseline and oxygenated gasoline. Prepared for
American Petroleum Institute. November 19, 2002. This document is available in Docket EPA-
HQ-OAR-2005-0036.
57 Rodes, C.; Sheldon, L.;  Whitaker, D.; et al. (1998) Measuring concentrations of selected air
pollutants inside California vehicles. Final report to California Air Resources Board. Contract
No. 95-339.
                                         3-145

-------
58 Fitz, D. R.; Winer, A. M.; Colome, S.; et al. (2003) Characterizing the Range of Children's
Pollutant Exposure During School Bus Commutes. Final Report Prepared for the California
Resources Board. This document is available in Docket EPA-HQ-OAR-2005-0036.
59 Sabin, L.D.; Behrentz, E.; Winer, A.M.; et al. (2005) Characterizing the range of children's air
pollutant exposure during  school bus commutes. J Expos Anal Environ Epidemiol 15: 377-387.
60 Behrentz, E.; Sabin, L.D.; Winer, A.M.; et al. (2005) Relative importance of school bus-
related microenvironments to children's pollutant exposure. J Air & Waste Manage Assoc 55:
1418-1430.
61 Batterman, S.A.; Peng, C.Y.; and Braun, J. (2002) Levels and composition of volatile organic
compounds on commuting routes in Detroit, Michigan. Atmos Environ 36: 6015-6030.
62 Fruin, S.A.; Winer, A.M.; Rodes, C.E. (2004) Black carbon concentrations in California
vehicles and estimation of in-vehicle diesel exhaust particulate matter exposures. Atmos Environ
38:4123-4133.
63 Adams, H.S.; Nieuwenhuijsen, M.J.; Colvile, R.N. (2001) Determinants of fine particle
(PM2.5) personal exposure levels in transport microenvironments, London, UK. Atmos Environ
35: 4557-4566.
64 Leung, P.-L.; Harrison, R.M. (1999) Roadside and in-vehicle concentrations of monoaromatic
hydrocarbons. Atmos Environ 33: 191-204.
65 Weinhold, B. (2001) Pollutants lurk inside vehicles. Environ Health Perspec  109 (9): A422-
A427.
66 Riediker, M.; Williams, R.; Devlin, R.; et al. (2003) Exposure to particulate matter, volatile
organic compounds, and other air pollutants inside patrol cars. Environ Sci Technol 37: 2084-
2093.
67 Van Wijnen J.H.; Verhoeff A.P.; Jans H.W.A.; Van Bruggen M. (1995) The exposure of
cyclists, car drivers and pedestrians to traffic-related air pollutants. Int Arch Occup Environ
Health 67: 187-193.
68 Chan C.-C.; Ozkaynak H.; Spengler J.D.;  Sheldon L. (1991) Driver Exposure to Volatile
Organic Compounds,  CO, Ozone, andNO2 under Different Driving Conditions. Environ. Sci.
Technol. 25: 964-972.
69 Shikiya, D.C., C.S.  Liu, M.I. Kahn, et al.  (1989) In-vehicle air toxics characterization study
in the south coast air basin.  South Coast Air Quality Management District, El Monte, CA. May,
1989. This document  is available in Docket EPA-HQ-OAR-2005-0036.
70 Chan C.-C., Spengler J.  D., Ozkaynak H., and Lefkopoulou M. (1991) Commuter Exposures to
VOCs in Boston, Massachusetts. J. Air Waste Manage. Assoc. 41,  1594-1600.
71 U.S. EPA (2000) Development of microenvironmental factors  for the HAPEM4 in support of
the National Air Toxics Assessment (NATA). External Review Draft Report Prepared by ICF
Consulting and TRJ Environmental,  Inc.  for the U.S. EPA, Office of Air Quality Planning and
Standards, Research Triangle Park, NC. This document is available in Docket EPA-HQ-OAR-
2005-0036.

                                        3-146

-------
72 U.S. EPA (2000) Determination of microenvironmental factors for diesel PM. An addendum
to: Development of microenvironmental factors for the HAPEM4 in support of the National Air
Toxics Assessment (NATA). External Review Draft Report Prepared by ICF Consulting and
TRJ Environmental, Inc. for the U.S. EPA, Office of Air Quality Planning and Standards,
Research Triangle Park, NC. This document is available in Docket EPA-HQ-OAR-2005-0036.
73 Personal communication with FACES Investigators Fred Lurmann, Paul Roberts, and
Katharine Hammond. Data is currently being prepared for publication.
74 Kim J.J.; Smorodinsky S.; Lipsett M.; et al. (2004) Traffic-related air pollution near busy
roads. Am J Respir Crit Care Med 170:  520-526.
75 Korenstein, S. and Piazza, B. (2002) An Exposure Assessment of PM10 from a Major
Highway Interchange: Are Children in Nearby Schools at Risk? J Environ Health 65(2): 9-17.
76 Janssen, N.A.H.; van Vliet, P.H.N.; Aarts, F.; et al. (2001) Assessment of exposure to traffic
related air pollution of children attending schools near motorways. Atmos Environ 35: 3875-
3884.
77 Roorda-Knape M.C.; Janssen N.A.H.; De Hartog J.J.; et  al. (1998) Air pollution from traffic in
city districts near major motorways. Atmos Environ 32: 1921-1930.
78 Kinney, P.L.; Chillrud, S.N.; Ramstrom, S.;  et al. (2002) Exposures to multiple air toxics in
New York City. Environ Health Perspect 110  (Suppl 4): 539-546.
79 Weisel, C.P. (2002) Assessing exposure to air toxics relative to asthma.  Environ Health
Perspect 110 (Suppl 4): 527-537.
80 Naumova, Y.Y.; Eisenreich, S.J.; Turpin, B.J.; et al. (2002) Polycyclic aromatic hydrocarbons
in the indoor and outdoor environment of three cities in the U.S.  Environ Sci Technol 36: 2552-
2559.
81 Chatzis, C.; Alexopoulos, E.G.; and Linos, A. (2005) Indoor and outdoor personal exposure to
benzene in Athens, Greece. Sci Total Environ 349: 72-80.
82 Gulliver J.; Briggs DJ. (2004) Personal exposure to particulate air pollution in transport
microenvironments. Atmos Environ 38: 1-8.
83 Van Wijnen J.H.; Verhoeff A.P.; Jans H.W.A.; Van Bruggen M. (1995) The exposure of
cyclists, car drivers and pedestrians to traffic-related air pollutants. Int Arch Occup Environ
Health 67: 187-193.
84 Chan C.-C.; Ozkaynak H.; Spengler J.D.; Sheldon L. (1991) Driver Exposure to Volatile
Organic Compounds, CO, Ozone, andNO2 under Different Driving Conditions. Environ. Sci.
Technol. 25: 964-972.
85 Chan C.-C., Spengler J. D., Ozkaynak H., and Lefkopoulou M.  (1991) Commuter Exposures
to VOCs in Boston, Massachusetts. J. Air Waste Manage. Assoc. 41, 1594-1600.
86 Duci, A.; Chaloulakou, A.; Spyrellis N. (2003) Exposure to carbon monoxide in the Athens
urban area during commuting. Sci Total Environ 309: 47-58.
87 Gulliver J.; Briggs D.J. (2004) Personal exposure to particulate air pollution in transport
microenvironments. Atmos Environ 38: 1-8.
                                        3-147

-------
  Ashmore, M.R.; Batty, K.; Machin, F.; et al. (2000) Effects of traffic management and
transport mode on the exposure of schoolchildren to carbon monoxide. Environ Monitoring and
Assessment 65: 49-57.
89 Atkinson, R.; Arey, J.; Hoover, S.; Preston, K. (2005) Atmospheric Chemistry of Gasoline-
Related Emissions: Formation of Pollutants of Potential Concern. Draft Report Prepared for
California Environmental Protection Agency. September 2005. This document is available in
Docket EPA-HQ-OAR-2005-0036.
90 U.S. EPA (1997) Exposure factors handbook. This document is available in Docket EPA-HQ-
OAR-2005-0036. http ://www.epa. gov/ncea
91 Sheltersource, Inc. (2002) Evaluating Minnesota homes.  Final report. Prepared for Minnesota
Department of Commerce.  U.S. Department of Energy grant DE-FG45-96R530335. This
document is available in Docket EPA-HQ-OAR-2005-0036.
92 Fugler, D.; Grande, C.; Graham, L. (2002) Attached garages are likely path for pollutants.
ASHRAE IAQ Applications 3(3): 1-4.
93 Emmerich, S.J.; Gorfain, I.E.; Howard-Reed, C. (2003) Air and pollutant transport from
attached garages to residential living spaces - literature review and field tests.  In J Ventilation 2:
265-276.
94 Isbell, M.; Gordian, M.E.; Duffy,  L. (2002) Winter indoor air pollution in Alaska: identifying
a myth. Environ Pollution 117:  69-75.
95 Wallace, L. (1996) Environmental exposure to benzene: an update. Environ Health Perspect
104 (Suppl6):  1129-1136.
96 Thomas, K. W.; Pellizzari, E. D.;  Clayton, C. A.; Perrit, R.; Dietz, R. N.; Goorich, R. W.;
Nelson, W.; Wallace, L. 1993. Temporal Variability  of Benzene Exposures in several New
Jersey homes with attached garages  or tobacco smoke. J. Expos.  Anal. Environ. Epi. 3: 49-73.
97 Batterman, S.; Hatzivasilis, G.; Jia, C. (2005) Concentrations and emissions of gasoline and
other vapors from residential vehicle garages. Atmos  Environ (in press).
98 George, M.; Kaluza, P.; Maxwell, B.; et al. (2002) Indoor air quality & ventilation strategies in
new homes in Alaska.  Alaska Building Science Network. [Online at http://www.cchrc.org] This
document is available in Docket EPA-HQ-OAR-2005-0036.
99 Graham, L.A.; Noseworthy, L.; Fugler, D.; et al. (2004) Contribution of vehicle emissions
from an attached garage to residential indoor air pollution levels.  J Air & Waste Manage Assoc
54: 563-584.
100 Schlapia, A.; Morris, S.S. (1998) Architectural, behavioral, and environmental factors
associated with VOCs in Anchorage homes.  Proceedings of the Air & Waste Management
Association's 94th Annual Conference & Exhibition.  Paper 98-A504.
101 Isbell, M.; Ricker, J.; Gordian, M.E.; Duffy, L.K. (1999) Use of biomarkers in an indoor air
study:  lack of correlation between aromatic VOCs with respective urinary biomarkers.  Sci Total
Environ 241:  151-159.
                                         3-148

-------
102 Isbell, M.A.; Stolzberg, R.J.; Duffy, L.K. (2005) Indoor climate in interior Alaska:
simultaneous measurement of ventilation, benzene and toluene in residential indoor air of two
homes.  Sci Total Environ 345: 31-40.
103 Gordon, S.M.; Callahan, P.J.; Nishioka, M.G.; et al. (1999) Residential environmental
measurements in the National Human Exposure Assessment Survey (NHEXAS) pilot study in
Arizona: preliminary results for pesticides and VOCs. J Exposure Analysis Environ Epidemiol
9: 456-470.
104 Bonanno, L.J.; Freeman, N.C.G.; Greenberg, M.; Lioy, PJ. (2001) Multivariate analysis on
levels of selected metals, particulate matter, VOC, and household characteristics and  activities
from the Midwestern states NHEXAS. Appl Occup Environ Hygiene 16: 859-874.
105 Tsai, P.; Weisel, C.P. (2000) Penetration of evaporative emissions into a home from an M85-
fueled vehicle parked in an attached garage.  J Air & Waste Manage Assoc 50:  371-377.
106 Bailey, C.R. (2005) Additional contribution to benzene exposure from attached garages.
Memorandum to docket. This document is available in Docket EPA-HQ-OAR-2005-0036.
107 Wilson, A.L.; Colome, S.D.; and Tian, Y. (1991) Air toxics microenvironment exposure and
monitoring study. Final Report. Prepared for South Coast Air Quality Management District and
U.S. Environmental Protection Agency. This document is available in Docket EPA-HQ-OAR-
2005-0036.
108 Zielinska, B.; Fujita, E.M.; Sagebiel, J.C.; et al. (2002) Interim data report for Section 211(B)
Tier 2 high end exposure screening study of baseline and oxygenated gasoline.  Prepared for
American Petroleum Institute. November 19, 2002. This document is available in Docket EPA-
HQ-OAR-2005-0036.
109 Lee,  S.C.; Chan, L.Y.; and Chiu, M.Y. (1999) Indoor and outdoor air quality investigation at
14 public places in Hong Kong. Environ Int 25: 443-450.
no
   Wong, Y.-c.; Sin, D.W.-m.; and Yeung L.L. (2002) Assessment of the air quality in indoor
car parks. Indoor & Built Environment 11:  134-145.
111 Chaloulakou, A.; Duci, A.; Spyrellis, N. (2002) Exposure to carbon monoxide in enclosed
multi-level parking garages in the central Athens urban area. Indoor & Built Environment 11:
191-201.
112 Srivastava, A.; Joseph, A.E.; and Nair, S. (2004) Ambient levels of benzene in Mumbai city.
Int J Environ Health Res 14 (3): 215-222.
113
   Schwar, M.; Booker, J.; Tait, L. (1997) Car Park Air Pollution Exposure of Operatives and
the General Public. Clean Air & Environ Protection 27 (5): 129-137
114
   Morillo, P.; Dos Santos, S.G.; Santamaria, J.; et al. (1998) A study of the atmospheric
pollution produced by vehicles in car parks in Madrid, Spain. Indoor Built Environ 7: 156-164.
115 U.S. EPA (1987) The Total Exposure Assessment Methodology (TEAM) Study: Summary
and Analysis: Volume I. Office of Research and Development, Washington, D.C. June 1987.
EPA Report No. EPA/600/6-87/002a.
                                         3-149

-------
116 Wilson, A.L.; Colome, S.D.; and Tian, Y. (1991) Air toxics microenvironment exposure and
monitoring study, Final Report. Prepared for South Coast Air Quality Management District and
U.S. Environmental Protection Agency. This document is available in Docket EPA-HQ-OAR-
2005-0036.
117 Hartle, R. (1993) Exposure to methyl tert-butyl ether and benzene among service station
attendants and operators. Environ Health Perspect Supplements: 101  (Suppl. 6): 23-26.
118 Northeast States for Coordinated Air Use Management (1999) RFG/MTBE Findings and
Recommendations. August 1999. This document is available in Docket EPA-HQ-OAR-2005-
0036.
119 Vayghani, S.A.; Weisel, C. (1999) The MTBE air concentrations in the cabin of automobiles
while fueling. J Expos Analysis Environ Epidem 9: 261-267.
120 Verma, O.K.; Johnson, D.M.; Shaw, M.L.; et al. (2001) Benzene and total hydrocarbon
exposures in the downstream petroleum industries.  Am Indust Hygiene Assoc J 62: 176-194.
121 Verma, O.K. and des Tombe, K. (2002) Benzene in gasoline  and crude oil: occupational and
environmental implications. Am Indust Hygiene Assoc J 63: 225-230.
122 Baldauf R.; Fortune C.; Weinstein J.; et al. (2005) Air contaminant exposures during the
operation of lawn and garden equipment. J Expos Anal Environ  Epidemiol in press.
123 Eriksson, K.; Tjarner, D.; Marqvardsen, I; et al. (2003) Exposure to benzene, toluene,
xylenes, and total hydrocarbons among snowmobile drivers in Sweden. Chemosphere 50: 1343-
1347.
124 Kado, N.Y.; Kuzmicky, P.A.; and Okamoto, R.A. (2001) Environmental and occupational
exposure to toxic air pollutants from winter snowmobile use in Yellowstone National Park.
Final report to the Yellowstone Park Foundation, Pew Charitable Trusts, and National Park
Service. This document is available in Docket EPA-HQ-OAR-2005-0036.
125 NESCAUM (2003) Evaluating the occupational and environmental impact of nonroad diesel
equipment in the Northeast. Interim Report June 9, 2003. This document is available in Docket
EPA-HQ-OAR-2005-0036. http://www.nescaum.org/resources/reports/index.html
126 U. S. EPA (2006) National-Scale Air Toxics Assessment for 1999.
http://www.epa.gov/ttn/atw/natal999.
127 U.S. EPA (2000) User's Guide for the Assessment System for Population Exposure
Nationwide (ASPEN, Version 1.1) Model. Office of Air Quality Planning and Standards,
Research Triangle Park,  NC, Report No. EPA-454/R-00-017. This document is available in
Docket EPA-HQ-OAR-2005-0036. http://www.epa.gov/scramOO 1/userg/other/aspenug.pdf
128 Rosenbaum, A. (2005) The HAPEM5 User's Guide: Hazardous Air Pollutant Exposure
Model, Version 5. Prepared by ICF, Inc. for U. S. EPA.  This document is available in Docket
EPA-HQ-OAR-2005-0036. http://www.epa.gov/ttn/fera/hapem5/hapem5_guide.pdf
129 U. S. EPA (2005) Risk - Air Toxics Risk Assessment.
http://www.epa.gov/ttn/fera/risk_atoxic.html.
                                        3-150

-------
130 U. S. EPA (1993) Motor Vehicle-Related Air Toxics Study. Report No. EPA420-R-93-005.
http://www.epa.gov/otaq/regs/toxics/tox_archive.htm#2
131 U. S. EPA (2000) Technical Support Document: Control of Hazardous Air Pollutants from
Motor Vehicles and Motor Vehicle Fuels.  Office of Transportation and Air Quality. Report No.
EPA-420-R-00-023. http://www.epa.gov/otaq/toxics.htm
132 Cook, R., Jones, B., Cleland, J. (2004) A Cohort Based Approach for Characterizing Lifetime
Inhalation Cancer Risk from Time-Varying Exposure to Air Toxics from Ambient Sources.
Environmental Progress 23(2): 120-125.
133 Cook, R., Strum, M., Touma, J., et al.  2002. Trends in Mobile Source-Related Ambient
Concentrations of Hazardous Air Pollutants, 1996-2007. SAE Paper No. 2002-01-1274. This
document is available in Docket EPA-HQ-OAR-2005-0036.
134 Byun, D. W., Ching, J. K. S. 1999. Science Algorithms of the EPA Models-3 Community
Multiscale Air Quality (CMAQ) Modeling System. U. S. EPA, Office of Research and
Development, Washington, DC. Report No. EPA/600/R-99/030. This document is available in
Docket EPA-HQ-OAR-2005-0036.
135 Luecken, D. J., Hutzell, W. T., Gipson, G. J. 2005.  Development and Analysis of air quality
modeling simulations for hazardous air pollutants. Submitted to Atmospheric Environment.
136 Seigneur, C., Pun, B., Lohman, K., Wu, S.-Y. 2003. Regional modeling of the atmospheric
fate and transport of benzene and diesel particles. Environ. Sci. Technol. 37: 5236-5246.
137 U. S. EPA. 2004.  User's  Guide for the Emissions Modeling System for Hazardous Air
Pollutants (EMS-HAP, Version 3.0), Office of Air Quality Planning and Standards, Research
Triangle Park, NC, Report No. EPA-454/B-00-007. This document is available in Docket EPA-
HQ-OAR-2005-0036. http://www.epa.gov/scram001/userg/other/emshapv3ug.pdf
138 Battelle. 2003. Estimated background concentrations for the National-Scale Air Toxics
Assessment. Prepared for U. S. EPA, Office of Air Quality Planning and Standards. Contract
No. 68-D-02-061. Work Assignment 1-03. This document is available in Docket EPA-HQ-OAR-
2005-0036.
139 U. S. EPA. 1993. Motor Vehicle-Related Air Toxics Study. Office of Mobile Sources, Ann
Arbor, MI.  Report No. EPA 420-R-93-005. This document is available in Docket EPA-HQ-
OAR-2005-0036. http://www.epa.gov/otaq/regs/toxics/tox_archive.htm
140 U. S. EPA. 1999. Analysis of the Impacts of Control Programs on Motor Vehicle Toxics
Emissions and Exposure in Urban Areas and Nationwide.  Prepared for U. S. EPA, Office of
Transportation and Air Quality, by Sierra Research, Inc., and Radian International
Corporation/Eastern Research Group. Report No. EPA 420 -R-99-029/030. This document is
available in Docket EPA-HQ-OAR-2005-0036.
http://www.epa.gov/otaq/regs/toxics/tox_archive.htm
141 U. S. EPA. 2002.  1996 National-Scale Air Toxics Assessment.
http ://www. epa. gov/ttn/atw/nata/
142 Glen, G., Lakkadi, Y., Tippett, J. A., del Valle-Torres M.  1997. Development of
NERL/CHAD: The National Exposure Research Laboratory Consolidated Human  Activity

                                        3-151

-------
Database. Prepared by ManTech Environmental Technology, Inc.  EPA Contract No. 68-D5-
0049. This document is available in Docket EPA-HQ-OAR-2005-0036.
143 Long, T,; Johnson, T. ; Laurensen, J.; Rosenbaum, A.  2004.  Development of Penetration and
Proximity Microenvironment Factor Distributions for the HAPEM5 in Support of the 1999
National-Scale Air Toxics Assessment (NATA).  Memorandum from TRJ Consulting and ICF
Consulting, Inc. to Ted Palma, U. S. EPA, Office of Air Quality Planning and Standards, RTF,
NC., April 5, 2004. This document is available in Docket EPA-HQ-OAR-2005-0036.
http ://www. epa. gov/ttn/fera/human_hapem.html
144 Rosenbaum, A. 2005. The HAPEM5 User's Guide: Hazardous Air Pollutant Exposure
Model, Version 5. Prepared by ICF, Inc. for Ted Palma, U. S. EPA. This document is available
in Docket EPA-HQ-OAR-2005-0036. http://www.epa.gov/ttn/fera/hapem5/hapem5_guide.pdf
145 U. S. EPA.  2001.  National-Scale Air Toxics Assessment for 1996: Draft for EPA Science
Advisory Board Review.  Report No. EPA-453/R-01-003 This document is available in Docket
EPA-HQ-OAR-2005-0036. http://www.epa.gov/ttn/atw/sab/natareport.pdf
146 U. S. EPA.  2004.  Benefits of the Proposed Inter-State Air Quality Rule.  Office of Air
Quality Planning and Standards, Research Triangle Park,  North Carolina. Report No. EPA 4527-
03-001.  This document is available  in Docket EPA-HQ-OAR-2005-0036.
http://www.epa.gov/CAIR/technical.html or http://www.epa.gov/CAIR/tsdO 175.pdf
147 Memorandum from Arlene Rosenbaum and Kevin Wright, ICF Consulting,  to Chad Bailey,
U.S.  Environmental Protection Agency.  Subject: Estimating near roadway populations and
areas for HAPEM6.  This document is available in Docket EPA-HQ-OAR-2005-0036.
148 Riediker  M, Williams, R., Devlin, R., et al. 2003. Exposure to particulate matter, volatile
organic compounds, and other air pollutants inside patrol  cars. Environ. Sci. Technol. 2003, 37,
2084-2093.
149 Rodes, C., Sheldon, L., Whitaker, D., et al.  1998. Measuring concentrations of selected air
pollutants inside California vehicles. Main Study Report for California ARE. Contract 95-339.
http ://www. arb. ca. gov/research/ab stracts/95 -3 3 9. htm
150 Zhu,  Y, Hinds, W. C., Kim, S., et al. 2002. Study of ultrafine particles near a major highway
with heavy-duty diesel traffic. Atmos Environ 36 (2002) 4323-4335.
151 Kwon, J. 2005. Development of a RIOPA database and  evaluation of the effect of proximity
on the potential residential exposure to VOCS from ambient sources. PhD. dissertation. Graduate
School, New Brunswick,  Rutgers, the State University of New Jersey and the University of
Medicine and Dentistry of New Jersey. This document is  available in Docket EPA-HQ-OAR-
2005-0036.
152 Meng, Q. Y., Turpin, B. J., Korn, L., et al. 2005. Influence of ambient (outdoor) sources on
residential indoor and personal PM2.5 concentrations: Analyses of RIOPA data. Journal of
Exposure Analysis and Environ Epidemiology  15: 17-28.
153 Weisel, C. P., Zhang, J. J., Turpin, B. J., et al.  2004. Relationship of Indoor, Outdoor and
Personal Air (RIOPA) study; study  design, methods and quality assurance / control results.
Journal of Exposure Analysis and Environ Epidemiology 15: 123-137.

                                        3-152

-------
154 Cohen, J.,. Cook, R., Bailey, C. R., Carr, E. 2005. Relationship between motor vehicle
emissions of hazardous pollutants, roadway proximity, and ambient concentrations on Portland,
OR. Environmental Modelling and Software 20: 7-12.
155 U. S. EPA (2003) Estimated Background Concentrations for the National-Scale Air Toxics
Assessment. Emissions, Monitoring, and Analysis Division, Office of Air Quality Planning and
Standards, Research Triangle Park, NC.  http://www.epa.gov/ttn/atw/natal999/background.html
This document is available in Docket EPA-HQ-OAR-2005-0036.
156 U. S. EPA (2005) Supplemental Guidance for Assessing Susceptibility from Early-Life
Exposure to Carcinogens.  Report No. EPA/630/R-03/003F. This document is available in
Docket EPA-HQ-OAR-2005-0036.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=l 16283
157 Science Advisory Board. 2001.  NATA - Evaluating the National-Scale Air Toxics
Assessment 1996 Data - An SAB Advisory. Report No. EPA-SAB-EC-ADV-02-001
158
   Chen, C. W.; Gibb, H.  (2003). Procedures for Calculating Cessation Lag.  Reg. Toxicol.
Pharmacol. 38: 157-165.
159 U. S. EPA.  (2004)  Benefits of the Proposed Inter-State Air Quality Rule. Office of Air
Quality Planning and Standards, Research Triangle Park, North Carolina. Report No. EPA 4527-
03-001. This document is available in Docket EPA-HQ-OAR-2005-0036.
http://www.epa.gov/CAIR/technical.html or http://www.epa.gov/CAIR/tsdO 175.pdf
160Bailey, C.R. (2005) Additional contribution to benzene exposure from attached garages.
Memorandum to docket. This document is available in Docket EPA-HQ-OAR-2005-0036.
161 Evans, J.S.; Wolff, S.K.; Phonboon, K.; Levy, J.I.; Smith, K.R. (2002) Exposure efficiency:
an idea whose time has come? Chemosphere 49:  1075-1091.
162 Marshall, J.D.; Behrentz, E. (2005) Vehicle self-pollution intake fraction:  children's
exposure to school bus emissions. Environ Sci Technol 39: 2559-2563.
163 Touma, J. S.; Isakov, V.; Ching, J.; Seigneur, C.  (2006). Air Quality Modeling of Hazardous
Air Pollutants: Current Status and Future Directions. J. Air and Waste Manage. Assoc., in press.
164 Isakov, V.; Venkatram, A. (2005) Resolving neighborhood scale in air toxics modeling: a
case study in Wilmington, California. J Air & Waste Manage Assoc submitted.
165 Pratt, G. C.; Wu, C. Y.;Bock, D.; et al. (2004) Comparing air dispersion model predictions
with measured concentrations of VOCs  in urban communities. Environ. Sci. Technol. 38:  1949-
1959.
166 Kinnee, E.J.; Touma, J.S.; Mason, R.; Thurman, J.; Beidler, A.; Bailey, C.; Cook,  R.  (2004)
Allocation of Onroad Mobile Emissions to Road Segments for Air Toxics Modeling in Harris
County, Texas.  Transport Res Part D: Transport and Environ 9:139-150.
167 Cohen, J.; Cook, R.; Bailey, C.R.; Carr, E. (2005) Relationship between motor vehicle
emissions of hazardous pollutants, roadway proximity, and ambient concentrations in Portland,
Oregon. Environ Modelling & Software 20: 7-12.
                                         3-153

-------
168 Touma, J. S.; Isakov, V.; Ching, J.; Seigneur, C.  (2006). Air Quality Modeling of Hazardous
Air Pollutants: Current Status and Future Directions. J. Air and Waste Manage. Assoc., in press.
169 Cook, R.; Beidler, A.; Touma, J.S.; Strum M. (2005) Preparing Highway Emissions
Inventories for Urban Scale Modeling: A Case Study in Philadelphia. Submitted to
Transportation Research Part D: Transport and Environment.
170 U.S. EPA. 1996.  Air Quality Criteria for Ozone  and Related Photochemical Oxidants,
EPA600-P-93-004aF. This document is available in Docket EPA-HQ-OAR-2005-0036.

171 U.S. EPA. 1996.  Air Quality Criteria for Ozone  and Related Photochemical Oxidants,
EPA600-P-93-004aF. This document is available in Docket EPA-HQ-OAR-2005-0036.

172 U.S. EPA. 1996. Review of National Ambient Air Quality Standards for Ozone, Assessment
of Scientific and Technical Information, OAQPS Staff Paper, EPA-452/R-96-007. This
document is available in Docket EPA-HQ-OAR-2005-0036.
173 U.S. EPA. 2005. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Second
External Review Draft). This document is available  in Docket EPA-HQ-OAR-2005-0036.
174 American Academy of Pediatrics, Committee on Environmental Health. 2004. "Ambient air
pollution: health hazards to children." Pediatrics 114:  1699-1707.
175 Mortimer, K. M.; Tager, I. B.; Dockery, D. W.; et al. 2000. "The effect of ozone on inner-city
children with asthma: identification of susceptible subgroups." Am. J. Respir. Crit. Care Med.
162: 1838-1845.

176 Gent, J. F.; Triche, E. W.; Holford, T. R.; et al. 2003. "Association of low-level ozone and
fine particles with respiratory symptoms in children with asthma." J. Am. Med. Assoc. 290:
1859-1867.
177 Romieu, I; Sienra-Monge, J. J.; Ramirez-Aguilar, M.; et al.  2002. "Antioxidant
supplementation and lung functions among children with asthma exposed to high levels of air
pollutants." Am. J. Respir. Crit. Care Med. 166:  703-709.

178 Romieu, I; Sienra-Monge, J. J.; Ramirez-Aguilar, M.; et al.  2004. "Genetic polymorphism of
GSTM1 and antioxidant supplementation influence lung function  in relation to ozone exposure in
asthmatic children in Mexico City."  Thorax 59: 8-10.
179 Bates, D.V.; Baker-Anderson, M.; Sizto, R. 1990. "Asthma attack periodicity: a study of
hospital emergency visits in Vancouver."  Environ. Res. 51: 51-70.
180 Thurston, G.D.; Ito, K.;  Kinney, P.L.; Lippmann, M. 1992. "A multi-year study of air
pollution and respiratory hospital admissions in three New York State metropolitan areas:
results for 1988 and  1989 summers." J. Exposure Anal. Environ.  Epidemiol. 2:429-450.

181 Thurston, G.D.; Ito, K.;  Hayes, C.G.; et al. 1994. "Respiratory hospital  admissions and
summertime haze air pollution in Toronto, Ontario:  consideration of the role of acid aerosols."
Environ. Res. 65: 271-290.
182 Lipfert, F.W.; Hammerstrom, T. 1992. "Temporal patterns in air pollution and hospital
admissions." Environ. Res.  59: 374-399.
                                         3-154

-------
183 Burnett, R.T.; Dales, R.E.; Raizenne, M.E.; et al. 1994. "Effects of low ambient levels of
ozone and sulfates on the frequency of respiratory admissions to Ontario hospitals." Environ.
Res. 65: 172-194.
184 U.S. EPA. 1996.  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF.  This document is available in Docket EPA-HQ-OAR-2005-0036.
185 Delfmo, R. 1; Zeiger, R. S.; Seltzer, J. M.; Street, D. H. 1998. "Symptoms in pediatric
asthmatics and air pollution: differences in effects by symptom severity, anti-inflammatory
medication use and parti culate averaging time." Environ. Health Per sped. 106: 751-761.
186 Gold, D. R.; Damokosh, A. I; Pope, C. A., Ill; et al. 1999. "Paniculate and ozone pollutant
effects on the respiratory function of children in  southwest Mexico City." Epidemiology 10: 8-
16.
187 Mortimer, K. M.; Tager, I. B.; Dockery, D. W.; et al. 2000. "The effect of ozone on inner-city
children with asthma: identification of susceptible subgroups." Am. J. Respir. Crit. Care Med.
162: 1838-1845.
188 Newhouse, C. P.; Levetin, B. S.; Levetin, E. 2004. "Correlation of environmental factors with
asthma and rhinitis symptoms in Tulsa, OK." Ann. Allergy Asthma Immunol. 92: 356-366.
189 Koken, P. J.; Piver, W. T.; Ye, F.; et al. 2003. "Temperature, air pollution, and hospitalization
for cardiovascular diseases among elderly people in Denver." Environ. Health Perspect. Ill:
1312-1317.
190 Gwynn, R. C.; Thurston, G. D. 2001. "The burden of air pollution: impacts among racial
minorities." Environ. Health Perspect. Suppl. 109(4): 501-506.
191 Wilson, A. M.; Wake, C. P.; Kelly, T.; Salloway, J. C. 2005. "Air pollution, weather, and
respiratory emergency room visits in two northern New England cities: an ecological time-series
study." Environ. Res. 97: 312-321.
192 Weisel, C. P.; Cody, R. P.; Georgopoulos, P.  G.; et al. 2002. "Concepts in developing health-
based indicators for ozone." Int. Arch. Occup. Environ. Health 75: 415-422.
193 U.S. EPA. 1996.  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF.  This document is available in Docket EPA-HQ-OAR-2005-0036.
194 Devlin, R. B.; McDonnell, W. F.; Mann, R.; et al. 1991. "Exposure of humans to  ambient
levels of ozone for 6.6 hours causes cellular and  biochemical changes in the lung." Am.  J. Respir.
CellMol.Biol.4:72-81.
195 Koren, H. S.; Devlin, R. B.; Becker, S.; et al.  1991. "Time-dependent changes of markers
associated with inflammation in the lungs of humans exposed to ambient levels of ozone."
Toxicol. Pathol. 19: 406-411.
196 Koren, H. S.; Devlin, R. B.; Graham, D. E.; et al. 1989. "Ozone-induced inflammation in the
lower airways of human subjects." Am. Rev. Respir. Dis. 139: 407-415.
197 Schelegle, E.S.; Siefkin, A.D.; McDonald, R.J. 1991. "Time course of ozone-induced
neutrophilia in normal humans." Am. Rev. Respir. Dis. 143:1353-1358.
                                         3-155

-------
198 Romieu, I; Meneses, F.; Ruiz, S.; et al. 1996. "Effects of air pollution on the respiratory
health of asthmatic children living in Mexico City." Am. J. Respir. Crit. Care Med. 154: 300-
307.
199 Romieu, I.; Meneses, F.; Ruiz, S.; et al. 1997. "Effects of intermittent ozone exposure on peak
expiratory flow and respiratory symptoms among asthmatic children in Mexico City." Arch.
Environ. Health 52: 368-376.
200 Gielen, M. H.; Van Der Zee, S. C.; Van Wijnen, J. H.; et al. 1997. "Acute effects of summer
air pollution on respiratory health of asthmatic children." Am. J. Respir. Crit. Care Med. 155:
2105-2108.
201 Just, J.; Segala, C.; Sahraoui, F.;  et al. 2002. "Short-term health effects of particulate and
photochemical air pollution in asthmatic children." Eur. Respir. J. 20: 899-906.
202 U.S. EPA. 1996. Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. This document is available  in Docket EPA-HQ-OAR-2005-0036.
203 Hodgkin, I.E.; Abbey, D.E.; Euler, G.L.; Magie, A.R. 1984. "COPD prevalence in
nonsmokers in high and low photochemical air pollution areas." Chest 86: 830-838.
204 Euler, G.L.; Abbey, D.E.; Hodgkin, I.E.; Magie, A.R. 1988. "Chronic obstructive pulmonary
disease symptom effects of long-term cumulative exposure to ambient levels of total  oxidants
and nitrogen dioxide in California Seventh-day Adventist residents."  Arch. Environ.  Health 43:
279-285.
205 Abbey, D.E.; Petersen, F.; Mills, P.K.; Beeson, W.L. 1993. "Long-term ambient
concentrations of total suspended particulates, ozone, and sulfur dioxide and respiratory
symptoms in a nonsmoking population."  Arch. Environ. Health 48: 33-46.
206 Frischer, T. M.; Kiihr, J.; Pullwitt, A.; et al. 1993. "Ambient ozone causes upper airways
inflammation in children." Am. Rev. Respir. Dis. 148: 961-964.
207 Kinney, P. L.; Nilsen, D. M.; Lippmann, M.; et al. 1996. "Biomarkers of lung inflammation in
recreational joggers exposed to ozone. "Am. J. Respir. Crit. Care Med. 154:  1430-1435.
208 U.S. EPA. 1996. Review of National Ambient Air Quality Standards for Ozone, Assessment
of Scientific and Technical Information, OAQPS Staff Paper, EPA452-R-96-007. This
document is available in Docket EPA-HQ-OAR-2005-0036.
209 Brauer, M.; Brook, J. R. 1997. "Ozone personal exposures and health effects for selected
groups residing in the Fraser Valley." In: Steyn, D. G.; Bottenheim, J. W., eds. The Lower Fraser
Valley Oxidants/Pacific '93 Field Study. Atmos.  Environ. 31: 2113-2121.
210 Chan, C.-C.; Wu, T.-H. 2005. "Effects of ambient ozone exposure on mail carriers' peak
expiratory flow rates." Environ. Health Perspect. 113: 735-738.
211
   Brauer, M.; Blair, J.; Vedal, S. 1996. "Effect of ambient ozone exposure on lung function in
farm workers." Am. J. Respir. Crit. Care Med. 154: 981-987.
212
   Korrick, S. A.; Neas, L. M.; Dockery, D. W.; et al. 1998. "Effects of ozone and other
pollutants on the pulmonary function of adult hikers." Environ. Health Perspect. 106: 93-99.
                                         3-156

-------
213 U.S. EPA. 1996.  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. This document is available in Docket EPA-HQ-OAR-2005-0036.
214 U.S. EPA. 1996.  Review of National Ambient Air Quality Standards for Ozone, Assessment
of Scientific and Technical Information, OAQPS Staff Paper, EPA452-R-96-007. This
document is available in Docket EPA-HQ-OAR-2005-0036.
215 Brauer, M.; Blair, J.; Vedal, S.  1996. "Effect of ambient ozone exposure on lung function in
farm workers." Am. J. Respir. Crit. Care Med. 154: 981-987.

216 Korrick, S. A.; Neas, L. M.; Dockery, D. W.; et al. 1998. "Effects of ozone and other
pollutants on the pulmonary function of adult hikers." Environ. Health Perspect. 106: 93-99.

217 U.S. EPA. 1996.  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. This document is available in Docket EPA-HQ-OAR-2005-0036.

218 U.S. EPA. 1996.  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. This document is available in Docket EPA-HQ-OAR-2005-0036.

219 Avol, E. L.; Trim, S. C.; Little, D. E.; et al. 1990. "Ozone exposure and lung function in
children attending a  southern California summer camp." Presented at: 83rd annual meeting and
exhibition of the Air & Waste Management Association; June; Pittsburgh,  PA. Pittsburgh,  PA:
Air & Waste Management Association; paper no. 90-150.3.

220 Higgins, I. T. T.;  D'Arcy, J. B.; Gibbons, D. L; et al. 1990. "Effect of exposures to ambient
ozone on ventilatory lung  function in children." Am. Rev. Respir. Dis. 141: 1136-1146.

221 Raizenne, M. E.;  Burnett, R. T.; Stern, B.; et al. 1989. "Acute lung function responses to
ambient acid aerosol exposures in children." Environ. Health Perspect. 79: 179-185.
222 Raizenne, M.; Stern, B.; Burnett, R.; Spengler, J. 1987. "Acute respiratory function and
transported air pollutants:  observational studies." Presented at: 80th annual meeting of the  Air
Pollution Control Association; June; New York, NY. Pittsburgh, PA: Air Pollution Control
Association; paper no. 87-32.6.
223 Spektor, D. M.; Lippmann, M.  1991. "Health effects of ambient ozone on healthy children at
a summer camp." In: Berglund, R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone
and the environment: papers from an international conference; March 1990; Los Angeles, CA.
Pittsburgh, PA: Air & Waste Management Association; pp. 83-89.  (A&WMA transaction series
no. TR-19).
224
225
226
   Spektor, D. M.; Thurston, G. D.; Mao, J.; et al. 1991. "Effects of single- and multiday ozone
exposures on respiratory function in active normal children." Environ. Res. 55: 107-122.
   Spektor, D. M.; Lippman, M.; Lioy, P. J.; et al. 1988. "Effects of ambient ozone on
respiratory function in active, normal children." Am. Rev. Respir. Dis. 137: 313-320.
   Thurston, G. D.; Lippmann, M.; Scott, M. B.; Fine, J. M. 1997. "Summertime haze air
pollution and children with asthma." Am. J. Respir. Crit. Care Med. 155: 654-660.
227 US EPA. 2002. New Ozone Health and Environmental Effects References, Published Since
Completion of the Previous Ozone AQCD.  This document is available in Docket EPA-HQ-
OAR-2005-0036.

                                        3-157

-------
228 Thurston, G.D., M.L. Lippman, M.B. Scott, and J.M. Fine. 1997. "Summertime Haze Air
Pollution and Children with Asthma." American Journal of Respiratory Critical Care Medicine,
155:654-660.
229 Ostro, B, M. Lipsett, J. Mann, et al. 2001. "Air pollution and exacerbation of asthma in
African-American children in Los Angeles." Epidemiology 12(2): 200-208.
230 Jalaludin, B. B.; Chey, T.; O'Toole, B. I.; et al. 2000. "Acute effects of low levels of ambient
ozone on peak expiratory flow rate in a cohort of Australian children." Int. J. Epidemiol. 29: 549-
557.
231 Desqueyroux, H.; Pujet, J.-C.; Prosper, M.; et al. 2002. "Short-term effects of low-level air
pollution on respiratory health of adults suffering from moderate to severe asthma." Environ.
Res. A 89: 29-37.
232 Just, J.; Segala, C.; Sahraoui, F.; et al. 2002. "Short-term health effects of particulate and
photochemical air pollution in asthmatic children." Eur. Respir. J. 20: 899-906.
233 McDonnell, W.F., D.E. Abbey, N. Nishino and M.D. Lebowitz.  1999. "Long-term ambient
ozone concentration and the incidence of asthma in nonsmoking adults:  the ahsmog study."
Environmental Research. 80(2 Pt 1): 110-121.
234 McConnell, R.; Berhane, K.; Gilliland, F.; et al. 2002. "Asthma in exercising children
exposed to ozone: a cohort study." Lancet 359: 386-391.
235 Burnett, R.  T.; Smith_Doiron, M.; Stieb, D.; et al. 2001. "Association between ozone and
hospitalization for acute respiratory diseases in children less than 2 years of age." Am. J.
Epidemiol. 153: 444-452.
236 Gouveia, N.; Fletcher, T. 2000. "Respiratory diseases in children and outdoor air pollution in
Sao Paulo, Brazil: a time series analysis." Occup. Environ. Med. 57: 477-483.
237 Petroeschevsky, A.; Simpson, R. W.;  Thalib, L.; Rutherford, S. 2001. "Associations between
outdoor air pollution and hospital admissions in Brisbane, Australia." Arch. Environ. Health 56:
37-52.
238 Wong, T. W.; Lau, T.  S.; Yu, T. S.; et al. 1999. "Air pollution and hospital admissions for
respiratory and cardiovascular diseases in Hong Kong." Occup. Environ. Med. 56: 679-683.
239 Burnett, R.  T.; Smith-Doiron, M.; Stieb, D.; et al. 2001. "Association between ozone and
hospitalization for acute respiratory diseases in children less than 2 years of age." Am. J.
Epidemiol. 153:444-452.
240 Chen, L.; Jennison, B. L.; Yang, W.; Omaye, S. T. 2000. "Elementary school absenteeism and
air pollution." Inhalation Toxicol. 12: 997-1016.
241 Gilliland, F.D.; Berhane, K.; Rappaport, E.B.; et al. 2001. "The effects of ambient air
pollution on school absenteeism due to respiratory illnesses." Epidemiology 12:43-54.
242 Park, H.; Lee, B.; Ha,  E.-H.; et al. 2002. "Association of air pollution with school
absenteeism due to illness." Arch. Pediatr. Adolesc. Med. 156:  1235-1239.
243 Chen, L.; Jennison, B. L.; Yang, W.; Omaye, S. T. 2000. "Elementary school absenteeism and
air pollution." Inhalation Toxicol. 12: 997-1016.
                                         3-158

-------
244 Gilliland, F.D.; Berhane, K.; Rappaport, E.B.; et al. 2001. "The effects of ambient air
pollution on school absenteeism due to respiratory illnesses." Epidemiology 12:43-54.
245 Park, H.; Lee, B.; Ha, E.-H.; et al. 2002. "Association of air pollution with school
absenteeism due to illness." Arch. Pediatr. Adolesc. Med. 156: 1235-1239.
246 Samet, J. M.; Zeger, S. L.; Dominici, F.; et al. 2000. The national morbidity, mortality, and
air pollution study. Part II: morbidity, mortality, and air pollution in the United States.
Cambridge, MA: Health Effects Institute; research report no. 94, part II. This document is
available in Docket EPA-HQ-OAR-2005-0036.
247 Dominici, F.; McDermott, A.; Daniels, M.; et al. 2003. "Mortality among residents of 90
cities." In: Revised analyses of time-series studies of air pollution and health. Special report.
Boston, MA: Health Effects Institute; pp. 9-24. Available:
http://www.healtheffects.org/Pubs/TimeSeries.pdf [12 May, 2004]. This document is available
in Docket EPA-HQ-OAR-2005-0036.
248 Bell, M. L.; McDermott, A.; Zeger, S. L.; et al. 2004. "Ozone and short-term mortality in 95
US urban communities, 1987-2000" JAMA J. Am. Med. Assoc. 292: 2372-2378.
249 Huang, Y.; Dominici, F.; Bell, M. L. 2005. "Bayesian hierarchical distributed lag models for
summer ozone exposure and cardio-respiratory mortality." Environmetrics 16: 547-562.
250
   Schwartz, J. 2005. "How sensitive is the association between ozone and daily deaths to
control for temperature?" Am. J. Respir. Crit. Care Med.  171: 627-631.
251 Gryparis, A.; Forsberg, B.; Katsouyanni, K.; et al. 2004. Acute effects of ozone on mortality
from the "Air Pollution and Health: A European Approach" project. Am J Respir Crit Care Med
170: 1080-1087.
252 Touloumi, G.; Katsouyanni, K.; Zmirou, D.; et al. 1997. "Short-term effects of ambient
oxidant exposure on mortality: a combined analysis within the APHEA project." Am. J.
Epidemiol. 146: 177-185.
253 Bell, M. L.; Dominici, F.; Samet, J. M. 2005. "A meta-analysis of time-series studies of ozone
and mortality with comparison to the national morbidity,  mortality, and air pollution study."
Epidemiology 16: 436-445.
254 Ito, K.; De Leon, S. F.; Lippmann, M. 2005. "Associations between ozone and daily
mortality, analysis and meta-analysis." Epidemiology 16: 446-457.
255
   Levy, J. L; Chemerynski, S. M.; Sarnat, J. A. 2005. "Ozone exposure and mortality, an
empiric Bayes metaregression analysis." Epidemiology 16: 458-468.
256
   U.S. EPA. 2005. Technical Support Document for the Final Clean Air Interstate Rule - Air
Quality Modeling.  This document is available in Docket EPA-HQ-OAR-2005-0036.
257
   U.S. EPA. 2005. Technical Support Document for the Proposed Mobile Source Air Toxics
Rule - Air Quality Modeling. This document is available in Docket EPA-HQ-OAR-2005-0036.
   U.S. EPA. 2005. Technical Support Document for the Final Clean Air Interstate Rule - Air
Quality Modeling. This document is available in Docket EPA-HQ-OAR-2005-0036.
258
                                         3-159

-------
259 U.S. EPA. 2005. Clean Air Interstate Rule Emissions Inventory Technical Support
Document. Available from: http://www.epa.gov/cair/technical.html.  This document is available
in Docket EPA-HQ-OAR-2005-0036.

260 U.S. EPA. 2005. Technical Support Document for the Proposed Mobile Source Air Toxics
Rule - Ozone Air Quality Modeling. This document is available in Docket EPA-HQ-OAR-2005-
0036.
261 Pielke, R.A., W.R. Cotton, R.L. Walko, et al. 1992. "A Comprehensive Meteorological
Modeling System - RAMS." Meteor. Atmos. Phys., Vol. 49, pp. 69-91.
262 U.S. EPA. 1999.  The Benefits and Costs of the Clean Air Act,  1990-2010. Prepared for U.S.
Congress by U.S. EPA, Office of Air and Radiation, Office  of Policy Analysis and Review,
Washington, DC, November; EPA report no. EPA410-R-99-001. This document is available in
Docket EPA-HQ-OAR-2005-0036.
263 U.S. EPA. 1996.  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. This document is available in Docket EPA-HQ-OAR-2005-0036.
264 Winner, W.E., and CJ. Atkinson. 1986. "Absorption of air pollution by plants, and
consequences for growth." Trends in Ecology and Evolution 1:15-18.
265 U.S. EPA. 1996.  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. This document is available in Docket EPA-HQ-OAR-2005-0036.
266 Tingey, D.T., and Taylor, G.E. 1982. "Variation in plant response to ozone: a conceptual
model of physiological events." In: Effects of Gaseous Air  Pollution in Agriculture and
Horticulture (Unsworth, M.H., Omrod, D.P., eds.) London, UK: Butterworth Scientific, pp. 113-
138.
267U.S. EPA. 1996. Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. This document is available in Docket EPA-HQ-OAR-2005-0036.
268 U.S. EPA. 1996.  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. This document is available in Docket EPA-HQ-OAR-2005-0036.
269 U.S. EPA. 1996.  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. This document is available in Docket EPA-HQ-OAR-2005-0036.
270 Ollinger,  S.V., J.D. Aber and P.B. Reich. 1997. "Simulating ozone effects on forest
productivity: interactions between leaf canopy and stand level processes." Ecological
Applications 7:1237-1251.

271 Winner, W.E., 1994. "Mechanistic analysis of plant responses to air pollution." Ecological
Applications, 4(4):651 -661.

272 U.S. EPA. 1996.  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. This document is available in Docket EPA-HQ-OAR-2005-0036.

273 U.S. EPA. 1996.  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600P-93-004aF. This document is available in Docket  EPA-HQ-OAR-2005-0036.

274 Fox, S., and R. A. Mickler, eds. 1996. Impact of Air Pollutants on Southern Pine Forests.
Springer-Verlag, NY, Ecol. Studies, Vol. 118, 513 pp.
                                        3-160

-------
275 National Acid Precipitation Assessment Program (NAPAP), 1991. National Acid
Precipitation Assessment Program. 1990 Integrated Assessment Report. National Acid
Precipitation Program. Office of the Director, Washington DC.
276 De Steiguer, I, J. Pye, C. Love. 1990. "Air Pollution Damage to U.S. Forests." Journal of
Forestry, Vol 88(8) pp. 17-22.
277 Pye, J.M. 1988. "Impact of ozone on the growth and yield of trees: A review." Journal of
Environmental Quality 17 pp.347-360.
278 U.S. EPA. 1996.  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF.  This document is available in Docket EPA-HQ-OAR-2005-0036.
279 U.S. EPA. 1996.  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF.  This document is available in Docket EPA-HQ-OAR-2005-0036.
280 McBride, J.R., P.R. Miller, and R.D. Laven.  1985. "Effects of oxidant air pollutants on forest
succession in the mixed conifer forest type of southern California." In:  Air Pollutants Effects
On Forest Ecosystems, Symposium Proceedings, St. P,  1985, p. 157-167.
281 Miller, P.R., O.C. Taylor, R.G. Wilhour. 1982. Oxidant air pollution effects on a western
coniferous forest ecosystem. Corvallis, OR: U.S. Environmental Protection Agency,
Environmental Research Laboratory (EPA600-D-82-276). This document is available in Docket
EPA-HQ-OAR-2005-0036.
282 U.S. EPA. 1996.  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF.  This document is available in Docket EPA-HQ-OAR-2005-0036.
283Kopp, R. J.; Vaughn, W. J.; Hazilla, M.; Carson, R. 1985.  "Implications of environmental
policy for U.S. agriculture: the case of ambient ozone standards."  J. Environ. Manage.  20:321-
331.
284 Adams, R. M.; Hamilton, S. A.; McCarl, B. A. 1986. "The benefits of pollution control: the
case of ozone and U.S. agriculture." Am. J. Agric. Econ.  34: 3-19.
285 Adams, R. M.; Glyer, J. D.; Johnson, S. L.; McCarl,  B. A.  1989. "A reassessment of the
economic effects of ozone on U.S. agriculture." JAPCA 39:960-968.
286 Abt Associates, Inc. 1995.  Urban ornamental plants: sensitivity to ozone and potential
economic losses. U.S. EPA, Office of Air Quality Planning and Standards, Research Triangle
Park. Under contract to RADIAN Corporation, contract no. 68-D3-0033, WA no. 6. pp. 9-10.
287 U.S. EPA. 2004. Air Quality Criteria for Particulate Matter.  EPA/600/P-99/002aF and
EPA/600/P-99/002bF.  This document is available in Docket EPA-HQ-OAR-2005-0036.
288 U.S. EPA. 2005. Review of the National Ambient Air Quality Standard for Particulate Matter:
Policy Assessment of Scientific and Technical Information, OAQPS Staff Paper. EPA-452/R-
05-005.  This document is available in Docket EPA-HQ-OAR-2005-0036.
289 Laden F; Neas LM; Dockery DW; et al. 2000. "Association of fine particulate matter from
different sources with daily mortality in six U.S. cities." Environ Health Perspectives
108(10):941-947.
                                         3-161

-------
290 Schwartz J; Laden F; Zanobetti A.  2002.  "The concentration-response relation between
PM(2.5) and daily deaths." Environ Health Perspect 110(10): 1025-1029.
291 Janssen NA; Schwartz J; Zanobetti A.; et al. 2002. "Air conditioning and source-specific
particles as modifiers of the effect of PM10 on hospital admissions for heart and lung disease."
Environ Health Perspect 110(l):43-49.
292 Kunzli, N; Medina, S; Kaiser, R; et al. 2001. "Assessment of deaths attributable to air
pollution: should we use risk estimates based on time series or on cohort studies?" Am J
Epidemiol 153(11): 1050-1055.
293 Dockery, DW; Pope, CA, III; Xu, X; et al. 1993. "An association between air pollution and
mortality in six U.S. cities." NEngUMed 329:1753-1759.
294 Pope, CA, III; Burnett, RT; Thun, MJ; Calle, EE; et al. 2002. "Lung cancer, cardiopulmonary
mortality, and long-term exposure to fine particulate air pollution."  J Am MedAssoc 287: 1132-
1141.
295 Krewski, D; Burnett, RT; Goldberg, M S; et al. 2000. "Reanalysis of the Harvard Six Cities
study and the American Cancer Society study of particulate air pollution and mortality. A special
report of the Institute's Particle Epidemiology Reanalysis Project." Cambridge, MA: Health
Effects Institute. This document is available in Docket EPA-HQ-OAR-2005-0036.
296 Jerrett, M; Burnett, RT; Ma, R; et al. 2005. "Spatial Analysis of Air Pollution and Mortality in
Los Angeles." Epidemiology. 16(6):727-736.
297 Kunzli, N.; Jerrett, M.; Mack, W.J.; et al. 2005. "Ambient air pollution and atherosclerosis in
Los Angeles." Environ Health Perspect. 113:201 -206.
298 Riediker, M.; Cascio, W.E.;  Griggs, T.R.; et al. 2004. "Particulate matter exposure in cars is
associated with cardiovascular effects in healthy young men." Am JRespir Crit Care Med 169:
934-940.
299 U.S. EPA. 2005. Technical Support Document for the Final Clean Air Interstate Rule - Air
Quality Modeling. This document is available in  Docket EPA-HQ-OAR-2005-0036.
300 U.S. EPA. 2005. Review of the National Ambient Air Quality  Standard for Particulate Matter:
Policy Assessment of Scientific and Technical Information, OAQPS Staff Paper.  EPA-452/R-
05-005.  This document is available in Docket EPA-HQ-OAR-2005-0036.
301 U.S. EPA. 1993. Effects of the 1990 Clean Air Act Amendments on Visibility in Class I
Areas: An EPA Report to Congress. EPA452-R-93-014.
302
   U.S. EPA (2002) Latest Findings on National Air Quality - 2002 Status and Trends.  EPA
454/K-03-001. This document is available in Docket EPA-HQ-OAR-2005-0036.
303 National Park Service. Air Quality in the National Parks, Second edition. NFS, Air
Resources Division. D 2266.  September 2002. This document is available in Docket EPA-HQ-
OAR-2005-0036.
304 U.S. EPA (2002) Latest Findings on National Air Quality - 2002 Status and Trends.  EPA
454/K-03-001. This document is available in Docket EPA-HQ-OAR-2005-0036.

                                         3-162

-------
305 U.S. EPA (2005). Technical Support Document for the Final Clean Air Interstate Rule - Air
Quality Modeling.  This document is available in Docket EPA-HQ-OAR-2005-0036.
306
   U.S. EPA (2000) Deposition of Air Pollutants to the Great Waters: Third Report to Congress.
Office of Air Quality Planning and Standards. EPA-453/R-00-0005.
307 U.S. EPA (2004) National Coastal Condition Report II. Office of Research and Development/
Office of Water. EPA-620/R-03/002.
308 Gao, Y., E.D. Nelson, M.P. Field, et al.  2002.  "Characterization of atmospheric trace
elements on PM2.5 particulate matter over the New York-New Jersey harbor estuary."  Atmos.
Environ. 36: 1077-1086.
309 Kim, G., N. Hussain, J.R. Scudlark, and T.M. Church. 2000.  "Factors influencing the
atmospheric deposit!onal fluxes of stable Pb, 210Pb, and 7Be into Chesapeake Bay." J. Atmos.
Chem. 36: 65-79.
310 Lu, R., R.P. Turco, K. Stolzenbach, et al.  2003. "Dry deposition of airborne trace metals on
the Los Angeles Basin and adjacent coastal waters." J. Geophys. Res. 108(D2, 4074): AAC 11-1
to 11-24.
311 Marvin,  C.H., M.N. Charlton, EJ. Reiner, et al. 2002. "Surficial sediment contamination in
Lakes Erie and Ontario: A comparative analysis." J. Great Lakes Res. 28(3): 437-450.
312 Smith, W.H. 1991. "Air pollution and Forest Damage." Chemical Engineering News, 69(45):
30-43.
313 Gawel, I.E.; Ahner, B.A.; Friedland, A.J.; and Morel, F.M.M. 1996. "Role for heavy metals in
forest decline indicated by phytochelatin measurements." Nature, 381: 64-65.
314 Cotrufo, M.F.; DeSanto, A.V.; Alfani, A.; et al. 1995. "Effects of urban heavy metal pollution
on organic matter decomposition in Quercus ilix L. woods." Environmental Pollution, 89: 81-87.
315 Niklinska, M.; Laskowski, R.; Maryanski, M. 1998. "Effect of heavy metals and storage time
on two types of forest litter: basal respiration rate and exchangeable metals." Ecotoxicological
Environmental Safety, 41: 8-18.
316 Mason, R.P. and Sullivan, K.A. 1997. "Mercury in Lake Michigan." Environmental Science
& Technology, 31: 942-947. (from Delta Report "Atmospheric deposition of toxics to the Great
Lakes")
317 Landis, M.S. and Keeler, GJ. 2002. "Atmospheric  mercury deposition to Lake Michigan
during the Lake Michigan Mass Balance Study." Environmental Science & Technology, 21:
4518-24.
318 U.S. EPA. 2000. EPA453/R-00-005, "Deposition of Air Pollutants to the Great Waters: Third
Report to Congress," Office of Air Quality Planning and Standards, Research Triangle Park,
North Carolina. This document is available in Docket EPA-HQ-OAR-2005-0036.
319 Callender, E. and Rice, K.C. 2000. "The Urban Environmental Gradient: Anthropogenic
Influences on the Spatial and Temporal Distributions of Lead and Zinc in Sediments."
Environmental Science & Technology, 34: 232-238.
                                         3-163

-------
320 Rice, K.C. 1999. "Trace Element Concentrations in Streambed Sediment Across the
Conterminous United States." Environmental Science & Technology, 33: 2499-2504.
321 Ely, JC; Neal, CR; Kulpa, CF; et al. 2001. "Implications of Platinum-Group Element
Accumulation along U.S. Roads from Catalytic-Converter Attrition." Environ. Sci. Technol. 35:
3816-3822.
322 U.S. EPA. 1998. EPA454/R-98-014, "Locating and Estimating Air Emissions from Sources
of Poly cyclic Organic Matter," Office of Air Quality Planning and Standards, Research Triangle
Park, North Carolina. This document is available in Docket EPA-HQ-OAR-2005-0036.
323 U.S. EPA. 1998. EPA454/R-98-014, "Locating and Estimating Air Emissions from Sources
of Poly cyclic Organic Matter," Office of Air Quality Planning and Standards, Research Triangle
Park, North Carolina. This document is available in Docket EPA-HQ-OAR-2005-0036.
324 Simcik, M.F.; Eisenreich, S.J.; Golden, K.A.; et al. 1996. "Atmospheric Loading of
Polycyclic Aromatic Hydrocarbons to Lake Michigan as Recorded in the Sediments."
Environmental Science and Technology, 30: 3039-3046.
325 Simcik, M.F.; Eisenreich, S.J.; and Lioy, PJ. 1999. "Source apportionment and source/sink
relationship of PAHs in the coastal atmosphere of Chicago and Lake Michigan." Atmospheric
Environment, 33: 5071-5079.
326 Arzayus, K.M.; Dickhut, R.M.; and Canuel, E.A. 2001. "Fate of Atmospherically Deposited
Polycyclic Aromatic Hydrocarbons (PAHs) in Chesapeake Bay." Environmental Science &
Technology,  35, 2178-2183.
327 Park, J.S.; Wade, T.L.; and Sweet, S. 2001. "Atmospheric distribution of polycyclic aromatic
hydrocarbons and deposition to Galveston Bay, Texas, USA." Atmospheric Environment, 35:
3241-3249.
328 Poor, N.;  Tremblay, R.; Kay, H.; et al. 2002.  "Atmospheric concentrations and dry deposition
rates of polycyclic aromatic hydrocarbons (PAHs) for Tampa Bay, Florida, USA." Atmospheric
Environment 38: 6005-6015.
329 Arzayus, K.M.; Dickhut, R.M.; and Canuel, E.A. 2001. "Fate of Atmospherically Deposited
Polycyclic Aromatic Hydrocarbons (PAHs) in Chesapeake Bay." Environmental Science &
Technology,  35, 2178-2183.
330 U.S. EPA. 2000. EPA453/R-00-005, "Deposition of Air Pollutants to the Great Waters: Third
Report to Congress," Office of Air Quality Planning and Standards, Research Triangle Park,
North Carolina. This document is available in Docket EPA-HQ-OAR-2005-0036.
331 Van Metre, P.C.; Mahler, B.J.; and Furlong, E.T. 2000. "Urban Sprawl Leaves its PAH
Signature" Environmental Science & Technology, 34: 4064-4070.
332 Cousins, IT.; Beck, A.J.; and Jones, K.C. 1999. "A review of the processes involved in the
exchange of semi-volatile organic compounds across the air-soil interface." The Science of the
Total Environment, 228: 5-24.
333 Tuhackova, J., Cajthaml, T.; Novak, K.; et al. 2001. "Hydrocarbon deposition and  soil
microflora as affected by highway traffic." Environmental Pollution, 113: 255-262.

                                        3-164

-------
334 U.S. EPA (2005) Review of the National Ambient Air Quality Standard for Parti culate
Matter: Policy Assessment of Scientific and Technical Information, OAQPS Staff Paper. EPA-
452/R-05-005. This document is available in Docket EPA-HQ-OAR-2005-0036.
335 Hoek, G.; Brunekreef, B.; Goldbohm, S.; et al. (2002) Association between mortality and
indicators of traffic-related air pollution in the Netherlands: a cohort study. Lancet 360: 1203-
1209.
336 Finkelstein, M.M.; Jerrett, M.; Sears, M.R. (2004) Traffic air pollution and mortality rate
advancement periods. Am JEpidemiol 160: 173-177.
337 Finkelstein, M.M.; Jerrett, M.; Sears, M.R. (2005) Environmental inequality and circulatory
disease mortality gradients. J Epidemiol Community Health 59: 481-486.
338 Hoek, G.; Brunekreef, B.; Goldbohm, S.; et al. (2002) Association between mortality and
indicators of traffic-related air pollution in the Netherlands: a cohort study. Lancet 360: 1203-
1209.
339 Maheswaran, M.;  Elliott, P. (2003) Stroke mortality associated with living near main roads in
England and Wales.  A geographical study. Stroke 34: 2776-2780.
340 Roemer, W.H.; van Wijnen, J.H. (2001) Daily mortality and air pollution along busy streets in
Amsterdam, 1987-1998. Epidemiol 12: 649-653.
341 Heinrich, J.; Wichmann, H-E. (2004) Traffic related pollutants in Europe and their effect on
allergic disease. Current Opin Clinical Epidemiol 4:  341-348.
342 Ryan, P.H.; LeMasters, G.; Biagnini, J.; et al. (2005) Is it traffic type, volume, or distance?
Wheezing in infants living near truck and bus traffic.  J Allergy Clin Immunol  116: 279-284.
343 Kim, J.J.; Smorodinsky, S.; Lipsett, M.; et al. (2004) Traffic-related air pollution near busy
roads:  the East Bay Children's Respiratory Health Study. Am JRespir Crit Care Med 170: 520-
526.
344 Lin, S.; Munsie, J.P.; Hwang, S-A.; et al. (2002) Childhood asthma hospitalization and
residential exposure to state route traffic. Environ Res 88:  73-81.
345 English, P.; Neutra, R.;  Scalf, R.; et al. (1999) Examining associations between childhood
asthma and traffic flow using a geographic information system. Environ Health Perspect 107:
761-768.
346 Garshick, E.; Laden, F.; Hart, I.E.; Caron, A.  (2003) Residence near a major road and
respiratory symptoms in U.S. veterans. Epidemiology 14: 728-736.
347
   Heinrich, J.; Wichmann, H-E. (2004) Traffic related pollutant in Europe and their effect on
allergic disease.  Current Opin Clinical Epidemiol 4:  341-348.
348
   Peters, A.; von Klot, S.; Heier, M.; et al. (2004) Exposure to traffic and the onset of
myocardial infarction. N Engl J Med 351:  1721-1730.
349 Riediker, M.; Cascio, W.E.; Griggs, T.R..; et al. (2003) Particulate matter exposures in cars is
associated with cardiovascular effects in healthy young men. Am J Respir Crit Care Med Vol.
169:934-940.

                                          3-165

-------
350 Schwartz, J.; Litonjua, L.; Suh, H.; et al. (2005) Traffic related pollution and heart rate
variability in a panel of elderly subjects. Thorax 60: 455-461.
351
   Gold, D.R.; Litonjua, A.A.; Zanobetti, A.; et al. (2005) Air pollution and ST-segment
depression in elderly subjects. Environ Health Perspect 113: 883-887.
352
   Wilhelm, M.; Ritz, B. (2002) Residential proximity to traffic and adverse birth outcomes in
Los Angeles County, California, 1994-1996. Environ Health Prospect 111:207-216.
353 Ritz B; Yu F. (1999) The effect of ambient carbon monoxide on low birth weight among
children born in southern California between 1989 and 1993. Environ Health Perspect 107:17-
25.
354 Ritz B; Yu F; Capa G; Fruin S. (2000) Effect of air pollution on premature birth among
children born in southern California between 1989 and 1993. Epidemiology 11:502-511.
355 Ritz B; Yu F; Fruin S; et al. (2002) Ambient air pollution and risk of birth defects in Southern
California. Am JEpidemiol 155:17-25.
356 Perera, F.P.; Rauh, V.; Tsai, W-Y.; et al. (2002) Effect of transplacental exposure to
environmental pollutants on birth outcomes in a multiethnic population.  Environ Health
Perspect 111: 201-205.
357 U. S. EPA (2005) Supplemental Guidance for Assessing Susceptibility from Early-Life
Exposure to Carcinogens. Report No. EPA/630/R-03/003F. This document is available in
Docket EPA-HQ-OAR-2005-0036.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid= 116283
358 U. S. EPA. 2002. Toxicological Review of Benzene (Noncancer effects).  Report No. EPA
635/R-02/001-F.
359 Savitz, D.A.; Feingold, L. (1989) Association of childhood cancer with residential traffic
density. Scand J Work Environ Health 15:  360-363.
360 Pearson, R.L.; Wachtel, H.; Ebi, K.L. (2000) Distance-weighted traffic density in proximity
to a home is a risk factor for leukemia and other childhood cancers. J Air Waste Manage Assoc
50: 175-180.
361 Feychting, M.; Svensson, D.; Ahlbom, A. (1998) Exposure to motor vehicle exhaust and
childhood cancer. Scan J Work Environ Health 24(1): 8-11.
362
   Langholz, B.; Ebi, K.L.; Thomas, D.C.; et al. (2002) Traffic density and the risk of childhood
leukemia in a Los Angeles case-control study.  Ann Epidemiol 12: 482-487.
363
   Raaschou-Nielsen, O.; Hertel, O.; Thomsen, B.L.; Olsen, J.H. (2001) Air pollution from
traffic at the residence of children with cancer. Am J Epidemiol 153: 433-443.
364 Reynolds, P.; Von Behren, J.; Gunier, R.B.; et al. (2003) Childhood cancer incidence rates
and hazardous air pollutants in California: an exploratory analysis. Environ Health Perspect 111:
663-668.
365 Reynolds, P.; Von Behren, J.; Gunier, R.B.; et al. (2004) Residential exposure to traffic in
California and childhood cancer. Epidemiol 15: 6-12.

                                         3-166

-------
366 Knox, E.G. (2005) Oil combustion and childhood cancers. J Epidemiol Community Health
59: 755-760.
367United States Census Bureau. (2004) American Housing Survey web page.  [Online at
http://www.census.gov/hhes/www/housing/ahs/ahs03/ahs03.html] Table IA-6.
368This statistic is based on the odds ratio of being near major transportation sources, compared
to not being near transportation sources for housing units located in different geographic regions.
369Garshick, E.; Laden, F.; Hart, I.E.; Caron, A. (2003) Residence near a major road and
respiratory symptoms in U.S. veterans. Epidemiol 14: 728-736.
370Green, R.S.; Smorodinsky, S.; Kim, J.J.; McLaughlin, R.; Ostro, B. (2004) Proximity of
California public schools to busy roads.  Environ Health Perspect 112: 61-66.
371Gunier, R.B.; Hertz, A.; Von Behren,  J.; Reynolds, P. (2003) Traffic density in California:
socioeconomic and ethnic differences among potentially exposed children. J Expos Anal
Environ Epidemiol 13: 240-246.
                                         3-167

-------
                              Chapter 4: Table of Contents

Chapter 4: Industry Characterization	2
   4.1  Light-Duty Vehicle and Light-Duty Truck Market Structure	2
     4.1.1  Domestic vs. Foreign Manufacturers	2
     4.1.2  Light-Duty Vehicles vs. Light-Duty Trucks	4
     4.1.3  Small Volume Manufacturers, Importers, and Alternative Fuel Vehicle Converters. 6
   4.2  Petroleum Refining Industry	6
     4.2.1  Gasoline Supply	7
     4.2.2  Gasoline Demand	7
     4.2.3  Industry Organization	7
     4.2.4  Gasoline Market Data	8
   4.3  Portable Gasoline Container (Gas Can) Industry	8
     4.3.1  Manufacture and Distribution	9
     4.3.2  Gas Can Use	9
     4.3.3  Market Structure	10
     4.3.4  Market Entry	10
                                          4-1

-------
                     Chapter 4: Industry Characterization

       An understanding of the nature of the affected industries is useful in assessing the potential
impact of the proposed emission control program. Information regarding the structure of the market,
including such things as the degree of concentration, entry barriers, and product differentiation, can
help explain the pricing and other policies that exist in that market. This chapter describes the
light-duty vehicle (LDV) and light-duty truck (LDT) manufacturers, the petroleum refining industry,
and the portable gasoline container (gas can) manufacturers.

4.1    Light-Duty Vehicle and Light-Duty Truck Market Structure

        The LDV/LDT market is fairly concentrated, with only five of the 19 total generally-recog-
nized manufacturers accounting for almost 82 percent of all sales. LDV/LDT sales numbered more
than 16.9 million vehicles in 2004, the latest year for which we have complete data. The top five
companies are the so-called "Big Three" (General Motors (GM), Ford, and Daimler-Chrysler) plus
Toyota and Honda. The remaining 18 percent of sales are split between the other 14 manufacturers,
with none of them achieving more than 2 percent of total sales. The bottom 10 manufacturers in fact
account for only about 4.5 percent of total sales. Four of these firms, Ferrari, Maserati, Lamborghini,
and Lotus, are considered small-volume manufacturers, since their sales are less than 15,000
vehicles per year.A Table 4.1.-1  provides sales figures by manufacturer.

       None of the major manufacturers are small businesses. (As discussed later in Chapter 14, the
Small Business Administration (SBA) criterion for a small business in the vehicle manufacturing
industry is 1,000 employees or less.) This is mainly because of the large outlay of capital and other
resources necessary to enter the  market. Become even a relatively minor player in the industry
requires a great deal of manufacturing capacity to achieve the necessary production volumes, as well
as an extensive distribution and marketing network. There is  also a significant amount of brand
loyalty on the part of consumers, because of tradition or perceived differences in the product. These
all combine to make market entry difficult, and the industry is basically dominated by the
established major manufacturers.

       As discussed later in Section 4.1.3, there  are also a few smaller, lesser-known LDV/LDT
small volume manufacturers, importers and alternative fuel vehicle converters. These have limited
product lines, and account for less than one-tenth of one percent of all U.S. sales. They primarily fill
niche markets of one kind or another. More than  half of these firms are small businesses.

4.1.1  Domestic vs. Foreign Manufacturers

        Previously, it has been relatively easy to characterize manufacturers as "domestic" or
  EPA defines small volume manufacturers to be those with total U.S. salesof less than 15,000 vehicles per year. This
status allows vehicle models to be certified under a slightly simpler certification process. For certification purposes,
small volume manufacturers also include independent commercial importers (ICIs) and alternative fuel vehicle
converters since they sell less than 15,000 vehicles per year.

                                            4-2

-------
  CO
  UJ
  tt
  CO

  k
  *:
  oc
  co
  uj
  O

 L CO

*3

13

I5 to"
  DC
  uj
  ct:
o


o
       => s£
       o

       M

       LU
          C/3
     O
       SB3
       ^ _i
       UM.


       K
NUFACTU

NAME

                CM
              en j* en 10

              r-_ CM_ to o_

              CM" CM" T-" T-"
                       en CM


                       •"
<£> 00 T- f-

CM" U3 co h-"
                                                           CD
                           CM

                           CM



                           Off
                                                         00




                                                         5>
                                  O
                                  o
                                  CM



                                  *"
                                  O

                                  m

                                  m
                                  «*-•
                                  (C

                                  Q



                                  1

                                  «5
              CM
                                                         oo



                                                         
-------
"foreign." However, this is currently much more difficult. For example, the Daimler-Chrysler
merger combined the former Chrysler divisions Chrysler, Dodge and Jeep with the imported
Mercedes line; but it also includes Maybach, a high-end German luxury car. Ford now includes not
only the traditional Ford, Mercury and Lincoln lines, but also the imported marques Jaguar, Volvo,
Land Rover and Aston-Martin. GM sales include the Swedish import Saab.

       Conversely, Toyota and Honda, as well as the six other Far Eastern manufacturers, all
maintain a substantial American manufacturing presence, and the majority of their vehicles sold
here, almost 80 percent on average, are manufactured in North America. Sales figures from North
American manufacturing facilities for individual firms range from 95 to 98 percent for Toyota and
Honda, to 52 to 72 percent for some of the smaller manufacturers. Volkswagen, which now also
includes Bentley, is the only European manufacturer with a North American manufacturing opera-
tion. About 55 percent of its sales are manufactured here.  BMW, which now includes the formerly
British Rolls-Royce and Mini lines, is 100 percent imported, as is Porsche.

       On the other hand, substantial portions of the Ford and GM "domestic" lines are also
imported. Actually, the term "North American-built," meaning "made in the United States, Canada
or Mexico," seems to have replaced the term "domestic" in the sales reports. About 28 percent of all
domestic LDVs sold in the U.S. are considered "imports," i.e., not North-American built, as opposed
to only about  13 percent of all LDTs.

4.1.2   Light-Duty Vehicles vs. Light-Duty Trucks

       In earlier years, light-duty vehicles tended to outsell light-duty trucks by a fairly wide
margin. In 1981, for example, LDTs comprised less than 20 percent of total sales, and this had only
grown to about 38 percent by 1993. However, in recent years the gap has been closing rapidly.
LDTs have made considerable gains in the last decade; by the 2000 model year LDVs outsold LDTs
by a margin of only about 52 to 48 percent. By 2001 the split was roughly 50/50, with LDT sales
actually moving slightly ahead by about 100,000 units.l As shown in Table 4.1-1, for the 2004
model year, LDTs outsold LDVs by a 55 to 45 percent margin. The rise of the Sport-Utility Vehicle
(SUV) accounts for much of this change, but stronger sales of the more traditional LDTs account for
a substantial amount of the increase as well.

       In general, LDTs and LDVs are produced by the  same manufacturers, both foreign and
domestic. The Big Three plus Toyota and Honda account for almost 90 percent of LDT sales. The
Big Three actually account for almost 75 percent of all LDT sales, but only about 45  percent of all
LDV sales. All of the Far Eastern manufacturers, except for Isuzu and Subaru, also make LDTs as
well as LDVs. Isuzu sells only LDTs, in the U.S., while Subaru sells only LDVs. Three European
manufacturers, Volkswagen, BMW, and Porsche, sell both LDTs and LDVs, while the remaining
four European manufacturers sell only LDVs. These four are all small-volume, high-end sports car
manufacturers (Ferrari, Maserati, Lamborghini and Lotus). Figures 4.1-1 and 4.1-2 show market
shares for LDV and LDT manufacturers.
                                          4-4

-------
                         Figure 4.1-1.
LDV MANUFACTURER SALES
       2.4%
         2.4% -,  r 1.4%
               0.8%
     2.9%
     4.0%
  5.9%
 7.0%
 11.0%
                       24.4%
GENERAL MOTORS
FORD MOTOR CO.
VW of AMERICA
BMW GROUP
DAIMLER CHRYSLER  • MAZDA
TOYOTA MOTOR CO.  • SUBARU
                          13.5%
AMERICAN HONDA
NISSAN MOTOR CO.
HYUNDAI GROUP.
        13.7%
                    10.5%
                                                   • MITSUBISHI
                                                   D ALL OTHERS
                         Figure 4.1-2.
LOT MANUFACTURER SALES
              0.8%
         0.3%
10.9%
   17.6%
                         30.1%
                               GENERAL MOTORS
                               FORD MOTOR CO.
                               DAIMLER CHRYSLER
                               AMERICAN HONDA
                               NISSAN MOTOR CO.
                  HYUNDAI GROUP.
                  VW of AMERICA
                  BMW GROUP
                               TOYOTA MOTOR CO.   D MAZDA
                  ALL OTHERS
                    24.8%
                               4-5

-------
       For regulatory purposes, LDVs and LDTs were formerly are divided into categories based
on their gross vehicle weight ratings (GVWR). This distinction was based on the premise that
heavier vehicles produce more pollutants than do lighter vehicles, making it more difficult to achieve
comparable emission reductions. Standards for the heavier vehicles were therefore less stringent.
However, modern emission-control technologies are virtually the same and equallyeffective for both
the lighter and the heavier vehicles. Therefore, the Tier 2 emission standards now make no
distinction between weight categories, except in some cases for medium duty passenger vehicles
(MDPVs), i.e. passenger vehicles over 10,000 Ibs.  GVW, certified to engine-dynamometer
standards. These are primarily the very large SUVs.

       Emission standards were also slightly less stringent for the LDTs than for LDVs, partly
because of weight considerations, and partly because of perceived differences in usage patterns.
Again, the Tier 2 emission standards now make no distinction between LDVs and LDTs, except for
some minor differences in the evaporative emissions standards. In large part this is because LDVs
and LDTs share the same basic emission-control technologies and are primarily used for the same
purpose, for personal transportation. Thus, there does not appear to be a strong rationale for making
distinctions between the two.

4.1.3  Small Volume Manufacturers, Importers, and Alternative Fuel Vehicle Converters

       There are a number of lesser-known small volume manufacturers who produce high
performance and other specialized vehicles, such as Roush Industries or the Panoz Auto Develop-
ment Company. These number less than a dozen, and about half are small businesses. In addition to
the manufacturers, there are a handful of Independent Commercial Importers (ICIs) who are issued
certificates to import a limited number of nonconforming vehicles for racing or other purposes, and
to modify these vehicles to meet U.S. standards.8 These ICIs are almost all considered small busi-
nesses, and total sales for all of them are fewer than 500 vehicles per year. There are also a small
number of converters who convert conventional gasoline- or diesel-fueled vehicles to operate on
alternative fuel (e.g., compressed natural gas and liquefied petroleum gas). These are also few in
number, and are almost all small businesses. Altogether, combined sales for these small-volume
manufacturers, importers, and converters accounted for less than one-tenth of one percent of total
sales of LDVs and LDTs for the 2004 model year.

4.2    Petroleum Refining Industry

       Early in this rulemaking process, EPA commissioned an analysis  of the U.S. gasoline pro-
duction and distribution system from RTI International in order to support economic analyses of the
proposal. The final report of the analysis, entitled "Characterizing Gasoline Markets: A Profile,"
discusses supply and demand issues associated with the refining industry and with gasoline market
behavior.2  The information contained in the report is summarized below, supplemented by addi-
tional information found in this RIA and in other sources.
  ICIs are not required meet the emission standards in effect when the vehicle is modified, but instead they must meet
the emission standards in effect when the vehicle was originally produced (with an annual production cap of a total of
50 light-duty vehicles and trucks).

                                           4-6

-------
4.2.1   Gasoline Supply

       Detailed descriptions of the refinery processes by which gasoline is produced can be found
in the final report mentioned above and in Chapter 6 of this RIA.  Gasoline is the dominant product
for most refineries, constituting almost half of the total product produced by U.S. refineries in 2002.3
 Federal and state regulations have resulted in a variety of gasoline formulations.  These include the
RFG and CG designations, oxygenated gasoline, octane-based gasoline grades, and volatility
distinctions.  Additional variation occurs when different oxygenates are used, though that difference
will lessen significantly in the coming years as MTBE use diminishes and the renewable oxygenate
requirements of the Energy Policy Act of 2005 cause a substantial increase in  ethanol use in
gasoline.  Some gasoline regulations, such as gasoline sulfur and MSAT1,  affect all gasoline and
impact refineries and gasoline production, but do not contribute to additional gasoline types.

       Gasoline supply is also affected by the types of crude oils available, and the refining indus-
try's ability to process the different crude types to maximize gasoline production while meeting all
applicable regulations. Sweet, or low sulfur, crude oils are more easily processed, but this factor
increases their cost compared to sour, or high sulfur, crude oils. Some refineries are optimized to
run based on a certain type of crude oil, and have little flexibility in processing other types. Crude
cost is the largest factor in total refining cost and the price of crude can significantly affect the total
cost of product on.

       Gasoline and other petroleum products are transported from the refineries to intermediate
points such as terminals, and to the final market by pipeline, truck and barge.  Most product is
moved via pipeline, as the cost is extremely low.  Pipelines have been able to accommodate the
many gasoline formulations that have resulted from federal and state gasoline regulations, but are
near their limit in handling additional formulations. Modifying schedules and flow rates in order to
get gasoline and non-gasoline products on and off the pipeline contributes to increased costs. The
final step for gasoline transport to retail outlets is via truck.

 4.2.2  Gasoline Demand

       Gasoline demand is  affected by gasoline use and factors that influence consumption. The
vast majority of gasoline is used for private and commercial highway use.  About 3  percent is used
in non-highway applications such as lawn and garden or marine use. Light duty transportation
accounts for over 90% of gasoline used, and most of this is attributable to private automobile use.
Transportation choices, and thus gasoline use, are affected by many factors, including personal
income, geography, gasoline prices and the prices of related goods.  Though daily travel increases
with household income, average annual expenditures for gasoline, as a percent of income, showed
little variation by geography or income class.  Consumers can respond to gasoline price increases in
many ways, such as reducing the number of miles traveled, or by adjusting their "capital stock," that
is, for example, by purchasing a car with better fuel economy.

4.2.3   Industry Organization

                                           4-7

-------
       The refining industry structure is critical to the implementation and impact of the proposed
regulation. Factors such as regional production and shipment patterns and industry concentration
can influence market price and product availability. For instance, because of current fuel formula-
tions and distribution patterns, consideration of regional (PADD) gasoline markets, rather than a
national gasoline market, may be more appropriate for evaluating certain impacts of the proposed
regulatory program.

       Market concentration refers to some measure of the market share of competitors in an area.
Fligh market concentration may indicate some ability of competitors in an area to influence prices by
coordinated action, thus resulting in less competition and higher product prices.  A recent Federal
Trade Commission analysis has shown that the refining industry is not concentrated or only
moderately concentrated. In addition, the possibility of increased gasoline imports, particularly into
PADDs I and III, can serve to moderate any attempts to set prices.

       Refiners serving the same market may have a wide range of total delivered costs. Cost to the
refiner is a function of distance to market, refinery-specific operating costs and gasoline formulation.
 Gasoline formulation, as discussed, depends on the crude oil, refinery configuration and
environmental or other gasoline controls. The market price for gasoline is set by the producers with
the highest costs, taking into consideration their full range of products produced.

4.2.4   Gasoline Market Data

       An analysis of the impacts of a policy change-in this case, from  current gasoline toxics re-
quirements to the proposed fuel benzene standard—requires consideration of the baseline case com-
pared to likely changes expected from the new policy.  National and regional (by PADD) consump-
tion and gasoline price, price volatility, international trade, and projected growth (in gasoline con-
sumption) are the primary factors considered in estimating economic impacts of the proposed rule.

       Gasoline consumption is estimated to increase by about 1.8 percent annually through 2025.
As discussed above, gasoline consumption, primarily influenced by personal light-duty vehicle use,
is affected by many factors,  including retail gasoline price. Gasoline price is a function of distribu-
tion and marketing costs, refining costs, profit, federal and state taxes, and crude oil cost. Crude oil
cost accounts for almost half of the retail price of gasoline. Price volatility is primarily due to the
magnitude of any supply and demand imbalance, and the speed with which new supply can be pro-
vided.  These imbalances can be caused by unexpected refinery shutdowns or pipeline disruptions,
or even by relatively planned activity, such as seasonal transitions. Isolated markets, or those re-
quiring unique gasoline blends, are likely to be more susceptible to such  supply  and demand
imbalances.

       International gasoline trade, that is, imports and exports of gasoline, account for an
extremely small part of all gasoline transactions.  However, regional activity, at the PADD level,
shows  significant variation.  PADD I received over 90% of all gasoline and gasoline blendstock
imports in 2002.4

4.3    Portable Gasoline Container (Gas Can) Industry
                                           4-8

-------
       EPA also contracted with RTI International for a characterization of the gas can industry in
support of our economic analyses of the proposal. The final analysis report, entitled "Characterizing
Gas Can  Markets: A Profile," discusses production and distribution issues associated with gas
cans.5  This report is also summarized below, and is again supplemented by additional information
found in this RIA and in other sources.

4.3.1   Manufacture and Distribution

       Portable gasoline containers (gas cans) are designed to transport, store and dispense fuel,
normally for refueling vehicles when they run out of gas, or for home applications such as refueling
lawnmowers, trimmers, etc. Gas cans range in capacity from a gallon or less to over 6 gallons.
Standard gas cans have three main components: a spout for pouring fuel, a tank with a fill port to
hold the gasoline, and a vent to make pouring the fuel easier. About 98 percent of all gas cans are
made of high-density polyethylene (HDPE) plastic, chosen mainly because of its fuel-resistant
properties. Two main manufacturing processes are used: extrusion blow molding, which is used for
the bodies, in which a molten tube of plastic is forced into a mold by compressed air; and injection
molding, which is used for spouts, caps and other tubes. In injection molding, plastic material is
forced through a heated injection chamber and through a nozzle into a cold mold. Because of safety
regulations in most states, all gas cans are colored red during the manufacturing process. Industry
and other sources indicate that gas cans are distributed by manufacturers through their distribution
centers to major retail establishments.

4.3.2   Gas Can Use

       Gas cans allow people to refuel a wide variety of equipment without the inconvenience of
taking it to a retail gasoline station. This equipment can range from lawn and garden equipment such
as tractors, lawnmowers, trimmers and chainsaws to recreational vehicles such as motorcycles,
ATVs and golf carts. We estimate that there are about 80 million gas cans in the U.S., which is
similar to other such estimates.6  Although publicly-available data on gas can usage are scarce,  a
California Air Resources Board (CARB) study performed in 1999  indicated that 94 percent of all
gas cans in California were used in households. The remaining 6 percent were used for such com-
mercial applications as farming,  logging, construction, lawn care, and automotive applications  such
as repair  shops and gasoline stations. State surveys in California and Texas indicated that between
46 and 72 percent of all households owned gas cans, and that 14 percent of those surveyed had
bought one during the past year.  The average number of gas cans ranged from 1.4 per household in
Texas to  1.8 per household in California. A typical plastic gas can will have a life expectancy of 3 to
5 years before it needs to be replaced.

       The demand for gas cans reflects the demand for other goods and services. The gas can
industry has suggested that the sales of gas cans are linked to the sales of gasoline-powered equip-
ment such as lawn and garden equipment or recreational vehicles. So factors that influence the sales
of these types of equipment will  also influence the sales of gas cans. These factors can include  such
things as price, population growth, or changes in personal income.

                                            4-9

-------
       Gas can sales for 2002, the latest year for which we were able to develop data, were about 22
million units. Although the gas can manufacturing industry has become fairly concentrated, with one
firm accounting for more than half of all U.S. sales, that firm does not exert significant influence
over market prices. This is because there are few barriers to market entry by other companies, and
the products are substantially the same, making for very limited brand loyalty. Other firms could
enter or re-enter the market should the economic conditions seem right. Imports from Canada, which
amount to about 10 percent of annual sales, would also tend to limit arbitrary pricing practices.

4.3.3  Market Structure

       As noted above, the gas can market is fairly concentrated, with only five firms accounting
for the vast majority of sales. These are Blitz USA, Midwest Can, Scepter Manufacturing, Ltd.
(Canadian), No-Spill Research, and Wedco Molded Products, which is owned by the Plastics Group.
All of these companies, except for the parent company Plastics Group, meet the primary Small
Business Administration (SBA)  criterion for small businesses (i.e., less than 500 employees). There
are other gas can manufacturers, but they have a very limited market share. Most of their products
are designed for industrial use or to fill a niche market (e.g., racing, or other safety cans used in an
industrial setting), which would not be covered by the proposed standards. These companies include
Eagle Manufacturing, Protectoseal Company, and Scribner Plastics. These firms all meet the SBA
criteria for small businesses. Table 4.3-1 provides relevant data about these firms.

4.3.4  Market Entry

       There are very few barriers to entering the gas can market. Only  about 2 percent of the gas
cans sold in the U.  S. in 2002 were of metal construction; the vast majority were plastic. These are
produced by a fairly straightforward molding process in much the same manner as hundreds if not
thousands of other plastic products. Plastic gas cans are in fact classified in the U.S. Economic
Census as "All other plastics product manufacturing." Since manufacturing such gas cans is similar
to manufacturing most other molded plastic products, any firm with that capability could freely enter
the market with a relatively low initial investment, if the economic conditions should appear
advantageous to do so.  Since most consumers tend to view gas cans as more or less all the same,
there is not a well-developed brand loyalty to one brand or other, so competition in the industry is
based primarily on price. Finally, safely regulations in most states prevent consumers from using old
paint thinner cans or other such containers as substitutes for gas cans, thus eliminating any potential
reduction in sales from that quarter.
                                           4-10

-------
























tfi
Vf
o:
LU
IV
3
^B
Q
<(
1










m
u
s
g
O
u





h-
Z
UJ
>-
o
1
a.
e



z"
t/M ^^
J-J
•
C/f 5
••
S



COMPANY
NAME



ULTIMATE
PARENT





i-
LU
•XL
Of
<
(CT
S
D£
LU
=3
CO
z
o
o




o
CM






o
ID

cb
CM





Z)
N
CO



to
N
i
CD
LU
LL
ft
CO
^r
1—
LU


_l
5
ji
*£
t£
O_



O>
CM
O
O






O
T™*
LO



O
Z
EAGLE MANUFACTURI
O


MANUFACTUR
LU
O
<£
LU


1-
LU
*L
Q;
<
"•ST
S
ry
LU
to ^
Z £Q
7; 2
O
o




LD
"*






o
ID
i
O
CM





MIDWEST CAN



O
CO
LU
1
•---i
i

z

1-
=3
m
of
h-
to
Q
Q
1 —

cE
Zi












in
m
CM



u
Z
NO-SPILL RESEARCH,
d
z

LL RESEARCH:
Q.
to
1
o
z

<
(V
UlM
h-
to
u
Q
Z

_l
E
2
^
a:
a.



en
CM
ci
o






o
ID
1
o
CM





PROTECTOSEAL CO.



CTOSEAL CO.
LU
5
Of
a.
Of
LU
^
^£
to
z
o
o

z
o

z

o




o
CM






o
CM

cb
t —





SCEPTER MFG., LTD.



ci
i-
d
LL
Of
LU
H-
Q_
LU
O
to
CO

LU
Z
O
o
x_
5
o
LU
Q.
CO




CJ5
^y
CD
CM






O
T™*
A





SCRIBNER PLASTICS



IER PLASTICS
S
QL
O
to


1-
LU
•XL
QL
<
^"
S
Of
LU
=3
CO
Z
o
o




o







o
ID
i
O
CM




Ci
WEDCO MOLDED PRO



a.
o
a:
O
O
h-
co
_J
a.
LU
I
h-
CD
Z
fy*
U«i
Z)
o
<
LL
Z)
z
2
1—
o
Z)
Q
O
o:
a.
to
O
i—
to
^
Q.

o:
LU 0
T 2
o s
j _3

*QT
of u_
^^ _l

CM <
\GED IN NAICS CODE 3
332431, METAL CAN M
O LU
Z Q
LU n
BUSINESSES
OR NAICS C(
*









RS
g
\tf
"(8
C
O
"J^
™~"
P
a:
(D
v=
O
LKM
Q.

(0 tj
a> ^
"<5 (j)
•*§>
" Q
s g
(d
0 13

-------
References for Chapter 4


1  Source: Ward's "World Motor Vehicle Data, 2004."

2 "Characterizing Gasoline Markets: A Profile," Final Report. EPA Contract Number 68-D-99-
024, prepared for Robert Johnson, USEPA, Office of Transportation and Air Quality, Ann Arbor,
MI by Brooks Depro, et al, RTI International, Research Triangle Park, NC, August 2004.

3 Table 1-1 in Final Report. Data Source: DOE, EIA Petroleum Supply Annual 2002.

4 Table 4-4. Final Report. Source DOE EIA Petroleum Supply Annual 2002.

5   "Characterizing Gas Can Markets: A Profile." Final Report, EPA Contract Number 68-D-99-
024, prepared for Robert Johnson, USEPA, Office of Transportation and Air Quality, Ann Arbor,
MI by Brooks Depro, et al, RTI International, Research Triangle Park, NC, August 2004.

6 Memorandum from Terrance R. Karels, Consumer Product Safety Commission, January 3, 2003.
                                        4-12

-------
                              Chapter 5: Table of Contents

Chapter 5:  Vehicle Technological Feasibility	2
   5.1    Feasibility of Cold Exhaust Emission Standards for Vehicles	2
     5.1.1   NMHC Emissions Control Technologies on Tier 2 Gasoline-Fueled Vehicles	2
        5.1.1.1  Calibration and Software Control Technologies	3
           5.1.1.1.1  Idle Speed and Air Flow Control	4
           5.1.1.1.2  Spark Control	4
           5.1.1.1.3  Secondary Air Injection Control	5
           5.1.1.1.4  Cold Fuel Enrichment	6
           5.1.1.1.5  Closed Loop Delay	7
           5.1.1.1.6  Transient Fuel Control	8
           5.1.1.1.7  Fuel Volatility Recognition	8
           5.1.1.1.8  Fuel Injection Timing	9
           5.1.1.1.9  Spark Delivery Control	10
           5.1.1.1.10  Universal Oxygen Sensor	10
        5.1.1.2  Tier 2 Engine and Exhaust Control Technologies	10
     5.1.2  Data Supporting Cold NMHC Standard Technical Feasibility	11
        5.1.2.1  Certification Emission Level	11
        5.1.2.2  EPA Test Program	16
           5.1.2.2.1  2004 Chevrolet Trailblazer Feasibility	17
           5.1.2.2.2  Additional Tier 2 Vehicle Feasibility	20
   5.2    Feasibility of Evaporative Emissions Standards for Vehicles	20
                                           5-1

-------
               Chapter 5:  Vehicle Technological Feasibility

5.1    Feasibility of Cold Exhaust Emission Standards for Vehicles

5.1.1   NMHC Emissions Control Technologies on Tier 2 Gasoline-Fueled Vehicles

       Emission control technology has evolved rapidly since the passage of the CAA
Amendments of 1990. Emission standards applicable to 1990 model year vehicles required
roughly 90 percent reduction in exhaust non-methane hydrocarbon (NMHC) emissions  compared
to uncontrolled emission levels. The Tier 2 program and before that, the NLEV program, contain
stringent standards for light-duty vehicles that have further resulted in additional NMHC
reductions. Tier 2 vehicles currently in production show overall reductions in NMHC of more
than 98 percent compared to uncontrolled emissions levels. These emission standards for NMHC
are measured under the EPA Federal Test Procedure (FTP), which measures exhaust emissions
from vehicles operating only in the ambient temperature range of 68° F to 86° F.

       Table 5.1-1 below lists specific types of NMHC emission controls that EPA projected in
the Tier 2 technological feasibility assessment could be used in order to meet the final Tier 2
standards.  It is important to point out that all of the following technologies have not necessarily
been needed to meet the Tier 2 standards. The choices and combinations of technologies have
depended on several factors, such as current engine-out emission levels, effectiveness of existing
emission control systems, and individual manufacturer preferences. In some cases, no additional
hardware from the NLEV level of hardware was  needed. Instead, many manufacturers focused
their efforts in the software and calibration controls to  achieve stringent emission levels.

       Table 5.1-1. Tier 2 Projected Emission Control Hardware and Technologies
Emission Control Technologies
Fast Light-off Exhaust Oxygen Sensors
Retarded Spark Timing at Start-up
More Precise Fuel Control
Individual Cylinder Control
Manifold with Low Thermal Capacity
Air Assisted Fuel Injection
Faster Microprocessor
Secondary Air Injection into Exhaust
Heat Optimized/Insulated Exhaust Pipe
Close-coupled Catalyst
Improved Catalyst Washcoats/Substrates
Increased Catalyst Volume and Loading
Engine Modifications
Universal Exhaust Oxygen Sensor
       A number of technological advances and breakthroughs have allowed these significant
emission reductions to occur without the need for expensive emission control equipment. For
example, the California Air Resources Board (ARE) originally projected that many vehicles
would require electrically-heated catalysts to meet their LEV I program requirements. Today,
with even more stringent standards than LEV I, no manufacturer needs to use these devices to
comply with program requirements.  Similarly, the Tier 2 and LEV II programs, currently being
phased-in, have projected that some  additional emission control hardware and techniques may be
required. However, initial indications from the Tier 2 vehicles already certified indicate that
increases in hardware content have been kept to a minimum, likely to minimize cost.
                                          5-2

-------
       The Tier 2 program requires reductions in all regulated pollutants, but the largest
reductions are required for NOx emissions. To achieve these NOx reductions, significant
improvements in catalyst technologies have been employed, largely in improved catalyst
substrates and washcoats containing the precious metals.  In fact, some manufacturers have even
been able to reduce precious metal loadings as compared to previous generation catalysts
because of the new substrate and washcoat improvements developed in response to Tier 2. These
catalyst technologies have generally also resulted in better emission performance of all  regulated
pollutants, largely because of improved catalyst light-off times.

       The Tier 2 program also includes new tighter non-methane organic gases (NMOG)
standards. Unlike tight NOx controls, manufacturers had significant experience in NMOG
controls, and therefore NMHC controls, primarily from the stringent NMOG standards  under the
NLEV and LEV I programs. In fact, the NMOG standards for a Tier 2 Bin5 package are the
same as the passenger car (PC) and light-duty truck (LDT1) as those established under the
NLEV program. The largest challenge manufacturers have encountered under the Tier 2 program
is possibly the program's weight neutral standards for all vehicles up to 8500 Ibs. gross vehicle
weight rating (GVWR) and medium-duty passenger vehicles (MDPV) up to 10,000 Ibs. GVWR.
These heavier vehicles may be where new hardware will more likely be required to meet Tier 2
weight neutral standards as they fully phase in to Tier 2.

       Some of the most significant technological advances that have facilitated low NMHC
emission levels  have occurred in calibration and software based controls. These controls have
been carefully designed to both minimize exhaust emissions before exhaust aftertreatment
readiness and accelerate the usage of the aftertreatment earlier in the operation of the engine.
Additionally, fuel metering controls during the critical period  prior to aftertreatment readiness is
more precise than previous systems, largely due to advances in software controls. While some
improvements also have been  made to base engine designs, which have resulted in lower overall
operating engine-out emissions, controls aimed at minimizing emissions during the critical
period before exhaust aftertreatment readiness have been done almost exclusively with  software
based controls.  Even with base engine and exhaust hardware improvements, calibration and
software controls of the emission control hardware remain the most important and powerful
emission control technique used by manufacturers. Calibrations and software controls will
continue to become more refined and sophisticated as manufacturers learn new ways to better
utilize existing hardware, particularly in the remaining Tier 2 phase-in vehicle models.

       Today, these emission control strategies are utilized at 75° F to meet stringent Tier 2 and
LEV II NMOG standards.  The potential exists for these same software and calibration  controls
to be utilized at 20° F and all other cold start temperatures to control NMHC emissions.  Most of
these controls are feasible and available today in Tier 2 and LEV II vehicles. With the
implementation of these controls at the colder start temperatures, significant reductions in
NMHC emissions (and therefore air toxics) can be realized. The following sections provide
details on these software and calibration control strategies, supporting certification results, and
feasibility studies utilizing these existing emission control opportunities.

5.1.1.1       Calibration and Software Control Technologies
                                           5-3

-------
       Tier 2 vehicles are equipped with very sophisticated emissions control systems. Table
5.1-1 above lists some of the technologies manufacturers have successfully used to meet
stringent Tier 2 emission standards. In addition to hardware technologies, manufacturers have
developed calibration and software control strategies to meet Tier 2 emission standards that also
can be effectively used at 20° F to achieve significant reductions in NMHC and other emissions.
We expect manufacturers will expand the use of these same emission control strategies already
in place on Tier 2 vehicles at 75° F to control NMHC emissions at 20° F. The following
descriptions provide an overview of the calibration and software technologies capable of
reducing exhaust emissions at 20° F.

5.1.1.1.1      Idle Speed and Air Flow Control

       Idle speed and air flow control have been utilized very successfully as a method to both
reduce emissions before the catalyst aftertreatment is considered active and to accelerate the
activity of the catalyst. Elevated idle speeds immediately following the start of a vehicle,
particularly in park and neutral, will result in more stable combustion resulting from the
improved air and fuel mixture motion. This is largely due to the higher air velocity entering the
combustion chamber which generally results in a more homogeneous mixture,  and therefore, a
more fully combustible air-fuel mixture. The higher engine speed may also increase heat created
from piston to cylinder wall friction, further assisting in transforming fuel  droplets to burnable
mixtures. The higher engine speeds cause additional combustion events, further assisting in the
rapid heating of the combustion chamber. The higher combustion stability can generally result in
the ability to run leaner air-fuel ratios, which reduces the percentage of unburned fuel that would
be exhausted from the engine.

       Air flow through the engine, which is exhausted after combustion,  provides the heat
required for the catalyst to become active.  Increased air flow through the engine, mainly through
elevated idle speeds, provides the catalyst with supplemental heat. Additionally, this extra
exhaust heat is carried to the catalyst at higher exhaust flow velocities, further shortening the
amount of time the catalyst is inactive.  The higher combustion stability from the increased air
flow provides the catalyst with a preferable mixture composed of less lost fuel in the form of
hydrocarbons, which can actually quench a catalyst and slow its warming.  The ability to run
leaner mixtures can provide the catalyst with the necessary oxygen for the catalyst to begin
oxidation of NMHC and carbon monoxide (CO).

       Elevated air flow used off idle can also produce significant emission benefits.  This
elevated air flow is achieved by allowing extra air flow primarily when the throttle is closed but
also during the transient period when the throttle is in the process of closing. This momentary air
flow increase has been referred to as "dashpot" effect. It typically has been used only for short
durations following a throttle closing to help provide additional air flow, and usually only during
the first few minutes of cold start engine operation. Elevated air flow has also been used to
provide slightly more closed throttle engine torque to overcome additional loads only
encountered following a cold start. This reduces risk of idle undershoots and stalling.

5.1.1.1.2      Spark Control
                                           5-4

-------
       Spark control has evolved with modern electronic controls to a highly precise tool to
carefully control when the combustion event is initiated in a spark ignition engine. Retarding the
spark delivery immediately after the start has been highly effective at reducing exhaust
emissions.  Retarding the spark, particularly after a cold start, generally reduces engine-out
emissions. This is generally believed to be a result of the longer period of time that the fuel is
under compression and absorbing combustion chamber heat.  This assists in more complete
combustion when the fuel is finally spark-ignited.  It also is believed that the retarded spark
timing results in lower cylinder peak pressures during the combustion of the air-fuel mixture,
reducing the opportunity for hydrocarbons to migrate to crevices and further helping lower
engine-out hydrocarbon emissions.

       Retarded timing  also has been used very effectively to accelerate the early usage of the
catalyst by providing supplemental heat, which reduces the time for the catalyst to begin
oxidation. The retarded timing results in peak combustion of the air-fuel mixture occurring later
in the engine operating cycle, leading to significant thermal energy being transferred into the
exhaust. This thermal energy very effectively provides a boost to the catalyst warm-up,
particularly at colder temperatures and for large mass catalyst systems or catalyst systems that
are further from the engine than desirable.

       The effectiveness of retarded timing can be enhanced significantly when used in
conjunction with elevated idle speeds and/or air flow control. The simultaneous use of the two
features generally is much more effective than either feature used independently,  and the
resulting emission reductions can be much higher than sum of each feature measured
independently.  Additionally, utilizing elevated idle speeds while retarding the timing can offset
any engine vacuum level concerns encountered when only retarding timing is used.

5.1.1.1.3      Secondary Air Inj ection Control

       Many Tier 2 vehicles produced today contain secondary air injection systems to comply
with stringent Tier 2 and LEV II standards.  These systems reduce vehicle emissions by injecting
ambient air into the rich engine exhaust upstream of the catalyst for a short period of time
immediately after a start. This reduces emissions in two ways. First, the oxygen in the ambient
air being pumped into the exhaust assists in oxidizing HC and CO prior to reaching the catalyst.
Second, this oxidation can result in the generation of highly desirable, large amounts of heat that
help bring the catalyst to effective temperatures much sooner.  As the catalyst reaches effective
temperature, the secondary air can continue to provide needed oxygen for oxidation in the
catalyst until the total system is ready to go "closed loop," at which time the secondary air
injection is ceased.

       The secondary air injection technology for controlling emissions is not a new technology.
For many years, manufacturers used secondary air injection systems that ran continuously from a
mechanical belt-driven pump to oxidize HC and CO  emissions produced from a rich exhaust
mixture. With the advent of the three way catalyst (TWC), manufacturers began to use engine
control modules (ECM)  to activate electric air pumps to reduce start emissions only at 75° F,
typically on vehicle packages with specific emission challenges. For example, vehicles that have
                                           5-5

-------
large mass catalysts or catalyst systems located relatively far from the engine have utilized
secondary injection to assist catalyst light-off. Further, many Tier 2 and LEV II packages
certified to the cleanest emission levels utilize secondary air injection to achieve these results.
Some Tier 2 packages that appear to have relatively high engine-out emissions, possibly due to
engine design limitations, also have implemented secondary injection to allow compliance with
Tier 2 emission standards.

       Many manufacturers that have equipped some of their Tier 2 vehicles with secondary air
injection systems do not appear to consistently utilize this emission control strategy across start
temperature ranges outside of the currently regulated cold start temperature (75° F for Tier 2 and
50° F for LEV II).  However, many vehicle models common to Europe and the U.S. that are
equipped with secondary air injection do appear to be using this technology at 20° F on models
sold in the U.S., based on our analysis of the certification data.  This is attributable to common
emission control technologies with the European market vehicles, where manufacturers are
already required to meet a 20° F NMHC standard.

       The activation of the secondary air system is a feasible and effective emission control
technology for 20° F as well as all other interim start temperatures. The use of secondary air
injection technology at 20° F is well  proven as an emission control technology,  as observed in
the European vehicles. Certain design criteria must be taken into account for the system to
operate robustly at these colder temperatures, but there  appears to be no technological challenge
that would prevent these vehicles already equipped with secondary air injection from activating
this emission control technology at 20° F.

       Some manufacturers, who do not use secondary air injection systems at 20° F but do
include the systems on some of their U.S.-only models, have expressed concerns with freezing
water in the system. We have investigated this concern with the manufacturers of the secondary
air injection components and found this to be a system design issue that has been addressed by
guidelines on the location and plumbing of the individual secondary air injection components.1

5.1.1.1.4      Cold Fuel Enrichment

       Gasoline-fueled spark ignition engines generally require rich air-fuel mixtures (i.e., a
larger amount of fuel for a given amount of air) for some amount of time immediately following
a cold start. Under normal operating  conditions, the amount of required enrichment always
increases as start temperature decreases. This is largely because low in-cylinder temperatures for
some period of time following the cold start lead to a lower percentage of liquid fuel vaporizing
to a burnable mixture. The level of enrichment and its duration following the start will vary with
many factors, including base engine  hardware design and fuel properties. Fuel property
interactions with engine combustion  chamber dynamics are quite complex and can vary with fuel
composition, but typical gasoline fuel available in the U.S. during the cold weather (e.g., 20° F)
is properly formulated for robust cold start operation.

       The level of enrichment should be calibrated to  closely match the "winter" grade fuel
properties that the overwhelming majority of vehicles will be experiencing during the colder start
conditions. Winter grade fuel is formulated to have a higher Reid vapor pressure  (RVP),
                                           5-6

-------
specifically to allow the fuel to vaporize at lower cold start temperatures and minimize the need
for additional enrichment. Any fuel enrichment beyond the minimum required level results in
proportional increases in cold start emissions, primarily NMHC and CO. Additionally, over-
fueling can hamper earlier use of the exhaust aftertreatment by quenching the catalyst with the
unburned fuel, effectively cooling the catalyst. This retards the warm-up rate of the catalyst and
also reduces the availability of any excess oxygen that would be used by the catalyst to oxidize
the NMHC and CO.

       The amount of required enrichment also can be reduced when used in conjunction with
the previously mentioned elevated idle speed emission control technology.  As stated earlier,
elevated idle speeds will result in a more homogeneous mixture which supports more stable
combustion. The improvements in the mixture will allow the enrichment levels to be reduced
accordingly.

5.1.1.1.5      Closed Loop Delay

       "Closed loop" operation refers to operation that allows the exhaust oxygen sensor to feed
back to the engine control module and  control the air-fuel mixture to an exhaust stoichiometric
ratio.  Following start-up of a modern gasoline fueled engine, operation in closed loop is delayed
for some amount of time based on a combination of engine and oxygen sensor readiness criteria.
As stated in the previous section, gasoline-fueled engines require rich air-fuel mixtures for some
amount of time immediately following a start. The amount of time requiring the rich operation
and, therefore, the delay of exhaust stoichiometric operation, will vary with the gasoline engine's
ability to operate smoothly at these air-fuel ratios.

       The delay also will be determined by the exhaust oxygen sensor's ability to properly
function. Modern exhaust oxygen sensors, including both conventional switching and universal
linear sensors, contain heating elements to allow them to maintain proper operating  sensor
temperatures and also to be used sooner following a cold start.  These internal heating elements
require careful control to prevent any potential thermal shock from water or fuel in the exhaust
stream. The water is generated from the combustion  process but also can be present in the
exhaust pipe from condensation of water, particularly during certain ambient temperature and
humidity operating conditions. Generally, cold starts  at 20° F only require a short delay to allow
the initial heating of the exhaust manifold to vaporize any combustion water.  This period is
followed by an electronically controlled and monitored heating of the sensor. Exhaust oxygen
sensors have been designed to have significant protection from water and are typically fully
operational well before the engine is prepared to use their information.

       Generally, within approximately one minute of 20° F cold start operation, combustion
chamber temperatures are at levels that vaporize sufficient amounts of the gasoline fuel to
command exhaust stoichiometric operation of the engine. Also within that minute, exhaust
oxygen sensors should have sufficient time to reach operating temperature with any thermal
issues mitigated, allowing closed loop stoichiometric operation. As stated earlier, operating a
gasoline-fueled engine at stoichimetry provides the exhaust aftertreatment with oxygen required
for oxidation of HC and CO. Therefore, the amount of time requiring enrichment should be
                                           5-7

-------
minimized and closed loop operation of the emission control system should be able to occur as
soon as physically possible.

5.1.1.1.6      Transient Fuel Control

       The control of the air-fuel ratio during transient maneuvers (i.e., operator-induced throttle
movement) has dramatically improved with modern hardware and software controls. This is
largely due to the improved accuracy of both the measurement sensors and the fuel delivery
devices, but also refined software modeling of both air flow and physical fuel characteristics.
Tier 2 vehicles have highly accurate sensors that measure changes in air flow to predict and
deliver the appropriate amount of metered fuel. Additionally, the software that interprets these
sensor signals  has evolved to predict transient behaviors with much higher accuracy than ever
before. Many  of these improvements were necessitated by increases in emission stringency in
the recent Tier 2 and LEV II programs, which were much less tolerant of transient errors that
were acceptable in past emission control systems.

       With the recent widespread penetration of electronic throttle controls  (ETC), partially in
response to the stringent Tier 2 and LEV II 75° F standards, manufacturers have been able to
further reduce  variability of transient errors. ETC applications remove the direct mechanical
connection from the accelerator pedal to the engine. Instead, the pedal is simply a sensor that
reports pedal movement to the ECM. The ECM interprets the pedal movement and provides a
corresponding controlled movement of the engine throttle.

       Transient air-fuel errors can be minimized through advanced approaches to ETC usage.
This is possible because the electronic  controls can better synchronize the introduction of the
transient maneuver and closely match required air and fuel amounts.  The controls can be
designed and programmed to prevent most of the transient errors experienced with older cable-
driven mechanical systems.  The older mechanical systems resulted in reactionary response to
throttle movements, making  it significantly more difficult to deliver precise dynamic air-fuel
control.  Since the ETC systems control the actual movement of the throttle, they  have the ability
to essentially eliminate transient errors by preceding the throttle movement with appropriate fuel
metering amounts. This is particularly important at colder temperatures (i.e.,  20° F cold start)
where transient errors can be exaggerated when the engine is operating rich of stoichiometry.

5.1.1.1.7      Fuel Volatility Recognition

       Improved modeling of the effect of fuel properties on engine and emission performance
has eliminated the need for a new sensor.  For instance, some manufacturers have successfully
designed software models that can determine the percentage of ethanol in the fuel on which the
vehicle is operating. These "virtual sensor" models take into account information  from sources
such as existing sensors and use historical data for the determinations. The models use this
information to adjust many outputs including fuel metering and spark ignition control.

       Currently, manufacturers have active software features  that are designed to recognize and
recover from a lean condition that can be a precursor to an engine stall. These features use
different input criteria to identify and actively change the air-fuel ratio when  an excessively lean
                                           5-8

-------
condition may be occurring.  These features may look at control parameters such as engine speed
(RPM), engine manifold absolute pressure (MAP), engine mass air flow (MAP), and even engine
misfire-related information to determine if a fuel metering change should occur.

       The approaches described above exemplify possible software-based control designs that
can achieve the desired emission and engine performance characteristics.  Manufacturers have
extensive experience designing and implementing software features to identify and react to
specific fuel parameters that are deemed important to engine operation. The ability to recognize
fuel volatility and actively adjust the fuel metering accordingly would allow the gasoline-fueled
engine to operate at the lean limit, reducing engine-out emissions, particularly NMHC and CO.
Much like the  "virtual sensor" model described above for ethanol content, this model would take
existing sensor information and other information available from the ECM and determine the
fuel volatility characteristics at any given cold start temperature.  The modern engine controllers
have the ability to maintain significant historical data that can help predict fuel properties. The
items of importance for fuel volatility may include ambient temperature exposure of fuel, amount
of time since previous start, and other related items.

5.1.1.1.8      Fuel Injection Timing

       Fuel injection timing control is another emission control technology that has evolved as a
result of increased  computing power of the engine. Depending on the engine design and the
thermal characteristics of the intake port design, significant opportunity may exist  for optimizing
fuel preparation prior to combustion.

       Generally, there are two fuel injection timing approaches used to optimize fuel
preparation:  closed valve injection and open valve injection. Closed valve injection is the
traditional method  of injecting fuel into the cylinder head intake port. As the name states, the
intake valve is closed during the injection time period.  This approach allows the fuel to have
residence time in the intake port prior to ingestion into the cylinder. Usually, the fuel injector is
targeted to spray the fuel on the back of the closed intake valve in order to allow the fuel to
absorb any heat conducted through the valve from the combustion events occurring inside the
cylinder chamber.  The heat absorbed by the fuel potentially allows more of the fuel to vaporize
either in the port or in the chamber, resulting in higher percentage of vaporized fuel that can be
combusted.  If the higher percentage of vaporized fuel burns, less liquid fuel will be exhausted,
effectively reducing the engine-out NMHC levels.

       Open valve injection involves carefully coordinating the fuel injection timing in  order to
inject fuel while the intake valve is in some state of opening. This approach attempts to take
advantage of the incoming air velocity as the air is  drawn through the port and also the intake air
pressure depression.  The mixture motion  and depression can help vaporize the fuel and assist in
better mixing of the air and fuel prior to combustion, resulting in improved fuel burn. This
approach is dependent on many aspects, including injector spray design, injector targeting, intake
valve timing, and intake valve lift. Open valve timing may be used initially after engine start
followed by  a closed valve approach, described previously, once the intake valve is heated.
Many similar approaches are detailed in past Society of Automotive Engineers (SAE) papers2.
                                           5-9

-------
5.1.1.1.9      Spark Delivery Control

       With the increases in the computing power of the engine controller, opportunities have
been created for new spark delivery related emission control features. Separate from the retarded
timing benefits described previously, there are other potential controls that may help reduce
engine-out emissions. Many new engines contain individual cylinder ignition coils. With these
individual coils comes the opportunity for individual cylinder-based spark control features
designed to promote more complete combustion. Additionally, some new engines have dual
spark plugs (i.e., two plugs for each cylinder). These dual spark plug systems may have
opportunities for new concepts targeted at emission reductions, particularly following cold start
operation.

       Spark energy, the amount of energy delivered to the spark plug that is used to ignite the
air-fuel mixture, can be carefully controlled by modifying the dwell time delivered to the ignition
coil. The dwell time is the amount of time that the ignition coil is allowed to be charged with
electrical energy. An increase in dwell time will generally result in an increase in spark energy
delivered to the spark plug.  Higher spark energy typically results in a higher burn rate
particularly in air-fuel mixtures that are not optimized, which is typical of mixtures at start-up.

       Other new concepts may include such ideas as multiple spark events on a single engine
cycle. The concept of delivering redundant spark events has been used in the past, primarily for
engine performance. While we do  not currently know if redundant spark events are beneficial in
reducing emissions, it could be explored for emissions control.  Similarly,  dual spark plug
engines or engines with individual  cylinder ignition coils can explore other spark delivery related
concepts that may prove to be effective emission control tools.

5.1.1.1.10     Universal Oxygen Sensor

       As listed in Table 5.1-1 above, universal oxygen sensors were projected to be an emission
control hardware that could be used to meet Tier 2 vehicle standards.  Several manufacturers did
in fact decide to replace their conventional switching oxygen sensors with these universal
oxygen sensors. Universal oxygen  sensors have  certain benefits over conventional switching
sensors that should prove substantially beneficial at 20° F.  While these sensors require a similar
delay to reach operating temperature following a start, universal oxygen sensors can accurately
control the air-fuel ratio during rich operating conditions prior to commanded closed loop
operation. Conventional switching  sensors cannot indicate the actual air-fuel ratio during rich
conditions, therefore preventing them from being used as a control sensor during critical rich
operation. Additionally, universal  oxygen sensors can be used to more accurately recover from
air-fuel transient errors during the warm-up due  to their ability to measure the magnitude of the
error.

5.1.1.2        Tier 2 Engine and Exhaust Control Technologies

       The Tier 2 technological feasibility assessment described several engine and exhaust
hardware control technologies that  could be used to meet stringent Tier 2 emission standards.3
These technologies continue to be very effective emission control strategies to meet Tier 2
                                           5-10

-------
standards. We believe that manufacturers will use these same Tier 2 technologies in order to
meet the proposed 20° F NMHC standard. We do not expect that manufacturers will need to
utilize additional emission control hardware. However, if a manufacturer chose to do so, most of
these  same Tier 2 technologies can also be used to meet the proposed 20° F NMHC standard.

5.1.2   Data Supporting Cold NMHC Standard Technical Feasibility

       Data to support the feasibility of complying with a 20° F NMHC standard are presented
in the following two sections.  The first section includes evidence from recent model year
certification emissions data submitted to EPA. Certification data are required to include cold
temperature carbon monoxide emissions data, and some manufacturers have also included
associated cold temperature total hydrocarbon emissions data. The second section provides
evidence from a feasibility evaluation program recently undertaken by EPA.  This program
examined the effects of making only calibration modifications to vehicles with 20° F NMHC
levels that were significantly higher than the industry average.

       When considering the supporting data, it should be noted that manufacturers generally
design vehicles to incorporate a compliance margin in their exhaust  emissions controls systems
to account for operational variability.  Specifically, they will design  controls to meet emissions
targets below the standard when using catalytic converters thermally aged to the full useful life.
By ensuring that emission targets are met when testing on artificially aged converters,
manufacturers reduce the probability that in-use vehicles will exceed the relevant standard
throughout the useful life of the vehicles.

       However, the data presented in the following sections do not explicitly incorporate a
compliance margin since the cold temperature NMHC data, at the time they were submitted to
the EPA, were not subject to EPA standards.  The data represent the cold NMHC emissions as
tested, and only suggest that a significant number of vehicles are within reach of today's
proposed standards

5.1.2.1       Certification Emission Level

       Manufacturers are required to report carbon monoxide (CO)  exhaust emissions test
results for compliance with cold temperature CO  standards (i.e., the  20° F FTP test) for light-
duty vehicles, light-duty trucks, and medium-duty passenger vehicles. Many manufacturers have
included total hydrocarbon (THC) cold temperature exhaust emission data that are collected
along with cold CO data.  In addition, several of these manufacturers also reported test results for
both the THC emission data and the matching NMHC emission data. Based on these data from
manufacturers who have included both THC and NMHC cold temperature data,  non-methane
hydrocarbons (NMHCs) account for approximately 95% of total hydrocarbon emissions at cold
temperatures. Therefore, a review of the more abundant THC data provides a reasonable means
of assessing manufacturers'  cold NMHC emissions performance.

       EPA analyzed 2004, 2005, and 2006 model year full useful life certification data for
vehicles certified to nationwide Tier 2 standards, NLEV program standards, and the California
program standards. Lists were compiled from certification data submissions that reported
                                          5-11

-------
unrounded cold THC results and for which an associated FTP full useful life deterioration factor
(DF) was available.  The DF is incorporated into the emissions result to estimate emissions at the
full useful life of the vehicle.  The DF was applied to the unrounded test result, and that result
was rounded to one decimal point.  This calculation was then compared to the cold temperature
NMHC standards of 0.3 g/mi for LDV/LLDTs, and 0.5 g/mi for HLDT/MDPVs.

       Table 5.1-2 shows the number of car lines for which the resulting calculation for total
hydrocarbons was at or below the 0.3 g/mi NMHC standard for LDV/LLDTs, and at or below
the 0.5 g/mi NMHC  standard for HLDT/MDPVs. Again, these data only reflect an analysis of
those car lines for which manufacturers voluntarily provide cold THC data.

       Tables 5.1-3, 5.1-4, and 5.1-5 show, by model year, the total hydrocarbon emission levels
(calculated according to the method described above) for LDV/LLDTs at or below 0.3  g/mi,  and
HLDT/MDPVs at or below 0.5  g/mi. For each manufacturer, the data were grouped according to
car lines with the same calculated cold THC emission result.  Where a range is shown for the
emission level, tests on multiple configurations within the car line yielded a range of results.

Table 5.1-2. Number of car lines with one or more engine families whose certification data
       for total hydrocarbons was at or below the proposed cold NMHC standards
Year
2004
2005
2006
LDV/LLDTs
41
42
44
HLDT/MDPVs
13
16
22
Total Car Lines
54
58
68
       As the tables suggest, there are already a significant number of vehicle configurations,
across a wide range of vehicle types and manufacturers, within reach of the proposed cold
temperature NMHC standards.  Though the number of LDV/LLDT configurations at or near the
proposed cold NMHC standards significantly outnumber the heavier HLDT/MDPVs, EPA is
proposing a later phase-in for HLDT/MDPVs due to the unique challenges related to these
heavier vehicles, as discussed in Section VI of the Preamble.  The number of configurations
approaching the proposed standard increases for both LDV/LLDTs and HLDT/MDPVs from
2004 to 2006, as vehicles have adopted more stringent emission controls to meet the Tier 2
standards.

       This analysis does not necessarily imply that manufacturers could have certified these
vehicles to meet the proposed cold NMHC standards.  But the data do support the feasibility of
meeting such standard levels. This analysis is conservative given that actual NMHC emissions
would be slightly less than that of the total hydrocarbon emissions, and not all of the vehicles
included here were certified to the more stringent Tier 2 standards. That is, some vehicles in the
certification data are interim non-Tier 2 vehicles.  We would expect hydrocarbon levels to be
somewhat lower as these vehicles fully phase-in to Tier 2.
                                         5-12

-------
     Table 5.1-3.  2004 model year vehicles with certification data
for total hydrocarbons at or below the proposed cold NMHC standard
MANUFACTURER
CAR LINE
COLD TOTAL HC LEVEL
LDV/LLDTs
ACURA
ACURA
ACURA
AUDI
BMW
BMW
CADILLAC
CHEVROLET
HONDA
HONDA
HONDA
HONDA
HONDA
HONDA
HYUNDAI
MAZDA
MAZDA
MERCEDES-BENZ
MERCEDES-BENZ
MITSUBISHI
MITSUBISHI
NISSAN
NISSAN
SATURN
TOYOTA
TOYOTA
VOLKSWAGEN
VOLVO
1.7EL, TL
MDX 4WD
RSX
A4 QUATTRO
3251 SPORT WAGON, 330CI CONVERT.
X3
CTS
CORVETTE
ACCORD
CIVIC
CIVIC HYBRID, INSIGHT
CR-V 4WD, ELEMENT 4WD, S2000
ODYSSEY 2WD
PILOT 4WD
XD-5DR
MAZDA 3
MAZDA 6, MAZDA 6 SPORT WAGON, MPV
C240 (WAGON), C-CLASS SEDAN/WAGON, S-CLASS
E320 4MATIC (WAGON), S500 (GUARD)
GALANT
LANCER SPORTBACK
ALTIMA
SENTRA
VUE AWD
CAMRY
PRIUS, RAV4 4WD
JETTA, JETTA WAGON, BEETLE CONVERT.
V70
0.1
0.2
0.3
0.3
0.1
0.2
0.2
0.2
0.1-0.3
0.1 -0.2
0-0.1
0.2
0.3
0.2-0.3
0.3
0.2-0.3
0.3
0.3
0.2
0.1 -0.2
0.3
0.3
0.2-0.3
0.2
0.3
0.2
0.2
0.2-0.3
HLDT/MDPVs
BENTLEY
BMW
CHEVROLET
CHEVROLET
GMC
HIREUS
MERCEDES-BENZ
PORSCHE
ROLLS-ROYCE
VOLKSWAGEN
VOLVO
CONTINENTAL GT
X5
ASTRO AWD(C) CONV
K15 SLV HYBRID 4WD
K1500 SIERRA AWD
RR01
G500, ML350
CAYENNE, CAYENNE S
PHANTOM
TOUAREG
XC90
0.3
0.3
0.5
0.4
0.4
0.3
0.4
0.3
0.3
0.4
0.3,0.5
                               5-13

-------
     Table 5.1-4.  2005 model year vehicles with certification data
for total hydrocarbons at or below the proposed cold NMHC standard
MANUFACTURER
CAR LINE
COLD TOTAL HC LEVEL
LDV/LLDTs
ACURA
ACURA
AUDI
BMW
BMW
BUICK
CADILLAC
HONDA
HONDA
HONDA
HONDA
HONDA
HYUNDAI
HYUNDAI
HYUNDAI
MAZDA
MAZDA
MERCEDES-BENZ
MERCEDES-BENZ
MERCEDES-BENZ
MITSUBISHI
MITSUBISHI
NISSAN
SATURN
SATURN
TOYOTA
TOYOTA
VOLKSWAGEN
1.7EL, MDX4WD
RL, RSX
A4 QUATTRO
3251 SPORT WAGON, 330CI CONVERTIBLE
X3
LACROSSE/ALLURE
CTS
ACCORD
ACCORD HYBRID
CIVIC
CIVIC HYBRID
CR-V 4WD, ODYSSEY 2WD, S2000
JM(2WD)
JM(4WD)
XD-5DR
MAZDA 3
MPV
C240 (WAGON), C32 AMG, E320 4MATIC (WAGON), S55 AMG
C320
S430 4MATIC
GAL ANT
LANCER, LANCER SPORTBACK
SENTRA
RELAY AWD
VUE AWD
CAMRY, SCION XB
PRIUS, RAV4 4WD
JETTA, JETTA WAGON, BEETLE CONVERT., V70
0.1
0.2
0.3
0.1
0.2
0.3
0.2
0.1 -0.2
0.2
0.1 -0.2
0-0.1
0.2
0.3
0.2
0.3
0.2-0.3
0.2
0.3
0.2
0.1
0.2-0.3
0.3
0.2
0.3
0.2
0.3
0.2
0.2
HLDT/MDPVs
BENTLEY
BMW
CHEVROLET
CHEVROLET
GMC
LAND ROVER LTD
LEXUS
MERCEDES-BENZ
MERCEDES-BENZ
PORSCHE
ROLLS-ROYCE
TOYOTA
VOLVO
CONTINENTAL GT
X5
ASTRO AWD(C) CONV, C2500 SLVRADO 2WD, K1500 SUB'N 4WD
K15SLV HYBRID 4WD
G3500 SAVANA(P), K1500 SIERRA AWD
LR3
GX470
G500, ML350
G55 AMG
CAYENNE
PHANTOM
TOYOTA TUNDRA 4WD
XC90
0.3
0.3
0.5
0.4
0.4
0.4
0.4
0.4
0.2
0.3
0.3
0.5
0.3
                               5-14

-------
     Table 5.1-5.  2006 model year vehicles with certification data
for total hydrocarbons at or below the proposed cold NMHC standard
MANUFACTURER
CAR LINE
COLD TOTAL HC LEVEL
LDV/LLDTs
ACURA
ACURA
AUDI
BUICK
CADILLAC
CHEVROLET
CHRYSLER
HONDA
HONDA
HONDA
HONDA
HONDA
HYUNDAI
HYUNDAI
LEXUS
MAZDA
MAZDA
MERCEDES-BENZ
MERCEDES-BENZ
MERCEDES-BENZ
MITSUBISHI
MITSUBISHI
NISSAN
SATURN
SATURN
SUZUKI
TOYOTA
VOLKSWAGEN
VOLKSWAGEN
VOLVO
MDX 4WD
RL, RSX
A4 QUATTRO
LACROSSE/ALLURE
CTS
COBALT, IMP ALA
TOWN & COUNTRY 2WD
ACCORD
CIVIC, CR-V 4WD, ODYSSEY 2WD
CIVIC HYBRID
INSIGHT
S2000
JM(2WD), XD-4DR/5DR
JM(4WD)
GS 300 4WD, RX 400H 4WD
MAZDA 3, MAZDA 5, MPV
MAZDA 6, MAZDA 6 SPORT WAGON
B200 TURBO, S350
S430 4MATIC
S55 AMG
GAL ANT
LANCER, LANCER SPORTBACK
ALTIMA, SENTRA
RELAY AWD
VUE AWD
FORENZA WAGON
CAMRY, CAMRY SOLARA, YARIS
JETTA WAGON
PASSAT WAGON
V70
0.1
0.2
0.3
0.3
0.3
0.3
0.3
0.1 -0.2
0.2
0.1
0-0.1
0.3
0.3
0.2
0.3
0.2
0.3
0.2
0.1
0.3
0.2-0.3
0.3
0.3
0.3
0.2
0
0.3
0.2
0.3
0.2
HLDT/MDPVs
CADILLAC
CHEVROLET
CHEVROLET
DODGE
GMC
GMC
HONDA
JEEP
LAND ROVER LTD
LEXUS
LEXUS
MERCEDES-BENZ
FUNERAL COACH/HEARS, SRX AWD
C2500 SLVRADO 2WD
K15SLV HYBRID 4WD
DAKOTA PICKUP 4WD, RAM 1500 PICKUP 2WD
ENVOY XUV 4WD, G1525 SAVANA CONV
Kl 5 YUKON XL AWD
RIDGELINE 4WD
GRAND CHEROKEE 4WD
LR3
GX470
LX470
R500
0.5
0.5
0.3
0.5
0.5
0.3
0.2
0.4
0.5
0.4
0.5
0.2
                               5-15

-------
PORSCHE
PORSCHE
ROLLS-ROYCE
TOYOTA
VOLKSWAGEN
VOLVO
CAYENNE, CAYENNE S
CAYENNE TURBO KIT
PHANTOM
TOYOTA TUNDRA 4WD
PHAETON
XC90
0.3
0.5
0.3
0.5
0.5
0.3
5.1.2.2
EPA Test Program
       To determine the feasibility of meeting the proposed NMHC standard with only changes
to the calibration, EPA performed a test program involving a Tier 2 vehicle that was deemed
very challenging. The vehicle selection criteria for a feasibility study include several key
aspects. First, the vehicle needs to currently produce 20° F NMHC levels that are significantly
higher than the industry average. Second, since vehicle weight was determined to be a potential
disadvantage, a heavier GVWR vehicle is preferable for feasibility testing. Finally, the
technological approach chosen by the manufacturer to meet stringent 75° F Tier 2 standards was
also considered. Specifications for the test vehicle are included in Table 5.1-6.

                          Table 5.1-6. EPA Test Vehicle Specifications
Vehicle
2004
Chevrolet
Trailblazer
Engine Family
4GMXT04.2185
Powertrain
4.2L 16
4-speed auto
2-WD
GVWR
5550 Ibs.
Emission
Class
Tier 2 Bin 5
Mileage
36,500
       The vehicle was tested at 20° F following EPA cold FTP test procedures established in 40
CFR 86.230-94. In addition to regulated pollutant measurements, additional measurements
included NMHC, oxides of nitrogen (NOx), and particulate matter (PM). NMOG analysis also
produced measurements of 13 carbonyls.  PM measurement was performed following 40 CFR
86.110-94 procedures.  A detailed diagram of the emission and PM sampling system can be seen
in the docket.A The road load force target coefficient settings, contained in Table 5.1-7, are 10%
higher than the vehicle's 75° F target coefficients as established procedure in EPA guidance
letter CD-93-01.B
                        Table 5.1-7. EPA 20° F Cold Test Vehicle Settings
Vehicle
2004 Chevrolet
Trailblazer
Test Weight
5000 Ibs.
20° F Target
Coefficients
A=38.97
B=1.2526
C=.02769
A "Cold Chamber Sampling System Diagram," PDF file from test lab.
B Available at www.epa.gov/otaq/cert/dearmfr/dearmfr.htm.
                                           5-16

-------
5.1.2.2.1      2004 Chevrolet Trailblazer Feasibility

       As indicated earlier, the selection criteria of the vehicle candidate for the feasibility study
were designed to meet several key goals. The 2004 Chevrolet Trailblazer was chosen as a
candidate because it met the desired criteria.  First, it is certified as a Tier 2 Bin 5 package, which
represents what can be considered the "typical" or average 75° F emission level once Tier 2
phase-in is complete. This is because the Bin 5 emission standards represent the required EPA
fleet average for NOx and therefore the hardware used on the Trailblazer to comply with Bin 5
standards represents what we might expect from many manufacturers and vehicle lines. Second,
while it was certified to the expected average Tier 2 emission levels, its NMHC emission
performance at 20° F was substantially worse than the industry averages. Finally, due to its
GVWR, it represents vehicles that are very close to 6000 Ibs. GVWR. Different Trailblazer
models fall above and below 6000 Ibs. GVWR, but do not have any discernable differences in
the emission control hardware.

       The Trailblazer engine control system  is representative of typical Tier 2 systems. The
system includes an electronic engine control module (ECM), individual cylinder fuel injectors,
individual cylinder ignition coils, heated exhaust gas oxygen sensors (FtEGO) before and after
the catalyst, electronic throttle control, variable valve timing and several other necessary
supporting sensors. The aftertreatment hardware consists of a single, under-floor catalyst and a
secondary air injection system.

       The secondary air injection system is composed of an electric air pump and an electric
solenoid valve. The air pump is located under the vehicle's  driver-side floor board where it is
mounted to a frame bracket. The electric solenoid valve is mounted to the engine cylinder head
directly above the exhaust manifold on the passenger side of the vehicle. Clean air is drawn by
the air pump from the air cleaner assembly in the engine compartment through a pipe, then it is
pumped back to the electric solenoid valve through a second pipe.  The two pipes used to
transport the air are fairly long, due primarily to the air pump location.

       The secondary air injection system on the Trailblazer appears to operate on cold starts
above 40° F only. The system operates for approximately 20 to 45 seconds after the start,
depending on start-up coolant temperature, and is deactivated when the emission control system
goes into closed loop operation. Some manufacturers have  indicated that operation of the
secondary air injection system is not currently performed at and below freezing cold start
temperatures due to potential water freezing in the system which would prevent proper system
operation. This is, however, not universal across all manufacturers, since several manufacturers
do, in fact, operate their secondary air injection system at 20° F cold start temperatures and
above.  They have addressed the issue of water collecting and freezing by design aspects
primarily concentrated around system plumbing and location of the components. On some
European vehicle models, these manufacturers effectively use the secondary air injection systems
to comply with a 20° F NMHC standard in Europe.4

       A key element of the feasibility test program was to imitate emission control system
behaviors observed at the currently regulated start temperatures of 75° F and 50° F (California-
only requirement).  In the case of the Trailblazer, while not all behaviors could be demonstrated,
                                           5-17

-------
several of the most important behaviors were replicated.  First, the operation of the secondary air
injection system was determined to be a requirement. Second, elevated idle speeds, similar to
what the Trailblazer currently uses after the start at the regulated start temperatures, were also
required.

       The activation of the secondary air injection was accomplished through circuit overrides
of the air pump and solenoid valve control circuits, completely external to the ECM. The air
pump and the solenoid valve are each powered by a relay normally only controlled by the ECM
output signals.  The two relays were forced on to activate the secondary air injection system
during the desired period following the cold start. Several delay periods from the start of the
engine until the secondary air system was activated were tested to measure benefits of earlier
introduction of the air injection. The secondary air was always run until ECM induced closed
loop operation (approximately 60 seconds after the start). At the completion of the desired period
of operation, control of the relays was  returned to the ECM.

       The elevated idle speed was performed by allowing a manually  controlled vacuum leak
into the intake manifold during the first 30 to 60 seconds following engine start. The controlled
vacuum leak targeted  1550 to!600 RPM idle speed in park/neutral, mimicking the same desired
idle speed the ECM commands at 50° F cold starts. Typically, idle speeds increase with drops in
start temperature, but the observed desired idle speeds in the Trailblazer were lower at 20° F
(1350 RPM) than at the warmer 50° F  starts (1550 RPM). Ideally, utilizing the electronically
controlled throttle to achieve a target idle speed would have been the best method, but control of
the  electronic throttle was not available.  Manufacturers today control to a desired idle speed
through control of electronic throttle or other air bleed devices.

       Table 5.1-8 below contains the weighted test total (3 bags) emission results of the
different test configurations attempted on the Trailblazer. Test #7  and #8 also included defroster
operation starting at 130 seconds into the test and remaining on for the rest  of the test. Since the
methods used to control cold start NMHC emissions were used only in  the first minute of
operation, prior to defroster activation, the NMHC and PM emission results with defroster
operation remain representative of emission control opportunities.  It is important to note the
consistent reductions in NMHC with early activation of the secondary air injection system as
seen in the test sequence from test #3 through test #6, but also in the defroster tests. The tests
with defroster operation were included to assess any emission impacts of defroster-on, which is
being proposed in a fuel economy rule.0

       While NOx emissions are not part of the controls investigation,  the NOx levels appeared
to increase with the NMHC control methods. After some modal investigation, it was determined
that the NOx increases were occurring after the NMHC controls had performed the majority of
their benefits.  The NOx emissions were brought back almost to the baseline levels by shortening
the  elevated idle speed and air bleed time. This can be observed in the results of test #6 and #7.
In fact, test #6 produced the largest NMHC reduction with essentially the same NOx levels as the
baseline tests.  Manufacturers would be able to better synchronize their controls through their
ECM to control NMHC and NOx emissions simultaneously, as compared to this test program's
limitations.
 Fuel Economy Final Rule XX Defroster Operation Requirement for Cold FTP.


                                           5-18

-------
       CO and PM measurements also indicate significant reductions when NMHC controls are
activated. CO, the only currently regulated pollutant at 20° F, demonstrated consistent
reductions over baseline levels with each of the control combinations.  PM generally also
indicated reductions; however, it is less obvious when reported as test total results. Since the
emissions are recorded over the three-phase test with each phase composed of an individual bag
measurement, PM reductions can be better evaluated in Table 5.1-9, which contains the emission
results for only the first phase (bag 1) of the three-phase emission test.

         Table 5.1-8. Trailblazer Test Configuration and 20° F FTP Weighted Test Total
                                         Results
Test
Number
Air
Injection
Elevated Idle &
air bleed time
Proposed Standard < 6000 Ibs GVWR
Proposed Standard > 6000 Ibs GVWR

1 -baseline
2-baseline

3 -controls
4-controls
5 -controls
6-controls

7-defrost on
8-defrost on

none
none

5 s delay
2 s delay
1 s delay
0 s delay

1 s delay
0 s delay

none
none

60s
60s
60s
30s

30s
45 s
NMHC
g/mi
.3
.5

1.08
1.03

.59
.42
.35
.29

.38
.32
CO
g/mi



7.8
9.5

5.2
5.5
5.2
5.1

6.9
6.4
NOx
g/mi



.05
.04

.15
.19
.17
.06

.08
.13
PM
g/mi



.024
.015

.025
.013
.014
.013

.012
.013
Fuel Economy
mi/gallon



13.82
13.64

13.87
13.56
13.71
13.64

13.17
13.25
       As can be seen in Table 5.1-8, control test #6 provided a NMHC level that would have
allowed the Trailblazer to comply with the proposed standard for the < 6000 Ibs GVWR class
(i.e., 0.3g/mi). While this vehicle was tested as the lower GVWR class at 5000 Ibs test weight,
the Trailblazer also is sold as an over 6000 Ibs. GVWR model that would have been tested at
5500 Ibs. We believe that if tested at the higher weight, the emission results likely would not
have increased much, reflecting a large margin (.2 g/mi) for this vehicle when certified to the
heavier class. We recognize that manufacturers will need to account for a compliance margin,
but we believe this vehicle can achieve a comfortable compliance margin for the more stringent
standard (i.e., 0.3g/mi) with some additional minor calibration changes.

       While emissions results for the 20° F cold CO test are reported as a weighted three-bag
average, bag one (the first 505 seconds of the test) provides a better indication of emission
reductions achieved with controls. Since almost all of the emissions at 20° F are emitted in the
first few minutes of operation and all control changes were attempted only during the first
minute of operation, Table 5.1-9 presents only the bag 1 emission results. This table highlights
the emission reductions from the control changes by not diluting the improvements over the
second and third phase (bag 2 and 3) of the emission test.
                                          5-19

-------
       As observed below in Table 5.1-9, the level of reductions in emissions with the different
control changes are more obvious as measured in the first phase of the test. NMHC, CO and PM
reductions can be clearly seen from the results.  NMHC and CO reductions occur with all the
control attempts but achieve the best results with control test #6 and #8, in which secondary air
injection was activated immediately upon engine cranking. PM reductions also follow similar
behavior as NMHC but appear to be very sensitive to delayed secondary air injection.

        Table 5.1-9. Trailblazer Test Configuration and 20° F FTP Phase 1 Only Results
Test
Number
1 -baseline
2-baseline

3 -controls
4-controls
5 -controls
6-controls

7-defrost on
8-defrost on
Air
Injection
none
none

5 s delay
2 s delay
1 s delay
0 s delay

1 s delay
0 s delay
Elevated Idle &
air bleed time
none
none

60s
60s
60s
30s

30s
45 s
NMHC
g/mi
5.18
4.92

2.81
1.96
1.63
1.34

1.75
1.47
CO
g/mi
27.3
31.7

18.6
15.0
13.6
13.3

14.8
13.2
NOx
g/mi
.22
.16

.72
.85
.81
.29

.35
.61
PM
g/mi
.055
.040

.043
.033
.026
.022

.010
.022
Fuel Economy
mi/gallon
11.55
11.47

11.29
11.30
11.40
11.45

11.23
11.27
       While the emissions reductions were fairly substantial with the best control combination
in test #6, we believe that even greater emission reductions can be achieved with more precise
use of the secondary air system and additional control measures described earlier in the
calibration and controls technology section. The ability to more precisely provide the ideal air-
fuel mixture for the secondary air injection system likely would have resulted in faster catalyst
light-off and subsequently even greater reductions in emissions, especially NMHC.
Additionally, retarded timing was not tested due to the limited capability to modify engine
operation. Typically this would further compound the rate of heating the catalyst, particularly on
secondary air injection systems, and thus, would be expected as an additional opportunity to
reduce NMHC.
5.1.2.2.2
Additional Tier 2 Vehicle Feasibility
       We are entertaining expanding the feasibility testing to additional Tier 2 vehicles utilizing
the technologies described earlier in the calibration and controls technology section.  Any
additional studies are contingent on our ability to access and modify these emission control
technologies in the time window of this rulemaking.

5.2    Feasibility of Evaporative Emissions Standards for Vehicles

       The proposed standards for evaporative emissions, which are equivalent to the California
LEV II standards, are technologically feasible now.  As discussed in Section VI of the preamble
for today's proposed rulemaking, the California LEV II program contains numerically more
stringent evaporative emissions standards compared to existing EPA Tier 2 standards, but
                                          5-20

-------
because of differences in testing requirements, some manufacturers view the programs as similar
in stringency.  See Section VI.B.2.C of today's proposed rule for further discussion of such test
differences (e.g., test temperatures and fuel volatilities). Thus, some manufacturers have
indicated that they will produce 50-state evaporative systems that meet both sets of standards
(manufacturers sent letters indicating this to EPA in 2000).5' 6' 7 In addition, a review of recent
model year certification results indicates that essentially all manufacturers certify 50-state
evaporative emission systems.8 Therefore, harmonizing with California's LEV-II evaporative
emission standards would streamline certification and be an "anti-backsliding" measure - that is,
it would prevent future backsliding as manufacturers pursue cost reductions.  It also would
codify the approach manufacturers have already indicated they are taking for 50-state
evaporative systems.
                                           5-21

-------
References for Chapter 5
1 Memo to docket "Discussions Regarding Secondary Air System Usage at 20° F with European
  Automotive Manufacturers and Suppliers of Secondary Air Systems," December 2005.

2 Meyer, Robert and John B. Heywood, "Liquid Fuel Transport Mechanisms into the Cylinder of
  a Firing Port-Injected SI Engine During Start-up," SAE 970865, 1997.

3 For a more detailed description of these technologies see the Tier 2 final rule at 65 FR 6698-
6822, February 10,2000, and Regulatory Impact Analysis Chapter IV: Technical Feasibility.

4 Memo to docket "Discussions Regarding Secondary Air System Usage at 20° F with European
  Automotive Manufacturers and Suppliers of Secondary Air Systems,"  December 2005.

5 DaimlerChrysler, Letter from Reginald R. Modlin to Margo Oge of U.S. EPA, May 30, 2000.
  A copy of this letter can be found in Docket No. EPA-HQ-OAR-2005-0036.

6 Ford, Letter from Kelly M. Brown to Margo Oge of U.S. EPA, May 26, 2000. A copy of this
  letter can be found in Docket No. EPA-HQ-OAR-2005-0036.

7 General Motors, Letter from Samuel A. Leonard to Margo Oge of U.S. EPA, May 30, 2000. A
  copy of this letter can be found in Docket No. EPA-HQ-OAR-2005-0036.

8 U.S. EPA, Evaporative Emission Certification Results for Model Years 2004 to 2006,
  Memorandum to Docket EPA-HQ-OAR-2005-0036 from Bryan Manning, February 9, 2006.
                                        5-22

-------
                             Chapter 6: Table of Contents

Chapter 6: Feasibility of Complying with a Benzene and Other Control Standards	3
   6.1  Overview of Refinery Flow	3
   6.2  What are the Benzene Levels in Gasoline Today?	6
   6.3  Where does Gasoline Benzene Come from?	16
     6.3.1     How do Reformers work?	17
     6.3.2     How can Benzene Levels be Reduced in Gasoline?	22
        6.3.2.1    Prefractionation to Reroute Benzene Precursors	22
        6.3.2.2    Benzene Saturation via Isomerization	23
        6.3.2.3    Reformate Postfractionation with Benzene Saturation	25
        6.3.2.4    Benzene Extraction	27
        6.3.2.5    Low-Pressure Reformer Operation	29
        6.3.2.6    Prefractionation Combined with Low-Pressure Reformer Operation	29
   6.4  Experience Using Benzene Control Technologies	29
     6.4.1     Benzene Levels Achievable through Reformate Benzene Control	30
     6.4.2     Other Benzene Controls	33
   6.5  Averaging, Banking, and Trading (ABT) Program	36
     6.5.1     Starting Gasoline Benzene Levels	36
     6.5.2     Refinery Compliance Strategies	37
     6.5.3     Benzene Reduction Strategies	38
        6.5.3.1   Early Process Changes Completed Prior to January 1,2011	38
        6.5.3.2   Final Process Changes Requiring a Large Capital Investment	49
     6.5.4     Ending Gasoline Benzene Levels	50
     6.5.5     Standard Credit Generation Opportunities	50
        6.5.5.1   How are Standard Credits Calculated?	51
        6.5.5.2   How Many Standard Credits would be Generated in 2011 and Beyond?	51
     6.5.6     Credit Use	52
        6.5.6.1 Credit Trading Area	52
        6.5.6.2   Credit Life	53
        6.5.6.3   Credit Availability	54
        6.5.6.4   Credit Value	56
   6.6  Feasibility for Recovering Octane	57
   6.7  Will the Proposed Benzene Standard Result in Any New Challenges to the Fuel
         Distribution System or End-Users?	60
   6.8  Impacts on the Engineering and Construction Industry	61
     6.8.1     Design and Construction Resources Related to Benzene  Reduction Equipment61
     6.8.2     Number and Timing of Benzene Reduction Units	62
     6.8.3     Timing of Projects Starting Up in the Same Year	63
     6.8.4     Timing of Design and Construction Resources Within a  Project	63
     6.8.5     Projected Levels of Design and Construction Resources	64
   6.9  Time Needed to Comply with a Benzene Standard	68
   6.10 Will the Proposed Fuel Standard Be More Protective Than Current Programs in All
         Areas?	70
     6.10.1     Modeling  Approach	71
        6.10.1.1      Choice of Analysis Cases and Data Sources	72

                                          6-1

-------
        6.10.1.2     Adjustment of Fuel Parameters for Future Years	74
        6.10.1.3     Conversion of Production Properties to In-Use Properties	77
        6.10.1.4     Running the MOBILE Model	81
     6.10.2    Interpretation of Results	84
     6.10.3    Conclusions	85
   6.11  Feasibility for Lower RVP	85
     6.11.1    Means for Reducing RVP	86
   6.12  Feasibility of Removing Sulfur from Gasoline	91
     6.12.1    Source of Gasoline Sulfur	91
     6.12.2    Complying with the Tier 2 Gasoline Sulfur Standard	92
        6.12.2.1     Background	92
        6.12.2.2     FCC Feed Hydrotreating	93
        6.12.2.3     FCC Naphtha Hydrotreating	94
        6.12.2.4     FCC Naphtha Desulfurization Technologies	94
     6.12.3    Meeting a 10 ppm Gasoline Sulfur Standard	98
        6.12.3.1     Feasibility of Meeting a 10 ppm Gasoline Sulfur Standard	101
Appendix 6A: Additional Background on Refining and Gasoline	102
   6A.1   Petroleum Refining	102
   6A.2   Crude Oil	103
     6A.2.1    Crude Desalting	104
     6A.2.2    Atmospheric Crude Unit	105
     6A.2.3    Preflash	105
     6A.2.4    Crude Unit	106
     6A.2.5    Atmospheric Tower Gasoil and Residuum; Vacuum Unit	107
     6A.2.6    Naphtha Splitter	110
     6A.2.7    Hydrotreating	110
     6A.2.8    Fluid Catalytic Cracker	Ill
     6A.2.9    Alkylation	113
     6A.2.10     Thermal Processing	113
   6A.3   Gasoline	113
     6A.3.1    Gasoline as a Complex Mixture	114
     6A.3.2    Octane	117
   6A.4   Kerosene and Diesel	120
                                         6-2

-------
    Chapter 6: Feasibility of Complying with a Benzene and Other
                              Control Standards

       This chapter summarizes our assessment of the feasibility of complying with a benzene
control standard.  It begins with an overview of refining followed by a summary of the benzene
levels of gasoline today and where that benzene comes from.  The various technologies which
reduce benzene levels in gasoline are described along with an assessment of the levels of
benzene achievable by the application of these technologies and their potential to be applied by
refineries.  This assessment of the benzene levels achieved by applying control technologies is
used to assess the feasibility of complying with the proposed benzene standard. Next the lead
time to apply the various control technologies and to comply with the proposed standard is
evaluated.  Finally, the energy and supply impacts of the proposed rule are assessed.

6.1    Overview of Refinery Flow

       Figure 6.1-1 shows a process flow diagram for a typical complex refinery, capable of
making a wide product slate (shown on the right side of the figure) from crude oil (input on the
left). Following the figure is a brief description of key units and streams focusing more on the
gasoline producing units.
                                         6-3

-------
           Figure 6.1-1. Process Flow Diagram for a Typical Complex Refinery
    Natural
      Gas
* Fuel Gas
                                                                           LPG
                                                                           Gasoline
                                                                           Aromatics
                                                                           Kerosine
                                                                         ,. Jet Fuel
                                                                          On-Highway
                                                                          Diesel
                                                                         Off-roadDiesel
                                                                        * Fuel Oil
                                                                         *• Resid
                   Vacuum Tower
                                                                           Coke
                                                               Coker
       Crude Tower

       The purpose of the crude tower is to perform a distillation separation of crude oil into
different streams for additional processing in the refinery and for the production of specific
products.  Crude oil is shipped to the refinery via pipeline, ship, barge, rail, or truck, whereupon
it is sampled, tested, and approved for processing.  The crude oil is heated to between 650° F and
700° F and fed to crude distillation tower. Crude components vaporize and flow upward through
the tower. Draw trays are installed at specific locations up the tower from which desired side
cuts or fractions are withdrawn. The first side-cut above the flash zone is usually atmospheric
gasoil (AGO), then diesel and kerosene/jet fuel are the next side-cuts, in that order The lightest
components, referred to here as straight run naphtha, remain in the vapor phase until they exit the
tower overhead, following which they are condensed and cooled and sent to the naphtha
splitter.l

       Naphtha Splitter

       The purpose of the naphtha splitter is to perform a distillation separation of straight run
naphtha into light straight run naphtha and heavy straight run naphtha. The feed  can be split
between the Cs's and Ce's in order to assure the Ce's and heavier were fed to the reformer.2
                                           6-4

-------
       Isomerization Unit

       The purpose for the isomerization unit is to convert the light naphtha from straight chain
hydrocarbons to branched chain hydrocarbons, increasing the octane of this stream. The
isomerate is sent to gasoline blending.3

       Reformer

       The purpose of the reformer unit is to convert C6 to C8 or C9 hydrocarbons into aromatic
and other higher octane compounds (benzene is one of the aromatic compounds produced).
Heavy straight run naphtha is hydrotreated and fed to the reformer. As the reformer converts the
feed hydrocarbons to aromatics, hydrogen and light gases are produced as byproducts. The liquid
product, known as reformate, is sent directly to gasoline blending, or to aromatics extraction.4

       Aromatics Extraction Unit

       The purpose of aromatics extraction is to separate the aromatic compounds from the rest
of the hydrocarbons in reformate using chemical extraction with a solvent to concentrate the
individual aromatic compounds, (mainly xylene and benzene) for sale to the chemicals market.5

       Vacuum Tower

       The purpose of the vacuum distillation tower unit is to enable a refinery to produce more
gasoline and diesel fuel out of a barrel of crude oil.  It separate the heavy vacuum gasoil
(HVGO), which is fed to the FCC unit, from the vacuum tower bottoms (VTB) which is sent to
the coker, or in other refineries is  made into asphalt.

       Fluidized Catalytic Cracker

       The purpose of the fluidized catalytic cracker is to convert heavy hydrocarbons, which
have very low value,  to higher value lighter hydrocarbons.  AGO and HVGO are the usual feeds
to a fluid catalytic cracker (FCC). The full boiling range cracked product leaves the reactor and
is sent to a fractionator. The overhead includes propane, propylene, butane, butylene, fuel gas
and FCC naphtha, which contains some benzene. There are two heavy streams; light cycle oil
(LCO), which can be hydrotreated and blended into diesel fuel or hydrocracked into gasoline;
and heavy cycle oil, sometimes called slurry oil, which can be used for refinery fuel.6

       Gas Plant

       The purpose of the gas plant is to use a series of distillation towers to separate various
light hydrocarbons for further processing in the alkylation or polymerization units or for sale.

       Alkylation Unit

       The purpose of the alkylation unit is to chemically react light hydrocarbons together to
produce a high quality, heavy gasoline product. Alkylation uses sulfuric or hydrofluoric acid as

                                           6-5

-------
catalysts to react butylene or propylene together with isobutane.  Following the main reaction
and product separation, the finished alkylate is sent to gasoline blending.  Alkylate is low in RVP
and high in octane.7

       Polymerization Unit

       The purpose of the polymerization unit is to react light hydrocarbons together to form a
gasoline blendstock.  A polymerization unit, often referred to as a "cat poly" is somewhat similar
to an alkylation unit, in that both use light olefms to produce gasoline blendstocks.  The feed is
generally propylene and/or butylene from the gas plant. The product, called polygas is sent to
gasoline blending.

       Coker Unit

       The purpose of the coker unit is to process vacuum tower bottoms (VTB) to coke and to
crack a portion to various lighter hydrocarbons. The hydrocarbons produced by the coker
include cracked gases, coker naphtha, coker distillate and gas oil. The gas is fed to the gas  plant,
the naphtha to the reformer hydrotreater,  and the distillate either to distillate hydrotreating or to
the hydrocracker.

       Hydrocracker

       The purpose of the hydrocracker is to crack and "upgrade" the feedstock into higher
value products.  The feedstock to the hydrocracker is usually light cycle oil (LCO) and coker
distillate, poor quality distillate blendstocks, which are upgraded to diesel fuel, or cracked to
gasoline. Heavier hydrocarbons such as AGO  and HVGO can be feedstocks as well.

       A more complete description for reforming is contained in Section 6.3. Other refinery
units are described in more detail in the Appendix.

6.2    What are the Benzene Levels in Gasoline Today?

       EPA receives information on gasoline quality, including benzene, from each refinery in
the U.S. under the reporting requirements of the Reformulated Gasoline and Antidumping
Programs. Benzene levels averaged 0.94 volume percent for gasoline produced in and imported
into the U.S. in 2003, which is the most recent  year for which complete data was available at the
time of this analysis. The benzene levels differ depending on different volumes of interest.  We
assessed the 2003 benzene levels by conventional versus reformulated gasoline, winter versus
summer, and with and without California and Imports. Table 6.2-1 contains the benzene levels
for these various gasoline types by season and  aggregated.
                                           6-6

-------
   Table 6.2-1.  Summary of U.S. Benzene Levels by Gasoline Type and Season for 2003

CG Summer
CG Winter
Total CG
% by total volume
RFC Summer
RFC Winter
Total RFC
% by total volume
Summer CG &RFG Avg.
Winter CG & RFC Avg.
CG & RFC Avg.
% by total volume
U.S. Production
(excluding CA)
1.129
1.086
1.107
65
0.598
0.637
0.620
20
1.009
0.966
0.991
85
Imports
1.022
0.826
0.914
2
0.682
0.715
0.701
2
0.850
0.768
0.804
4
Production plus
Imports
1.126
1.078
1.101
67
0.605
0.645
0.627
22
1.002
0.965
0.982
89
CA



0
0.620
0.620
0.620
11
0.62
0.62
11
Production plus
Imports Plus CA
1.126
1.078
1.101
67
0.610
0.636
0.625
33
0.965
0.923
0.942
100
       Individual refinery gasoline benzene levels can vary significantly from the national
average with trends forming in specific regions of the country. Therefore, it is useful to
understand how the benzene levels vary by individual refinery as well as regionally.  Figure 6.2-
1 contains a summary of annual average gasoline benzene levels by individual refinery for
conventional gasoline and reformulated gasoline versus the cumulative volume of gasoline
produced (not including California refineries for which EPA does not receive data).
                                           6-7

-------
            Figure 6.2-1. Benzene Content of RFG and Conventional Gasoline.
             10      20      30      40      50      60     70      80
                        Cumulative Gasoline Supplied (Billion Gallons)
90
100
       Figure 6.2-1 shows that the annual average benzene levels of conventional gasoline
produced by individual refineries varies from 0.3 to 3.5 volume percent. The volume-weighted
average is 1.10 volume percent. As expected, the annual average benzene levels of reformulated
gasoline as produced by individual refineries are lower ranging from 0.1 to 1.0 volume percent.
The volume-weighted average benzene content for U.S. reformulated gasoline (not including
California) is 0.62 volume percent.

       The information presented for annual average gasoline benzene levels does not indicate
the variability in gasoline batches produced by each refinery. We also evaluated the batch-by-
batch gasoline benzene levels for individual  refineries.  This information is obtainable from data
provided to EPA under the reporting requirements of the RFG program. To illustrate the degree
of variability within different refineries, in Figure 6.1-2 through 6.2-7 we provide the data for 3
different refineries which produce both conventional  and reformulated gasoline and 3 refineries
which produce solely conventional gasoline.  For the RFG producing refineries we summarize
the data by gasoline type as these refineries produce both RFG and CG.  For the CG refineries
we break out the data by premium grade, regular grade and midgrade gasoline, if the refinery
produces it.  We arbitrarily labeled the refineries in these figures refineries A through F to
facilitate the discussion about this data.

-------
     Figure 6.2-2. RFG and CG Batch-by-Batch Benzene Levels for Refinery "A'
     	(volume percent benzene in 2003 gasoline)	
  2.5
  2.0
2 1.5
  1.C
  o.e
  o.o
   Jan-03
                                                                    CG   X RFG
X X ^ • x Xx   ,XX
               Apr-03
 Jun-03
Batch Date
Sep-03
Dec-03
                                         6-9

-------
      Figure 6.2-3.  RFG and CG Batch-by-Batch Benzene Levels for Refinery "Br
     	(volume percent benzene in 2003 gasoline)	
   3.5
   3.0
   2.5
0)
   2.0
0)
.a
S  1.5
                                                                    CG   X RFG
    Jan-03
Apr-03
 Jun-03
Batch Date
Sep-03
Dec-03
                                        6-10

-------
 Figure 6.2-4. Batch-by-Batch Benzene Levels for Refinery "C" that Produces both RFG
	and CG Gasoline (volume percent benzene in 2003 gasoline)	
   2.5
   2.C
 2 1.5
   1.0
   o.e
   0.0
    Jan-03
                                                          CG
                                             XRFG
Apr-03
                                               X^ffi^^>M«fc!
 Jun-03
Batch Date
Sep-03
Dec-03
                                      6-11

-------
 Figure 6.2-5. Premium and Regular Grade Gasoline Batch-by-Batch Benzene Levels for
	Refinery "D" (volume percent benzene in 2003 gasoline)	
   3.0
   2.5
   2.0
nzene
 J| 1.5
   1.0
   0.0
    Jan-03
                                                 x Premium Grade    • Regular Grade
Apr-03
 Jun-03
Batch Date
Sep-03
Dec-03
                                         6-12

-------
Figure 6.2-6. Premium, Midgrade and Regular Grade Batch-by-Batch Benzene Levels for
               Refinery "E" (volume percent benzene in 2003 gasoline)

c
A
enzene
0 4
& J <
ss
0
o
1


• Regular Grade A Mid-Grade x Premium Grade

*
A A A A
A A A
A A A A
A A* * A* A^ A^
* "V* A « * A
' V '. . /* .- ^AA ^ %'*.AA •
• •.•-.' A-"- •*_£. -* A/ 'AV ^ «
' '.. - • ' - .' ,-•. \ *V~ ." ' - ..
x x • ,
XvX ^.X XXX.
*x*xXx* xxx xx>
-------
 Figure 6.2-7. Premium and Regular Grade Gasoline Batch-by-Batch Benzene Levels for
 	Refinery "F" (volume percent benzene in 2003 gasoline)	
    1
    1.2
  0)

  I
  0)
  .a
                                                     x Premium Grade   • Regular Grade
     Jan-03
Apr-03
 Jun-03
Batch Date
Sep-03
Dec-03
       Most of the refineries that we studied produced substantially different batch-to-batch
benzene levels. As expected, the RFG batches were consistently lower than the CG batches.
Two of the RFG producing refineries had a wide variability in benzene levels.  The gasoline
batch benzene levels for refineries A and B varied by over an order of magnitude.  Refinery C's
gasoline batch benzene levels varied less than those of refinery A and B.  Most all of refinery
C's batches were under 0.5 volume percent benzene except for a very few which were much
higher. Also, refinery C's gasoline batches had similar benzene levels for both RFG and CG, a
very different trend than refineries A and B.

       Of the three CG refineries, refineries labeled E and F have widely varying gasoline batch
benzene levels. Refinery  E's gasoline batch benzene levels were consistently higher than the
rest, ranging from under 1 percent to over 4 percent.  Refinery F had no clear trend for either the
regular or premium grade of gasoline; the benzene levels varied for both by about an order of
magnitude. Refinery E did have an interesting trend for specific refinery grades. Premium grade
tended to have lower benzene levels than the other grades, midgrade had the  highest benzene
levels and regular grade's benzene levels were in between the other two grades. Evaluated all
together, the various grades of refinery E also varied by an order of magnitude.  The gasoline
                                          6-14

-------
batch benzene levels for refinery D were consistently under 0.5 volume percent for most of the
batches, although a very small fraction of the batches had much higher benzene levels. The
lower variability in refinery D's batches was similar for both premium and regular grades of
gasoline.

       There are several reasons for the variability in refinery gasoline benzene levels across all
the refineries.  First, crude oil varies greatly in aromatics content. Since benzene is an aromatic
compound, its concentration tends to vary consistent with the aromatics content of crude oil.  For
example Alaskan North Slope (ANS) crude oil contains a high percentage of aromatics. A
refiner processing ANS crude oil in their refineries shared with us that their straight run naphtha
off the atmospheric crude distillation column contains on the order of 3 volume percent benzene.
 This is one reason why the gasoline in PADD 5 outside of California is high in benzene.
Conversely, refiners with very paraffmic crude oils (low in aromatics) may have benzene levels
as low as 0.3 volume percent benzene in their straight run naphtha.

       The second reason why benzene levels vary is due to the types of units in  their refinery.
Different refinery streams contain widely different concentrations of benzene, with reformate
typically contributing the most.  If a refinery relies on the reformer for virtually all of their
octane needs, especially the type which operates at higher pressures and  temperatures that tends
to produce more benzene, they will likely have a high benzene level in their gasoline. Refineries
with a reformer and without an FCC unit are particularly prone to higher benzene levels.
However,  refineries which can rely on several different units or means for boosting their
gasoline octane can usually run their reformers at a lower severity resulting in less benzene in
their gasoline pool. Examples of octane-boosting refinery units include the alkylation unit, the
isomerization unit, and units which produce oxygenates. Refiners may have these units in their
refineries, or in many cases,  the gasoline blendstocks produced by these units can be purchased
from other refineries or third-party  producers. The blending of alkylate,  isomerate, and
oxygenates into the gasoline pool provides a significant octane contribution which would allow
refiners to rely less on the octane from reformate. The  variation in gasoline blendstock content
across different batches of gasoline is likely the reason for the drastically differing benzene
levels between batches of gasoline.

       Finally, many refiners may be operating their refinery today to intentionally have less
benzene in their gasoline.  They could be doing this by operating the refinery with that end in
mind such as for the Federal or California RFG programs.  Refiners which are currently
producing reformulated gasoline are targeting to reduce their gasoline benzene levels to less than
0.95 volume percent for the Federal RFG program or lower for the California RFG program, and
are using benzene control technologies to produce gasoline with lower benzene levels. If they
are producing  conventional gasoline along with the reformulated gasoline, their conventional
gasoline is usually lower in benzene as well compared with the conventional gasoline produced
by other refineries. Alternatively, some refiners add specific refinery units such as benzene
extraction which intentionally removes  benzene and concentrates it for the profit  it earns.  The
profit gained by extraction is due to the much higher price that benzene earns on the benzene
chemical market compared to the price of gasoline.  In  most cases, refineries with extraction
units are also marketing their low benzene  gasoline in the RFG areas.
                                           6-15

-------
       Table 6.2-2 shows the variations in gasoline benzene levels as produced by refineries in,
and as imported into, refining regions called Petroleum Administrative for Defense Districts
(PADD) for 2003.8  The information is presented for both conventional gasoline and
reformulated gasoline.
 Table 6.2-2. 2003 Benzene Levels by Gasoline Type and by PADD as Supplied in the U.S

Conventional
Gasoline
Reformulated
Gasoline
Gasoline
Average
PADD1
0.84
0.60
0.70
PADD 2
1.39
0.82
1.28
PADD 3
0.94
0.56
0.87
PADD 4
1.54
N/A
1.54
PADD 5
1.79
N/A
1.79
CA
0.63
0.62
0.62
U.S.
1.11
0.62
0.94
       Table 6.2-2 shows that benzene levels vary fairly widely across different regions of the
country. PADD 1 and 3 benzene levels are lower because the refineries in these regions produce
a high percentage of reformulated gasoline for both the Northeast and Gulf Coast.  About 60
percent of PADD 1's gasoline is reformulated, while 20 percent of PADD 3's gasoline is
reformulated. Reformulated gasoline must meet a 0.95 volume percent average benzene
standard, and a  1.3  volume percent cap standard.  Another reason why the benzene levels are so
low in these two regions is because 35 percent of the refineries in these two regions, are
extracting benzene for sale to the petrochemicals market. When refiners are extracting benzene
from their gasoline, they extract as much benzene as possible to take maximum advantage of the
expensive cost of capital associated with extraction units.  This is likely the  reason why the CG
in PADDs 1 and 3 is low in benzene  as well. In other parts of the U.S., where little to no
reformulated gasoline is being produced and little extraction exists, the benzene levels are much
higher.

6.3    Where does Gasoline Benzene Come from?

       The portion of the crude oil barrel which boils within the gasoline boiling range is called
naphtha.  There are two principal sources of naphtha.  The first principal source of naphtha is
straight run naphtha which comes directly off of the crude oil atmospheric tower. The second
principal source of naphtha is from the cracking reactions. Each type of naphtha provides a
source  of benzene to gasoline.

        Straight run naphtha which comes directly from the distillation of crude oil contains
anywhere from  0.3  to 3 volume percent benzene. While straight run naphtha is in the correct
distillation range to be usable as gasoline, its octane value is typically 70 octane numbers which
is too low for blending directly into gasoline. Thus, the octane value of this material must be
increased to enable it to be sold as gasoline.  The primary means for increasing the octane of
naphtha is reforming. In the process of increasing the octane of this straight run material, the
                                          6-16

-------
reformer increases the benzene content of this stream.

       There are two primary cracking processes in the refinery. One is called the fluidized
catalytic cracking (FCC) unit and the second is called hydrocracking. Other cracking units
include cokers and thermal crackers. These various cracked naphthas contain anywhere from 0.5
to 5 volume percent benzene.

       The attached table summarizes the range in benzene content and typical percentage of
gasoline of the various refinery intermediate streams used to blend up gasoline.
     Table 6.3-1. Benzene Content and Typical Gasoline Fraction of Various Gasoline
                                     Blendstocks.
Process or Blendstock Name
Reformate
FCC Naphtha
Alky late
Isomerate
Hydrocrackate
Butane
Light Straight Run
MTBE/Ethanol
Natural Gasoline
Coker Naphtha
Benzene Level
(volume percent)
3-11
0.5-2
0
0
1-5
0
0.3-3
0.05
0.3-3
3
Typical Volume in Gasoline
(percent)
30
36
12
4
3
4
4
3
3
1
       Table 6.3-1 shows that the principal contributor of benzene to gasoline is reformate. This
is due both to the high benzene content of reformate and the relatively large gasoline fraction
that it comprises of the gasoline pool.  For this reason, reducing the benzene in reformate is the
focus for the various benzene reduction technologies available to refiners.

6.3.1   How do Reformers work?

       Reformers have been the dominant gasoline high octane producing units since they first
came into operation in the 1940's.9  An indication of their importance in refining is that every
U.S. refinery except one has a reformer. Prior to the lead phase-down in the early 1980's
reformers operated at fairly moderate severities and produced product octane numbers around 85
RON (see the Appendix for a discussion of octane). After the phase-down and eventual phase-
out of lead from gasoline, and as the demand for high-octane premium fuel grew, octane
numbers for reformate increased to a range from a RON in the low 90s to 104. The reforming
process works by rearranging, e.g., "reforming" the chemical structure of straight-chain and
cycloparaffin molecules in a given feedstock, to produce a variety of high-octane benzene,
substituted aromatic, and isoparaffmic molecules.  The reforming process uses a combination of
heat, pressure, and catalyst, to produce high octane, high-value finished blendstocks from a low-
octane, (about 50 RON in some cases) low-value feedstock.
                                          6-17

-------
       Reformer Chemical Reactions

       The chief means by which reformers increase octane is through the formation of aromatic
compounds, including benzene. Aromatic compounds are distinguished from other hydrocarbon
compounds by their structure which cannot be described without at least a very rudimentary
discussion of organic chemistry.  All hydrocarbons can be categorized into two groups, saturated
and unsaturated.  Saturated compounds have single bonds between carbons with the other bonds
to carbon being made with hydrogen. Unsaturated hydrocarbons contain a double bond between
one or more carbon atoms thus, there are fewer hydrogen atoms attached to the carbons.
Aromatic compounds are unsaturated ring hydrocarbons with six carbons forming the ring.
Benzene is the most basic of the aromatic compounds having a structure of CeHe.  Other
aromatic compounds are variants of the benzene ring.  Toluene has a methyl group replacing one
hydrogen molecule attached to the six carbon ring of benzene. Xylenes have two methyl groups
replacing two of the hydrogens of the benzene ring.

       Five reactions take place in  a reformer: 1) The dehydrogenation (hydrogen removal) of
naphthenes; 2) The dehydroisomerization (hydrogen removal and conversation of hydrocarbons
from straight chain to branched chain) of alkyl cyclopentanes; 3) The isomerization (conversion
of hydrocarbons from straight chain to branched chain) of paraffins and aromatics; 4) The
dehydrocyclization (hydrogen removal and conversion of hydrocarbons from straight chain to
cyclic) of paraffins; and 5) The hydrocracking (conversion of hydrocarbons to smaller molecules
with hydrogen as a reactant)  of paraffins and naphthenes. Reactions numbered 1, 2 and 4 form
aromatic compounds, while reaction number 3 can alter aromatic types. There are two very
important reactions which result in  the formation of benzene. Reaction number 1 forms benzene
from cyclohexane. Reaction number 2 forms benzene from methyl cyclopentane.  Reactions
numbered 1, 2, & 4 produce hydrogen as a by-product.  Reaction number 3 neither produces nor
consumes hydrogen. Reaction number 5 consumes hydrogen.10,11

       Reformer Feed and Operations

       The feed to the reformer comes from the splitter bottom as we described previously; in
some cases, the feed may come directly from  the crude tower. Until recently, the reformer feed
boiling point range was  about 180°  F to 370°  F. The 180° F initial boiling point temperature sets
the cut between the hexanes and pentanes in the crude tower overhead. If the initial boiling point
of the feed is lower than 180° F, pentanes that are  normally not considered good feed will be
pulled into the reformer. The 180°  F temperature has varied somewhat according to the crude
from which the feed comes and also according to a particular refiner's economics.

       Feed boiling point (FBP) adjustments  often have to do with economics. The maximum
FBP for reformer feed is about 390° F to 400° F. The catalyst will coke (accumulate carbon) at
370° F, but as the feed FBP's rise above 370° F the coking rate rises increasingly more rapidly,
until at the 390° F to 400° F range,  the catalyst cycle length is far to short to  even be considered.
 On the other hand, the reformer feed portion  that boils above about 340° F could be cut into
kerosene, jet fuel, or diesel. In other words, the price-spread between gasoline and diesel may
warrant cutting some of the heavy straight run into diesel. Under other economics, it may pay to
run the reformer feed FBP up as high as possible in order to maximize gasoline make.  During

                                         6-18

-------
summer months the demand for gasoline grows while the demand for diesel fuel drops. To stay
in balance, a refiner may raise the FBP of the HSR to as high as 390° F.  This move would
increase the reformer feed volume and at the same time reduce the kerosene and ultimately the
diesel make.  If the refiner has a jet fuel contract, he may not be able to make such a change.
Increasing the initial boiling point can reduce the benzene make in the reformer. This is covered
in the next section discussing the technologies for reducing gasoline benzene levels.

       Different crude oil types affect the quality and volume of feed to the reformer. Light,
sweet crude, such as that produced in southwestern Wyoming, is reported to have had as much as
35% to 45% by volume of heavy  straight run (HSR) naphtha that is high naphthenes and
aromatics and consequently a fairly rich feed.  By contrast, there are heavy asphaltic crudes
produced from off the California  coast with almost no HSR.A  Reformer feed often includes
intermediate streams from hydrocrackers and cokers. Coker naphtha ordinarily must be
hydrotreated at conditions well beyond the severity of the common reformer hydrotreater before
it is fed to a reformer.  HSR from a hydrocracker is usually very clean with regard to most
critical contaminants, but as a rule must be reformed because it has a very low octane.
Occasionally a refiner must consider reforming a poorer feed (e.g., feed from paraffinic crude).
In such cases, the refiner may need to load two or three different catalysts into his reactors in
stacked-beds in order to provide for all the necessary reactions.  Paraffinic feedstocks are
ordinarily difficult to reform.

       A reformer consists  of essentially three separate components: the naphtha hydrotreater
section, the reformer section, and the product stabilization section.  The reformer section
contains a catalyst which  is usually bi-metallic; platinum and rhenium are two that are often
used. Consequently, the catalyst  is quite expensive.

       The feed to the reformer is hydrotreated to reduce contaminants, such as sulfur, nitrogen,
and arsenic. Arsenic poisons the catalyst, from which the catalyst activity cannot be recovered;
sulfur and nitrogen deactivate the catalyst and to some degree activity can be regained through
regeneration. The process conditions of the hydrotreater are ordinarily not severe; using
common hydrotreating catalysts,  temperatures around 600° F and pressures of around 400 psi.

       The hydrotreater reactor effluent is fed to a stabilizer/splitter to remove light products
and gaseous contaminants, such as hydrogen sulfide formed in the hydrotreating process. The
stabilizer bottoms are heated against reformer reactor effluent in feed/effluent exchangers, and
subsequently fed to the first pass  of the reformer feed furnace.  There are typically four reactors
IA & IB, II, and III, in series.  The feed is heated to a feed temperature of about 930° F in the
first pass and fed down-flow to reactors IA & IB, where several endothermic reactions take
place; the reactor effluent is then  fed to the second furnace pass and reheated to the same reactor
inlet temperature as for the first set of reactors. It is subsequently fed to reactor II.  The effluent
is heated once again, and  fed to the third furnace to be reheated and fed to the third reactor.

       Effluent from the third reactor is cooled against first-pass furnace feed in the
feed/effluent exchangers and fed  to the high pressure separator.  One of the principal byproducts
        Internal document.

                                           6-19

-------
of the reforming reactions is hydrogen.  Volumes in excess of 1000 scf per barrel of feed have
been reported. The high pressure separator is used to separate the hydrogen from the cooled
reactor effluent liquid.  Part of the hydrogen is recycled back to the reformer; mole ratios of five
moles of hydrogen to one mole of feed are usually required to suppress catalyst coking.  Some of
the excess hydrogen is fed to the naphtha hydrotreater and the balance is available for other units
in the refinery that may need it;  e.g., cat feed hydrotreaters or distillate hydrotreaters are
examples. The liquid reactor effluent is reheated and fed to a stabilizer to control the Reid Vapor
Pressure (RVP) of the final reformate.  The stabilizer is ordinarily a total-reflux unit, the pressure
of which is controlled by a gas controller on the tower overhead drum. Light hydrocarbons in
the off-gas, released to maintain pressure control, are sent to either the gas plant or to fuel gas.
The light hydrocarbons in the off-gas includes methane,  ethane, propane and butanes in small
volumes.

       Different reformer operating conditions result in  the production of different qualities of
reformate, different hydrogen production levels and can  change the reformer cycle length (time
between catalyst replacements or regeneration).  For example, low reactor pressure increases
yield and octane but increases the production of coke. Increased hydrogen partial pressure, that
is the ratio of hydrogen to hydrocarbon, suppresses coke formation, it promotes hydrogen yield
and product octane, but it also promotes hydrocracking.  Reducing the space-velocity, that is the
rate at which the reactor volume of the hydrocarbon changes per unit time, favors aromatic
production, but also promotes cracking. Higher activity  catalysts increase cycle lengths and
usually yields, but sometimes they are more expensive.12

       Certain tools are available to refiners to tailor the reforming process to their needs. There
are several proprietary processes, including catalysts, from which refiners can choose to treat the
specific qualities of their heavy  naphtha. In most cases,  a few laboratory tests allow vendors to
estimate, with reasonable accuracy, how well their processes can reform a given feedstock.
However, in some cases, vendors insist on running pilot  plant tests before they will guarantee
their process's performance. A  common lab test, known as a PONA, is used to determine
paraffin, olefm, aromatic, and naphthene content; API gravity,  sulfur, nitrogen,  and metals are
also important.  From these test  results, most vendors have computer-based process simulators
that, for a given RON, can estimate the finished product  and hydrogen yield, off-gas composition
at several different Reid Vapor Pressures (RVP), reformate octanes, and catalyst cycle lengths, if
a unit already exists with suitable reactors and compressors in place. In nearly all cases, vendors
supply the above test results for a range of RON's.  For example, the lowest RON a refiner may
decide to produce might be 85 RON. A vendor could provide process design services to
determine the cycle length requested by that refiner for a set of specified equipment design
criteria. This, of course, is based on, among other criteria, the type of reformer.

       Types of Reformers

       There are two types of reformers in use today, the semi-regenerative reformer, and the
continuous reformer. The predominant operating differences between the two are the pressure
and the means for regenerating the catalyst.
                                           6-20

-------
       The semi-regenerative reformer gets its name from the need to periodically shut down the
unit to regenerate and reactivate the catalyst. The catalyst, usually carrying a specific weight
percent platinum and rhenium on a common base material, is loaded in a series of down-flow
reactors.  The process pressure is higher in this type of reformer, at around 200 psi to 350 psi.
Reactor inlet temperatures begin at around 930° F. This start-of-run inlet temperature may vary
from process to process, as will the final end-of-run temperature. A delta temperature from start
to end of about 40° F is common.  Over time, as a result of some of the reforming/hydrocracking
reactions, coke builds up on the surface and the catalyst deactivates.  As coke is gradually
deposited on the catalyst, the reforming reactions slow down somewhat and the reformate or
product octane begins to drop a little below the desired set point. To compensate, the feed
temperature is raised until  the desired octane is reached again.  These steps are repeated
periodically over the cycle length of the particular catalyst. Contaminants such as  sulfur can
speed up the deactivation,  as can other problems.  When the maximum allowable feed
temperature is reached, the refiner must shut the unit down and regenerate the catalyst.

       Regeneration may take place "in situ" or the catalyst may be removed from the unit and
sent to a regeneration contractor for regeneration. Briefly, regeneration involves carefully
burning the coke off of the catalyst surface, and then chemically treating the clean  catalyst to
reactivate it. Regeneration is a fairly delicate operation, in that, for example, if too much oxygen
is allowed into the process, the temperature may get high enough to damage the catalyst and
prevent it from being reused. Regeneration, whether in situ or away from the refinery, is
generally done the same way. The one significant difference is that the catalyst is not reduced
with hydrogen directly following the burn phase at the off site plant. If carried out in situ, the
process can go forward without interruption. Some refiners insist on burning in  situ.
Regardless, the catalyst still must periodically be dumped, screened to remove fines, and
reloaded. The burn phase  also usually takes place before the unit is shutdown for other
maintenance.  Startup following a regeneration period also requires patience and may take
several days before a specified product octane can be reached.  An important step is to dry out
the catalyst before attempting to raise the reactor inlet temperatures to achieve the desired
octane. As the catalyst "life" shortens, the start-of-run temperature will gradually increase, so
that the usual delta T will gradually become narrower and eventually the catalyst cycle length
becomes too short to be economical.

       This regeneration process can be burdensome on refiners. For this reason, refiners
choose to operate this unit at a higher operating pressure to reduce the frequency of regeneration
cycles. The higher operating pressure reduces the formation of coke on the catalyst which
extends the cycles between regeneration. Higher pressure also reduces hydrogen make and
increases the cracking of heavier aromatics to benzene.

       The second type of reformer uses continuous catalyst regeneration, wherein the catalyst
is continuously withdrawn from the process, the coke burned off, the catalyst is reduced, and fed
back into the process without shutting the unit down for long operating periods.  In some ways,
the process is similar to the FCC.  The reactors are stacked rather than lined up separately in
series so that the catalyst can flow under gravity. From the bottom of the reactor stack, the
'spent' catalyst is 'lifted' by nitrogen to the top of the regenerator stack. In the regenerator, the
above mentioned "regeneration" steps of coke burning, chlorination and drying are done in

                                           6-21

-------
different sections, separated by a system of valves, screens, and other equipment. From the
bottom of the regenerator stack, catalyst is lifted by hydrogen to the top of the reactor stack, in a
special area called the reduction zone, where once heated is brought into contact with hydrogen,
which reduces (changes the valence) the catalyst surface to restore its activity. A continuous
regeneration process can be maintained without unit shutdown for run lengths of about 4 to 5
years.

       The continuous reformer's regeneration process is much more streamlined than the semi-
regenerative reformer. For this reason, the continuous reformers are operated at a considerably
lower pressure, from as low as 90 to 120 psi, than the semi-regen process and the hydrogen make
is considerably higher. For the same reason, the severity of continuous reformers can be higher
and product octane in the range of 104 RON is not uncommon.  The lower pressure of the
continuous reformer also causes less benzene make from the cracking of heavy aromatic
compounds.

       The above information has been presented from a conceptual point of view.  For an
informative discussion see13

6.3.2   How can Benzene Levels be Reduced in Gasoline?

       There are several ways available to refiners to reduce the benzene in their finished
gasoline.8  One is to prefractionate the feed, thus the benzene precursors out of the reformer.
The other is to post-fractionate reformate into light and heavy cuts, and either saturate the
benzene in the light cut or extract it for sale in the chemical feed market.

6.3.2.1        Prefractionation to  Reroute Benzene Precursors

       The heavy straight run naphtha can be cut differently to reduce gasoline benzene levels.
As discussed earlier, the heavy straight run naphtha is cut to prevent the C5s from being sent to
the reformer.  This means that most of the C6s are sent to the reformer along with the C7s, C8s
and sometimes the C9s. The cut-point could be changed from between the Cs's and Ce's to
between the Ce's and Cy's; in so doing the benzene precursors are also cut out of the reformer.
To assure that most of the Ce's are cut out of the reformer feed, the initial boiling point of the
feed would need to be raised from 180° F to around 215° F to 220° F. by changing the draw
temperatures  on the units. The cut adjustments can be made in the pre-flash column (a simple
unit before the crude tower which removes the lightest compounds before entering the crude
tower), the crude tower overhead, or the naphtha splitter. These various distillation columns are
usually designed to make a fairly imprecise cut between the C6s and C7s, which would also cut
some Cy's out of the reformer feed.  Cutting some of the C7s out of the heavy straight run going
   The benzene reduction technologies are discussed here in the context of the feasibility for
reducing the benzene levels of gasoline to meet a gasoline benzene content standard. However,
this section could also substitute for a feasibility discussion of complying with a total air toxics
standard since benzene control would be the means refiners would choose for complying with
such a standard.
                                          6-22

-------
to the reformer would, of course, reduce the production of C7 aromatics (toluene), and further
reduce the make of hydrogen. This would be costly to the refiner, so the refiner pursuing this
strategy would be expected to increase the ability to make a sharper cut between the C6s and
C7s. They would accomplish this by adding height or adding trays to their existing naphtha
splitter. In many cases, the refinery would replace the existing naphtha splitter with a new taller
tower.  The naphtha splitter in some refineries would already be outfitted to make such a cut.

        A few other concerns would need to be addressed as a result of removing the benzene
precursors. Benzene has a fairly high octane blending value; well in excess of 100 RON.
Simple arithmetic demonstrates that for each one-percent benzene removed, the reformate octane
is reduced by at least one number. Most refiners can't tolerate this, particularly if other high
octane blendstocks are not readily available. An obvious means to recover the lost octane would
be to increase reformer severity; while this seems reasonable, there are generally additional
consequences. Increased severity will  likely convert more of the  Cy's, Cg's, and Cg's into
compounds that could finally end up as benzene. For example, methyheptane can also be
converted into benzene, through paraffin dehydrocyclizaion (the methylated paraffin is
converted into a cycloparaffm and dehydrogenated) and demethylization (the methyl group is
removed) the possibility of which is more likely in semi-regen reformers.  Similar reactions can
be predicted for other Cg and Cg alkanes, all of which reduces the net effect of the original
reduction. Even so, the benzene content will be lower than prior to prefractionation.  Addressing
the octane loss due to benzene precursor rerouting can be addressed through other means
described below in Section 6.6.  Other potential problems are that hydrogen production will be
reduced and that the increased severity naturally shortens the catalyst  cycle length; this is
particularly important for semi-regeneration units, but also affects the continuous regeneration
units.

       Cutting the benzene precursors out of the reformer feed would definitely reduce the
benzene content in gasoline, but it would not completely eliminate it.  As discussed above, some
of the benzene in reformate is formed by the cracking of heavy aromatics, thus some benzene
would remain in reformate. Also the naturally occurring benzene present with the benzene
precursors would still be present in the rerouted C6 stream.

6.3.2.2       Benzene Saturation via Isomerization

       The rerouted benzene precursor stream contains the naturally occurring benzene from
crude oil. An existing isomerization unit could be used to saturate this naturally occurring
benzene in the rerouted C6 stream.  The role of the isomerization unit is to convert straight chain
compounds to branched chain compounds using a catalyst and in  the presence of hydrogen,
which increases the octane of the treated stream. The isomerization reactor saturates benzene
using the hydrogen present in the reactor for the isomerization reactions. However, isomerate
has a fairly high RVP (in the range of 13 psi to 15 psi) which could make it difficult for the
refiner to add more isomerization capacity in that refinery while still meeting the RVP
requirement that applies to its gasoline. As such, a safe assumption could be made that the
refinery would be capable to use the existing isomerization unit up to the listed capacity of the
unit. The refiner presumably sized the isomerization unit to be able to use that capacity in the
first place.  Treating the benzene in the rerouted benzene precursor stream could be

                                          6-23

-------
accomplished by giving a higher priority to treating the rerouted C6 stream in the isomerization
unit. If the isomerization unit's capacity is reached before it can treat all the C5 and C6s, then
the original C5 stream could be backed out until all the C6s are treated. Even so, adding an
isomerization unit may be possible, which also may require the refiner to add some RVP
reduction capacity elsewhere in the refinery to compensate for increased isomerate.

       A potential drawback to isomerization is that as benzene is saturated, it  produces heat
(exothermic reaction).  Isomerization reactions are all equilibrium reactions.  As such, as the
temperature in the reactor increases, it changes the equilibrium and shifts the isomerization
reactions back, which could lower the product octane. The licenser of the Penex isomerization
process has provided a recommendation that the isomerization unit be limited to 6 volume
percent benzene in the feed for this reason. The refinery could still treat this C6 stream using
this means, it would, however, need an additional reactor installed before the isomerization
reactor solely designed for saturating the benzene in this stream.  The combined benzene
saturation reactor with the isomerization reactor is called a Penex Plus unit.

       Another potential drawback to the benzene saturation option is that it requires at least
three moles of hydrogen (as H^) per mole of benzene saturated; this of course would require
additional hydrogen production. Providing additional hydrogen would add additional operating
cost to supply this hydrogen and could require capital investment.

       The naphtha splitter overhead (typically light straight run gasoline, LSR, most of which
is C5's with some C6's) is routinely fed to an isomerization unit (otherwise it is blended  directly
into gasoline). Most refiners run the feed through a deisopentanizer to remove  isopentane, since
it won't need to be treated (it is already a branched chain compound and would only use up
existing capacity). The deisopentanizer bottoms are mixed with hydrogen, which helps
minimize coke formation on the catalyst; hydrogen is neither generated nor consumed in the
isomerization reactions.

       The reactor effluent, known as unstabilized isomerate, is fed to a stabilizer where the
vapor pressure is controlled.  Any light gas produced by minor cracking reactions is typically
scrubbed and blended into the refinery fuel gas system. Isomerate, at this point, would probably
have a clear octane number 10 points higher than the LSR feed; perhaps 80 to 82 RON.

       The overall severity of isomerization process conditions is relative low; the temperature,
and the total and hydrogen partial pressures are all relatively low, compared with, say, reforming
or some other refinery processes. Isomerization is a vapor-phase process which uses hydrogen
to suppress dehydrogenation and coking. The catalyst is ordinarily an  alumina  type onto which
organic chlorides have been deposited. In that the chlorides are sensitive to moisture, the feed
must be very dry. Some organic chloride is added to the feed in order to maintain catalyst
activity.

       Increasing the severity of the isomerization unit will likely increase the  product octane
but may likewise produce more light ends. Yields are highly dependent on feedstock
characteristics, which naturally are closely related to the characteristics of the original crude;
paraffmicity, aromaticity, etc.  Poor feed quality will usually yield net liquid percent recovered

                                           6-24

-------
in the mid-80's or less, while good feed quality may yield net liquid percent recovered in the
mid- to upper 90's (the rest being cracked to gaseous hydrocarbons). The key control variable is
probably the process temperature, in that raising it increases severity and promotes
hydrocracking side reactions.  Raising the process pressure may increase catalyst life but will
also likely promote hydrocracking reactions, which reduce the net liquid yield. While increased
hydrogen partial pressure may extend catalyst life, it nevertheless promotes hydrocracking side-
reactions that reduce net liquid yield. Run lengths can be extended using as low temperature as
possible with moderate hydrogen partial pressure  and reduced space velocity. This may or may
not seem obvious, but extending run lengths this way has drawbacks as far as product quality
and net yield  of octane-barrels is concerned.14

6.3.2.3        Reformate Postfractionation with Benzene Saturation

       Another method for reducing reformate benzene is to postfractionate reformate into
heavy and light cuts; the light, Ce, cut would contain the reformate benzene which could be
treated to remove benzene, while the C7+ stream would be blended directly into gasoline. An
important question associated with this methodology is the efficiency that the benzene could be
removed from the rest of the reformate, preserving the C7s. Based on vendor information, a
typical reformate splitter would be designed to capture about 96 percent of the benzene while
only capturing 1  percent of the toluene in the C6 stream. The refinery would design this unit as
appropriate for the refinery considering their particular economics and refinery situation. The
C6 stream would then be sent to a benzene saturation unit to saturate the benzene into
cyclohexane.  There are two technologies for doing this. One is named Bensat and is licensed by
UOP. The other is named CDHYDRO and is licensed by CDTech,

       Bensat


       UOP has put their Bensat™ process forward as a way to reduce the benzene content of
gasoline. The process was originally developed to reduce to below six percent the benzene
concentration in  the feedstock to their Penex™ isomerization unit (the Penex unit is capable of
saturating the rest).  The process saturates the benzene converting it into cyclohexane, which can
then be fed to the Penex™  unit.

       Although the process was originally designed for Penex™ feed, the vendor has modified
it to be used to saturate the benzene in a light reformate cut. UOP reported in a bulletin
published on  one of their websites15  that a Bensat™ unit can be designed to handle from 5% to
30% benzene in the feed. Although  not stated, it was implied that the benzene content could be
reduced to below six percent.  We have received personal communications indicating that while
the benzene content of light reformate will normally vary, an average range would be about 15%
to!8%.

       The process is carried out in a standalone reactor and according to UOP the process uses
a commercially proven noble metal catalyst that is benzene-selective with no side reactions.
Since there is essentially no cracking there is also  essentially no coke lay-down on the catalyst to
cause deactivation.  Sulfur in the feed can deactivate the catalyst, but activity  can be restored by
                                          6-25

-------
removing the sulfur.  Of course, light reformate would be very low in sulfur; other feedstocks
may need to be hydrotreated.

       During start-up, hydrogen is mixed with the feed and pumped through feed/effluent
exchangers and a start-up preheater.  Once the unit is up and running, the heat generated by the
process provides heat to the feed via the feed/effluent exchangers.  Benzene saturation requires
three moles of H2 per mole of benzene, so makeup hydrogen is continually added to the reactor
feed.  The reactor effluent is routed to a stabilizer to remove light ends. As noted previously,
some octane loss due to benzene saturation can be regained by feeding the resulting cyclohexane
to an isomerization unit.16

       CDHYDRO

       Catalytic Distillation Technologies (CDTECH®) has two processes for reducing the
benzene content of gasoline by converting it into cyclohexane. Both are referred to as
CDHYDRO™ technologies, but one is actually specified for the selective hydrogenation of
benzene in the entire reformate to cyclohexane in a catalytic distillation column, while the other
is advertised to hydrogenate a benzene-only  stream to cyclohexane in a catalytic distillation
column.
       They advertise both processes online; we note that if a refiner finds it necessary to extract
the benzene from his reformate to saturate it, the  process advertised to convert benzene to
cyclohexane may be of interest17. However, we will focus on the process they put forward for
reducing the benzene content of reformate, in that they claim it is possible to do without
fractionating the reformate prior to the saturation step18. This has a clear advantage by
combining a splitting column with a benzene saturation reactor which would be expected to
reduce the capital cost for this technology.

       According to CDTECH® in excess of 90% of the benzene in reformate can be hydrated
and the treated Ce's removed from the final product, all in a single catalytic distillation tower;
the tower they recommend is a benzene-toluene splitter, either refitted or new. The feed appears
to be a mixture of low pressure hydrogen and reformate. The feed is sent to the column and the
benzene saturation reaction occurs in the reactor.  The overhead stream is condensed, cooled,
and collected in a reflux or overhead accumulator drum. The accumulator off gas, mainly
unreacted hydrogen, is recycled to feed. There also  appears to be an off-gas purge  stream.  The
reflux drum liquid is said to be primarily treated Ce's.  Part of the overhead is used  for tower
reflux while the balance is pumped back into the  C?+ treated reformate tower bottoms.  Since
this reaction process takes place in a conventionally designed C6/C7 splitter column, this column
could presumably be designed to treat the same benzene/ toluene split that a Bensat unit would
be designed for.
                                          6-26

-------
6.3.2.4        Benzene Extraction

       The extraction of benzene from reformate for use as a petrochemical feed can be a useful
way to remove the benzene from the gasoline pool.  This method is more attractive when the
refinery is located near to petrochemical complexes which use benzene as a feedstock.

       Benzene extraction involves three different steps.  The first step is to separate a C6
stream from the rest of reformate using a reformate splitter. This C6, benzene-rich stream is sent
to a liquid/liquid extraction unit where the benzene and any other aromatic compounds, such as
any toluene which may captured along with the benzene in the reformate, are extracted  from the
rest of the hydrocarbons.  This aromatic stream is then sent to a very robust distillation process
for concentrating the benzene for sale into the chemicals market.

       The reformate would be split to separate the C6s from the rest of reformate. This cut
would likely be made similar to the splitter unit used for the benzene saturation unit, although
since the toluene would only be separated and not be chemically treated,  refiners would have
more leeway to capture more of the benzene in this case with less effect on the rest of the stream
then with benzene saturation.

       After separation, the Ce light reformate cut, containing a  fairly complex mixture of
paraffins, isoparaffin, and benzene, would be fed to an extraction unit.  This type of operation,
commonly known as liquid-liquid extraction is one variation on  a whole host of extraction
processes used in the petrochemical industry.

       The essence of the benzene extraction process is to bring the light-reformate cut into
intimate contact with a slightly miscible to completely immiscible solvent,  into which the
benzene may be selectively transferred (absorbed or dissolved) from the light-reformate. Liquid-
liquid extraction is applied by several industries, including the pharmaceutical and perfume
businesses, in a variety of vessels, such as stirred mixer-settlers, high-speed rotary centrifugal
extractors, and various  columns, each of which is designed for a particular type of extraction.
There are several column types from which an engineer could choose, such as static or agitated,
along with spray, sieve plate, and packed columns. For the purposes of this discussion, we will
be referring to a static column.

       For our general case, the extraction column has essentially two inlet streams and two
outlet streams. One inlet stream, fed at the top of the column is the light-reformate from which
the benzene aromatic components are to be extracted.  The other inlet stream  is the lean solvent
(solvent with no aromatics in solution) which will extract the aromatics from  the light-reformate.
 The solvent flows upward, while the light-reformate flows downward, during which time the
two streams come into intimate contact on the surface of the tower internals.

       As designed, the solvent, containing the extracted aromatics, leaves the top of the column
as the extract or "aromatic-rich" stream. The light-reformate leaves the column bottom with
only a small residual volume of aromatics remaining and may be referred to as the raffinate. It
will consist mostly of paraffins and isoparaffins that can be sent to the gasoline blending pool.
                                          6-27

-------
       The aromatic-rich stream is then separated from the solvent, after which the solvent is
recycled back to the extractor for reuse. The benzene, subsequently separated from the other
aromatics, can be sold into the chemicals market. The benzene-free aromatics, consisting of
toluene and in some cases xylene, which have high octane blending values, can be sent to
gasoline blending or to the chemicals market as well.

       Despite only being regulated to reduce the benzene content of gasoline, the refiner may
choose to also extract toluene and xylenes. Taking such a step would cause a much larger
impact on the octane level of the refinery's gasoline and this octane loss would have to be
recovered.  This may be possible using the octane recovery technologies summarized below.
This may improve the economics for reducing benzene levels, particularly because xylenes are
valued more than benzene. Extracting the C6 - C8 aromatics may allow omitting the reformate
splitter since  refineries omitting the heavy straight run naphtha from the reformer feed (omitting
the C9+ fraction) could send all the reformate to the extraction unit.  The extraction unit would
have to be designed to be much larger and of course the downstream distillation unit would have
to be much larger as well.

       There are three proprietary extraction processes available.  They are the Udex, the
Sulfolane, and the Carom processes. The di-, tri-, and tetra-ethylene glycol isomers are used as
solvents.

       Extractive distillation provides what appears to be a very reasonable alternative to full
liquid-liquid aromatics extraction. According to one source, "Liquid-liquid extraction (LLE)
was for many years the primary choice for aromatics recovery, because the solvents available
during that time were not suitable for separating a wide-boiling range feedstock in the extractive
distillation mode of operation.  To do so required making narrow boiling feed fractions sent to
separate extractive distillation units." "However, solvent technology has improved, and the
availability of new solvent blends makes it feasible and more profitable to employ extractive
distillation to aromatics separation."19

       In short, when certain mixtures cannot be easily separated by ordinary distillation, either
because of low relative volatility or the presence of a homogeneous azeotrope, it may be possible
to effect a separation by the use of extractive distillation.  According to Perry's "In extractive
distillation, the agent or 'solvent' is considerably less volatile than the regular feed components
and is added near the top of the  column. Because of its low volatility,  the agent behaves as a
typical heavier-than-heavy key component and is also readily separated from the product
streams... A typical extractive distillation might be a unit for separating benzene and
cyclohexane using phenol as the separating agent. "Benzene and cyclohexane have nearly
identical boiling points and form a homogeneous azeotrope containing about 45 wt.%
cyclohexane. However, with the phenol present, the cyclohexane volatility is nearly twice that
of benzene."20  The benzene/cyclohexane mixture is fed at or near the  center of the distillation
column, while the phenol separating agent is fed into the tower a few trays below the top... The
phenol remains in the liquid phase and flow downward over the trays and out the bottom. The
overhead vapor is essentially pure cyclohexane.. .The bottom phenol/benzene stream is sent to a
second tower for separation.  Another source suggested using aniline for the
benzene/cyclohexane separating agent.21 A full-boiling range light reformate may be more

                                          6-28

-------
complicated, but the principles are essentially the same. It appears that the choice of separating
agent is critical. As demonstrated by the benzene/cyclohexane example we just described, using
two different solvents, it should be clear that there will likely be more than one choice available
for any given system.  An economic argument for using extractive distillation as opposed to
liquid-liquid extraction is that fewer pieces of processing equipment are usually required.

6.3.2.5       Low-Pressure Reformer Operation

      Lowering the pressure at which the reformer operates is another means of controlling the
benzene content. Lower pressure operation would provide some benzene reduction by reducing
the benzene formed from the hydrodealkylation (cracking) of heavier aromatics to benzene.
Beyond retarding the hydrodealkylation reaction, low pressure is an effective means of
increasing hydrogen and liquid yields, but can hurt catalyst cycle lengths.  Lowering process
pressure in a semi-regen unit is reported to provide from 50% to 70% benefits of a continuous
catalyst regeneration reformer.

      However, it is somewhat difficult to  lower the pressure of an early-design semi-regen unit
below a certain level.  The early generations of reformers were designed for pressures in the
range of 350 psi (as an example).  Higher pressure usually allowed design engineers to specify
small diameter pipe. Lowering the pressure changes the hydraulics, restricts flow, and the
reformer simply won't operate.  The recycle compressors would also likely need to be changed
in order to reduce the pressure.  In short, it is not a simple fix to change a unit from high-pressure
to low-pressure. Continuous regen reformers already operate at pressures considerably lower
than semi-regen units, in the range of say, 90 psi and therefore have little room for improvement.

6.3.2.6       Prefractionation Combined with Low-Pressure Reformer Operation

       Pre-fractionation of benzene precursors combined with low pressure reformer operation
(< 100 psi) will usually produce less than 1 vol% benzene in the reformate regardless of the feed
composition. If octane can be obtained through other means, this appears to be a useful
approach.

6.4    Experience Using Benzene Control Technologies

       All these benzene reduction technologies and octane generating technologies described
above have been demonstrated in refineries  in the U.S. and abroad.  Each of these technologies
have been used for compliance purposes for the federal Reformulated Gasoline program, which
requires that benzene levels be reduced to an average of 0.95 volume percent or lower starting in
1995. The two primary means used by refiners to produce low benzene gasoline for the RFG
program is routing benzene precursors around the reformer and benzene extraction. Benzene
saturation is another technology used to achieve benzene reductions for the reformulated
gasoline program on a limited basis.

       According to the Oil and Gas Journal's worldwide refining capacity report for 2003,
there are 27 refineries  in the U.S. with extraction units. Those refineries which chose extraction
often reduced their benzene to levels well below 0.95 volume percent because the value of

                                          6-29

-------
benzene as a chemical feedstock is high. The reformulated gasoline program also caused the
installation of a couple of benzene saturation units. There are two benzene saturation units in the
Midwest installed in refineries there to produce RFG for the markets there. California has its
own reformulated gasoline program which also put into place a stringent benzene standard for
the gasoline sold there. The Oil and Gas Journal's Worldwide Refining Report shows that four
California refineries have benzene saturation units. If we assume that those refineries producing
RFG that do not have extraction or saturation units are routing their precursors around their
reformer, then there are 28 refineries using benzene precursor rerouting as their means to reduce
benzene levels.  Personal conversations with several refiners have revealed that some of the
refineries which are routing the benzene precursors around the reformer are sending that rerouted
stream to their isomerization unit for saturating the benzene and recovering lost octane. Thus,
these four technologies have been demonstrated in many refineries since the mid-90s in the U.S.
and should be considered by the refining community as commercially proven technologies.

       A vendor of benzene control technology has shared with us how the refining companies
in other countries have controlled the benzene levels of their gasoline in response to the benzene
standards put in place there. In Europe, benzene control is achieved by routing the benzene
precursors around the reformer and feeding that rerouted stream to an isomerization unit.  In
Japan, much of the benzene is extracted from gasoline and sold to the chemicals market.  Finally,
in Australia and New Zealand, refiners use benzene saturation to reduce the benzene levels in
their gasoline.

6.4.1   Benzene Levels Achievable through Reformate Benzene Control

       We evaluated the benzene levels achievable by refineries applying benzene control in
two different ways. One way was to evaluate the benzene levels of refineries in 2003 which are
producing low benzene gasoline to comply with the RFG requirements. The second way  was to
use the refinery-by-refinery cost model to evaluate the benzene levels achievable by the various
benzene control technologies.

       Refiners today are producing gasoline with low benzene levels for sale into the RFG
market. The RFG program requires that gasoline must meet a 0.95 benzene control standard.
While the benzene standard is much less stringent than the proposed 0.62 benzene control
standard, many refiners comply at a much lower level probably because they are using benzene
extraction to comply. When extracting benzene from gasoline, the high capital costs associated
with extraction provides a strong incentive to maximize the extraction of as much benzene as
possible.  The low benzene levels achieved by today's refineries provide an indication of the
feasibility of complying with the proposed 0.62 benzene standard.  RFG averages 0.62 volume
percent benzene - the same level as the proposed benzene standard.

       There are 17 refineries today producing gasoline which currently averaged 0.62 volume
percent benzene or below.  Of these 17 refineries with very low benzene levels, 11 are located in
PADD 3, four are located in PADD 1, and one each are located in PADDs 2 and 4. The benzene
levels for these refineries range from 0.29 to 0.62 volume percent and average of 0.51 volume
percent. The  average benzene level for these refineries is well below the proposed 0.62 benzene
standard. We reviewed the list of refinery unit capacities from EIA and the Oil and Gas Journal

                                          6-30

-------
to determine if these refineries have benzene saturation or extraction benzene control
technologies.  Of the 17 refineries with benzene levels at or below 0.62 volume percent, 14 of
these have benzene extraction or saturation units, while two more are assumed to be selling
reformate to other refineries with extraction units.  While this demonstrates that achieving the
proposed benzene is feasible for a portion of U.S. refiners, this does not indicate that all U.S.
refiners are capable of achieving a 0.62 volume percent benzene level.

       To assess the ability for the  rest of the refineries to achieve a benzene level of 0.62 or
below, we used the refinery-by-refinery model. For each benzene control technology, we
assessed its ability to achieve benzene reductions.  Routing the benzene precursors around the
reformer is the least severe benzene control technology.  The refinery by refinery cost model
shows that refineries using this technology can reduce their gasoline benzene levels from an
average of about 1.6 volume percent to 1.1 volume percent, a 30 percent reduction. The
refinery-by-refinery cost model shows that only two refineries would be able to meet or exceed
the proposed 0.62 benzene standard using this technology.  This technology is clearly
insufficient for achieving the proposed benzene control standard by itself.

       Those refineries with isomerization units would be able to route their rerouted  benzene
precursors to this unit further reducing their benzene levels by saturating the naturally occurring
benzene in this stream.  The refinery-by-refinery cost model shows that on average these
refineries would be able to reduce their gasoline benzene levels to 0.75 volume percent using this
technology combined with benzene precursor rerouting.  Of these refineries, 9 would  be able to
achieve the proposed 0.62 benzene  standard. Averaged across the U.S. refineries, benzene
precursor rerouting can achieve about a 60 percent reduction in reformate benzene levels. When
benzene precursor rerouting is combined with isomerization, about an 80 percent reduction in
reformate benzene levels  is possible.  While this benzene precursor rerouting combined with
isomerization can achieve a significant reduction in refinery benzene levels, the application of
further benzene control technologies is still required to enable the U.S. refining industry to
achieve the proposed benzene control standard. The reason why these combined benzene control
technologies are incapable of achieving a significant enough benzene reduction is because they
do not address the benzene formed from reforming the heavy part of reformate.

       We assessed the benzene reduction capacity of benzene saturation and benzene
extraction.  These two technologies are able to achieve a deeper reduction in gasoline benzene
levels because they treat all the benzene in reformate - that formed from the six carbon
hydrocarbons, that formed from the cracking of heavier aromatics to benzene in heavy reformate,
and the naturally occurring benzene which is in the feed to the reformer. Our analysis of these
benzene control technologies reveals that they are able to reduce reformate benzene levels by 96
percent. The refinery-by-refinery model shows that for those refineries that were found eligible
for using benzene saturation, they were able to reduce their gasoline benzene levels from about
1.6 volume percent to 0.5 volume percent, a 60 percent reduction. For refineries identified as
eligible as using benzene  extraction, the refinery-by-refinery cost model estimates that they are
capable of reducing their  gasoline benzene levels from 0.9 volume percent to 0.5 volume
percent, a 40 percent reduction.  The refineries eligible for benzene extraction are already low in
benzene because many of them are using extraction today, or they are selling a benzene-rich
reformate stream to a neighboring refinery which is extracting the benzene from this stream.

                                           6-31

-------
However, the refinery-by-refinery cost model estimates that they are able to achieve further
benzene reduction, by revamping their benzene extraction units to do so. While the use of
benzene extraction is limited to refineries on the East and Gulf Coasts, where they have access to
the petrochemical markets, the use of benzene saturation is not limited. Therefore, each refinery
in the U.S. is able to install one of these two benzene control technologies.  We assessed the
benzene reduction capacity of using these two maximum reformate control technologies.

       We found that, on average, U.S. refineries could achieve a benzene level of 0.52 volume
percent based on the maximum level of benzene control from reformate, assuming that benzene
saturation or extraction was applied in  each refinery in the country. However, this average was
obtained by  averaging refineries with benzene levels both above and below 0.52 volume percent
ranging between 0.29 to 0.78 volume percent benzene. To illustrate the benzene levels
achievable by the application of benzene extraction and benzene saturation in each refinery in
the U.S., we plotted the estimated final benzene level for each refinery against their cumulative
gasoline volume from low to highest benzene level in Figure 6.4-1. To provide a perspective for
how the gasoline benzene levels for U.S. refineries compare to the proposed 0.62  benzene
standard, we provided a line at 0.62 volume percent benzene.

 Figure 6.4-1.  Benzene Levels achievable by U.S. Refineries Applying  Benzene Extraction
                                     and Saturation
     0.90
     0.80
     0.00
       0.00     1000.00    2000.00    3000.00    4000.00    5000.00    6000.00
                               Cumulaltive Volume (Thousand barrels/day)
                                                                  7000.00
                                                                          8000.00
                                                                                   9000.00
       As shown in Figure 6.4-1, the refinery-by-refinery cost model estimates that if reformate
                                           6-32

-------
were treated with benzene saturation and benzene extraction, 13 refineries would continue to
have benzene levels above 0.62 volume percent benzene. Under the ABT program, this would
not be an issue since those refineries with benzene levels above 0.62 could purchase credits from
refineries with benzene levels below the 0.62 benzene standard. However, credits must always
be available for these refineries to show compliance with the proposed 0.62 benzene standard.
While we believe that credits would be available, it is still possible to show that each refinery
could attain the benzene standard with additional benzene control options available to them.

6.4.2   Other Benzene Controls

       We have identified other technologies that could be used to reduce gasoline benzene
levels in addition to the reductions modeled in the refinery-by-refinery cost model. Although we
have not quantified their costs, they are expected to be more expensive and therefore less
attractive for achieving benzene reductions than the reformate treating technologies identified
above.

       One of these less attractive opportunities would apply for those refineries using benzene
saturation or extraction.  They could achieve additional benzene reduction with these units by
capturing more of the benzene in the reformate splitter and  sending this additional benzene to
their saturation or extraction unit.  Normally refiners attempt to optimize the capital and
operating costs with the amount of benzene removed when  splitting a benzene-rich stream out of
the reformate stream for treating in a benzene saturation or  extraction unit.  To do this, they
optimize the distillation cut between benzene and toluene, thus achieving a benzene reduction of
about 96 percent in the reformate while preserving all but about 1 percent of the high-octane
toluene. However, if a refiner was to be faced with the need for additional benzene reductions, it
could change the distillation cut in their existing reformate  splitter to send the last 4 percent of
the benzene to the saturation or extraction units. This action though would also capture more of
the seven carbon hydrocarbons, resulting  in the saturation of the toluene contained in the seven
carbon hydrocarbons.  Refiners using this strategy to capture more of the benzene in the
reformate  splitter would have to have sufficient capacity downstream in the saturation or
extraction units to process this additional  volume, although refiners normally design their units
with some excess capacity. They  could design either their reformate splitter, or their benzene
saturation or extraction units with this end in mind.  On the one hand, they could design their
reformate  splitter to be larger to make a "hard cut" thus capturing virtually all the benzene and
rejecting virtually all the toluene;  sending only the additional volume of benzene to their
downstream saturation or extraction unit.  This option would entail increased capital and
operating costs for their reformate splitter. On the other hand, they could maintain the optimized
reformate  splitter but design additional excess capacity in their downstream saturation and
extraction units to handle the additional seven carbon hydrocarbons that would be sent to these
units. In the case of benzene saturation, the benzene saturation reaction would have to be sized
larger. In  the case  of benzene extraction,  the benzene extraction unit would have to be designed
to handle the increased six and seven carbon hydrocarbons  forwarded to it by the reformate
splitter. The aromatics distillation equipment downstream of the extraction unit would also have
to be sized larger to separate the additional toluene and benzene sent to this unit. For each of the
13 refineries which the refinery-by-refinery cost shows could not achieve 0.62 volume percent
benzene, we estimate the extent that benzene levels could be further reduced by capturing the

                                           6-33

-------
remaining reformate benzene and treating it in a saturation unit or extracting it from gasoline,
and summarize this in Table 6.4-1 below.

       Another means for further reducing the benzene levels for 5 of these 13 refineries which
have hydrocrackers or cokers is to reduce the benzene content of one of the products of the
hydrocracker or coker units, the light hydrocrackate naphtha or light coker naphtha streams.
Light hydrocrackate and light coker naphtha are normally blended directly into gasoline.  These
streams are estimated to contain on average 2 volume percent benzene.  While this level of
benzene is moderate relative to the benzene levels of reformate, its benzene contribution to the
gasoline pool for these refineries is significant.  Light hydrocrackate or light coker naphtha could
be treated by routing these streams to an isomerization unit, similar to how refiners isomerize the
six-carbon straight run naphtha as discussed above.  Isomerizing this stream would increase its
vapor pressure and could require additional steps to counter the vapor pressure increase by
lowering the vapor pressure of the FCC naphtha as described below discussing the methodology
for achieving vapor pressure reductions. Alternatively, the refiners could use  additional
distillation equipment to cut the light hydrocrackate and coker naphtha more finely. In this way,
more of the benzene could be shifted to the "medium" hydrocrackate and coker streams, which
are sent to the reformer and thus would be treated along with the rest of reformate in benzene
saturation or extraction units. For each of the 6 refineries with a hydrocracker or coker which
the refinery-by-refinery cost model shows could not achieve 0.62 volume percent benzene, we
estimate the extent that benzene levels could be further reduced by addressing the benzene in
light hydrocrackate and summarize this in Table 6.4-1.

       Another way that the gasoline benzene levels of most of these refineries could be further
reduced would be to treat the benzene in natural gasoline.  Many U.S. refiners, especially in
PADDs 4 and 5, and to a lesser extent in PADDs 2 and 3, blend some light gasoline-like
material, which is a by-product of natural gas wells, into their gasoline. We assume that this
material is blended directly into gasoline by each refinery in each PADD where natural gasoline
is a feedstock for refineries.  The benzene concentration in this stream is estimated to be 1.3
volume percent which, because it is not high, would be costly to treat for reducing its benzene
content.  However, by 2011 which is when this rule would take effect, refiners may be treating
this stream in the refinery to reduce its sulfur level.  To comply with the 30 ppm Tier 2 sulfur
standard, refiners may be treating this stream in a way to reduce its sulfur. Because natural
gasoline is fairly low in octane, it could be treated in the reformer to both reduce its sulfur as
well  as improve its octane.  If this stream is treated in the reformer to treat its sulfur, it would
also be treated for benzene if reformate benzene control are later added to meet a benzene
control standard. Another way that the sulfur of the light portion (that which contains the
benzene) could be treated for reducing its sulfur is with an extractive caustic treater such as a
Merox unit (see the section below on sulfur control). While  this technology would address the
sulfur in this stream it would not reduce, nor would it place this stream in the position to reduce,
the benzene level of this stream. Another way that these refineries with high benzene levels
could deal with the benzene of natural gasoline is to simply stop purchasing all or a part of the
natural gasoline that it currently purchases.  This volume of natural gasoline that could be
rejected by these refineries could then be purchased by other refineries. For each of the
refineries which are assumed to be purchasing natural gasoline in the refinery-by-refinery cost
model, and which could not achieve 0.62 volume percent benzene with reformate benzene

                                           6-34

-------
control, we estimated the extent that treating the benzene in natural gasoline could lower their
gasoline benzene levels in Table 6.4-1.

       Another possible option for these refineries to further control benzene might be to control
the benzene content in naphtha from the fluidized catalytic cracker, or FCC unit. As shown in
Table 6.3-1 above, FCC naphtha contains less than 1 percent benzene on average. Despite the
low concentration of benzene in FCC naphtha, the large volumetric contribution of this stream to
gasoline results in this stream contributing a significant amount of benzene to gasoline as well.
There are no proven processes which treat benzene in FCC naphtha.  This is likely because its
benzene concentration is low as well as because FCC naphtha contains a high concentration of
olefins. Segregating a benzene-rich stream from FCC naphtha for sending to a benzene
saturation unit would saturate the olefins in this stream, in addition to the benzene, causing an
unacceptable loss in octane value. Such a stream could probably be sent to an extraction unit,
but this would be expensive to treat because of the low benzene concentration in this stream.
There may be another way that a few refiners could further reduce their benzene levels. We
learned that one refinery is operating their FCC unit very severely to produce a high octane (92
octane number) gasoline blendstock.  This resulted in this particular FCC naphtha having a
benzene content of 1.2 volume  percent. This refiner could change the operations of their FCC
unit (change the catalyst and operating characteristics) to reduce the severity and produce
slightly less benzene and make up the octane loss in other ways, such as blending in ethanol.22
We do not know if any of the refineries which the refinery-by-refinery  cost model has identified
as not being able to achieve the 0.62 benzene standard using reformate benzene control are
operating their FCC units this way. Thus, we cannot estimate that any  of these refineries could
reduce their gasoline benzene levels by reducing the severity of their FCC units.
    Table 6.4-1. Additional Benzene Reduction Achievable by non-Reformate Means of
  Control for Refineries Unable to Achieve the Proposed 0.62 Standard using Reformate
                                        Control
Refinery Number
1
2
3
4
5
6
7
8
9*
10
11
12
13
Gasoline Benzene
Level after
Reformate Benzene
Control
0.78
0.77
0.70
0.75
0.66
0.64
0.63
0.67
0.77
0.64
0.70
0.74
0.65
Treating last 4% of
Reformate Benzene
-0.06
-0.11
-0.06
-0.10
-0.05
-0.07
-0.06
-0.11
-0.07
-0.08
-
-
-0.06
Treating 96% of
Light Hydrocrackate
and Coker Naphtha
Benzene
-0.03
-0.37
N/A
-0.38
N/A
N/A
N/A
-0.37
N/A
N/A
-0.23
-0.42
N/A
Treating 96% of
Natural Gasoline
Benzene
-0.07
-0.13
-0.07
-0.12
-0.07
-0.09
-0.09
-0.15
-0.03
-0.03
-0.27
-0.02
-0.07
* Refinery #9 is shown to have added an isomerization unit after 2003 that is estimated to reduce its gasoline
benzene level 0.12 volume percent.  This will be modeled in the final rule.
                                          6-35

-------
6.5    Averaging, Banking, and Trading (ABT) Program

       We are proposing that refiners and importers could use credits generated under the
averaging, banking, and trading program (ABT) to meet the 0.62 vol% benzene standard in 2011
and beyond. This regulatory impact analysis0 begins with a discussion of starting refinery
benzene levels then explains the strategies refineries would take to meet the standard. For
refineries that plan to reduce actual benzene levels, we have explained when the benzene
reducing steps would occur and how early process changes made prior to 2011 would generate
early credits that could provide the refining industry with additional lead time to make their final
investments. We also explain the basis and derivation of early credit  baselines, early credit
trigger points, and the trigger point value. We have provided  an analysis of how the early credit
program would enable a gradual phase in of the standard and an amortization of refinery
compliance costs. We also explain which refinery improvements would be postponed until 2011
or later as early credits permit. We conclude with a discussion of ending refinery benzene levels
and an explanation of how program credits would be generated and traded to meet the 0.62vol%
standard on an average nationwide basis.

6.5.1   Starting Gasoline Benzene Levels

       In order to begin the ABT analysis, it was first necessary to establish a baseline benzene
level for each refinery. Batch benzene concentrations are provided to EPA as part of the existing
RFG/anti-dumping refinery requirements. In summer 2003, the benzene content of gasoline
produced by 115 U.S. refineries ranged from 0.41 to 3.81 vol% with an overall volume-weighted
average of 0.97 vol% as shown in Table 6.5-1.
                          Table 6.5-1. Starting Benzene Levels

PADD1
PADD2
PADD3
PADD4
PADD 5****
California
Total
No. of Refineries by Gasoline Benzene Level (vol%)
<0.5
4
0
4
0
0
0
8
0.5-<1.0
3
5
18
1
0
12
39
1.0X1.5
3
8
10
4
1
0
26
1.5-<2.0
0
11
7
6
3
0
27
2.0-<2.5
2
1
0
3
2
0
8
>=2.5
0
1
2
2
2
0
7
Benzene Level (vol%)*
MIN
0.41
0.60
0.41
0.60
1.36
0.51
0.41
MAX
2.19
2.85
3.10
3.56
3.81
0.77
3.81
RANGE**
1.77
2.25
2.69
2.96
2.44
0.26
3.39
AVG***
0.62
1.32
0.86
1.60
2.06
0.63
0.97
    * Starting benzene levels based on summer 2003 batch data
    ** Range in benzene level (MIN-MAX)
    *** Average volume-weighted benzene level
    **** PADD 5 excluding California
       The ABT analysis for this proposal includes all U.S. refineries including California since
the decision to exclude California gasoline from this proposal was made subsequent to this
analysis. For the final rule, the analysis presented here would be redone using the best available
 This analysis includes small refiners
                                           6-36

-------
batch gasoline data and excluding California refineries. We predict that there would be some
changes in the results of the analysis (i.e. who/where the benzene reductions come from,
compliance costs, etc.) however, we believe the overall outcome would be relatively unaffected.
 We anticipate very few changes as a result of using more current batch data since there have not
been any changes in gasoline benzene regulation that would significantly impact starting
benzene levels. We also believe there would be few changes associated with excluding
California refineries from the analysis since their average starting benzene levels are already
near the proposed 0.62 vol% standard based on existing state fuel programs. Our current ABT
analysis does not predict them to make very many changes in benzene level nor does it suggest
they would be a key player in the proposed credit generation and trading program. As such,
removing them from the analysis should have very little impact.

       There is currently a wide variation in nationwide gasoline benzene levels. The variation
(explained in more detail in 6.2) is primarily attributed to crude oil quality, use of low-benzene
blendstocks, benzene control technology, and refinery operating procedures.
The variation or range in starting benzene  levels has been calculated to equal 3.39 vol% overall
or 1.77, 2.25, 2.69, 2.96, and 2.44 vol% for PADDs 1-5, respectively as shown in Table 1.

       In part due to this variation in starting benzene level, we predict that it would be much
more difficult for some refiners to comply with the 0.62 vol% gasoline benzene standard in 2011
and beyond based on actual levels than others.  As such, we are proposing an ongoing
nationwide averaging, banking, and trading (ABT) program that would allow some refineries to
maintain gasoline benzene levels above 0.62 vol%, provided they are equivalently offset by
refineries below the standard.  Refineries that elect to maintain gasoline benzene levels above the
standard would have to purchase benzene credits generated by refineries for early reduction
efforts and/or overcompliance with the standard.

6.5.2   Refinery Compliance Strategies

       As discussed in  Chapter 9, our cost analysis assumes that refiners would choose the most
economical strategy for complying with the gasoline benzene standard in 2011  and beyond.  We
predict that the majority of refinery compliance strategies would involve making at least some
sort of process change to reduce benzene levels. For some refineries, it is economical to reduce
gasoline benzene levels to < 0.62 vol%, while for others it is more economical to make
incremental reductions  in gasoline benzene level to > 0.62 vol% and rely partially upon benzene
credits. For the refineries whose compliance strategies do not involve reducing benzene levels,
most are already below the standard so no further action is required. For the remaining
refineries, it is more economical to rely solely upon credits than to make any process
improvements to reduce gasoline benzene.  A summary the model-predicted refinery compliance
strategies are presented in Table 6.5-2.
                                          6-37

-------
                  Table 6.5-2. Predicted Refinery Compliance Strategies
Refinery Compliance Strategy
Make process improvement to reduce
gasoline benzene levels?
Yes, reduce Bz levels to <= 0.62 vol%
Yes, reduce Bz levels to > 0.62 vol%
No, Bz levels are already <= 0.62 vol%
No, maintain Bz levels > 0.62 vol%
Rely on
Credits?
No
Yes
No
Yes
Total Number of Refineries
No. of Refineries by PADD
PADD1
4
4
4
0
12
PADD 2
7
18
0
1
26
PADD 3
23
8
7
3
41
PADD 4
1
14
1
0
16
PADD 5*
2
5
0
1
8
CA
2
0
7
3
12
Total
39
49
19
8
115
    Refers to PADD 5 excluding the State of California
6.5.3   Benzene Reduction Strategies

       We believe that most refiners planning on reducing gasoline benzene levels would focus
on reformate control, since the majority of the benzene found in gasoline comes from the
reformer as explained in 6.3.1. We predict most refiners would choose this strategy since it is
capable of getting the greatest benzene reductions and the technology is known and readily
available. For our ABT analysis, we have specifically focused on the following forms of
reformate control: light naphtha splitting, isomerization, benzene extraction, and benzene
saturation. These technologies are discussed in more detail in 6.3.2.

       Our refinery cost model predicts which benzene reducing step(s) each individual refinery
would take based on the lowest overall cost strategy to meet the proposed 0.62 vol% standard
nationwide.  The benzene control strategy a refinery selects depends on existing equipment,
proximity to the petrochemical s market, and technology costs compared to the cost of buying
credits.  The cost model also contains estimates of the timing necessary for each refinery to make
the predicted refinery process changes.  A refinery's ability to make benzene reductions earlier
than required is dependent on the nature of the improvement(s), required planning time, and
associated capital  costs.

6.5.3.1 Early Process Changes Completed Prior to January 1, 2011

       In many cases there are benzene reductions strategies consistent with refineries' overall
compliance strategies that could be implemented earlier than required. To encourage early
introduction of benzene control technology, we are proposing that refiners could generate early
benzene credits from June 1, 2007 to December 31, 2010 by making qualifying reductions from
their pre-determined refinery baselines.  A discussion of how refinery baselines are established
and what constitutes a qualifying benzene reduction is found in the paragraphs to follow.

       The early reductions we are predicting to occur would be consistent with  each refinery's
ultimate benzene control strategy but simply completed sooner than required. As discussed in
the subsections that follow, we predict that prior to January 1, 2011, refiners could implement
operational changes and/or make small capital investments to reduce gasoline benzene.  These
                                          6-38

-------
actions would create a two-step phase down in gasoline benzene levels prior to 2011 as shown in
Figure 6.5-1. The early credits generated could be used to postpone refiners' final, most
expensive, benzene control technology investments.

                Figure 6.5-1. ABT Program with Early Credit Generation
                                  Benzene Level vs. Time
     1.00
     0.95 -
     0.90 -
  0.85 -
   2! 0.80
   o
S 0.75 -
0)
c
HI
£ 0.70 -
0)
m

< 0.65


  0.60 -



  0.55 -
     0.50
            Early       v
            Operational    •
            Changes
                             Early Small
                             Capital
                             Investments
          2005    2006    2007   2008   2009    2010    2011    2012   2013   2014    2015   2016   2017
                                                Year
                                  •No Early Credit Program —•— Early Credit Program
       Early Operational Changes

       We estimate that the first phase of early benzene reductions could occur as early as June
1, 2007 after the rule is signed, published, and congressional review is complete.  These refinery
modifications would consist of operational changes made to the reformer that could be
implemented with virtually no capital investment.  The early operational changes we predict to
occur are light naphtha splitting and isomerization. For refineries that already have light naphtha
splitters in place, we assume that operational changes could be made to re-route up to 75% of the
benzene precursors around the reformer.  If the refinery is equipped with an isomerization unit,
we predict that this re-routed light naphtha would also be isomerized.  If no isomerization unit
exists, we predict that the light naphtha would simply be combined with the light straight run to
make gasoline.

       Based on our refinery cost model, we predict that 48 of the 115 U.S. refineries would
take advantage of the early credit opportunity and make the early operational changes described
                                           6-39

-------
above.  These operational changes would result in an overall 13% reduction in gasoline benzene
levels from 0.97 vol% to 0.84 vol%.  The changes would also result in an overall 28% reduction
in benzene level variation from 3.39 vol% to 2.43 vol%.  A summary of these reductions and
resulting benzene levels are found in Table 6.5-4.

PADD1
PADD2
PADD3
PADD4
PADD 5****
California
Total
Table 6.5-4. Benzene Levels after Early Operational Changes
No. of Refineries by Gasoline Benzene Level (vol%)
<0.5
4
0
4
0
0
0
8
0.5-<1.0
4
13
21
2
0
12
52
1.0X1.5
2
11
12
10
3
0
38
1.5-<2.0
0
1
3
3
2
0
9
2.0-<2.5
2
0
0
0
3
0
5
>=2.5
0
1
1
1
0
0
3
Benzene Level (vol%)*
MIN
0.41
0.56
0.41
0.60
1.01
0.51
0.41
MAX
2.19
2.85
2.71
2.51
2.19
0.77
2.85
RANGE**
1.77
2.28
2.30
1.91
1.18
0.26
2.43
AVG***
0.61
0.99
0.80
1.27
1.57
0.63
0.84
    * Starting benzene levels based on summer 2003 batch data
    ** Range in benzene level (MIN-MAX)
    *** Average volume-weighted benzene level
    **** PADD 5 excluding California
       Early Technology Changes Requiring a Small Capital Investment

       We estimate that a second phase of early benzene reductions would occur 2-3 years after
the rule is signed or by about the end of 2009.  These refinery modifications would consist of
upgrades in reformate benzene control technology which require a relatively small capital
investment. For the purpose of this analysis, the refinery cost model defines a small capital
investment as investments that cost up to $8MMD. The early technology changes we predict to
occur include light naphtha splitting, isomerization, and benzene extraction.  For refineries that
already have light naphtha splitters in place or those that do not, we assume that technological
upgrades could be made to re-route 100% of the benzene precursors around the reformer. As
with the operational changes mentioned above, if the refinery is equipped with an isomerization
unit, we predict that the re-routed light naphtha would also be isomerized.  If no isomerization
unit exists, we predict that the light naphtha would be combined with the light straight run to
make gasoline. We also predict that refineries currently extracting benzene could make
modifications to their existing extraction units (up to $8MM) to improve the benzene separation
and in turn reduce the concentration of benzene in the final gasoline product.

       Based on our refinery cost model, we predict that 55 of the 115 U.S. refineries would
make early technology changes which require a small capital investment.  These changes along
with the operational changes discussed above would result in an overall 22% reduction in
gasoline benzene levels from 0.97 vol% to 0.76 vol%.  These changes would also result in an
overall 51% reduction in benzene level variation from 3.39 vol% to 1.67 vol%.  A summary of
these reductions and resulting benzene levels are found in Table 6.5-5.
       D At a revamped extraction unit cost of $8MM and above, the investment was judged to be sufficiently
complicated that the revamp would require the full lead time period to complete. Revamping an extraction unit can
be complicated because they are comprised of several major refinery units combined together and all of them could
require a significant revamp above the identified investment cost threshold.
                                           6-40

-------

PADD1
PADD2
PADD3
PADD4
PADD 5****
California
Total
rable 6.5-5. Benzene Levels after Early Small Capital Investments
No. of Refineries by Gasoline Benzene Level (vol%)
<0.5
4
1
7
0
0
0
12
0.5-<1.0
4
21
21
6
1
12
65
1.0X1.5
2
2
11
9
4
0
28
1.5-<2.0
1
2
1
1
3
0
8
2.0-<2.5
1
0
1
0
0
0
2
>=2.5
0
0
0
0
0
0
0
Benzene Level (vol%)*
MIN
0.41
0.49
0.41
0.60
0.81
0.51
0.41
MAX
2.09
1.95
2.07
1.94
1.84
0.77
2.09
RANGE**
1.67
1.46
1.65
1.34
1.04
0.26
1.67
AVG***
0.58
0.79
0.75
1.09
1.48
0.63
0.76
    * Starting benzene levels based on summer 2003 batch data
    ** Range in benzene level (MIN-MAX)
    *** Average volume-weighted benzene level
    **** PADD 5 excluding California
       What factors impact refiners' decisions to make early process changes?

       As mentioned before, a refinery's ability to make early benzene reductions depends on
the nature of the improvement(s), required lead time, and associated capital costs. However, a
refinery's decision to make early improvements depends on the trigger point and the company's
need for early credits.  Our ABT analysis assumes that refiners would only make reductions
predicted by the refinery cost model early if both of the following conditions were satisfied:

   1.  The reduction was significant enough to allow them to generate early credits.  A refiner
       would not make a model-predicted early benzene reduction if it did not satisfy the 10%
       reduction trigger point (discussed in more detail in the sections to follow). Applying this
       assumption reduced the number of predicted early operational changes from 58 to 49 and
       the number of early small capital investments from 61 to 56.

   2.  The company had a need for early credits because their average starting benzene
       concentration was higher than the standard. To prove this point, consider the  opposite.  If
       a company's average benzene level was 0.62 vol% or lower to begin with, they would not
       have a need to generate early credits to postpone compliance since they could do nothing
       and comply with the standard in 2011 via company averaging. Applying this  assumption,
       one refinery which  the model predicted to make both operational and small  capital
       investments was assumed not to do so early. This further reduced the number of early
       operational changes from 49 to 48 and the number of early small capital investments
       from 56 to 55.

For refiners whose decision to make early reductions was impacted by these two provisions, our
ABT analysis assumes that the model-predicted benzene reductions would eventually occur, just
not earlier than required.

       How are early credits calculated?

       Before we can calculate early credits we must first explain how early credit baselines and
                                          6-41

-------
annual average benzene levels are computed as well as how the proposed trigger point would
impact credit generation.  Additionally, we will explain the assumptions made to perform this
preliminary ABT analysis.

       We are proposing that any refiner planning on making early reductions establish
individual refinery benzene baselines in order to provide a starting point for early credit
calculations. Refinery baselines would be defined as the annualized volume-weighted benzene
content of gasoline produced at a refinery from January 1, 2004 to December 31, 2005.  For the
purpose of this ABT analysis, we used the summer 2003 starting gasoline benzene levels
reflected in Table 6.5-1 to represent refinery baselines.

       The benzene level from which early credits are calculated is the average volume-
weighted benzene concentration of all batches  of gasoline produced during a given averaging
period.  This is referred to as the annual average benzene concentration. For the purpose of this
ABT analysis,  we have used the benzene levels predicted by the refinery cost model to represent
annual average benzene levels. For 2007, 2008, and 2009, we have used the post-operational
change benzene levels reflected in Table 6.5-4. For 2010, we have used the benzene levels
following the early small  capital investments reflected in Table 6.5-5.

       In order to qualify to generate early credits, refiners would first need to reduce gasoline
benzene levels to 0.90 times their refinery benzene baseline during a given averaging period.  A
further explanation of how we arrived at the 10% reduction trigger point can be  found in
subsections to follow. Once the 10% reduction trigger point was met, refineries could generate
early credits based on the entire benzene reduction. For example, if in 2008 a refinery reduced
its annual benzene level from a baseline of 2.00 vol% to 1.50 vol% (below the trigger of 0.90 x
2.00 = 1.80 vol%), its benzene credits would be determined based on the difference in annual
benzene content (2.00 - 1.50 = 0.50 vol%) divided by 100 and multiplied by the gallons of
gasoline produced in 2008 (credits expressed in gallons of benzene).

       How many early credits does our refinery cost model predict?

       By applying these criteria to the refinery cost model, we estimate that refineries making
early operational changes and small capital investments in reformate technology from June 1,
2007 to December 31, 2010 would generate over 650 million gallons of early benzene credits as
shown in Table 6.5-6.
                                          6-42

-------
                     Table 6.5-6. Early Credits Generated by PADD

PADD 1
PADD 2
PADD 3
PADD 4
PADD 5*
California
Total
Early Credits Generated by Year (gal Bz)
2007
1,276,497
53,145,796
16,919,006
7,512,220
12,361,833
0
91,215,351
2008
2,188,280
91,107,079
29,004,010
12,878,091
21,191,714
0
156,369,173
2009
2,188,280
91,107,079
29,004,010
12,878,091
21,191,714
0
156,369,173
2010
6,143,596
148,719,615
57,451,088
20,115,709
25,268,439
0
257,698,447
Total
11,796,653
384,079,568
132,378,113
53,384,110
80,013,701
0
661,652,145
    Refers to PADD 5 excluding the State of California
       How much lead time would be generated by early credits?

       Under the proposed ABT program, we assume that early credits generated prior to 2011
could be used to provide refineries with additional lead time to postpone their final investments
in benzene control technology. This would essentially postpone the full implementation of the
0.62 vol% benzene standard by a certain period of time, providing a more gradual phase-in of the
standard.

       To calculate the potential "lag"  in compliance, we first calculated the demand for early
credits by refineries which the cost model predicted would still be above the 0.62 vol% standard
in 2010 after the early small capital investment period. This included refineries which the cost
model predicted to make future investments as well as those predicted to rely on credits as part
of their ongoing compliance strategy.

       The early credit demand was calculated individually for each refinery above the standard
as demonstrated in the following example.  If in 2010 a refinery's annual average benzene level
was 0.80, it's early credit demand would be determined based on the difference between the
annual benzene level and the standard (0.80 - 0.62 = 0.18 vol%) divided by 100 and multiplied
by it's annual average production volume (early credit demand expressed in gallons of benzene
per year).  The total early credit demand by PADD is found in Table 6.5-7.
                                          6-43

-------
                    Table 6.5-7.  Demand for Early Credits by PADD

PADD 1
PADD 2
PADD 3
PADD 4
PADD 5*
California
Total
Refineries with Bz Levels >0.62 vol% in 2010
Total
Number of
Refineries
7
20
26
15
8
4
80
Average
Benzene
Level (vol%)
0.83
0.84
0.99
1.13
1.48
0.74
0.93
Gasoline
Production
(MMgal/yr)
5,394
22,566
28,791
3,550
4,341
7,073
71,716
Early Credit
Demand
(gal Bz/yr)
11,176,350
49,124,851
94,888,243
18,190,371
37,276,799
8,139,253
218,795,867
            Refers to PADD 5 excluding the State of California
       Finally, the length of the early credit lag was computed as the total number of early
credits generated (661,652,145 gal Bz) divided by the early credit demand (218,795,867 gal/yr).
 The lag was found to be 3.02 years which could postpone compliance with the 0.62 vol%
standard from 2011 to 2014 as shown in Figure 6.5-1. Based on this theoretical early credit lag,
a matching 3-year early credit life was proposed.

       What is the value of the proposed early credit program?

       Not only does the early credit program result in sooner benzene emission reductions for
the environment, it also results in a cost savings to the refining industry. With no early credit
program, all refiners would implement their benzene control strategies around the same time
causing a sharp $168 million increase in compliance costs in 2011 (annualized capital plus
operating costs).  With the early credit program, refineries would have incentive to implement
some of their technologies sooner. The early  credits generated could be used to delay final
investments as much as three years, as calculated above and allowed by the three-year  early
credit expiration date. This would spread out industry-wide demand for recourses and  total
compliance costs over time. This gradual phase in of costs is represented in Figure 6.5-2 and
would result in a net savings of $86 million to the refining industry during the 2007-2014 period.
 This net cost savings has been computed as the difference between the areas under the curves.
                                          6-44

-------
                Figure 6.5-2. ABT Program with Early Credit Generation
                	Annual Compliance Costs vs. Time	
     $180
     $160 -
  _ $140 -
     $120 -
     $100 -
   c
   .5
   "5.
   o
   o
     $40 -
     $20 -
            2006
                     2007
                             2008
                                      2009
                                                        2011
                                                                 2012
                                                                          2013
                                                                                   2014
                                •No Early Credit Program
• Early Credit Program
       Early Credit Trigger Points

       What is the purpose of an early credit trigger point?

       In order to qualify to generate early credits, refiners would first need to reduce gasoline
benzene levels to 0.90 times their refinery benzene baseline during a given averaging period.
The purpose of setting an early credit generation trigger point is to ensure that changes in
benzene level are representative of real process improvements. Without a trigger point,
refineries could generate credits based on operational fluctuations in benzene level from year to
year.  This would compromise the environmental benefits of an ABT program because the early
credits generated would have no associated benzene emission reduction value.

       What trigger points did we consider?

       In designing the early credit generation program, we considered a variety of different
types of trigger points. We performed sensitivity analyses around absolute level trigger points
(refineries must reduce gasoline benzene levels to a certain concentration in order to generate
credits), fixed reduction trigger points (refineries must reduce gasoline benzene levels by a
certain concentration in order to generate credits), and percent reduction trigger points (refineries
                                            6-45

-------
must reduce gasoline benzene by a percentage in order to generate). The results of these analyses
are found in Table 6.5-8, Table 6.5-9, and Table 6.5-10, respectively. For comparison purposes,
we have focused on trigger points resulting in an approximate three-year early credit lag.
           Table 6.5-8.  Absolute Level Trigger Point (ALTP) Credit Generation
Absolute Level
Trigger Point
(vol%)
2.00
1.90
1.80
1.70
1.60
1.50
1.40
1.30
1.20
1.10
1.00
Early Credit Generation by Starting* Bz Level (vol%)
"0.5 to <1
57,435,070
57,435,070
57,435,070
57,435,070
57,435,070
57,435,070
57,435,070
57,435,070
57,435,070
57,435,070
57,435,070
1 to<1.5
212,079,916
212,079,916
212,079,916
212,079,916
212,079,916
212,079,916
212,079,916
212,079,916
209,454,644
195,161,525
187,483,551
1.5to<2
290,561,782
290,561,782
290,561,782
290,561,782
290,561,782
290,561,782
241 ,777,402
207,685,666
206,244,587
172,872,517
113,702,251
2 to <2.5
75,954,122
75,842,055
75,842,055
75,842,055
50,651,118
47,932,394
28,052,007
18,460,791
2,977,994
2,977,994
1,943,107
>=2.5
29,226,711
26,344,612
19,571,551
6,267,344
6,267,344
6,267,344
1,045,758
1,045,758
1 ,045,758
1 ,045,758
0
TOTAL
665,257,600
662,263,434
655,490,374
642,186,166
616,995,229
614,276,505
540,390,152
496,707,200
477,158,052
429,492,864
360,563,979
Early Credit
Lag
(Years)
3.09
3.04
2.94
2.83
2.64
2.63
2.15
1.89
1.73
1.51
1.23
    * Starting benzene levels based on summer 2003 batch data
    ** Model does not predict any early credits to be generated by refineries with starting benzene levels <0.5 vol%
       As shown in Table 6.5-8, for a 1.90 vol% absolute level trigger point (ALTP), the
number of early credits generated by refineries with starting benzene levels >2.5 vol% is 26
million.  This is about half the amount of early credits generated by the same group of refineries
under the proposed 10% reduction trigger point (51 million).  In addition, early credit generation
is reduced to zero as the absolute level trigger point decreases. As such, we conclude that
absolute level trigger points are too restrictive towards refineries with high starting benzene
levels. It is important not to restrict early credit generation  for this class of refineries because
they could arguably benefit the most from early reductions. They have the highest starting
benzene levels and thus the greatest need for real benzene reductions.  They would also have the
greatest amount of work to do to meet the 0.62 vol% standard, so they could benefit significantly
from the additional lead time provided by early credits. The lead time could be used to spread
out subsequent benzene technology investments making compliance with the benzene  standard
more affordable.  Another disadvantage of an ALTP is that  there could potentially be a
"windfall" of early credits generated by refineries with starting benzene levels near the trigger
point.  For example a refinery with a starting benzene level  of 1.91 vol% could generate early
credits based on minor operation fluctuations in benzene level from year to year.  This would
essentially generate "artificial" credits with no associated benzene reduction value.
                                           6-46

-------
          Table 6.5-9. Fixed Reduction Trigger Point (FRTP) Credit Generation
Fixed
Reduction
Trigger Point
(vol%)
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.55
0.60
Early Credit Generation by Starting* Bz Level (vol%)
**0.5 to <1
49,322,559
39,520,923
30,425,825
20,941,241
15,524,718
7,727,474
7,727,474
0
0
0
0
0
1 to <1.5
211,538,905
211,113,794
198,861,358
175,558,970
173,501,315
172,244,773
170,093,278
161,526,161
155,290,562
124,921,489
107,289,504
59,186,172
1.5to<2
290,561,782
290,561,782
289,662,459
289,662,459
287,020,226
287,020,226
284,660,705
265,100,388
265,100,388
198,630,694
177,787,494
176,112,996
2 to <2.5
75,954,122
75,954,122
75,900,117
75,900,117
75,900,117
75,842,055
75,842,055
75,842,055
75,842,055
75,842,055
75,842,055
75,842,055
>=2.5
53,011,708
53,011,708
53,011,708
53,011,708
53,011,708
51,018,812
51,018,812
49,952,616
49,952,616
49,952,616
49,952,616
49,770,570
TOTAL
680,389,075
670,162,328
647,861 ,466
615,074,494
604,958,083
593,853,340
589,342,324
552,421,221
546,185,622
449,346,855
410,871,670
360,911,793
Early Credit
Lag
(Years)
3.15
3.08
2.93
2.62
2.57
2.45
2.41
2.10
2.06
1.58
1.41
1.23
    * Starting benzene levels based on summer 2003 batch data
    ** Model does not predict any early credits to be generated by refineries with starting benzene levels <0.5 vol%

       As  shown in Table 6.5-9, for a 0.10 vol% fixed reduction trigger point (FRTP), the
number of early credits generated by refineries with starting benzene levels <1 vol% is under 40
million.  Not only does this trigger point generate less credits than the 10% reduction trigger
point (42 million), early credit generation is reduced to zero as the fixed reduction trigger point
increases.  Fixed reduction trigger points are biased towards refineries with higher starting
benzene levels because it is easier for them to achieve a fixed reduction than it is for a lower
benzene level refinery to achieve the same reduction.  Therefore, we conclude that fixed
reduction trigger points are too restrictive towards refineries with low starting benzene levels.
We do not feel that these innovative refineries should be penalized for already being "cleaner".
         Table 6.5-10. Percent Reduction Trigger Point (PRTP) Credit Generation
Percent
Reduction
Trigger Point
(%)
5%
10%
15%
20%
25%
30%
35%
Early Credit Generation by Starting* Bz Level (vol%)
**0.5 to <1
44,888,175
42,364,574
33,656,028
25,559,561
20,941,241
15,524,718
10,523,099
1 to <1.5
211,538,905
202,706,184
190,891,588
173,501,315
172,244,773
159,933,137
147,465,199
1.5to<2
290,561,782
289,662,459
287,020,226
284,660,705
265,100,388
183,845,616
163,978,824
2 to <2.5
75,954,122
75,900,117
75,842,055
75,842,055
75,842,055
50,651,118
28,052,007
>=2.5
53,011,708
51,018,812
49,952,616
49,952,616
49,770,570
49,770,570
23,157,227
TOTAL
675,954,691
661,652,145
637,362,514
609,516,253
583,899,027
459,725,159
373,176,357
Early Credit
Lag
(Years)
3.10
3.02
2.87
2.60
2.40
1.65
1.27
    * Starting benzene levels based on summer 2003 batch data
    ** Model does not predict any early credits to be generated by refineries with starting benzene levels <0.5 vol%
       As shown in Table 6.5-10, a 10 percent reduction trigger point (PRTP) tends to moderate
credit generation better than the absolute level and fixed reduction trigger points we have
                                            6-47

-------
considered.  This is especially true for the extreme cases where refinery starting benzene levels
are <1 vol% or > 2.5 vol%.  For the 47 refineries with starting benzene levels < 1 vol %, a 10
PRTP generates 42 million credits which is more than a 0.10 vol% FRTP (40 million) but less
than a 1.90 ALTP (58 million). For the 7 refineries with starting benzene levels > 2.5 vol%, a 10
PRTP generates 51 million credits which is less than a 0.10 vol% FRTP (53 million) but more
than a 1.90 ALTP (26 million). As such, we concluded that a percent reduction trigger point
would be the most appropriate early credit validation tool to address the wide range in starting
benzene levels.

       How did we decide on a value for the trigger point?

       Once we decided that a percent reduction trigger point (PRTP) was the  most suitable type
of early credit trigger point, the next step was to determine the optimum value for the trigger
point. In assessing the appropriate PRTP value, there were two main objectives. The first was to
set a trigger point that was stringent enough to require refineries to make real improvements in
benzene control technology in order to generate credits. A less stringent trigger point could
potentially allow refineries to generate artificial or "windfall" credits based on  normal
operational fluctuations in gasoline benzene level from year to year.  The second objective was
to ensure that the trigger point was not too stringent as to discourage refiners from making early
reductions in gasoline benzene. As mentioned in 6.2.2.9.3.1.3, we predict that  refiners would not
make reductions in gasoline benzene earlier than required if the trigger point was credit
prohibitive.  Accordingly, the closer the trigger point was to corresponding with real achievable
benzene reductions, the more refineries would pursue making early process improvements.  As
such, a carefully selected early credit trigger point would enhance early credit generation and
result in a more reliable market for trading.

       To make an educated decision on the most appropriate trigger point, we evaluated the
model-predicted early benzene reductions and compared them to the "normal"  year-to-year
variation in refinery benzene level.  We started by examining the benzene reductions resulting
from our model-predicted refinery process changes.  Our model predicts that some refiners  could
make early improvements in reformate benzene control  technology resulting in 2-70% benzene
reductions.  This indicates that any trigger point above 2% would restrict early  credit generation
to some degree. As such, based on credit generation alone, we would want to choose the lowest
possible trigger point.  However, if we were to choose a 2% trigger point, the potential for
refineries to generate "windfall" credits would be high.  To get a better understanding of how
gasoline benzene levels currently fluctuate from year to year, we reviewed the 2002-2004 batch
reports required under the RFG/antidumping regulations.  As a reference point, we chose to use
the 2002-2003 calendar years as the baseline period, along the same lines as the two-year early
credit baseline provision in this proposal. From there, we calculated each refinery's change in
benzene level in 2004 compared to their baseline. Changes in refinery benzene level were found
to range from 42% (net decrease in benzene level) to  -48% (net increase in benzene level).  From
here, we chose to focus our analysis on only those refineries which made reductions in benzene
levels, since that is how early credits would be generated under the proposed ABT program.
Refineries' 2004 benzene reductions ranged from 0.28 to 42% percent with an  average refinery
reduction of 11.4%.  Based on this limited data, to eliminate any chances of "windfall" credit
generation we considered a trigger point on the magnitude of 40%. However, as shown in Table

                                          6-48

-------
6.2.2.9-11, this would have a detrimental effect on refiners' decisions to make early process
changes and resulting ability to generate early credits
         Table 6.5-11. Impact of Trigger Point Value on Early Reductions/Credits
Percent Reduction
Trigger Point
(%)
0%
5%
10%
15%
20%
25%
30%
35%
40%
Affect on Refineries Early Process Changes
Operational
Changes
57
53
48
41
38
37
29
19
8
Small
Technology
Changes
60
58
55
52
49
44
41
37
34
Total Early
Changes
117
111
103
93
87
81
70
56
42
%
Reduction
N/A
5%
12%
21%
26%
31%
40%
52%
64%
Affect on Early Credits
Early Credits
(gal Bz)
682,596,896
675,954,691
661,652,145
637,362,514
609,516,253
583,899,027
459,725,159
373,176,357
222,727,472
%
Reduction
N/A
1%
3%
7%
11%
14%
33%
45%
67%
       As shown in Table 6.5-11, as the value of the trigger point increases from 0% (no trigger
point) to 40%, the number of refinery-predicted process changes decreases from 117 to 42 by
64%.  Accordingly, the number of early credits generated drops drastically by 67% compared to
unrestricted credit generation. The proposed 10 PRTP roughly coincides with the average
fluctuation in benzene level from 2002/2003 to 2004 and is also the same as that finalized in the
Tier 2 gasoline sulfur rulemaking.  In response to this competing relationship between windfall
credits and early credits, we are proposing a 10% reduction trigger point because it strikes a
balance that errs of the side of encouraging early credit generation.

6.5.3.2 Final  Process Changes Requiring a Large Capital Investment

       We estimate that the final phase of benzene reductions would begin in 2011. This phase
of refinery upgrades would include modification or installation of some of the more expensive
reformate  control technologies - benzene extraction and benzene saturation. For refineries
pursuing benzene extraction, this would include upgrades in existing benzene extraction units
exceeding $8MM and  installation of new benzene extraction units. This would also include
installation of new benzene saturation units. Finally, this phase of refinery improvements would
also include small capital investments that were predicted to occur early but were postponed
based on the value of the trigger point.

       Based on our refinery cost model, we predict that 33 of the 115 U.S. refineries would
make technology improvements at this time. More specifically, 16 refineries would pursue
extraction and 11 refineries would pursue benzene saturation requiring a large capital
investment. Additionally, 6 refineries would pursue light naphtha splitting, isomerization, or
extraction requiring a small capital investment that were postponed based on lack of early credit
                                          6-49

-------
incentives. These final refinery technology upgrades would be completed in 2011 or up to three
years later as early credits permit.  These 33 total technology changes would result in an overall
36% reduction in gasoline benzene levels from 0.97 vol% to 0.62 vol%. The changes would also
result in an overall 50% reduction in benzene level variation from 3.39 vol% to 1.71 vol%.  A
summary of these reductions and resulting benzene levels are found in Table 6.2.2.9-12.
          Table 6.5-12. Benzene Levels after Final Capital Investments by PADD

PADD1
PADD 2
PADD 3
PADD 4
PADD 5****
California
Total
No. of Refineries by Gasoline Benzene Level (vol%)
<0.5
4
1
10
0
0
2
17
0.5-<1.0
5
22
27
8
4
10
76
1.0X1.5
1
1
3
7
2
0
14
1.5-<2.0
2
2
0
1
2
0
7
2.0-<2.5
0
0
1
0
0
0
1
>=2.5
0
0
0
0
0
0
0
Benzene Level (vol%)*
MIN
0.41
0.49
0.36
0.53
0.54
0.46
0.36
MAX
1.96
1.95
2.07
1.94
1.84
0.77
2.07
RANGE**
1.54
1.46
1.71
1.40
1.30
0.31
1.71
AVG***
0.51
0.73
0.55
0.95
1.04
0.60
0.62
    ' Starting benzene levels based on summer 2003 batch data
    '* Range in benzene level (MIN-MAX)
    '** Average volume-weighted benzene level
    '*** PADD 5 excluding California
6.5.4   Ending Gasoline Benzene Levels

   As summarized in Table 6.5-12, after full implementation of the program, the benzene
content of gasoline produced by the 115 U.S. refineries would range from 0.36 to 2.07 vol% with
an overall volume-weighted average of 0.62 vol%.

6.5.5   Standard Credit Generation Opportunities

       We are proposing that benzene credits (referred to hereafter as standard credits) could be
generated by any refinery or importer that overcomplies with the 0.62 vol% gasoline benzene
standard on an annual volume-weighted basis in 2011 and beyond.

       The refinery cost model discussed further in Chapter 9, predicts which refineries would
reduce benzene levels in an order of precedence based on cost until the 0.62 vol% refinery
average standard is achieved.  Accordingly, the model predicts which refineries would
overcomply with the standard in 2011 and beyond and in turn generate standard credits. Credits
would be generated by two main sources.

       First, standard credits would be generated by refineries whose current gasoline benzene
levels are already below the 0.62 vol% standard. According to the model, 19 refineries are
predicted to maintain current gasoline benzene levels and overcomply with the standard without
making any additional process improvements. These refineries would generate approximately
42 million gallons of benzene credits per year without making any investment in technology.
Additionally, the model predicts that 5 other refineries would reduce gasoline benzene levels
even further below 0.62 vol% resulting in deeper overcompliance and an additional 6 million
                                          6-50

-------
gallons of benzene credits per year.

       Second, standard credits would be generated by refineries whose current gasoline
benzene levels are above 0.62 vol% but are predicted by the model to overcomply with the
standard based on existing refinery technology, liquid capital, and/or proximity to the benzene
chemical market.  The model predicts that 34 refineries with gasoline benzene levels above 0.62
vol% would make process improvements to reduce benzene levels below the standard and in turn
generate approximately 40 million gallons of benzene credits per year.

       For the refineries which the model predicts to make process changes to overcomply with
the standard, the incremental cost to overcomply is relatively small or even profitable in some
cases of benzene extraction. As expected, refineries with the lowest compliance costs would
have the greatest incentive to overcomply based on the value of the credits to the refining
industry.

6.5.5.1 How are  Standard Credits Calculated?

       We are proposing that benzene credits could be generated by  any refinery or importer
that overcomplies with the 0.62 vol% gasoline benzene standard on an annual volume-weighted
basis in 2011 and beyond. For example, if in 2011 a refinery's annual average benzene level was
0.52, its benzene credits would be determined based on the margin of overcompliance with the
standard (0.62  - 0.52 = 0.10 vol%) divided by 100 and multiplied by the gallons of gasoline
produced during the 2011 calendar year (credits expressed in gallons of benzene). Likewise, if
in 2012 the same refinery produced the same amount of gasoline with the same benzene content
they would earn the same amount of credits.  The credit generation opportunities for
overcomplying with the standard would continue indefinitely.

6.5.5.2 How Many Standard Credits would be Generated in 2011 and Beyond?

   As mentioned above, standard credits would be generated beginning January  1, 2011 by
refineries that overcomply with the 0.62 vol% standard on an annual, volume-weighted basis.
According to our refinery cost model we predict that approximately 88 million would be
generated in 2011 and indefinitely thereafter as summarized in Table 6.5-13.
                                          6-51

-------
           Table 6.5-13. Standard Credits Generated/Needed in 2011 & Beyond

PADD1
PADD2
PADD3
PADD4
PADD 5*
California
Total
Standard Credits Generated
by Refineries < 0.62 vol%
(gal/yr)
21,069,691
4,997,840
50,492,943
347,760
820,766
10,102,342
87,831,343
Standard Credits Needed** by
Refineries > 0.62 vol%
(gal/yr)
3,033,093
34,592,643
11,785,856
12,939,012
18,884,725
6,596,015
87,831,343
           Refers to PADD 5 excluding the State of California
           **After early credit lag
   As shown in Table 6.5-13, PADDs 1 and 3 would have the highest annual standard credit
generation. That is because refineries in these geographic regions are located in close proximity
to the petrochemicals market making benzene extraction (resulting in very low benzene levels) a
viable compliance strategy.

6.5.6   Credit Use

       We are proposing that refiners and importers could use benzene credits generated or
purchased under the provisions of the ABT program to comply with the 0.62 vol% gasoline
benzene standard in 2011 and indefinitely thereafter.  All credits are to be used towards
compliance on a one-for-one basis, applying each benzene gallon credit to offset the same
volume of benzene produced in gasoline above the standard.  For example, if in 2011 a refinery's
annual average benzene level was 0.72, the number of benzene credits needed to comply would
be determined based on the margin of under-compliance with the standard (0.72 - 0.62 = 0.10
vol%) divided by 100 and multiplied by the gallons of gasoline produced during the 2011
calendar year (number of credits expressed in gallons of benzene).

   Early credits may be used equally and interchangeably with standard credits to comply with
the 0.62 vol% benzene  standard in 2011 and beyond.  However, based on the credit life
provisions described further in 6.2.2.9.6.2.2, we predict that refiners would chose to use early
credits first before relying on standard credits.  By the beginning of 2014, or once all early
credits have been used, terminated, or become otherwise unavailable, we predict that refiners
would begin relying solely on standard credits.  Our refinery cost model projects that at this
point the  credit supply produced by refineries that overcomply with the standard would be
sufficient to meet the credit demand of refineries that under-comply with the standard.  The
ongoing credit demand would be approximately 88 billions gallons of benzene credits per year
which equals the supply as shown in Table 6.5-12.

6.5.6.1 Credit Trading Area
                                          6-52

-------
       We are proposing a nationwide credit trading area.  We have not placed any geographic
restrictions on where credits may or may not be traded.  If PADD restrictions were placed on
credit trading, there would be an imbalance between the supply and demand of credits.  If there
was no inter-PADD trading allowed, PADDs 1 and 3 would have a surplus of standard credits
while PADDs 2, 4, and 5 would have a shortage of credits as shown in Table 6.5-12.  This would
result in surplus credits expiring and becoming worthless in PADDs 1 and 3 while at the same
time PADDs 2, 4, and 5 would experience insufficient credit availability. This would force
refineries with more expensive benzene technology costs in PADDs 2, 4, and 5 to comply
increasing total compliance costs.  Overall, restricting  credit trading by PADD would result in a
more expensive, less  flexible, and less efficient  program.

       Additionally,  we believe that restricting  credit trading could reduce refiners' incentive to
generate credits and hinder trading essential to this program. In other fuel standard ABT
programs (e.g., the highway diesel sulfur program), fuel credit trading restrictions were
necessary to ensure there was adequate low-sulfur fuel available in each geographic area to meet
the corresponding vehicle standard.  Since there is no vehicle emission standard being proposed
that is dependent on gasoline benzene content, we do not believe there is a crucial need for
geographic trading restrictions. We project that under the proposed nationwide ABT program,
all areas of the country (PADDs) would still experience large reductions in gasoline benzene
levels as shown in Table 6.5-14.

            Table 6.5-14. Total Percent Reductions in Benzene Level by PADD

PADD1
PADD 2
PADDS
PADD 4
PADD 5***
California
Total
Starting*
Benzene
Levels (vol%)
0.62
1.32
0.86
1.60
2.06
0.63
0.97
Ending**
Benzene
Levels (vol%)
0.51
0.73
0.55
0.95
1.04
0.60
0.62
% Reduction
in Benzene
Level
18.82%
44.92%
36.19%
41.12%
49.69%
4.80%
36.03%
              * Starting benzene levels based on summer 2003 batch data
              ** Ending benzene levels based on model-predicted benzene reductions
              *** PADD 5 excluding California
6.5.6.2 Credit Life

       We are proposing that early credits generated prior to 2011 would have a three-year
credit life from the start of the program. In other words, early credits would have to be applied
to the 2011, 2012, and/or 2013  compliance years or they would expire.

       We are proposing that standard credits generated in 2011 and beyond would have to be
used within five years of the year in which they were generated.  If standard credits were traded
                                          6-53

-------
to another party they would still have to be used during the same five-year period.  In other
words, standard credit life would be tied to time of generation, not the time of transfer. Standard
credits not used within five years would expire.

       These proposed credit life provisions are similar to those finalized in gasoline sulfur
program, except the early credit life is three years instead of two. This three-year early credit
expiration period corresponds with the early credit lag calculated above in Section 6.5.3.1.
Additionally, we believe that three years would be more than sufficient time for all early credits
generated to be utilized. We believe that this certainty that all credits could be utilized would
strengthen refiners' incentive to generate early credits and subsequently establish a more reliable
credit market for trading.

       In addition to the above-mentioned provisions, we are proposing that credit life may be
extended by two years  for early credits and/or standard credits traded to approved small refiners.
 We are offering this provision as a mechanism to encourage more credit trading to small
refineries. Small refiners are often technologically challenged, so they would tend to have more
of a need to rely on credits. At the same time, they have less business affiliations than other
refiners, so they could have difficulty obtaining credits. We believe this provision would be
equally beneficial to refiners generating credits. This additional credit life for credits traded to
small refiners would give refiners generating credits a greater opportunity to fully utilize the
credits before they expire. For example, a refiner who was holding on to credits for emergency
purposes or other reasons later found to be unnecessary, could trade these credits at the end of
their life to small refiners who could utilize them for two more years.

6.5.6.3 Credit Availability

       Our ABT analysis presented here assumes perfect nationwide credit trading.  In reality,
we  recognize that not all credits generated may necessarily be available for sale. Since EPA is
not proposing to manage the credit market, credit trading would be at the generating  refiners'
discretion. With such a program concerns are always expressed that credits may not be made
available on the market. This is always a concern of single refinery refiners.  To determine the
likelihood of credit availability, we have expressed credit generation and trading by company
using our refinery-cost model.  The results preserve refiner identity, are segregated by early
credits and standard credits, and are found in Tables 6.5-15 and 6.5-16, respectively.
                                           6-54

-------
                    Table 6.5-15. Early Credit Trading by Company

Company
Company 1
Company 2
Company 3
Company 4
Company 5
Company 6
Company 7
Company 8
Company 9
Company 10
Company 11
Company 12
Company 13
Company 14
Company 15
Company 16
Company 17
Company 18
Company 19
Company 20
Company 21
Company 22
Company 23
Company 24
Company 25
Company 26
Company 27
Company 28
Company 29
Company 30
Company 31
Company 32
Company 33
Company 34
Company 35
Total
Generation
(2007-2010)
0
103,072,091
32,759,678
15,613,470
0
54,779,242
7,674,171
9,823,659
12,246,166
4,729,316
10,345,379
112,371,363
2,659,661
5,197,754
17,329,072
26,996,329
3,093,255
14,858,489
2,700,053
61,377,633
96,304,724
7,686,770
1,388,498
58,061
3,361,260
3,590,867
13,304,208
13,443,033
2,166,784
12,607,342
0
0
0
1,034,887
9,078,930
661,652,145
Need
(3-Year Lag)
0
70,718,784
11,654,558
27,590,955
8,072,835
80,868,167
1,883,932
75,786,123
4,671,250
9,790,231
11,495,180
29,269,755
81,605,213
8,063,391
927,373
40,533,634
1,803,271
8,057,316
17,987,381
42,898,986
82,271,317
2,620,612
0
919,079
3,037,674
0
13,387,601
992,077
4,632,876
11,542,289
6,317,414
542,056
0
1,205,920
504,894
661,652,145
Net Early
Credits
0
32,353,307
21,105,120
-11,977,485
-8,072,835
-26,088,925
5,790,239
-65,962,464
7,574,916
-5,060,915
-1,149,801
83,101,608
-78,945,551
-2,865,637
16,401,699
-13,537,305
1,289,984
6,801,173
-15,287,328
18,478,647
14,033,407
5,066,158
1,388,498
-861,018
323,586
3,590,867
-83,393
12,450,955
-2,466,092
1,065,053
-6,317,414
-542,056
0
-171,034
8,574,036
0
% of Credit
Supply

13.51%
8.82%



2.42%

3.16%


34.71%


6.85%

0.54%
2.84%

7.72%
5.86%
2.12%
0.58%

0.14%
1.50%

5.20%

0.44%




3.58%
100.00%
% of Credit
Need



5.00%
3.37%
10.90%

27.55%

2.11%
0.48%

32.98%
1 .20%

5.65%


6.39%




0.36%


0.03%

1.03%

2.64%
0.23%

0.07%

100.00%
Credits Used
Internally
0
70,718,784
11,654,558
15,613,470
0
54,779,242
1,883,932
9,823,659
4,671,250
4,729,316
10,345,379
29,269,755
2,659,661
5,197,754
927,373
26,996,329
1,803,271
8,057,316
2,700,053
42,898,986
82,271,317
2,620,612
0
58,061
3,037,674
0
13,304,208
992,077
2,166,784
11,542,289
0
0
0
1,034,887
504,894
422,262,892
       As shown in Table 6.5-15, 17 of the 35 companies have the potential to generate more
early credits than they could use up in the three-year period allowed. The refinery concentration
of early credits ranges from <1% to 35%.  Consequently, there does not appear to be substantial
credit market concentration so there should be significant potential for the 16 refiners that seek
early credits to postpone future investments to find them.   Additionally, intra-company trading
accounts for approximately two thirds of all early credit trades which equates to a high
likelihood that the predicted transfers would actually occur.
                                          6-55

-------
                   Table 6.5-16. Standard Credit Trading by Company

Company
Company 1
Company 2
Company 3
Company 4
Company 5
Company 6
Company 7
Company 8
Company 9
Company 10
Company 11
Company 12
Company 13
Company 14
Company 15
Company 16
Company 17
Company 18
Company 19
Company 20
Company 21
Company 22
Company 23
Company 24
Company 25
Company 26
Company 27
Company 28
Company 29
Company 30
Company 31
Company 32
Company 33
Company 34
Company 35
Total
Generation
(Per Year)
7,399,928
7,049,962
284,168
720,022
7,141,365
13,265,539
205,489
8,313,793
1,243,281
0
3,273,055
7,859,848
7,478,875
0
446,425
2,542,138
0
0
8,056,730
1,988,254
8,445,411
0
326,669
0
0
68,855
0
0
0
643,791
272,972
0
804,773
0
0
87,831,343
Need
(Per Year)
0
23,352,267
1,295,626
5,009,084
471,475
5,878,620
568,094
8,298,569
1,542,508
2,807,751
3,795,859
3,319,185
0
2,662,637
306,231
3,704,126
595,464
2,660,631
5,713,982
6,809,039
3,685,330
865,360
0
303,492
1,003,080
0
581,573
327,597
1,529,836
0
0
178,994
0
398,211
166,723
87,831,343
Net Standard
Credits/Yr
7,399,928
-16,302,306
-1,011,458
-4,289,062
6,669,891
7,386,920
-362,605
15,224
-299,226
-2,807,751
-522,804
4,540,663
7,478,875
-2,662,637
140,194
-1,161,988
-595,464
-2,660,631
2,342,747
-4,820,785
4,760,080
-865,360
326,669
-303,492
-1,003,080
68,855
-581,573
-327,597
-1,529,836
643,791
272,972
-178,994
804,773
-398,211
-166,723
0
% of Credit
Supply
17.27%



15.57%
17.24%

0.04%



10.60%
17.45%

0.33%



5.47%

11.11%

0.76%


0.16%



1.50%
0.64%

1.88%


100.00%
% of Credit
Need

38.04%
2.36%
10.01%


0.85%

0.70%
6.55%
1.22%


6.21%

2.71%
1.39%
6.21%

1 1 .25%

2.02%

0.71%
2.34%

1.36%
0.76%
3.57%


0.42%

0.93%
0.39%
100.00%
Credits Used
Internally
0
7,049,962
284,168
720,022
471,475
5,878,620
205,489
8,298,569
1,243,281
0
3,273,055
3,319,185
0
0
306,231
2,542,138
0
0
5,713,982
1,988,254
3,685,330
0
0
0
0
0
0
0
0
0
0
0
0
0
0
44,979,761
       As shown in Table 6.5-16, 14 of the 35 companies have the potential to generate more
standard credits than they could use up in a given year.  The refinery concentration of standard
credits ranges from <1% to 17%.  Consequently, there does not appear to be substantial credit
market concentration so there should be significant potential for the 21 refiners that need
standard credits to ensure compliance to find them.  Additionally, intra-company trading
accounts for approximately one half of all standard credit trades which equates to a good
likelihood that the predicted transfers would actually occur.

6.5.6.4 Credit Value

       Credits generated under the proposed ABT program would have an associated monetary
value to the refining industry. This value (price) would be based on the cost to generate the
credits (selling price) and the cost avoided from not having to invest in benzene control
                                          6-56

-------
technologies (buying price). Although EPA is not proposing to control the price of benzene
credits, we can estimate that the cost of a credit based on our refinery cost model. Based on
individual refinery compliance costs, we estimate the price of a credit to be around $60 per
barrel of benzene reduced.  This value is between the highest cost of compliance or the last
refinery to come in ($59.40/bbl Bz) and next refinery to come in using BenSat ($61.39/bbl Bz).
A further discussion of how refinery compliance costs were calculated is found in Chapter 9.

6.6    Feasibility for Recovering Octane

       The use of the various benzene control technologies modeled would affect each
refinery's octane in various ways. Rerouting the benzene precursors, adding a benzene
saturation unit, adding a new extraction unit, or revamping an existing one, all would reduce the
octane of gasoline. In the case that the rerouted benzene precursors are sent to an isomerization
unit, there would be  a slight increase in octane for the rerouted stream.  We evaluated the
average octane impacts of each of these technologies on reformate and on the gasoline pool for
those refineries assumed to be taking action under the proposed benzene control standard. As we
compiled these figures, we observed that there is a large variance in octane impacts for these
technologies. The reason for much of the variance in octane impacts is that many refineries are
estimated to be using benzene precursor rerouting or some benzene extraction today. These
technologies reduce  the octane of reformate today. Thus when the reformate treating
technologies are applied the octane loss is smaller than if the refinery is not already using
benzene precursor rerouting or benzene extraction.  Since the refineries with large octane
impacts would need  to recover all of their octane loss caused by benzene controls, we provide
the maximum octane impacts in addition to the average octane impacts.  The average and
maximum octane impacts on gasoline for each benzene control technology are summarized in
Table 6.6-1.
Table 6.6-1. Summary of the Average and Maximum Octane Number Impacts for Benzene
     Control Technologies Under the Proposed Benzene Control Program ((R+M)/2)

Average Octane Impacts
Maximum Octane Impacts
Estimated Number of
Benzene Control
Technologies under the
Proposed Program
Benzene
Precursor
Rerouting
0.13
0.35
26
Benzene Precursor Rerouting
followed by Isomerization of
Benzene Precursors
0.12
0.34
28
Benzene
Saturation
0.25
0.40
11
Extraction
0.13
0.20
23
       We assessed the extent to which various means for recovering octane would have to be
applied to recover the octane reduced by the application of benzene control technologies. The
various octane recovery means we evaluated included revamping certain octane producing units
to produce more of that blendstock, purchasing and blending in high octane blendstocks, and
reducing the production of premium gasoline.  As shown in Table 6.6-1, depending on a refiner's
benzene control technology selection, the volume-weighted average octane impact for those
refineries which take steps to reduce their benzene levels averaged 0.13 octane numbers. When
                                          6-57

-------
weighted across the entire gasoline pool, this decreases to 0.08 octane numbers.  The maximum
octane loss that we observed over all the technologies is a loss of 0.40 octane numbers. We will
assess the ability for differing octane recovery means to recover 0.13, and 0.40 octane number
reductions in the gasoline pool, which represents the average and maximum reduction in octane
numbers.

       Alkylate averages about 93 octane numbers and because it is very low in benzene it is an
ideal blendstock for recovering lost octane.  Alkylate can be produced within a refinery or it
could be purchased from outside sources. Other blendstocks similar to alkylate are isooctane
and isooctene.  Depending on the feedstocks, isooctane and isooctene can have an octane as high
as 100.  Along with alkylate, isooctane and isooctene are likely replacements for the phase-out  of
MTBE by reusing the MTBE feedstocks. Because isooctane and isooctene will largely be
produced when MTBE is phased out of gasoline and used to explicitly replace MTBE, it will not
be considered in this analysis, although it could still play a marginal role for octane recovery. In
Table 6.6-2 below, we estimate the amount of alkylate which would have to be blended into a
refiner's gasoline pool to recover the various octane losses described above.

       Isomerization converts straight chain hydrocarbons into branched chain hydrocarbons
and can also saturate benzene.  The isomerization unit increases the octane of light straight run, a
gasoline blendstock which averages an octane number of 70, into a gasoline blendstock with an
average octane number of about 80. While isomerate is not a high octane blendstock and is
generally not sold as one, it is very useful for increasing the  octane of a refiner's gasoline while
saturating benzene at the same time.  In Table 6.6-2,  we estimate the volume of light straight run
that would have to be isomerized to recover the various octane losses described above.

       Ethanol's very high octane number, which is  115 octane numbers, allows making up the
octane loss using a smaller volume than the other blendstocks. Ethanol is an economical source
of octane in part due to the federal 51 cents per gallon subsidy. It contains a very small amount
of benzene (benzene is present in ethanol only because gasoline is added as a denaturant).  The
Energy Policy  Act of 2005 (EPAct) recently established a renewable fuels requirement that is
expected to predominantly be met with the addition of ethanol into gasoline.  An estimated 4
billion gallons of ethanol was blended into gasoline nationwide in 2005.  By 2012, the EPAct
calls for 7.5 billion gallons of renewable fuels to be blended into gasoline.  The increasing
renewable requirement in EPAct provides a synergistic match with the octane needs of the
proposed benzene standard.  In Table 6.6-2 we summarize the volume of ethanol that would have
to be blended into gasoline to recover a range of octane losses.

       Finally premium  gasoline usually meets either a 91 or 93 octane number rating, while
regular grade gasoline must meet an 87 octane number requirement, although for high altitude
areas the requirement is relaxed to an 85 octane number requirement. The recent increase in
energy prices resulted in a reduced demand for premium grade gasoline.  From 2000 to 2005, the
fraction that premium gasoline comprises of total gasoline consumed in the U.S. decreased from
20 percent to 12 percent. Considering that this reduced demand for premium grade gasoline may
continue, we evaluated the extent that the demand in premium grade gasoline would have to
continue to be  supplanted by regular  grade gasoline to make up for the projected loss of octane
due to benzene reduction in gasoline (this supplanted octane production means that these

                                          6-58

-------
refineries producing less premium gasoline would have the potential to increase their octane
production potential by this same amount). This shift in premium gasoline demand to regular
grade demand to recover the range of octane losses is described in Table 6.6-2.
    Table 6.6-2. Percent Changes in Gasoline Content for Recovering Octane Shortfalls
                              (volume percent of gasoline)

Isomerizing Light Straight
Run Naphtha
Blending in Alky late
Blending in Ethanol
Reduced 9 lor 93 ON
Premium Grade Gasoline
0.13 Octane Number Loss
1
2
0.5
3
0.40 Octane Number Loss
4
7
2
10
       Isomerizing a refinery's gasoline blendstocks is effective because in addition to
addressing octane, it can also treat the benzene normally found in gasoline.  It would not be an
available technology in those refineries that are already fully using isomerization.  The refinery -
by-refmery cost model estimates that light straight run feedstock to the isomerization unit
typically comprises about 7 percent of each refinery's gasoline pool so it potentially could meet
the octane needs of even the greatest octane needs caused by this rulemaking if isomerization is
not already being used. Even those refineries that will be isomerizing all their light straight run
prior to the implementation of the proposed benzene standard could reroute the six carbon
hydrocarbons around the reformer and send this stream to an isomerization unit to recover at
least a part of the octane loss associated with the benzene reduction.  An average octane loss of
0.14 octane numbers and the refinery-specific maximum 0.40 octane numbers would require an
additional 1 volume percent and 4  percent of the light straight run currently blended into
gasoline to be isomerized, respectively.

       Alkylate's moderate octane value requires a relatively large volume to make up for the
octane losses associated with the removal of benzene.  At the estimated highest octane loss, the
volume of alkylate would have to nearly double relative to the 12 percent typically blended into
gasoline in 2003. Additional alkylate may be able to be produced by increasing the severity of
the FCC unit, if there is capacity to do so, that would increase the production of feedstocks to the
alkylate unit. Alkylate's very  desirable gasoline blending properties, which is high octane, low
RVP and sulfur and very low benzene, encourages its use. To replace an average octane loss  of
0.14 octane numbers and the refinery-specific maximum 0.40 octane numbers, a refinery would
need to produce or purchase and blend in an additional 2 volume percent and 7 percent of
alkylate into their gasoline, respectively.

       Ethanol is very high in octane which allows the recovery of lost octane caused by the
treating of benzene with a smaller volume than the other octane recovery means considered. The
additional volume of ethanol expected to be blended into gasoline under EPAct makes it a likely
candidate for an  octane replacement for a benzene standard. If all of EPAct's renewable
requirement is met with the blending of ethanol into gasoline, the 31A> additional billion gallons
of ethanol that would be blended into gasoline between today and 2012 would increase ethanol's
                                          6-59

-------
content in gasoline from 2.8 to 4.7 volume percent, a 1.9 volume percent increase in all U.S.
gasoline. To replace an average octane loss of 0.14 octane numbers and the refinery-specific
maximum 0.40 octane numbers, a refiner would need to blend in an additional 0.5 volume
percent and 2 percent of ethanol in their gasoline, respectively. This provides far more than the
octane number increase needed to recover the average octane loss of refineries that reduce their
benzene levels to comply with the proposed benzene standard.  The phasing-in, under the ABT
program, of the benzene standard and its associated octane loss would coincide with the period
that EPAct's renewable requirement phases in.

       The decreasing demand for premium grade gasoline would provide another means for
refiners to recover the octane lost from benzene control.  The demand for premium has been
supplanted by a higher demand for lower octane regular, freeing up octane producing potential in
refineries. Between 2000 and 2005, premium gasoline demand decreased by 8 volume percent.
This decrease represents nearly a 0.4 octane number decrease in the gasoline pool. To replace an
average octane loss of 0.14 octane numbers and the maximum refinery-specific 0.40 octane
numbers, a refiner would need to have shifted 3 volume percent and 10 percent of their gasoline
production from premium grade to regular grade, respectively.  This indicate