Control of Hazardous Air Pollutants
   from Mobile Sources

   Regulatory Impact Analysis
United States
Environmental Protection
Agency

-------
                  Control of Hazardous Air Pollutants
                           from Mobile Sources


                        Regulatory Impact Analysis
                              Assessment and Standards Division
                             Office of Transportation and Air Quality
                             U.S. Environmental Protection Agency
v>EPA
United States                                       EPA420-R-07-002
Environmental Protection                                  February 2007
Agency

-------
                                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 the Benzene Control Program	6-1

Chapter 7: Portable Fuel Container Feasibility and Test Procedures	7-1

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

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

Chapter 10: Portable Fuel Container 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

-------
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
ANS               Alaska North Slope
APHEA            Air Pollution and Health: A European Approach
API                American Petroleum Institute
AQCD             Air Quality Criteria Document
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
CHD               Coronary Heart Disease
CI                 Compression Ignition
CUT               Chemical Industry Institute of Toxicology
CIMT              Carotid Intima-Media Thickness
                                                11

-------
CM 15              Gasoline with 15 percent methanol content
CMAI              Chemical Market Associates Incorporated
CMAQ             Community Multiscale Air Quality Model
CNG               Compressed Natural Gas
CNS               Central Nervous System
CO                Carbon Monoxide
CO2               Carbon Dioxide
COI                Cost-of-Illness
COPD              Chronic Obstructive Pulmonary Disease
COV               Coefficient of Variation
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
DF                Deterioration Factor
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
EPEFE             European Programme on Emissions, Fuels, and Engine Technology
EPAct              Energy Policy Act of 2005
ETBE              Ethyl Tertiary Butyl Ether
ETC               Electronic Throttle Control
ETS                Environmental Tobacco Smoke
EU                European Union
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
FEV               Functional Expiratory Volume
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
                                                ill

-------
GPA
GVW
GVWR
H2
HAD
HAP
HAPEM5
HC
HCO
HDN
HOPE
HDS
HOT
HEGO
HHC
HHDDT
HI
HLDT
HQ
HSR
HVGO
IBP
ICAO
ICD-9
ICI
IFF
IMO
IMPROVE
IRFA
IRIS
ISBL
ISC
ISCST
JAMA
K
KBBL
Kfoeb/day
KWH
LB
LCG
LCN
LCD
LDGT
LDGV
LOT
LDV
LEV I
LEV II
LEV
LHC
LLDT
LLE
LNS
LP
LPG
LRS
Geographic Phase-in Area
Gross Vehicle Weight
Gross Vehicle Weight Rating
Hydrogen gas
Diesel Health Assessment Document
Hazardous Air Pollutant
Hazardous Air Pollutant Exposure Model version 5
Hydrocarbon
Heavy Cycle Oil (a refinery stream)
Naphtha Hydrotreater (also Hydro-Denitrogenation Unit)
High density polyethylene
Hydro-Desulfurization Unit
Hydrotreater
(Heated) Exhaust Gas Oxygen
Heavy Hydrocrackate
Heavy Heavy-Duty Diesel Truck
Hazard Index
Heavy Light-Duty Truck
Hazard Quotient
Heavy Straight Run (a refinery stream)
Heavy Vacuum Gas Oil (a refinery stream)
Initial Boiling Point
International Civil Aviation Organization
International Classification of Diseases - Ninth Revision
Independent Commercial Importer
Institute Francais du Petrole
International Maritime Organization
Interagency Monitoring of Protected Visual Environments
Initial Regulatory Flexibility Analysis
Integrated Risk Information System
Inside Battery Limits
Integrated Source Complex
Industrial Source Complex Short Term
Journal of the American Medical Association
Thousand
Thousand barrels
Thousands of Fuel Oil Equivalent Barrels per Day
Kilowatt Hour
Pound
Light Cracked Gasoline
Light Coker Naphtha
Light Cycle Oil (a refinery stream)
Light Duty Gasoline Truck
Light Duty Gasoline Vehicle
Light-Duty Truck
Light-Duty Vehicle
Low Emission Vehicle I
Low Emission Vehicle II
Low Emission Vehicle
Light Hydrocrackate
Light Light-Duty Truck
Liquid-Liquid Extraction
Light Naphtha Splitter
Linear Programming (a type of refinery model)
Liquefied Petroleum Gas
Lower Respiratory Symptom
                                                 IV

-------
LSR
MACT
MAP
MAP
MC
MDPV
mg/m3
MHDDT
MHDT
MI
MILY
MLE
MM
MNCPES
MOBILE6.2
MON
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
OBD
OEHHA
OGJ
OMB
OP
OSHA
Light Straight Run (a refinery stream)
Maximum Available Control Technology
(Engine) Mass Air Flow
(Engine) Manifold Absolute Pressure
Motorcycle
Medium-Duty Passenger Vehicle
Milligrams per cubic meter
Diesel-Fueled Medium Heavy-Duty Truck
Medium Heavy-Duty Truck
Myocardial Infarction
Morbidity Inclusive Life Years
Maximum Likelihood Estimate
Million
Minnesota Children's Pesticide Exposure Study
EPA's Highway Vehicle Emission Model
Motor Octane Number
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
On-board Diagnostics
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

-------
OTAQ             Office of Transportation and Air Quality
PADD             Petroleum Administration for Defense District
PAH               Polycyclic Aromatic Hydrocarbon
PC                 Particle Count
PC                 Passenger car
PFC               Portable Fuel Containers
PFT               perfluorocarbon tracer
PM                Paniculate Matter
PM/NMHC         Paniculate Matter to Non-Methane Hydrocarbon Ratio
PM10               Coarse Particle
PM2 5               Fine Particle
POM               Polycyclic Organic Matter
PONA             Paraffin, Olefin, Naphthene, Aromatic
PPM               Parts Per Million
PRTP              Percentage Reduction Trigger Point
PSI                Pounds per Square Inch
PSR               Power Systems Research
QALY             Quality-Adjusted Life Year
R&D               Research and Development
RAMS             Regional Atmospheric Modeling System
RBOB             Reformulated Blendstock for Oxygenate Blending
RECS              Residential Energy Consumption Survey
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
RPM               Revolutions Per Minute
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
SIP                State Implementation Plan
SOA               Secondary Organic Aerosols
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
                                                VI

-------
TEAM             Total Exposure Assessment Methodology
THC               Total Hydrocarbon
TMP               2,2,4-Trimethylpentane
TWA              Time-weighted Average
TSP               Total Suspended Paniculate Matter
TWC              Three-Way Catalyst
UC                Unified Cycle Emission Test Procedure from ARB
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
                                                Vll

-------
Final Regulatory Impact Analysis
                               Executive Summary

       EPA is adopting new standards to reduce emissions of mobile source air toxics (MSATs)
including benzene and overall hydrocarbons from motor vehicles, motor vehicle fuels, and
portable fuel containers (PFCs). This Regulatory Impact Analysis provides technical, economic,
and environmental analyses of the 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 adopting, providing a technical justification for the
feasibility of the standards for vehicles, fuels, and PFCs, respectively.  Chapters 8 throughlO
present cost estimates of complying with the new standards or vehicles, fuels, and PFCs,
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 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 finalizing and the
estimated impacts.

Emissions Standards

Vehicles

       We are adopting new standards for both exhaust and evaporative emissions from
passenger vehicles.  The new exhaust emissions standards will 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. We project that these cold-temperature NMHC controls  will also
result in lower direct PM emissions at cold temperatures.

       Accordingly, we are requiring that light-duty vehicles, light-duty trucks, and medium-
duty passenger vehicles be subject to a new NMHC  exhaust emissions standard at 20° F.
                                          ES-1

-------
Final Regulatory Impact Analysis
Vehicles at or below 6,000 pounds gross vehicle weight rating (GVWR) will 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 will be subject to a sales-weighted fleet
average NMHC level of 0.5 grams/mile.  For lighter vehicles, the standard will phase in between
2010 and 2013. For heavier vehicles, the new standards will phase in between 2012 and 2015.
We are also adopting 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 adopting a set of nominally more stringent evaporative  emission standards
for all light-duty vehicles, light-duty trucks, and medium-duty passenger vehicles.  The new
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 will codify the approach that manufacturers are already taking for 50-state evaporative
systems, and thus the standards will prevent backsliding in the future. The  new evaporative
emission standards begin in 2009 for lighter vehicles and in 2010 for the heavier vehicles.

Gasoline Fuel Standards

       We are requiring that beginning January 1, 2011, refiners and fuel importers will meet a
refinery 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 new fuel standard will result in air toxics emissions reductions that are greater than
required under all existing gasoline toxics programs.  As a result, EPA is establishing that upon
full implementation in 2011, the regulatory provisions for the benzene control program will
become the single regulatory mechanism used to implement the  reformulated gasoline (RFG)
and Anti-dumping  annual average toxics requirements.  The current RFG and Anti-dumping
annual average provisions will be replaced by the new benzene control program. The MSAT2
benzene control program will also replace the MSAT1 requirements.  In addition, the program
will satisfy certain fuel MS AT conditions of the Energy Policy Act of 2005. In all  of these ways,
we will significantly consolidate and simplify the existing national fuel-related MSAT regulatory
program.

       We are also allowing 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 can generate benzene credits by taking  early  steps to reduce gasoline benzene levels.
Beginning in 2011  and continuing indefinitely,  refiners can generate credits by producing
gasoline with benzene levels below the 0.62 vol% refinery average standard.  Refiners can apply
the credits towards company compliance, "bank" the credits for  later use, or transfer ("trade")
them to other refiners nationwide (outside of California) under the new program. Under this
program, refiners can use credits to achieve compliance with the benzene content standard. In
addition, to the 0.62 vol% standards, refiners must also meet a maximum average benzene
                                          ES-2

-------
Final Regulatory Impact Analysis
standard of 1.3 vol% beginning on July 1, 2012.  A refinery's or importer's actual annual
average gasoline benzene levels may not exceed this maximum average standard.
Portable Fuel Container Controls

       Portable fuel containers (PFCs) include gasoline containers (gas cans) and kerosene and
diesel containers. PFCs are consumer products used to refuel a wide variety of equipment,
including lawn and garden equipment, generators, heaters, recreational equipment, and passenger
vehicles that have run out of gas. We are adopting standards that will reduce hydrocarbon
emissions from evaporation, permeation, and spillage. These standards will significantly reduce
benzene and other toxics, as well as VOC more generally. VOC is an ozone precursor.  We are
also applying the new requirements to kerosene and diesel containers, which are identical to gas
cans except for their color and could be used for gasoline.

       We are adopting 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 will
apply to PFCs manufactured on or after January 1, 2009. We are also adopting test procedures
and a certification and compliance program, in order to ensure that PFCs will meet the emission
standard over a range of in-use conditions.  The new requirements will 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 PFCs in 2001, and since then,
several other states have adopted the program.  In 2005, California adopted a revised program,
which will take effect July 1, 2007. The revised California program is very similar to the
program we are adopting. Although a few aspects of the program we are adopting are different,
we believe  manufacturers will be able to meet both EPA and California requirements with the
same container 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
                                          ES-3

-------
Final Regulatory Impact Analysis
sources as well due to lower benzene levels in gasoline.  Annual benzene emissions from
   Table 1: Estimated Reductions in Benzene Emissions from New Control Measures by
                         Sector, 2020 and 2030 (tons per year)

Fuels
Vehicles
PFCs
2020
17,618
27,097
718
2030
19,643
45,037
814
Total 42,760 61,035
   Table 2: Estimated Reductions in MSAT Emissions from New Control Measures by
                         Sector, 2020 and 2030 (tons per year)

Fuels
Vehicles
PFCs
2020
17,618
177,007
18,553
2030
19,643
294,284
21,036
Total 210,303 330,844
voc

      VOC emissions will be reduced by the hydrocarbon emission standards for both light-
duty vehicles and PFCs.  Annual VOC emission reductions from these sources will be about 34%
lower in 2030 because of the new rule.

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

Vehicles
PFCs
2020
529,363
216,294
2030
882,762
245,255
                                       ES-4

-------
Final Regulatory Impact Analysis
  Total
745,658
1,128,017
PM2.5

       We expect that only the vehicle control will reduce emissions of direct PM2.5. As shown
in Table 4, we expect this control to reduce direct PM2.5 emissions by about 19,000 tons in 2030.
In addition, the VOC reductions  from the vehicle and PFC standards will also reduce secondary
formation of PM2.5.

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

PM2 5 Reductions from Vehicle
Standards (tons)
2020
11,646
2030
19,421
Costs

Fuels

       The refinery model estimates that the benzene standard will cost 0.27 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 will be 0.40 cents per
gallon.) This per-gallon cost will result from an industry-wide investment in capital equipment
of $1,110 million to reduce gasoline benzene levels.  This will amount to an average of $14
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 Benzene Standard, 2020 and 2030
                                          2020
                                         2030
 Fuels program
$398 million
$441 million
Vehicles
                                         ES-5

-------
Final Regulatory Impact Analysis


       We project that the average incremental costs associated with the new cold temperature
standards will be less than $1 per vehicle. We are not projecting changes to vehicle hardware as
a result of the new standard. Costs are 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 new evaporative 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 will streamline
certification and be an "anti-backsliding" measure.  It also will 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, we project the costs will be fully amortized by 2020.

   Table 6.  Estimated Aggregate Annual Cost for the Vehicle Standards, 2020 and 2030
                                          2020
                                         2030
 Vehicles program
$0
$0
PFCs

       Table 7 summarizes the projected near-term and long-term per unit average costs to meet
the new emission standards. Long-term impacts on PFCs 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 PFC when used with gasoline.

            Table 7  Estimated Average PFC Costs and Lifetime Fuel Savings

Near-Term Costs
Long-Term Costs
Cost
$2.69
$1.52
               Gasoline Savings (NPV)
              $4.24
       We have also estimated aggregate costs and gasoline fuel savings which are projected to
peak in 2013 at about $61 million and then drop to about $33 million once fixed costs are
recovered. The aggregate annual costs and gasoline savings estimates for 2020 and 2030 are
provided in Table 8.

 Table 8. Estimated Aggregate Annual Cost and Gasoline Savings for the PFC Standards,
                                     2020 and 2030
                                         ES-6

-------
Final Regulatory Impact Analysis

PFC Costs
2020
$37,542,748
2030
$45,764,401
 PFC Gasoline Savings         $109,589,064                  $124,264,434
Cost Per Ton

       We have calculated the cost per ton of HC, benzene, total MSATs, and PM emissions
reductions associated with the new fuel, vehicle, and PFC programs. We have calculated the
costs per ton using the net present value of the annualized costs of the program, including PFC
gasoline 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 PFCs 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

-------
Final Regulatory Impact Analysis
       Table 9. HC Aggregate Cost per Ton and Long-Term Annual Cost Per Ton
                                     ($2003)

Vehicles
PFCs (without
fuel savings)
PFCs (with fuel
savings)
Combined (with
fuel savings)
Discounted
Lifetime
Cost per ton at 3%
$14
$240
$0
$0
Discounted
Lifetime
Cost per ton at 7%
$18
$270
$0
$0
Long-Term Cost
per Ton in 2020
$0
$170
$0
$0
Long-Term Cost
per Ton in 2030
$0
$190
$0
$0
    Table 10. Benzene Aggregate Cost per Ton and Long-Term Annual Cost Per Ton
                                     ($2003)

Fuels
Vehicles
PFCs (without
fuel savings)
PFCs (with fuel
savings)
Combined (with
fuel savings)
Discounted
Lifetime
Cost per ton at 3%
$22,400
$270
$74,500
$0
$8,200
Discounted
Lifetime
Cost per ton at 7%
$23,100
$360
$82,900
$0
$8,600
Long-Term Cost
per Ton in 2020
$22,600
$0
$52,200
$0
$7,600
Long-Term Cost
per Ton in 2030
$22,500
$0
$56,200
$0
$5,900
                                      ES-8

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

Fuels
Vehicles
PFCs (without
fuel savings)
PFCs (with fuel
savings)
Combined (with
fuel savings)
Discounted
Lifetime
Cost per ton at 3%
$22,400
$42
$2,800
$0
$1,700
Discounted
Lifetime
Cost per ton at 7%
$23,100
$54
$3,100
$0
$1,800
Long-Term Cost
per Ton in 2020
$22,600
$0
$2,000
$0
$1,600
Long-Term Cost
per Ton in 2030
$22,500
$0
$2,200
$0
$1,100
    Table 12 Direct PM Aggregate Cost per Ton and Long-Term Annual Cost Per Ton
                                        ($2003)

Vehicles
Discounted
Lifetime
Cost per ton at 3%
$650
Discounted
Lifetime
Cost per ton at 7%
$870
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 new standards through 2030. When translating emission benefits to health
effects and monetized values, however, we only quantify the PM-related benefits associated with
the new cold temperature vehicle standards.  The reductions in PM from the cold temperature
vehicle standards will 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 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 $6.3 billion using a 3 percent discount rate and $5.7 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 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 rule, which include both the PFC, vehicle, and fuel
standards, are $400 million in 2030 (in 2003$, including fuel  savings).
                                         ES-9

-------
Final Regulatory Impact Analysis
       The PM2.5 benefits are scaled based on relative changes in direct PM emissions between
this rule and the proposed Clean Air Nonroad Diesel (CAND) rule.  As explained in Section
12.2.1 of the RIA, the PM2.5 benefits scaling approach is limited to those studies, health impacts,
and assumptions that were used in the proposed CAND analysis. As a result, PM-related
premature mortality is based on the updated analysis of the American Cancer Society cohort
(ACS; Pope et al., 2002).  However, it is important to note that since the CAND rule, EPA's
Office of Air and Radiation (OAR) has adopted a different format for its benefits analysis in
which characterization of the uncertainty in the concentration-response function is integrated into
the main benefits analysis. Within this context, additional data sources are available, including a
recent expert elicitation and updated analysis of the Six-Cities Study cohort (Laden et al., 2006).
Please see the PM NAAQS RIA for an indication of the sensitivity of our results to use of
alternative concentration-response functions.

       The analysis presented here assumes a PM threshold of 3 ug/m3, equivalent to
background.  Through the RIA for the Clean Air Interstate Rule  (CAIR), EPA's consistent
approach had 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 had been supported by advice from EPA's
technical peer review panel, the Science Advisory Board's Health Effects Subcommittee (SAB-
HES). However, EPA's most recent PM2.s Criteria Document concludes that "the available
evidence does not either support or refute the existence  of thresholds for the effects of PM on
mortality across the range of concentrations in the studies," (p. 9-44).  Furthermore, in the RIA
for the PM NAAQS we used a threshold of 10 ug/m3 based on recommendations by the Clean
Air Scientific  Advisory Committee (CAS AC) for the Staff Paper analysis.  We consider the
impact of a potential, assumed threshold in the PM-mortality concentration response function in
Section 12.6.2.2 of the RIA
   Table 13 Estimated Monetized PM-Related Health Benefits of the 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
$3.3 + B
$3.0 + B
2030
$6.3 + B
$5.7 + B
a  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 of the RIA.
b  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 12.1-2 of the RIA.
0  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 two significant
   digits for ease of presentation and computation.

Economic Impact Analysis
                                           ES-10

-------
Final Regulatory Impact Analysis
       We prepared an Economic Impact Analysis (EIA) to estimate the economic impacts of
the emission control program on the PFC, gasoline fuel, and light-duty vehicle markets.  Our
estimates of the net social costs of the program for 2020 and 2030 are provided in Table 14
below. These estimates reflect the estimated costs associated with the gasoline, PFC, and vehicle
controls and the expected gasoline fuel  savings from better evaporative controls on PFCs.  The
results of the economic impact modeling performed for the gasoline fuel and PFC control
programs suggest that the social costs of those two programs are expected to be about $440.1
million in 2020 with consumers of these products expected to bear about 58 percent of these
costs. We estimate fuel savings of about $80.7 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 Program (Millions of 2003$)
1
2020
2030 |
Net Social Costs $359.4 $400.0
Impact on Small Businesses

       We prepared a Regulatory Flexibility Analysis, which evaluates the potential impacts of
new standards and fuel controls 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 are adopting most of the Panel's
recommendations.  These provisions will reduce the burden on small entities that will be subject
to this rule's requirements. We have included provisions that give small light-duty vehicle
manufacturers, small gasoline refiners, and small PFC manufacturers several compliance options
aimed specifically at reducing the burden on these small entities. In general, for vehicles and
fuels, the options are similar to small entity provisions adopted in prior rulemakings where EPA
set vehicle and fuel standards. The options included for small PFC manufacturers are unique to
this rulemaking since we are adopting PFC standards for the first time.
                                         ES-11

-------
Final Regulatory Impact Analysis
                              Chapter 1: Table of Contents
Chapter 1: Mobile Source Air Toxics Health Information	2
   1.1.   What Are MSATs?	2
      1.1.1.  Compounds Emitted by Mobile Sources and Identified in IRIS	2
      1.1.2.  Compounds Emitted by Mobile Sources and Included on Section 112(b) List of
            Hazardous Air Pollutants	5
      1.1.3.  Other Sources of Information on Compounds with Potential Serious Adverse
            Health Effects	6
      1.1.4.  Which Mobile Source Emissions Pose the Greatest Health Risk at Current Levels??
       1.1.4.1.     Risk Drivers in 1999 National-Scale Air Toxics Assessment	7
       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	18
      1.3.4.  Acetaldehyde	19
      1.3.5.  Acrolein	20
      1.3.6.  Naphthalene	20
      1.3.7.  2,2,4-Trimethylpentane	21
      1.3.8.  Ethylbenzene	21
      1.3.9.  n-Hexane	22
      1.3.10. Methyl Tertiary Butyl Ether (MTBE)	23
      1.3.11. Styrene	23
      1.3.12. Toluene	23
      1.3.13.Xylenes	24
      1.3.14. Polycyclic Organic Matter (POM)	25
      1.3.15. Diesel Exhaust	25
   1.4.   Emerging Issues	26
      1.4.1.  Gasoline PM	26
      1.4.2.  Metals	29
                                          1-1

-------
Final Regulatory Impact Analysis



        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 or not quantitatively assessed 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 (http://www.epa.gov/otaq/toxics.htm ).
Table l.l.-l lists those compounds from the Master List that currently meet those 2001 MSAT
criteria, based on the current IRIS.

       Table l.l.-l 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 Chapter 2 Table 2.2.-1. There  are
                                          1-2

-------
Final Regulatory Impact Analysis
several compounds for which IRIS assessments are underway and therefore are not included in
Table 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.
                                           1-3

-------
Final Regulatory Impact Analysis
      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
                                         1-4

-------
Final Regulatory Impact Analysis
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
                                          1-5

-------
Final Regulatory Impact Analysis
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
                                            1-6

-------
Final Regulatory Impact Analysis
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. The limitations and uncertainties associated with NAT A are
discussed in Section 3.2.1.3 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.0 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.0  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.
       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.
       c Benzene was assigned an overall confidence level of "higher" based on consideration of the combined
uncertainties from the modeling estimates.
       D Acrolein was assigned an overall confidence level of "lower" based on consideration of the combined
uncertainties from the modeling estimates.


                                            1-7

-------
Final Regulatory Impact Analysis
Sixteen regional-scale noncancer risk drivers were identified in the 1999 NATA (see Table 1.1.-
3.).

-------
Final Regulatory Impact Analysis
Table 1.1.-3.  National and Regional Cancer and Noncancer Risk Drivers in 1999 NATA
Cancer l
National drivers 2
Benzene11
Regional drivers 3
Arsenic compounds1"
BenzidineL
l,3-ButadieneL
Cadmium compounds1"
Carbon tetrachloride11
Chromium VIL
Coke ovenM
Ethylene oxide11
HydrazineM
NaphthaleneM
Perchl oroethy 1 eneM
Polycyclic organic matterM




Noncancer
National drivers 4
AcroleinL
Regional drivers 5
Antimony11
Arsenic compounds1"
l,3-ButadieneL
Cadmium compounds1"
Chlorine1"
Chromium VIL
Diesel PMM
FormaldehydeM
Hexamethylene l-6-diisocyanateM
Hydrazine11
Hydrochloric acidL
Maleic anhydride1"
Manganese compounds1"
Nickel compounds1"
2,4-Toluene diisocyanate1"
Triethylamine1"
          :The 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
          EPA has assigned an overall confidence level for each pollutant in NATA based on consideration of the
          combined uncertainties from emissions estimation, ambient concentration modeling, and exposure
          modeling. These judgments refer to the relative confidence between two air toxics compounds. A
          judgment of "Higher" (H) means the confidence is higher for this compound than for compounds
          assigned a "Medium" (M) or "Lower" (L).
                                                 1-9

-------
Final Regulatory Impact Analysis
       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 NAT A, EPA rated its confidence
inrisk 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 based on 1999 NATA data.

           Table 1.1.-4. Mobile Source Contribution to 1999 NATA Risk Drivers
1999 NATA Risk Drivers



Benzene11
l,3-ButadieneL
FormaldehydeM
AcroleinL
Polycyclic organic matter*M
NaphthaleneM
Diesel PM and Diesel
exhaust organic gasesM
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.
       EPA has assigned an overall confidence level for each pollutant in NATA based on consideration of the
       combined uncertainties from emissions estimation, ambient concentration modeling, and exposure
       modeling. These judgments refer to the relative confidence between two air toxics compounds. A judgment
       of "Higher" (H) means the confidence is higher for this compound than for compounds assigned a
       "Medium" (M) or "Lower" (L).
                                            1-10

-------
Final Regulatory Impact Analysis
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

-------
Final Regulatory Impact Analysis
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, if needed, would lead to higher estimates of lifetime risk.
                                       1-12

-------
Final Regulatory Impact Analysis
1.2.2.   Chronic Exposure and Noncancer Health Effects

       Noncancer effects resulting from chronic exposures include a wide range of
effects in many 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, EPA reviews available data on different health endpoints and target organs
and describes the range of effects observed and the related dose/exposure levels. EPA
focuses particular attention 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 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 for a few selected compounds including acrolein and hydrogen sulfide.
                                       1-13

-------
Final Regulatory Impact Analysis
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 occupational inhalation exposure to benzene has been shown to cause
cancers of the hematopoetic (blood cell) system in adults.  Among these are acute
nonlymphocytic leukemia,E and chronic lymphocytic leukemia.17'18  A doubling of risk
for acute nonlymphocytic 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
       E 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 blood cell that is responsible for antibody and cell-mediated 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

-------
Final Regulatory Impact Analysis
                                         20, 21
chromosomal changes in humans and animals  '   and increased proliferation of mouse
                 fjrj ry^
bone marrow cells.  '

       The latest assessment by EPA estimates the excess risk of developing 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 an estimated 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 reflects
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 (7.8 x  10"6 per |ig/m3) was used
in the 1999NATA.

             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 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. This relationship could support the inference
that cancer risks are as high, or higher than the estimates provided in the existing EPA
assessment.26 However, since the mode of action for benzene carcinogenicity is
unknown, the current cancer unit risk estimate assumes linearity of the low-dose response.
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.28'29'30 These data highlight the
importance of ambient exposure levels and their contribution to benzene-related adducts.
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".31'32'33 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
how they might influence low-dose risk estimates. A better understanding of the
                                       1-15

-------
Final Regulatory Impact Analysis
biological mechanism of benzene-induced leukemia is needed.

       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
been occupationally exposed to benzene.34'35  Data from animal studies have shown
benzene exposures result in damage to the hematopoietic (blood cell formation) system
during development.36'37'38  Also, key changes related to the development of childhood
leukemia occur in the developing fetus.39 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.40  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.41'42  People with long-term occupational exposure to benzene  have experienced
harmful effects on the blood-forming tissues,  especially in the 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,F  a condition characterized by
decreased numbers of circulating erythrocytes (red blood cells), leukocytes (white blood
cells), and thrombocytes (blood platelets).43'44 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.45' 46  The most sensitive noncancer effect observed
in humans, based on current data, is the depression of the absolute lymphocyte count in
blood.47' 48

       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
       F 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 skeletal system. 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

-------
Final Regulatory Impact Analysis
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
hematological endpoints that are triggered at occupational exposures of less than 5 ppm
(about 16 mg/m3)49 and, more significantly, at air levels of 1 ppm (about 3 mg/m3) or less
among genetically susceptible populations.50  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.51 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
          rrj r--) 	
inhalation.   '    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.54'55'56 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.57

       A recent study extended the investigation of 1,3-butadiene exposure and leukemia
among synthetic rubber industry workers.58 The results of this study strengthen the
evidence for the relationship between 1,3-butadiene exposure and lymphohematopoietic
cancer. This relationship was found to persist after controlling for exposure  to other
toxics in this work environment.

       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.59  Based on this critical effect
                                       1-17

-------
Final Regulatory Impact Analysis
and the benchmark concentration methodology, an RfC for chronic health effects was
calculated at 0.9 ppb (approximately 2 jig/m3). Confidence in the inhalation RfC is
medium.

1.3.3.   Formaldehyde

      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.60  EPA's
current IRIS summary provides an upper bound cancer unit risk estimate of 1.3xlO"5 per
|ig/m3.G  In other words, there is an estimated risk of about thirteen excess leukemia
cases in one million people exposed to 1 |ig/m3 of formaldehyde over a lifetime. EPA is
currently reviewing recently published epidemiological data. For instance, research
conducted by the National Cancer Institute (NCI) found an increased risk of
nasopharyngeal cancer and lymphohematopoietic malignancies such as leukemia among
workers exposed to formaldehyde.61' 62 NCI is currently performing an update of these
studies.  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.63 Extended follow-up of a cohort of British chemical workers
did not find evidence of an increase in nasopharyngeal or lymphohematopoeitic cancers,
but a continuing statistically significant excess in lung cancers was reported.64

      Based on the developments of the last decade, 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." Recently the German Federal Institute for
Risk Assessment determined that formaldehyde is a known human carcinogen.65

      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.66'67'68
CIIT's risk assessment of formaldehyde incorporated mechanistic and  dosimetric
information on formaldehyde. The risk assessment analyzed carcinogenic risk from
       G U.S. EPA (1989). Integrated Risk Information System File for Formaldehyde. This material is
available electronically at http://www.epa.gov/iris/subst/0419.htm.
                                      1-18

-------
Final Regulatory Impact Analysis
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.69  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).70  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.71

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 toxic by inhalation.72  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.73'74 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.75  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.76 Some asthmatics have been shown to be a sensitive
                                       1-19

-------
Final Regulatory Impact Analysis


subpopulation to decrements in functional expiratory volume (FEV1 test) and broncho-
constriction upon acetaldehyde inhalation.77

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.78 As such, it readily absorbs into airway fluids in the respiratory tract
when inhaled. The toxicological data base demonstrating the highly irritating nature of
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 2003.79 The Agency has developed an RfC for acrolein of 0.02 |ig/m3 and
an RfD of 0.5 ug/kg-day.80 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.81

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 of a reassessment of the inhalation
carcinogenicity  of naphthalene.82 The draft reassessment (External Review Draft, IRIS
Reassessment of the Inhalation Carcinogenicity of Naphthalene) completed external peer
review in 2004 by Oak Ridge Institute for Science and Education.83  Based on external
comments, additional analyses are being considered.  California EPA has released a new
risk assessment for naphthalene with a cancer unit risk estimate of 3xlO"5 per |ig/m3.84
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:
                                       1-20

-------
Final Regulatory Impact Analysis


possibly carcinogenic to humans.85  Current risk estimates for naphthalene are based on
extrapolations from rodent studies conducted at higher doses.  At present, human data are
inadequate for developing estimates.

       The current EPA IRIS assessment includes noncancer data on hyperplasia and
metaplasia in nasal tissue that form the basis of the inhalation RfC of 3 |ig/m3.86  The
principal study was given medium confidence because adequate numbers of animals were
used, and the severity of nasal effects increased at the higher exposure concentration.
However, the study produced high mortality and hematological evaluation was not
conducted beyond 14 days. The database was given a low-to-medium confidence rating
because there are no chronic  or subchronic inhalation studies in other animal species, and
there are no reproductive or developmental studies for inhalation exposure. In the
absence of human or primate toxicity data, the  assumption is made that nasal responses in
mice to inhaled naphthalene are relevant to humans; however,  it cannot be said with
certainty that this RfC for naphthalene based on nasal effects will be protective for
hemolytic anemia and cataracts, the more well-known human effects from naphthalene
exposure. As a result, we have medium confidence in the RfC.

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
exposures in humans.87 Overall, there was "inadequate information to assess
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 kidneys.88 These effects were not seen in the female rat test subjects.
These renal effects, specific to the male rat, are not thought to be relevant to humans.
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.89 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
                                       1-21

-------
Final Regulatory Impact Analysis
data from animal bioassays and human studies
                                           90
       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.91

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.92

       Under the 2005 Guidelines for Carcinogen Risk Assessment, there is inadequate
information to assess the carcinogenic potential of n-hexane.93 Chronic exposure to n-
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.94  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.  Nervous system
effects include dizziness, giddiness, slight nausea, and headache in humans.95
                                       1-22

-------
Final Regulatory Impact Analysis
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."96  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."97 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 rats98. 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
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
central nervous system effects in humans. 99' 10°

1.3.12.  Toluene

       Toluene is found in evaporative as well as exhaust emissions from  motor vehicles.
                                      1-23

-------
Final Regulatory Impact Analysis
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, toluene was not carcinogenic in adequate
inhalation cancer bioassays of rats and mice exposed for life, and increased incidences of
mammary cancer and leukemia were reported in a lifetime rat oral bioassay.101

       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.102

       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.103  The overall
confidence in the RfD is medium.

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.104

       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.105  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
                                       1-24

-------
Final Regulatory Impact Analysis
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.106

       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.107

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. At least eight of the compounds included in the class of
compounds known as POM are classified by EPA as probable human carcinogens based
on animal data. These include acenaphthene, benzo(a)anthracene, benzo(b)fluoranthene,
benzo(k)fluoranthene, benzo(a)pyrene, chrysene, dibenz(a,h)anthracene, and
indeno(l,2,3-cd)pyrene. One POM,  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.108
These studies are discussed later in Chapter 3.

1.3.15. Diesel Exhaust

       In EPA's Diesel Health Assessment Document (HAD),109 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.  An exploratory analysis
was used to characterize a possible risk range 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.
                                       1-25

-------
Final Regulatory Impact Analysis
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.

       Noncancer health effects of acute and chronic exposure to 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.110'
in, 112, 113 The Rfc ig 5 ^m3 for diegel exhaust as measurecj 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 EPA Diesel HAD states, "With DPM [diesel paniculate matter] being a
ubiquitous component of ambient PM, there is an uncertainty about the adequacy of the
existing DE [diesel exhaust] noncancer database to identify all of the pertinent DE-caused
noncancer health hazards" (p. 9-19).

       The Diesel HAD also briefly summarizes health effects associated with ambient
PM and discusses 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 PM2.s NAAQS is designed to provide protection
from the noncancer and premature mortality effects of PM2.5 as a whole, of which diesel
PM is a constituent.

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.
Gasoline exhaust is a ubiquitous source of particulate matter, contributing to the health
effects observed for ambient PM which is discussed extensively in the EPA Particulate
Matter Criteria Document.114  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.s.115  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).  There is also emerging evidence for the mutagenicity and cytotoxicity of
gasoline exhaust and gasoline PM. Seagrave et al. investigated the combined parti culate
                                       1-26

-------
Final Regulatory Impact Analysis
and semivolatile organic fractions of gasoline and diesel engine emissions in various
animal and bioassay tests.116  The authors suggest that emissions from gasoline engines
(including both the semi-volatile organic compounds and the particulate matter) are
mutagenic and can induce inflammation and have cytotoxic effects.

       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 large 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 shortly.  This work
shows how PM emissions vary for light-duty gasoline vehicles (automobiles and light-
duty trucks) for different model years. It also shows how colder temperatures increase
gasoline PM emissions. The  data from this program are being evaluated.  Some source
apportionment studies in various areas of the country, including Denver and California,
show gasoline and diesel PM can result in larger contributions to ambient PM than
predicted by EPA emission inventories.117'118 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 is formed in the atmosphere 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.119'120'121'122'123'124'125

       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 such as toluene (and probably higher molecular
weight non-aromatic hydrocarbons).

       Smog chamber studies show that benzene forms SOA  possibly through reactions
with NOx. Prior smog chamber work126 suggested benzene might be  relatively inert in
forming SOA, although this early study may not be conclusive.  However, the more
recent work shows that benzene does form SOA in smog chambers. This new smog
chamber work  shows that benzene can be oxidized in the presence of NOx to form SOA
                                      1-27

-------
Final Regulatory Impact Analysis


with maximum mass of SOA being 8-25% of the mass of benzene.127 Work is needed to
determine if a tracer compound can be found for benzene SOA which might indicate how
much of ambient SOA comes from benzene.

       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.  128' 129'130' 131; 132>133  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, to be 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 (about 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 be
used to assess whether current treatment of aromatic SOA in the EPA CMAQ model need
to be modified. Along with most of the other state-of-the-science air quality models,
CMAQ predicts low levels of aromatic SOA.

       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 overestimate the amount of SOA in the
atmosphere to the gaseous hydrocarbons studied to date.
                                      1-28

-------
Final Regulatory Impact Analysis
       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.134  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.  Other metals are also being evaluated in various studies.

       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

-------
Final Regulatory Impact Analysis


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://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=12479.

11  EPA.  1996. Guidelines for reproductive toxicity risk assessment. EPA/630/R-96/009.
Federal Register 56274-56322, 31 October 1996.
http: //cfpub. epa. gov/ncea/cfm/recordi spl ay. cfm? dei d=2 838.

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

-------
Final Regulatory Impact Analysis
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) Carcinogenic Effects of Benzene: An Update, National Center for
Environmental Assessment, Washington, DC. EPA600-P-97-001F.  Enter report number
at the following search page,
http://yosemite.epa.gOv/ncepihom/nsCatalog.nsf//SearchPubs7Openform.

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) Carcinogenic Effects of Benzene: An Update, National Center for
Environmental Assessment, Washington, DC. EPA600-P-97-001F. Enter report number
at the following search page,
http://vosemite.epa.gOv/ncepihom/nsCatalog.nsf//SearchPubs7Openform.
                                      1-31

-------
Final Regulatory Impact Analysis
24 U.S. EPA (1998) Carcinogenic Effects of Benzene: An Update, National Center for
Environmental Assessment, Washington, DC. EPA600-P-97-001F. Enter report number
at the following search page,
http://yosemite.epa.gOv/ncepihom/nsCatalog.nsf//SearchPubs7Openform.

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 U.S. EPA (1998) Carcinogenic Effects of Benzene: An Update. EPA/600/P-97/001F. .
Enter report number at the following search page,
http://vosemite.epa.gOv/ncepihom/nsCatalog.nsf//SearchPubs7Openform.

27 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 Rappaport, S.M.; Waidyanatha, S.; Qu, Q.; Yeowell-O'Connell, K.; Rothman, N.;
Smith M.T.; Zhang, L.; Qu, Q.; Shore,  R.; Li, G.; Yin, S. (2005) Protein adducts as
biomarkers of human enzene metabolism. Chem Biol Interact. 153-154:103-109.

30 Lin, Y-S., Vermeulen, R., Tsai, C.H., Suramya,  W., Lan, Q., Rothman, N.,  Smith,
M.T., Zhang, L., Shen, M., Songnian, Y., Kim, S., Rappaport, S.M. (2006) Albumin
adducts of electrophilic benzene metabolitesin benzene-exposed and control workers.
Environ Health Perspec.

31 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.

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

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

34 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.

35 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.
                                      1-32

-------
Final Regulatory Impact Analysis
36 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.

37 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.

38 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.

39 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-tr.pdf.

40 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.

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

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

43 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.

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

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

46 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.

47 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.
                                      1-33

-------
Final Regulatory Impact Analysis
48 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.

49 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.

50 Lan, Q.;, Zhang, L., Li, G., Vermeulen, R., et al. (2004).  Hematotoxicity in Workers
Exposed to Low Levels of Benzene. Science 306:  1774-1776.
51
  Turtletaub, K.W. and Mani, C. (2003). Benzene metabolism in rodents at doses
relevant to human exposure from urban air. Health Effects Inst. Research Report No. 113.

52 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.

53 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.

54 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.

55 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/recordisplav.cfm?deid=54499.

56 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.

57 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.
                                      1-34

-------
Final Regulatory Impact Analysis
58 Delzell, E., Sathiakumar, N., Graff, I, Macaluso, M., Maldonado, G., Matthews, R.
(2006) An updated study of mortality among North American synthetic rubber industry
workers.  Health Effects Institute Report Number 132.

59 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.

60 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.

61 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.

62 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.

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

64 Coggon, D, EC Harris, J Poole, KT Palmer. 2003. Extended follow-up of a cohort of
British chemical workers exposed to formaldehyde. J National Cancer Inst. 95:1608-
1615.

65 Bundesinstitut fur Risikobewertung (BfR) Toxicological Assessment of Formaldehyde.
Opinion of BfR No.  023/2006 of 30 March 2006.
www.bfr.bund. de/cm/290/toxicological_assessment_of_formal dehyde.pdf

66 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.

67 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.

68 Chemical Industry Institute of Toxicology (CUT). 1999. Formaldehyde: Hazard
characterization and dose-response assessment for carcinogenicity by the route of
inhalation.  CUT, September 28, 1999. Research Triangle Park, NC.

69 Health Canada (2001) Priority Substances List Assessment Report. Formaldehyde.
Environment Canada, Health Canada, February 2001.  The document may be accessed at
                                      1-35

-------
Final Regulatory Impact Analysis
http://www.hc-sc.gc.ca/ewh-setnt/pubs/contatninants/psl2-
Isp2/formaldehyde/index_e.html .

70 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)

71 Agency for Toxic Substances and Disease Registry (ATSDR). Toxicological Profile
for Formaldehyde. Public Health Service, U.S. Department of Health and Human
Services, Atlanta, GA. Available at http://www.atsdr.cdc.gov/toxprofiles/tpl 11 .html.

72 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.

73 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.

74 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.

75 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.

76 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.

77 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.

78 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

79 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.
                                      1-36

-------
Final Regulatory Impact Analysis
80 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.

81 Agency for Toxic Substances and Disease Registry (ATSDR). Draft Toxicological
Profile for Acrolein. 2005 Public Health Service, U.S. Department of Health and Human
Services, Atlanta, GA. Available at http://www.atsdr.cdc.gov/toxprofiles/.

82US EPA .2004. Toxicological Review of Naphthalene. (Reassessment of inhalation
cancer risk), Environmental Protection Agency, Integrated Risk Information System
(IRIS), Research and Development, National Center for Environmental Assessment,
Washington, DC http://www.epa.gov/iris/subst/0436.htm

83 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

84 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

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

86 USEPA. 1998. Toxicological Review of Naphthalene. Environmental Protection
Agency, Integrated Risk Information System (IRIS), Research and Development,
National Center for Environmental Assessment, Washington, DC
http://www.epa.gov/iris/subst/0436.htm.

87 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.

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

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

90 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,
                                      1-37

-------
Final Regulatory Impact Analysis
OH http://www.epa.gov/iris/subst/0051 .htm.

91 Agency for Toxic Substances and Disease Registry (ATSDR). Toxicological Profile for
Ethylbenzene. 1999 Public Health Service, U.S. Department of Health and Human
Services, Atlanta, GA. Available at http://www.atsdr.cdc.gov/toxprofiles/.

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

93 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.

94 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.

95 Agency for Toxic Substances and Disease Registry (ATSDR). Toxicological Profile for
n-Hexane. 1999 Public Health  Service, U.S. Department of Health and Human Services,
Atlanta, GA. Available at http://www.atsdr.cdc.gov/toxprofiles/.

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

97 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/regs/fuels/ostpfm.pdf.

98 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.

99 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.

100 Agency for Toxic Substances Disease Registry (1992) Toxicological profile for
styrene. Atlanta: ATSDR. http://www.atsdr.cdc.gov/toxprofiles/tp53.html.

101 EPA 2005 "Full IRIS Summary for Toluene (CASRN 108-88-3)" Environmental
Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
                                      1-38

-------
Final Regulatory Impact Analysis
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati,
OH http://www.epa.gov/iris/subst/0118.htm.

102 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.
103 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.
104 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.
105 EPA Toxicological Review of Xylenes. January 2003. EPA 635/R-03/001.

106 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.
107 Agency for Toxic Substances and Disease Registry (ATSDR). Toxicological Profile
for Xylene. 2005 Public Health Service, U.S. Department of Health and Human
Services, Atlanta, GA. Available at http://www.atsdr.cdc.gov/toxprofiles/.

108 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.

109 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 .

110 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.

111 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.

112 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.
113
   Nikula, KJ; Snipes, MB; Barr, EB; et al. (1995) Comparative pulmonary toxicities and
                                      1-39

-------
Final Regulatory Impact Analysis
carcinogenicities of chronically inhaled diesel exhaust and carbon black in F344 rats.
Fundam Appl Toxicol 25:80-94.

114 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.

115 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.

116 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.

117 Watson,!., Fujita,E., Chow,J., Zielinska,B., Richards,L., Neff,W., Dietrich,D.
Northern Front Range Air Quality Study Final Report: Volume 1. June 30, 1998. For
Colorado State University, Cooperative Institute for Research in the Atmosphere, by
Desert Research Institute, Reno, NV. This document is available in the EPA Docket as
EPA-HQ-OAR-2005-0036-1179.

118 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.

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

120 Watson,!., Fujita,E., Chow,J., Zielinska,B., Richards,L., Neff,W., Dietrich,D.
Northern Front Range Air Quality Study Final Report: Volume 1. June 30, 1998. For
Colorado State University, Cooperative Institute for Research in the Atmosphere, by
Desert Research Institute, Reno, NV. This document is available in the EPA Docket as
EPA-HQ-OAR-2005-0036-1179.

121 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.

122 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.
                                      1-40

-------
Final Regulatory Impact Analysis
123 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.

124 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.

125 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.

126 Izumi, L. and T. Fukuyama.  1990, Photochemical aerosol formation from aromatic
hydrocarbons in the presence of NOx., Atmospheric Environment, 24A, 1433.

127 Martin-Reviego, M. and K. Wirtz, 2005.  "Is benzene a precursor for secondary
organic aerosol?" Environmental Science and Technology, 39, 1045-1054.
1 98
   Kleindienst, T.E., T.S. Conver, C.D. Mclver, and E.G. Edney. 2004. Determination of
secondary organic aerosol products from the photooxidation of toluene and their implications in
ambient PM2 5. J. Atmos. Chem. 47, 79-100.

   Edney. E.O., T.E. Kleindienst, M. Jaoui, M. Lewandowski, J.H. Offenberg, W. Wang, M.
Claeys. 2005. Formation of 2-methyl tetrols and 2-methylglyceric acid in secondary organic
    J                         J                 J O J                   JO
aerosol from laboratory irradiated isoprene/NOx/SO2/air mixtures and their detection in ambient
PM25 samples collected in the Eastern United States. Atmospheric Environment 39: 5281-5289.


   Claeys, M., R. Szmigielski, I. Kourtchev, P. Van der Veken, R. Vermeylen, W. Maenhaut, M.
Jaoui, T.E. Kleindienst, M. Lewandowski, J.H. Offenberg, E.O. Edney. 2007.
Hydroxydicarboxylic acids: Markers for secondary organic aerosol from the photooxidation of ot-
pinene. Environ. Sci.  Technol. (Web Edition, 01/23/2007).

   Jaoui, M., T.E. Kleindienst, M. Lewandowski, J.H. Offenberg, E.O. Edney. 2005.
Identification and quantification of aerosol polar oxygenated compounds bearing carboxylic or
hydroxyl groups. 2. Organic tracer compounds from monoterpenes. Environ. Sci.
Technol. 39:5661-5673.

132 Edney. E.G., T.E. Kleindienst, T.S. Conver, C.D. Mclver, and W.S. Weathers. 2003. Polar
organic oxygenates in PM2 5 at a southeastern site in the United States. Atmospheric Environment
37, 3947-3965.

   Lewandowski, M., M. Jaoui, T.E. Kleindienst, J.H. Offenberg, E.O. Edney. 2007.
Composition of PM25 during the summer of 2003 in Research Triangle Park, North Carolina.
Atmos. Environ,  (in press; available on line 01/30/2007).
                                        1-41

-------
Final Regulatory Impact Analysis
134 Schauer, James J., Glynis C Lough, Martin M Shafer, William F Christensen, Michael
F Arndt, Jeffrey T DeMinter and June-Soo Park. 2006.  Characterization of metals
emitted from motor vehicles. Health Effects Institute Research Report Number 133.
                                       1-42

-------
Final Regulatory Impact Analysis
                           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	4
        2.1.1.2  Portable Fuel Containers	9
     2.1.2 Emission Reductions of Criteria Pollutants Resulting From Controls	10
     2.1.3 Strengths and Limitations of Criteria Pollutant Inventories	20
   2.2  Air Toxics	22
     2.2.1 Emission Inventories Used in Air Quality Modeling	22
           2.2.1.1.1  Highway Vehicles	23
           2.2.1.1.2  Nonroad Equipment in theNonroad Model	31
           2.2.1.1.3  Commercial Marine Vessels, Locomotives and Aircraft	32
           2.2.1.1.4. Portable Fuel Containers	33
           2.2.1.1.5. Gasoline Distribution	35
           2.2.1.1.6. Other Stationary Sources	36
           2.2.1.1.7  Precursor Emissions	38
        2.2.1.2  Trends in Air Toxic Emissions	42
           2.2.1.2.1  Emission Trends Without Controls	42
           2.2.1.2.2  Impact on Inventory of Controls	58
     2.2.2 Emission Reductions from Controls	64
        2.2.2.1  Methodology Changes from Air Quality Inventories	64
           2.2.2.1.1  Highway Vehicles	64
           2.2.2.1.2  Nonroad Equipment	64
           2.2.2.1.3  Portable Fuel Containers	65
           2.2.2.1.4  Gasoline Distribution	65
        2.2.2.2  Estimated Reductions for Air Toxic Pollutants of Greatest Concern.... 65
           2.2.2.2.1  Fuel Benzene Standard	66
           2.2.2.2.2  Cold Temperature VOC Emission Control	75
           2.2.2.2.3  Portable Fuel Container Control	79
           2.2.2.2.4  Cumulative Reductions of Controls	82
   2.3  Potential  Implications of New Emissions Data for Inventories	89
     2.3.1 Newer Technology Light Duty Vehicles	89
     2.3.2 Heavy-Duty Vehicles (CRC E-55/E-59)	90
     2.3.3 Small Spark Ignition Engines	91
     2.3.4 Nonroad CI engines	93
   2.4  Description of Current Mobile Source Emissions Control Programs that Reduce
          MSATs	94
     2.4.1 Fuels Programs	94
        2.4.1.1  RFG	94
        2.4.1.2  Anti-dumping	95
        2.4.1.3  2001 Mobile Source Air Toxics Rule (MSATI)	95
        2.4.1.4  Gasoline Sulfur	96
        2.4.1.5  Gasoline Volatility	96
        2.4.1.6  Diesel Fuel	96
                                       2-1

-------
Final Regulatory Impact Analysis


        2.4.1.7 Phase-Out of Lead in Gasoline	97
     2.4.2 Highway Vehicle and Engine Programs	97
     2.4.3 Nonroad Engine Programs	99
        2.4.3.1 Land-based Diesel Engines	100
        2.4.3.2 Land-Based SI Engines	100
           2.4.3.2.1  Large Land-Based SI Engines	100
           2.4.3.2.2  Recreational Vehicles	101
           2.4.3.2.3  Small Land-Based SI Engines	101
        2.4.3.3 Marine Engines	101
           2.4.3.3.1  Marine SI Engines	102
           2.4.3.3.2  Marine Diesel Engines	102
        2.43 A Locomotives	103
        2.4.3.5 Aircraft	103
     2.4.4 Voluntary Programs	103
                                       2-2

-------
Final Regulatory Impact Analysis
                    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 final rule promulgation, because
of the lead time required to conduct air quality, exposure, and risk analyses. Thus, these
inventories do not include revised estimates of emissions using new fuel quality estimates
developed for the Renewable Fuel Standard Program, as discussed below.  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

       For the final rule, we have revised the emission inventories to reflect conditions
anticipated under the Renewable Fuel Standard (RFS) program.  The RFS program was
mandated by the Energy Policy Act of 2005 in order to increase national consumption of
renewable fuels. In September 2006, EPA issued a proposed rule to implement the RFS
program for 2007 and beyond.1 The RFS proposal analyzed several different scenarios of
increased ethanol use and developed county-level fuel properties specific to each
scenario.

       In one particular RFS scenario, we estimated county-level fuel properties by
allocating the Energy Information Agency's forecast of 9.6 billion gallons of national
ethanol consumption in 2012, attributing as much as possible for use as an oxygenate in
reformulated gasoline.  For purposes of this rule, we have selected this scenario as the
most likely ethanol volume and distribution in 2012, and have therefore adopted those
fuel properties as the new baseline fuel for MSAT inventories used to evaluate the cost-
effectiveness  of the standards being finalized in this rule. In the discussion that follows,
the new MSAT baseline fuel is referred to as the "RFS fuel". The RFS Draft Regulatory
Impact Assessment (DRIA) contains a detailed discussion of the effects of ethanol fuel
on gasoline properties and the methods by which we derived RFS county-level fuel
properties.2

       Though cost-effectiveness inventories in both the RFS proposal and the MSAT
final rule reflect RFS fuels, there are slight differences in other methodologies used to
estimate the emissions inventories. However, the differences are minor and have  little
impact on emission reductions used to evaluate cost-effectiveness.
                                       2-3

-------
Final Regulatory Impact Analysis
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).3
MOBILE6.2 uses emission factors obtained through the analysis of emissions data
collected from vehicle emission research.4 The emission factors reflect impacts of
vehicle standards as well as current and planned inspection and maintenance programs.
They also reflect impacts of changes in properties of gasoline and diesel fuels.  Impacts
of alternative fueled vehicles and engines (e.g. liquid propane, compressed natural gas,
methanol) are negligible in NMIM. 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. NEMS projects VMT for personal
travel based on demographic effects and economic influences such as estimated fuel costs
and disposable income, and projects commercial truck travel based on economic factors
such as industrial output and demand. This is the same approach used in the Clean Air
Interstate Air Quality (CAIR) rule.5 As mentioned  above, county-level fuel properties
contained in the public release version of NMIM were revised to RFS fuel.

       Analysis of vehicle emission certification data submitted by vehicle
manufacturers to EPA as part of requirements to comply with requirements for cold
temperature carbon monoxide (CO) standards, as well as surveillance testing data from
the California Air Resources Board, indicated that MOBILE6.2 was substantially
underestimating start emissions at cold temperatures for Tier 1 and later vehicles. This
data was supplemented with test data collected by EPA at Southwest Research Institute
(SwRI)6 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.7

       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.
Available bag emission data indicated that at 20 °F, as in the standard Federal Test
Procedure (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.

       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 EPA testing 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.
                                       2-4

-------
Final Regulatory Impact Analysis
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
                                              2-5

-------
Final Regulatory Impact Analysis
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 documentation.8  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 0°, 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 of Tier 2 high emitting vehicles
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 to develop the additive factor for all Tier 2 high-emitting
                                       2-6

-------
Final Regulatory Impact Analysis
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.
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
                                         2-7

-------
Final Regulatory Impact Analysis
       The above tables and the new emission standard were used to determine the
effects of the cold temperature 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. With the exception of using RFS fuel, no
modification to any other components of NMEVI is needed to calculate these inventories.
The inventories are presented in Chapter 2.1.2.
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
3
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
                                        2-8

-------
Final Regulatory Impact Analysis
2.1.1.2 Portable Fuel Containers

       In 1999, California's Air Resources Board (ARE) proposed a methodology to
estimate annual VOC 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 NONROAD2005
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 stored in each PFC, EPA
estimated the number of PFCs in use (each season) with gasoline 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.  (The estimated RVPs for future years include
the effects of the Renewable Fuels Standard.) 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.9 This analysis does not consider usage of PFCs with diesel or
kerosene fuels, as these fuels contribute minimally to evaporation emissions  due to the
very low volatilities of these fuels.

       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.

       Twelve states plus the District of Columbia (California, Connecticut, Delaware,
Maine, Maryland, New Hampshire, New Jersey, New York, Ohio, Pennsylvania, Texas,
Virginia, and Washington DC) already have or will implement controls on the design of
PFCs that will reduce HC emissions. Additionally, three other states (Massachusetts,
Rhode Island, and Vermont) are also planning to adopt the California PFC program.
Inventories include the impacts of these programs, as described in a technical support
document (EPA, 2006, Report No. EPA420-R-07-001).

       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.
                                       2-9

-------
Final Regulatory Impact Analysis
       To estimate the VOC emissions from PFCs upon implementation of the final rule,
we made the following three changes to our inventory estimates:

1.      Since the final rule makes it unlikely for a newly designed PFC 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 15 states plus the
       District of Columbia that are adopting the California PFC rules already had this
       change applied.  So, this affected the VOC emissions from only PFCs in the other
       35 states.)

2.      This final rule also produces changes to the design of the individual PFCs that are
       expected to reduce the spillage by 50 percent when these PFCs 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 PFCs in the remaining 35 states contributed to our estimated
       reductions of spillage.

3.      Finally, the final rule includes a maximum emission rate of 0.3 grams per gallon
       per day for the new PFCs. We used this emission standard to estimate the total
       permeation plus evaporative emissions from each newly designed PFC.  Only
       California has adopted this requirement. Thus, the effect of this final 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. The national pre- and post-control inventories are presented in Chapter 2.1.2
below.
2.1.2  Emission Reductions of Criteria Pollutants Resulting From Controls

2.1.2.1 Light-Duty Gasoline Vehicles

       We are finalizing as proposed 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 (including medium-duty
passenger vehicles; i.e., "MDPVs"). 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
                                      2-10

-------
Final Regulatory Impact Analysis
2 vehicles only.  MOBILE6.2 was then used with NMIM (using fuel properties which
reflect RFS, as described in Section 2.1.1) to generate county and nationwide inventories
for the control case. When the standard is fully phased in we expect a 60% reduction in
start emissions in gasoline-fueled vehicles that have a gross vehicle weight rating
(GVWR) of less than or equal to 6000 Ibs and a 30% reduction for 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. Table 2.1 .-10 shows the impacts on a state-by-
state basis in year 2030.
                                       2-11

-------
Final Regulatory Impact Analysis
   Table 2.1-7.  PFC Emissions (Tons VOC per Year) by Source Type (for 1990)

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
Rl
SC
SD
TN
TX
UT
VA
VT
WA
Wl
WV
WY
50-State
Refilling PFC at Pump
Vapor Displ
224.8
24.8
279.1
105.7
1,532.1
202.7
123.2
36.6
7.6
933.1
390.9
58.1
50.6
405.1
241.4
99.6
93.5
129.1
168.9
40.7
226.0
199.0
316.9
181.4
97.0
212.6
26.4
55.0
123.4
44.1
332.9
58.0
517.1
407.9
17.2
507.3
139.6
133.8
419.5
28.3
207.8
20.9
237.0
875.0
70.8
18.7
309.8
225.6
65.4
166.3
14.8
11,403.3
Spillage
15.0
1.9
22.9
8.5
133.9
18.9
12.0
3.1
0.7
72.5
32.4
4.0
4.9
36.3
19.8
8.3
8.5
11.1
12.1
4.3
21.1
19.1
29.6
15.5
7.4
19.0
2.6
5.2
10.6
4.5
30.0
5.2
47.5
31.5
1.8
41.1
12.1
12.8
38.5
2.7
15.1
2.0
18.6
67.6
6.7
1.9
27.6
20.5
5.4
16.4
1.5
972.1

Spillage
During
Transport
447.1
60.1
647.9
262.7
3,760.8
536.5
342.7
89.1
25.0
2,055.5
930.8
112.9
146.3
1,058.5
578.2
248.5
247.9
340.2
370.7
130.7
597.8
556.1
886.3
463.1
230.6
560.9
81.9
154.0
295.8
131.0
857.4
155.6
1,414.3
911.8
54.1
1,188.5
352.5
373.6
1,132.0
80.8
438.3
62.1
553.4
1,954.5
201.4
57.6
786.6
595.1
170.6
488.5
48.1
28,226.3
Refueling Equipment
Vapor
Displ
224.8
24.8
279.1
105.7
1,532.1
202.7
123.2
36.6
7.6
933.1
390.9
58.1
50.6
405.1
241.4
99.6
93.5
129.1
168.9
40.7
226.0
199.0
316.9
181.4
97.0
212.6
26.4
55.0
123.4
44.1
332.9
58.0
517.1
407.9
17.2
507.3
139.6
133.8
419.5
28.3
207.8
20.9
237.0
875.0
70.8
18.7
309.8
225.6
65.4
166.3
14.8
11,403.3
Spillage
1,010.8
103.2
1,630.1
533.4
9,284.9
1,319.4
837.2
217.9
56.6
5,050.7
2,234.7
285.3
316.0
2,458.1
1,353.8
541.9
567.2
727.8
771.4
297.7
1,520.6
1,322.3
1,966.1
992.3
476.0
1,271.6
160.5
336.4
759.6
299.6
2,041.2
358.7
3,095.2
2,179.0
103.9
2,843.0
824.4
864.5
2,644.5
188.9
1,066.9
124.8
1 ,245.3
4,645.6
418.4
127.6
1,986.7
1,399.7
345.8
1,089.3
92.7
66,389.1

Permeation Plus
Evaporation
4,286.7
776.6
3,936.1
2,813.4
19,682.1
2,137.2
1,422.5
514.9
235.1
14,664.5
5,918.5
1 ,208.2
780.1
5,764.9
3,914.8
1 ,886.4
1 ,457.6
2,914.7
5,178.9
620.4
2,528.1
2,561.3
5,253.7
3,281.1
2,997.4
3,427.2
506.1
911.4
1 ,362.6
572.4
4,049.9
1 ,050.8
8,473.6
6,950.0
302.1
7,500.9
2,322.6
1 ,889.7
6,498.5
422.5
3,981.0
398.1
4,944.1
15,730.9
1,208.1
296.3
3,853.6
3,174.0
1 ,700.5
2,512.5
265.7
181,040.0

Totals by
State
6,209.2
991.3
6,795.1
3,829.4
35,925.8
4,417.5
2,860.8
898.2
332.7
23,709.5
9,898.3
1,726.6
1,348.5
10,127.9
6,349.3
2,884.3
2,468.2
4,252.1
6,670.9
1,134.5
5,119.6
4,856.7
8,769.4
5,114.9
3,905.5
5,704.1
803.9
1,516.8
2,675.3
1,095.5
7,644.2
1,686.2
14,064.8
10,888.3
496.2
12,588.0
3,790.6
3,408.1
11,152.6
751.5
5,916.9
628.9
7,235.5
24,148.7
1,976.3
520.9
7,274.1
5,640.6
2,353.1
4,439.2
437.7
299,434.1
                                      2-12

-------
Final Regulatory Impact Analysis
         Table 2.1.-8. 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
2,990,760
2,614,987
2,538,664
2,878,836
Tons With Standard
N. A.
2,839,012
2,293,703
2,009,301
1,996,074
Reduction
N.A.
151,748
321,284
529,363
882,762
                                    2-13

-------
 Final Regulatory Impact Analysis
  Table 2.1.-10.  Impacts on State Light Duty Vehicle and Truck VOC Emissions of
       20 °F FTP Emission Standard for Non-Methane Hydrocarbons in 2030.
State
AL
AK
AZ
AR
CA
CO
CT
DE
DC
FL
GA
HI
ID
IL
IN
IA
KS
KY
LA
ME
MD
MA
MI
MN
MS
MO
MT
NE
NV
NH
NJ
NM
NY
NC
ND
OH
OK
OR
PA
RI
SC
SD
TN
TX
UT
VT
VI
WA
WV
WI
WY
Reference Case
Tons
52,985
11,605
50,655
30,893
259,253
61,855
28,766
7,213
3,229
123,002
100,284
7,835
21,439
107,467
85,144
38,982
31,740
48,011
36,806
16,942
45,754
44,407
133,830
86,476
25,290
71,439
16,326
22,897
28,102
15,434
54,869
31,625
112,589
94,614
11,222
115,095
46,290
66,957
105,046
9,036
47,950
11,920
70,526
159,952
36,024
9,873
80,579
108,386
16,993
64,663
10,566
Control Case
Tons
41,636
6,299
39,988
23,185
185,702
40,187
17,706
4,639
2,146
110,498
75,270
7,626
13,588
67,221
57,529
25,254
22,190
32,867
30,134
10,247
29,230
25,717
86,171
51,148
19,642
49,467
10,015
15,077
20,771
9,413
35,834
22,152
67,387
69,429
6,752
73,824
34,712
46,226
67,864
5,641
36,058
7,443
51,999
126,799
24,050
5,906
53,729
74,481
10,833
37,816
6,574
Reduction
in Tons (a)
11,349
5,306
10,667
7,708
73,551
21,667
11,059
2,574
1,082
12,504
25,014
209
7,851
40,245
27,615
13,729
9,550
15,144
6,672
6,695
16,525
18,690
47,659
35,328
5,648
21,972
6,311
7,819
7,330
6,022
19,035
9,473
45,202
25,185
4,470
41,271
11,578
20,731
37,183
3,395
11,892
4,476
18,528
33,154
11,974
3,967
26,850
33,905
6,160
26,847
3,992
Percent
Reduction
21
46
21
25
28
35
38
36
34
10
25
3
37
37
32
35
30
32
18
40
36
42
36
41
22
31
39
34
26
39
35
30
40
27
40
36
25
31
35
38
25
38
26
21
33
40
33
31
36
42
38
(a) Values calculated prior to rounding reference and control values.
                                        2-14

-------
Final Regulatory Impact Analysis


       Test data show that the 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.10

       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, 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.n  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.
                                       2-15

-------
Final Regulatory Impact Analysis
     Figure 2.1.-1. FTP Bag 1 PM Emissions vs. Temperature, Tier 2 Vehicles
                      I    I   I   I
0  20 40 60
    ill    l   l    l   l
              E
              O)
             CL   -2 -
                  -4 -
              as
             m
                  -6 -
                 -10 -
Vehicle 4
0
8
0

o
Vehicle 1
8
8 o
0 9
o
o
Vehicle 5
0
8


Vehicle 2
o
o
o
Vehicle 6
8.
o o
o
o o

Vehicles
o
©
o
o

                      i   i   i   r
                                                               - -2
                                                               - -6
                                                               - -10
                                                 i   i   i    r
                      0  20  40 60                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.
                                       2-16

-------
Final Regulatory Impact Analysis
  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.12 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 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.
                                      2-17

-------
Final Regulatory Impact Analysis
       Using this number, the expected reductions in PM from this rule are estimated to
be 7,068 tons in 2015, 11,646 tons in 2020 and 19,421 tons in 2030. These calculations
provide initial evidence that the potential public health impacts of this final rule are
substantial.

       In subsequent test programs demonstrating the feasibility of the NMHC standards
in this final rule, the test vehicles exhibited substantial reductions in PM emissions as
well.  The test results from the two feasibility vehicles fall within the range of those
derived from the SwRI test program. These PM emission reductions at 20° F were of
similar magnitude as those predicted by the above calculation.  Furthermore, examining
the feasibility demonstration results, the PM/NMHC ratio of the emission reductions
were both close to the value of 0.022 used in the above calculation, spanning either side
of the original number (0.010-0.025).A In the first feasibility test program, the vehicle
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.  That feasibility program was a "proof of concept" study that did not
have the ability to fully explore ideal control coordination  and sizing of the emission
control  system. In the second feasibility study, the vehicle only received recalibration to
achieve emission reductions, which is likely to be more representative of the emission
control technologies that will be employed for the majority of vehicles. Despite different
technologies being used in the feasibility tests, 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 two feasibility test programs, 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).
2.1.2.2  Portable Fuel Containers

      The PFC controls in this final rule will also reduce emissions of hydrocarbons.  As
noted in Section 2.1.1.2, fifteen states plus the District of Columbia have adopted
controls on PFCs independent of the controls being finalized in this rule. 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,
 ' This range derives from the feasibility tests with the lowest measured NMHC emissions.


                                       2-18

-------
Final Regulatory Impact Analysis
      a scenario in which only those 15 states plus DC have implemented PFC controls
      illustrated with the solid (blue) line, and

      a scenario in which the PFC controls are implemented nationwide illustrated with
      the dashed (red) line

                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
o
n
o
o
n


— * —


- - • Bas
_ O*~


r - - - * *
v
\

e
:e Programs
^™ ~ Nationwide


•
— «
* *
_^^^

\
\
\- — —

^ * *
- *
^^

— — "" ""


1990 2000 2010 2020 2030
Calendar Year
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 control program to a reference case scenario that includes only State level
controls.
                                       2-19

-------
Final Regulatory Impact Analysis
           Table 2.1.-11.  Nationwide Annual PFC VOC Emissions (tons)
Calendar
Year
1999
2007
2010
2015
2020
2030
With NO EPA
PFC Controls
325,030
327,320
316,756
329,504
353,470
402,916
With EPA
PFC Controls
NA
NA
256,175
127,157
137,175
157,661
Reduction
NA
NA
60,580
202,347
216,294
245,255
2.1.3  Strengths and Limitations of Criteria Pollutant Inventories

       As previously discussed, the MSAT final rule inventories were estimated using
fuel properties developed for the RFS proposed rule. Because the RFS and MSAT
inventories were developed in relatively close proximity, we highlight in this section
some minor differences in methodologies, as well as uncertainties related to the RFS fuel.
Though these methodologies contribute to different baseline RFS and MSAT inventories,
they have little impact on our estimates of emission reduction benefits associated with
this MSAT final rule.

       Future Volume of Renewable Fuel - Under the RFS mandate, a minimum volume
of ethanol must be blended in gasoline. However, the Energy Information Agency (EIA)
has forecasted that market forces alone will push ethanol use well beyond the minimum
volume required by the RFS mandate13. The volume of renewable fuel forecasted by
EIA, and not the RFS program mandate, was used as the baseline for developing RFS
fuel properties used in MSAT  inventories. Though there are uncertainties related to the
future volume of renewable fuel use (and regional allocation), the effects on the emission
reduction benefits achieved by the MSAT final rule are likely minimal.  Furthermore, as
presented in the RFS Draft RIA (DRIA)2, inventories for criteria pollutants never differ
by more than a few percent between the RFS mandate volume scenario (7.2 billion
gallons of national ethanol  use) and the EIA-predicted scenario (9.6 billion gallons of
national ethanol use).

       Ethanol Effects on Gasoline Properties - The MSAT rule inventories are based
on fuel properties  estimated under the RFS program. In the RFS draft regulatory impact
analysis, we based our assessment of the effects of ethanol on gasoline fuel properties on
annual fuel survey data provided by the Alliance of Automobile Manufacturers. We
limited the analysis to cities for which data from both ethanol-blend and non-ethanol
gasoline samples were available.  These criteria reduced the data to samples from four
cities, limiting the national geographic representation of the fuel effects.  In addition,
there was no distinction to indicate whether the fuel came from multiple refineries within
any given city, which eliminates refinery-specific effects. However, we checked the
results against the AAM data from all U.S. cities, comparing all conventional gasoline
                                      2-20

-------
Final Regulatory Impact Analysis
non-ethanol blends to all conventional gasoline ethanol blends. The results were very
similar to those from the four cities.

       Seasonal and Permeation Effects- For MSAT inventories, we interpolated
summer and winter fuel properties to all 12 months and ran each month in NMEVI
individually.  Due to time constraints during development of RFS proposal inventories,
we ran NMEVI for only January and July, using fuel survey data collected in summer and
winter, and assumed that emissions for those two months could be extrapolated to
represent the entire year. We estimated RFS annual emissions inventories by summing
the two monthly inventories and multiplying by six. For RFS, we also added the effect of
ethanol on permeation from onroad non-exhaust emissions. Again, these different
methodologies have minimal effect on the emissions benefits associated with this final
rule.

       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.14'15

       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.  Furthermore, emission rates  were
modeled for only a subset of the total number of counties.  Therefore, we do not fully
capture all local conditions, introducing additional uncertainty into the inventories.

       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.

       Finally, the MSAT inventories used the fuel effects contained in MOBILE6.2.  In
the RFS proposal, we accounted for uncertainties in MOBILE6.2 fuel effects by adjusting
                                       2-21

-------
Final Regulatory Impact Analysis


the model output for exhaust VOC and NOx emissions by applying EPA Predictive
Model fuel effects instead. The MSAT inventories do not use the Predictive Model
effects since the use of the Predictive Model would have little impact on estimates of
emission benefits of this rule.

      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.

      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 PFC population in the rest of the country might not be
exactly like California's, 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.16 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 final rule are discussed in
detail in the EPA Technical Report, "National Scale Modeling of Air Toxics for the Final
Mobile Source Air Toxics Rule; Technical Support Document," Report Number EPA-
454/R-07-002. All underlying data and summary statistics are included in the docket for
                                      2-22

-------
Final Regulatory Impact Analysis


this rule. Major revisions have been made to the inventories used for air quality,
exposure, and risk modeling since proposal.  These revisions include:

          •  Revisions to cold temperature start emissions for gaseous air toxics in Tier
              1 and later highway gasoline vehicles
          •  Estimation of air toxic emissions for nonroad equipment using
             NMEVI2005 rather than NMIM2004
          •  Inclusion of air toxic emissions from portable fuel containers
          •  Revision of the benzene and naphthalene inventories for gasoline
             distribution based on recent analysis of benzene in gasoline vapor emitted
             during the distribution process

While cold temperature emissions and portable fuel container emissions were included in
analyses of emission benefits and cost-effectiveness for the proposal, the proposal did not
use NMTM2005 for nonroad equipment or include changes to the gasoline distribution
emissions estimates based on recent analyses. While the air quality modeling inventories
for the final rule included the above improvements, it did not include impacts of the
renewable fuel standard, as the inventories were developed well in advance of the
proposal for that standard.  Furthermore, the modeling accounted only for the 0.62
percent  standard, but not the 1.3 vol% maximum average. Thus, the emission reductions
from highway vehicles and other sources attributable to the fuel benzene standard are
underestimated  in many areas of the country, particularly in areas where fuel benzene
levels were highest without control, such as the Northwest. It should be noted that the
inventories used in the proposal were presented in a peer reviewed journal article.17

       The following sections summarize the methods used to develop the air quality
modeling inventories, including details of the major revisions listed above, and also
present inventory 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 2010 in order
to better assess  emission trends over time. Control case modeling which included
cumulative impacts of the controls being finalized in this rule was done for 2015, 2020
and 2030. 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.

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

2.2.1.1.1 Highway Vehicles

       Highway gasoline vehicle inventories for all emission types except refueling were
developed using a modified version of NMIM2005.18'19'20 NMIM develops inventories
using EPA's MOBILE6.2 emission factor model for highway vehicles, EPA's
                                       2-23

-------
Final Regulatory Impact Analysis
NONROAD 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.
Refueling emission estimates for 2015 and later years were carried over from the
inventories used for air quality modeling in the proposal. For 1999 and 2010, benzene
refueling emission estimates were not available, so benzene refueling emissions were
backcast from 2015, using ratios of VOC refueling emissions in 1999 or 2010 to 2015
VOC.  The approach used to do this is discussed in detail in the technical support
document.

       NMEVI was modified to include the hydrocarbon start emission adjustment factors
discussed in Section 2.1.1.1.   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.21'22 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 late model
vehicles at the same temperature ranges cited above.23'24'25 These results can be used to
reevaluate the validity of the assumption discussed above. 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, 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.
                                       2-24

-------
Final Regulatory Impact Analysis
       A third source of data is testing done by Southwest Research Institute for EPA on
four model year 2005 vehicles - a Ford F-150, a Mazda 3, a Honda Odyssey and a
Chevrolet Equinox.26 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  be 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 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.-1 depicts the relationship between
carbonyl compounds and NMHC.

   Figure 2.2.-1. Regressions of Carbonyl Emissions Versus NMHC for Chevrolet
          Trailblazer Recalibrated to Meet Cold Temperature Standard.
                               Acetaldehyde vs. NMHC
                        1000   2000
3000
NMHC
4000   5000   6000
                                      2-25

-------
Final Regulatory Impact Analysis
                                Acrolein vs NMHC
                        1000   2000   3000   4000
                                 NMHC (mg/mi)
5000  6000
Formaldehyde vs. NMHC
10 -, 	 -

E °
| 6
O 4
O
I 2
o
•
+ v- 7F OCIY+ fi P17 a. '

f » R^= 0.0082
* I










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.

       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.  These parameters are stored in
the NMIM database.  The fuel parameters are: sulfur content, olefins content, aromatics
content, benzene content, E200 value, E300 value, oxygenate content by type, and
oxygenate sales fraction by type.8 Since these fuel parameters are area-specific, EPA
developed county-level inputs for each of these parameters. 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
 ! E200 and E300, represent the percentage of vapor that gasoline produces at 200 and 300 °F, respectively.
                                       2-26

-------
Final Regulatory Impact Analysis


surveys.  Documentation for the National Emissions Inventory (NEI)27 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/

       Although fuel parameter data were prepared for only two seasons (summer and
winter), NMIM uses monthly rather than seasonal fuel parameters, and parameters for
spring and fall months are estimated by interpolating from summer and winter data.  In
addition,  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
NMIM, 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). These adjustment factors
were developed using results of several refinery modeling  analyses conducted to assess
                                              o£ OQ ^n
impacts of fuel control programs on fuel properties.  '  '   The database used for this
assessment assumes no Federal ban  on MTBE, but does include State bans. Also, it did
not account for recent increases in the use of ethanol oxygenated gasoline, the renewable
fuels mandate in  the recent Energy Policy Act, or the 1.3 vol% maximum average fuel
benzene level.

       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 stored in the NMIM database, 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. Polycyclic 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. 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 (i.e., non-oxygenated) gasoline and  gasoline oxygenated with
MTBE or ethanol, separate ADDITIONAL HAPS input data 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).
                                      2-27

-------
Final Regulatory Impact Analysis
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
Vehicle miles traveled used in this assessment were those developed for the Clean Air
Interstate Air Quality Rule (CAIR).
31
       For years 2015, 2020, and 2030, inventories were developed that reflected the
cumulative impacts of the controls being finalized in this rule. These control case
inventories included all the pollutants in Table 2.2-1.

       To develop these inventories, NMIM was rerun with revised gasoline fuel
parameter inputs for fuel benzene and aromatics levels, as well as estimated emission
reductions from new start emission standards for hydrocarbons.  The fuel parameter
inputs were revised based on refinery modeling done for the proposed rule, rather than
the final rule refinery modeling discussed in Chapter 6 of the this document.  As part of
the refinery modeling, average fuel properties under a 0.62% fuel benzene standard, with
no maximum average level, 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 standards
were used to develop multiplicative factors which were applied to the reference case fuel
benzene levels for each county in the NMTM 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.  Table 2.2.-3
compares average fuel benzene levels for each PADD used in the air quality modeling
                                       2-28

-------
Final Regulatory Impact Analysis
inventories, compared to levels predicted by refinery modeling for the final rule, which
assumes a 1.3 vol% maximum average.  If the refinery modeling data had been available
to be used in the air quality modeling inventories, benzene emission reductions from the
fuel standard would have been significantly greater in PADDs 2 and 5, but slightly lower
inPADDs 1 and 3.
       The refinery modeling also indicated that the reduction in fuel benzene levels
would result in small proportional decreases in aromatics levels as well.32  Thus
aromatics levels were adjusted using the additive factors calculated as follows:

              Additive Factor = 0.77 (BZ(control) - BZ(ref))            (1)

              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
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 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, NMTM was rerun with
data files that included new start emission standards for hydrocarbons. Output included
exhaust emissions, non-refueling evaporative emissions, and refueling evaporative
emissions.

       It should be noted that the inventory used for air quality modeling included an
error in contractor-supplied input files for 13 Northeastern states. This error had a small
impact on reference case inventories, but the impact on estimates of emission reductions
with controls was insignificant.  In addition, the control case inventory for 2015 assumes
that the fuel program is fully phased in, which is a simplification of the actual phase-in.
For more information about fuel program phase-in, refer to Chapter 6 of the Regulatory
Impact Analysis.
                                       2-29

-------
Final Regulatory Impact Analysis
    Table 2.2.-2. Average Fuel Benzene Level (Volume Percent) by PADD with
  Implementation of Fuel Benzene Standard (CG - Conventional Gasoline; RFG -
                           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
   Table 2.2-3.  Comparison of Average Fuel Benzene Level (Volume Percent) by
           PADD In Inventories Versus Final Rule Refinery Modeling..

Average Fuel
Benzene Level
Assumed in
Inventories
(0.62% standard)
Average Fuel
Benzene Level,
Final Rule
Refinery
Modeling, with
0.62% Standard
and 1.3 vol%
Maximum
Average

CG
RFG



CG
RFG







PADD1
0.55 %
0.54%



0.61%
0.54%







PADD 2
0.68%
0.71%



0.62%
0.60%







PADD 3
0.54%
0.55%



0.63%
0.55%







PADD 4
0.93%
0.62%



0.85%
0.62%







PADD 5
0.85%
0.60%



0.65%
0.60%







Calif.
0.61%
0.61%



0.61%
0.61%







      For highway diesel vehicles, we used a different approach than we used for
gasoline vehicles. NMIM2004 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:
                                    2-30

-------
Final Regulatory Impact Analysis
              f ± 20XX
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. Highway diesel
vehicle air toxic emission estimates remained unchanged from the proposal.

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 final rule
modeling, we used 1999 and future year inventories developed using NMIM2005, which
includes NONROAD2005. NONROAD2005 includes a number of improvements over
NONROAD 2004, which was used in the proposed rule. These improvements include
new evaporative categories for tank permeation, hose permeation, hot soak, and running
loss emissions, a revised methodology for calculating diurnal emissions, and
improvements to allocation of emissions from recreational marine and construction
equipment.

      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.

      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
and evaporative emissions of nonroad gasoline equipment were proportional to changes
in highway light-duty gasoline vehicle emissions.
       PF nonroad exhaust20XX = —	              (3)
                               h,LDGVExhaustNMIMReference20XX


                              EjLDGVevap ^nfr^fr  t 120XX
       PF nonroad evap20XX =	                  (4)
       The nonroad refueling associated with PFCs was subtracted from the nonroad
inventory prior to air quality modeling, and the inventory summaries presented in Section
2.2.1.2.1 include this subtraction.
                                      2-31

-------
Final Regulatory Impact Analysis
2.2.1.1.3 Commercial Marine Vessels, Locomotives and Aircraft

       These source sectors will not be impacted by the fuel benzene standards being
finalized in this rule. Final rule inventories are unchanged from those used to model the
proposal.

       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
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.33  More detailed information on methods used to develop air toxic
inventories for these sectors can be found in the documentation for the 1999 NEI.34 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.35
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.
                                       2-32

-------
Final Regulatory Impact Analysis
2.2.1.1.4. Portable Fuel Containers

       Any MSATs contained in the liquid gasoline will be present as a component of
the VOCs. 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, and
       2,2,4-trimethylpentane.

       While MSAT inventories for portable fuel containers (PFCs) were developed at
the State level (benzene) or national level (other MSATs above) to estimate emission
benefits and cost-effectiveness of the proposed rule, county-level inventories for portable
fuel containers were not developed for use in air quality modeling for the proposal.  In
this section, we describe the methodology used to develop such inventories for 1999,
2010, 2015, 2020, and 2030, for the final rule.

As discussed in Section 2.1.1., VOC inventories were developed at the State level for the
following years - 1990, 2005, 2010, 2015, 2020, and 2030. Thus an inventory had to be
developed for 1999. This was done by linear interpolation of the 1990 and 2005
inventories. Permeation and evaporative emissions had to be separated as well, since
they were combined in the State-level VOC inventories. Based on analyses done by the
California Air Resources Board, 33.87 percent of combined permeation plus evaporative
emissions was assumed to be permeation (see Section 2.1.1.2). This percentage was
applied to all the State inventory estimates.

       Statewide total annual VOC inventories were allocated to counties using county
level fuel consumption ratios for calendar year 2002, obtained from the public release
version of NONROAD2005:

PFC VOC Emissions Emsswn Type ^ scc YYY^ County z  =

PFC VOC Emissions                    x County Fml ConsumPtion^nroad^    (5)
.r.rU VV^. ^ml^lon^                    X
                                           State Fuel ConsumptionNonroad
                                                                      2005
       For all compounds except benzene and naphthalene, the fraction of total PFC emissions
that is composed of each of those HAPs was assumed to be directly proportional to the ratio of
each of those HAPs at the county level in total evaporative emissions from light-duty gasoline
vehicles (Equation 8).
                                      2-33

-------
Final Regulatory Impact Analysis
                             MSAT Emissions LDGveVap7riYY
       PFCMSATevap7nYY = - -^^   x PFCVOC Emissions (6)
                     riOXX    VOC EmiSSiOnS LDGVEvap20xx

These ratios were obtained from the database of toxic to VOC ratios in the NMIM model,
discussed in previous sections. NMIM has ratios that vary by fuel type (conventional or baseline
gasoline, ethanol oxygenated gasoline, and MTBE oxygenated gasoline).

       Another approach was used to estimate emissions of benzene with and without
PFC control, and also with and without the fuel benzene standard. First, we divided
county-level benzene refueling emissions by county-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.
These ratios were then adjusted based on a recent analysis of average nationwide
percentage of benzene in gasoline vapor from gasoline distribution with an RVP of 10 psi
at 60 degrees Fahrenheit.36  That analysis shows that the percentage  of benzene in
gasoline vapor from gasoline distribution is 0.27%, in contrast to 0.74% benzene on
average nationwide in vehicle refueling emissions from highway vehicles. The reason
for this difference is that the refueling algorithm in MOBILE6.2 is based on a
temperature of 90 degrees, whereas temperatures for gasoline marketing emissions will
typically be lower.  Thus a ratio of 0.36 was applied to the gasoline  vehicle  refueling
ratios.  For all emission types except permeation, the equation used was:

         PFC Benzene Emissions Emission Type ^ scc YYY^ Comfy z =
         PFC VOC Emissions Emisswn Type x^ scc 777> Comty z x            (7)

          Re fueling Benzene LDGVi Comty z
            V.Q fueling VOCLDG^Comty
                                         xO.36
                                  ity Z
       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. A recent study37 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:

       PFC  Benzene Emissions Emjssjon Type x, scc YYY , county  z =
       PFC  VOC Emissions Emjsslon Type x, scc YYY , county z x            (8)

       (Re fueling  Benzene WG^County z]
       1                      —:	—  x 0.36 x 1.77
          Re fueling VOC
                          LDOV , County Z
       A similar adjustment was applied to naphthalene emissions with and without fuel
benzene control, based on a recent analysis of average nationwide percentage of
naphthalene in gasoline vapor from gasoline distribution with an RVP of 10 psi at 60
                                       2-34

-------
Final Regulatory Impact Analysis


degrees Fahrenheit.38'39 The percentage is 0.00027, in contrast to 0.05% naphthalene on
average nationwide in vehicle refueling emissions from highway vehicles.  Thus a ratio
of 0.0054 was applied to the gasoline vehicle refueling ratios:

       PFC Naphthalen e Emissions Emisslon Type x^ scc YYY^ County z =
       PFC VOC Emissions Emssion Type x> scc YYY^ County z  x              (9)

       (Refueling Naphthalen e LDGV ^ County z^\ ^  ^^
       {     Re fueling VOCLDGVtCountyZ     }

2.2.1.1.5.  Gasoline Distribution

       EPA's estimates of gasoline distribution emissions reflect improvements in its
methodology developed for the 2002 National Emissions Inventory (NEI).  The key changes are:

1)  Vehicle refueling emissions are estimated as part of the highway vehicle inventory using
    NMTM2004, as discussed previously, and included in the highway vehicle inventory. Details
    of how the modeling was done can be found in the documentation for the mobile source 2002
    NEI.40 The previous methodology is described in the nonpoint 1999 NEI documentation.41
    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 is using 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.42

       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.-4 compares
benzene estimates in the 1999 NEI, final version 3, and the final 2002 NEI.
                                       2-35

-------
Final Regulatory Impact Analysis
 Table 2.2.-4.  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 fuel benzene control being
finalized in this rule, EPA developed updated air toxic inventories for vehicle refueling and
gasoline distribution to reflect the changes made in the 2002 NEI. In addition, the same
adjustment factors for benzene and naphthalene described above for PFC emissions were also
applied to gasoline distribution emissions.

Inventories were developed as follows:

1)   Vehicle refueling emissions were estimated using NMIM2004. Refueling emissions were
     estimated for reference case inventories in 1999, 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 a  2007 projection for the proposed rule.

3)   EPA developed new county-level reference case inventories for these pollutants by  applying
     these adjustment factors to county-level gasoline distribution emissions for 1999 and future
     years.  The gasoline distribution projections were based on projection information (growth
     factors, closures, reductions, etc.) from the 1999 NEI.43 Revised inventories were
     developed for years 1999, 2015, and 2020. 2030 was assumed to be the same as 2020.

4)   Additional nationwide adjustments of 0.36 and 0.0054, respectively, were applied to
     emissions of benzene and naphthalene. The basis for these adjustments is discussed in the
     Section 2.2.1.1.4.

5)   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 proposed rule control scenario/emissions proposed rule reference case

       These ratios reflect reductions estimated based on the assumption that reductions
       are proportional to reductions in vehicle refueling emissions.

2.2.1.1.6. Other Stationary Sources
                                       2-36

-------
Final Regulatory Impact Analysis
       Stationary source estimates for 1999, for all source categories except gasoline
distribution, were obtained from the National Emissions Inventory.44' 45
       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.46  EMS-
HAP has the capability of projecting emissions to 2020. After 2020, stationary source
emissions were assumed to remain constant.

       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  (10)

       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;47'48
       Regional and National fuel-use forecast data from the Energy Information
       Administration, U.S. Department of Energy, Annual Energy Outlook (AEO)49
                                       2-37

-------
Final Regulatory Impact Analysis
       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.50

       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.
       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.

2.2.1.1.7 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.-5,
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.51  Stationary
source precursor emissions were assumed to remain at 1999 levels in future year
modeling since the impact  of growth and control is unknown. However, mobile source
precursor emissions are expected to increase along with VOCs. To account for this in
modeling done to support the final rule, we estimated secondary concentrations from
mobile sources in future years by assuming they increased proportionally with primary
concentrations. For the proposed rule, we had projected precursor emissions for 1999 to
future years using ratios of VOCs for future years versus 1999, then used these projected
emissions to model secondary concentrations. A comparison of the two approaches,
                                      2-38

-------
Final Regulatory Impact Analysis
using modeling data from the proposal, yielded very similar results. A more detailed
discussion of the comparison can be found in EPA Technical Report Number EPA-
454/R-07-002

2.2.1.1.8 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, olefin 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.52

       Finally, as discussed in Section 2.1.3, there are greater uncertainties in projection
year estimates.
                                       2-39

-------
Final Regulatory Impact Analysis
                         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)
                                         2-40

-------
Final Regulatory Impact Analysis
       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.

       Portable Fuel Containers -- Since no direct measurements of air toxic emissions
from evaporation of gasoline in portable fuel containers were available, they were
estimated based on toxic to VOC ratios obtained from evaporative emissions
measurements taken from light-duty gasoline vehicles.  However, since evaporation of
fuel occurs at higher temperatures in vehicles than in PFCs, speciation profiles are
different. An effort to account for these differences was made for benzene and
naphthalene based on recent analyses done for the gasoline distribution sector.

       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.
                                       2-41

-------
Final Regulatory Impact Analysis
       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
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 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.-2).
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, emissions from all air toxics 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 2.2-2). It should be noted that these estimates do not account for
higher estimates of cold temperature hydrocarbon emissions in vehicles, PFC emissions,
or categories of nonroad gasoline evaporative emissions included in NONROAD2005
and discussed in Section 2.2.1.1.2.

       If higher estimates of cold temperature hydrocarbon emissions and vehicles and
evaporative emissions from nonroad gasoline equipment are accounted for, air toxic
emissions emitted from mobile sources will be reduced 46% between 1999 and 2020
without the controls in this proposal, from 2.38 million to 1.29 million tons (Figure 2.2-
3). This reduction will occur despite a projected 57% increase in vehicle miles traveled,
and a 47% projected increase in nonroad activity (See Figures 2.2.-4 and 2.2.-5).  It
should be noted, however, that EPA anticipates mobile source air toxic emissions will
begin to increase after 2020, from about 1.29 million tons in 2020 to 1.42 million tons in
2030. Benzene emissions from all sources decrease from about 366,000 tons in 1999 to
279,000 tons in 2020, and as is the case with total air toxic emissions, begin to increase
between 2020 and 2030 (Figure 2.2.-5).
                                       2-42

-------
Final Regulatory Impact Analysis
  Figure 2.2.-2. 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. Does not Account for Higher Estimates
   of Cold Temperature Hydrocarbon Emissions in Vehicles, PFC Emissions, or
        Categories of Nonroad Gasoline Evaporative Emissions Included in
                               NONROAD2005.




t~
o
tf>
c
o
1
w
0
w
E
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 	 U Major
I 	 | Area and Other
HI Fires - Prescribed and Wild
I I Non-Road Mobile
| 	 | On-Road Mobile
^
r^
f^r
*v^
7.X "!

"














2.69


0.91
0.34


0.75




2.55




5.03













1.26




0.29

0.76


1.44



^

^9V92 ' ^-\

,,-






4.09










0.95
1.45
0.29
0.65
0.77










,--'
-^







3.9
0.9


1 51



0.29
058

0.63



s
ci

«

^'
CO
*"




CO
CO



^,J









3.88










0.97



1.64



0.29
049
0.49

























4.01
1.06



1.78



0.29
0.44
0.44
'-""'



T-
00


^
s


-







9^1

-
-^




5





^














                1990
                       1999
                               2007
                                      2010
                                                 2015
                                                         2020
                                      2-43

-------
Final Regulatory Impact Analysis
    Figure 2.2.-S.  Contribution of Source Categories to Mobile Source Air Toxic
 Emissions, 1999 to 2030 (Not Including Diesel Particulate Matter). Includes Higher
 Estimates of Cold Temperature Hydrocarbon Emissions and Vehicles, Evaporative
 Emisions from Nonroad Gasoline Equipment, and PFC Emissions as Part of Area
                                  Source Inventory.
    4,000,000 i
    3,500,000
    3,000,000
    2,500,000
    2,000,000
    1,500,000
    1,000,000
     500,000
• Nonroad Equipment
D Highway Vehicles
D Fires — Prescribed and Wild
• Area and Other
• Major
                1999
                                                         2030
                                        2-44

-------
Final Regulatory Impact Analysis
Figure 2.2.-4.  Trend in Highway Vehicle Air Toxic Emissions Versus VMT, 1999 to
                                       2030.
   Figure 2.2.-S. Trend in Emissions of Nonroad Equipment Air Toxic Emissions
 (Excluding Commercial Marine Vessels, Locomotives and Aircraft) versus Activity,
                                   1999 to 2030.
          § 500000 - -
                         2010     2015
                                Year
3E+11 £  ^TonsMSATs
   .£•  —•—Nonroad Equipment Activity (hp-hrs)
                                      2020    2030
                                        2-45

-------
Final Regulatory Impact Analysis
              Figure 2.2.-6. Trend in Benzene Emissions, 1999 to 2030.
    400,000
    350,000
    300,000
    250,000
    200,000
    150,000
    100,000
     50,000
• Nonroad Equipment
D Highway Vehicles
D Fires — Prescribed and Wild
• Area and Other
D Major
               1999
                                                         2030
       Highway Vehicle Trends - Table 2.2.-6 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, forty percent of the chromium from highway vehicles and eighteen percent of
the chromium from nonroad sources was assumed to be the highly toxic hexavalent form.
The estimate for highway vehicles is based on data from utility boilers,53 and the estimate
for nonroad equipment is based on combustion data from stationary combustion turbines
that burn diesel fuel.54
                                        2-46

-------
Final Regulatory Impact Analysis
 Table 2.2.-6.  Nationwide Emissions (Tons) of Individual Air Toxic Pollutants from
                               Highway Vehicles.
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
1999
23,876
182,120
29,821
3,845
183,661
8
5
73,439
80,458
66,267
57,801
5
4,056
10
497
4,288
14,284
489,873
277,285
2010
11,473
101,880
17,169
1,824
110,526
10
7
40,732
38,885
39,801
29,886
6
2,261
13
255
2,327
7,652
268,871
152,046
2015
10,763
94,469
16,149
1,650
105,956
11
8
37,528
35,857
33,481
23,089
6
2,022
14
234
2,154
7,368
250,646
141,710
2020
11,355
96,315
16,893
1,704
110,317
12
8
38,080
37,153
30,727
18,372
7
1,986
16
239
2,222
7,814
257,367
145,473
2030
13,378
111,783
19,879
1,981
129,290
15
10
44,055
43,404
33,468
17,957
9
2,259
19
278
2,574
9,253
299,677
169,369
       Table 2.2.-7 summarizes total tons of air toxic emissions from highway vehicles
by vehicle class in 1999, 2007, 2010, 2015, 2020, and 2030. Table 2.2.-S provides the
percentage of total highway vehicle emissions associated with each vehicle class. In
1999, 55% 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 34% of highway vehicle HAP emissions will be from LDGVs and 60% 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.
                                      2-47

-------
Final Regulatory Impact Analysis
  Table 2.2.-7. 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
36,958
66,672
1,215
688
353,671
188,134
836,995
7,267
1,491,600
2010
22,622
21,323
589
41
279,674
144,254
349,220
7,899
825,624
2015
19,605
14,812
528
23
287,644
141,165
290,746
8,595
763,117
2020
19,469
11,638
470
16
319,974
144,247
270,956
9,291
776,062
2030
22,172
10,188
389
16
375,603
159,682
319,395
11,213
898,659
HDDV: Heavy Duty Diesel Vehicles
HDGV: Heavy Duty Gasoline Vehicles
LDDT: Light Duty Diesel Trucks
LDDV: Light Duty Diesel Vehicles
LDGT1 : Light Duty Gasoline Trucks 1
LDGT2: Light Duty Gasoline Trucks 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 2030 (Not Including Diesel Particulate Matter).
Vehicle
LDGV
LDGT1 and 2
HDGV
HDDV
Other (motorcycles and
light-duty diesel
vehicles and trucks)
1999
56%
36%
5%
2%
1%
2010
42%
51%
3%
3%
1%
2015
38%
56%
2%
2%
2%
2020
35%
60%
1%
2%
2%
2030
35%
60%
1%
2%
2%
       Tables 2.2.-9 through 2.2.-14 summarize total tons of emissions nationwide for
benzene, 1,3-butadiene, formaldehyde, acetaldehyde, naphthalene, and acrolein from
highway vehicles. About 90% 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.
                                      2-48

-------
Final Regulatory Impact Analysis
  Table 2.2.-9. Tons of Benzene Emissions from Highway Vehicle Classes, 1999 to
                                     2030.
Vehicle Type
HDDV
HDGV
LDDT
LDDV
LDGT1
LDGT2
LDGV
MC
Total Highway
Emissions (tons/yr)
1999
2,564
6,665
200
112
46,358
21,392
105,724
646
183,661
2010
1,574
2,383
97
7
39,456
19,742
46,598
669
110,526
2015
1,366
1,715
87
4
41,796
20,074
40,186
728
105,956
2020
1,358
1,399
78
3
47,352
21,083
38,257
787
110,317
2030
1,547
1,213
64
3
56,290
23,737
45,489
947
129,290
Table 2.2.-10. Tons of 1,3-Butadiene Emissions from Highway Vehicle Classes, 1999
                                    to 2030.
Vehicle Type
HDDV
HDGV
LDDT
LDDV
LDGT1
LDGT2
LDGV
MC
Total Highway
Emissions (tons/yr)
1999
1,489
1,177
90
50
5,307
3,526
12,034
202
23,876
2010
915
197
44
3
3,820
1,991
4,280
224
11,473
2015
794
99
39
2
3,929
1,913
3,743
243
10,763
2020
789
78
35
1
4,520
2,064
3,605
263
11,355
2030
899
63
29
1
5,411
2,344
4,312
318
13,378
                                     2-49

-------
Final Regulatory Impact Analysis
Table 2.2.-11.  Tons of Formaldehyde Emissions from Highway Vehicle Classes,
1999 to 2030.
Vehicle Type
HDDV
HDGV
LDDT
LDDV
LDGT1
LDGT2
LDGV
MC
Total Highway
Emissions (tons/yr)
1999
19,094
6,142
386
217
15,666
9,916
28,522
516
80,458
2010
1 1 ,724
1,213
188
13
9,702
4,851
10,627
567
38,885
2015
10,176
688
168
7
10,030
4,656
9,515
617
35,857
2020
10,114
556
150
5
11,487
4,961
9,213
667
37,153
2030
1 1 ,522
460
124
5
13,790
5,652
1 1 ,044
806
43,404
Table 2.2.-12. Tons of Acetaldehyde Emissions from Highway Vehicle Classes, 1999
                                   to 2030.
Vehicle Type
HDDV
HDGV
LDDT
LDDV
LDGT1
LDGT2
LDGV
MC
Total Highway
Emissions (tons/yr)
1999
7,032
1,411
123
69
6,050
3,429
11,555
152
29,821
2010
4,318
390
60
4
4,808
2,367
5,043
180
17,169
2015
3,748
248
54
2
5,068
2,329
4,504
196
16,149
2020
3,725
204
48
2
5,836
2,502
4,364
213
16,893
2030
4,243
173
40
2
7,039
2,880
5,246
258
19,879
                                     2-50

-------
Final Regulatory Impact Analysis
 Table 2.2.-13. Tons of Acrolein Emissions from Highway Vehicle Classes, 1999 to
                                     2030.
Vehicle Type
HDDV
HDGV
LDDT
LDDV
LDGT1
LDGT2
LDGV
MC
Total
Highway
Emissions (tons/yr)
1999
855
689
35
20
623
326
1,286
13
3,845
2010
525
76
17
1
457
231
503
14
1,824
2015
455
24
15
1
472
226
442
15
1,650
2020
453
17
14
0
538
240
425
16
1,704
2030
516
12
11
0
644
271
508
20
1,981
 Table 2.2.-14.  Tons of Naphthalene Emissions from Highway Vehicle Classes, 1999
                                    to 2030.
Vehicle Type
HDDV
HDGV
LDDT
LDDV
LDGT1
LDGT2
LDGV
MC
Total
Highway
Emissions (tons/yr)
1999
167
773
7
7
760
489
1,830
23
4,056
2010
65
400
2
0
640
267
861
25
2,261
2015
32
248
1
0
697
273
743
27
2,022
2020
19
195
1
0
769
281
693
29
1,986
2030
16
176
1
0
900
315
817
35
2,259
       NonroadEquipment Trends — Table 2.2.-15 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.-16
summarizes total tons of air toxic emissions from categories of nonroad equipment by
equipment type in 1999, 2010, 2015, 2020, and 2030. Table 2.2.-17 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
about 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.
                                      2-51

-------
Final Regulatory Impact Analysis
    Table 2.2.-15. 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 III
Chromium VI
Ethyl Benzene
Formaldehyde
Hexane
Manganese
MTBE
Naphthalene
Nickel
POM
Propionaldehyde
Styrene
Toluene
Xylenes
Annual Total Nonroad Emissions (Tons)
1999
10,333
109,793
21,952
2,754
74,902
15
3
46,072
52,083
36,925
2
78,585
1,212
31
347
4,968
3,055
234,558
208,728
2010
7,136
83,546
16,208
2,264
54,763
15
4
33,435
38,213
29,758
2
28,464
1,182
34
305
3,462
2,297
189,605
147,242
2015
6,586
71,362
14,459
2,179
49,985
16
4
29,489
34,406
27,430
2
27,238
1,228
36
287
3,036
2,003
164,871
126,825
2020
6,518
62,991
13,663
2,168
48,453
16
4
27,057
32,678
26,083
2
27,245
1,291
37
275
2,824
1,807
146,220
114,252
2030
7,004
62,250
14,153
2,340
51,647
16
4
28,033
33,994
27,439
2
29,865
1,440
41
287
2,865
1,835
145,330
116,764
                                     2-52

-------
Final Regulatory Impact Analysis
 Table 2.2.-16. 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
Annual Total Nonroad Emissions (Tons)
1999
21,397
14,276
325
59,302
8,736
42,496
1 1 ,422
261,635
3,578
332,631
4,412
125,933
177
886,318
2010
12,512
14,965
198
33,977
9,742
22,280
4,247
129,932
2,094
202,760
3,972
201,118
138
637,934
2015
9,686
16,081
157
35,994
10,213
18,688
2,793
130,157
2,228
163,953
3,886
167,488
114
561 ,439
2020
7,875
17,256
141
39,207
10,973
16,439
2,239
139,762
2,452
148,746
3,752
124,640
101
513,583
2030
6,567
19,603
152
46,503
13,354
15,207
2,093
160,669
2,960
147,720
3,533
106,845
102
525,309
 Table 2.2.-17. Contribution of Equipment Types to Nonroad Air Toxic Emissions,
              1999 to 2030 (not Including Diesel Particulate Matter).
Equipment
Type
Lawn and
Garden
Pleasure Craft
Recreational
All Others
1999
30%
38%
14%
18%
2010
20%
32%
32%
16%
2015
23%
29%
30%
18%
2020
27%
29%
24%
19%
2030
31%
28%
20%
21%
       Over 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,
over 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.-18 through 2.2.-23 summarize total tons of emissions nationwide for
benzene, 1,3-butadiene, formaldehyde, acetaldehyde, naphthalene, and acrolein from
nonroad equipment types.
                                      2-53

-------
Final Regulatory Impact Analysis
Table 2.2.-18. 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
Annual Total Nonroad Emissions (Tons)
1999
2,105
1,102
33
7,931
644
3,945
1,498
25,753
202
24,963
162
6,548
15
74,902
2010
1,283
1,163
19
5,140
719
2,111
524
15,996
131
16,698
143
10,825
12
54,763
2015
1,020
1,247
15
5,478
753
1,786
335
15,540
130
14,101
139
9,430
10
49,985
2020
855
1,335
14
6,010
809
1,595
263
16,644
140
13,145
134
7,502
9
48,453
2030
736
1,511
15
7,178
982
1,494
233
19,133
168
13,264
126
6,798
9
51,647
  Table 2.2.-19. 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
Annual Total Nonroad Emissions (Tons)
1999
236
824
4
1,324
6
455
242
4,034
35
2,069
114
990
1
10,333
2010
145
859
2
774
6
231
76
2,240
21
1,291
104
1,385
1
7,136
2015
116
924
2
820
e7
198
47
2,085
21
1,034
102
1,230
11
6,586
2020
98
993
2
901
6
180
37
2,225
23
928
99
1,025
1
6,518
2030
85
1,131

1,080

171
31
2,558
28
909
94
907

7,004
                                    2-54

-------
Final Regulatory Impact Analysis
 Table 2.2.-20. 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
Annual Total Nonroad Emissions (Tons)
1999
8,890
6,549
123
3,516
4,715
12,103
2,487
7,050
334
2,345
1,895
1,990
87
52,083
2010
5,051
6,809
83
2,331
5,252
7,352
1,212
3,902
153
1,548
1,721
2,731
68
38,213
2015
3,759
7,333
66
2,122
5,499
5,662
837
3,633
117
1,274
1,683
2,365
56
34,406
2020
2,915
7,885
58
2,019
5,899
4,541
697
3,816
109
1,160
1,624
1,904
50
32,678
2030
2,296
8,990
63
2,080
7,152
3,858
718
4,328
116
1,147
1,527
1,669
50
33,994
  Table 2.2.-21. 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
Annual Total Nonroad Emissions (Tons)
1999
3,986
2,019
55
1,390
2,364
5,433
1,087
2,381
133
1,615
850
599
39
21,952
2010
2,265
2,098
37
999
2,639
3,308
539
1,522
59
1,098
772
843
30
16,208
2015
1,685
2,259
30
902
2,768
2,550
372
1,410
41
920
755
743
25
14,459
2020
1,306
2,430
26
850
2,974
2,046
310
1,476
37
844
728
613
22
13,663
2030
1,028
2,770
28
866
3,619
1,739
320
1,670
37
834
685
533
23
14,153
                                    2-55

-------
Final Regulatory Impact Analysis
Table 2.2.-22. 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
Annual Total Nonroad Emissions i
1999
232
968
3
143
98
326
72
398
11
218
130
151
2
2,754
2010
132
1,005
2
89
112
195
33
206
5
134
119
231
2
2,264
2015
99
1,083
2
85
118
151
23
195
5
106
117
194
1
2,179
2020
77
1,165
2
86
129
123
19
207
5
95
113
148
1
2,168
Tons)
2030
61
1,329
2
94
161
105
19
236
5
93
107
128
1
2,340
  Table 2.2.-2S.  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
Annual Total Nonroad Emissions i
1999
42
456
1
104
65
56
26
305
2
34
61
59
0
1,212
2010
26
496
1
103
68
37
13
245
2
36
44
112
0
1,182
2015
21
530
0
113
72
30
9
246
2
37
42
127
0
1,228
2020
17
566
0
125
79
22
6
264
1
39
40
132
0
1,291
Tons)
2030
12
638
0
149
102
17
4
303
2
42
35
136
0
1,440
      Portable Fuel Containers - Table 2.2.-24 summarizes nationwide emissions of
individual air toxics from gasoline in portable fuel containers (PFCs), from 1999 to 2030.
                                     2-56

-------
Final Regulatory Impact Analysis


   Table 2.2.-24. Tons of Air Toxic Emissions from Portable Fuel Containers, 1999 to
                                     2030.
Pollutant
2,2,4-Trimethylpentane
Benzene
Ethyl Benzene
Hexane
MTBE
Naphthalene
Toluene
Xylenes
Total
1999
4,870
853
2,135
5,417
6,969
1
10,733
6,189
37,166
2010
4,461
833
1,900
5,176
4,763
1
9,668
5,432
32,232
2015
4,741
889
2,027
5,532
4,987
1
10,329
5,800
34,306
2020
5,088
953
2,175
5,935
5,007
1
11,082
6,223
36,464
2030
5,805
1,086
2,480
6,766
5,503
1
12,636
7,096
41,374
About 75% of all HAP emissions and benzene emissions from PFCs are associated with
residential use, and the rest are from commercial use. As can be seen in Figure 2.2.-7,
most commercial PFC air toxic emissions are associated with equipment refueling, and
most residential emissions are associated with evaporation and permeation.
   Figure 2.1.-7. Distribution of air toxic emissions (tons) among emission types for
                     commercial versus residential PFCs, 1999.
           Air Toxic Emissions from Commercial PFCs, 1999
                                    (Tons)
                         Evaporation,        Permeation, 240,
                Spillage During61'5%
               Transport, 2124,^
                    21%
               Refueling
              Equipment:
                Vapor
             Displacement,
               957,10%
Refilling at Pump:
Spillage, 84,1%

  Refilling at Pump:
^_     Vapor
   Displacement,
     957,10%
                                                      Refueling
                                                     Equipment:
                                                   Spillage, 5172,
                                                        51%
                                      2-57

-------
Final Regulatory Impact Analysis
               Air Toxic Emissions from Residential PFCs, 1999,
              Refueling
             Equipment:
           Spillage, 2644,
               10%
          Refilling at
         Pump: Vapor
         Displacement, ^x\
           497, 2%

            Refilling at
          Pump: Spillage,^
              44, 0%
   Refueli ng
  Equipment:
    Vapor
 Displacement,
\  497,2%
                                      Tons
  Spillage During
^Transport, 1557,
       6%
                     Permeation
                     7505, 28%
                              Evaporation,
                              14429,52%
       Diesel Paniculate Matter - The inventory estimates presented above for mobile
source air toxics do not include diesel particulate matter.  Table 2.2.-25 summarizes the
trend in diesel particulate matter between 1999 and 2030, by source category.  These
inventory estimates were obtained from EPA's recently proposed national ambient air
quality standard for particulate matter.55 Diesel particulate matter emissions will be
reduced by 75% between 2001 and 2030. As controls on highway diesel engines and
nonroad diesel engines phase in, diesel-powered locomotives and commercial  marine
vessels increase from 13% of the inventory in 2001 to 55% in 2030.

     Table 2.2.-2S. Percent Contribution of Mobile Source Categories to Diesel
  Particulate Matter (PMio) Emissions, 2001 to 2030 in Tons Per Year (Percent of
                                      Total).
Source
Highway Vehicles
Commercial
Marine Vessels
Locomotives
Other Nonroad
Equipment
2001
125,162
(36.7%)
20,541
(6%)
25,173
(7.4%)
170,357
(49.9%)
2015
37,463
(24.8%)
17,085
(11.3%)
17,521
(11.6%)
78,930
(52.3%)
2020
26,471
(24.4%)
16,984
(15.7%)
16,535
(15.3%)
48,284
(44.6%)
2030
18,135
(21.6%)
21,388
(25.5%)
25,086
(29.9%)
19,285
(23.0%)
2.2.1.2.2 Impact on Inventory of Controls

       The controls being finalized in this rule would reduce air toxic emissions from
highway gasoline vehicles, nonroad gasoline equipment, gasoline distribution and
                                       2-58

-------
Final Regulatory Impact Analysis
portable fuel containers. The total air toxic emissions reduced in the 2030 inventories
used for air quality modeling for these sectors are 335,000 tons, and the total benzene
emissions reduced are 65,000 tons.  It should be emphasized that the air quality, exposure
and risk modeling inventory does not account for recent increases in the use of ethanol
oxygenated gasoline or implementation of the renewable fuels standard.  For inventories
which include these emissions, see Section 2.2.2.2.

       Table 2.2.-26  summarizes the nationwide impact of the controls on emissions of
key air toxics from highway vehicles in 2015, 2020, and 2030. The reductions in
highway vehicle air toxic emissions by 2030 are dramatic, about 35%. Benzene
reductions are over 40%. Nonroad equipment emissions are impacted by fuel benzene
control, which result in reductions of about 14% for that pollutant (Table 2.2.-27).
Emissions from PFCs will be impacted by both controls on the containers themselves as
well as the fuel benzene standard (Table 2.2-28), with reductions in total air toxic
emissions of over 60% in 2030, and reductions in benzene of about 80%.  In addition,
fuel benzene controls would reduce emissions within the gasoline distribution sector.
Table 2.2.-29 presents estimated reductions for this source in 2015 and 2020, which total
over 30%, due to the fuel benzene standard.  Figures 2.2.-8 and 2.2.-9  depict the trend in
total MSAT and benzene emissions for all sources with the controls being finalized in
this rule. More detailed summaries of emissions by individual pollutant, by State, and for
urban versus rural areas can be found in Excel workbooks included in  the docket for this
rule.
                                      2-59

-------
Final Regulatory Impact Analysis
 Table 2.2.-26.  Nationwide Impact of Controls on Emissions of Gaseous Air Toxics from Highway Vehicles in 2015, 2020, and
                                                        2030.

Pollutant
1,3 -Butadiene
2,2,4-
Trimethylpentane
Acetaldehyde
Acrolein
Benzene
Ethyl Benzene
Formaldehyde
Hexane
MTBE
Propionaldehyde
Styrene
Toluene
Xylenes
Total
Annual Emissions (tons) by Vehicle Type
2015
Reference
Case
10,763
94,469
16,149
1,650
105,956
37,528
35,857
33,481
23,089
2,154
7,368
250,646
141,710
760,821
2015
Control
Case
9,160
80,630
13,970
1,458
79,034
32,189
31,475
30,802
22,363
1,925
6,134
212,901
120,444
642,486
2015
Reduction
1,602
13,840
2,180
192
26,922
5,339
4,382
2,679
725
230
1,234
37,745
21 ,266
118,336
2020
Reference
Case
11,355
96,315
16,893
1,704
110,317
38,080
37,153
30,727
18,372
2,222
7,814
257,367
145,473
773,793
2020
Control
Case
8,655
73,103
13,222
1,382
73,141
29,117
29,877
26,241
17,226
1,837
5,743
194,002
109,772
583,319
2020
Reduction
2,700
23,212
3,671
322
37,176
8,962
7,276
4,486
1,146
385
2,071
63,365
35,701
190,474
2030
Reference
Case
13,378
110,895
19,879
1,981
129,290
43,676
43,404
32,435
17,109
2,574
9,253
297,748
168,285
889,908
2030
Control
Case
8,707
72,262
13,677
1,434
72,673
28,770
31,196
25,832
16,080
1,919
5,720
191,607
108,480
578,358
2030
Reduction
4,670
38,634
6,202
548
56,617
14,906
12,207
6,602
1,029
655
3,533
106,141
59,805
311,549
                                                         2-60

-------
Final Regulatory Impact Analysis
Table 2.2.-21. Nationwide Impact of Controls 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)
2015
Reference
Case
6586
14459
2179
49985
34406
107615
2015
Control
Case
6599
14468
2179
43220
34433
101418
2015
Reduction
-13
-9
0
6265
-27
6197
2020
Reference
Case
6518
13663
2168
48453
32678
103480
2020
Control
Case
6530
13671
2168
41736
32703
97339
2020
Reduction
-12
-8
0
6717
-25
6141
2030
Reference
Case
7004
14153
2340
51647
33994
109138
2030 Control
Case
7017
14162
2340
44427
34020
102528
2030
Reduction
-13
-9
0
7220
-26
6610
                                                        2-61

-------
Final Regulatory Impact Analysis
 Table 2.2.-2S. Nationwide Impact of Controls on Emissions of Air Toxics from Portable Fuel Containers in 2010, 2015, 2020,
                                                     and 2030.

Pollutant
2,2,4-
Trimethylpentane
Benzene
Ethyl Benzene
Hexane
MTBE
Naphthalene
Toluene
Xylenes
Total

2010
Reference
Case
4,461
833
1,900
5,176
4,763
1
9,668
5,432
32,232

2010
Control
Case
4,003
752
1,700
4,622
4,295
1
8,646
4,862
28,880
Annual Emissions (tons)
2015
Reference
Case
4,741
889
2,027
5,532
4,987
1
10,329
5,800
34,306
2015
Control
Case
1,864
179
756
1,932
2,360
0
3,752
2,157
13,000
2020
Reference
Case
5,088
953
2,175
5,935
5,007
1
11,082
6,223
36,464
2020
Control
Case
2,012
193
816
2,085
2,382
0
4,050
2,328
13,867
2030
Reference
Case
5,805
1,086
2,480
6,766
5,503
1
12,636
7,096
41,374
2030
Control
Case
2,315
222
939
2,399
2,638
0
4,658
2,678
15,849
    Table 2.2.-29. Nationwide Impact of Controls Emissions of Benzene from Gasoline Distribution in 2015 and 2020 (2030
                                          Assumed to be the Same as 2020).
Annual Emissions (tons)
2015
Reference
Case
2,160
2015
Control
Case
1,460
2015
Reduction
700
2020
Reference
Case
2,234
2020
Control
Case
1,516
2020
Reduction
719
                                                        2-62

-------
Final Regulatory Impact Analysis
    Figure 2.2.-S. Contribution of Source Categories to Mobile Source Air Toxic
  Emissions, 1999 to 2030, with Final Rule Standards in Place (Not Including Diesel
                                 Particulate Matter).
                                                                   • Nonroad Equipment
                                                                   DHighway Vehicles
                                                                   D Fires -- Prescribed and Wild
                                                                   • Area and Other
                                                                   D Major
     Figure 2.2.-9.  Contribution of Source Categories to Mobile Source Benzene
            Emissions, 1999 to 2030, with Final Rule Standards in Place.
                                                                • Nonroad Equipment
                                                                D Highway Vehicles
                                                                D Fires - Prescribed and Wild
                                                                •Area and Other
                                                                D Major
                                          2-63

-------
Final Regulatory Impact Analysis
2.2.2  Emission Reductions from Controls

       Section 2.2.2 describes revisions made to emission inventories after we developed
MSAT inventories for air quality modeling ("air quality inventories").  The primary
revision is accounting for the impacts of implementing the federal Renewable Fuel
Standard (RFS).  The revised inventories were used to estimate emission benefits of the
rule and the cost-effectiveness of the control strategies. We refer to the revised
inventories as "cost-effectiveness inventories" in this section to distinguish them from the
air quality inventories discussed in Section 2.2.1.

2.2.2.1 Methodology Changes from Air Quality Inventories

2.2.2.1.1  Highway Vehicles

       The fundamental difference between the air quality and cost-effectiveness
inventories is the use of fuel parameters that reflect implementation of the Renewable
Fuel Standard (RFS fuel), as described in Section 2.1.1.  We also corrected a minor error
which addresses how MOBILE6.2 calculates benzene evaporative emissions with ethanol
oxygenated fuel. In addition,  for the control case, aromatics levels were adjusted using a
different  algorithm to calculate additive adjustment factors:

              Additive Factor = 1.0 (BZ(control) - BZ(ref))             (11)

              Where BZ = benzene
We assume that with increased ethanol use, when fuel benzene is reduced there will be no
increase in other aromatic levels to help compensate for octane loss.  An Excel workbook
with all the fuel parameters used, "MSAT Fuels Cost Effectiveness.xls," is included in
the docket for this rule. Also, we estimated vehicle refueling emissions using NMIM
2005, instead of projecting them from the 1999 NEI, as discussed in Section 2.2.1.
Finally, it should be noted that inventories do not account increased permeation due to
ethanol use, nor do they account for the 1.3 vol% maximum average fuel benzene level.

2.2.2.1.2 Nonroad Equipment

       Unlike the air quality inventories, the cost-effectiveness inventories for nonroad
equipment used the RFS fuel as described in Section 2.1.  As with the air quality
inventories, we assumed that changes in county-level exhaust and evaporative emissions
of nonroad gasoline equipment were proportional to changes in highway light-duty
gasoline vehicle emissions. It should be noted that our inventories did not account for
increased hose and tank permeation associated with increased ethanol use. As a result,
our estimates of emission reductions from fuel benzene control may be slightly
underestimated.
                                       2-64

-------
Final Regulatory Impact Analysis
2.2.2.1.3 Portable Fuel Containers

       The RFS fuel was used to develop cost-effectiveness inventories for PFCs, as
described in Section 2.1. Air toxic inventories for PFCs for the reference and control
cases were developed by speciating VOC, following the same approach used for the air
quality modeling inventories (See Section 2.2.1.1.4). However, since the air quality
modeling inventories did not account for RFS fuel, we used revised highway gasoline
vehicle inventories for benzene and VOC from refueling that did account for RFS fuel to
develop benzene to VOC ratios, and total evaporative emission ratios for other air toxics.

2.2.2.1.4 Gasoline Distribution

       Gasoline distribution inventories were also revised to account for the RFS fuel.
The reference case (RC) inventory was estimated for each source category (SCC code) at
the county  level as follows:
GasDistr. Benzene Emissions RCSCC ^ CounlyZ: RFSMMal =

^   r,-   D     D/-Z7 •  •                        ( Nonrefueling evap Benzene RC1DGrcz
GasDistr.BenzeneRCEmisswnsSCCYYY                  — - — — - - - —
                       SCCYYYCa  z FinalRtlleAQIn    . — - — — - - - —
                                            {Nonrefuelmg evap Benzene RC
                                                                   LDGVCountyZAQImentca
Where,
Final Rule AQ Inventory = the inventory for SCC code YYY in county Z from the air
quality inventory, as discussed in Section 2.2.1
RFS Max 9.6  = the inventory for SCC code YYY in county Z assuming 9.6 billion
gallons of national ethanol consumption nationwide, attributing as much as possible for
use as  an oxygenate in reformulated gasoline.

       The air quality inventory was adjusted using ratios of non-refueling evaporative
emissions, because the methodology for estimating  refueling emissions differed for the
air quality inventory versus the final rule inventory, as discussed above.

       The control case (CC) inventory was estimated using the following equation:

Gas Distr. Benzene Emissions CC

Gas Distr. Benzene RC Emissions
JSCC YYY, County Z, RFS 9.6Max ~

                    "Re fueling BenzeneCCWQVtCountyZfRFS96Max
 SCC YYY, County Z,RFS9.6Max ~
                                                          '(13)
                                                Re fueling Benzene RCLDGr^ County z ^ 96Max
2.2.2.2  Estimated Reductions for Air Toxic Pollutants of Greatest Concern

       The following sections present control case inventories and reductions for each
individual control, and then cumulative reductions for all controls combined.
                                        2-65

-------
Final Regulatory Impact Analysis
2.2.2.2.1 Fuel Benzene Standard

       Highway Gasoline Vehicles - The fuel benzene standard will reduce emissions
from light-duty gasoline vehicles and trucks, motorcycles, and heavy-duty gasoline
trucks. Tables 2.2.-30, 2.2.-31, and 2.2.-32 present nationwide benzene emissions for
these vehicle classes with and without the 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.-30. Impact of Fuel Benzene Control on Benzene Emissions from
                        Highway Vehicles, by Class, 2015.
Vehicle Class
LDGV
LDGT1
LDGT2
MC
HDGV
TOTAL
Reference Case
Tons
37,881
39,657
17,696
773
1,782
97,789
Control Case Tons
33,766
35,279
15,658
663
1,509
86875
Reduction
4,115
4,378
2,037
110
273
10914
     Table 2.2.-31.  Impact of Fuel Benzene Control on Benzene Emissions from
                        Highway Vehicles, by Class, 2020.
Vehicle Class
LDGV
LDGT1
LDGT2
MC
HDGV
TOTAL
Reference Case
Tons
35,987
44,611
18,627
833
1,456
101514
Control Case Tons
32,213
39,849
16,572
714
1,240
90,588
Reduction
3,774
4,762
2,056
118
215
10926
                                      2-66

-------
Final Regulatory Impact Analysis
     Table 2.2.-S2.  Impact of Fuel Benzene Control on Benzene Emissions from
                        Highway Vehicles, by Class, 2030.
Vehicle Class
LDGV
LDGT1
LDGT2
MC
HDGV
TOTAL
Reference Case
Tons
42,752
52,993
20,996
1,002
1,273
119016
Control Case Tons
38,345
47,477
18,738
861
1,081
106502
Reduction
4,407
5,516
2,259
141
192
12514
       Reductions from the 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.-33 summarizes impacts of fuel
benzene control on the benzene emission inventory for gasoline vehicles in each State in
2030.
                                       2-67

-------
Final Regulatory Impact Analysis
  Table 2.2.-3S. 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
2,260.4
1,304.4
1,788.9
1,349.2
9,422.4
2,728.3
1,033.1
269.6
112.0
4,175.1
4,176.9
189.7
1,149.1
4,075.3
3,392.9
1,580.3
1,385.9
1,988.4
1,540.8
668.5
1,716.1
1,690.7
5,642.0
4,086.7
967.8
2,839.5
795.0
985.5
1,021.5
641.0
1,858.8
1,739.4
4,519.9
3,922.7
515.0
4,619.9
1,808.7
3,724.9
4,102.3
324.4
2,038.9
523.5
2,545.4
6,294.5
1,731.2
463.7
3,312.0
5,856.9
862.5
2,693.8
581.2
2030 Control
Case Tons
2,013.9
895.8
1,631.6
1,197.4
9,387.8
2,359.2
1,019.1
265.7
110.5
3,687.3
3,781.8
188.9
969.7
3,740.6
2,956.3
1,354.9
1,154.5
1,749.9
1,356.0
639.8
1,667.0
1,667.2
4,827.3
3,349.5
856.5
2,471.7
675.0
831.7
969.3
614.6
1,833.1
1,448.3
4,278.1
3,521.1
430.1
4,005.6
1,609.1
3,108.3
3,821.3
320.0
1,833.7
447.6
2,239.0
5,651.4
1,488.9
428.1
3,109.4
4,888.8
759.9
2,397.2
491.5
2030 Tons
Reduced
246.4
408.7
157.4
151.8
34.6
369.2
14.0
3.9
1.5
487.8
395.1
0.7
179.5
334.8
436.6
225.4
231.4
238.5
184.8
28.7
49.2
23.4
814.8
737.2
111.3
367.8
120.0
153.8
52.2
26.4
25.6
291.0
241.8
401.6
84.9
614.3
199.6
616.6
281.0
4.4
205.2
75.8
306.3
643.1
242.3
35.6
202.6
968.1
102.6
296.6
89.7
% Change
10.9
31.3
8.8
11.3
0.4
13.5
1.4
1.4
1.3
11.7
9.5
0.4
15.6
8.2
12.9
14.3
16.7
12.0
12.0
4.3
2.9
1.4
14.4
18.0
11.5
13.0
15.1
15.6
5.1
4.1
1.4
16.7
5.3
10.2
16.5
13.3
11.0
16.6
6.8
1.4
10.1
14.5
12.0
10.2
14.0
7.7
6.1
16.5
11.9
11.0
15.4
                                      2-68

-------
Final Regulatory Impact Analysis
Gasoline Nonroad Equipment - Table 2.2.-34 summarizes the nationwide impact of the
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.-35, these benefits vary from 1 to 31% by State in 2030.

  Table 2.2-34. Nationwide Impact of Fuel Benzene Control on Benzene Emissions
                       from Nonroad Gasoline Equipment.

2015 Reference Case
2015 Control Case
20 15 Reduction
2020 Reference Case
2020 Control Case
2020 Reduction
2030 Reference Case
2030 Control Case
2030 Reduction
Tons
41,343
35,825
5,518
40,161
34,717
5,444
42,994
37168
5,826
       Portable Fuel Containers -Table 2.2.-36 summarizes MSAT emissions from
PFCs with no fuel benzene or federal PFC control (but including State control programs).
The fuel benzene control will reduce benzene emissions from PFCs, as summarized in
Table 2.2.-37.  Again, emission benefits vary across the U. S., as seen in Table 2.2.-38.

       Gasoline Distribution -Table 2.2.-39 presents the benzene inventory from
gasoline distribution (not including refueling) in 2015  and 2020 with and without the fuel
benzene control.  Table 2.2.-40 presents the  inventory  for 2020 at the State level with and
without fuel benzene control. More detailed inventory estimates by county are available
in the docket for the rule.
                                      2-69

-------
Final Regulatory Impact Analysis
 Table 2.2.-3S.  Gasoline Nonroad Equipment Benzene Emission Reductions (Tons)
                       from 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 Tons
1,024.8
188.3
715.3
637.6
4,055.2
623.4
412.9
98.2
28.1
3,752.5
1,576.1
127.6
285.1
1,298.3
722.9
489.1
309.2
483.1
1,133.7
257.3
789.2
678.8
1,585.6
900.5
534.4
778.8
133.6
209.1
310.3
270.3
1,053.7
266.5
2,366.1
1,654.7
107.2
1,329.4
596.0
639.2
1,516.0
110.5
907.6
109.5
739.7
3,156.9
356.5
134.2
1,105.1
1,039.3
326.6
962.0
107.4
2030 Control
Case Tons
830.6
129.4
615.1
514.1
4,032.6
525.5
403.5
95.8
27.4
3,070.6
1,324.2
126.7
232.3
1,192.6
605.6
404.2
245.4
403.6
896.6
238.2
737.3
663.1
1,288.9
708.8
428.8
654.7
110.3
168.4
279.1
245.4
1,029.4
208.0
2,154.0
1,382.2
86.0
1,100.3
483.6
514.6
1,353.4
108.0
749.8
89.0
606.4
2,688.5
294.5
116.4
990.5
841.3
269.3
815.1
87.8
2030 Tons
Reduced
194.2
58.9
100.2
123.5
22.6
97.9
9.4
2.4
0.8
682.0
251.9
0.9
52.8
105.7
117.4
84.9
63.7
79.5
237.1
19.0
52.0
15.7
296.7
191.7
105.6
124.1
23.3
40.8
31.2
24.9
24.3
58.5
212.1
272.5
21.2
229.1
112.4
124.7
162.6
2.5
157.8
20.5
133.3
468.5
62.0
17.8
114.7
198.0
57.3
146.9
19.6
% Change
19.0
31.3
14.0
19.4
0.6
15.7
2.3
2.5
2.8
18.2
16.0
0.7
18.5
8.1
16.2
17.4
20.6
16.4
20.9
7.4
6.6
2.3
18.7
21.3
19.8
15.9
17.5
19.5
10.0
9.2
2.3
21.9
9.0
16.5
19.8
17.2
18.9
19.5
10.7
2.2
17.4
18.7
18.0
14.8
17.4
13.3
10.4
19.0
17.5
15.3
18.2
                                     2-70

-------
Final Regulatory Impact Analysis
 Table 2.2.-S6. MSAT Emissions (Tons) from Uncontrolled PFCs (No Fuel Benzene
        Control, No Federal PFC Control, But Including State Programs)
Pollutant
2,2,4-Trimethylpentane
Benzene
Ethylbenzene
n-Hexane
MTBE
Naphthalene
Toluene
Xylenes
TOTAL
1999
4,870
853
2,135
5,417
6,969
1
10,733
6,189
37,167
2010
4,994
943
1,805
4,679
0
1
8,764
5,004
21,186
2015
5,195
992
1,884
4,895
0
1
9,161
5,226
27,354
2020
5,573
1,063
2,021
5,250
0
1
9,825
5,605
29,338
2030
6,353
1,210
2,303
5,981
0
1
11,195
6,387
33,430
 Table 2.2.-S7. Reduction in Benzene PFC Emissions (Tons) with Fuel Control (No
                         Control on PFC Emissions).
Year
1999
2015
2020
2030
Reference Case
853
992
1,063
1,210
Control Case
N.A.
619
664
756
Reduction
N.A.
373
399
454
                                    2-71

-------
Final Regulatory Impact Analysis
  Table 2.2.-3S.  Reduction in Benzene PFC Emissions (Tons) with Fuel Control in
            2030 by State (No Additional Control on PFC Emissions).
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
Tons
33.1
19.3
25.8
23.7
36.4
31.5
3.5
1.1
0.4
138.2
42.5
5.1
12.8
42.6
38.0
20.8
20.8
25.3
39.2
1.7
7.0
5.9
64.1
40.9
24.6
31.7
8.0
12.7
11.9
1.7
8.7
14.0
17.9
54.9
4.5
33.3
20.8
29.9
16.3
0.9
31.5
4.8
41.8
42.5
16.7
0.8
11.5
44.5
14.3
26.0
4.2
Control Case
Tons
19.1
11.6
15.5
13.5
35.7
19.8
3.2
1.0
0.4
82.9
25.5
5.0
8.1
32.9
21.4
11.2
11.2
16.1
22.4
1.3
6.2
5.4
34.6
22.1
14.0
19.9
5.1
6.8
7.8
1.4
8.0
8.0
14.4
32.9
2.4
18.0
11.3
17.9
11.5
0.9
18.9
2.6
22.6
30.6
10.5
0.5
8.9
26.7
8.6
16.9
2.7
Reduction
14.0
7.7
10.3
10.2
0.7
11.7
0.3
0.1
0.0
55.3
17.0
0.1
4.7
9.6
16.6
9.6
9.6
9.2
16.9
0.3
0.9
0.5
29.5
18.8
10.6
11.8
3.0
5.8
4.1
0.3
0.7
6.0
3.5
21.9
2.1
15.3
9.6
12.0
4.8
0.1
12.6
2.2
19.2
12.0
6.2
0.3
2.6
17.8
5.7
9.1
1.6
% Change
42.3
40.0
40.0
43.0
2.0
37.0
8.0
8.0
8.0
40.0
40.0
2.0
37.0
22.6
43.7
46.0
46.0
36.4
43.0
20.0
12.5
8.0
46.0
46.0
43.0
37.2
37.0
46.0
34.5
16.5
8.0
43.0
19.6
40.0
46.0
46.0
46.0
40.0
29.5
8.0
40.0
46.0
46.0
28.1
37.0
40.0
22.5
40.0
40.0
34.9
37.0
                                     2-72

-------
Final Regulatory Impact Analysis
    Table 2.2.-S9.  Nationwide Impact of Controls on Emissions of Benzene from
                     Gasoline Distribution in 2015 and 2020.

Tons of
Benzene
2015
Reference
Case
2,445
2015
Control
Case
1,635
2015
Reduction
810
2020
Reference
Case
2,621
2020
Control
Case
1,772
2020
Reduction
849
                                      2-73

-------
Final Regulatory Impact Analysis
Table 2.2.-40. Reduction in Gasoline Distribution Emissions of Benzene (Tons) with
                       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
Tons
34.0
3.0
40.9
16.5
98.0
27.5
18.6
3.4
3.4
88.4
36.6
3.1
20.2
98.6
41.2
53.6
70.0
48.4
108.9
21.4
34.1
25.0
82.3
74.3
42.6
26.6
12.0
11.1
7.8
3.9
41.7
26.3
359.7
36.2
9.3
89.7
61.0
118.6
72.7
4.7
19.9
6.5
57.9
344.2
25.5
1.2
42.1
57.1
55.2
23.7
11.8
Control Case
Tons
19.6
1.8
24.6
9.4
96.1
17.3
17.1
3.1
3.1
53.1
22.0
3.1
12.7
69.0
23.7
29.0
37.8
28.4
62.1
16.3
28.6
23.0
44.5
40.1
24.3
16.6
7.5
6.0
5.7
3.2
38.3
15.0
313.4
21.7
5.0
48.4
32.9
71.2
46.1
4.4
11.9
3.5
31.3
243.6
16.0
0.7
29.8
34.3
33.1
15.1
7.4
Reduction
14.4
1.2
16.4
7.1
2.0
10.2
1.5
0.3
0.3
35.4
14.6
0.1
7.5
29.6
17.5
24.7
32.2
20.0
46.8
5.1
5.5
2.0
37.9
34.2
18.3
9.9
4.4
5.1
2.1
0.6
3.3
11.3
46.3
14.5
4.3
41.3
28.1
47.4
26.6
0.4
8.0
3.0
26.7
100.6
9.4
0.5
12.3
22.8
22.1
8.6
4.4
% Change
42.4
40.0
40.0
43.0
2.0
37.0
8.0
8.0
8.0
40.0
40.0
2.0
37.0
30.0
42.5
46.0
46.0
41.3
43.0
23.9
16.1
8.0
46.0
46.0
43.0
37.4
37.0
46.0
27.1
16.7
8.0
43.0
12.9
40.0
46.0
46.0
46.0
40.0
36.6
8.0
40.0
46.0
46.0
29.2
37.0
40.0
29.2
40.0
40.0
36.4
37.0
                                      2-74

-------
Final Regulatory Impact Analysis
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 from these vehicles
operating 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. The
emissions reduced are those created by the engine start following the vehicle soak.  These
parameters are currently modeled by vehicle class and vehicle age in MOBILE6.2.56'57'
58'59 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.  This estimate does not include the effects of fuel benzene control. Effects on the
trends in the inventories for the affected MSATs are shown in Table 2.2.-41 through
Table 2.2.-4S.

   Table 2.2.-41.  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
Propionaldehyde
Styrene
Toluene
Xylenes
Total MSATs
Emissions in Tons
20,868
175,241
21,035
2,234
173,474
69,299
54,104
61,664
54,990
2440
13,070
464,646
262,298
1,376,002
                                      2-75

-------
Final Regulatory Impact Analysis
 Table 2.2.-42. 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
Propionaldehyde
Styrene
Toluene
Xylenes
Total MSATs
Reference Case
Tons in Calendar
Year 2010
9,159
95,194
16,680
1,132
99,559
36,001
23,466
32,850
0
1254
6,688
238,683
134,742
695,408
Vehicle Control
Case Tons in
Calendar Year 2010
8,417
88,628
15,220
1,041
91,621
33,489
21,371
31,590
0
1144
6,107
220,939
124,744
644,312
Reduction
in Tons
742
6,566
1,460
91
7,939
2,512
2,095
1,260
0
110
581
17,744
9,998
51,987
Percent
Reduction
8
7
9
8
8
7
9
4
0
9
9
7
7
7
   Table 2.2.-4S. 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
Propionaldehyde
Styrene
Toluene
Xylenes
Total MSATs
Reference Case
Tons in Calendar
Year 2015
8,635
87,857
16,253
1,080
95,234
33,276
22,657
27,699
0
1216
6,481
223,510
126,114
650,012
Vehicle Control
Case Tons in
Calendar Year 2015
7,083
73,956
13,123
887
78,664
27,970
18,298
25,034
0
985
5,254
186,031
104,997
542,281
Reduction
in Tons
1,552
13,901
3,131
193
16,570
5,305
4,359
2,665
0
231
1,227
37,480
21,117
107,731
Percent
Reduction
18
16
19
18
17
16
19
10
0
19
19
17
17
17
                                     2-76

-------
Final Regulatory Impact Analysis
   Table 2.2.-44. 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
Propionaldehyde
Styrene
Toluene
Xylenes
Total MSATs
Reference Case
Tons in Calendar
Year 2020
9,131
89,711
17,345
1,139
99,225
33,992
24,007
25,765
0
1293
6,898
230,933
130,267
669,707
Vehicle Control
Case Tons in
Calendar Year 2020
6,592
66,807
12,143
822
72,128
25,268
16,922
21,380
0
914
4,880
169,303
95,543
492,700
Reduction
in Tons
2,539
22,904
5,203
317
27,097
8,724
7,086
4,385
0
379
2,018
61,630
34,725
177,007
Percent
Reduction
28
26
30
28
27
26
30
17
0
29
29
27
27
26
   Table 2.2.-4S. 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
Propionaldehyde
Styrene
Toluene
Xylenes
Total MSATs
Reference Case
Tons in Calendar
Year 2030
10,798
104,511
20,663
1,347
116,742
39,603
28,529
28,437
0
1540
8,207
270,625
152,647
783,648
Vehicle Control
Case Tons in
Calendar Year 2030
6,540
66,317
12,064
818
71,704
25,053
16,897
21,125
0
907
4,841
167,829
94,728
488,824
Reduction
in Tons
4,257
38,194
8,599
529
45,037
14,551
11,632
7,312
0
633
3,366
102,796
57,919
294,824
Percent
Reduction
39
37
42
39
39
37
41
26
0
41
41
38
38
38
                                      2-77

-------
Final Regulatory Impact Analysis
       State-level reductions in calendar year 2030 benzene inventories are reported in
Table 2.2.-46.
 Table 2.2.-46. Impacts of Vehicle Control on Light-Duty Gasoline Vehicle Benzene
                            Emissions, 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 Benzene Totals
2030 Reference
Case Benzene Tons
in Calendar 2030
2,199.4
1,293.1
1,734.6
1,323.0
9,286.2
2,674.7
1,019.9
264.8
109.9
4,032.2
4,076.8
183.7
1,132.6
4,004.2
3,335.4
1,564.5
1,362.3
1,952.6
1,502.6
657.6
1,689.2
1,649.1
5,560.1
4,038.3
946.8
2,787.8
785.0
970.8
989.8
632.5
1,815.8
1,698.0
4,421.5
3,836.4
509.0
4,536.7
1,771.7
3,631.9
4,026.6
320.3
1,998.9
516.9
2,495.0
6,111.9
1,705.9
457.1
3,271.3
5,778.8
852.8
2,651.4
573.9
116,741.6
2030 Control Case
Benzene Tons in
Calendar 2030
1,562.7
652.7
1,190.5
900.4
5,634.4
1,560.6
540.4
148.9
62.1
3,374.8
2,698.4
174.7
653.4
2,255.1
2,035.6
927.2
859.7
1,197.3
1,122.2
358.9
943.6
816.5
3,279.8
2,190.2
652.0
1,722.2
443.4
582.1
627.4
347.0
976.9
1,079.3
2,201.2
2,474.0
282.7
2,595.6
1,206.9
2,270.3
2,235.7
175.6
1,358.6
296.1
1,623.0
4,356.2
1,041.8
246.7
1,977.9
3,564.6
489.6
1,407.8
329.5
71,704.3
Reduction (Tons)
636.6
640.4
544.1
422.5
3,651.7
1,114.1
479.6
115.9
47.7
657.4
1,378.4
9.0
479.3
1,749.0
1,299.8
637.3
502.6
755.3
380.4
298.7
745.6
832.6
2,280.3
1,848.0
294.8
1,065.6
341.6
388.7
362.4
285.5
838.9
618.7
2,220.3
1,362.4
226.3
1,941.1
564.8
1,361.5
1,790.9
144.6
640.3
220.7
872.1
1,755.7
664.2
210.4
1,293.4
2,214.2
363.2
1,243.6
244.4
45,037.3
Percent Reduction
28.9
49.5
31.4
31.9
39.3
41.7
47.0
43.8
43.5
16.3
33.8
4.9
42.3
43.7
39.0
40.7
36.9
38.7
25.3
45.4
44.1
50.5
41.0
45.8
31.1
38.2
43.5
40.0
36.6
45.1
46.2
36.4
50.2
35.5
44.5
42.8
31.9
37.5
44.5
45.2
32.0
42.7
35.0
28.7
38.9
46.0
39.5
38.3
42.6
46.9
42.6
38.6
                                       2-78

-------
Final Regulatory Impact Analysis
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.-47 through 2.2.-50.  Table 2.2.-51 reports benzene reductions by State in 2030 as a
result of federal PFC control.

 Table 2.2-41.  Estimated Reductions in MSAT Emissions from PFC Control, 2010
                           (No Fuel Benzene Control).
Pollutant
2,2,4-Trimethylpentane
Benzene
Ethyl Benzene
n-Hexane
MTBE
Naphthalene
Toluene
Xylenes
Total
Reference Case
4,994
943
1,805
4,679
0
1
8,764
5,004
26,189
Control Case
4,039
743
1,450
3,742
0
1
7,021
4,015
21,010
Reduction in
Tons
955
201
355
937
0
0
1,743
989
5,179
Percent
Reduction
19
21
20
20
0
19
20
20
20
 Table 2.2.-4S. Estimated Reductions in MSAT Emissions from PFC Control, 2015
                           (No Fuel Benzene Control).
Pollutant
2,2,4-Trimethylpentane
Benzene
Ethyl Benzene
n-Hexane
MTBE
Naphthalene
Toluene
Xylenes
Total
Reference Case
5,195
992
1,884
4,895
0
1
9,161
5,226
27,355
Control Case
2,005
320
695
1,750
0
0
3,316
1,912
9,998
Reduction in
Tons
3,190
672
1,189
3,145
0
0
5,846
3,314
17,357
Percent
Reduction
61
68
63
64
0
61
64
63
63
                                     2-79

-------
Final Regulatory Impact Analysis
 Table 2.2.-49. Estimated Reductions in MSAT Emissions from PFC Control, 2020
                          (No Fuel Benzene Control).
Pollutant
2,2,4-Trimethylpentane
Benzene
Ethyl Benzene
n-Hexane
MTBE
Naphthalene
Toluene
Xylenes
Total
Reference Case
5,573
1,063
2,021
5,250
0
1
9,825
5,605
29,338
Control Case
2,163
345
750
1,888
0
0
3,577
2,063
10,785
Reduction in
Tons
3,410
718
1,271
3,362
0
0
6,248
3,543
18,553
Percent
Reduction
61
68
63
64
0
61
64
63
63
 Table 2.2.-50. Estimated Reductions in MSAT Emissions from PFC Control, 2030
                          (No Fuel Benzene Control).
Pollutant
2,2,4-Trimethylpentane
Benzene
Ethyl Benzene
n-Hexane
MTBE
Naphthalene
Toluene
Xylenes
Total
Reference Case
6,353
1,210
2,303
5,981
0
1
11,195
6,387
33,430
Control Case
2,486
396
862
2,169
0
0
4,110
2,370
12,394
Reduction in
Tons
3,867
814
1,442
3,812
0
1
7,085
4,017
21,036
Percent
Reduction
61
67
63
64
0
61
63
63
63
                                    2-80

-------
Final Regulatory Impact Analysis
 Table 2.2.-51. Reductions in Benzene Emissions (Tons) with PFC Control by State,
                        2030 (No Fuel Benzene Control).
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
33.1
19.3
25.8
23.7
36.4
31.5
3.5
1.1
0.4
138.2
42.5
5.1
12.8
42.6
38.0
20.8
20.8
25.3
39.2
1.7
7.0
5.9
64.1
40.9
24.6
31.7
8.0
12.7
11.9
1.7
8.7
14.0
17.9
54.9
4.5
33.3
20.8
29.9
16.3
0.9
31.5
4.8
41.8
42.5
16.7
0.8
11.5
44.5
14.3
26.0
4.2
Control Case
6.8
3.7
7.3
4.3
36.4
11.1
3.0
0.9
0.3
30.0
11.7
0.9
3.8
14.3
9.7
4.7
5.3
5.9
6.3
1.2
5.5
4.8
16.9
11.0
3.7
8.4
2.1
3.1
3.2
1.3
7.0
3.8
13.7
13.1
1.2
20.0
5.0
9.2
11.6
0.8
6.5
1.2
8.2
26.2
4.5
0.6
8.8
14.7
3.2
8.1
1.2
Reduction
26.3
15.6
18.6
19.4
0.0
20.4
0.4
0.3
0.1
108.3
30.8
4.2
9.0
28.3
28.3
16.1
15.5
19.3
32.9
0.4
1.6
1.0
47.1
29.9
20.9
23.3
6.0
9.5
8.7
0.4
1.7
10.2
4.2
41.7
3.4
13.3
15.8
20.6
4.7
0.2
25.0
3.6
33.6
16.3
12.1
0.2
2.7
29.8
11.1
17.9
3.0
% Change
79.5
80.9
71.8
82.0
0.0
64.9
12.8
22.5
32.3
78.3
72.4
82.4
70.7
66.5
74.4
77.5
74.5
76.6
83.9
26.4
22.1
17.6
73.6
73.1
85.1
73.4
74.2
75.2
73.1
24.4
19.5
73.0
23.5
76.1
73.9
39.9
75.9
69.1
28.7
18.9
79.4
75.1
80.3
38.4
72.7
26.0
23.6
66.9
77.5
69.0
70.5
                                     2-81

-------
Final Regulatory Impact Analysis
2.2.2.2.4 Cumulative Reductions of 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 PFCs
depend on both fuel benzene content and the PFC emission controls.  Tables 2.2.-52 and
2.2.-53 summarize the expected reductions in benzene and MSAT emissions,
respectively, from the combined effects of our vehicle, fuel, and PFC controls.

       Table 2.2.-54 summarizes the cumulative benzene emission reductions from these
controls on highway gasoline vehicles, nonroad gasoline vehicles, PFCs, and gasoline
distribution at the State level in 2030.

       Table 2.2.-55 presents the impact of controls on total benzene emissions from
mobile sources and PFCs, and the impacts on total benzene emissions from all sources.
Table 2.2.-56 presents the cumulative  impact of controls on total emissions of MSATs
from mobile sources and PFCs, as well as the impact on total emissions of MSATs from
both mobile and stationary sources.  As discussed previously, the fuel benzene control
reduces stationary source emissions of benzene associated with gasoline distribution.
                                      2-82

-------
Final Regulatory Impact Analysis
        Table 2.2-52. Estimated Reductions in Benzene Emissions from All Control Measures by Sector, 2015 to 2030.
Benzene
Gasoline
Onroad Mobile
Sources
Gasoline
Nonroad
Mobile Sources
PFCs
Gasoline
Distribution
Total
1999

183,660
68,589
853
1,984
255,086
2015
Without
Rule (tons)
97,789
41,343
992
2,445
142,569
With Rule
(tons)
71,688
35,825
215
1,635
109,363
Reduction
(tons)
26,101
5,518
111
810
33,206
2020
Without
Rule (tons)
101,514
40,161
1,063
2,621
145,359
With Rule
(tons)
65,878
34,717
232
1,772
102,599
Reduction
(tons)
35,636
5,444
831
849
42,760
2030
Without
Rule (tons)
119,016
42,994
1,210
2,621
165,841
With Rule
(tons)
65,601
37,167
267
1,772
104,807
Reduction
(tons)
53,415
5,827
944
849
61,035
                                                          2-83

-------
Final Regulatory Impact Analysis
         Table 2.2.-5S.  Estimated Reductions in MSAT Emissions from All Control Measures by Sector, 2015 to 2030
MSAT
Gasoline
Onroad Mobile
Sources
Gasoline
Nonroad
Mobile Sources
PFCs
Gasoline
Distribution
Total
1999

1,452,739
806,725
37,166
57,765
2,354,395
2015
Without
Rule (tons)
675,781
449,422
27,355
62,870
1,215,428
With Rule
(tons)
558,666
443,973
9,893
62,059
1,074,591
Reduction
(tons)
117,115
5,449
17,462
811
140,837
2020
Without
Rule (tons)
693,189
406,196
29,338
64,942
1,193,665
With Rule
(tons)
507,782
400,816
10,672
64,092
983,362
Reduction
(tons)
185,408
5,380
18,666
850
210,303
2030
Without
Rule (tons)
808,141
412,617
33,430
64,942
1,319,130
With Rule
(tons)
505,074
406,856
12,264
64,092
988,286
Reduction
(tons)
303,067
5,761
21,166
850
330,844
                                                          2-84

-------
Final Regulatory Impact Analysis
Table 2.2.-S4. Cumulative Benzene Emission Reductions From All Controls at the State Level 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
Gasoline Highway
Vehicles
Tons Reduced
826.3
849.8
665.3
534.7
3,675.9
1,348.7
488.0
118.3
48.7
1,090.6
1,667.6
9.8
588.6
1,966.1
1,591.8
780.9
658.6
915.2
528.2
315.5
776.5
845.7
2,799.1
2,270.1
378.9
1,316.8
413.0
487.8
402.7
%
36.6
65.1
37.2
39.6
39.0
49.4
47.2
43.9
43.5
26.1
39.9
5.1
51.2
48.2
46.9
49.4
47.5
46.0
34.3
47.2
45.2
50.0
49.6
55.5
39.2
46.4
52.0
49.5
39.4
Nonroad Gasoline Engines
Tons
Reduced
194.2
58.9
100.2
123.5
22.6
97.9
9.4
2.4
0.8
682.0
251.9
0.9
52.8
105.7
117.4
84.9
63.7
79.5
237.1
19.0
52.0
15.7
296.7
191.7
105.6
124.1
23.3
40.8
31.2
%
19.0
31.3
14.0
19.4
0.6
15.7
2.3
2.5
2.8
18.2
16.0
0.7
18.5
8.1
16.2
17.4
20.6
16.4
20.9
7.4
6.6
2.3
18.7
21.3
19.8
15.9
17.5
19.5
10.0
PFCs
Tons
Reduced
29.1
17.1
21.5
21.3
0.7
24.5
0.7
0.3
0.2
120.2
35.4
4.2
10.4
31.5
32.5
18.3
18.0
21.5
35.6
0.7
2.2
1.4
54.9
35.0
22.5
26.4
6.7
11.0
9.8
%
88.2
88.5
83.1
89.7
2.0
77.9
19.8
28.7
37.7
87.0
83.4
82.7
81.5
74.0
85.6
87.9
86.3
85.1
90.8
41.1
31.8
24.2
85.7
85.5
91.5
83.3
83.7
86.6
82.4
Gasoline Distribution
Tons
Reduced
14.4
1.2
16.4
7.1
2.0
10.2
1.5
0.3
0.3
35.4
14.6
0.1
7.5
29.6
17.5
24.7
32.2
20.0
46.8
5.1
5.5
2.0
37.9
34.2
18.3
9.9
4.4
5.1
2.1
%
42.4
40.0
40.0
43.0
2.0
37.0
8.0
8.0
8.0
40.0
40.0
2.0
37.0
30.0
42.5
46.0
46.0
41.3
43.0
23.9
16.1
8.0
46.0
46.0
43.0
37.4
37.0
46.0
27.1
Total
Tons
Reduced
1,064.0
926.9
803.4
686.5
3,701.2
1,481.2
499.6
121.4
49.9
1,928.2
1,969.5
15.0
659.3
2,132.9
1,759.2
908.8
772.5
1,036.1
847.8
340.3
836.2
864.8
3,188.6
2,530.9
525.4
1,477.3
447.5
544.6
445.8
%
31.7
61.2
31.2
33.9
27.2
43.4
34.0
32.6
34.7
23.6
33.8
4.6
44.9
38.7
41.9
42.4
43.3
40.7
30.0
35.9
32.8
36.0
43.2
49.6
33.5
40.2
47.2
44.7
33.0
                                                           2-85

-------
Final Regulatory Impact Analysis

State
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
301.2
855.1
814.4
2,354.8
1,648.3
276.2
2,326.7
717.4
1,774.1
1,963.3
147.3
791.2
267.2
1,093.5
2,255.3
821.5
230.9
1,427.6
2,848.3
425.9
1,417.3
298.3
%
47.0
46.0
46.8
52.1
42.0
53.6
50.4
39.7
47.6
47.9
45.4
38.8
51.0
43.0
35.8
47.5
49.8
43.1
48.6
49.4
52.6
51.3
Nonroad Gasoline Engines
Tons
Reduced
24.9
24.3
58.5
212.1
272.5
21.2
229.1
112.4
124.7
162.6
2.5
157.8
20.5
133.3
468.5
62.0
17.8
114.7
198.0
57.3
146.9
19.6
%
9.2
2.3
21.9
9.0
16.5
19.8
17.2
18.9
19.5
10.7
2.2
17.4
18.7
18.0
14.8
17.4
13.3
10.4
19.0
17.5
15.3
18.2
PFCs
Tons
Reduced
0.6
2.3
11.8
6.9
47.0
3.9
22.5
18.1
24.3
8.1
0.2
27.6
4.2
37.4
23.7
13.8
0.5
4.7
35.7
12.3
20.7
3.4
%
36.9
26.0
84.6
38.5
85.6
85.9
67.5
87.0
81.5
49.7
25.4
87.6
86.6
89.4
55.7
82.8
55.6
40.8
80.2
86.5
79.8
81.4
Gasoline Distribution
Tons
Reduced
0.6
3.3
11.3
46.3
14.5
4.3
41.3
28.1
47.4
26.6
0.4
8.0
3.0
26.7
100.6
9.4
0.5
12.3
22.8
22.1
8.6
4.4
%
16.7
8.0
43.0
12.9
40.0
46.0
46.0
46.0
40.0
36.6
8.0
40.0
46.0
46.0
29.2
37.0
40.0
29.2
40.0
40.0
36.4
37.0
Total
Tons
Reduced
327.4
885.1
896.0
2,620.1
1,982.3
305.6
2,619.6
876.0
1,970.6
2,160.6
150.4
984.6
294.9
1,290.9
2,848.0
906.7
249.7
1,559.2
3,104.8
517.7
1,593.5
325.7
%
35.7
29.9
43.8
36.1
35.0
48.0
43.1
35.2
43.7
37.9
34.1
32.8
45.8
38.1
28.9
42.6
41.6
34.9
44.4
41.1
43.0
46.2
                                                                   2-86

-------
Final Regulatory Impact Analysis
            Table 2.2-55.  Impact of Controls on Total Benzene Emissions From Mobile Sources and All Sources.

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

16,804
16,570
32,396

16,768
27,097
41 ,91 1

18,796
45,037
60,186
Mobile Source
and PFC Tons,
Reference Case

140,124
140,124
140,124

142,738
142,738
142,738

163,220
163,220
163,220
% of Mobile
Source and
PFC Tons
Reduced

12
12
23

12
19
29

12
28
37
Total Tons
Reduced From
All Sources

17,614
16,570
33,206

17,617
27,097
42,760

19,645
45,037
61,035
Total Mobile
and Stationary
Tons, Reference
Case

256,755
256,755
256.755

262,828
262,828
262,828

283,310
283,310
283,310
% of Total
Benzene
Reduced

7
7
13

7
10
16

7
16
22
                                                          2-87

-------
Final Regulatory Impact Analysis
  Table 2.2.-56.  Cumulative Impact of Controls on Total MSAT Emissions From Mobile Sources and PFCs, and From All Sources.

2015
2020
2030
Mobile
Source and
RFC Tons
Reduced
140,726
210,173
330,713
Mobile Source
and RFC Tons,
Reference
Case
1,215,146
1,193,281
1,318,746
% of Mobile
and RFC
Tons
Reduced
12
18
25
Total Tons
Reduced
141,536
211,022
331,562
Total Mobile
and
Stationary
Tons
2,636,063
2,733,020
2,858,485
% of Mobile and
Stationary Tons
Reduced
5
8
12
                                                          2-88

-------
Final Regulatory Impact Analysis


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.60'61
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.62

       Eastern Research Group, under contract to EPA,  recently compared exhaust
emissions data from newer technology vehicles to see if the toxics 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.63  The data from EPA's Office of Research and Development have been
published.64

       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.
                                      2-89

-------
Final Regulatory Impact Analysis
   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. As part of this ongoing
effort, EPA is reviewing engine exhaust data, which includes air toxic emissions, from
the CRC (Coordinating Research Council) E-67 study on engine emissions from light-
duty vehicles using ethanol fuels.65 Likewise, 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
                                      2-90

-------
Final Regulatory Impact Analysis
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,
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 (NATA).  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 blower.72 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 NMIM 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
NMEVI. 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
                                      2-91

-------
Final Regulatory Impact Analysis
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.

       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
                                     2-92

-------
Final Regulatory Impact Analysis
       Table 2.S.-2. NMIM MSAT Fractions versus Fractions from EPA Testing of
18 Handheld SI Engines
Type Standard Fuel
BLOWER Ph2 CG
SAW CG
SAW Phi CG
SAW Phi RFG
SAW Ph2 CG
SAW Ph2 RFG
TPJMMER CG
TRIMMER RFG
TRIMMER Phi CG
TRIMMER Phi CG
TRIMMER Phi RFG
TRIMMER Phi RFG
TRIMMER Ph2 CG
TRIMMER Ph2 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

0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002

0.0002

0.0002
0.0002
0.0002
Toluene
NMTM
0.0978
0.0598
0.0598
0.0598
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
NMTM
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
equipment
          73,74,75
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 this new emission data, the Agency has begun an effort to update the toxics
                                      2-93

-------
Final Regulatory Impact Analysis
portion of its NMIM model.  EPA will also address differences between Tier 1 and
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 for creating category- and power-specific TAFs for NONROAD.

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 75% between 2001 and 2030. 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.SM

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. Some fuel programs will have a mixed impact on the species and quantity of
MSAT emissions expected with the introduction of these new fuels into commerce.

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%
                                      2-94

-------
Final Regulatory Impact Analysis
reduction is required. These reductions are determined using the Complex Model. As
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.0  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 EPA's Complex Model) were 104.5 mg/mileD;
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, MS ATI 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.E 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
     cSee RFG rule for why evaporative emissions are not included in the anti-dumping toxics
determination.
     DPhase II
     EExcept 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).
                                       2-95

-------
Final Regulatory Impact Analysis
importers in business during the baseline period had sufficient data to establish an
individual baseline.  An MSAT1 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

       Beginning in 2006, EPA's gasoline sulfur program77 requires that sulfur levels in
gasoline can be no higher than 80 ppm as a per gallon cap and must average 30 ppm
annually. When the program is 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 as well as
emissions of NOx. 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 to 9 psi and wintertime RVP, when additional volatility is required
for starting in cold temperatures, ranges  from about 9 to 14 psi.  Gasoline vapors contain a
subset of the liquid gasoline components and thus can contain toxics compounds, such as
benzene. Since 1989, EPA has controlled summertime gasoline RVP primarily as a VOC
and ozone precursor control, resulting in additional 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 during the transition to the 15 ppm
level, 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.
                                      2-96

-------
Final Regulatory Impact Analysis
       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
provisions will 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.

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.
                                      2-97

-------
Final Regulatory Impact Analysis
    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
and enhanced evaporative
emissions regulations)
February 10, 2000
NLEV (National Low-Emitting
Vehicle)
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*
X
X
X
X
X

X
PM
Standards
X
X

X
X
X

* Standards in various forms including HC, NMHC, NMOG, and NOx+NMHC
                                     2-98

-------
Final Regulatory Impact Analysis
              Table 2.4-2 Current Nonroad Engine/Vehicle Programs.
Category
^and-based diesel
^ocomotives
Vlarine



Large spark-
ignition engines
Small spark-
ignition engines

Aircraft
(NOx Std in 2005;
Smoke Std in 1982)
Recreational
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 - 2006
2007, and later
1997-2007
2002 - 2007
2001 -2007
No current/recent
standards for VOC or
PM
2006-2012
VOC
Standards*
X
X
X
X
X

X
X
X
X
X


X
PM
Standards
X
X
X
X

X
X
X






* 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).
Table 2.4-2 shows current mobile source programs for nonroad engines. Brief summaries
                                       2-99

-------
Final Regulatory Impact Analysis
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 engines, mining
equipment or 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 Land-Based SI Engines

       The category of land-based nonroad SI engines  is comprised of a broad mix of
service and recreational equipment with engines which range from less than 10
horsepower to several hundreds of horsepower. Most of these  engines have been subject
to one or more tiers  of engine emission controls for some time, while others in the
category, such as recreational vehicles, are just coming under engine emission control
regulations in 2006.

2.4.3.2.1 Large Land-Based SI Engines

       Since the MS ATI rule was published, we have also finalized emissions standards
                                                       OT
for SI engines above 25 hp used in commercial applications.   Such engines are used in a
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
                                      2-100

-------
Final Regulatory Impact Analysis
will provide about a 90 percent reduction in HC emissions on average for new engines
versus Tier 1 controlled engines.

2.4.3.2.2 Recreational Vehicles

       In 2006, new recreational vehicles, which include snowmobiles, off-road
motorcycles and "all terrain vehicles", began a first tier of engine emission standards.
These standards require significant reductions in HC emissions from new engines,
ranging from 50 to 86 percent compared to pre-controlled engines.84

2.4.3.2.3 Small Land-B ased SI Engines

       Small land-based spark-ignition (small SI) engines at or below 25 hp may be
either handheld or non-handheld and are used primarily in lawn and garden equipment
such as walk-behind and tractor mowers, string trimmers, chain saws and other similar
equipment.  Our Phase 1  exhaust 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.85  We also
have Phase 2 regulations for these engines which, when fully phased-in, are projected to
result in additional combined HC and NOx exhaust emission reductions beyond the Phase
1 levels of 60 percent for new non-handheld engines and of 70 percent for new handheld
engines.86 We are currently developing a proposal for new combined HC and NOx
exhaust standards for Phase 3 non-handheld small SI engines that should be
approximately  35 to 40 percent lower than present Phase 2 standards for this class of
engine.  Further, we also expect to propose new evaporative emission standards for small
SI engines and equipment to control fuel hose permeation, fuel tank permeation, diurnal
and running loss emissions.

       Phase 3 standards for Small SI engines are expected to achieve toxics benefits
through reduction of engine VOC emissions from three sources. The new standards
would result in fewer evaporative VOC and, therefore, air toxics emissions by lowering
hose and tank permeation losses for these types of equipment. Phase 3 engines will also
have lower exhaust VOC emissions under these new standards. Finally, Phase 3 Small  SI
engines are expected to achieve a small fuel economy benefit during operation. While
small, VOC emission savings from increased fuel economy will feed back through a
reduced number of gallons of fuel kept onboard these engines during operation.  This will
result in less VOC from tank/hose permeation, and less fuel burned overall will mean
fewer exhaust emissions

2.4.3.3 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.
                                      2-101

-------
Final Regulatory Impact Analysis
2.4.3.3.1 Marine SI Engines

       For gasoline-fueled outboard and personal-watercraft engines, we adopted an
initial phase of exhaust standards which became fully implemented with the 2006 model
year.  These standards have led to a major technology shift in this category of engines to
four-stroke engines and advanced-technology two-stroke engines, for an estimated 75
percent reduction in HC emissions from uncontrolled levels.87 We are developing a
proposal to adopt new, more stringent exhaust standards for these engines that would
further reduce emissions from this class of engines by an additional 60 percent or more
from the initial phase of standards  .

       Another class of marine SI engine, referred to as stern drive and inboard engines,
uses automotive-type engines. 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
                                                                             00
new emission standards for these engines in an upcoming marine SI engine proposal.
These new standards would likely be based on the application of catalyst technology to
substantially reduce HC and NOx emissions from the operation of these engines.

       The proposals mentioned above will also cover fuel evaporative emission
standards for fuel lines, fuel tanks and diurnal venting emissions for vessels powered by
gasoline-fueled engines in both of these engine classes.

2.4.3.3.2 Marine Diesel Engines

       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
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.
                                      2-102

-------
Final Regulatory Impact Analysis
       EPA is also investigating the possibility of designating U.S. coastal areas as SOx
(oxides of sulfur) 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.

2.4.3.4 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."F  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.

2.4.3.5 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.

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
     F To "remanufacture" an engine is to rebuild that engine to "new condition" at the end of four-to-
eight year long maintenance cycles.
                                      2-103

-------
Final Regulatory Impact Analysis
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 paniculate 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.  Almost 300 clean diesel projects have been initiated through the
Campaign. These projects will reduce more than 20,000 PM lifetime tons. PM and NOx
reductions from these programs will provide nearly $5 billion in health benefits.

       Reducing vehicle idling provides important environmental benefits. As a part of
their daily routine, truck drivers often keep their vehicles running 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 to take
advantage of proven systems that provide drivers with basic necessities without idling the
main engine. To date, there are 80 mobile and stationary idle-reduction projects
throughout the country. Emission reductions, on an annual basis, from these programs
are in excess of 157,000 tons of CO2, 2,000 tons of NOX and 60 tons of PM; over 14
million gallons of fuel are being saved annually. The SmartWay Transport Partnership
also works with the freight industry 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 recognizes employers that have met the National Standard of Excellence for
Commuter Benefits by adding them to the List of Best Workplaces for Commuters™.
These companies offer superior commuter benefits such as transit subsidies for rail, bus,
and vanpools and promote flexi-place and telework.  Emergency Ride Home programs
provide a safety net for  participants. More than 1,600 employers representing 3.5 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 over 14
major metropolitan areas to increase the penetration of commuter benefits in the
marketplace and the visibility of the companies that have received this distinguished
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
                                      2-104

-------
Final Regulatory Impact Analysis


granted the Best Workplaces for Commuters™ "District" designation to over twenty
locations across the country including sites in downtown Denver, Houston, Minneapolis,
Tampa, and Boulder.
                                      2-105

-------
Final Regulatory Impact Analysis


References for Chapter 2


1 Federal Register. Regulation of Fuels and Fuel Additives: Renewable Fuel Standard
Program; Proposed Rule, 70(184): Page 55551-55600, September 22, 2006.

2 Renewable Fuel Standard Program, Draft Regulatory Impact Analysis, EPA420-D-06-
008, September 2006. http://www.epa.gov/otaq/renewablefuels/index.htm.

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

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

5 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.

6 "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

7 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.

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

9 Landman, L. C. (2007) 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-R-07-001. This document is
available in Docket EPA-HQ-OAR-2003-0053.

10 Watson,!., Fujita,E., Chow,J., Zielinska,B., Richards,L., Neff,W., Dietrich,D. Northern
Front Range Air Quality Study Final Report: Volume 1.  June 30, 1998. Colorado State
University. This document is available in the EPA Docket as EPA-HQ-OAR-2005-0036-
1179.

11 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.
                                     2-106

-------
Final Regulatory Impact Analysis
12 Bailey, C.R. (2005) Cold-temperature exhaust particulate matter emissions.
Memorandum to Docket EPA-HQ-OAR-2005-0036.

13 Energy Information Agency, Annual Energy Outlook 2006, Table 17.
http://www.eia.doe.gov/oiaf/forecasting.html

14 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
Coordinating Research Council  and U. S. EPA.  March, 2004.
http://www.epa.gov/otaq/m6.htmtfm60

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

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

17 Strum, M., Cook, R., Pope, A., Palma, T., Shedd, S., Mason, R., Michaels, H.,
Thurman, J., Ensley, D. 2006. Projection of Hazardous Air Pollutant Emissions to
Future Years. Science of the Total Environment, 366: 590-601.

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/ei 13/index.html. This document is available in
Docket EPA-HQ-OAR-2005-0036.

20 U. S. EPA. NMEVI2005. http://www.epa.gov/otaq/nmim.htm

21 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.

22 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

                                     2-107

-------
Final Regulatory Impact Analysis
Emissions from 1985-1987 Model Year Light-Duty Gasoline Vehicles — II. Atmospheric
Environment 24A: 2105-2112.

23 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.

24 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.

25 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.

26 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.

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

28 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.

29 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.
Prepared for the American Petroleum Institute.  February 26. This document is available
in Docket EPA-HQ-OAR-2005-0036.

30 MathPro. 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.

31 Mullen, M., Neumann, J. Technical Memorandum: Documentation of 2003  VMT
Projection Methodology, Prepared by E. H. Pechan and Associates and Industrial
                                     2-108

-------
Final Regulatory Impact Analysis
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.

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

33 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.

34 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.

35 Federal Aviation Administration, 2004.  Terminal Area Forecast System.
http://www.apo.data.faa.gov/main/taf.asp.

36 Hester, Charles. 2006.  Review of Data on HAP Content in Gasoline. Memorandum
from MACTEC to Steve Shedd, U. S. EPA, March 23, 2006. This document is available
in Docket EPA-HQ-OAR-2003-0053.

37 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.

38 Hester, Charles. 2006.  Review of Data on HAP Content in Gasoline. Memorandum
from MACTEC to Steve Shedd, U. S. EPA, March 23, 2006. This document is available
in Docket EPA-HQ-OAR-2003-0053.

39 U. S. EPA.  2006.  Determination that Gasoline Distribution Stage 1 Area Source (GD
AS) Category Does Not Need to Be Regulated Under Section 112(c)6.  Memoradum
from Stephen Shedd to Kent Hustvedt, May  9, 2006.  This document is available in
Docket EPA-HQ-OAR-2005-0036.

40 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.
                                    2-109

-------
Final Regulatory Impact Analysis
41 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.

42 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.

43 Strum, M., Cook, R., Pope, A., Palma, T., Shedd, S., Mason, R., Michaels, H.,
Thurman, J., Ensley, D. 2006. Projection of Hazardous Air Pollutant Emissions to
Future Years. Science of the Total Environment., 366: 590-601.

44 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.

45 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.

46 U. S. Environmental Protection Agency. User's Guide for the Emissions Modeling
System for Hazardous Air Pollutants (EMS-HAP) Version 3.0 EPA-454/B-03-006
August 2004 by: Madeleine Strum, Ph.D. U.S. EPA, Office of Air Quality Planning and
Standards Emissions, Monitoring and Analysis Division Research Triangle Park, NC
And, Under Contract to the U.S. This document is available in Docket EPA-HQ-OAR-
2005-0036 and at http://www.epa.gov/scram001/userg/other/emshapv3ug.pdf
47
  Regional Economic Models, Inc.  2004. REMI Policy Insight, http://www.remi.com.
48 Fan W, Treyz F, Treyz G.  An evolutionary new economic geography model. J
Regional Sci 2000 40: 671- 696.

49 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.
                                     2-110

-------
Final Regulatory Impact Analysis
50 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.

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

52 Cook, R., Beidler, A., Touma, J., Strum, M.  2006. Preparing Highway Emissions
Inventories for Urban Scale Modeling: A Case  Study in Philadelphia. Transportation
Research Part D: Transport and Environment, 11: 396-407.

53 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

54 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

55 U. S. EPA.  2006. Regulatory Impact Analysis for the Review of the Particulate Matter
National Ambient Air Quality Standards. Office of Air Quality Planning and Standards,
Research Triangle Park, NC.  http://www.epa.gov/ttn/naaqs/standards/pm/s_pm_cr_ria.html.  This
document is available in Docket EPA-HQ-OAR-2006-0834.

56 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.

57 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.

58 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.

59 Glover, E.; Brzezinski, D.  2001.  Exhaust Emission Temperature
                                     2-111

-------
Final Regulatory Impact Analysis
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.

60 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.

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

62EPA. 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.

63 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.

64 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.

65 Durbin,T., Miller,!.W., Younglove,T., Huai,T. and Cocker,K.  "Effects of Ethanol and
Volatility Parameters on Exhaust Emissions: CRC Project No. E-67." January 30, 2006.
Coordinating Research Council # 06-VE-59596-E67.

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.
                                     2-112

-------
Final Regulatory Impact Analysis
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.

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)

  65 FR 6697, February 10, 2000.
                                    2-113

-------
Final Regulatory Impact Analysis
81 66 FR 5001, January 18, 2001.




82 69 FR 38958, June 29, 2004.




83 67 FR 68241, November 8, 2002.




84 67 FR 68241, November 8, 2002.




85 60 FR 34582, July 3, 1995.




86 64 FR 15208, March 30, 1999 and 65 FR 24267, April 25, 2000.




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.
                                   2-114

-------
Final Regulatory Impact Analysis

                              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	9
     3.1.3   Elevated Concentrations and Exposures in Mobile Source-Impacted Areas	12
        3.1.3.1 Concentrations Near Major Roadways	14
           3.1.3.1.1     Particulate Matter	14
           3.1.3.1.2     Gaseous Air Toxics	15
        3.1.3.2 Exposures Near Major Roadways	16
           3.1.3.2.1     In Vehicles	16
           3.1.3.2.2     InHomesand Schools	19
           3.1.3.2.3     Pedestrians and Bicyclists	21
        3.1.3.3 Concentrations and Exposure in Homes with Attached Garages	22
        3.1.3.4 Concentrations and Exposure in Parking Garages	25
        3.1.3.5 Concentrations and Exposure at Service Stations	28
        3.1.3.6 Occupational Exposure	29
     3.1.4   Uncertainties in Air Toxics Measurements	31
   3.2    Modeled Air Quality, Exposures, and Risks for Air Toxics	31
     3.2.1   National-Scale Modeled Air Quality, Exposure, and Risk for Air Toxics	31
        3.2.1.1 Air Quality Modeling	34
           3.2.1.1.1     Methods	34
           3.2.1.1.2     Air Quality Trends for Air Toxics: Reference Case	35
           3.2.1.1.3     Distributions of Air Toxic Concentrations across the U. S.: Reference
                 Case 38
           3.2.1.1.4     Impacts of Controls on Ambient Concentrations	39
        3.2.1.2 Exposure and Risk Modeling	44
           3.2.1.2.1     Methods	44
           3.2.1.2.2     Exposure and Risk Trends for Air Toxics: Reference Case	48
           3.2.1.2.3     Distributions of Air Toxics Risk across the U. S.: Reference Case	57
           3.2.1.2.4     Impacts of Controls on Average Inhalation Cancer Risks and
                 Noncancer Hazards	67
        3.2.1.3 Strengths and Limitations	74
        3.2.1.4. Perspective  on Cancer Cases	77
     3.2.2   Local-Scale Modeling	79
   3.3 Ozone	83
     3.3.1   Science of Ozone Formation	84
     3.3.2   Health Effects of Ozone	85
     3.3.3   Current 8-Hour Ozone Levels	86
     3.3.4   Projected 8-Hour Ozone Levels	88
        3.3.4.1 CAIR Ozone Air Quality Modeling	88
        3.3.4.2 Ozone Response Surface Metamodel Methodology	93
                                          3-1

-------
Final Regulatory Impact Analysis

        3.3.4.3 Ozone Response Surface Metamodel Results	96
     3.3.5  Environmental Effects of Ozone Pollution	97
        3.3.5.1 Impacts on Vegetation	97
   3.4ParticulateMatter	99
     3.4.1  Science of PM Formation	99
     3.4.2  Health Effects of Paniculate Matter	100
        3.4.2.1 Short-Term Exposure Mortality and Morbidity Studies	100
        3.4.2.2 Long-Term Exposure Mortality and Morbidity Studies	101
        3.4.2.3 Roadway-Related Pollution Exposure	102
     3.4.3  Current and Projected PM Levels	102
        3.4.3.1 Current PM2.5 Levels	102
        3.4.3.2 Current PMio Levels	103
        3.4.3.3 Projected PM2.5  Levels	104
           3.4.3.3.1    PM Modeling Methodology	104
           3.4.3.3.2   Areas at Risk of Future PM2.5 Violations	104
     3.4.4  Environmental Effects of PM Pollution	107
        3.4.4.1 Visibility Impairment	107
           3.4.4.1.1    Current Visibility Impairment	108
           3.4.4.1.2   Current Visibility Impairment at Mandatory Class I Federal Areas.. 108
           3.4.4.1.3    Future Visibility Impairment	109
           3.4.4.1.4   Future Visibility Impairment at Mandatory Class I Federal Areas.... 109
        3.4.4.2 Atmospheric Deposition	110
           3.4.4.2.1    Heavy Metals	Ill
           3.4.4.2.2   Polycyclic Organic Matter	112
        3.4.4.3 Materials Damage and Soiling	112
   3.5     Health and Welfare Impacts of Near-Roadway Exposure	113
     3.5.1  Mortality	114
     3.5.2  Non-Allergic Respiratory Symptoms	116
     3.5.3  Development of Allergic Disease and Asthma	117
     3.5.4  Cardiovascular Effects	118
     3.5.5  Birth Outcomes	119
     3.5.6  Childhood Cancer	120
     3.5.7  Summary  of Near-Roadway Health Studies	121
       Appendix 3 A: Influence of Emissions in Attached Garages on Indoor Air Benzene
Concentrations and Human Exposure	124
Appendix 3B: 8-Hour Ozone Nonattainment	144
Appendix 3C: PM Nonattainment	159
Appendix 3D: Visibility Tables	162
                                          5-2

-------
Final Regulatory Impact Analysis

 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 are 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 have 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. In 2004, EPA published a draft National Air Toxics
Monitoring Strategy to advance this goal.3 The National Air Toxics Trends Station (NATTS)
monitoring network is currently in place,  consisting of 23 sites in 22 urban areas nationally.4

       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
                                                      Detroit        •
                                           Cedar Rapids     ^      Providence
                           Grand Junction
                                                       Charleston
                   San Jacinto     Rio Rancho
                                                                San Juan
                                                      Tampa •       |
                                           5-3

-------
Final Regulatory Impact Analysis

       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 the NATTS,
and numerous community-scale monitoring studies.5  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
                  c 4
                  o
                  o
                           90% of sites have concentrations below this line
                           of sites have concentrations below this line
94      95      Q6      97      99
             1994-00: 47% decrease
                                                             99
00
       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.

       EPA recently completed a study of the spatial and temporal trends in ambient air toxics
data within the NATTS and other networks from 1990-2003.6  Most data came from urban
monitors.  Nationally, citywide average annual concentrations  of benzene,  formaldehyde, and
                                          3-4

-------
Final Regulatory Impact Analysis

acetaldehyde, varied by about a factor of five, and 1,3 -butadiene by more than 10 times.  The
coefficient of variation21 of annual average concentrations between different monitors within the
same city averaged 0.37 for benzene, 0.45 for 1,3-butadiene. Between cities, the coefficient of
variation could vary substantially.  Different pollutants showed different seasonal trends, with
average concentrations of benzene and 1,3-butadiene being highest in colder seasons, while
average concentrations of formaldehyde and acetaldehyde were higher during warm seasons,
reflecting the high photochemical production of aldehydes.  The concentrations of benzene,
butadiene, and acetaldehyde fell substantially over different time periods.  From 1990-2003,
benzene concentrations fell by 57%. Insufficient data existed in earlier years to analyze 1,3-
butadiene and acetaldehyde.  Formaldehyde increased by 134% over this period, although
changes in sampling methodology at some sites around 1995 make this quantification suspect.
From 1995-2003, the average annual changes in benzene, 1,3-butadiene, formaldehyde, and
acetaldehyde were -47%, -54%, +11%, and -12%. From 1998-2003, the changes were -21%, -
46%, +17%, and -4%, respectively. For benzene, these trends were statistically significant, but
formaldehyde and acetaldehyde trends after 1995 were not.

       One recent publication evaluated the trends in ambient concentrations of benzene and
1,3-butadiene in the Houston, TX metropolitan area.7 Using data from two air monitoring
networks, a state-based network and the Photochemical Assessment Monitoring Systems, the
study constructed a statistical model, controlling for meteorology and seasonality, to evaluate
trends in ambient toxics over the 1997-2004 time period. Averaged over state monitoring sites
with data across the time period, the model estimated 1.7% and 3.7% average annual decrease in
ambient benzene and 1,3,-butadiene, respectively. Mobile source and point source emission
reductions contributed roughly equally to this change. Examining long-term average
concentration differences across monitoring sites, benzene varied by roughly two-fold across
monitors while  1,3-butadiene varied roughly six-fold across monitors.  This may be attributable
to the substantial contribution of industrial sources to the local 1,3-butadiene inventory, while the
benzene inventory is dominated by mobile sources.  The study also evaluated differences in
weekday and weekend concentration, with the model predicting significant meteorologically-
adjusted concentration weekday increases relative to weekend only during the 6-9 A.M. morning
rush hour period.

       A recent study from San Francisco, CA evaluated trends in ambient benzene emissions
and air quality throughout the 1990's.8 The study noted substantial decreases in benzene
emissions and ambient concentrations. Unique to the study was the attribution of components of
these reduction to specific regulatory programs related to vehicles and fuels.  In particular,  the
study attributed a 1-year drop of 54% in benzene emission rates to a combination of the
introduction of California phase 2 RFG (attributed a 50% decrease) and fleet turnover (attributed
a 4% decrease). During the same year (1995-1996), a 42% reduction in the ambient
concentration of benzene was also observed. Fleet turnover effects were shown to be cumulative
over time. The  study indicates that in San Francisco both fuel and vehicle effects are important
a A "coefficient of variation" is a measure of variability, and for a set of data is defined as the standard deviation
over the mean.
                                           3-5

-------
Final Regulatory Impact Analysis

contributors to changes in emissions and ambient concentrations of benzene.

       New York State has a systematic program in place that has been measuring air toxics
since the 1990s.9 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
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 and moving average (KZ filter) on the concentration data.  The analysis
indicated that site-specific ambient concentration levels of benzene decreased by 50% or more
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.

       A similar downward trend was observed in California.  In California, the Air Resources
Board (ARB) maintains an Almanac of Emissions and  Air Quality.10 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
                                          3-6

-------
Final Regulatory Impact Analysis

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.
       Another recent evaluation of hazardous air pollutant (HAP) trends was conducted for
selected metropolitan areas.11  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 in reducing ambient concentrations of air toxics over
time. Additional reductions are expected with the implementation of additional regulatory
measures such as this one.  It should be noted that due to the limited spatial and temporal
coverage of air toxics monitoring networks, using ambient monitors to represent exposure adds
substantial uncertainty in exposure assessment.
                                           5-7

-------
Final Regulatory Impact Analysis
 Table 3.1-1. Site Descriptions of the Monitoring Stations Along with Mean Benzene Concentration from 1990-2003 and 2001-
                                     2003, for Monitoring Stations in New York State.

Site Character
Location Area
2000 Population
(thousands)
Annual Vehicle
Miles Traveled
(million miles)
Period 1990-2003
Mean Concentration
(ns/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

-------
Final Regulatory Impact Analysis
3.1.2   Population-Based (Representative) Exposure Measurements

       In addition to measurements of outdoor concentrations, an important component of
understanding human exposure to air toxics is the body of studies that employ survey techniques
to assess microenvironmental and 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-1990.12 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
preliminary results of the NHEXAS pilot study in Arizona,  another study area, indicate that
among the 179 statistically-sampled homes, median indoor concentrations were 1.3 ug/m3 during
the mid-1990's, while outdoor concentrations  were 1.0 ug/m3.13 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.14  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, were also obtained for the study
area.  Overall, median outdoor modeled concentrations of benzene and other fuel-related VOCs
corresponded well with measured data in the area (correlation coefficient of median VOC
concentrations = 0.97).  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 Relationships of 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.15  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
                                          3-9

-------
Final Regulatory Impact Analysis

for indoor air were 3.50 ug/m3 and 10.0 ug/m3, while outdoor statistics were 2.15 and 5.16
ug/m3.

       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,16 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 toward higher concentrations.
The 95th percentile concentration was 12.7 ug/m3.  Homes with attached garages had
significantly higher concentrations of benzene indoors (p < 0.0001). 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, personal
exposures to benzene and toluene were elevated for children living in a home with an attached
garage, but only the relationship for toluene was significant at the 95% confidence level.
                                          3-10

-------
Final Regulatory Impact Analysis




                  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
                                                          5-11

-------
Final Regulatory Impact Analysis
       In another study, students recruited from an inner-city school in Minneapolis, MN
participated in an exposure study called SHIELD.17 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.18

       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.19  Average winter
benzene personal concentrations were 4.70 ug/m3, while indoor and outdoor concentrations
averaged 5.97 and 2.55 ug/m3. Average indoor concentrations exceeding average personal
concentrations is unique to the TEACH winter results. 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. There was no substantial evidence for indoor sources of benzene.20 In a subsequent
publication, personal exposure concentrations for both cities were reported, averaged across both
seasons.  New York City average exposure concentrations were 3.82 ug/m3, while Los Angeles
average exposure concentrations were 4.64 ug/m3.21

       Overall, these studies show that personal and indoor concentrations  of benzene and other
VOCs are substantially higher than those found outdoors (see Table 3.1-3).  In general, these
differences are statistically significant. Some of the factors leading to these elevated
concentrations are likely 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.
                                          3-12

-------
Final Regulatory Impact Analysis
        Table 3.1-3. Personal Exposure to Benzene from Population-Based Studies"


Location
EPA Region
5
Baltimore,
MD

Elizabeth,
NJ,
Houston, TX,
Los Angeles
CA
Elizabeth,
NJ,
Houston, TX,
Los Angeles
CA
Minneapolis
St. Paul, MN

Minneapolis,
MN



New York,
NY


Los Angeles,
CA


Year(s)
1995-
1996
2000-
2001


1999-
2001


1999-
2001

1997


2000




1999-
2000


1999-
2000

Includes
Smokers
Yes
No


No


No

Yese


Yese




No



No
Personal
Average
(ug/m3)
7.52
4.06


3.64


4.16

4.8

2.1
Winter
1.5
Spring
4.7
Winter
3.1
Summer
3.8
Total
4.64
"Upper
Bound"
(Hg/m3)
13.71b
7.30C


10.7C,
27.4g


12.0C,
43.6s

9.1

6.5
Winter"
4.2
Spring
11.4
Winte/
7.0
Summer
12.3
Totaf
11.27
Indoor
Average
(Hg/m3)
7.21
3.70


3.50


N/Rh

3.9

2.2
Winter
2.1
Spring
6.0
Winter
1.8
Summer
3.6
Total
3.87
Outdoor
Average
(Hg/m3)
3.61
1.84


2.15


N/Rh

3.3

1.3
Winter
1.1
Spring
2.5
Winter
1.3
Summer
1.8
Total
3.32


Reference
Clayton et
al. (1999)
Payne-
Sturges et
al. (2004)

Weisel et
al. (2005)


Weisel et
al. (2005)

Adgate et
al. (2004a)

Adgate et
al. (2004b)


Kinney et
al. (2002);
Sax et al.
(2006)

Sax et al.
(2006)
   a Children's studies in italics
   b 90th percentile
   c 95th percentile
   d Mean +2 standard deviations
   e Smoking in homes
   f Maximum measured value
   g 99th percentile
   h Not reported
                                         3-13

-------
Final Regulatory Impact Analysis


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.22

         Table 3.1-4.  Descriptions and Particle Sizes of Each Category of Particles
Description
Supercoarse
Coarse (or Thoracic Coarse Mode)
Fine (or Accumulation Mode)
Ultrafine (or Nuclei Mode)a
Particle Size, dp (um)
dp>10
2.5 < dp < 10
0.1 
-------
Final Regulatory Impact Analysis


ultrafme mode particles dominate the number concentration in close proximity to the roadway,
while fine mode dominates farther from the road.  Particle size distributions, mass and elemental
carbon concentrations have been examined near roads in Los Angeles.25'26  Researchers observed
a four-fold increase in particle number concentrations, when comparing measurements 300 m
and 20 m from LA highways. Other studies have similarly shown that ultrafme mode particles
show a sharp decrease in particle number concentrations as the distance from major roadways
increases.27'28 Evidence was recently found of increased exposures to ultrafme particles near
roads when it was discovered that children living near major roads had elevated levels of
particle-containing alveolar macrophages.29  Additionally, roadside monitoring has shown that
particle number varies with vehicle type and vehicle operating conditions. For example, elevated
ultrafme mode particle concentrations have been identified when operating speeds on the road
increase as well as when the proportion of heavy-duty diesel vehicles increases.30'31

       An increase in coarse particles near roads could originate from engine deterioration,
brake and tire wear, and secondary aerosol formation.32'33'34'35 Engine deterioration is generally
a function of vehicle age and maintenance condition. Brake wear emissions are highly
dependent on brake pad materials.36 Secondary aerosol formation is dependent on fuel
composition, 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.37  Studies suggest that ambient temperature
variation can also affect particle number gradients near roads substantially.38 Wind direction
also affects traffic-related air pollution mass concentrations inside and outside of schools near
motorways.39'40 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
methods 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.

       Air quality monitoring is one way of evaluating pollutant concentrations at locations near
sources such as roadways. 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.41  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
                                          3-15

-------
Final Regulatory Impact Analysis
BTEX concentrations around homes within 200 m of roadways and gas stations are 1.5 to 4
times higher than urban background levels. In a subsequent study, proximity to major roadways,
meteorology, and photochemistry were all found to be significant determinants of ambient
concentration of a range of aldehyde species, including formaldehyde, acetaldehyde, acrolein,
and others. For most aldehydes, spring and summer concentrations were significantly higher
than those from colder seasons.42 However, formaldehyde concentrations were significantly
lower in summertime, suggesting greater photochemical destruction than production.  On colder
days, when photochemical activity was lower, concentrations of formaldehyde, acetaldehyde,
acrolein, and other aldehydes were significantly higher with increasing proximity to high-traffic
roads.

       Several other 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 43>44>45>46>47>48 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.49  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/m3.50
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 hardly any traffic (<
50 automobiles/day) at 1.3 ug/m3.51 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.

       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.52'53'54'55'56  Researchers have demonstrated exponential
reductions in concentrations of CO, as well as PM number, and black carbon (as measured by an
aethalometer), with increasing downwind distance from a freeway in Los Angeles.57'58 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 ofin-
                                          3-16

-------
Final Regulatory Impact Analysis
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).59 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.60  The study included
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. The final report of this study is expected to be released in 2007.

       In 1998, the California Air Resources Board published an extensive study of
concentrations of in-vehicle air toxics in Los Angeles and Sacramento, CA.61 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.62'63 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.64 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.
                                          3-17

-------
Final Regulatory Impact Analysis
   Table 3.1-5.  Mean Concentrations of Black Carbon (BC), Particle Bound PAH, NO2,
  Particle Count (PC), and PM2.5 in Three School Bus Commute Microenvironments and
                                    Background Air

BC fag/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
25"
Bus
Commutes21
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.65 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.66'67'68'69'70'71

       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.72'73'74  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.
                                          3-18

-------
Final Regulatory Impact Analysis
  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 concentrations of outdoor pollutants can therefore influence
indoor concentrations. A review of the literature determined that approximately 100% of
gaseous compounds,  such as benzene, and 80% of diesel PM can penetrate indoors.75'76

       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.77 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
                                          3-19

-------
Final Regulatory Impact Analysis
limited access). PAH concentrations also appear to increase with increase in annual average
daily traffic on nearest major collector.

       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 2001,78  Concentrations of
the traffic pollutants PMio, PM2 5, black carbon, total NOX, and NC>2 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 associated with roads with
heavy traffic within a relatively small geographic area.

       A study to assess children's exposure to traffic-related  air pollution while attending
schools near roadways was performed in the Netherlands.79  Investigators measured PM2.5, NO2
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 1995.80 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.

       In another Dutch study, researchers monitored children's personal exposure
concentrations, and home indoor and home  outdoor levels of "soot" (particle blackness), NO,
and NO2.81 Four-month average concentrations were calculated for each pollutant. Personal
exposure to "soot"  was 35-38% higher in students living within 75 meters of roads with 10,000
average annual  daily traffic, a statistically significant result.  Nonsignificant elevations in
personal exposure to NO, NO2, and NOx were also found.

       The TEACH study (Toxic Exposure Assessment - Columbia/Harvard) measured the
concentrations of VOCs, PM2.s, black carbon, and metals outside the homes of high school
                         Qn
students in New York City.   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, spatial and temporal patterns of
MTBE concentrations, used as  a tracer for motor vehicle pollution, were consistent with traffic
patterns.
                                          3-20

-------
Final Regulatory Impact Analysis
       Average benzene concentrations were determined in a recent evaluation of the exposure
of urban inhabitants to atmospheric benzene in Athens, Greece.83  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.84  Similarly, researchers found that traffic-related
pollutant exposure concentrations of car drivers were higher than for cyclists.85 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
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.86 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 higher than the fixed-site measurement (1.9  ug/m3) on
urban roadways 100-300 m from streets.

       The same researchers studied the exposure of commuters in Boston to VOCs during car
driving, subway travel, walking, and biking.87  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.
                                          3-21

-------
Final Regulatory Impact Analysis
A survey of CO concentration was conducted for various transport modes along heavy traffic
routes in Athens, Greece.88 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 PMi0, PM2.s, and PMi were made
during walking and in-car journeys on two suburban routes.89  In-car measurements were highest
(43.16, 15.54, and 7.03 ug/m3 for PMi0, 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.90 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.

3.1.3.3 Concentrations and Exposure 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.91 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.  A recent study from Fairbanks, Alaska used
perfluorocarbon tracer (PFT) gases to estimate that 12.2% of air entering a mechanically
ventilated energy efficient home and 47.4% of the air entering the living spaces of an older
passively ventilated home originated in the homes'  attached garages.92 In an Ann Arbor,
Michigan home, researchers used PFT gases to estimate that 16% of the air entering the home
entered through the garage.93  A recent study of a representative sample of homes in Anchorage,
Alaska employing PFT estimated that in homes with a forced air furnace in an attached garage,
36.7% of indoor air originated in the garage.94 In homes that had  forced air furnaces indoors or
hytronic heat, 17.0% and 18.4% of indoor air originated in the garage, respectively. A study
from Minnesota examined homes constructed in 1994, 1998, and 2000.95 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
                                          3-22

-------
Final Regulatory Impact Analysis


average of 13% of home air intake came from the garage.96 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.97  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.98

       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)
Study was completed in 1985.99 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 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. In the study, a large fraction of an average nonsmoker's benzene
exposure originated from sources in attached garages.100  Work done as part of the TEAM study
also identified stored gasoline as an important source  of elevated benzene levels indoors.101 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. Lawn  and garden equipment fuel tank
emissions, however, are significantly lower than evaporative emissions from gas cans, 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 outside the occupational setting.  The garage also
supplies contaminated air to the home to which it is attached.  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.102 In Alaska, where fuel benzene levels tend to be very high and homes may be
built very airtight, garage concentrations have been measured at even higher levels.  One study
from Anchorage measured average garage benzene concentrations of 103 ug/m3, with a standard
deviation of 135 ug/m3.103  More recently, a two-home study in Fairbanks found garage benzene
average concentrations  of 119 ug/m3 during summer and 189 ug/m3 during winter in one well-
ventilated home with an air-to-air heat exchanger.104  In an older home with passive ventilation
summer and winter garage benzene concentrations were 421 and 103 ug/m3, respectively.
                                          3-23

-------
Final Regulatory Impact Analysis
       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.105  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.

       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.106 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.107 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.108 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
                                               1 flQ
significantly higher in homes with attached garages.    Homes  with attached garages had
                                          3-24

-------
Final Regulatory Impact Analysis


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.110

       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.1U 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.

       Appendix 3 A presents an EPA analysis of the effect of attached garages on indoor air
under various scenarios.  This study was undertaken to evaluate the magnitude of exposure
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.  Because it is
unclear how well the  homes studied to  date represent the housing nationally, it is not currently
feasible to provide a highly precise estimate of the effect of attached garages on benzene
exposure nationally.  Depending on how the available data are summarized, overall modeled
exposure concentrations would be expected to increase between 1.2 and 6.6 ug/m3 above average
inhalation exposure concentrations to benzene from ambient sources (1.4 ug/m3, as discussed in
Section 3.2).  It should be noted that there is considerable uncertainty associated with this
estimated range,  as discussed in Appendix 3 A.
       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 Concentrations and Exposure 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 substantially higher
than found in outdoor air. Because of the limited amount of data, we include results from some
non-U.S. studies, although differences in fuels and control technology limited their applicability
to the U.S.

       In November  1990, a study of microenvironments, partially funded by the US EPA,
evaluated the potential range in concentrations of selected air toxics.112 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
                                          3-25

-------
Final Regulatory Impact Analysis
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.113
An interim report is available. 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 with traffic level and reached concentrations that
were significantly higher than ambient levels outside the garage.  The final report of the 21 l(b) is
expected in 2007.

       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.114 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 2000.115  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, PMio,  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
                                          3-26

-------
Final Regulatory Impact Analysis
ventilation and location, since the occupancy rates and fleet mixes were similar.  The authors
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 1999.116  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 showed 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.

       In Mumbai, India, ambient levels of benzene were determined during different seasons at
several different locations, including two parking areas.117 Parameters of the parking areas 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.118 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.119 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
                                          3-27

-------
Final Regulatory Impact Analysis
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 Concentrations and Exposure 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 TEAM study from the 1980's, described above, pumping gas and exposure to auto
exhaust were significantly associated with elevated 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 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.  Since this study, implementation of fuel controls,
onboard vapor recovery, 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.120 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
                                                                          101
with and without Stage II vapor recovery in Cincinnati, Phoenix, and Los Angeles.   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.122 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
                                               1 Ol
determined inside automobile cabins during fueling.   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
                                          3-28

-------
Final Regulatory Impact Analysis
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.

       In a study conducted between summer 1998 and spring 1999, self-service gas station
customers took part in a study to measure personal and breath concentrations of benzene at gas
stations in New Jersey.124 Benzene exposure concentrations during refueling (with a median
duration of three minutes) averaged 2.9 mg/m3 (SD = 5.8 mg/m3).  Breath concentrations
averaged 160 ug/m3 (SD = 260 ug/m3).  Breath benzene concentrations were significantly
correlated with refueling exposure concentrations, which was itself significantly associated with
refueling duration, time of year, and fuel octane grade.

       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.60 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.

3.1.3.6 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.125 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 fuel benzene
reductions to reduce their exposures as well.  This statement is echoed by researchers in the
occupational literature.126 Occupational exposures in this range have been associated with
increased risk of certain leukemias in occupational epidemiology studies (Section 1.3.1).

       Handheld and non-handheld equipment operators may also be exposed to elevated
concentrations of air toxics. As discussed below, 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.
                                          3-29

-------
Final Regulatory Impact Analysis
EPA recently conducted a study of occupational exposures among lawn and garden workers
using riding tractors, walk-behind lawn mowers, string trimmers, and chainsaws.127 Results
demonstrated that equipment operators can experience highly variable exposures, with short-
term personal concentrations of CO and PM2.s ranging over two orders of magnitude. The study
also reported operator breathing-zone concentrations of formaldehyde and acetaldehyde that
were higher than background levels in all tests. 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.128  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, PMi0, and CO.129
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 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. 13° 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 one recently-published study of diesel exhaust exposures in a representative sample of
trucking terminals nationally, investigators applied structural equation modeling to data on
personal exposure to diesel exhaust (as elemental carbon).131 The study found that worker
exposure to elemental carbon depended on work area concentrations and worker tobacco use.
Work area concentrations depended on the size and type of the trucking terminal, whether the
work site was a mechanical shop, work site ventilation, and terminal yard concentrations.
Terminal yard concentrations in turn were related to local meteorology, the proximity of
interstate highways, surrounding industrial land uses, and region of the country. This study is
valuable in showing how personal occupational exposures are a complicated function of many
factors. Sophisticated statistical methods are needed to properly estimate models with highly
complex covariance structures.
                                          3-30

-------
Final Regulatory Impact Analysis
       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.

3.1.4   Uncertainties in Air Toxics Measurements

       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 and exposure 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
or in personal breathing zones. Air quality monitoring at these central sites often do not
represent actual exposures, especially for populations living near roads or with substantial
occupational exposure.

       Monitoring samples are often integrated and therefore lack time resolution. This can
result in difficulty in determining source contributions. Additionally, some compounds are hard
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.132 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.

       Personal exposure monitoring provides greater realism in describing a person's actual
exposure to air toxics.  However, given the limitations on size of equipment, detection limits in
personal  exposure monitoring  studies are sometimes greater than those found in studies using
other techniques.

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 (NAT A).  It assesses lifetime risks assuming
continuous  exposure to levels of air toxics estimated for a particular point in time. The most
recent NATA was done for the year 1999.133 It had four steps:

       1) Compiled 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.
                                          3-31

-------
Final Regulatory Impact Analysis
       2) Estimated ambient concentrations based on emissions as input to an air dispersion
       model (the Assessment System for Population Exposure Nationwide, or ASPEN
       model).134
       3) Estimated population exposures based on a screening-level inhalation exposure model
       (Hazardous Air Pollutant Exposure Model, version 5, or HAPEM5) and the estimated
       ambient concentrations (from the ASPEN model) as input to the exposure model.135
       4) Characterized 1999 potential public health risks due to inhalation of air toxics. This
       included cancer and noncancer effects, using available information on air toxics health
       effects, current EPA risk assessment and risk characterization guidelines, and estimated
       population exposures.136

       For this final rule, we have conducted air quality, exposure and risk modeling for the
years 1999, 2015, 2020, and 2030, using the same general approach as the 1999 NAT A.  We
modeled all the pollutants in Table 2.2-1 for both the reference case, which includes all control
programs currently planned by EPA in regulations, and the control case, which includes the
cumulative impacts of the standards proposed in this rule. 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. EPA has
previously done future-year projections  of the mobile source contribution to air toxics
concentrations, exposure, and risk for selected air toxics,137'138'139' 140but prior to the proposal
for this rule, had never done a comprehensive assessment that includes projections for all mobile
source air toxics, as well as the stationary source contribution for those pollutants. It should be
noted that the reference case assessment results  developed for the proposal have been published
in a peer reviewed journal article.141

       As discussed in Chapter 2, a number of major revisions to inventory methodology have
been made relative to what was done for both the 1999 NATA, and air quality exposure and risk
modeling for the proposal.  These include revisions to cold start emissions, use of NMIM2005
for nonroad equipment, addition of portable fuel container emissions, and changes to gasoline
distribution inventories.  Also, this final rule modeling for 1999 does not include data submitted
by States for the 1999 NEI. In addition, the modeling for the final rule relied on an updated
version of the HAPEM model, HAPEM6.142  HAPEM6 improves on HAPEM5 by accounting
for the spatial variability of outdoor concentrations of air toxics within a census tract due to
higher outdoor concentrations at locations near major roadways. Other improvements to
HAPEM are discussed in section  3.2.1.2.1. This modeling work is discussed in more detail in an
EPA technical report, "National Scale Modeling of Air Toxics for the Final Mobile Source Air
Toxics Rule; Technical Support Document," Report Number EPA-454/R-07-002.  It should be
noted that the control case modeling accounted only for the 0.62 percent standard, but not the 1.3
vol% maximum average. Thus, the emission reductions from highway vehicles and other
sources attributable to the fuel benzene standard are underestimated in many areas of the
country, particularly in areas where fuel benzene levels were highest without control, such as the
Northwest.
                                          3-32

-------
Final Regulatory Impact Analysis
       The NATA 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
ambient "hot spots," as mobile sources are not represented explicitly as roads or other locations
of mobile source activity. 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 potentially significant
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 and some reactive
compounds. The cancer unit risk estimates for most 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.

       Although we do not use it in this modeling, another tool that EPA uses to assess
distributions of concentrations of air toxics at the national scale is the Community Multiscale Air
Quality (CMAQ) modeling system.143 CMAQ can account for photochemical destruction and
production, deposition and regional transport of toxic air pollutants, and thus can be used to
predict the concentrations of HAPs with significant atmospheric production. 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.144 CMAQ underpredicts
monitored benzene levels more than ASPEN, because ASPEN values contain a large, added-on
concentration based on monitored values of benzene. 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 concentration
gradients, such as benzene.145 Finally, CMAQ is requires more computational resources, 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 other potentially significant sources of inhalation exposure,
such as benzene emissions from sources in attached garages (such as vehicles, snowblowers,
lawnmowers and gas cans).
                                          3-33

-------
Final Regulatory Impact Analysis


3.2.1.1 Air Quality Modeling

3.2.1.1.1     Methods

       Prior to performing air quality modeling of 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
input files used by ASPEN to calculate air quality concentrations.146 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 are processed, they are 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 kilometers away. The model then interpolates concentrations
to census tract centroids.  For 1999, meteorological conditions in 1999 and 2000 census tract
data were used.

       In using ASPEN to estimate projected concentrations in 2015, 2020, and 2030 for this
final rule, 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 Final 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.147 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.
Thus, for the proposal, a sensitivity analysis was done to evaluate the potential impact of not
changing the background concentration (see Section 3.2.1.4).

       It should be noted that in the control case scenarios, we have modeled the cumulative
impacts on air quality, exposure, and risk for all of the programs finalized today, not the impacts
of individual programs. Were we to model each program individually, we anticipate that
changes in air quality, exposure, and risk would track the patterns of emission changes closely.

       Also, for the final rule, we estimated the contribution of secondary formation to ambient
concentrations of MSATs by applying ratios of secondary to primary concentrations from 1999
NATA to the modeled primary concentrations for this rule. This is different from the approach
                                          3-34

-------
Final Regulatory Impact Analysis
used in the proposal where we projected precursor emissions and then modeled secondary
formation. When we applied the ratio approach to the proposal's primary concentrations, the
results were very similar to the full modeling approach (see Section 3.2.1.3). The comparisons
are discussed in the technical document cited above.

       We estimated the contributions to ambient concentrations for the following source
sectors: major, area and other, onroad, nonroad, and background.13

3.2.1.1.2       Air Quality Trends for Air Toxics: Reference Case

       Table 3.2-1 summarizes nationwide mean census tract ambient concentrations, without
the controls being finalized in this rule,  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 over 40% by 2015, with a decrease in ambient
benzene concentration from all sources  of about 25%.  Subsequently, increases in vehicle miles
traveled (VMT) are projected to produce increasing concentrations.  Summary tables providing
data by State, and for reformulated and non-reformulated (i.e., conventional) 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 non-reformulated gasoline areas - about 1.9 |ig/m3 versus 1.2 |ig/m3.  However the
percent reduction in average 2015 ambient concentration is similar regardless of fuel type - 22%
for non-reformulated gasoline counties versus 29% for reformulated gasoline counties.
b 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.


                                             3-35

-------
Final Regulatory Impact Analysis


 Table 3.2-1.  Mean Ambient Concentrations of Mobile Source Air Toxics in 1999, 2015, 2020, and 2030, Without Controls in
                                                       this Rule.

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
(MS m 3)
5.10E-02
O.OOE+00
5.17E-01
O.OOE+00
3.94E-01
O.OOE+00
O.OOE+00
O.OOE+00
7.62E-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
1.70E-01
1999 average concentrations (jig in 3)
major
1.97E-03
2.16E-02
2.94E-02
3.21E-03
2.20E-02
8.22E-04
1.07E-04
1.84E-02
3.99E-02
6.68E-02
1.30E-02
2.71E-03
4.56E-03
7.76E-04
4.93 E-03
1.01E-02
2.52E-02
2.03E-01
9.98E-02
area &
other
2.05E-02
2.32E-02
5.49E-02
2.93E-02
1.40E-01
4.53E-04
1.98E-04
9.00E-02
8.77E-02
4.30E-01
6.04E-02
2.22E-03
4.11E-02
1.42E-03
1.61E-02
2.33E-02
1.40E-02
8.05E-01
5.59E-01
onroad
5.20E-02
7.29E-01
6.78E-01
5.63E-02
6.89E-01
3.22E-05
2.15E-05
2.73E-01
4.65E-01
2.34E-01
4.00E-01
1.73E-05
1.46E-02
3.96E-05
1.73 E-03
1.68E-01
2.98E-02
1.81E+00
1.01E+00
nonroad
1.81E-02
1.96E-01
1.47E-01
2.27E-02
1.77E-01
5.53E-05
1.25E-05
9.73E-02
2.21E-01
8.56E-02
4.04E-01
5.46E-06
4.36E-03
9.98E-05
8.60E-04
4.27E-02
3.65E-03
4.18E-01
3.99E-01
total
(including
background)
1.44E-01
9.70E-01
1.43E+00
1.11E-01
1.42E+00
1.36E-03
3.39E-04
4.79E-01
1.58E+00
8.17E-01
8.77E-01
4.95E-03
6.46E-02
2.33E-03
2.37E-02
2.45E-01
7.27E-02
3.24E+00
2.23E+00
2015 annual average concentrations (jig m"3)
major
2.17E-03
1.09E-02
2.97E-02
3.53E-03
1.55E-02
1.04E-03
1.36E-04
1.24E-02
4.98E-02
5.94E-02
1.38E-02
3.23 E-03
3.97E-03
8.87E-04
3.79E-03
9.31E-03
3.00E-02
1.43E-01
8.22E-02
area &
other
2.05E-02
2.69E-02
5.71E-02
2.62E-02
1.63E-01
6.16E-04
2.72E-04
1.19E-01
9.82E-02
5.21E-01
6.52E-02
2.92E-03
5.01E-02
1.62E-03
1.86E-02
2.39E-02
1.89E-02
1.06E+00
7.60E-01
onroad
2.28E-02
3.66E-01
3.86E-01
2.42E-02
3.79E-01
4.40E-05
2.94E-05
1.35E-01
1.92E-01
1.16E-01
1.05E-01
2.36E-05
7.90E-03
5.43E-05
9.13E-04
8.24E-02
1.50E-02
9.00E-01
4.98E-01
nonroad
1.08E-02
1.15E-01
1.10E-01
1.81E-02
1.14E-01
5.85E-05
1.32E-05
5.66E-02
1.63E-01
5.93E-02
1.08E-01
6.46E-06
4.49E-03
1.15E-04
7.66E-04
2.83E-02
2.18E-03
2.50E-01
2.18E-01
total
(including
background)
1.07E-01
5.19E-01
1.10E+00
7.20E-02
1.07E+00
1.76E-03
4.50E-04
3.24E-01
1.27E+00
7.56E-01
2.93 E-01
6.17E-03
6.65E-02
2.67E-03
2.40E-02
1.44E-01
6.61E-02
2.35E+00
1.73E+00

-------
Final Regulatory Impact Analysis


   Table 3.2-1 (cont'd). Mean Ambient Concentrations of Mobile Source Air Toxics in 1999, 2015, 2020, and 2030, Without
                                                  Controls in this Rule.

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
(MS m 3)
5.10E-02
O.OOE+00
5.17E-01
O.OOE+00
3.94E-01
O.OOE+00
O.OOE+00
O.OOE+00
7.62E-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
1.70E-01
2020 annual average concentrations (jig m 3)
major
2.34E-03
1.17E-02
3.10E-02
3.96E-03
1.70E-02
1.17E-03
1.54E-04
1.39E-02
5.65E-02
6.53E-02
1.55E-02
3.59E-03
4.46E-03
9.61E-04
4.21E-03
9.35E-03
3.44E-02
1.60E-01
9.29E-02
area &
other
2.05E-02
2.84E-02
5.83E-02
2.54E-02
1.69E-01
6.96E-04
3.07E-04
1.31E-01
1.03E-01
5.62E-01
6.67E-02
3.21E-03
5.32E-02
1.78E-03
1.90E-02
2.45E-02
2.09E-02
1.16E+00
8.38E-01
onroad
2.37E-02
3.66E-01
3.98E-01
2.50E-02
3.88E-01
4.84E-05
3.23E-05
1.35E-01
1.97E-01
1.07E-01
8.48E-02
2.60E-05
7.86E-03
5.97E-05
9.47E-04
8.45E-02
1.57E-02
9.11E-01
5.04E-01
nonroad
1.14E-02
1.14E-01
1.09E-01
1.91E-02
1.18E-01
5.90E-05
1.34E-05
5.78E-02
1.64E-01
6.13E-02
1.12E-01
6.83E-06
4.80E-03
1.20E-04
7.71E-04
2.78E-02
2.21E-03
2.50E-01
2.18E-01
total
(including
background)
1.09E-01
5.20E-01
1.11E+00
7.34E-02
1.09E+00
1.97E-03
5.07E-04
3.38E-01
1.28E+00
7.96E-01
2.79E-01
6.83E-03
7.03E-02
2.92E-03
2.49E-02
1.46E-01
7.32E-02
2.48E+00
1.82E+00
2030 annual average concentrations (jig m"3)
major
2.34E-03
1.17E-02
3.10E-02
3.96E-03
1.70E-02
1.17E-03
1.54E-04
1.39E-02
5.65E-02
6.53E-02
1.55E-02
3.59E-03
4.46E-03
9.61E-04
4.21E-03
9.35E-03
3.44E-02
1.60E-01
9.29E-02
area &
other
2.05E-02
2.84E-02
5.83E-02
2.54E-02
1.69E-01
6.96E-04
3.07E-04
1.31E-01
1.03E-01
5.62E-01
6.67E-02
3.21E-03
5.32E-02
1.78E-03
1.90E-02
2.45E-02
2.09E-02
1.16E+00
8.38E-01
onroad
2.78E-02
4.24E-01
4.69E-01
2.94E-02
4.54E-01
5.94E-05
3.96E-05
1.57E-01
2.31E-01
1.18E-01
8.42E-02
3.19E-05
9.11E-03
7.34E-05
1.12E-03
9.84E-02
1.85E-02
1.06E+00
5.86E-01
nonroad
1.30E-02
1.24E-01
1.18E-01
2.18E-02
1.32E-01
6.04E-05
1.37E-05
6.45E-02
1.80E-01
6.87E-02
1.25E-01
7.59E-06
5.51E-03
1.31E-04
8.57E-04
2.99E-02
2.47E-03
2.75E-01
2.40E-01
total
(including
background)
1.15E-01
5.88E-01
1.19E+00
8.05E-02
1.17E+00
1.98E-03
5.15E-04
3.66E-01
1.33E+00
8.14E-01
2.92E-01
6.84E-03
7.23E-02
2.95E-03
2.52E-02
1.62E-01
7.63E-02
2.65E+00
1.93E+00

-------
Final Regulatory Impact Analysis
Figure 3.2-1. Nationwide Average Benzene Concentration, 1999-2030, Without Controls in
                                         this Rule.

    1.60
    1.40
    1.20
    1.00
    0.80
    0.60
    0.40
    0.20
    0.00
               1999
                                2015
                                                  2020
                                                                   2030
                                         Year
3.2.1.1.3      Distributions of Air Toxic Concentrations across the U. S.: Reference Case

       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, absent the
controls being finalized in this rule. 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 percentiles 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
                                           3-38

-------
Final Regulatory Impact Analysis


Table 3.2-2.  National Distribution of Census Tract Concentrations for Mobile Source Air
Toxics in 2020, Without Controls in this Rule.

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
3.03E-03
3.83E-02
5.45E-01
6.04E-03
3.42E-01
5.73E-06
3.52E-06
2.04E-02
4.08E-01
3.27E-02
3.34E-03
1.33E-05
2.88E-03
1.38E-05
1.72E-03
1.24E-02
2.52E-03
1.54E-01
2.66E-01
10th
percentile
5.60E-03
7.00E-02
5.82E-01
9.78E-03
4.15E-01
1.52E-05
8.79E-06
3.79E-02
5.29E-01
6.16E-02
7.88E-03
4.35E-05
5.91E-03
3.80E-05
2.94E-03
2.13E-02
4.88E-03
2.83E-01
3.43E-01
25th
percentile
3.12E-02
1.74E-01
6.99E-01
2.09E-02
6.33E-01
6.40E-05
3.56E-05
1.01E-01
8.08E-01
1.90E-01
2.39E-02
2.04E-04
1.86E-02
1.67E-04
5.73E-03
4.81E-02
1.23E-02
7.34E-01
6.35E-01
Median
8.36E-02
3.79E-01
9.41E-01
4.41E-02
9.37E-01
2.41E-04
1.22E-04
2.30E-01
1.16E+00
4.76E-01
7.22E-02
8.68E-04
4.48E-02
6.65E-04
1.19E-02
1.07E-01
2.70E-02
1.64E+00
1.22E+00
75th
percentile
1.30E-01
6.80E-01
1.29E+00
8.64E-02
1.32E+00
7.31E-04
3.32E-04
4.06E-01
1.52E+00
8.93E-01
2.44E-01
3.53E-03
8.82E-02
2.01E-03
2.08E-02
1.93E-01
5.39E-02
2.96E+00
2.06E+00
90th
percentile
1.98E-01
1.12E+00
1.84E+00
1.71E-01
1.90E+00
2.34E-03
9.08E-04
6.70E-01
2.12E+00
1.70E+00
8.80E-01
1.42E-02
1.67E-01
4.78E-03
3.62E-02
3.26E-01
1.06E-01
5.31E+00
3.61E+00
95th
percentile
3.28E-01
1.50E+00
2.49E+00
2.71E-01
2.36E+00
4.89E-03
1.55E-03
9.60E-01
2.67E+00
2.81E+00
1.30E+00
2.10E-02
2.37E-01
8.17E-03
5.78E-02
4.33E-01
1.75E-01
7.43E+00
5.38E+00
indicate higher than average fuel benzene levels in these areas.  These areas also have higher
benzene emissions in winter due to cold starts. 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.

       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 Controls on Ambient Concentrations
       The standards being finalized in this rule will substantially reduce ambient concentrations
of air toxics across the United States.  As noted above, these results reflect the cumulative effects
of all of the programs finalized in today's rule, not the individual programs. Table 3.2-3 shows
the reduction in nationwide average census tract concentrations of MSATs from all sources in
2015, 2020 and 2030.  Table 3.2-4 shows the reduction in the highway vehicle contribution to
nationwide average census tract concentrations of MSATs. Table 3.2-5 shows that in 2030, the
highway vehicle portion of ambient benzene concentrations will be reduced almost 45% across
the U.S., the nonroad equipment contribution will be reduced about 10%, and
                                          3-39

-------
Final Regulatory Impact Analysis
    Figure 3.2-2. Geographic Distribution of County Median Concentrations (ug/m ) of
                       Benzene in 2020 Without Controls in this Rule.
                                                                               0.063 - 0.338

                                                                               0.339 - 0.567

                                                                               0.568 - 0.880

                                                                               0.881 -1.316

                                                                               1.317-2.114

                                                                               2.115-4.929

-------
Final Regulatory Impact Analysis
  Table 3.2-3. Nationwide Average Census Tract Concentrations of MSATs, With and Without Controls in this Rule, 2015,
                                                   2020, and 2030.
2015 2020 2030
Reference Control % Reduction Reference Control % Reduction Reference Control % Reduction
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
1.07E-01
5.19E-01
1.10E+00
7.20E-02
1.07E+00
1.76E-03
4.50E-04
3.24E-01
1.27E+00
7.56E-01
2.93E-01
6.17E-03
6.65E-02
2.67E-03
2.40E-02
1.44E-01
6.61E-02
2.35E+00
1.73E+00
1.03E-01
4.53E-01
1.04E+00
6.79E-02
9.56E-01
1.76E-03
4.50E-04
2.99E-01
1.24E+00
7.37E-01
2.82E-01
6.17E-03
6.65E-02
2.67E-03
2.40E-02
1.33E-01
6.33E-02
2.18E+00
1.64E+00
3.6
12.7
5.8
5.7
10.3
0.0
0.0
7.5
2.3
2.5
3.5
0.0
0.0
0.0
0.0
7.8
4.3
7.1
5.3
1.09E-01
5.20E-01
1.11E+00
7.34E-02
1.09E+00
1.97E-03
5.07E-04
3.38E-01
1.28E+00
7.96E-01
2.79E-01
6.83E-03
7.03E-02
2.92E-03
2.49E-02
1.46E-01
7.32E-02
2.48E+00
1.82E+00
1.03E-01
4.19E-01
1.01E+00
6.69E-02
9.38E-01
1.97E-03
5.07E-04
3.01E-01
1.24E+00
7.70E-01
2.66E-01
6.83E-03
7.03E-02
2.92E-03
2.49E-02
1.28E-01
6.87E-02
2.22E+00
1.68E+00
5.7
19.5
9.1
8.9
13.6
0.0
0.0
11.1
3.6
3.2
4.6
0.0
0.0
0.0
0.0
12.2
6.2
10.4
7.8
1.15E-01
5.88E-01
1.19E+00
8.05E-02
1.17E+00
1.98E-03
5.15E-04
3.66E-01
1.33E+00
8.14E-01
2.92E-01
6.84E-03
7.23 E-02
2.95E-03
2.52E-02
1.62E-01
7.63E-02
2.65E+00
1.93E+00
1.04E-01
4.26E-01
1.03E+00
6.97E-02
9.50E-01
1.98E-03
5.15E-04
3.07E-01
1.26E+00
7.76E-01
2.74E-01
6.84E-03
7.23E-02
2.95E-03
2.52E-02
1.33E-01
6.89E-02
2.24E+00
1.70E+00
9.0
27.6
13.7
13.4
18.5
0.0
0.0
16.3
5.6
4.7
6.0
0.0
0.0
0.0
0.0
18.0
9.7
15.7
11.8
                                                         5-41

-------
Final Regulatory Impact Analysis
    Table 3.2-4.  Nationwide Highway Vehicle Contribution to Average Census Tract Concentrations of MSATs, With and
                                  Without Controls in this Rule, 2015, 2020, and 2030.
2015 2020 2030
% %
Reference Control Reduction Reference Control % Reduction Reference Control Reduction
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
2.28E-02
3.66E-01
3.86E-01
2.42E-02
3.79E-01
4.40E-05
2.94E-05
1.35E-01
1.92E-01
1.16E-01
1.05E-01
2.36E-05
7.90E-03
5.43E-05
9.13E-04
8.24E-02
1.50E-02
9.00E-01
4.98E-01
1.89E-02
3.06E-01
3.22E-01
2.01E-02
2.83E-01
4.40E-05
2.94E-05
1.14E-01
1.62E-01
1.05E-01
1.01E-01
2.36E-05
7.90E-03
5.43E-05
9.13E-04
7.12E-02
1.22E-02
7.47E-01
4.14E-01
17.0
16.3
16.5
17.0
25.3
0.0
0.0
16.0
15.3
9.8
4.3
0.0
0.0
0.0
0.0
13.6
18.8
17.1
16.9
2.37E-02
3.66E-01
3.98E-01
2.50E-02
3.88E-01
4.84E-05
3.23 E-05
1.35E-01
1.97E-01
1.07E-01
8.48E-02
2.60E-05
7.86E-03
5.97E-05
9.47E-04
8.45E-02
1.57E-02
9.11E-01
5.04E-01
1.74E-02
2.71E-01
2.97E-01
1.85E-02
2.55E-01
4.84E-05
3.23E-05
1.01E-01
1.50E-01
8.89E-02
7.77E-02
2.60E-05
7.86E-03
5.97E-05
9.47E-04
6.66E-02
1.12E-02
6.66E-01
3.69E-01
26.3
25.9
25.4
26.2
34.2
0.0
0.0
25.6
23.6
16.9
8.3
0.0
0.0
0.0
0.0
21.1
28.8
26.9
26.7
2.78E-02
4.24E-01
4.69E-01
2.94E-02
4.54E-01
5.94E-05
3.96E-05
1.57E-01
2.31E-01
1.18E-01
8.42E-02
3.19E-05
9.11E-03
7.34E-05
1.12E-03
9.84E-02
1.85E-02
1.06E+00
5.86E-01
1.75E-02
2.70E-01
3.06E-01
1.87E-02
2.54E-01
5.94E-05
3.96E-05
l.OOE-01
1.56E-01
8.84E-02
7.30E-02
3.19E-05
9.11E-03
7.34E-05
1.12E-03
6.92E-02
1.11E-02
6.62E-01
3.67E-01
37.0
36.4
34.8
36.6
44.0
0.0
0.0
36.1
32.4
25.0
13.4
0.0
0.0
0.0
0.0
29.6
39.8
37.7
37.5
                                                        5-42

-------
Final Regulatory Impact Analysis
   Table 3.2-5. Contributions of Source Sectors to Nationwide Average Census Tract Concentrations of Benzene, With and
                                  Without Controls in this Rule, 2015, 2020, and 2030.


Reference
Control
% Difference

2015 annual average concentrations I
major
1.55E-02
1.54E-02
0

area &
other
1.63E-01
1.61E-01
-1

highway
vehicles
3.79E-01
2.83E-01
-25

nonroad
1.14E-01
1.02E-01
-10

fig m'3)
total
(including
background)
1 .07E+00
9.56E-01
-10

2020 annual average concentrations (fig m"3)
major
1 .70E-02
1 .69E-02
0

Average Nationwide Difference in Ambient Benzene Concentration - Non RFC Areas

Reference
Control
% Difference


1.08E-02
1.08E-02
0


1.43E-01
1.41E-01
-2


2.96E-01
2.17E-01
-27


8.15E-02
6.82E-02
-16


8.93E-01
7.99E-01
-11


1 .20E-02
1 .20E-02
0

Average Nationwide Difference in Ambient Benzene Concentration - RFC Areas

Reference
Control
% Difference

2.39E-02
2.38E-02
0

1.99E-01
1.97E-01
-1

5.29E-01
4.02E-01
-24

1.72E-01
1.63E-01
-5

1 .38E+00
1 .24E+00
-10

2.58E-02
2.58E-02
0
area & other
1.69E-01
1.67E-01
-1



1.48E-01
1.46E-01
-2



2.08E-01
2.05E-01
-1
highway
vehicles
3.88E-01
2.55E-01
-34



3.06E-01
2.00E-01
-35



5.34E-01
3.54E-01
-34
nonroad
1.18E-01
1.05E-01
-10



8.34E-02
6.95E-02
-17



1.79E-01
1.70E-01
-5
total
(including
background)
1 .09E+00
9.38E-01
-14



9.11E-01
7.89E-01
-13



1 .40E+00
1.21E+00
-14
2030 annual average concentrations (fig m'3)
major
1 .70E-02
1 .69E-02
0



1 .20E-02
1 .20E-02
0



2.58E-02
2.58E-02
0
area&
other
1.69E-01
1.67E-01
-1



1.48E-01
1.46E-01
-2



2.08E-01
2.05E-01
-1
highway
vehicles
4.54E-01
2.54E-01
-44



3.57E-01
1.97E-01
-45



6.29E-01
3.57E-01
-43
nonroad
1.32E-01
1.18E-01
-10



9.29E-02
7.72E-02
-17



2.03E-01
1.92E-01
-5
total
(including
background)
1.17E+00
9.50E-01
-19



9.71 E-01
7.94E-01
-18



1 .52E+00
1 .23E+00
-19
                                                         5-43

-------
Final Regulatory Impact Analysis


the area source contribution will be reduced about 1 to 2%. The reduction for area sources is due
to the impacts of fuel benzene control on gasoline distribution emissions, and reductions in
portable fuel container (PFC) emissions from PFC and fuel benzene controls. Reductions in
non-reformulated gasoline areas are even larger.  It should be noted that the estimated total
reductions in ambient concentrations from all sources 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.  Figure 3.2-3 presents the distribution of percent
reductions in median ambient benzene concentrations for U.S. counties with the controls being
finalized in 2030. Again, since the 1.3% maximum average fuel benzene standard is not
included in the modeling, reductions in some parts of the country, including the Pacific
Northwest, are underestimated.  Summary tables providing data by State, as well as maps of
MSAT concentrations with controls and percent reductions with controls, can be found in the
docket for the rule.
       Figure 3.2-3. Distribution of Percent Reductions in Median Ambient Benzene
          Concentrations, 2030, for U. S. Counties with the Controls in this Rule.
                                                                      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 HAPEM6 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.148'149' 15°  HAPEM6 is designed to assess average
                                           3-44

-------
Final Regulatory Impact Analysis
long-term inhalation exposures of the general population, or a specific sub-population, over
spatial scales ranging from urban to national. HAPEM6 uses the general approach of tracking
representatives of 6 specified age groups as they move among indoor and outdoor
microenvironments and among geographic locations (a total of 14, HAPEM5 had 37). 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.
HAPEM calculates 30 replicates with different exposures for each demographic group. These
data can be used to develop a distribution of exposures for the entire U. S. population.

      HAPEM6 uses five primary sources of information: year 2000 population data from the
U.S. Census, population activity data, air quality data, roadway locations, and
microenvironmental data.  The population data used are obtained from the U.S. 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).151 The commuting data
contained in the HAPEM6 default file were derived from the year 2000 U.S. Census, and
includes the number of residents of each tract that work in that tract and every other U.S. Census
tract, as well as data on  commuting times and distances.  The air quality data come from ASPEN
(after  background has been added). The road locations are determined from geographic
information system files from the U.S. Census. 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.152

      New to HAPEM6 are algorithms which account for the gradient in concentrations of
primary  (directly emitted) mobile source air toxics within 200 meters of major roadways. 153
HAPEM6 adjusts ambient concentrations generated by ASPEN for each census tract using
concentration gradients  developed with the CALPUFF dispersion model.154 For locations within
75 meters and from 75 to 200 meters from major roads, ambient concentrations are adjusted
upward, while locations further from major roadways are adjusted downward. These
adjustments are consistent with results from prior modeling studies that explicitly accounted for
concentration gradients  around major roads within census tracts.155 These adjusted
concentrations are then  employed in microenvironmental concentration calculations.

      HAPEM6 has a  number of other technical improvements over the previous version of
HAPEM. These improvements, along with other details of the model, are described  in the
HAPEM6 User's Guide.156 In short, HAPEM6 reduces the number of demographic groups to 6
age-based groups from  10 age-gender groups in HAPEM5, and reduces the  number of
microenvironments modeled, from 37 to  14. This reduces modeling run time significantly with
little impact on results.  HAPEM6 also accounts for commuting time better, basing commute
times  and travel modes  for each census tract on distributions reported in the 2000 Census. The
HAPEM runs used year 2000 census 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, nonroad, and background.
                                         3-45

-------
Final Regulatory Impact Analysis
       Once HAPEM runs were completed, cancer risk and noncancer risk were calculated for
each of the mobile source air toxic pollutants, based on population exposure distributions. In the
HAPEM6 output, for each source category, there are 30 replicate exposure concentrations for
each of the six demographic groups (180 concentrations per census tract for each source
category). For each source category and each of the 30 replicates, a lifetime exposure
concentration was calculated.  A risk estimate was then calculated for each of the 30 replicates.
The resulting data were used to develop distributions of population risks at various summary
levels (census tract, county, state, national).  More detail is provided in the technical support
document.  Table 3.2-6 lists the pollutants with their respective unit risk estimates (UREs) for
cancer calculations and reference concentrations (RfCs) for noncancer calculations.  These are
the same values used in the 1999 NAT A, 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 noncancer calculations.

    Table 3.2-6.  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
POM1
POM2
POMS
POM4
POMS
POM6
POM7
POMS
Styrene
Toluene
Xylenes

Carcinogen
Class
A
N/A
B2

A
N/A
A

B



C
A
B2
B2
B2
B2
B2
B2
B2
B2




URE
(per jig/m3)
S.OxKT5
N/A
2.2xl(r6
0
7.8xl(r6*
N/A
1.2xl(r2
0
5.5xl(r9
N/A
N/A
N/A
3.4xlO'5
1.6xl(r4
5.5X10'5
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/
OAQPS
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








Neurological
Respiratory,
Neurological
Neurological

RfC(mg/
m3)
2.0xl(r3
N/A
9.0xl(r3
2.0xl(r5
S.OxKT2
N/A
l.OxKT4
1.0
9.8xl(r3
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
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
                                          3-46

-------
Final Regulatory Impact Analysis
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 noncancer effects - used as RfC.
       3)  California Office of Environmental Health Hazard Assessment (OEHHA) 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.-6 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
       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 3.2.-6 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 NATA and in Appendix H of the 2001 EPA draft
       report to the Science Advisory Board on the 1996 National-Scale Assessment.157

       Individual 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
                                          3-47

-------
Final Regulatory Impact Analysis
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 NAT A, 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. An HI greater than one can  be best described as indicating that a potential
may exist for adverse health effects.

3.2.1.2.2      Exposure and Risk Trends for Air Toxics: Reference Case

       Tables 3.2-7 and 3.2-8 summarize nationwide averages of median and 90th percentile
census tract exposure concentrations of mobile source air toxics in 1999, 2015, 2020, and 2030,
without the controls being finalized in this rule.  It should be noted that all the other non-
inventoried  sources, as well as the contribution from transport, contribute to background levels.
Overall, exposure to ambient concentrations tends to be less than ambient concentrations because
penetration rates to indoor microenvironments are typically less than one.c  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.
c In the exposure monitoring studies discussed in section 3.1.2, average measured personal exposure concentrations
are greater than those in both indoor and outdoor air. These differences may be attributable to several factors.  First,
HAPEM6 does not include pollution sources within indoor microenvironments, such as attached garages,
environmental tobacco smoke, and solvent storage. Second, measured personal breathing zone concentrations are
integrated measurements that account for time-weighted average (TWA) concentrations that incorporate every
source, activity, and location with which a monitor comes into contact. Microenvironmental models like HAPEM6
simplify individual time budgets so they fit within the microenvironments modeled or monitored.
                                            3-48

-------
Final Regulatory Impact Analysis
       Table 3.2-9 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 25 in a million
in 1999.

       In all projection years, benzene emissions are by far the largest contributor to cancer risk
from mobile sources (see Figure 3.2-4). Other significant contributors to cancer risk from
mobile source air toxics include 1,3-butadiene, acetaldehyde, naphthalene, and hexavalent
chromium. It should be noted, however, that we have no actual measurements of hexavalent
chromium emissions from mobile sources, and that the risk estimate for this pollutant is based on
an assumption that forty percent of the chromium from highway vehicles and eighteen percent of
the chromium from nonroad sources was assumed to be the highly toxic  hexavalent form. The
estimate for highway vehicles is based on data from utility boilers,158 and the estimate for
nonroad equipment is, based on combustion data from stationary combustion turbines that burn
diesel fuel.159  Thus there is a great deal of uncertainty in estimates for this pollutant.

       Despite significant reductions in risk from mobile source air toxics, average inhalation
cancer risks for these pollutants in 2030, accounting for both mobile and stationary source
contributions, remain well above 20  in 1,000,000 (Figure 3.2-5). In addition, average risk from
exposure to benzene remains above 9 in 1,000,000.
                                          3-49

-------
Final Regulatory Impact Analysis
  Table 3.2-7. National Means  of Census Tract Median Population Exposure Concentrations of Mobile Source Air Toxics in
                                1999, 2015, 2020, and 2030, Without Controls in this Rule.

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
(Hg m 3)
3.96E-02
O.OOE+00
4.00E-01
O.OOE+00
3.05E-01
O.OOE+00
O.OOE+00
O.OOE+00
6.12E-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
1.28E-01
1999 annual average concentrations (ng in 3)
major
1.54E-03
1.70E-02
2.34E-02
2.56E-03
1.76E-02
3.23E-04
4.25E-05
1.45E-02
3.29E-02
5.50E-02
1.05E-02
1.05E-03
3.82E-03
3.02E-04
2.87E-03
7.73E-03
2.04E-02
1.61E-01
8.08E-02
area&
other
1.66E-02
1.86E-02
4.33E-02
2.35E-02
1.16E-01
1.79E-04
7.94E-05
7.49E-02
7.20E-02
3.60E-01
4.84E-02
8.93E-04
3.37E-02
5.78E-04
l.OOE-02
1.80E-02
1.14E-02
6.57E-01
4.66E-01
onroad
6.39E-02
8.23E-01
8.08E-01
6.62E-02
8.08E-01
1.93E-05
1.30E-05
3.22E-01
5.78E-01
2.85E-01
4.61E-01
1.08E-05
1.79E-02
2.38E-05
1.56E-03
1.93E-01
3.40E-02
2.14E+00
1.21E+00
nonroad
1.64E-02
1.57E-01
1.18E-01
1.83E-02
1.51E-01
2.21E-05
5.06E-06
8.02E-02
1.88E-01
7.13E-02
3.40E-01
2.40E-06
3.85E-03
4.17E-05
5.48E-04
3.35E-02
3.03E-03
3.42E-01
3.33E-01
total
(including
background)
1.38E-01
1.02E+00
1.39E+00
1.10E-01
1.40E+00
5.43E-04
1.40E-04
4.91E-01
1.48E+00
7.71E-01
8.59E-01
1.96E-03
5.92E-02
9.46E-04
1.50E-02
2.52E-01
6.88E-02
3.30E+00
2.22E+00
2015 annual average concentrations (|j,g m"3)
major
1.72E-03
8.68E-03
2.41E-02
2.91E-03
1.25E-02
4.11E-04
5.40E-05
9.92E-03
4.15E-02
4.94E-02
1.26E-03
1.13E-02
3.37E-03
3.47E-04
2.26E-03
7.24E-03
2.40E-02
1.16E-01
6.79E-02
area&
other
1.69E-02
2.18E-02
4.60E-02
2.14E-02
1.37E-01
2.43E-04
1.09E-04
l.OOE-01
8.26E-02
4.41E-01
1.17E-03
5.35E-02
4.18E-02
6.50E-04
1.16E-02
1.89E-02
1.56E-02
8.80E-01
6.43E-01
onroad
2.88E-02
4.16E-01
4.70E-01
2.90E-02
4.53E-01
2.64E-05
1.78E-05
1.62E-01
2.46E-01
1.44E-01
1.48E-05
1.24E-01
9.89E-03
3.29E-05
8.33E-04
9.56E-02
1.73E-02
1.09E+00
6.11E-01
nonroad
1.01E-02
9.26E-02
9.07E-02
1.49E-02
9.87E-02
2.34E-05
5.38E-06
4.69E-02
1.38E-01
4.98E-02
2.84E-06
8.90E-02
4.02E-03
4.80E-05
4.97E-04
2.28E-02
1.83E-03
2.06E-01
1.85E-01
total
(including
background)
9.71E-02
5.39E-01
1.03E+00
6.83E-02
1.01E+00
7.03E-04
1.86E-04
3.19E-01
1.12E+00
6.85E-01
2.45E-03
2.78E-01
5.91E-02
1.08E-03
1.52E-02
1.45E-01
5.86E-02
2.29E+00
1.63E+00
                                                         5-50

-------
Final Regulatory Impact Analysis


  Table 3.2-7 (cont'd). National Means of Census Tract Median Population Exposure Concentrations of Mobile Source Air
                           Toxics in 1999, 2015, 2020, and 2030, Without Controls in this Rule.

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
(M^m3)
3.96E-02
O.OOE+00
4.00E-01
O.OOE+00
3.05E-01
O.OOE+00
O.OOE+00
O.OOE+00
6.12E-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
1.28E-01
2020 annual average concentrations (jig in 3)
major
1.86E-03
9.37E-03
2.52E-02
3.27E-03
1.37E-02
4.59E-04
6.14E-05
1.11E-02
4.71E-02
5.44E-02
1.27E-02
1.40E-03
3.78E-03
3.77E-04
2.51E-03
7.27E-03
2.74E-02
1.30E-01
7.68E-02
area &
other
1.69E-02
2.31E-02
4.70E-02
2.07E-02
1.42E-01
2.74E-04
1.23E-04
1.10E-01
8.68E-02
4.77E-01
5.48E-02
1.29E-03
4.44E-02
7.15E-04
1.18E-02
1.94E-02
1.72E-02
9.68E-01
7.10E-01
onroad
2.98E-02
4.16E-01
4.85E-01
2.99E-02
4.64E-01
2.90E-05
1.96E-05
1.62E-01
2.52E-01
1.33E-01
1.01E-01
1.62E-05
9.84E-03
3.62E-05
8.63E-04
9.81E-02
1.80E-02
1.10E+00
6.18E-01
nonroad
1.07E-02
9.21E-02
9.01E-02
1.58E-02
1.02E-01
2.37E-05
5.45E-06
4.83E-02
1.38E-01
5.19E-02
9.25E-02
3.00E-06
4.31E-03
5.02E-05
5.01E-04
2.25E-02
1.87E-03
2.09E-01
1.87E-01
total
(including
background)
9.88E-02
5.41E-01
1.05E+00
6.97E-02
1.03E+00
7.86E-04
2.09E-04
3.32E-01
1.14E+00
7.17E-01
2.61E-01
2.71E-03
6.23E-02
1.18E-03
1.57E-02
1.47E-01
6.45E-02
2.41E+00
1.72E+00
2030 annual average concentrations (fig m"3)
major
1.86E-03
9.37E-03
2.52E-02
3.27E-03
1.37E-02
4.59E-04
6.14E-05
1.11E-02
4.71E-02
5.44E-02
1.27E-02
1.40E-03
3.78E-03
3.77E-04
2.51E-03
7.27E-03
2.74E-02
1.30E-01
7.68E-02
area &
other
1.69E-02
2.31E-02
4.70E-02
2.07E-02
1.42E-01
2.74E-04
1.23E-04
1.10E-01
8.68E-02
4.77E-01
5.48E-02
1.29E-03
4.44E-02
7.15E-04
1.18E-02
1.94E-02
1.72E-02
9.68E-01
7.10E-01
onroad
3.49E-02
4.81E-01
5.68E-01
3.51E-02
5.40E-01
3.56E-05
2.40E-05
1.87E-01
2.94E-01
1.46E-01
l.OOE-01
1.99E-05
1.14E-02
4.45E-05
1.02E-03
1.14E-01
2.13E-02
1.28E+00
7.17E-01
nonroad
1.21E-02
l.OOE-01
9.78E-02
1.79E-02
1.15E-01
2.43E-05
5.62E-06
5.41E-02
1.51E-01
5.83E-02
1.04E-01
3.35E-06
4.94E-03
5.47E-05
5.58E-04
2.42E-02
2.10E-03
2.30E-01
2.06E-01
total
(including
background)
1.05E-01
6.14E-01
1.14E+00
7.70E-02
1.12E+00
7.93 E-04
2.14E-04
3.62E-01
1.19E+00
7.36E-01
2.72E-01
2.71E-03
6.45E-02
1.19E-03
1.59E-02
1.65E-01
6.80E-02
2.61E+00
1.84E+00
                                                         5-51

-------
Final Regulatory Impact Analysis
   Table 3.2-8. National Means of Census Tract 90  Percentile Population Exposure Concentrations of Mobile Source Air
                           Toxics in 1999, 2015, 2020, and 2030, Without Controls in this Rule.

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
(Hg m 3)
5.88E-02
O.OOE+00
5.82E-01
O.OOE+00
4.50E-01
O.OOE+00
O.OOE+00
O.OOE+00
8.03E-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
2.04E-01
1999 annual average concentrations (ng in 3)
major
2.03E-03
2.65E-02
3.48E-02
3.68E-03
2.57E-02
4.55E-04
6.16E-05
2.33E-02
4.21E-02
7.54E-02
1.46E-02
1.44E-03
4.81E-03
4.25E-04
3.68E-03
1.30E-02
2.87E-02
2.52E-01
1.23E-01
area&
other
2.23 E-02
3.12E-02
6.34E-02
3.36E-02
1.71E-01
2.59E-04
1.15E-04
1.19E-01
9.22E-02
5.11E-01
7.13E-02
1.18E-03
4.39E-02
8.25E-04
1.21E-02
2.79E-02
1.78E-02
1.05E+00
7.05E-01
onroad
l.OOE-01
1.42E+00
1.27E+00
1.07E-01
1.24E+00
2.88E-05
1.92E-05
5.49E-01
7.89E-01
4.32E-01
7.22E-01
1.47E-05
2.44E-02
3.52E-05
2.04E-03
3.36E-01
5.90E-02
3.61E+00
1.95E+00
nonroad
2.49E-02
2.65E-01
1.80E-01
2.82E-02
2.28E-01
3.15E-05
7.20E-06
1.35E-01
2.52E-01
1.07E-01
5.16E-01
3.25E-06
5.09E-03
6.04E-05
7.05E-04
5.58E-02
5.23E-03
5.66E-01
5.25E-01
total
(including
background)
2.08E-01
1.75E+00
2.13E+00
1.72E-01
2.12E+00
7.74E-04
2.03E-04
8.27E-01
1.98E+00
1.13E+00
1.32E+00
2.64E-03
7.83E-02
1.35E-03
1.85E-02
4.33E-01
1.11E-01
5.48E+00
3.50E+00
2015 annual average concentrations (|j,g m"3)
major
2.15E-03
1.30E-02
3.32E-02
3.72E-03
1.70E-02
5.81E-04
7.88E-05
1.51E-02
4.93E-02
6.51E-02
1.45E-02
1.72E-03
4.07E-03
4.94E-04
2.89E-03
1.15E-02
3.31E-02
1.70E-01
9.59E-02
area&
other
2.16E-02
3.56E-02
6.27E-02
2.79E-02
1.91E-01
3.51E-04
1.57E-04
1.51E-01
9.74E-02
5.95E-01
7.24E-02
1.55E-03
5.13E-02
9.09E-04
1.38E-02
2.72E-02
2.31E-02
1.32E+00
9.14E-01
onroad
4.11E-02
7.08E-01
6.89E-01
4.36E-02
6.52E-01
3.97E-05
2.63E-05
2.63E-01
3.03E-01
2.04E-01
1.92E-01
2.01E-05
1.25E-02
4.77E-05
1.04E-03
1.60E-01
2.87E-02
1.71E+00
9.13E-01
nonroad
1.39E-02
1.54E-01
1.28E-01
2.11E-02
1.39E-01
3.35E-05
7.67E-06
7.53E-02
1.67E-01
7.06E-02
1.34E-01
3.85E-06
4.99E-03
6.89E-05
6.14E-04
3.57E-02
3.01E-03
3.21E-01
2.72E-01
total
(including
background)
1.38E-01
9.10E-01
1.49E+00
9.64E-02
1.45E+00
1.01E-03
2.70E-04
5.04E-01
1.42E+00
9.34E-01
4.13E-01
3.30E-03
7.29E-02
1.52E-03
1.84E-02
2.34E-01
8.79E-02
3.52E+00
2.40E+00
                                                         5-52

-------
Final Regulatory Impact Analysis
 Table 3.2-8 (cont'd). National Means  of Census Tract 90  Percentile Population Exposure Concentrations of Mobile Source
                         Air Toxics in 1999, 2015, 2020, and 2030, Without Controls in this Rule.

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
(Hg in 3)
5.88E-02
O.OOE+00
5.82E-01
O.OOE+00
4.50E-01
O.OOE+00
O.OOE+00
O.OOE+00
8.03E-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
2.04E-01
2020 annual average concentrations (ng in 3)
major
2.32E-03
1.40E-02
3.47E-02
4.15E-03
1.86E-02
6.51E-04
8.98E-05
1.68E-02
5.60E-02
7.12E-02
1.61E-02
1.92E-03
4.55E-03
5.39E-04
3.21E-03
1.16E-02
3.78E-02
1.88E-01
1.08E-01
area&
other
2.16E-02
3.75E-02
6.42E-02
2.70E-02
1.99E-01
3.97E-04
1.77E-04
1.65E-01
1.02E-01
6.39E-01
7.34E-02
1.71E-03
5.44E-02
l.OOE-03
1.42E-02
2.78E-02
2.55E-02
1.44E+00
1.01E+00
onroad
4.28E-02
7.09E-01
7.14E-01
4.53E-02
6.68E-01
4.37E-05
2.90E-05
2.62E-01
3.11E-01
1.86E-01
1.52E-01
2.21E-05
1.24E-02
5.25E-05
1.08E-03
1.65E-01
2.99E-02
1.73E+00
9.22E-01
nonroad
1.48E-02
1.54E-01
1.28E-01
2.23 E-02
1.44E-01
3.39E-05
7.78E-06
7.71E-02
1.67E-01
7.27E-02
1.37E-01
4.07E-06
5.33E-03
7.19E-05
6.21E-04
3.52E-02
3.07E-03
3.23E-01
2.74E-01
total
(including
background)
1.40E-01
9.14E-01
1.52E+00
9.87E-02
1.48E+00
1.13E-03
3.04E-04
5.21E-01
1.44E+00
9.69E-01
3.78E-01
3.65E-03
7.66E-02
1.66E-03
1.91E-02
2.39E-01
9.64E-02
3.68E+00
2.51E+00
2030 annual average concentrations (jig m 3)
major
2.32E-03
1.40E-02
3.47E-02
4.15E-03
1.86E-02
6.51E-04
8.98E-05
1.68E-02
5.60E-02
7.12E-02
1.61E-02
1.92E-03
4.55E-03
5.39E-04
3.21E-03
1.16E-02
3.78E-02
1.88E-01
1.08E-01
area&
other
2.16E-02
3.75E-02
6.42E-02
2.70E-02
1.99E-01
3.97E-04
1.77E-04
1.65E-01
1.02E-01
6.39E-01
7.34E-02
1.71E-03
5.44E-02
l.OOE-03
1.42E-02
2.78E-02
2.55E-02
1.44E+00
1.01E+00
onroad
5.11E-02
8.25E-01
8.55E-01
5.37E-02
7.93E-01
5.40E-05
3.58E-05
3.06E-01
3.70E-01
2.06E-01
1.50E-01
2.72E-05
1.45E-02
6.45E-05
1.29E-03
1.94E-01
3.57E-02
2.04E+00
1.09E+00
nonroad
1.72E-02
1.69E-01
1.41E-01
2.56E-02
1.65E-01
3.48E-05
8.04E-06
8.71E-02
1.86E-01
8.21E-02
1.53E-01
4.54E-06
6.16E-03
7.86E-05
6.97E-04
3. 83 E-02
3.47E-03
3.61E-01
3.07E-01
total
(including
background)
1.51E-01
1.05E+00
1.68E+00
1.10E-01
1.63E+00
1.14E-03
3.11E-04
5.75E-01
1.52E+00
9.98E-01
3.93E-01
3.66E-03
7.96E-02
1.68E-03
1.94E-02
2.72E-01
1.03E-01
4.03E+00
2.71E+00
                                                         5-53

-------
Final Regulatory Impact Analysis
    Table 3.2-9.  National Average Cancer Risk Across Census Tracts for 1999, 2015, 2020, and 2030 by Pollutant, Without
                                                 Controls in this Rule.

Pollutant
1,3-Butadiene
Acetaldehyde
Benzene
Chromium VI
Formaldehyde
Naphthalene
Nickel
POM
1999 average Individual risk
major
4.36E-08
5.65E-08
1.49E-07
5.32E-07
1.81E-10
1.21E-07
4.81E-08
1.77E-07
area &
other
4.85E-07
1.10E-07
9.82E-07
9.43E-07
4.51E-10
1.22E-06
9.79E-08
1.06E-06
onroad
2.06E-06
1.96E-06
6.79E-06
1.69E-07
3.36E-09
6.38E-07
4.17E-09
1.05E-07
nonroad
5.39E-07
2.89E-07
1.30E-06
7.18E-08
1.11E-09
1.37E-07
6.20E-09
3.62E-08
total
(including
background)
4.43E-06
3.39E-06
1.18E-05
1.72E-06
8.69E-09
2.11E-06
1.56E-07
1.38E-06
2015 annual average individual risk
major
4.62E-08
5.59E-08
l.OOE-07
6.67E-07
2.10E-10
1.01E-07
5.53E-08
1.46E-07
area &
other
4.50E-07
1.16E-07
1.13E-06
1.25E-06
5.18E-10
1.46E-06
1.07E-07
1.25E-06
onroad
8.69E-07
1.08E-06
3.66E-06
2.29E-07
1.35E-09
3.43E-07
5.65E-09
5.39E-08
nonroad
3.20E-07
2.12E-07
8.25E-07
8.11E-08
7.69E-10
1.39E-07
6.87E-09
3.25E-08
total
(including
background)
2.97E-06
2.43E-06
8.33E-06
2.23E-06
6.43E-09
2.04E-06
1.75E-07
1.48E-06

Pollutant
1,3-Butadiene
Acetaldehyde
Benzene
Chromium VI
Formaldehyde
Naphthalene
Nickel
POM
2020 annual average individual risk
major
4.95E-08
5.80E-08
1.09E-07
7.53E-07
2.34E-10
1.12E-07
6.02E-08
1.61E-07
area &
other
4.38E-07
1.19E-07
1.17E-06
1.40E-06
5.47E-10
1.54E-06
1.16E-07
1.30E-06
onroad
8.92E-07
1.10E-06
3.71E-06
2.50E-07
1.38E-09
3.39E-07
6.19E-09
5.54E-08
nonroad
3.39E-07
2.08E-07
8.54E-07
8.34E-08
7.63E-10
1.48E-07
7.10E-09
3.27E-08
total
(including
background)
3.00E-06
2.46E-06
8.45E-06
2.49E-06
6.49E-09
2.14E-06
1.90E-07
1.55E-06
2030 annual average individual risk
major
4.82E-08
5.75E-08
1.08E-07
7.48E-07
2.28E-10
1.09E-07
6.01E-08
1.61E-07
area &
other
4.19E-07
1.19E-07
1.16E-06
1.38E-06
5.54E-10
1.52E-06
1.15E-07
1.31E-06
onroad
1.03E-06
1.28E-06
4.29E-06
3.05E-07
1.59E-09
3.91E-07
7.55E-09
6.52E-08
nonroad
3.86E-07
2.23E-07
9.59E-07
8.78E-08
8.22E-10
1.69E-07
7.60E-09
3.59E-08
total
(including
background)
3.16E-06
2.65E-06
9.13E-06
2.52E-06
6.76E-09
2.19E-06
1.90E-07
1.57E-06
                                                         5-54

-------
Final Regulatory Impact Analysis
 Figure 3.2-4. Contributions to Average Inhalation Cancer Risk from Air Toxics Emitted
  by Mobile Sources, 2020 (Not Including Diesel PM and Diesel Exhaust Organic Gases),
                             Without Controls in this Rule.
                               Naphthalene  Nickel
                                  6%   ~
                      Formaldehyde
                         0%    ~"\
                       Chromium VI
                          4%
1,3-Butadiene
   15%
                                                        Acetaldehyde
                                                           16%
                              Benzene
                               58%
   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), Without Controls in this Rule.
                     O.OE+00
                                1999      2015     2020     2030
                                              Year
                                         3-55

-------
Final Regulatory Impact Analysis
       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 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),160 and is
also discussed in Chapter 12 of the RIA). These statistics do not include populations in Alaska
and Hawaii; thus populations in these States were assumed to remain at year 2000 levels. More
details on the methodology used to project the U. S. population above various cancer risk
benchmarks are provided  in the technical support document "National-Scale Modeling of Mobile
Source Air Toxic Emissions, Air Quality, Exposure and Risk for the Final Mobile Source Air
Toxics Rule." From Figure 3.2-6 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 223 million in 1999 to 272 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, Without Controls in  this Rule.
Ann
fUU
OCfl
oou
= 1 E-04




       Tables 3.2-10 and 3.2-11 summarize national average population hazard quotients 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 33% between 1999 and 2030 (Figure 3.2-7), it is
still over 4 in 2030, indicating a potential for adverse health effects. The reduction in hazard
index occurs despite large increases in activity for highway and nonroad sources. In addition,
                                          3-56

-------
Final Regulatory Impact Analysis
about 90% of this non-cancer risk is attributable to acrolein in all projection years.  It should be
noted that the 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) are 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 and increasing vehicle miles traveled will increase the number of Americans
with a respiratory hazard index for mobile source air toxics above one, from 258 million in 1999
to 307 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.: Reference Case

       Table 3.2-12 gives the distribution of nationwide individual cancer risks for mobile
source air toxics in 2020, absent the controls  being finalized in this rule. 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 three 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, and separately for 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.  Higher risks
are also seen in States with colder winters, due to elevated cold start emissions.

       Previously discussed changes to the HAPEM exposure model, to account for near road
impacts, can impact distributions of risk.  In order to evaluate the effect of switching to
HAPEM6 from HAPEM5 on individual risks nationally, we conducted model runs using
identical input data. Figure 3.2-13 depicts the national distribution of individual cancer risks
from benzene, comparing HAPEM6 and HAPEM5. Note that the graph is on a logarithmic
scale. As the graph illustrates, when HAPEM6 is used, there are fewer individuals with lower
benzene cancer risk levels (e.g. lxlO"4) is higher with  HAPEM6 than HAPEM5.  In general, the distribution of cancer
risks shifts slightly higher when comparing HAPEM6 to HAPEM5, but the largest effects are
observed in the  populations with the highest and lowest risk levels, which are generally small
fractions of the total population.
                                           3-57

-------
Final Regulatory Impact Analysis
       Table 3.2-13 gives the distribution of nationwide individual 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 over 20 times that at the 5th
percentile, and about 4 times the median.  Thus, some populations are experiencing much higher
hazard indices than others. Figure 3.2-14 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.
This error was discovered at too late a date to produce and update emissions inventories for use
in the analyses undertaken for this rule.  The errors are not expected to affect the analyses of the
impacts of controls undertaken for this rule.
                                           3-58

-------
Table 3.2-10.  National Average Population Hazard Quotient for Chronic Noncancer Effects Across Census Tracts, 1999 -
                                      2030, Without Controls in this Rule.

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
1999 average Hazard Quotient
major
7.27E-04
2.86E-03
1.44E-01
6.35E-04
4.43E-04
1.60E-05
3.36E-03
2.76E-04
3.86E-06
2.04E-02
1.19E-03
4.62E-03
2.38E-05
4.55E-04
8.47E-04
area &
other
8.08E-03
5.54E-03
1.28E+00
4.20E-03
7.86E-04
8.09E-05
8.37E-03
1.89E-03
1.76E-05
1.93E-02
1.19E-02
9.42E-03
1.28E-05
1.82E-03
5.00E-03
onroad
3.43E-02
9.92E-02
3.70E+00
2.90E-02
1.41E-04
3.60E-04
6.23E-02
1.55E-03
1.72E-04
2.27E-04
6.25E-03
4.01E-04
3.77E-05
5.96E-03
1.32E-02
nonroad
8.98E-03
1.46E-02
1.03E+00
5.55E-03
5.98E-05
9.17E-05
2.05E-02
3.95E-04
1.28E-04
4.59E-05
1.35E-03
5.96E-04
3.46E-06
9.69E-04
3.72E-03
total
(including
background)
7.39E-02
1.71E-01
6.16E+00
5.06E-02
1.43E-03
5.48E-04
1.61E-01
4.11E-03
3.21E-04
3.99E-02
2.07E-02
1.50E-02
7.78E-05
9.20E-03
2.43E-02
2015 average Hazard Quotient
major
7.69E-04
2.82E-03
1.58E-01
4.29E-04
5.56E-04
1.05E-05
3.90E-03
2.43E-04
3.94E-06
2.65E-02
9.88E-04
5.32E-03
2.85E-05
3.12E-04
6.85E-04
area &
other
7.49E-03
5.84E-03
1.13E+00
4.83E-03
1.04E-03
1.05E-04
9.62E-03
2.21E-03
1.88E-05
2.56E-02
1.43E-02
1.03E-02
1.76E-05
2.39E-03
6.69E-03
onroad
1.45E-02
5.46E-02
1.54E+00
1.56E-02
1.90E-04
1.74E-04
2.51E-02
7.58E-04
4.43E-05
3.07E-04
3.36E-03
5.43E-04
1.84E-05
2.88E-03
6.38E-03
nonroad
5.34E-03
1.07E-02
8.10E-01
3.53E-03
6.76E-05
5.30E-05
1.43E-02
2.71E-04
3.20E-05
5.32E-05
1.36E-03
6.61E-04
2.05E-06
5.72E-04
2.02E-03
total (including
background)
4.96E-02
1.23E-01
3.63E+00
3.56E-02
1.86E-03
3.42E-04
1.19E-01
3.48E-03
9.90E-05
5.24E-02
2.00E-02
1.68E-02
6.66E-05
6.16E-03
1.72E-02
                                                     3-59

-------
Final Regulatory Impact Analysis


  Table 3.2-10 (cont'd).  National Average Population Hazard Quotient for Chronic Noncancer Effects Across Census Tracts,
                                             Without Controls in this Rule.

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
2020 average Hazard Quotient
major
8.25E-04
2.93E-03
1.78E-01
4.67E-04
6.28E-04
1.17E-05
4.34E-03
2.66E-04
4.36E-06
2.99E-02
1.09E-03
5.78E-03
3.29E-05
3.47E-04
7.69E-04
area&
other
7.30E-03
5.99E-03
1.09E+00
4.99E-03
1.17E-03
1.14E-04
1.02E-02
2.37E-03
1.91E-05
2.80E-02
1.51E-02
1.12E-02
1.96E-05
2.63E-03
7.35E-03
onroad
1.49E-02
5.58E-02
1.57E+00
1.58E-02
2.09E-04
1.72E-04
2.55E-02
6.92E-04
3.53E-05
3.37E-04
3.33E-03
5.95E-04
1.90E-05
2.89E-03
6.39E-03
nonroad
5.64E-03
1.05E-02
8.52E-01
3.65E-03
6.95E-05
5.44E-05
1.42E-02
2.82E-04
3.28E-05
5.59E-05
1.45E-03
6.83E-04
2.09E-06
5.78E-04
2.04E-03
total
(Including
background)
5.00E-02
1.24E-01
3.69E+00
3.61E-02
2.07E-03
3.52E-04
1.20E-01
3.61E-03
9.16E-05
5.83E-02
2.10E-02
1.83E-02
7.36E-05
6.45E-03
1.80E-02
2030 average Hazard Quotient
major
8.03E-04
2.90E-03
1.78E-01
4.63E-04
6.23 E-04
1.15E-05
4.23 E-03
2.65E-04
4.26E-06
3.08E-02
1.07E-03
5.78E-03
3.32E-05
3.44E-04
7.59E-04
area &
other
6.98E-03
6.02E-03
1.08E+00
4.96E-03
1.15E-03
1.12E-04
1.03E-02
2.32E-03
1.90E-05
2.81E-02
1.49E-02
1.10E-02
1.97E-05
2.63E-03
7.26E-03
onroad
1.72E-02
6.47E-02
1.82E+00
1.83E-02
2.54E-04
1.96E-04
2.95E-02
7.53E-04
3.44E-05
4. 11 E-04
3. 83 E-03
7.26E-04
2.22E-05
3.32E-03
7.33E-03
nonroad
6.43E-03
1.13E-02
9.62E-01
4.10E-03
7.32E-05
6.09E-05
1.53E-02
3.17E-04
3.62E-05
6.15E-05
1.65E-03
7.30E-04
2.32E-06
6.37E-04
2.25E-03
total (including
background)
5.26E-02
1.34E-01
4.04E+00
3.90E-02
2.10E-03
3.81E-04
1.25E-01
3.66E-03
9.38E-05
5.94E-02
2.14E-02
1.83E-02
7.74E-05
6.93E-03
1.90E-02
                                                         5-60

-------
Final Regulatory Impact Analysis
  Table 3.2-11.  National Respiratory Hazard Index for Chronic Noncancer Effects across
                     Census Tracts, Without Controls in this Rule.
Respiratory System Average Hazard Index
Year
1999
2015
2020
2030
background
0.12
0.12
0.12
0.11
major
0.16
0.17
0.19
0.19
area & other
1.32
1.17
1.14
1.13
on road
3.88
1.63
1.66
1.92
n on road
1.07
0.84
0.88
0.99
total (including
background)
6.54
3.92
3.99
4.35
Figure 3.2-7. Average Respiratory Hazard Index for U.S. Population (Aggregate of Hazard
           Quotients for Individual Pollutants), Without Controls in this Rule.
7

«
£
x °
0
^ 4
C *r
-3
« °
N
T 9
-*- L.
1
n
u -


























_
1999













i













2015













i
Year













2020













i













2030



















...
D rnoDiie
• stationary
D background





                                        3-61

-------
Final Regulatory Impact Analysis
    Figure 3.2-8. U. S. Population at Various Non-Cancer Hazard Benchmarks due to
    Exposure to Mobile Source Air Toxics, 1999 - 2030, Without Controls in this Rule.
      400
                                                                     DHQ<0.1
                                                                     D0.1 <= HQ< 1
                                                                     • 1 <= HQ< 10
                                                                     DHQ>= 10
               1999
2015
2020
2030
                                     Year
  Table 3.2-12. Distribution of Individual Cancer Risks for Mobile Source Air Toxics in
                         2020, Without Controls in this Rule.

Pollutant
Total Risk: All HAPs
1,3-Butadiene
Acetaldehyde
Benzene
Chromium VI
Formaldehyde
Naphthalene
Nickel
POM
2020 risk distribution
5th
percentile
4.71E-06
1.52E-07
1.09E-06
2.72E-06
3.85E-08
2.29E-09
1.59E-07
1.84E-09
1.26E-07
10th
percentile
6.08E-06
2.96E-07
1.19E-06
3.36E-06
7.93 E-08
2.89E-09
2.80E-07
4.09E-09
1.90E-07
25th
percentile
9.78E-06
1.06E-06
1.46E-06
4.84E-06
2.38E-07
4.12E-09
6.72E-07
1.39E-08
3.48E-07
Median
1.53E-05
2.30E-06
1.96E-06
6.93E-06
7.01E-07
5.75E-09
1.39E-06
4.60E-08
6.78E-07
75th
percentile
2.37E-05
3.60E-06
2.81E-06
l.OOE-05
1.81E-06
7.67E-09
2.61E-06
1.31E-07
1.19E-06

3.79E-05
5.47E-06
4.20E-06
1.48E-05
4.54E-06
1.05E-08
4.73E-06
3.04E-07
1.99E-06
95th
percentile
4.93E-05
7.70E-06
5.35E-06
1.86E-05
7.29E-06
1.29E-08
6.68E-06
5.06E-07
3.07E-06
                                        3-62

-------
Final Regulatory Impact Analysis
Figure 3.2-9. 2020 County Median Cancer Risk for All Mobile Source Air Toxics, Without
                                 Controls in this Rule.
  Figure 3.2-10. 2020 County Median Cancer Risk for Benzene, Without Controls in this
                                        Rule.
                                         3-63

-------
Final Regulatory Impact Analysis
  Figure 3.2-11.  2020 County Median Cancer Risk for Acetaldehyde, Without Controls in
                                      this Rule.
 Figure 3.2-12. 2020 County Median Cancer Risk for 1,3-Butadiene, Without Controls in
                                      this Rule.
                                        3-64

-------
Final Regulatory Impact Analysis
 Figure 3.2-13. 1999 Comparison Between HAPEM6 and HAPEM5 Nationwide Individual
                  Benzene Cancer Risk, Without Controls in this Rule.
     100,000,000


      10,000,000


       1,000,000


        100,000


   o     10,000
   J5

   §•      1,000


           100


            10

rfl

1
r^ r^ h
o o c
LJJ LJJ LJ
- t
3 t
J L

^ h
D C
JJ U

- r
3 C
J L
rfl
~- h-
•? *•?
JJ LLJ
r :
1^ 1^ CD CD
CD CD CD CD
LLJ LLJ LLJ LLJ
'
'
:

CD CD CD C£
CD CD CD C
LJJ LJJ LJJ U



3 C!
3 C
J U
•
'•

3 C!
3 C
J U



3 C£
3 C
J U



3 U
3 C
J U

'•

3 U
3 C
J U



3 U
3 C
J L


•3 u
p c
LJ U


3 U
3 C
J U

3 U
3 C
U U

3 LT
3 C
U U

3 U
3 C
J U
i n
3 LO •*
3 CD C
J LJJ L

a- ^r ^r
p cp cp
U LLJ LLJ
                                      Population with Risk > Risk bin
                                        IHAPEM6 DHAPEM5
                                         3-65

-------
Final Regulatory Impact Analysis
   Table 3.2-13. Distribution of Individual Hazard Quotients/Hazard Indices for Mobile
Source Air Toxics (from both Mobile and Stationary Sources) in 2020, Without Controls in
                                      this Rule.

Pollutant
Acrolein
Respiratory System
2020 average Hazard Quotient or Hazard Index
5th
percentile
0.41
0.53
10th
percentile
0.61
0.75
25th
percentile
1.18
1.36
Median
2.31
2.57
75th
percentile
4.47
4.83
90th
percentile
8.05
8.54
95th
percentile
11.3
11.9
Figure 3.2-14. 2020 County Median Non-Cancer Hazard Index Respiratory Mobile Source
                       Air Toxics, Without Controls in this Rule.
                                        3-66

-------
Final Regulatory Impact Analysis


3.2.1.2.4      Impacts of Controls on Average Inhalation Cancer Risks and Noncancer Hazards

       The standards being finalized in this rule will substantially reduce inhalation cancer and
noncancer risk from exposure to air toxics emitted by mobile sources across the United States.
Table 3.2-14 shows that in 2030, the highway vehicle contribution to MSAT cancer risk will be
reduced on average 36% across the U.S., and the nonroad equipment contribution will be
reduced about 6%. In 2030, the highway vehicle contribution to benzene cancer risk will be
reduced on average by 43% across the U.S., and the nonroad contribution will be reduced by
11%.   Table 3.2-15 summarizes the change in median and 95th percentile inhalation cancer risks
from benzene and all MSATs attributable to all outdoor sources in 2015, 2020, and 2030, with
the controls being finalized in this rule.   Reductions are  significantly larger for individuals in
the 95th percentile than in the 50th percentile. Thus, this rule is providing bigger benefits to
individuals experiencing the highest levels of risk. In states with high fuel benzene levels and
high cold start emissions, the cancer risk reduction from total MSATs is  about 40% or higher
(Table 3.2-16).d Figure 3.2-15 depicts the impact on the mobile source contribution to
nationwide average population cancer risk from all MSATs and benzene in 2030. Nationwide,
the cancer risk attributable to total MSATs would be reduced by 30%, and the risk from mobile
source benzene would be reduced by 37%. Figures 3.2-16 and 3.2.-17 present the distribution of
percent reductions in average MSAT and benzene cancer risk, respectively,  from all sources in
2030 with the controls being finalized in 2030.  Table 3.2-17 shows reductions in hazard
quotients and hazard indices for acrolein and respiratory effects, respectively.  Nationwide, the
mobile source contribution to the acrolein hazard quotient and respiratory hazard index would
both be reduced about 23%, and the highway vehicle contribution will be reduced about 35%.
Summary tables providing exposure and risk data by State, as well as maps of cancer risks and
noncancer hazards with controls and percent reductions with controls, can be found in the docket
for the rule.

       It should be noted that the estimated total relative  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.  Again, as noted previously, since this
modeling did not include the 1.3 vol% maximum average fuel benzene level, reductions in risk
for some parts of the country,  such as the Pacific Northwest, are underestimated.
d Reductions are likely to be higher than estimated by this modeling, due to the 1.3% maximum average fuel
benzene level.
                                          3-67

-------
     Final Regulatory Impact Analysis
Table 3.2-14. Contributions of Source Sectors to Nationwide Average Cumulative MSAT Cancer Risk, With and Without Controls,
                                                   2015, 2020, and 2030


Total MSATs
Reference
Control
% Difference
Benzene
Reference
Control
% Difference
201 5 Average Risks
major

1.17E-06
1.17E-06
0.0

1 .OOE-07
1 .OOE-07
0.3
area & other

5.76E-06
5.74E-06
0.3

1.13E-06
1.12E-06
1.3
total onroad

6.24E-06
4.98E-06
20.2

3.66E-06
2.73E-06
25.4
total
nonroad

1 .62E-06
1 .53E-06
5.3

8.25E-07
7.38E-07
10.5
total (including
background)

1 .97E-05
1 .83E-05
6.9

8.33E-06
7.30E-06
12.3
2020 Average Risks
major

1 .30E-06
1 .30E-06
0.0

1 .09E-07
1 .09E-07
0.3
area & other

6.08E-06
6.06E-06
0.3

1.17E-06
1.15E-06
1.3
total onroad

6.35E-06
4.58E-06
27.9

3.71 E-06
2.45E-06
34.0
total
nonroad

1 .67E-06
1 .58E-06
5.5

8.54E-07
7.62E-07
10.8
total (including
background)

2.03E-05
1 .84E-05
9.3

8.45E-06
7.09E-06
16.2
2030 Average Risks
major

1 .29E-06
1 .29E-06
0.0

1 .08E-07
1 .08E-07
0.3
area & other

6.02E-06
6.01 E-06
0.3

1.16E-06
1.15E-06
1.3
total onroad

7.37E-06
4.69E-06
36.3

4.29E-06
2.43E-06
43.4
total
nonroad

1 .87E-06
1 .77E-06
5.6

9.59E-07
8.54E-07
10.9
total (including
background)

2.14E-05
1 .86E-05
13.1

9.13E-06
7.15E-06
21.7
                                                             5-68

-------
                                   -th
Table 3.2-15. Change in Median and 95  Percentile Inhalation Cancer Risk from Benzene
and all MSATs Attributable to Outdoor Sources in 2015, 2020, and 2030 with the Controls
                            Being Finalized in this Rule.


All
MSATs
Without
Controls
With
Controls
Percent
Change
Benzene
Without
Controls
With
Controls
Percent
Change
2015
median

l.SOxlO'5
1.41xlO'5
6

6.86xlO'6
6.17xlO'6
10
95th

4.75xlO'5
4.37xlO'5
8

1.82xlO'5
1.53xlO'5
16
2020
median

1.53xlO'5
1.40xlO'5
8

6.93xlO'6
6.02xlO'6
13
95th

4.93xlO'5
4.40xlO'5
11

1.86xlO'5
1.47xlO'5
21
2030
median

1.61xlO'5
1.42xlO'5
12

7.37xlO'6
6.06xlO'6
18
95th

5.28xlO'5
4.49xlO'5
15

2.06xlO'5
1.49xlO'5
28
 Table 3.2-16.  States with Highest Reductions in Average Benzene Cancer Risk Resulting
                        from Mobile Source Emissions, 2030.
State
Alaska
North Dakota
Washington
Minnesota
Wyoming
Montana
Idaho
Michigan
South Dakota
Oregon
Average Risk -
Reference Case
l.OlxlO'5
2.92xlO'6
1.39xlO'5
1.21xlO'5
2.38xlO'6
3.12xlO'6
5.03xlO'6
1.09xlO'5
2.73xlO'6
l.OlxlO'5
Average Risk -
Control Case
4.23xlO'6
1.68xlO'6
S.lOxlO'6
7.08xlO'6
1.39xlO'6
1.87xlO'6
3.02xlO'6
6.55xlO'6
1.66xlO'6
6.17xlO'6
Percent Difference
-58%
-42
-42
-42
-41
-40
-40
-40
-39
-39
                                       3-69

-------
 Figure 3.2-15. Contribution to Nationwide Average Population Cancer Risk from Mobile
   Source MSATs and Benzene Emitted by Mobile Sources in 2030, Without and With
                                Controls in this Rule.


Q F OR
8P OR
v 7 P OR

i- b.b-Ub -
d)
o

O
d)
O) 4 F OR
35 ^t.t-uo -
5 ^ P OR
9 F OR
1 P OR
n F+nn


































Total \
















/I SATs

































Ben
















:ene
























D Without Controls

• With Controls







Figure 3.2-16.  Distribution of Percent Reductions in Median MSAT Cancer Risk, 2030, for
                       U.S. Counties with Controls in this Rule.
                                       3-70

-------
 Figure 3.2-17. Distribution of Percent Reductions in Median Benzene Cancer Risk, 2030,
                       for U.S. Counties With Controls in this Rule.
                                                                     -46.519%--21.861%

                                                                     -21.860%--15.843%

                                                                     -15.842%--10.893%

                                                                     -10.892%--5.932%

                                                                     -5.931%-0.747%
As a result of the controls being finalized in this rule, 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 11 million in 2020 and by about 17 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 about 30 million in 2020 and 46 million in 2030 (Table 3.2-18).
                                           3-71

-------
Table 3.2.-17.  Reductions in Hazard Quotients and Hazard Indices for Acrolein and Respiratory Effects Due to MSAT Controls.
2015 Average Hazard Index/ Quotient
area &
other

1.17
1.17
0.0

1.13
totalonroad

1.63
1.35
16.7

1.54
total
nonroad

0.84
0.84
0.0

0.81
total (Including
background)

3.92
3.65
6.9

3.63
2020 Average Hazard Index/ Quotient
major

0.19
0.19
0.0

0.18
area &
other

1.14
1.14
0.0

1.09
total onroad

1.66
1.24
25.6

1.57
total
nonroad

0.88
0.88
0.0

0.85
total
(Including
background)

3.99
3.56
10.7

3.69
2030 Average Hazard Index/ Quotient
major

0.19
0.19
0.0

0.18
area &
other

1.13
1.13
0.0

1.08
total onroad

1.92
1.24
35.4

1.82
total
nonroad

0.99
0.99
0.0

0.96
total (Including
background)

4.35
3.67
15.6

4.04
                                                        5-72

-------
   Table 3.2-18.  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
                                Controls in this Rule.
Year
2015
2020
2030
Benzene
21,697,000
30,031,000
46,360,000
All Mobile Source Air
Toxics
8,149,000
11,257,000
16,737,000
The standards being finalized will also impact on the number of people above various respiratory
hazard index levels (Table 3.2-19).

Table 3.2-19. Decrease in Number of People with Inhalation Exposure above a Respiratory
Hazard Index of One due to Inhalation Exposure from Ambient Sources, With Controls in
                                     this Rule.
Year
2015
2020
2030
Decrease in Population
with Respiratory HI > 1
5,639,000
10,227,000
16,919,000
                                       3-73

-------
3.2.1.3 Strengths and Limitations

       Air quality, exposure, and risk were assessed using the best available suite of tools for
national-scale analysis of air toxics. The same general suite of tools was used in 1996 and 1999
NATA. The 1996 NATA 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:

   •   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.161
   •   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.

   In addition to the improvements for the 1999 NATA, improvements were made in analyses
for this rule, including inventory improvements and updates to HAPEM discussed earlier.

   The SAB expressed their belief that due to the limitations inherent in the analysis, the 1996
NATA should not be used to support regulatory action. However, the use of the improved
analyses in this rule 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 are 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. Uncertainties in dose-response parameters are discussed in Chapter 1 of the RIA.
The modeling also has certain key limitations: results are most accurate for large geographic
areas, 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

                                          3-74

-------
include default adjustments for early life exposures recently recommended in the Supplemental
Guidance for Assessing Susceptibility from Early-Life Exposure to Carcinogens.162 If
warranted, 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 NAT A, 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
for acetaldehyde was almost as good as benzene, but data suggest that ASPEN could be
underpredicting for other mobile source air toxics (see Table 3.2-20).

More detailed discussion of modeling  limitations and uncertainties can be found on the 1999
NATA website.

   Table 3.2-20. 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. Assumptions
about future-year meteorology introduce additional uncertainty in ambient concentrations and
resulting exposures. Another limitation is the use of 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. For the proposed rule
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 proposed rule 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 3.2-21 compares national average total concentrations from the
                                          3-75

-------
proposed rule using 1999 versus scaled backgrounds. More details are provided in the technical
support document for the proposed rule.
163
  Table 3.2-21. National Average Total Concentrations (All Sources and Background) for
 2015, 2020, and 2030 using both the 1999 Background and the Scaled Backgrounds (Data
                                 from Proposed Rule).
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
9.56X10"1
1.50
2020
7.50xlO"2
7.47X10"1
7.40X10"1
9.68X10"1
1.56
2030
7.86xlO"2
7.78X10"1
7.71 xlO'1
1.01
1.60
       The largest impacts were in the Midwest as can be seen in Figure 3.2-19, 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
equipment and fuels in attached garages and increased risks from early lifetime exposures,
estimated risks and risk reductions from fuel benzene control would be larger.
                                         3-76

-------
      Figure 3.2-19. Ratios of Benzene Concentrations with and without an Adjusted
            Background, 2020 (from modeling done to support proposed rule).
                                                                         0.516-0.638
                                                                         0.639-0.716
                                                                         0.717-0.786
                                                                         0.787 - 0.893
                                                                         0.894-1.068
3.2.1.4. 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.164
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
risk to "steady state level."165  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.  Cancer cases were estimated by
summing the distribution of individual cancer risks from the national-scale modeling done to
support this rule.

       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
                                          3-77

-------
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.

       In 2030, the cumulative excess average individual cancer risk from outdoor emissions of
mobile source air toxics is estimated at 2. IxlO"5. If the entire U. S. population (projected to be
about 364 million)166 were exposed to this level of risk over a 70-year lifetime, it would result in
about 7700 cancer cases, which translates into 110 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. If we used the
lower end of this range, incidence estimates would be lower by a factor of about 3.5.  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.

       In 2030, the national average inhalation individual cancer risk from outdoor mobile and
stationary sources of benzene, in the absence of the standards being finalized in this rule, is
estimated at approximately 9.1xlO"6, based on the modeling done for this rule. If the entire U. S.
population were exposed to that level of risk over a 70-year lifetime, it would result in
approximately 47 excess cancer cases per year (Equation 1).

(1) Excess Cancer Cases at 2030 Exposure Level =
(Average Individual Cancer Risk] x (2030 Population)
   = 9.1xlO~6  x3.64x!08 ^3300
   Annual Cancer Cases = 3300/70 = 47

       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, time-weighted  average
individual exposures to benzene could increase by roughly 1.2 to 6.6 |ig/m3 (Appendix 3 A).
There is a great deal of uncertainty associated with these estimates. This could result in about
another 3400 to 18700 excess cancer cases (equation 3). The numerical ranges expressed here
may not fully address all sources of uncertainty involved in making these projections.

(3) Attached Garage Excess Cancer Cases =
(Average Exposure] x (Benzene LIRE) x (Population)
= (\.2-6.6jUg/m3)x(7.8xW-6/jUg/m3) x(3.64x!08)= 3400-18700
Annual Cancer Cases = 49 - 268
                                          3-78

-------
Thus, including attached garages would increase the number of benzene-related excess cancer
cases to somewhere between 96 and 315 annually.  This estimate would still not include higher
exposure levels from 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.

       With the controls being finalized in this rule, average individual risk, not including
attached garage exposures, is reduced to 7.3xlO"6, which results in approximately 37 cancer
cases per year.  Thus, excess leukemia cases would be reduced by 10 annually.  A roughly 40%
reduction in overall benzene emissions could reduce attached garage exposures by approximately
0.5-2.6 |ig/m3 as well, thus reducing excess annual cancer cases from this source of exposure by
another estimated 20 to 100 excess cancer cases. Thus, this rule would prevent roughly 30 to
110 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 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:
   •  As discussed in Chapter 1, the current unit risk estimate for benzene may underestimate
       risk from 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.
   •  Geographically heterogeneous percentage emissions reductions do not translate directly
       into changes in ambient levels, exposure, and risk.
   •  The U.S. population would have experienced higher average exposures in previous years,
       but this is  not accounted for.
   •  The extent to which available studies of indoor air homes in with attached garage are
       representative of the national housing stock is unknown.
   •  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.

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

                                          3-79

-------
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
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.167 A case study of diesel exhaust particulate matter in Wilmington, CA was
recently conducting employing some of these advances.168 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.169 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

                                           3-80

-------
individual road links. 17° 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 3.2-20. Most of the gridded model emissions show lower
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.

       Recent air quality modeling in Portland, OR using the CALPUFF dispersion model
assigned emissions to specific roadway links.171 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.
                                          3-81

-------
    Figure 3.2-20.  Model to Monitor Comparisons of Houston Benzene Concentrations



5
— f
^
1
^
^





6.0-j
5.0-

4.5-
4.0-
3.5-
3.0-
2.0-
1.5-
1.0-
0.5-
0.0-
ISCST3 gndded D ISCST3 link-based A Moni:or
"''•" r»"-*

!',,^L^!?.

A ' A
D
i
D




I I I I
4567
If 0! if fore
, ;, p. _,!
W1 1 : > 1
V. .^ - * !



o
A




i


A recent review of local-scale modeling studies concluded that:172
       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
       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 1999.173 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
                                          3-82

-------
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 the
Philadelphia area differed significantly between the top-down and the bottom-up methodologies
as shown in Table 3.2-22.

  Table 3.2-22.  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 County, using local registration distribution data resulted in
significantly lower air toxics emission factors and resultant emissions, while Montgomery
County showed higher emissions. In the 1999 National Air Toxics Assessment, higher county-
level emissions were generally associated with higher county-level average concentrations, so it
is anticipated that county-level concentrations will follow similar trends.  However, in
microscale settings near specific road links, these results may not apply.

       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
                                          3-83

-------
future. Emissions of volatile organic compounds (VOCs) from the gas cans subject to this final
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 pollution is formed by the reaction of VOCs and 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.

       The science of ozone formation, transport, and accumulation is complex.174 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. 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.

       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.

                                          3-84

-------
       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
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.6  These
health effects are well documented and are critically assessed in the EPA ozone Air Quality
Criteria Document (ozone AQCD) and EPA staff paper.175'176 We are relying on the data and
conclusions in the ozone AQCD and staff paper, regarding the health effects associated with
ozone exposure.

       Ozone-related health effects  include lung function  decrements, respiratory symptoms,
aggravation of asthma, increased hospital and emergency room visits, increased asthma
medication usage, inflammation of the lungs, and a variety of other respiratory effects and
cardiovascular effects. People who  are more susceptible to effects associated with exposure to
ozone include children, asthmatics and the elderly.  There  is also suggestive evidence that certain
people may have greater genetic susceptibility. Those with greater exposures to ozone, for
instance due to time spent outdoors (e.g., outdoor workers), are also of concern.

       Based on a large number of scientific studies, EPA has identified several key health
effects associated with 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.177'178'179'180'181'182
Repeated exposure to ozone can increase susceptibility to  respiratory infection and lung
inflammation and can aggravate preexisting respiratory diseases, such as asthma.183'184'185'186'187
Repeated exposure to sufficient concentrations of ozone can also cause inflammation of the lung,
impairment of lung defense mechanisms, and possibly irreversible changes in lung structure,
which over time could lead to premature aging of the lungs and/or chronic respiratory illnesses,
such as emphysema and chronic bronchitis.188'189'190'191

       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.192 Children and outdoor workers tend to have higher ozone exposures because they
typically are active  outside, working, playing and exercising, during times of day and seasons
(e.g., the summer) when ozone levels are highest.193 For example, summer camp studies in the
Eastern United States and  Southeastern Canada have reported significant reductions in lung
e Human exposure to ozone varies over time due to changes in ambient ozone concentration and because people
move between locations which have notable different ozone concentrations. Also, the amount of ozone delivered to
the lung is not only influenced by the ambient concentration but also by the individuals breathing route and rate.
                                           3-85

-------
function in children who are active outdoors.194'195'196'197'198'199'200'201  Further, children are
more at risk of experiencing health effects from ozone exposure than adults because their
respiratory systems are still developing. 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 chest pain and cough, when exposed to relatively
low ozone levels during prolonged periods of moderate exertion.202'203'204'205
3.3.3  Current 8-Hour Ozone Levels

       The gas can emission reductions will 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/

       As of October 26, 2006, approximately 157 million people live in the 116 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 461 full or partial counties that make
up the 116 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 3B to this RIA.
f An ozone design value is the concentration that determines whether the ozone levels recorded at a monitoring site
meet the NAAQS for ozone. The level of a design value is 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. Greater detail on how these values are
determined (including how to account for missing values and other complexities) is given in Appendices H and I of
40 CFR Part 50. Due to the precision with which the standards are expressed (0.08 ppm for the 8-hour NAAQS
value), a violation of the 8-hour standard is defined as any design value greater than or equal to 0.085 ppm, or 85
ppb. For any particular county, the design value is the highest design value from amongst all the monitors having
valid design values within that county.  If there are no ozone monitors located in a particular county, that county is
not assigned a design value. However, readers should note that ozone design values represent air quality over a
broad area and the absence of a design value for a specific county does not imply that that county is in compliance
with the NAAQS for ozone.  Therefore, our analysis may underestimate the number of counties with ozone levels,
i.e., design values, which are above the level of the ozone NAAQS.
                                            3-86

-------
   Figure 3.3.-1. 8-Hour Ozone and PM2.s Nonattainment Areas and Mandatory Class I
                                      Federal Areas
             ^-  --  *       '
             *   _  y  T
             *                   -i
                        S _  k~    C?
         '•*-  *        *   "Cfe
         u"^7-       <  r?
                    -i—-_     \
          Legend
          ^^| PM and Ozone NonAttainment
            ^ Ozone NonAttainment
              pm25 NonAttainment
          ^^1 Class I Areas
       Counties designated as 8-hour ozone nonattainment were categorized, on the basis of
their one-hour ozone design value, as Subpart 1 or Subpart 2 (69 FR 23951, April 30, 2004).
Areas categorized 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 3B-1 presents the 8-hour ozone nonattainment areas, their 8-hour design values,
and their category or classification. States with 8-hour ozone nonattainment areas are required to
take action to bring those areas into compliance prior to the ozone season in the attainment year.
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 2013 time frame and then be required to maintain the 8-hour ozone NAAQS thereafter.8 The
gas can emission standards being finalized  in this action will become effective in 2009.  Thus,
g The Los Angeles South Coast Air Basin 8-hour ozone nonattainment area will have to attain before June 15, 2021.
                                           3-87

-------
the expected ozone precursor emission inventory reductions from the standards finalized in this
action will be useful to States in attaining and/or maintaining the 8-hour ozone NAAQS.

       EPA's review of the ozone NAAQS is currently underway and a proposed decision in
this review is scheduled for June 2007 with a final rule scheduled for March 2008.  If the ozone
NAAQS is revised then new nonattainment areas could be designated.  While EPA is not relying
on it for purposes of justifying this rule, the emission reductions from this rulemaking would also
be helpful to states if there is an ozone NAAQS revision.

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
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 final rule.206

       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.h 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 rule may help
ensure that these counties continue to maintain their attainment status and the emission
reductions from this final rule will be included by the states in their baseline inventory modeling
for their ozone maintenance plans.
h 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 ozone standard by their statutory date.
                                           3-88

-------
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
Maine
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
Hancock Co
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
80.5
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
51,791
2010popc
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
53,886
                                       3-89

-------
State
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
New Jersey
New Jersey
2010 Projected
8-hour Ozone
County Concentration (ppb)a 2000 popb
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
Mercer Co
Middlesex Co
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
95.2
92.1
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
350,761
750,162
2010 pop0
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
359,912
805,537
3-90

-------
State
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
Pennsylvania
Pennsylvania
2010 Projected
8-hour Ozone
County Concentration (ppb)a 2000 popb
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
Beaver Co
Berks Co
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
79.6
81.7
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
181,412
373,637
2010 pop0
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
183,693
388,194
3-91

-------
State
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
Wisconsin
Wisconsin
2010 Projected
8-hour Ozone
County Concentration (ppb)a 2000 popb
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
Door Co
Kenosha Co
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
82.1
91.0
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
27,961
149,577
2010 pop0
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
30,508
166,359
3-92

-------
2010 Projected
8-hour Ozone
State County Concentration (ppb)a 2000 popb
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%
79.9
80.0
82.1
85.8
83.9
87.7
37
148

20,187
82,887
940,164
82,317
188,831
112,646
22,724,010
58,453,962
2010 pop0
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 rule.207

       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.208  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.209
This inventory does not include the gas can emissions that are being controlled in this rule. 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. Finally, the inventory includes an error in mobile source
NOx for 13 Northeastern states. The impact of this error is minimized as the model is used in a
relative way.  Because the base years of our air quality modeling analysis are 2020 and 2030, we
                                          3-93

-------
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 final rule.210

       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.211  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 gas can controls.  Section 2.1.1.2
also details the states that have their own gas can control programs and how the controls
finalized here impact states which already  have gas can control programs.
                                          3-94

-------
      Figure 3.3-2. Map of the CAMx Domain used for MSAT Ozone Metamodeling
       The ozone RSM used for assessing the impacts of 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
154 runs (a base case plus 139 control runs plus  10 evaluation runs plus 4 boundary condition
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 gas can standards, the
                                          3-95

-------
experimental design also included oversampling 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 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 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 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.1
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-hour
1 Appendix I of 40 CFR Part 50.
                                          3-96

-------
ozone design values due to the gas can controls, when weighted by population.  The AQMTSD,
contained in the docket for this final rule, 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
ambient air.212  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 ozone AQCD notes that "ozone affects vegetation throughout the United States,
impairing crops, native vegetation, and ecosystems more than any other air pollutant."213  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.214 Once sufficient levels of ozone, a
highly reactive  substance, (or its reaction products) reaches the interior of plant cells, it can
inhibit or damage essential cellular components and functions, including enzyme activities,
lipids, and cellular membranes, disrupting the plant's osmotic (i.e., water) balance and energy
utilization patterns.215'216 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.217 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.  Furthermore, there is some 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.218

       Ozone can produce both acute and chronic injury in sensitive species depending on the
concentration level and the duration of the exposure. Ozone effects also tend to accumulate  over
the growing season of the plant, so that even lower concentrations experienced for a longer
duration have the potential to create chronic stress on sensitive vegetation. 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).219'220'221  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.222  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

                                           3-97

-------
numerous other environmental factors modify plant uptake and response to ozone, it is not
possible to identify threshold values above which ozone is consistently toxic for all plants. 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.223'224
In terms of forest productivity and ecosystem diversity, ozone may be the pollutant with the
greatest potential for regional-scale forest impacts.225  Studies have demonstrated repeatedly that
ozone concentrations commonly observed in polluted areas can have substantial impacts on plant
function.226'227

       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 impacts 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.228  In most instances, responses to chronic or
recurrent exposure in forested ecosystems are subtle and not observable for many years. These
injuries can cause stand-level forest decline in sensitive ecosystems.229'230'231  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."232 In addition, economic studies have shown
reduced economic benefits as a result of predicted reductions in crop yields associated with
observed ozone levels.233'234'235

       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. 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.236 This is therefore a potentially costly
environmental effect. However, 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.  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.
                                          3-98

-------
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 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 further described by
breaking it down into size fractions. PMio refers to particles generally less than or equal to 10
micrometers (|im) in diameter.  PM2 5 refers to fine particles, those particles generally less than
or equal to 2.5 jim in diameter. Inhalable (or "thoracic") coarse particles refer to those particles
generally greater than 2.5 jim but less than or equal to 10 jim in diameter. Ultrafine PM refers to
particles with diameters generally less than 100 nanometers (0.1  jim). Larger particles (>10 |im)
tend to be removed by the respiratory clearance mechanisms, whereas smaller particles are
deposited deeper in the lungs.

       Fine particles are produced primarily by combustion processes and by transformations of
gaseous emissions (e.g.,  SOx, NOx and VOCs) in the atmosphere. The chemical and physical
properties of PM2.5 may vary greatly with time, region,  meteorology and source category. Thus,
PM2.s, may include a complex mixture of different pollutants including sulfates, nitrates, organic
compounds, elemental carbon and metal compounds. These particles can remain in the
atmosphere for days to weeks and travel through the atmosphere hundreds to thousands of
kilometers.

       The vehicles that will be covered by the 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
will be covered by the standards also emit VOC which contribute to secondary PM2.s.  Both
types  of directly and indirectly  formed particles from vehicles and gas cans are found principally
in the fine fraction.

       EPA has recently amended the PM NAAQS (71 FR 61144, October 17, 2006). The final
rule, signed on September 21, 2006 and published on October 17, 2006,  addressed revisions to
the primary and secondary NAAQS for PM to provide increased protection of public health and
welfare, respectively. The primary PM2.5 NAAQS include a short-term (24-hour) and a long-
term (annual) standard. The level of the 24-hour PM2 5 NAAQS has been revised from 65ug/m3
to 35  ug/m3 to provide increased protection against health effects associated with short-term

                                          3-99

-------
exposures to fine particles.  The current form of the 24-hour PM2.5 standard was retained (e.g.,
based on the 98th percentile concentration averaged over three years).  The level of the annual
PM2.5 NAAQS was retained at 15ug/m3' continuing protection against health effects associated
with long-term exposures. The current form of the annual PM2.5 standard was retained as an
annual arithmetic mean averaged over three years, however, the following two aspects of the
spatial averaging criteria were narrowed: (1) the annual mean concentration at each  site shall be
within 10 percent of the spatially averaged annual mean, and (2) the daily values for each
monitoring site pair shall  yield a correlation coefficient of at least 0.9 for each calendar quarter.
With regard to the primary PMio standards, the 24-hour PMio NAAQS was retained at a level of
150 ug/m3 not to be exceeded more than once per year on average over a three-year period.
Given that the available evidence does not suggest an association between long-term exposure to
coarse particles at current ambient levels and health effects, EPA has revoked the annual PMio
standard.

       With regard to the secondary PM standards, EPA has revised these standards to be
identical in all respects  to the  revised primary standards. Specifically, EPA has revised the
current 24-hour PM2.5 secondary standard by making it identical to the revised 24-hour PM2.5
primary standard, retained the annual PM2 5 and 24-hour PMio secondary standards,  and  revoked
the annual PMio secondary standards. This suite of secondary PM standards is intended to
provide protection against PM-related public welfare effects, including visibility impairment,
effects on vegetation and  ecosystems, and material damage  and soiling.

3.4.2  Health Effects of Particulate Matter

       As stated in the  EPA Particulate Matter Air Quality  Criteria Document (PMAQCD),
available scientific findings "demonstrate well that human health outcomes are associated with
ambient PM."J We are  relying primarily on the data and conclusions in the PM AQCD and PM
staff paper, which reflects EPA's analysis of policy-relevant science from the PM AQCD,
                                                          OT7 01£
regarding the health effects associated with particulate matter.   '    We also present additional
recent studiesk published  after the cut-off date for the PM AQCD.239 Taken together this
information supports the conclusion that PM-related emissions such as those controlled in this
action are associated with adverse health effects.

3.4.2.1 Short-Term Exposure Mortality and Morbidity Studies

       As discussed in  the PM AQCD, short-term exposure to PM2.5 is associated with
premature mortality from cardiopulmonary diseases (PM AQCD, p.  8-305), hospitalization and
J Personal exposure includes contributions from many different types of particles, from many sources, and in many
different environments.  Total personal exposure to PM includes both ambient and nonambient components; and
both components may contribute to adverse health effects.
k These additional studies are included in the 2006 Provisional Assessment of Recent Studies on Health Effects of
Particulate Matter Exposure.  The provisional assessment did not and could not (given a very short timeframe)
undergo the extensive critical review by EPA, CASAC, and the public, as did the PM AQCD. The provisional
assessment found that the "new" studies expand the scientific information and provide important insights on the
relationship between PM exposure and health effects of PM. The provisional assessment also found that the "new"
studies generally strengthen the evidence that acute and chronic exposure to fine particles and acute exposure to
thoracic coarse particles are associated with health effects.
                                            3-100

-------
emergency department visits for cardiopulmonary diseases (PMAQCD, p. 9-93), increased
respiratory symptoms (PM AQCD, p. 9-46), decreased lung function (PM AQCD Table 8-34)
and physiological changes or biomarkers for cardiac changes (PM AQCD, Section 8.3.1.3.4).  In
addition, the PM AQCD 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 (PM AQCD, 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 (PM AQCD, 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 PM25 and mortality.240'241
Another recent study in 14 U.S. cities examined the effect of PMio exposures on daily hospital
admissions for cardiovascular disease. They found that the effect of PMi0 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.242  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

       Long-term exposure to elevated ambient PM2 5 is associated with mortality from
cardiopulmonary diseases and lung cancer (PM AQCD, p. 8-307), and effects on the respiratory
system such as decreased lung function or the development of chronic respiratory disease (PM
AQCD, pp. 8-313, 8-314).  Of specific importance to  this rule, the PM AQCD also notes that the
PM components of gasoline and diesel engine exhaust represent one class of hypothesized likely
important contributors to observed ambient PM-related increases in lung cancer incidence and
mortality (PM AQCD, p. 8-318).

       The PM AQCD and PM Staff Paper emphasize the results of two long-term studies, the
Six Cities and American Cancer Society (ACS) prospective cohort studies, based on several
factors - the inclusion of measured PM data, the  fact that the study populations were similar to
the general population, and the fact that these studies have undergone extensive reanalysis (PM
AQCD, p. 8-306,  Staff Paper, p.3-18).243'244'245   These studies indicate that there are significant
associations for all-cause, cardiopulmonary, and  lung cancer mortality with long-term exposure
to PM2.s. A variety of studies have been published since the completion of the AQCD.  One such
study, which was  summarized in EPA's provisional assessment, was an analysis of a subset of
the ACS  cohort data, which was published after the PM AQCD was finalized but in time for the
2006 Provisional Assessment, 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.246  EPA is
assessing the significance of this study within the context of the broader literature.

       As discussed in the PM AQCD, the morbidity studies that combine the features of cross-

                                         3-101

-------
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 (PM AQCD,
Section 8.3.3.2.3).  A variety of studies have been published since the completion of the AQCD.
One such study, which was summarized in EPA's provisional assessment, reported the results of
a cross-sectional study of outdoor PM2 5 and measures of atherosclerosis in the Los Angeles
basin.247 The study found significant associations between ambient residential PM2.5 and carotid
intima-media thickness (CEVIT), an indicator of subclinical atherosclerosis, an underlying factor
in cardiovascular disease. EPA is assessing the significance of this study within the context of
the broader literature.

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
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
                                                                     9zlS
concentrations of PM2 5 inside police cars  on North Carolina state highways.   A number of
studies of traffic-related pollution have shown associations between fine particles and adverse
respiratory outcomes in children who live  near major roadways.249'250'251  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  rule will 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.  In 2005,
EPA designated 39 nonattainment areas for the 1997 PM2.5 NAAQS based on air quality design
values (using 2001-2003 or 2002-2004 measurements) and a number of other factors.1 (70 FR
943, January  5, 2005; 70 FR 19844, April  14, 2005). These areas are comprised of 208 full or
partial counties with a total population exceeding 88 million. The 1997 PM2.5 nonattainment
areas and populations, as of October 2006, are listed in Appendix 3C to this RIA.  As mentioned
in Section 3.4.1,  the 1997 PM2.5 NAAQS was recently revised and the 2006 PM2.5 NAAQS
became effective on December 18, 2006.  Nonattainment areas will be designated with respect to
the 2006 PM2.s NAAQS in early 2010. Table 3.4-1 presents the number of counties in areas
1 The full details involved in calculating a PM2 5 design value are given in Appendix N of 40 CFR Part 50.
                                         3-102

-------
currently designated as nonattainment for the 1997 PM2.5 NAAQS as well as the number of
additional counties which have monitored data that is violating the 2006 PM2 5 NAAQS.
Table 3.4-1. PM2.s Standards: Current Nonattainment Areas and Other Violating Counties

1997PM2.5 Standards:
2006 PM2.5 Standards:
39 areas currently designated
Counties with violating monitors2
Total
Number of
Counties
208
49
257
Population1
88,394,000
18,198,676
106,592,676
1) Population numbers are from 2000 census data.
2) This table provides an estimate of the counties violating the 2006 PM25 NAAQS based on 2003-05 air quality
data. The areas designated as nonattainment for the 2006 PM2 5 NAAQS will be based on 3 years of air quality data
from later years. Also, the county numbers in the summary table includes only the counties with monitors violating
the 2006 PM2 5 NAAQS.  The monitored county violations may be an underestimate of the number of counties and
populations that will eventually be included in areas with multiple counties designated nonattainment.

       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 2010 to 2015 time frame and then be required to maintain the 1997 PM2.5
NAAQS thereafter.™  The attainment dates associated with the potential nonattainment areas
based on the 2006 PM2.5 NAAQS would likely be in the 2015 to 2020 timeframe. The emission
standards being finalized in this action will become effective between 2009 and 2015.  The
expected PM2 5 and PM2 5 precursor inventory reductions from the standards finalized in this
action will be useful to states in attaining or maintaining the PM2.5 NAAQS.

3.4.3.2  Current PMio Levels

       EPA designated PMio nonattainment areas in 1990." As of October 2006, approximately
28 million people live in the 46 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 46 full or partial counties that make up the PMio nonattainment areas.  The PMio
nonattainment areas and populations are listed in Appendix 3C to this RIA.

       As mentioned in Section 3.4.1, the 1997 PM NAAQS was recently revised and the 2006
PM NAAQS became  effective on December 18, 2006. The annual PMio NAAQS was revoked
and the 24 hour PMio NAAQS was not changed.  The projected reductions in emissions from the
controls finalized in this action will be useful to states to maintain the PMio NAAQS.
m The EPA finalized PM2 5 attainment and nonattainment areas in April 2005. The EPA proposed the PM
Implementation rule in November 2005 (70 FR 65984).
n 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.
                                           3-103

-------
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 PM NAAQS final rule, which
was promulgated by EPA in 2006.  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 PM NAAQS air quality assessment. The PM NAAQS analysis
included all final federal rules up to and including Clean Air Interstate Rule (CAIR) and all final
mobile source rule controls as of October 2006.  Details on the air quality modeling are provided
in the Regulatory Impact Analysis (RIA) for the Final PM NAAQS Rule, included in the docket
for this final rule.252

3.4.3.3.2      Areas at Risk of Future PM2.5 Violations

    Air quality modeling performed for the final PM NAAQS indicates that in the absence of
additional local, regional or national controls, there will likely continue to be counties that will
not attain some combination of the  annual 2006 PM2.5 standard (15 |ig/m3) and the daily 2006
PM2.s standard (35 |ig/m3). The PM NAAQS analysis provides estimates of future PM2.s levels
across the country.  For example, in 2015 based on emission controls currently adopted or
expected to be in place0, we project that 53 million people will live in 52 counties with projected
PM2.5 design values at and above the 2006 standard, see Table 3.4-2.p The rule will assist these
counties in attaining the PM2.5NAAQS.  Table 3.4-2 also lists the 54 counties, where 27 million
people are projected to live, with 2015 projected design values that do not violate the PM2.5
NAAQS but are within ten percent  of it. The rule may help ensure that these counties continue
to maintain their attainment status.

     Table 3.4-2. Counties with 2015 Projected Annual and Daily PM2.s Design  Values
                    Above and within 10%  of the 2006 PM2.5 Standard"
2015
Projected
Annual
PM2.5
Design
Value
State County (M9/m3)
2015
Projected
Daily
PM2.5
Design
Value 2015
(ug/m3) Population15
0 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 rule will help
these areas attain the PM standards by their statutory date.
p Note that this analysis identifies only counties projected to have a violating monitor; the number of counties to be
designated and the associated population would likely exceed these estimates.
                                           3-104

-------
Alabama
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California

California
California
California
California

California
California
California
California
California
California
California
California
California
California
Connecticut
Georgia
Georgia
Georgia
Georgia
Georgia
Georgia
Georgia
Idaho
Idaho
Idaho
Idaho
Idaho
Illinois
Illinois
Illinois
Illinois
Indiana
Indiana
Jefferson Co
Alameda Co
Butte Co
Colusa Co
Contra Costa Co
Fresno Co
Imperial Co
Inyo Co
Kern Co
Kings Co
Los Angeles Co
Merced Co
Orange Co
Placer Co
Riverside Co
Sacramento Co
San Bernardino
Co
San Diego Co
San Francisco Co
San Joaquin Co
San Luis Obispo
Co
San Mateo Co
Santa Clara Co
Solano Co
Sonoma Co
Stanislaus Co
Sutter Co
Tulare Co
Ventura Co
Yolo Co
Fairfield Co
Bibb Co
Clayton Co
DeKalb Co
Floyd Co
Fulton Co
Muscogee Co
Wilkinson Co
Ada Co
Bannock Co
Canyon Co
Power Co
Shoshone Co
Cook Co
Madison Co
St. Clair Co
Will Co
Clark Co
Lake Co
15.9
13.3
13.4
9.5
12.6
20.1
14.8
6.1
21.3
17.2
23.7
15.8
20.0
11.4
27.8
12.2
24.6
15.8
11.3
15.4
9.4
10.5
10.7
11.7
10.0
16.6
11.2
21.2
14.1
10.2
11.0
13.7
13.9
13.6
14.0
15.5
13.4
13.6
8.9
9.1
9.2
10.5
12.4
15.5
15.2
14.6
13.2
13.6
13.4
36.9
59.4
50.7
33.5
61.3
73.0
45.7
38.1
81.4
70.6
62.2
54.4
41.1
38.1
73.5
49.8
65.7
40.7
52.5
51.1
35.8
41.9
48.5
57.7
38.9
61.9
39.3
77.2
38.8
33.0
31.6
27.0
28.7
31.5
30.9
32.2
34.2
29.3
32.2
40.2
32.6
36.6
36.2
37.1
35.5
30.4
32.0
31.1
40.8
669,850
1,628,698
242,166
23,066
1,155,323
960,934
173,482
19,349
804,940
161,607
9,910,805
250,152
3,467,120
403,624
2,015,955
1,488,456
2,157,926
3,489,368
765,846
675,362
304,079
785,949
1,899,727
529,784
569,486
547,041
99,716
441,185
923,205
206,388
893,629
160,468
280,476
715,947
97,674
877,365
197,634
11,259
397,456
88,033
154,137
8,932
15,646
5,362,931
271,854
251,612
634,068
112,523
490,795
                                            3-105

-------
Indiana
Kentucky
Maryland
Maryland
Maryland
Massachusetts
Michigan
Michigan
Michigan
Michigan
Michigan
Montana
Montana
New Jersey
New Jersey
New Jersey
New York
New York
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Oregon
Oregon
Oregon
Oregon
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Tennessee
Utah
Utah
Utah
Utah
Utah
Washington
Washington
Washington
Washington
Washington
Marion Co
Jefferson Co
Anne Arundel Co
Baltimore city
Baltimore Co
Hampden Co
Kalamazoo Co
Kent Co
Oakland Co
St. Clair Co
Wayne Co
Lincoln Co
Missoula Co
Camden Co
Hudson Co
Union Co
Bronx Co
New York Co
Cuyahoga Co
Franklin  Co
Hamilton Co
Jefferson Co
Lucas Co
Scioto Co
Trumbull Co
Jackson Co
Klamath Co
Lane Co
Washington Co
Allegheny Co
Beaver Co
Berks Co
Dauphin Co
Lancaster Co
Lehigh Co
Mercer Co
Northampton Co
Philadelphia Co
York Co
Knox Co
Box Elder Co
Cache Co
Salt Lake Co
Utah Co
Weber Co
Clark Co
King Co
Pierce Co
Snohomish Co
Thurston Co
13.5
13.8
11.1
13.0
11.3
11.6
12.8
12.0
13.0
12.5
17.4
15.0
10.6
11.1
12.0
12.2
12.8
14.0
15.4
13.7
14.3
14.2
12.5
15.6
12.1
10.9
10.1
12.9
9.0
16.5
12.1
12.0
11.0
12.2
10.5
11.0
10.9
13.3
12.3
13.6
8.6
12.5
12.6
9.3
9.1
9.2
10.8
11.1
11.3
8.9
33.1
33.4
33.2
35.5
32.6
32.9
32.7
31.9
33.2
32.5
39.0
42.4
32.1
32.1
32.8
32.8
33.2
33.2
40.0
33.5
34.2
34.2
32.2
34.3
34.2
37.6
39.1
53.6
32.0
53.4
33.2
35.5
33.3
33.7
34.7
31.6
35.0
35.2
35.9
29.6
39.0
51.9
49.3
36.7
36.2
34.3
34.0
43.0
40.1
34.9
889,645
710,231
574,322
596,076
810,172
452,055
257,817
654,449
1,355,670
185,970
1,921,253
19,875
118,303
512,135
604,036
525,096
1,283,316
1,551,641
1,325,507
1,181,578
841,858
68,909
443,230
81,013
227,546
250,169
69,423
387,237
639,839
1,245,917
184,648
396,410
272,748
535,622
328,523
123,577
286,838
1,372,037
417,408
448,931
49,878
114,729
1,133,410
508,106
229,807
479,002
2,013,808
879,363
782,319
264,364
                                         3-106

-------
Washington Yakima Co
West Virginia Berkeley Co
West Virginia Hancock Co
West Virginia Kanawha Co
Wisconsin Milwaukee Co
Wisconsin Waukesha Co
Wyoming Sheridan Co
Number of Violating Counties
Population of Violating Counties
Number of Counties within 10%
Population of Counties within 10%
9.6
12.0
13.4
13.9
12.1
11.8
10.5




34.9
32.7
32.7
28.9
32.1
32.4
31.8
52

54

261,452
99,349
30,857
196,498
908,336
441,482
28,623

53,468,515

26,896,926
a) Bolded concentrations indicate levels above the PM2 5 standard.
b) 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 Impairment

       Visibility can be defined as the degree to which the atmosphere is transparent to visible
light.253 Visibility impairment manifests in two principal ways:  as local visibility impairment
and as regional haze.q  Local visibility impairment may take the form of a localized plume, a
band or layer of discoloration appearing well above the terrain as a result from complex local
meteorological conditions. Alternatively, local visibility impairment may manifest as an urban
haze, sometimes referred to as a "brown cloud." This urban haze is largely caused by emissions
from multiple sources in the urban areas and is not typically attributable to only one nearby
source or to long-range transport.  The second type of visibility impairment, regional haze,
usually results from multiple pollution sources spread over a large geographic region. Regional
haze can impair visibility over large regions and across states.

       Visibility is important because it has direct significance to people's enjoyment of daily
activities in all parts of the country. Individuals value good visibility  for the well-being it
provides them directly, 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,  and special emphasis is given to protecting  visibility in these areas.  For
more information on visibility see the 2004 PMAQCD as well as the 2005 PM Staff Paper.254'255

       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 which would
q See discussion in U.S. EPA, National Ambient Air Quality Standards for Paniculate Matter; Proposed Rule;
January 17, 2006, Vol71 p 2676. This information is available electronically at http://epa.gov/fedrgstr/EP A-
AIR/2006/Januarv/DaY-17/al77.pdf.
                                           3-107

-------
act in conjunction with the establishment of a regional haze program. In setting this secondary
standard, EPA concluded that PM2 5 causes adverse effects on visibility in various locations,
depending on PM concentrations and factors such as chemical composition and average relative
humidity. The secondary (welfare-based) PM2.5 NAAQS was established as equal to the suite of
primary (health-based) NAAQS. Furthermore, Section 169A of the Act provides additional
authority to address existing visibility impairment and prevent future visibility impairment in the
156 national parks, forests and wilderness areas categorized as mandatory class I federal areas
(62 FR 38680-81, July 18, 1997).r In July 1999 the regional haze rule (64 FR 35714) was put in
place to protect the visibility in mandatory class I federal areas.  Visibility can be said to be
impaired in both PM2 5 nonattainment areas  and mandatory class I federal areas.

       Data showing PM2.5 nonattainment areas and visibility levels above background at the
Mandatory Class I Federal Areas demonstrate that 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.256

       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 3C.  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.257'258 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.259

       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 ensure no
degradation  occurs on the cleanest days (20% least impaired days). Although there have been
   r 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.
                                          3-108

-------
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. 26°

3.4.4.1.3       Future Visibility Impairment

       Recent modeling for the final PM NAAQS rule was used to project PM2.5 levels in the
U.S. in 2015. The results suggest that PM2.5 levels above the 2006 NAAQS will persist in the
future. We predicted that in 2015, there will be 52 counties with a population of 53 million
where annual PM25 levels will exceed the 2006 PM25 NAAQS, 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.s levels above the annual 1997 PM2 5 NAAQS. Thus, the emissions from these sources
contribute to the current and anticipated visibility impairment and the emission reductions
finalized here may help improve future visibility impairment.

3.4.4.1.4      Future Visibility Impairment at Mandatory Class I Federal Areas

       Achieving the 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 the PM2.5 NAAQS 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 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 169A and 169B  of the CAA.

       Recent modeling for CAIR was also used to project visibility conditions in mandatory
class I federal areas across the country in 2015.  The results for the mandatory class I federal
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 3D it is projected that there will continue to be mandatory
class I federal areas with visibility levels above background in 2015.261  Additional emission
reductions will be needed from the broad set of sources that contribute, including the vehicles
and gas cans subject to this rule. The reductions being finalized in this action are a part of the

                                          3-109

-------
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.262

       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 those related to both nutrient
and toxic inputs. Adverse effects to human health and welfare can occur from the addition of
excess particulate nitrate nutrient enrichment which  contributes to toxic algae blooms and zones
of depleted oxygen, which can lead to fish kills, frequently in coastal waters.  Particles
contaminated with heavy metals or other toxins may lead to the 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.263'264'265'266'267

       Adverse impacts on soil chemistry and plant life have been observed for areas heavily
impacted by atmospheric deposition of nutrients, metals and acid species, resulting in species
shifts, loss of biodiversity, 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.
                                          3-110

-------
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 (PM AQCD, p. 4-87).268 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 (PM AQCD, 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 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 (PM AQCD, p. 4-76).269 Contamination of
plant leaves by heavy metals can lead to elevated soil levels. Some trace metals absorbed into
the plant and  can bind to the leaf tissue (PM AQCD, p. 4-75). When these leaves fall and
decompose, the heavy metals are transferred into the soil.270'271

       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 a portion of emitted mercury to travel
far  from the primary source before being deposited and accumulating in the aquatic ecosystem.
Localized or regional impacts are also observed for mercury emitted from combustion sources.
The major source of mercury in the Great Lakes is from atmospheric deposition, accounting for
                                                          r)Hr) T71
approximately eighty percent of the mercury in Lake Michigan.   '    Over fifty percent of the
mercury in the Chesapeake Bay has been attributed to atmospheric deposition.274 Overall, the
National Science and Technology Council (NSTC, 1999) identifies atmospheric deposition as the
primary source of mercury to aquatic systems. Forty-four 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.275'276 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
elevated levels along roadsides.277 Plant uptake of platinum has been observed at these
locations.
                                         3-111

-------
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.278 Polycyclic aromatic hydrocarbons (PAHs) are a class of POM that
contains 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 jam 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.279

       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.280'281  Analyses of PAH deposition to Chesapeake and
Galveston Bay indicate that dry deposition and gas exchange from the atmosphere to the surface
water predominate.282'283  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.284 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.285 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.286  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.287

       Cousins et al. (1999) estimates that greater than ninety percent of semi-volatile organic
compound (SVOC) emissions in the United Kingdom deposit on soil.288 An analysis  of
polycyclic aromatic hydrocarbon (PAH) concentrations near a Czechoslovakian roadway
indicated that concentrations were thirty times greater than background.289

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.290 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

                                         3-112

-------
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
conditions. Pollutants measured with elevated concentrations include benzene, polycyclic
aromatic hydrocarbons, carbon monoxide, nitrogen dioxide, black carbon, and coarse, fine, and
ultrafme particles. In addition, resuspended road dust, and wear particles from tire and brake use
also show concentration increases in proximity of major roadways.

       As noted in section 3.2, HAPEM6 estimates the changes in time-weighted exposures
associated with proximity to roadways for individual pollutants.  The studies discussed in this
section address exposures and health effects that are at least partially captured by our modeling,
but there may be additional exposures and health effects associated with pollutants, singly or in
combination, that are not explicitly quantified. However, because the studies discussed in this
section often employ exposure estimation metrics associated with multiple pollutants, exposure-
response information from these studies may not be suitable for risk assessment geared around
one or several chemicals.

       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

                                          3-113

-------
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 discussion below addresses the
studies in detail. However, in general terms, 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
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

       The quantifiable effects of this rule on premature mortality  associated with  exposure to
PM2.5 are assessed as part of the benefits estimates for this rule. In addition to studies that have
documented the relationship between ambient 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 ambient 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.
                                          3-114

-------
       A total of three cohort studies have examined premature mortality in relation to residence
near traffic, another examined county-level traffic density, 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 three recent cohort studies for all-cause and cardiopulmonary mortality from the
Netherlands, Ontario,  Canada, and most recently, Germany.291'292'293 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.294

       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).
A subsequent follow-up  study found elevated mortality rates from circulatory causes in the
Canadian study population.

       Most recently, German investigators followed up a series of cross-sectional studies on
women age 50-59 living in the North Rhine-Westphalia region during the late 1980's and
1990's, tracking vital status and migration to the years 2002-2003.295 In total, the cohort
consisted of approximately 4800 women. Exposures were categorized using ambient NO2 and
PMio (estimated from TSP), and an indicator of residence within 50 m of a "major road", defined
at > 10,000 cars/day. Overall, living within 50 meters of a major road was associated with a
significant 70% increase in the rate of cardiopulmonary mortality. Nearest-monitor NO2 and
PMio were also associated with a 57% and 34% increase in the rate of cardiopulmonary
mortality. Exposure to NO2 was also associated with a 17% increase in all-cause mortality.

       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.

       In another study evaluating a cohort of older, hypertensive male U.S. veterans, county-
level traffic index and pollution estimates were employed in estimating exposure to traffic
activity and other air pollutants.296 Area-based traffic density was significantly associated with
increased mortality rates, as were constituents of motor vehicle exhaust,  such as elemental
carbon.
                                          3-115

-------
       One cohort study conducted in the United Kingdom examined cardiocerebral (stroke)
mortality in relation to living near traffic.297 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.298 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
the European literature.299 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.300  Infants with at least one atopic parent qualified for
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.

                                          3-116

-------
       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.301 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.302  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.303 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.304 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 rule.  Taken together, these studies
provide 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,
primarily generated in European studies, was recently published.305 Overall, the review
concluded that there is some limited evidence of an association between traffic-generated
pollutants and asthma incidence. More recent studies have also found significant associations
between prevalent asthma and living near major roads.306  Toxicological evidence provides some
evidence that particles from diesel engine exhaust may serve as adjuvants to IgE-mediated

                                          3-117

-------
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 set of mechanisms
whereby PM2 5 may be leading to premature mortality. In Chapter 12, we estimate the
quantifiable benefits of PM-related non-fatal acute myocardial infarction and cardiovascular
hospital admissions. The studies described in Section 3.5.1 found higher relative risks for
cardiopulmonary causes of death.

       In addition to cardiopulmonary mortality, some studies have looked at morbidity. A
recent study from Germany also found significant increased odds of coronary heart disease
(CHD) in a cohort of approximately 3400 participants.307 Residents living within 150 meters of
major roads were compared to those living further ways.  Overall, controlling for background air
pollution and individual risk factors, the adjusted odds ratio for CHD prevalence was
significantly elevated (1.85).  Subgroup analyses indicated stronger effects in men, in
participants under 60 year of age, and in never-smokers.

       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.308  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.309
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.
    310
       Heart rate variability has also been measured in a study of elderly residents of the Boston
area.    In the study, ambient PM2.5 was associated with changes consistent with reduced

                                          3-118

-------
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.311 These studies further document a hypothesized mechanism associated with
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

       A few studies examining birth outcomes in populations living near major traffic sources
have 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.312 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.313 In another study, preterm birth was
associated with  ambient PMio and CO.314  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.315 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.316  Subsequent follow-up of the
same birth cohort to age three found evidence of neurodevelopmental deficits associated with
maternal exposure to PAHs during pregnancy, particularly in cognitive development.317

       Overall, although the number of studies examining perinatal exposures is 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
                                          3-119

-------
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
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 for compounds with a mutagenic mode of action to
account for early life exposures in the Supplemental Guidance for Assessing Susceptibility from
Early-Life Exposure to Carcinogens.318

       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.319

       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.320  The study was
reanalyzed using an approach to combine traffic volume with residential distance from major
roads to assess "distance-weighted traffic volume."321 The study found that the significant,
monotonically increasing risks associated with increased distance-weighted traffic volume.

       NC>2 has been used  as an indicator of traffic emissions in some studies; however, it is
important to note that NC>2 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
        IT)
Sweden.    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

                                         3-120

-------
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.323  A large  study of nearly 2,000 Danish children with cancer found no
association between modeled  concentrations of benzene and NO2 at home and the risk of
leukemia, central nervous system tumors, or total cancers.324 However, the study did find a
dose-dependent relationship between Hodgkin's disease and modeled air pollution from traffic.

       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.325 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.326

       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.327 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, PMio, 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 standards will  reduce a broad range of pollutants present in
higher concentrations near roadways. It is unclear to what extent these health effects are
attributable to PM versus other components of the complex mixture. Note that the benefits
associated with the direct PM reductions from the cold temperature vehicle standards are
presented in Chapter 12 of this RIA.
                                          3-121

-------
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. A study of the
populations nationally using geographic information systems indicated that more than half of the
population  lives within 200 meters of a major road (see file USbytract.txt in the docket for this
rule).8 It should be noted that this analysis relied on the Census Bureau definition of a major
road, which is not based on traffic volume. Thus, some of the roads designated as
"major" may carry a low volume of traffic. Detailed analyses of data were conducted in three
states, Colorado, Georgia, and New York. 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%.328

       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.329  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.330  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.331  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
       s Major roads are defined as those roads defined by the U.S. Census as one of the following: "limited
access highway," "highway," "major road (primary, secondary and connecting roads)," or "ramp."
                                          3-122

-------
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.332 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.333 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.
                                          3-123

-------
     Appendix 3A: Influence of Emissions in Attached Garages on
       Indoor Air Benzene Concentrations and Human Exposure

Introduction

       Measurement studies provide strong evidence that VOC sources in attached garages can
significantly increase VOC concentrations inside homes.334 Preliminary analyses of data from a
pilot study for the National Human Exposure Assessment Survey (NHEXAS) in Arizona also
found indoor concentrations of mobile source-related VOC compounds significantly higher in
homes with attached garages than in homes without them.335 This population-based exposure
study included measurements from 187 homes. A study in 50 Alaska residences found that in
homes with attached garages, indoor benzene levels averaged 70.8 ng/m3, while in homes
without attached garages, concentrations averaged 8.6 |j,g/m3.336 Multiple factors, including
house architecture, ventilation design, garage configuration, and climate can all play roles as
well.
       National-scale air toxics modeling efforts, such as those discussed in RIA Section 3.2.1.2,
employ Gaussian dispersion models in combination with human exposure models to calculate the
concentrations of air toxics in various microenvironments. Exposure models calculate an
average exposure resulting from the movement of a simulated population through a time-activity
pattern that brings them into contact with air in the various microenvironments.
       At this point, the NATA and the analyses performed for this rulemaking have only
included exposures from outdoor sources.  Although the HAPEM6 exposure model is capable of
addressing indoor sources, more thorough analyses of the prevalence and use of emission sources
within attached garages are required to develop quantitative estimates of model parameters to
address attached garage contributions across the U.S. population.
       This appendix addresses the potential impact of all benzene sources within an attached
garage on residential indoor air quality.

Methods

Calculation ofWithin-garage Source Emission Rate

       Emission rates for indoor sources of VOCs can be derived by several methods. Most
accurately, the actual emission rates of an indoor VOC source can be measured through the use
of a Sealed Housing for Evaporative Determination (SHED). However, test conditions must  be
representative of real world applications.  Short of SHED-based measurement, several surrogate
approaches may be employed. For evaporative losses from a sealed container, the change in
weight of a container over time may be used to calculate a total mass loss rate, which can be
assumed to be in the form of VOC. Alternatively, if the  air concentrations and ventilation
conditions of a defined indoor space are known, mass balance equations can be employed to
derive a "virtual" emission rate for all sources within the space.
       This appendix employs the latter approach in calculating source emission factors. The
general approach of a mass balance equation is to calculate the change in mass over a given time,
accounting  for the mass of a pollutant transported into a  space, the mass of pollutant transported
out of a space, the emission rate  of a source within the space, and the decay of any pollutants

                                        3-124

-------
within the space, which can be treated as a first-order decay. A simple space like a garage can be
treated as a single zone. The differential equation representing this mass balance is as follows:
                                    ^^ = ckd_L+^_cjv_
                                     dt      °  dt     dt     '  dt
Here, dMt/dt represents the rate of change of total indoor mass, Q is the indoor concentration,
C0 is the outdoor concentration, dV/dt is the volumetric air flow through the space, k is the
penetration fraction indicating the proportion of mass that passes through the wall of the
compartment, and dM/dt represents the mass emission rate inside the space.  Note that all air
entering the garage is assumed to enter from outdoors.
       Assuming steady-state conditions, dMt;i/dt assumes the value of zero, meaning that the
concentration in the garage does not change over time. Algebraically, this allows the equation
above to be represented as:

                             (2)    ffk-C.*)^
                                    \dt j             dt
In other words, the indoor source terms can be calculated if the volumetric flow through the
space and concentrations indoor and outdoor are known. Any gradient in concentration between
indoor and outdoor concentrations is explained by indoor sources and the fraction of mass that
does not penetrate from indoors to outdoors.
       The volumetric flow can be calculated by multiplying the volume of the space by the
number of times per hour that the air within the space is turned over. As such:

                                    (3)     ^--aV
                                           dt
Here, a is the "air exchange rate," expressed in air changes per hour (ACH).  Combining
equations (2) and (3), the mass emission rate is represented as:

                              (4)    av(C,-C0k)-^-

       A recent study in Ann Arbor, MI measured the air exchange rates and the in-garage and
outdoor concentrations of VOCs needed to perform these calculations.337 The homes in the
study were based on a convenience sample, and so may not be generally representative of the
local or national housing stock.  All garages but one adjoined a house.  All attached garages had
between one and three walls adjoining a residence. The distributions of garage benzene
concentration and ACH are shown in Figure 3 A-l. The distributions of each were not
significantly different from lognormal, judging by the Kolmogorov-Smirnov Z statistic.
       Values of k, the penetration factor, are dependent on the physical pathways through
which air passes into  a garage, as well as the presence and  chemical composition of any
insulating material through which air passes.  In the case of garages, the infrequency of insulated
garages and the low reactivity of benzene justifies the assumption that k=l.338
       These data from the Ann Arbor, MI study were used to solve equation (2) to derive  a
distribution of benzene mass emission rates in each garage in the study, based on variability in
measurements of outdoor concentrations.  Equation 4 was implemented using a Microsoft Excel
spreadsheet with the @Risk probabilistic simulation add-in (version 4.5).339  Monte Carlo
sampling was used for all terms in deriving the emission rates.
       As described below, this distribution can be used to evaluate the effect of various fuel
control measures on indoor benzene concentrations. A single lognormal distribution was used to

                                         3-125

-------
represent C0 in equation 4, based on other studies of ambient air, which have found that many
pollutants' concentrations are lognormally distributed.

Calculation of Garage Contributions to Indoor Air

       In the same way that a mass balance calculation can be used to calculate emission rates
for sources within garages, a mass balance equation can be used to estimate the additional
concentration in a home that will occur as a result of elevated concentrations in the garage.
However, unlike the garage case, it is not valid to assume that all air entering the home comes
directly from outdoors.
       Recent studies have provided indications that over multiple sequential days, variability in
within-home benzene concentration is relatively small. A recent study from Ann Arbor, MI
found a coefficient of variation (COV) of 4.6% for benzene.340  Furthermore, recent data
obtained by EPA through the Environmental and Occupational Health Sciences Institute
(EOHSI) on homes  in the Elizabeth, NJ area indicates no significant differences in within-home
concentrations at a 95% confidence level.1'341 These data are preliminary, and analyses are still
in progress.
       Given the fraction of air entering the home through the home-garage interface, the
appropriate mass balance equation for a single-compartment (e.g. well-mixed) home can be
represented as such:
                                               r.dV         dV     dV
Here, Q is the in-house concentration, C0 is the outdoor concentration, Cg is the concentration in
the garage, dV/dt is the volumetric air flow through the house, and^is the fraction of air entering
the home from the garage. One assumption made here is that the penetration factor for the air
moving through the house-garage interface is the same as air moving through the house-outdoors
interface.  Reactive decay is assumed to be zero. Such mass balance equations are standard
approaches in environmental science and engineering, and are frequently found in textbooks on
these subjects.342
       Again assuming steady-state conditions, dM^/dt = 0, the equation above simplifies to:

                          (6)            Cl=kC0(l-fg) + kCgfg
Or more simply, the indoor concentration under steady state conditions is proportional to the
fraction of air entering the house through the garage.
       Figure 3A-2 is a contour plot illustrating the range of average indoor air concentrations
that could plausibly arise given  a range of values of Cg and^g, with a background concentration
of zero. However, Figure 3A-2 does not answer the question of what the likely indoor air values
are in a sample of real homes.
       The text below describes procedures and results of a small-scale modeling study.

Modeling Approach
' In that study, one air sample was obtained in the room adjacent to an attached garage in each home and another was
obtained in another location. The New Jersey Department of Environmental Protection and EPA provided joint
funding for the study. A two-sided paired t-test was applied to data obtained from 36 homes over approximately 24
hours.
                                          3-126

-------
       All modeling analyses employed Equation 6 in a Microsoft Excel spreadsheet with the
@Risk probabilistic modeling add-in was the software employed in all modeling analyses.
Where appropriate, each of the terms in Equation 6 was treated as a random variable represented
as either a parametric distribution or as an empirical distribution based on measured data.
       Often, in employing data obtained from more than one study, combining data into a
single distribution was not justified on a priori grounds.  In ventilation studies, ambient
conditions such as temperature and geography can substantially affect air flow patterns and
building constructions.  For instance, residential air exchange rates differ significantly between
regions with substantially different climates.343 Furthermore, based on the limited number of
studies available, combining data from multiple studies into a single data set had the potential to
apply de facto weights to data, potentially shifting the fitted model parameters away from truly
"representative" distributions.
       Another consideration is the potential for independence of the^ and Cg variables.  There
is no a priori reason why the "leakiness"  of the house-garage envelope should be related to the
concentration of benzene in the garage.
       Because of these considerations, data onfg or Cg from studies in different areas were not
formally combined.  Rather, distributions fit separately to data from each study were used to
develop several model "scenarios." As described below, four different studies provided data for
Cg and three different studies provided data forfg. As such, a minimum of 12 (3 x 4)  scenarios
were needed to represent the totality of available data.
       For each scenario modeled, @Risk sampled from each distribution 20,000 times using a
proprietary Latin Hypercube sampling framework.  The large number of samples and Latin
Hypercube strategy were employed to ensure that modeled concentration distributions achieved
stability.
       Lastly, for comparison to the current approaches for exposure modeling, the following
equation was used, paralleling the approach taken by HAPEM5 with no garage emissions:
                                 (7)            Ct  = kC0

Data for Populating Model Parameters

Fraction of Air Entering Home through the Garage (fg)

       Several studies have examined the fraction of air entering the home from the garage.
Except for one,  all of these studies took place in northern states and Canada, where homes are
built with more insulation. A recent study of a set of homes in  Ontario, Canada found that
approximately 13% of the air entering the home came from the garage.344 One study from
Minnesota found that in newer homes, houses built in the year  1994 had an average of 17.4% of
total air leakage coming through their garages, houses build in  1998 had an average leakage
fraction of 10.5%, and houses built in 2000 had an average leakage fraction of 9.4%.345 Two
recent studies have employed perfluorocarbon tracer (PFT) gases to estimate air transport
between different "zones" of houses with attached garages. A recent study by Isbell et al. (2005)
based in Fairbanks, Alaska found that in a modern air-tight Alaskan home ventilated with an air-
to-air heat exchanger, 12.2% of the air entering a home entered through the garage, while 47.4%
of the air entering an older home ventilated passively by structural defects came through the
attached garage.346 Another study of a home in Ann Arbor, Michigan built in 1962 found that

                                          3-127

-------
16% of the air entering the home originated in the garage.347 In a more recent study from Ann
Arbor, investigators deployed PFT tracers in 15 homes and calculated the fraction of air entering
each home through an attached garage, with an average of 6.5±5.3% of the air entering through
the garage.348 From these studies, it is apparent that across homes, the fraction of air entering
through the garage is highly variable, making it necessary to acknowledge significant
uncertainties in characterizing "typical" infiltration patterns.

Benzene Concentrations in Garage Air (Cg)

       Four sources of in-garage concentration data are available in the format relevant for
steady-state modeling over extended periods of time.  First, there is the study by Batterman et al.
(2005), in which average garage concentrations of benzene were measured over a period of four
days in each of 15 homes using passive sampling badges.  The average garage concentration
reported was 36.6 |J,g/m3, with a standard deviation of 38.5 |J,g/m3.
       Second, a study in Alaska by George et al. (2002) measured benzene concentrations in 28
Alaska homes and 48 garages with passive diffusion badges.349 One disadvantage of this study
is the relatively high detection limit for benzene, 7 ppb (22 ng/m3). As a result, many of the data
available are based on a reported value of 50% of the  detection limit.  In the Alaska study, in-
garage benzene concentrations averaged 103 ng/rn3, and the standard deviation was 135 ng/rn3.
The study included concurrent in-home measurement of benzene in homes with attached
garages, allowing evaluation of the modeled indoor concentrations. However, it is not apparent
that this study underwent scientific peer review.
       A third study in one New Jersey home also evaluated garage and indoor benzene, as part
of an investigation into in-garage emissions of vehicles fueled  with methanol blends.350  Only
one home was sampled, but it was sampled multiple times inside the garage and at multiple
locations inside the residence.  A fourth study from Fairbanks, Alaska conducted measurements
in  12-hour periods on four separate days in two houses in two seasons, summer and winter.351
The study obtained two daily measurements of benzene concentration within each garage over a
12-hour sampling period.  One home was a modern, well-insulated home with an air-to-air heat
exchanger for ventilation. The other was an older home ventilated passively by structural  defects
in the building envelope. Because of the large differences in concentrations between homes and
seasons, data  from each home-season combination was treated as a separate distribution within
the indoor air model (Equation 6 in Excel/@Risk).  Treating these data as separate distributions
increased the  number of modeled "scenarios" to 21  (3fgx7 Cg).
                                         3-128

-------
Penetration Factor (k)

       The values of k in this case were obtained from the HAPEM5 user's manual, using the
PEN-1 factor, representing the fraction of benzene from outdoor air penetrating indoors.  The
values in HAPEM5 are presented as a distribution that assigns a 2/3 weight to the value 0.8 and a
1/3 weight to the value 1. These estimates are based on a comprehensive review of indoor and
outdoor air quality studies.

Outdoor Ambient Concentration (C0)

       The 1999 National Air Toxics Assessment (NATA) provided ambient concentration
estimates for every census tract in the U.S. For this modeling exercise, a lognormal distribution
was fit to these data.

Results

Within-Garage Emission Rates

       Equation 4 was used with Monte Carlo sampling to calculate a distribution of emission
factors for each home, based on the variability in outdoor concentrations reported in Batterman
et al. (2005). As shown in Figure 3-A3, the within-garage variation was a very small component
of overall variability compared to between-garage variation.  This finding implies that the factors
in individual garages, such as storage of vehicles, nonroad equipment, and fuels,  have a major
effect on in-garage concentrations.
       In aggregate, the mean emission rate for all garages sampled fell along a lognormal
distribution (p > 0.05). The  mean emission rate was 3049 ng/hr (73 mg/day), with a standard
deviation of 4220 ng/hr (101 mg/day).
       To evaluate the plausibility of these steady-state emission factors, known emission
factors for other emission sources were evaluated.  The California Air Resources Board (CARB)
conducted a study of emissions from portable fuel containers, finding that volume-specific
emissions rates for total VOC due to evaporation and permeation was 0.37 g/gal-day. Assuming
an average fuel container volume of two gallons, the average emission factor per can would be
0.74 g VOC/day.
       To evaluate the derived emission rates relative to CARB's measurements, a benzene fuel
vapor pressure fraction of 0.5-1% was assumed, based on MOBILE6.2 evaporative emission
factors.  Given that assumption, the average benzene emission rate from CARB's study is 3.7-7.4
mg/day.  This value is in the lower range of emission rates shown in Figure 3A-3. This
comparison suggests that emissions due to permeation and evaporation from portable fuel
containers may be a relatively  small fraction of overall garage benzene.
       Subsequently, one additional study used perfluorocarbon tracers (PFT) and VOC
measurement in two Fairbanks, Alaska homes to estimate two garages' "source strengths" for
benzene.352 For a new, energy efficient "tight" home with an air-to-air heat exchanger, median
garage emission estimates for benzene were 21 mg/h in summer and 14 mg/h in winter. In an
older home with passive ventilation due to structural defects, median benzene source strengths
were calculated at 40 and 22 mg/h in summer and winter, respectively. These values are
substantially higher than those calculated based on Batterman et al. (2005). However, the

                                         3-129

-------
difference may be attributable to higher fuel benzene in Fairbanks than in Michigan.  Study
design may also play a key role. In the Fairbanks study, the measurement periods were 12 hours
each in duration. In the Michigan study, measurement periods lasted four days each. The
Michigan study's longer duration may have allowed for a broader range of emissions activities
than the Fairbanks study.

Garage Contributions to Benzene in Indoor Air

       Figures 3A-4 to 3A-8 display the results of @Risk simulations of indoor air. Each figure
represents the modeled outputs  as cumulative probability distributions. In the legend of each
figure, the label of each distribution describes its^ and Cg sources. For instance, "George et al.
(2002) / Fugler FG Ci" indicates a distribution using garage concentration data from George et
al. (2002) andfg data from Fugler et al.
       Figure 3A-4 presents the output of Equation 6, a daily average indoor benzene
concentration including contributions from outdoor air and from attached garages.  As noted in
the "Methods" section of this appendix, it was necessary to run a large number of scenarios to
account for different combinations  offg and Cg data sources. The figure depicts results using
studies that contain Cg data from multiple homes as bold solid lines, while the model simulations
based on studies that employ Cg data from only one home are shown in dashed lines. As
indicated in the figure, there is no major difference in the Ct distributions predicted by using Cg
data from multiple homes or by using Cg measured from a single home. The average modeled
indoor benzene concentrations ranged from 2.9 to 16.4 |j,g/m3.
       For comparison, Figure  3A-5 presents cumulative distributions of the observed results
from several studies that measured indoor air concentrations in homes with attached garages.
Schlapia and Morris (1998) measured integrated 24-hour benzene concentrations inside 91
homes with attached garages in Anchorage, Alaska between 1994 and 1996.353 George et al.
(2002) reported average benzene concentrations in 36 homes in Anchorage,  Alaska, but no
distributional data.  Mentioned above, Isbell et al. (2005) also measured integrated 12-hour
benzene in two seasons in one modern air-tight home ("Home V" in Figures 3 A-4 to 3 A-8) and
one older passively-ventilated home ("Home NV" in Figures 3A-4 to 3A-8).354 Both homes
were located in Fairbanks, Alaska.  Batterman et al. (2006) measured indoor air benzene
concentrations in 15 homes in southeastern Michigan over four-day  sampling periods throughout
spring and summer of 2005.355  Lastly, Weisel (2006) conducted a study of indoor air in 21
homes in Union County, NJ between April 2005 and January 2006.  One monitor in each home
was sited in the room adjacent to the garage, while another was located in another part of the
house.356
       Comparing Figures 3 A-4 and 3A-5, it is apparent that the distributions of modeled indoor
air concentrations of benzene are very similar to those observed in monitoring studies. Both
figures indicate that there  is substantial variability in concentrations between homes and between
studies.
       Figure 3 A-6 presents the mean concentrations from modeling scenarios and from
monitoring studies.  In general,  the range of mean concentrations is close to  the values monitored
in the indoor air studies. Notable exceptions are the indoor air values by George et al. (2002),
the winter data from the passively-ventilated "NV" home from Isbell et al. (2005), and by
Schlapia and Morris (1998).  All of these studies took place in Alaska, which may have uniquely
high benzene fuel levels or housing architectures that create higher garage air infiltration indoors.

                                         3-130

-------
Of particular note, all of these studies included substantial numbers of homes with "tuck-under"
garages where one or more rooms of a house are situated above a garage. Schlapia and Morris
(1998) reported a very high average value that was not matched by the "average" conditions of
any other run.  It is notable that this high value is the average across 91 homes with attached
garages.
       Another consistent trend shown in Figure 3 A-6 is that scenarios employing^ data from
Batterman et al. (2006) produced consistently lower average benzene concentrations than
scenarios employing other sources.  This trend is attributable to the lower average fg reported in
Batterman et al. (2006), 6.5%, as compared to values found in Sheltersource (11.7%) and Fugler
etal. (13.6%).
       It is unclear whether the studies measuring Cg,fg, and Ct constitute a representative
sample of homes. In general Alaskan studies report higher concentrations, but not consistently.
The relatively greater prevalence of homes with "tuck under" garages in some Alaskan studies
may explain this discrepancy.
       In comparison to the values reported in Figures 3 A-4 and 3 A-5, indoor air concentrations
calculated with the default d = kC0 approach, similar to that employed in the national-scale
modeling for this rule, averaged 1.2 ng/m3.
       Overall, modeled concentrations presented here appear to provide a  credible estimate of
indoor benzene concentration in homes with attached garages. However, it is unclear whether
the homes included in the studies employed herein may be considered "representative."

Implications

Effect on Exposures Nationwide

       In calculating the hypothetical effect of attached garage on national  estimates of chronic,
time-weighted average  (TWA) human exposure, precise estimates are not possible. As noted
previously, the extent to which available studies of indoor air of homes with attached garages is
representative of the entire population of such homes is unclear.  Furthermore, the distribution of
housing stock by climate and meteorology is not well understood.  However, despite these
limitations, a bounding exercise is still feasible.
       One simple approach for such a bounding exercise is determined by the following
equation:
              (8)   ES = C1S*P*TS
              \  /      &     l'&   &  &
       Here, Eg represents the national average exposure to benzene in air attributable to
attached garages.  Ciig represents the average indoor concentration attributable to an attached
garage, Pg represents the fraction of the population living in a home with an attached garage, and
Tg represents the time spent in a home with an attached garage.
       C!;gis derived from Equation 6, and can be derived by setting the outdoor concentration
term (C0) to zero. An estimate of the  attached garage contribution to indoor air can be made for
studies with only indoor measurements as well. This can be accomplished by substituting
ASPEN concentration estimates for the county in which each study took place. For Equation 6,
C0 estimates from NATA for each census tract in the relevant county were assembled into a
lognormal distribution.  With this data and the other assumptions of Equation 6, an estimate of
d:g could be derived from the measurement studies.
                                          3-131

-------
       To estimate Pg, an estimate of the national fraction of homes with attached garages is
required. The Residential Energy Consumption Survey (RECS), run by the U.S. Energy
Information Administration, provides an estimate of the fraction of homes with attached
garages.357 RECS estimates a total of 107.0 million housing units nationally, 37.1 million
(34.7%) of which are homes with attached garages. Assuming that the population is uniformly
distributed across housing units allows this figure to serve as an estimate ofPg.
       Information on the fraction of time spent in a residence (Tg) is required to determine how
the microenvironmental concentration in homes with attached garages affects overall time-
weighted exposure concentrations. As cited in EPA's Exposure Factors Handbook, the average
person studied by the National Human Activity Pattern Survey (NHAPS) spent 1001.39 minutes
(16.68 hours) per day indoors within any room of a residence.358
       Results of model simulations using Equation 7 are shown in Figure 3A-7. As before, the
results of each combination of Cg andfg data source are shown.  For each study, the legend lists
the source for both Cg andfg data. As described above, the estimates C!;g derived from indoor air
measurements are also presented in Figure 3A-7. In the legend of Figure 3A-6, these studies are
denoted by the term "Direct Ci Measure."  As shown, there tends to be a greater degree of
agreement between modeling scenarios for lower concentration estimates, but less agreement for
higher concentration estimates.
       As described above, it is unclear to what the extent to which the homes studied for
benzene related to attached garages are representative of homes nationally. As such, in
summarizing the scenarios, several different approaches to "averaging" across scenarios are
presented here.  Figure 3A-8  shows the results of these different averaging scenarios. In the "All
Data" distribution shown in the figure, all scenarios are averaged together.  In the "Weighted
Average" distribution, weights are equal to the number of homes included in each study.  In the
"Model Only" distribution, only scenarios involving modeling C, are shown. In the "Measure
Only" distribution, only those studies in which C, was measured directly are shown. In the "AK
Only" distribution, only scenarios employing Alaskan Cg orfg studies are shown. In the "Non-
AK Only" distribution, only scenarios excluding Alaskan Cg orfg data are shown. These
scenarios are intended to span a range of estimates for the national estimate.
       The average concentrations from these "summary scenarios" are shown in Table 3A-1.
As shown in Table 3A-1 and  in Figure 3A-8, scenarios employing only measured indoor data
resulted in higher predicted benzene TWA exposure concentrations than the studies employing
only modeling.  Scenarios employing Alaskan data result in higher benzene concentrations than
scenarios excluding Alaskan  data.  Also weighting scenarios by the number of homes resulted in
higher benzene  concentrations.
       Accordingly, the national average TWA exposure concentration attributable to attached
garages is estimated to be 1.2 - 6.6 ng/m3. This range is intended to span possible values of
average TWA exposure from attached garages, given currently available information.  The actual
average TWA exposure concentration due to attached garages could be outside of this range.
Because of limited information on the representativeness of the homes studied, a more precise
"central estimate" is not appropriate at this time.  The width of the range, with the upper end
being 5.5 times  the lower end, is an indicator of the magnitude of uncertainty in the estimate.  It
is not a confidence interval in the traditional sense.  As more data become available, more
precise estimates will hopefully emerge.
       In comparison, the national average exposure concentration of census tract median
exposure concentrations in this rule is estimated at 1.4 |J,g/m3 for calendar year 1999.

                                         3-132

-------
Accordingly, if the attached garage exposure contribution is considered, the estimate of national
average exposure to benzene rises to 2.6 - 8.0  |J,g/m3, corresponding to an increase of 85-471%.

Effects of Emission Standards

       Several limitations prevent precise estimation of the effect of the standards in this rule on
garage-related exposures. First, cold temperature vehicle ignition and evaporative vehicle,
engine, and fuel container emissions can occur either in a garage or outdoors. Second, detailed
tracking of the time during which people are inside a house during cold vehicle starting or hot
vehicle soaking, when a majority of benzene emissions are likely to occur, is limited. However,
a bounding exercise can provide some estimates as to the effect of the standards in this rule.
       First, assuming full mixing and steady-state conditions, concentrations within a garage  is
estimable359 as:
                              (9)           Cg = (dMr/df) I aV
Here, the terms are similar to Equations 1-7.
       Given a change in the mass benzene emission rates from vehicle cold temperature
ignition, fuel evaporation from vehicles, engines, and fuel containers, an estimate of a change in
Cg is feasible. Table 2.2-52 of the RIA displays the emission reductions attributable to  each
program.  By splitting the emission reductions into evaporative  and exhaust emissions and
applying several simple assumptions about where emissions occur (in garage vs.  outdoors), the
fraction of emission reductions occurring within attached garages can be estimated.  This
estimate is calculated by assuming ranges of values for the fraction of evaporative and exhaust
emissions from each program that occur within an attached garage." As such, while the total
benzene mobile source and PFC emission reductions occurring as a result of the rule in 2030 are
37% less than the projected  emissions without controls (Table 2.2-52 of the RIA), emissions
inside attached garages are reduced by an estimated 43-44%.
       Applying this fraction to Equation 8 and Equation 7, for the "average" scenarios modeled
presented in Table 1, this amounts to a national average exposure reduction of approximately 0.5
-2.6
Limitations

       As apparent in the wide range of "scenario" averages, there remains considerable
uncertainty in ascertaining the true magnitude of attached garage exposure contributions
nationally. There are a number of limitations in the approaches undertaken here. First, although
comparison with measured indoor data shows reasonable performance for the modeling approach
employed here, the selection of simple one-compartment mass balance models for both garage
and home modeling may substantially understate the variation in concentrations within these
microenvironments.  All estimates here assumed steady-state conditions, and this may not be
u The assumed fraction of evaporative and exhaust emission reductions from each source occurring within an
attached garage are as follows. Ranges are represented as [min, max].  For LDGV, about 90% of emission
reductions are exhaust-related, of which Pg*[25%,75%] occur within attached garages; the fraction of evaporative
reductions occurring within attached garages are Pg *[25%,50%]. For small nonroad gasoline equipment, about
72% of emissions are from exhaust, of which Pg *[0%,2%] occur in attached garages; 24% are evaporative, of which
Pg *[90%,100%] occur in attached garages, 4% are refilling-related, of whichPg *[25%,75%] occur in attached
garages. For portable fuel containers, Pg *[25%,75%] of emissions are assumed to occur in attached garages.
                                           3-133

-------
appropriate for a source like a garage, where door opening, car entry and ignition, and other
major sources of benzene are likely to produce short-term spikes in exposure not accounted for
with steady-state assumptions.
       Second, the preponderance of these data were collected in locations with cold climates,
so the results may not be applicable to warmer locations where houses are not built with the
same degree of weather-tightness. Furthermore, studies suggest that indoor concentrations
arising from attached garages vary considerably in response to emission-related activities in a
garage such as cold vehicle ignition and parking a hot vehicle.360 Ambient temperatures may
affect the magnitude of emissions from these activities.
       Lastly, the extent to which the houses studied in the publications cited here are
"representative" of the  national housing stock is unknown.

Conclusions

       Modeled indoor benzene concentrations indicate that indoor air concentrations in homes
with attached garages may be substantially higher than in homes without attached garages.
       Garage concentrations of benzene appear to be a major source of indoor benzene in
homes with attached garages.  According to the modeling conducted here, this source could
explain the majority of exposures experienced by typical residents of such homes. Given this
finding, interventions that result in a reduction in emissions within the garage would be a
relatively efficient means of reducing overall personal exposure, particularly in areas
geographically similar to the areas of the studies upon which this analysis relies.  Given the
proximity of this source to homes, one major set of beneficiaries of the rule's emission controls
is likely to be people with homes with attached garages, particularly in areas with high fuel
benzene levels. Emissions from vehicles and fuel containers also may have greater relative
impacts on those with attached garages.  An elementary calculation of the intake fraction (iF) of
emissions occurring within attached garages with very basic assumptions  indicates that for
benzene emitted in a garage, approximately 3-18 parts per thousand are inhaled by a person in an
attached garage. This estimate is far in excess of estimated iF from ambient sources, and similar
to estimated iF estimates for indoor sources.361
                                          3-134

-------
Table 3 A-l.  Summary of National Average Exposure Estimates Attributable to Attached
Garages.  Different "averaging" assumptions shown.
"Averaging"
Scenario
All Data
Weighted Average
Measure Only
Model Only
AK Only
Non-AK Only
Benzene TWA
(ug/m3)
4.3
6.6
6.1
3.4
5.5
1.2
                                        3-135

-------
Figure 3A-la. Density of Garage Benzene Concentrations from Batterman et al. (2005)
        Density of Garage Concentrations
 Q
     CD

     O
     CN

     Q
     O
     p

     o
     o
     o
     o
            0       50       100      150


              Benzene Concentration (ug/m3)





Figure 3A-lb. Density of Air Exchange Rates (ACH)




              Density of Garage ACH
     oq

     o
     
-------
Figure 3A-2.  Additional Indoor Air Concentrations from Garage as a Function of Cg andfg
              Indoor Concentration as a Function of Garage Concentration (Cg) and %lntake Air
                                       from Garage (fg)
                                                            Cg=110
                                                              1=100
                                                                               • 40-45
                                                                               D 35-40
                                                                               • 30-35
                                                                               D 25-30
                                                                               • 20-25
                                                                               D15-20
                                                                               D10-15
                                                                               • 5-10
                                                                               DO-5
                                             3-137

-------
            Figure 3A-3. Distributions of Individual Garage Emission Factors


                                 Emission Rate (ug/hr)
              O
              O
              O
O
O
O
CD
O
O
O
00
O
O
O
O
O
O
O
O
O
O
O
O
O
CD
O
O
O
00
O
O
O
   p
   CO
  35



  H



  3£



  35



  3E



  35
1  5  i
   p
   CO
   p
   bo
   p
   CD
                                             IH
                       Q)Q)Q)Q)Q)Q)Q)Q)Q)Q)Q)Q)Q)Q)Q)
 Q)Q)
 CQCQ
 CDCD
                            Q)Q)Q)Q)Q)Q)Q)Q)Q)Q)Q)Q)Q)
                            CQCQCQCQCQCQCQCQCQCQCQCQCQ
                            CDCDCDCDCDCDCDCDCDCDCDCDCD
                                         3-138

-------
Figure 3A-4. Cumulative Distribution of Modeled Indoor Benzene Concentrations

                              Benzene Concentration (ug/m3)
         p
         o
   8
   rt-
   0"
   a
             p
             o
   -n  N>
   (D  O
=  OJ
fl>  o
"O   cn
c   o
1   *
8   §
   <  °
   aT  gs
   O"
   =  oo
   Zr  O
       CD
       O
       O
       O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
o
p
o

                                                                                  o_
                                                                                  OL
                                                           (D
                                                           0)

                                                           S
                                                           c_
                                                           ^
                                                          -a_
                                                           n>
                                                           I
                                                           o
                                                           (D
                                                           W
                                                                                  (D
                                                                                  (/>
                                                                                 s
                                                                                 O
                                                                                 Q.
                                                                                 2.
                                                                                 ffi
                                                                                 Q.
                                                                                    Q.
                                                                                    O
                                                                                    O
                                                                                    O
                                                                                  ^ s-
                                                                                  w
                                                                            n>  =• n
                                                                            -o  » is!
                                                                            fl>  g (/)
                                                                            I s =
                                                                            Q. 0) CT

                                                                            ll O
                                                                                    o
                                                                                    Q.
                                                                               D
                                                                               fi>
                                                                               0)
                                                             O
*• »

CD CD

ra ra
00)/FuglerFC
00) / Shelterso
£= u'
S 0
CD
Tl
O
O









CD

-g
O
?
0)
S
o








(fl
S
Si

o
Home NV, W
nter / Batt

3
=3
-n
O
o

i
C/)
CD
Si

O
g
ter / Batte
g


O




(fl
CD
Si

o
Home NV, Si
g
g
CD
CD
1

g
ZJ
-n
(,)
O

C/)
CD
£

§
HomeV, Sun
nmer/ Bat
CD
g


0



C/)
S
Si

O
Home NV, W
oT
~n
CQ
CD
~n
O
O




C/)
CD
Si

O
^
of
Tl
£=
(Q
CD

o
o




1
C/)
S
Si

O
Home NV, Si
3
3
CD
Tl
(Q

~n

n



f
^>

§
CO
£=
CD
Tl
£=
CQ
CD

o
o




• S-
CD
£

§
Home NV, W
nter /She
sr
S


CD
-n
G)
i
01
o- G
= f
Si

O
Home V, Win
ter/Shelte
t/5
o
^
CD
T|
O
o
I
^&^
T E
Si

o
Home NV, Si
mmer/Sh
CD
crT

°
a
CD

i
'I
Si

O
CO
£=
inner/She
oT
S
o
-1
CD

C)
}
O
CD
O


To
2) / Sheltersoi
rce FG Ci







It
O O
CD CD
O O


To To
S 1
O -n
O
O







1
1
=

Q)
ff
source FG
o






II
Q) Q)
1 1
= =

Q) Q)
CD" fD
an FG Ci
rGCi







                                                                              ffl_
                                                                               (D
                                                                               I
                                                                               O
                                                                                    O
                                                                                    o
                                                                                    3
                                                                                    JD
                                             3-139

-------
Figure 3 A-5. Cumulative Distributions of Observed Benzene Levels in Homes with Attached

Garages


                          Benzene Concentration (ug/m3)
                                                                                  TJ
                                                                                  O

                                                                             I    I
                                                                             a.    3
                                                                             i~    =:
                                                                             3'    o

                                                                             8    D
                                                                             «>     =•
                                                                             o     -•
                                                                             3     2.
                                                                             (D     -»
                                                                             (D
                                                                             U
                                                                             (/)
Q.
O
O
^

DO
                                                                             m
                                                                             S.    3
                                                                             X    N
                                                                             O     O

                                                                             q     o

                                                                             O O O
                                                                             DUO
                                                                               CD
§ J
as
measured ii
Maxmimum
NVsummei
r: __ zj
N) g- jn
^ 1 a

eg 9L^-
3 S _„
"° i
C/) ,—
It
CD. CO
9L

o
3
CD


i; ^®
CD
CD
Q_
5'
JQ.
g_
^
^
laximum i
Q_
o
o
9L

CD




artithmetic i
=i
CD
Q)
Q)

Q_
CQ
CD
O
3
a
o
3
CD
Q)

Q)
cn




Q) ~Z_
CD g.
ffS
cn >,
II

Q) §
o" -13
tq co
— i f\
0 §-
3 of
Q) 15.
— Q)
%'^
3". Q.
&. o
o g
-^ cn
— h
3 Q-
3- a
cr
c±
o

1
S
CD
z
D
m
~n
'/EPA (2006
CD
0)
m
0
O




CO
O
ST
•o.
^
o
^
w'
CD
Q)
Q_
o






DO
CD
i
Q)
CD_

CD
Q)
Q.
Q









CD
g
CD"
^
CD
0)
W
C
i
o



*

CD
Q.
o
O

-------
Figure 3A-6.  Comparison of Modeled and Observed Indoor Benzene Concentrations
          Batterman et al.

             (2005/6)
                                      Benzene Concentration (ug/m3)
                                            CO
                                            O
cn
o
CD
o
oo
o
   o
   ID

   D

   a
   fi>
   V)
   O
George et al. (2002)
Isbelletal. (2005),
Home NV, Summer
Isbelletal. (2005),
Home NV, Wnter
Isbelletal. (2005),
Home V, Summer
Isbelletal. (2005),
Home V, Wnter
Schlapia & Morris
(1998)
I

|

I

|
|
	 ,
I

I

I
I

	 ,
I

|

I

      Tsai & Weisel (2000)
     Weisel - NJDEP/EPA

           (2006)
                                               D D D D
                                               2 co ~n CD
                                               0 =T £= Q)
                                               oJ CD. (Q. ft

                                               £ CD  * 5

                                               3 I  a 1


                                                 I  " r
                                            3-141

-------
Figure 3A-7. Multiple Scenario Output of Predicted National Average Benzene Exposure
Attributable to Attached Garages
                            Benzene TWA Concentration (ug/m3)
                                           CO
                                           o
Ol
O
CD
O
      CO
      O
   o
   o
   TJ
      cn
      o
      CD
      o
      oo
      o
      CD
      o
      o
      o
\
1 §•
$ -g
i Q)'
£ <"
0 3
m o
ffl %
5 "
398) / Direct
k (2006) / Direct

tiiti
(/)(/)(/)(/)(/)(/)(/)(/)(/)
o-o-o-o-o-o-o-cro-
000000000
Q)Q)Q)Q)Q)Q)Q)Q)Q)

NJNJNJNJNJNJNJNJNJ
OOOOOOOOO
OOOOOOOOO
 00 3
0 0 0 0 ^^,^,^3
^^^^DCD~nw^
5g>?»S-S<§?^
(Ss(Q0a?t)stS
1 PS it
= 1 8
0
Ht
t/)t/)t/)t/)t/)t/)t/)^^
S"S"S"S"S"S"S"e. <8
S-S-S-0,-0,-0,-0,-^^
Q)Q)Q)Q)Q)Q)Q)55
U) U)
NJNJNJNJNJNJNJ0 0
ooooooo — —
ooooooo-T^-r^
o^o^o^cncncncn^^
	 o o
IIIIIIIOO
3333333m~n
0000000QjC
| | | < < < < s a
- - - w w w w 3 -1
cccllll™
3333333 =
0 00 1^ ^ ^ ^
-^ -^ -^ D CD ~n w
g> 2 » s S <§ J
^ CQ 0 0 0 Q r-4-
1 ?! If
fl) 0 c
1 8
0
Ht tl
-lOOOrorororo
U)000Q)Q)^)Q)
w. Soo^SSS
Poc3c3c3^^^^
-000^333
5000^)^)^)'u

_^-j^-j^-j^fl) 0) 0) 0)
Orvjrvj|vjoooo
5^?»xxxx
ogtSQrojm
ff3-'!"S3
-------
       Figure 3A-8.  Average "Summarized" Benzene Exposure Distributions
                   Benzene TWA Concentration (ug/m3)


             en         o         en
p
k>
p

CO
O
c
3
c_

~

o>
p
en
o
0>
D
^

TrT  P*
   0>
p
bo
p
CD
                                     -
                            O
                            3_

                            ^<
                                          Q)
                                  A  O  CD  m" —
                                  ^  Q-  Q) ,~ n
                                  O  CD.  « CO D


                                  •<  o  s
                                       CQ
                                       CD
                                                                                0)
                                                                                2.
                                                                                0)
                                                                                I-K

                                                                                o'
                                                                                3

                                                                                5'

                                                                                D
                                                                             O 51
                                                                             D) C
                                                                          8 3

                                                                          CD -"
                                                                          0) z
                                                                          (/) Q)
                                                                          (D S-
                                                                          tt 5'
                                                                          O =
                                                                          3 OJ.
                                                                             I
                                                                           : O
                                                                             5- m

                                                                               = 1
                                                                             > (/>
                                                                             w c
                                                                             g S
                                                                             3
                                                                          o
                                                                          3
                                                                                S-
                                                                                D)
                                                                                CT
                                                                                (D
                                                                                I-K
                                                                                O


                                                                                I
                                                                                D)
                                                                                O

                                                                                (D
                                                                                Q.
                                       3-143

-------
            Appendix 3B: 8-Hour Ozone Nonattainment
Table 3B-1. 8-Hour Ozone Nonattainment Areas, Counties and Populations (Data is
Current through October 2006 and Population Numbers are from 2000 Census Data)
8-hour Ozone Nonattainment
Area
Albany-Schenectady-Troy Area
Albany-Schenectady-Troy Area
Albany-Schenectady-Troy Area
Albany-Schenectady-Troy Area
Albany-Schenectady-Troy Area
Albany-Schenectady-Troy Area
Albany-Schenectady-Troy Area
Allegan County Area
Allentown-Bethlehem-Easton Area
Allentown-Bethlehem-Easton Area
Allentown-Bethlehem-Easton Area
Altoona Area
Amadorand Calaveras Counties
(Central Mountain Counties) Area
Amadorand Calaveras Counties
(Central Mountain Counties) Area
Atlanta Area
Atlanta Area
Atlanta Area
Atlanta Area
Atlanta Area
Atlanta Area
Atlanta Area
Atlanta Area
Atlanta Area
Atlanta Area
Atlanta Area
Atlanta Area
Atlanta Area
Atlanta Area
State
NY
NY
NY
NY
NY
NY
NY
Ml
PA
PA
PA
PA
CA
CA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
Classification3'15
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
County Name
Albany Co
Greene Co
Montgomery Co
RensselaerCo
Saratoga Co
Schenectady Co
Schoharie Co
Allegan Co
Carbon Co
Lehigh Co
Northampton Co
Blair Co
AmadorCo
Calaveras Co
Barrow Co
Bartow Co
Carroll Co
Cherokee Co
Clayton Co
CobbCo
Coweta Co
De Kalb Co
Douglas Co
Fayette Co
Forsyth Co
Fulton Co
Gwinnett Co
Hall Co
Whole
/Part
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
2000 Cty
Pop
294,565
48,195
49,708
152,538
200,635
146,555
31,582
105,665
58,802
312,090
267,066
129,144
35,100
40,554
46,144
76,019
87,268
141,903
236,517
607,751
89,215
665,865
92,174
91,263
98,407
816,006
588,448
139,277
                                 3-144

-------
Atlanta Area
Atlanta Area
Atlanta Area
Atlanta Area
Atlanta Area
Atlanta Area
Baltimore Area
Baltimore Area
Baltimore Area
Baltimore Area
Baltimore Area
Baltimore Area
Baton Rouge Area
Baton Rouge Area
Baton Rouge Area
Baton Rouge Area
Baton Rouge Area
Beaumont-Port Arthur Area
Beaumont-Port Arthur Area
Beaumont-Port Arthur Area
Benton Harbor Area
Benzie County Area
Berkeley and Jefferson Counties
Area
Berkeley and Jefferson Counties
Area
Boston-Lawrence-Worcester (E.
Mass) Area
Boston-Lawrence-Worcester (E.
Mass) Area
Boston-Lawrence-Worcester (E.
Mass) Area
Boston-Lawrence-Worcester (E.
Mass) Area
GA
GA
GA
GA
GA
GA
MD
MD
MD
MD
MD
MD
LA
LA
LA
LA
LA
TX
TX
TX
Ml
Ml
WV
WV
MA
MA
MA
MA
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart 1
Subpart 1
Subpart 1 - EAC
Subpart 1 - EAC
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Henry Co
Newton Co
Paulding Co
Rockdale Co
Spalding Co
Walton Co
Anne Arundel
Co
Baltimore (City)
Baltimore Co
Carroll Co
Harford Co
Howard Co
Ascension Par
East Baton
Rouge Par
Iberville Par
Livingston Par
West Baton
Rouge Par
Hardin Co
Jefferson Co
Orange Co
Berrien Co
Benzie Co
Berkeley Co
Jefferson Co
Barnstable Co
Bristol Co
Dukes Co
Essex Co
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
119,341
62,001
81,678
70,111
58,417
60,687
489,656
651,154
754,292
150,897
218,590
247,842
76,627
412,852
33,320
91,814
21,601
48,073
252,051
84,966
162,453
15,998
75,905
42,190
222,230
534,678
14,987
723,419
3-145

-------
Boston-Lawrence-Worcester (E.
Mass) Area
Boston-Lawrence-Worcester (E.
Mass) Area
Boston-Lawrence-Worcester (E.
Mass) Area
Boston-Lawrence-Worcester (E.
Mass) Area
Boston-Lawrence-Worcester (E.
Mass) Area
Boston-Lawrence-Worcester (E.
Mass) Area
Boston-Manchester-Portsmouth
(SE) Area
Boston-Manchester-Portsmouth
(SE) Area
Boston-Manchester-Portsmouth
(SE) Area
Boston-Manchester-Portsmouth
(SE) Area
Buffalo-Niagara Falls Area
Buffalo-Niagara Falls Area
Canton-Massillon Area
Cass County Area
Charlotte-Gastonia-Rock Hill Area
Charlotte-Gastonia-Rock Hill Area
Charlotte-Gastonia-Rock Hill Area
Charlotte-Gastonia-Rock Hill Area
Charlotte-Gastonia-Rock Hill Area
Charlotte-Gastonia-Rock Hill Area
Charlotte-Gastonia-Rock Hill Area
Charlotte-Gastonia-Rock Hill Area
Chattanooga Area
Chattanooga Area
Chattanooga Area
Chicago-Gary-Lake County Area
Chicago-Gary-Lake County Area
Chicago-Gary-Lake County Area
Chicago-Gary-Lake County Area
Chicago-Gary-Lake County Area
MA
MA
MA
MA
MA
MA
NH
NH
NH
NH
NY
NY
OH
Ml
NC
NC
NC
NC
NC
NC
NC
SC
GA
TN
TN
IL
IL
IL
IL
IL
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart 1
Subpart 1
Subpart 1
Subpart
2/Marginal
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Middlesex Co
Nantucket Co
Norfolk Co
Plymouth Co
Suffolk Co
Worcester Co
Hillsborough Co
Merrimack Co
Rockingham Co
Stratford Co
Erie Co
Niagara Co
Stark Co
Cass Co
Cabarrus Co
Gaston Co
Iredell Co
Lincoln Co
Mecklenburg Co
Rowan Co
Union Co
York Co
Catoosa Co
Hamilton Co
Meigs Co
Cook Co
Du Page Co
Grundy Co
Kane Co
Kendall Co
W
W
W
W
W
W
P
P
P
P
W
W
W
W
W
W
P
W
W
W
W
P
W
W
W
W
W
P
W
P
1,465,396
9,520
650,308
472,822
689,807
750,963
336,518
11,721
266,340
82,134
950,265
219,846
378,098
51,104
131,063
190,365
39,885
63,780
695,454
130,340
123,677
102,000
53,282
307,896
11,086
5,376,741
904,161
6,309
404,119
28,417
3-146

-------
Chicago-Gary-Lake County Area
Chicago-Gary-Lake County Area
Chicago-Gary-Lake County Area
Chicago-Gary-Lake County Area
Chicago-Gary-Lake County Area
Chico Area
Cincinnati-Hamilton Area
Cincinnati-Hamilton Area
Cincinnati-Hamilton Area
Cincinnati-Hamilton Area
Cincinnati-Hamilton Area
Cincinnati-Hamilton Area
Cincinnati-Hamilton Area
Cincinnati-Hamilton Area
Cincinnati-Hamilton Area
Clearfield and Indiana Counties
Area
Clearfield and Indiana Counties
Area
Cleveland-Akron-Lorain Area
Cleveland-Akron-Lorain Area
Cleveland-Akron-Lorain Area
Cleveland-Akron-Lorain Area
Cleveland-Akron-Lorain Area
Cleveland-Akron-Lorain Area
Cleveland-Akron-Lorain Area
Cleveland-Akron-Lorain Area
Columbia Area
Columbia Area
Columbus Area
Columbus Area
Columbus Area
Columbus Area
Columbus Area
Columbus Area
Dallas-Fort Worth Area
Dallas-Fort Worth Area
IL
IL
IL
IN
IN
CA
IN
KY
KY
KY
OH
OH
OH
OH
OH
PA
PA
OH
OH
OH
OH
OH
OH
OH
OH
SC
SC
OH
OH
OH
OH
OH
OH
TX
TX
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart
2/M ode rate
Subpart
2/M ode rate
Lake Co
Me Henry Co
Will Co
Lake Co
Porter Co
Butte Co
Dearborn Co
Boone Co
Campbell Co
Kenton Co
Butler Co
Clermont Co
Clinton Co
Hamilton Co
Warren Co
Clearfield Co
Indiana Co
Ashtabula Co
Cuyahoga Co
Geauga Co
Lake Co
Lorain Co
Medina Co
Portage Co
Summit Co
Lexington Co
Richland Co
Delaware Co
Fairfield Co
Franklin Co
Knox Co
Licking Co
Madison Co
Collin Co
Dallas Co
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
P
P
W
W
W
W
W
W
W
W
644,356
260,077
502,266
484,564
146,798
203,171
10,434
85,991
88,616
151,464
332,807
177,977
40,543
845,303
158,383
83,382
89,605
102,728
1,393,978
90,895
227,511
284,664
151,095
152,061
542,899
181,265
313,253
109,989
122,759
1,068,978
54,500
145,491
40,213
491,675
2,218,899
3-147

-------
Dallas-Fort Worth Area
Dallas-Fort Worth Area
Dallas-Fort Worth Area
Dallas-Fort Worth Area
Dallas-Fort Worth Area
Dallas-Fort Worth Area
Dallas-Fort Worth Area
Dayton-Springfield Area
Dayton-Springfield Area
Dayton-Springfield Area
Dayton-Springfield Area
Denver-Boulder-Greeley-Ft. Collins-
Love. Area
Denver-Boulder-Greeley-Ft. Collins-
Love. Area
Denver-Boulder-Greeley-Ft. Collins-
Love. Area
Denver-Boulder-Greeley-Ft. Collins-
Love. Area
Denver-Boulder-Greeley-Ft. Collins-
Love. Area
Denver-Boulder-Greeley-Ft. Collins-
Love. Area
Denver-Boulder-Greeley-Ft. Collins-
Love. Area
Denver-Boulder-Greeley-Ft. Collins-
Love. Area
Denver-Boulder-Greeley-Ft. Collins-
Love. Area
Detroit-Ann Arbor Area
Detroit-Ann Arbor Area
Detroit-Ann Arbor Area
Detroit-Ann Arbor Area
Detroit-Ann Arbor Area
Detroit-Ann Arbor Area
Detroit-Ann Arbor Area
Detroit-Ann Arbor Area
Door County Area
Erie Area
TX
TX
TX
TX
TX
TX
TX
OH
OH
OH
OH
CO
CO
CO
CO
CO
CO
CO
CO
CO
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Wl
PA
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart 1
Subpart 1
Denton Co
Ellis Co
Johnson Co
Kaufman Co
Parker Co
Rockwall Co
Tarrant Co
Clark Co
Greene Co
Miami Co
Montgomery Co
Adams Co
Arapahoe Co
Boulder Co
Broomfield Co
Denver Co
Douglas Co
Jefferson Co
Larimer Co
Weld Co
Lenawee Co
Livingston Co
Macomb Co
Monroe Co
Oakland Co
St Clair Co
Washtenaw Co
Wayne Co
Door Co
Erie Co
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
W
W
W
432,976
111,360
126,811
71,313
88,495
43,080
1,446,219
144,742
147,886
98,868
559,062
348,618
487,967
269,814
38,272
554,636
175,766
525,507
239,000
172,000
98,890
156,951
788,149
145,945
1,194,156
164,235
322,895
2,061,162
27,961
280,843
3-148

-------
Essex County (Whiteface Mtn.)
Area
Fayetteville Area
Flint Area
Flint Area
Fort Wayne Area
Franklin County Area
Frederick County Area
Frederick County Area
Grand Rapids Area
Grand Rapids Area
Greater Connecticut Area
Greater Connecticut Area
Greater Connecticut Area
Greater Connecticut Area
Greater Connecticut Area
Greene County Area
Greensboro-Winston-Salem-High
Point Area
Greensboro-Winston-Salem-High
Point Area
Greensboro-Winston-Salem-High
Point Area
Greensboro-Winston-Salem-High
Point Area
Greensboro-Winston-Salem-High
Point Area
Greensboro-Winston-Salem-High
Point Area
Greensboro-Winston-Salem-High
Point Area
Greensboro-Winston-Salem-High
Point Area
Greenville-Spartanburg-Anderson
Area
Greenville-Spartanburg-Anderson
Area
Greenville-Spartanburg-Anderson
Area
Hancock, Knox, Lincoln and Waldo
NY
NC
Ml
Ml
IN
PA
VA
VA
Ml
Ml
CT
CT
CT
CT
CT
PA
NC
NC
NC
NC
NC
NC
NC
NC
SC
SC
SC
ME
Subpart 1
Subpart 1 - EAC
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1
Subpart 1
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart 1
Subpart
2/Marginal -
EAC
Subpart
2/Marginal -
EAC
Subpart
2/Marginal -
EAC
Subpart
2/Marginal -
EAC
Subpart
2/Marginal -
EAC
Subpart
2/Marginal -
EAC
Subpart
2/Marginal -
EAC
Subpart
2/Marginal -
EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1
Essex Co
Cumberland Co
Genesee Co
Lapeer Co
Allen Co
Franklin Co
Frederick Co
Winchester
Kent Co
Ottawa Co
Hartford Co
Litchfield Co
New London Co
Tolland Co
Windham Co
Greene Co
Alamance Co
Caswell Co
Davidson Co
Davie Co
Forsyth Co
Guilford Co
Randolph Co
Rockingham Co
Anderson Co
Greenville Co
Spartanburg Co
Hancock Co
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
1,000
302,963
436,141
87,904
331,849
129,313
59,209
23,585
574,335
238,314
857,183
182,193
259,088
136,364
109,091
40,672
130,800
23,501
147,246
34,835
306,067
421,048
130,454
91,928
165,740
379,616
253,791
29,805
3-149

-------
Counties (Central Maine Coast)
Area
Hancock, Knox, Lincoln and Waldo
Counties (Central Maine Coast)
Area
Hancock, Knox, Lincoln and Waldo
Counties (Central Maine Coast)
Area
Hancock, Knox, Lincoln and Waldo
Counties (Central Maine Coast)
Area
Harrisburg-Lebanon-Carlisle Area
Harrisburg-Lebanon-Carlisle Area
Harrisburg-Lebanon-Carlisle Area
Harrisburg-Lebanon-Carlisle Area
Haywood and Swain Counties
(Great Smoky NP) Area
Haywood and Swain Counties
(Great Smoky NP) Area
Hickory-Morganton-Lenoir Area
Hickory-Morganton-Lenoir Area
Hickory-Morganton-Lenoir Area
Hickory-Morganton-Lenoir Area
Houston-Galveston-Brazoria Area
Houston-Galveston-Brazoria Area
Houston-Galveston-Brazoria Area
Houston-Galveston-Brazoria Area
Houston-Galveston-Brazoria Area
Houston-Galveston-Brazoria Area
Houston-Galveston-Brazoria Area
Houston-Galveston-Brazoria Area
Huntington-Ashland Area
Huron County Area
Imperial County Area
Indianapolis Area
Indianapolis Area
Indianapolis Area
Indianapolis Area
Indianapolis Area
Indianapolis Area
Indianapolis Area
Indianapolis Area
Indianapolis Area

ME
ME
ME
PA
PA
PA
PA
NC
NC
NC
NC
NC
NC
TX
TX
TX
TX
TX
TX
TX
TX
KY
Ml
CA
IN
IN
IN
IN
IN
IN
IN
IN
IN

Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart 1
Subpart 1
Subpart
2/Marginal
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1

Knox Co
Lincoln Co
Waldo Co
Cumberland Co
Dauphin Co
Lebanon Co
Perry Co
Haywood Co
Swain Co
Alexander Co
Burke Co
Caldwell Co
Catawba Co
Brazoria Co
Chambers Co
Fort Bend Co
Galveston Co
Harris Co
Liberty Co
Montgomery Co
Waller Co
Boyd Co
Huron Co
Imperial Co
Boone Co
Hamilton Co
Hancock Co
Hendricks Co
Johnson Co
Madison Co
Marion Co
Morgan Co
Shelby Co

P
P
P
W
W
W
W
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

33,563
28,504
604
213,674
251,798
120,327
43,602
28
260
33,603
69,970
64,254
141,685
241,767
26,031
354,452
250,158
3,400,578
70,154
293,768
32,663
49,752
36,079
142,361
46,107
182,740
55,391
104,093
115,209
133,358
860,454
66,689
43,445
3-150

-------
Jamestown Area
Jefferson County Area
Johnson City-Kingsport-Bristol Area
Johnson City-Kingsport-Bristol Area
Johnstown Area
Kalamazoo-Battle Creek Area
Kalamazoo-Battle Creek Area
Kalamazoo-Battle Creek Area
Kent and Queen Anne's Counties
Area
Kent and Queen Anne's Counties
Area
Kern County (Eastern Kern) Area
Kewaunee County Area
Knoxville Area
Knoxville Area
Knoxville Area
Knoxville Area
Knoxville Area
Knoxville Area
Knoxville Area
La Porte County Area
Lancaster Area
Lansing-East Lansing Area
Lansing-East Lansing Area
Lansing-East Lansing Area
Las Vegas Area
Lima Area
Los Angeles and San Bernardino
Counties (W Mojave Desert) Area
Los Angeles and San Bernardino
Counties (W Mojave Desert) Area
Los Angeles-South Coast Air Basin
Area
Los Angeles-South Coast Air Basin
Area
Los Angeles-South Coast Air Basin
Area
Los Angeles-South Coast Air Basin
Area
Louisville Area
Louisville Area
Louisville Area
Louisville Area
Louisville Area
Macon Area
Macon Area
Manitowoc County Area
NY
NY
TN
TN
PA
Ml
Ml
Ml
MD
MD
CA
Wl
TN
TN
TN
TN
TN
TN
TN
IN
PA
Ml
Ml
Ml
NV
OH
CA
CA
CA
CA
CA
CA
IN
IN
KY
KY
KY
GA
GA
Wl
Subpart 1
Subpart
2/M ode rate
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart
2/Marginal
Subpart
2/Marginal
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart
2/Marginal
Subpart
2/Marginal
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/Severe 1 7
Subpart
2/Severe 1 7
Subpart
2/Severe 1 7
Subpart
2/Severe 1 7
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Chautauqua Co
Jefferson Co
Hawkins Co
Sullivan Co
Cambria Co
Calhoun Co
Kalamazoo Co
Van Buren Co
Kent Co
Queen Annes
Co
Kern Co
Kewaunee Co
Anderson Co
Blount Co
Cocke Co
Jefferson Co
Knox Co
Loudon Co
Sevier Co
La Porte Co
Lancaster Co
Clinton Co
Eaton Co
Ingham Co
Clark Co
Allen Co
Los Angeles Co
San Bernardino
Co
Los Angeles Co
Orange Co
Riverside Co
San Bernardino
Co
Clark Co
Floyd Co
Bullitt Co
Jefferson Co
Oldham Co
Bibb Co
Monroe Co
Manitowoc Co
W
W
W
W
W
W
W
W
W
W
P
W
W
W
P
W
W
W
W
W
W
W
W
W
P
W
P
P
P
W
P
P
W
W
W
W
W
W
P
W
139,750
111,738
53,563
153,048
152,598
137,985
238,603
76,263
19,197
40,563
99,251
20,187
71,330
105,823
20
44,294
382,032
39,086
71,170
110,106
470,658
64,753
103,655
279,320
1,348,864
108,473
297,058
359,350
9,222,280
2,846,289
1,194,859
1,330,159
96,472
70,823
61,236
693,604
46,178
153,887
50
82,887
3-151

-------
Mariposa and Tuolumne Counties
(Southern Mountain Counties) Area
Mariposa and Tuolumne Counties
(Southern Mountain Counties) Area
Mason County Area
Memphis Area
Memphis Area
Milwaukee-Racine Area
Milwaukee-Racine Area
Milwaukee-Racine Area
Milwaukee-Racine Area
Milwaukee-Racine Area
Milwaukee-Racine Area
Murray County (Chattahoochee Nat
Forest) Area
Muskegon Area
Nashville Area
Nashville Area
Nashville Area
Nashville Area
Nashville Area
Nevada County (Western part) Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
CA
CA
Ml
AR
TN
Wl
Wl
Wl
Wl
Wl
Wl
GA
Ml
TN
TN
TN
TN
TN
CA
CT
CT
CT
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
Subpart 1
Subpart 1
Subpart 1
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart 1
Subpart
2/Marginal
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Mariposa Co
Tuolumne Co
Mason Co
Crittenden Co
Shelby Co
Kenosha Co
Milwaukee Co
Ozaukee Co
Racine Co
Washington Co
Waukesha Co
Murray Co
Muskegon Co
Davidson Co
Rutherford Co
SumnerCo
Williamson Co
Wilson Co
Nevada Co
Fairfield Co
Middlesex Co
New Haven Co
Bergen Co
Essex Co
Hudson Co
Hunterdon Co
Middlesex Co
Monmouth Co
Morris Co
Passaic Co
W
W
W
W
W
W
W
W
W
W
W
P
W
W
W
W
W
W
P
W
W
W
W
W
W
W
W
W
W
W
17,130
54,501
28,274
50,866
897,472
149,577
940,164
82,317
188,831
117,493
360,767
1,000
170,200
569,891
182,023
130,449
126,638
88,809
77,735
882,567
155,071
824,008
884,118
793,633
608,975
121,989
750,162
615,301
470,212
489,049
3-152

-------
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
NewYork-N. New Jersey-Long
Island Area
Norfolk-Virginia Beach-Newport
News (Hampton Roads) Area
Norfolk-Virginia Beach-Newport
News (Hampton Roads) Area
Norfolk-Virginia Beach-Newport
News (Hampton Roads) Area
Norfolk-Virginia Beach-Newport
News (Hampton Roads) Area
Norfolk-Virginia Beach-Newport
News (Hampton Roads) Area
Norfolk-Virginia Beach-Newport
News (Hampton Roads) Area
Norfolk-Virginia Beach-Newport
News (Hampton Roads) Area
Norfolk-Virginia Beach-Newport
News (Hampton Roads) Area
Norfolk-Virginia Beach-Newport
News (Hampton Roads) Area
Norfolk-Virginia Beach-Newport
News (Hampton Roads) Area
Norfolk-Virginia Beach-Newport
News (Hampton Roads) Area
Norfolk-Virginia Beach-Newport
News (Hampton Roads) Area
Norfolk-Virginia Beach-Newport
News (Hampton Roads) Area
Parkersburg-Marietta Area
Parkersburg-Marietta Area
NJ
NJ
NJ
NJ
NY
NY
NY
NY
NY
NY
NY
NY
NY
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
OH
WV
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart 1
Subpart 1
Somerset Co
Sussex Co
Union Co
Warren Co
Bronx Co
Kings Co
Nassau Co
New York Co
Queens Co
Richmond Co
Rockland Co
Suffolk Co
Westchester Co
Chesapeake
Gloucester Co
Hampton
Isle Of Wight Co
James City Co
Newport News
Norfolk
Poquoson
Portsmouth
Suffolk
Virginia Beach
Williamsburg
York Co
Washington Co
Wood Co
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
297,490
144,166
522,541
102,437
1,332,650
2,465,326
1,334,544
1,537,195
2,229,379
443,728
286,753
1,419,369
923,459
199,184
34,780
146,437
29,728
48,102
180,150
234,403
1 1 ,566
100,565
63,677
425,257
11,998
56,297
63,251
87,986
3-153

-------
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Philadelphia-Wilmington-Atlantic
City Area
Phoenix-Mesa Area
Phoenix-Mesa Area
Pittsburgh-Beaver Valley Area
Pittsburgh-Beaver Valley Area
Pittsburgh-Beaver Valley Area
Pittsburgh-Beaver Valley Area
Pittsburgh-Beaver Valley Area
Pittsburgh-Beaver Valley Area
Pittsburgh-Beaver Valley Area
Portland Area
Portland Area
Portland Area
Portland Area
DE
DE
DE
MD
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
PA
PA
PA
PA
PA
AZ
AZ
PA
PA
PA
PA
PA
PA
PA
ME
ME
ME
ME
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
Kent Co
New Castle Co
Sussex Co
Cecil Co
Atlantic Co
Burlington Co
Camden Co
Cape May Co
Cumberland Co
Gloucester Co
Mercer Co
Ocean Co
Salem Co
Bucks Co
Chester Co
Delaware Co
Montgomery Co
Philadelphia Co
Maricopa Co
Pinal Co
Allegheny Co
Armstrong Co
Beaver Co
Butler Co
Fayette Co
Washington Co
Westmoreland
Co
Androscoggin
Co
Cumberland Co
Sagadahoc Co
York Co
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
P
W
P
126,697
500,265
156,638
85,951
252,552
423,394
508,932
102,326
146,438
254,673
350,761
510,916
64,285
597,635
433,501
550,864
750,097
1,517,550
3,054,504
31,541
1,281,666
72,392
181,412
174,083
148,644
202,897
369,993
3,390
252,907
35,214
164,997
3-154

-------

Poughkeepsie Area
Poughkeepsie Area
Poughkeepsie Area
Providence (all of Rl) Area
Providence (all of Rl) Area
Providence (all of Rl) Area
Providence (all of Rl) Area
Providence (all of Rl) Area
Raleigh-Durham-Chapel Hill Area
Raleigh-Durham-Chapel Hill Area
Raleigh-Durham-Chapel Hill Area
Raleigh-Durham-Chapel Hill Area
Raleigh-Durham-Chapel Hill Area
Raleigh-Durham-Chapel Hill Area
Raleigh-Durham-Chapel Hill Area
Raleigh-Durham-Chapel Hill Area
Reading Area
Richmond-Petersburg Area
Richmond-Petersburg Area
Richmond-Petersburg Area
Richmond-Petersburg Area
Richmond-Petersburg Area
Richmond-Petersburg Area
Richmond-Petersburg Area
Richmond-Petersburg Area
Richmond-Petersburg Area
Riverside County (Coachella
Valley) Area
Roanoke Area
Roanoke Area
Roanoke Area
Roanoke Area
Rochester Area
Rochester Area

NY
NY
NY
Rl
Rl
Rl
Rl
Rl
NC
NC
NC
NC
NC
NC
NC
NC
PA
VA
VA
VA
VA
VA
VA
VA
VA
VA
CA
VA
VA
VA
VA
NY
NY
2/Marginal
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Serious
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1
Subpart 1

Dutch ess Co
Orange Co
Putnam Co
Bristol Co
Kent Co
Newport Co
Providence Co
Washington Co
Chatham Co
Durham Co
Franklin Co
Granville Co
Johnston Co
Orange Co
Person Co
Wake Co
Berks Co
Charles City Co
Chesterfield Co
Colonial Heights
Hanover Co
Henrico Co
Hopewell
Petersburg
Prince George
Co
Richmond
Riverside Co
Botetourt Co
Roanoke
Roanoke Co
Salem
Genesee Co
Livingston Co

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
P
W
W
W
W
W
W

280,150
341,367
95,745
50,648
167,090
85,433
621,602
123,546
21,320
223,314
47,260
48,498
121,965
118,227
35,623
627,846
373,638
6,926
259,903
16,897
86,320
262,300
22,354
33,740
33,047
197,790
324,750
30,496
94,911
85,778
24,747
60,370
64,328
3-155

-------
Rochester Area
Rochester Area
Rochester Area
Rochester Area
Rocky Mount Area
Rocky Mount Area
Sacramento Metro Area
Sacramento Metro Area
Sacramento Metro Area
Sacramento Metro Area
Sacramento Metro Area
Sacramento Metro Area
San Antonio Area
San Antonio Area
San Antonio Area
San Diego Area
San Francisco Bay Area
San Francisco Bay Area
San Francisco Bay Area
San Francisco Bay Area
San Francisco Bay Area
San Francisco Bay Area
San Francisco Bay Area
San Francisco Bay Area
San Francisco Bay Area
San Joaquin Valley Area
San Joaquin Valley Area
San Joaquin Valley Area
San Joaquin Valley Area
San Joaquin Valley Area
San Joaquin Valley Area
San Joaquin Valley Area
NY
NY
NY
NY
NC
NC
CA
CA
CA
CA
CA
CA
TX
TX
TX
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart
2/Serious
Subpart
2/Serious
Subpart
2/Serious
Subpart
2/Serious
Subpart
2/Serious
Subpart
2/Serious
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1 - EAC
Subpart 1
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Marginal
Subpart
2/Serious
Subpart
2/Serious
Subpart
2/Serious
Subpart
2/Serious
Subpart
2/Serious
Subpart
2/Serious
Subpart
Monroe Co
Ontario Co
Orleans Co
Wayne Co
Edgecombe Co
Nash Co
El Dorado Co
Placer Co
Sacramento Co
Solano Co
Sutter Co
Yolo Co
Bexar Co
Comal Co
Guadalupe Co
San Diego Co
Alameda Co
Contra Costa Co
Marin Co
Napa Co
San Francisco
Co
San Mateo Co
Santa Clara Co
Solano Co
Sonoma Co
Fresno Co
Kern Co
Kings Co
Madera Co
Merced Co
San Joaquin Co
Stanislaus Co
W
W
W
W
W
W
P
P
W
P
P
W
W
W
W
P
W
W
W
W
W
W
W
P
P
W
P
W
W
W
W
W
735,343
100,224
44,171
93,765
55,606
87,420
124,164
239,978
1,223,499
197,034
25,013
168,660
1,392,931
78,021
89,023
2,813,431
1,443,741
948,816
247,289
124,279
776,733
707,161
1,682,585
197,508
413,716
799,407
550,220
129,461
123,109
210,554
563,598
446,997
3-156

-------

San Joaquin Valley Area
Scranton-Wilkes-Barre Area
Scranton-Wilkes-Barre Area
Scranton-Wilkes-Barre Area
Scranton-Wilkes-Barre Area
Sheboygan Area
South Bend-Elkhart Area
South Bend-Elkhart Area
Springfield (W. Mass) Area
Springfield (W. Mass) Area
Springfield (W. Mass) Area
Springfield (W. Mass) Area
St. Louis Area
St. Louis Area
St. Louis Area
St. Louis Area
St. Louis Area
St. Louis Area
St. Louis Area
St. Louis Area
St. Louis Area
State College Area
Steubenville-Weirton Area
Steubenville-Weirton Area
Steubenville-Weirton Area
Sutter County (part) (Sutter Buttes)
Area
Tioga County Area
Toledo Area
Toledo Area
Ventura County (part) Area
Washington Area
Washington Area
Washington Area

CA
PA
PA
PA
PA
Wl
IN
IN
MA
MA
MA
MA
IL
IL
IL
IL
MO
MO
MO
MO
MO
PA
OH
WV
WV
CA
PA
OH
OH
CA
DC
MD
MD
2/Serious
Subpart
2/Serious
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart
2/M ode rate
Subpart 1
Subpart 1
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart

Tula re Co
Lackawanna Co
Luzerne Co
Monroe Co
Wyoming Co
Sheboygan Co
Elkhart Co
St Joseph Co
Berkshire Co
Franklin Co
Hampden Co
Hampshire Co
Jersey Co
Madison Co
Monroe Co
St Clair Co
Franklin Co
Jefferson Co
St Charles Co
St Louis
St Louis Co
Centre Co
Jefferson Co
Brooke Co
Hancock Co
Sutter Co
Tioga Co
Lucas Co
Wood Co
Ventura Co
Entire District
Calvert Co
Charles Co

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
W
W
P
W
W
W

368,021
213,295
319,250
138,687
28,080
112,646
182,791
265,559
134,953
71,535
456,228
152,251
21,668
258,941
27,619
256,082
93,807
198,099
283,883
348,189
1,016,315
135,758
73,894
25,447
32,667
1
41,373
455,054
121,065
753,197
572,059
74,563
120,546
3-157

-------

Washington Area
Washington Area
Washington Area
Washington Area
Washington Area
Washington Area
Washington Area
Washington Area
Washington Area
Washington Area
Washington Area
Washington Area
Washington County (Hagerstown)
Area
Wheeling Area
Wheeling Area
Wheeling Area
York Area
York Area
Youngstown-Warren-Sharon Area
Youngstown-Warren-Sharon Area
Youngstown-Warren-Sharon Area
Youngstown-Warren-Sharon Area

MD
MD
MD
VA
VA
VA
VA
VA
VA
VA
VA
VA
MD
OH
WV
WV
PA
PA
OH
OH
OH
PA
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart
2/M ode rate
Subpart 1 - EAC
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1
Subpart 1

Frederick Co
Montgomery Co
Prince George's
Co
Alexandria
Arlington Co
Fairfax
Fairfax Co
Falls Church
Loudoun Co
Manassas
Manassas Park
Prince William
Co
Washington Co
Belmont Co
Marshall Co
Ohio Co
Adams Co
York Co
Columbiana Co
Mahoning Co
Trumbull Co
Mercer Co

W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W

195,277
873,341
801,515
128,283
189,453
21,498
969,749
10,377
169,599
35,135
10,290
280,813
131,923
70,226
35,519
47,427
91,292
381,751
112,075
257,555
225,116
120,293

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 has 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.

b) Boston-Manchester-Portsmouth (SE), NH has the same classification as Boston-Lawrence- Worcester (E. MA),
MA.
                                                3-158

-------
                  Appendix 3C: PM Nonattainment
Table 3C-1. PMi.s Nonattainment Areas and Populations (data is current through October
2006 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
                                  3-159

-------
Table 3C-2. PMi0 Nonattainment Areas and Populations (data is current through March
             2006 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
Lame Deer, MT
Lane Co, OR
Libby, MT
Los Angeles South Coast Air Basin, CA
Medford-Ashland, OR
Missoula, MT
Mono Basin, CA
Mun. of Guaynabo, PR
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
Salt Lake Co, UT
San Bernardino Co, CA
San Joaquin Valley, CA
Moderate
Moderate
Moderate
Moderate
Serious
Serious
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Serious
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Serious
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Serious
Moderate
Serious
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Moderate
Serious
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
1
1
1
1
1
1
1
1
4
1
1
1
1
1
1
1
1
1
2
1
1
2
1
1
1
1
1
7
8
3
37
35
1,376
182
4
7
16
195
564
179
5
1
4
120
14
15
12
3
1
3
3
14,594
78
52
0
92
1,537
25
77
7
1
3,112
2
4
66
1
3
1,223
898
199
3,080
9
6
10
8
9
9
8
9
9
10
6
10
8
10
9
9
10
8
10
10
8
10
8
9
10
8
9
2
2
9
8
9
9
9
10
8
10
9
8
9
8
9
9
AZ
NM
ID
MT
NV
CA
MT
CA
AZ
AK
TX
OR
MT
ID
AZ
CA
AK
MT
OR
OR
MT
OR
MT
CA
OR
MT
CA
PR
NY
AZ
UT
CA
AZ
AZ
ID
MT
ID
AZ
MT
CA
UT
CA
CA
                                       3-160

-------
Sanders County (part);Thompson Falls and vicinity,
MT
Sheridan, WY
Shoshone Co, ID
Trona, CA
Utah Co, UT
Washoe Co, NV
Weirton, WV
Yuma, AZ
Moderate

Moderate
Moderate
Moderate
Moderate
Serious
Moderate
Moderate
1

1
1
1
1
1
2
1
1

16
10
4
369
339
15
82
8

8
10
9
8
9
3
9
MT

WY
ID
CA
UT
NV
WV
AZ
51 Total Areas                                                     51          28,674
                                          3-161

-------
                     Appendix 3D: Visibility Tables
Table 3D-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
                                    3-162

-------
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
                                      3-163

-------
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
                                      3-164

-------
 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.
                                          3-165

-------
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
                                            3-166

-------
Table 3D-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
                                       3-167

-------
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
3-168

-------
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).
                                                 3-169

-------
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 Environmental Protection Agency (2006) National Air Toxics  Trends Stations. [Online at
http ://www. epa. gov/ttn/amtic/natts.html]
5 Kenski, D; Koerber, M.; Hafner, H. et al. (2005) Lessons learned from air toxics data.  A
national perspective. Environ Manage. June 2005:  19-22.
6 Touma, J.S.; Cox, W.M.; Tikvart, J.A. (2006) Spatial and temporal variability of ambient air
toxics data. J Air & Waste Manage Assoc (in press).
7 Reiss, R. (2006) Temporal trend and weekend-weekday differences for benzene and 1,3-
butadiene in Houston, Texas. Atmos Environ 40:  4711-4724.
8 Harley, R.A.; Hooper, D.S.; Kean, A.J.; Kirchstetter, T.W.; Hesson, J.M.; Balberan, N.T.;
Stevenson, E.D.; Kendall, G.R. (2006) Effects of reformulated gasoline and motor vehicle fleet
turnover on emissions and ambient concentrations  of benzene. Environ  Sci Technol 40:  5084-
5088.
9 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.
10 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.
11 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.
12 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.
13 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.
                                         3-170

-------
14 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.
15 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.
16 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.
17 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.
18 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.
19 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.
20 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.
21 Sax, S.N.; Bennett, D.J.; Chillrud,  S.N.; Ross, J.; Kinney, P.L.; Spengler, J.D. (2006) A cancer
risk assessment of inner-city teenagers living in New York City and Los Angeles.  Environ
Health Perspect (in press) doi: 10.1289/ehp.8507.  [Online at http://dx.doi.org/1
22 http://www.epa.gov/eogaptil/module3/categorv/category.htm. This document is available in
Docket EPA-HQ-OAR-2005-0036.
23Kittleson, 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.
24 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.
25 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.
26 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.
27 Kittelson, D.B.; Watts, W.F.; Johnson, J.P. (2004) Nanoparticle emissions on Minnesota
highways. Atmos Environ 38: 9-19.
28 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.
                                         3-171

-------
29 Bunn, H.J.; Dinsdale, D.; Smith, T.; Grigg, J. (2001) Ultrafme particles in alveolar
macrophages from normal children. Thorax 56: 932-934.
30Kittelson, D.B.; Watts, W.F.; Johnson, J.P. (2004) Nanoparticle emissions on Minnesota
highways. Atmos Environ 38: 9-19.
31 Westerdahl, D.; Fruin, S.; Sax, T.; Fine, P.M.; Sioutas, C. (2005) Mobile platform
measurements of ultrafme particles and associated pollutant concentrations on freeways and
residential streets in Los Angeles. Atmos Environ 39: 3597-3610.
32Kittelson, D.B.; Watts, W.F.; Johnson, J.P. (2004) Nanoparticle emissions on Minnesota
highways. Atmos Environ 38: 9-19.
33 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.
34 Kamens, R.M.; Jang, M.; Lee, S.; et al. (2003) Secondary organic aerosol formation: some
new and exciting insights. American Geophysical Union 5: 02915.
35Kupiainen, 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.
36 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.
37 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.
38 Zhang, K.M.; Wexler, A.S. (2004) Evolution of particle number distribution near roadways -
Part I:  analysis of aerosol dynamics and its implications for engine emission measurement.
Atmos Environ 38: 6643-6653.
39 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.
40 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.
41 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.
42 Liu,  W.; Zhang, J.; Kwon, J.; Weisel, C.; Turpin, B.; Zhang, L.; Korn, L.; Morandi, M.; Stock,
T.; Colome,  S. (2006) Concentrations and source characteristics of airborne carbonyl compounds
measured outside urban residences. J Air & Waste Manage Assoc 56:  1196-1204.
43 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.
                                         3-172

-------
44 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.
45 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.
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 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.
48 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.
49 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.
50 http://www.mde.state.md.us/Programs/AirPrograms/airData/. This document is available in
Docket EPA-HQ-OAR-2005-0036.
51 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.
52 Hoek G.; Meliefste K.; Cyrys J.; et al.  (2002) Spatial variability of fine particle concentrations
in three European areas. Atmos. Environ. 36: 4077-4088.
53 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.
54 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.
55
  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.
56 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.
57 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.
58
  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.

                                         3-173

-------
59 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.
60 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.
61 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.
62 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.
63 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.
64 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.
65 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.
66 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.
67 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.
68 Leung, P.-L.; Harrison, R.M. (1999) Roadside and in-vehicle concentrations of monoaromatic
hydrocarbons. Atmos Environ 33: 191-204.
69 Weinhold, B. (2001) Pollutants lurk inside vehicles. Environ Health Perspec 109 (9): A422-
A427.
70 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.
71 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.
72 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.

                                         3-174

-------
73 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.
74 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.
75 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.
76 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.
77 Personal communication with FACES Investigators Fred Lurmann, Paul Roberts, and
Katharine Hammond. Data is currently being prepared for publication.
78 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.
79 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.
80 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.
81 Van Roosbroeck, S.; Wichmann, J.; Janssen, N.A.H.; Hoek, G.; van Wijnen, J.H; Lebret, E.;
Brunekreef, B. (2006) Long-term personal exposure to traffic-related air pollution among school
children, a validation study.  Sci Total Environ 368: 565-573.
82
  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.
83
  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.
84 Gulliver, J.; Briggs, D.J. (2004) Personal exposure to particulate air pollution in transport
microenvironments. Atmos Environ 38: 1-8.
85 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.
86 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.

                                        3-175

-------
87 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.
88
  Duci, A.; Chaloulakou, A.; Spyrellis N. (2003) Exposure to carbon monoxide in the Athens
urban area during commuting. Sci Total Environ 309: 47-58.
89 Gulliver, J.; Briggs, DJ. (2004) Personal exposure to parti culate air pollution in transport
microenvironments. Atmos Environ 38: 1-8.
90 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.
91 U.S. EPA (1997) Exposure factors handbook. This document is available in Docket EPA-HQ-
OAR-2005-0036. http ://cfpub . epa. gov/ncea/cfm/recordi splav . cfm?deid= 1 2464 .
92 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.
93 Batterman, S.;  Jia, C.; Hatzivasilis, G.; Godwin, C. (2006) Simultaneous measurement of
ventilation using  tracer gas techniques and VOC concentrations in homes, garages and vehicles.
J Environ Monit 8: 249-256.
94 Morris, S.S. "Influence of attached garages on indoor VOC concentrations in Anchorage
homes." Presented at the annual meeting of the Air & Waste Management Association's Pacific
Northwest Section, November 4, 2004.  This document is available in Docket EPA-HQ-OAR-
2005-0036.
95 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.
96 Fugler, D.; Grande, C.; Graham, L. (2002) Attached garages are likely path for pollutants.
ASHRAE IAQ Applications 3(3):  1-4.
97 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.
98 Isbell, M.; Gordian, M.E.; Duffy, L. (2002) Winter indoor air pollution in Alaska: identifying
a myth. Environ  Pollution 117: 69-75.
99 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.
100 Wallace, L. (1996) Environmental exposure to benzene: an update. Environ Health Perspect
104 (Suppl6):  1129-1136.
101 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.

                                        3-176

-------
102 Batterman, S.; Hatzivasilis, G.; Jia, C. (2005) Concentrations and emissions of gasoline and
other vapors from residential vehicle garages.  Atmos Environ (in press).
103 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.absn.com/]
This document is available in Docket EPA-HQ-OAR-2005-0036.
104 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.
105 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.
106 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.
107 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.
108 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.
109 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.
110 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.
111 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.
112 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.
113 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.
114 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.


                                         3-177

-------
115 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.
116 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.
117 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.
118 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
119 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.
120 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.
121 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.
122 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.
123 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.
124 Egeghy, P.P.; Tornero-Velez, R.T.; Rappaport, S.M. (2000) Environmental and biological
monitoring of benzene during self-service automobile refueling.  Environ Health Perspect 108:
1195-1202.
125 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.
126
   Verma, D.K. and des Tombe, K. (2002) Benzene in gasoline and crude oil: occupational and
environmental implications.  Am Indust Hygiene Assoc J 63: 225-230.
127
   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 16: 362-370
128 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.
129 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.
                                        3-178

-------
130 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/focus-areas/mobile-sources/mobile-
sources-documents.
131 Davis, M.E.; Smith, T.J.; Laden, F.; Hart, I.E.; Ryan, L.M.; Garshick, E. (2006) Modeling
particle exposure in U.S. trucking terminals. Environ Sci Technol 40:  4226-4232.
132 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.
133 U. S. EPA (2006) National-Scale Air Toxics Assessment for 1999.
http://www.epa.gov/ttn/atw/natal999.
134 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
135 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.
136 U. S. EPA (2005) Risk - Air Toxics Risk Assessment.
http://www.epa.gov/ttn/fera/risk_atoxic.html.
137 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.
138 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 and at
http://www.epa.gov/otaq/regs/toxics/r00023.pdf
139 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.
140 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.
141 Cook, R., Strum, M., Touma, J., Palma, T., Thurman, J., Ensley, D., Smith, R.  2006.
Inhalation Exposure and Risk from Mobile Source Air Toxics in Future Years. Journal of
Exposure Science and Environmental Epidemiology, in press.  [Advance  online publication, 27
September 2006; doi:10.1038/sj.jes.  7500529,
http ://www.nature. com/j es/j ournal/vaop/ncurrent/index.html. ]
142 U.S. EPA. 2007. The HAPEM6 User's Guide. Prepared for Ted Palma, Office of Air
Quality Planning and Standards, Research Triangle Park, NC, by Arlene Rosenbaum and

                                         3-179

-------
Michael Huang, ICF International, January 2007.
http://www.epa.gov/ttn/fera/human_hapem.html. This document is available in Docket EPA-
HQ-OAR-2005-0036.
143 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.
144 Luecken, D. J., Hutzell, W. T., Gipson, G. J.  2005. Development and Analysis of air quality
modeling simulations for hazardous air pollutants. Atmospheric Environment.  40: 5087-5096.
145 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.
146 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.
147 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.
148 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
149 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
150 U. S. EPA. 2002. 1996 National-Scale Air Toxics Assessment.
http://www.epa.gov/ttn/atw/nata/
151 Glen, G., Lakkadi, Y., Tippett, J. A., del Valle-Torres M. 1997. Development of
NERL/CHAD: The National Exposure  Research Laboratory Consolidated Human Activity
Database. Prepared  by ManTech Environmental Technology, Inc.  EPA Contract No. 68-D5-
0049. This document is available in Docket EPA-HQ-OAR-2005-0036.
152 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
                                        3-180

-------
153 U.S. EPA. 2007. The HAPEM6 User's Guide.  Prepared for Ted Palma, Office of Air
Quality Planning and Standards, Research Triangle Park, NC, by Arlene Rosenbaum and
Michael Huang, ICF International, January 2007.
http://www.epa.gov/ttn/fera/human_hapem.html. This document is available in Docket EPA-
HQ-OAR-2005-0036.
154 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.
155 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.
156 U.S. EPA. 2007. The HAPEM6 User's Guide.  Prepared for Ted Palma, Office of Air
Quality Planning and Standards, Research Triangle Park, NC, by Arlene Rosenbaum and
Michael Huang, ICF International, January 2007.
http://www.epa.gov/ttn/fera/human_hapem.html. This document is available in Docket EPA-
HQ-OAR-2005-0036.
157 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
158 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
159 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?obiectId=090007d480271237
160 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/pdfs/tsdO 175.pdf
161 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. This document is available in Docket EPA-HQ-OAR-
2005-0036. http ://www. epa.gov/ttn/atw/nata 19997background.html.
162 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= 160003

                                        3-181

-------
163 U. S. EPA.  2006.  National Scale Modeling of Air Toxics for the Mobile Source Air Toxics
Rule (Proposed); Technical Support Document," Report Number EPA-454/R-06-002.
164
   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
165 Chen, C. W.; Gibb, H.  (2003). Procedures for Calculating Cessation Lag.  Reg. Toxicol.
Pharmacol. 38: 157-165.
166 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 and at http://www.epa.gov/CAIR/pdfs/tsd0175.pdf
167 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., 56:547-
558..
168 Isakov, V.; Venkatram, A. (2005) Resolving neighborhood scale in air toxics modeling: a
case  study in Wilmington, California. J Air & Waste Manage. Assoc. 56:559-568.
169 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.
170 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.
171 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.
172 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., 56: 547-
558.
173 Cook, R.; Beidler, A.; Touma, J.S.; Strum M. (2006) Preparing Highway Emissions
Inventories for Urban Scale Modeling: A Case Study in Philadelphia. Transportation Res. Part
D: Transport and Environ. 11: 396-407.
174 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.

175 U.S. EPA.  Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006.
176 U.S. EPA (2007) Review of National Ambient Air Quality Standards for Ozone, Assessment
of Scientific and Technical Information, OAQPS Staff Paper, EPA-452/R-07-003.  This
document is available in Docket EPA-HQ-OAR-2005-0036.

177Bates, D.V.; Baker-Anderson, M.;  Sizto, R. (1990) Asthma attack periodicity: a study of

                                        3-182

-------
hospital emergency visits in Vancouver.  Environ. Res. 51: 51-70.

178Thurston, 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.

179Thurston, G.D.; Ito, K.; Hayes,  C.G.; Bates, D.V.; Lippmann, M. (1994) Respiratory hospital
admissions and summertime haze  air pollution in Toronto, Ontario: consideration of the role of
acid aerosols. Environ. Res. 65: 271-290.

180Lipfert, F.W.; Hammerstrom, T. (1992) Temporal patterns in air pollution and hospital
admissions. Environ. Res.  59: 374-399.

181Burnett, R.T.; Dales, R.E.; Raizenne, M.E.; Krewski, D.; Summers, P.W.; Roberts, G.R.;
Raad-Young, M.; Dann,T.; Brook, J. (1994) Effects of low ambient levels of ozone and sulfates
on the frequency of respiratory admissions to Ontario hospitals. Environ. Res. 65: 172-194.

182U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006.

183 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006.

184Devlin, R. B.; McDonnell, W. F.; Mann, R.; Becker, S.; House, D. E.; Schreinemachers, D.;
Koren, H. S. (1991) Exposure of humans to ambient levels of ozone for 6.6 hours causes cellullar
and biochemical changes in the lung. Am. J. Respir. CellMol. Biol. 4: 72-81.

185Koren, H. S.; Devlin, R. B.; Becker, S.; Perez, R.; McDonnell, W. F. (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.

186Koren, H. S.; Devlin, R. B.; Graham, D. E.; Mann, R.; McGee, M. P.; Horstman, D. H.;
Kozumbo, W. J.; Becker, S.; House, D. E.; McDonnell, W. F.; Bromberg, P. A. (1989a) Ozone-
induced inflammation in the lower airways of human subjects. Am. Rev. Respir. Dis.  139: 407-
415.

187Schelegle, E.S.; Siefkin, A.D.; McDonald, RJ. (1991)  Time course of ozone-induced
neutrophilia in normal humans. Am. Rev. Respir. Dis. 143:1353-1358.

188U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006.

189Hodgkin, 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.

190Euler, 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

                                        3-183

-------
and nitrogen dioxide in California Seventh-day Adventist residents.  Arch. Environ. Health 43:
279-285.

191 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.

192U.S. EPA.  Review of the National Ambient Air Quality Standards for Ozone: Policy
Assessment of Scientific and Technical Information, OAQPS Staff Paper- Second Draft. EPA-
452/D-05-002.

193 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006.

194U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006.

195Avol, E. L.; Trim, S. C.; Little, D. E.; Spier, C. E.;  Smith, M. N.; Peng, R.-C.; Linn, W. S.;
Hackney, J. D.; Gross, K. B.; D'Arcy, J. B.; Gibbons, D.; Higgins, I. T. T. (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.

196Higgins, I.  T. T.; D'Arcy, J. B.; Gibbons, D. L; Avol, E. L.; Gross, K. B. (1990) Effect of
exposures to ambient ozone on ventilatory lung function in children. Am. Rev. Respir. Dis. 141:
1136-1146.

197Raizenne, M. E.; Burnett, R. T.; Stern, B.; Franklin, C. A.; Spengler, J. D. (1989) Acute lung
function responses to ambient acid aerosol exposures in children. Environ. Health Per sped. 79:
179-185.

198Raizenne, 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.

199 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).

200Spektor, D. M.; Thurston, G. D.; Mao, J.; He, D.; Hayes, C.; Lippmann, M. (1991) Effects of
single- and multiday ozone exposures on respiratory function in active normal children. Environ.
Res.  55: 107-122.

201Spektor, D. M.; Lippman, M.; Lioy, P. J.; Thurston, G. D.;s Citak, K.; James, D. J.; Bock, N.;
                                         3-184

-------
Speizer, F. E.; Hayes, C. (1988a) Effects of ambient ozone on respiratory function in active,
normal children. Am. Rev. Respir. Dis.  137: 313-320.

202U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006.

203Hazucha, M. J.; Folinsbee, L. J.; Seal, E., Jr. (1992) Effects of steady-state and variable ozone
concentration profiles on pulmonary function. Am. Rev. Respir. Dis. 146: 1487-1493.

204Horstman, D.H.; Ball, B.A.; Folinsbee, L.J.; Brown, J.; Gerrity, T. (1995) Comparison of
pulmonary responses of asthmatic and nonasthmatic subjects performing light exercise while
exposed to a low level of ozone. Toxicol. Ind. Health.

205Horstman, D.H.; Folinsbee, L.J.; Ives, P.J.; Abdul-Salaam, S.; McDonnell, W.F. (1990) Ozone
concentration and pulmonary response relationships for 6.6-hour exposures with five hours of
moderate exercise to 0.08, 0.10, and 0.12 ppm. Am. Rev. Respir. Dis. 142: 1158-1163.
206 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.
207 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.
208 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.
209 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.
210 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.
211 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.
212 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.
213 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006. This
document is available in Docket EPA-HQ-OAR-2005-0036.
214 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.
215 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006.This
document is available in Docket EPA-HQ-OAR-2005-0036.
                                        3-185

-------
216 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.
217 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006.This
document is available in Docket EPA-HQ-OAR-2005-0036.
218 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006. This
document is available in Docket EPA-HQ-OAR-2005-0036.
219 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006. This
document is available in Docket EPA-HQ-OAR-2005-0036.
220 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.
221 Winner, W.E., 1994. "Mechanistic analysis of plant responses to air pollution." Ecological
Applications, 4(4):651 -661.
222 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006. This
document is available in Docket EPA-HQ-OAR-2005-0036.
223 U U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final).
U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006. This
document is available in Docket EPA-HQ-OAR-2005-0036.
224 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.
225 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.
226 De Steiguer, I, J. Pye, C. Love. 1990. "Air Pollution Damage to U.S. Forests." Journal of
Forestry, Vol 88(8) pp. 17-22.
227 Pye, J.M. 1988.  "Impact of ozone on the growth and yield of trees: A review." Journal of
Environmental Quality 17 pp.347-360.
228 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006. This
document is available in Docket EPA-HQ-OAR-2005-0036.
229 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006. This
document is available in Docket EPA-HQ-OAR-2005-0036.

                                         3-186

-------
230 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.
231 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.
232 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-05/004aF-cF, 2006. This
document is available in Docket EPA-HQ-OAR-2005-0036.
233Kopp, 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.
234 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.
235 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.
236 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.
237 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.
238 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.
239 U.S. EPA 2006. Provisional Assessment of Recent Studies on Health Effects of Particulate
Matter Exposure. EPA/600/R-06/063. This document is available in Docket EPA-HQ-OAR-
2005-0036.
240 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 Per sped. 108(10):941-
947.
241 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.
242 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.
243 Dockery, DW; Pope, CA, III; Xu, X; et al. 1993. "An association between air pollution and
mortality in six U.S. cities."  N. Engl. J. Med. 329:1753-1759.
                                         3-187

-------
244 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. Med. Assoc. 287:
1132-1141.
245 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.
246 Jerrett, M; Burnett, RT; Ma, R; et al. 2005. "Spatial analysis of air pollution and mortality in
Los Angeles. Epidemiology 16(6):727-736.
247 Kiinzli, N.; Jerrett, M.; Mack, W.J.; et al. 2005. Ambient air pollution and atherosclerosis in
Los Angeles. Environ Health Perspect. 113:201 -206.
248 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.

249 Van Vliet, P.; Knape, M.; de Hartog, J.; Janssen, N.; Harssema, H.; Brunekreef, B. (1997).
Motor vehicle exhaust and chronic respiratory symptoms in children living near freeways. Env.
Research 74: 122-132.
250 Brunekreef, B., Janssen, N.A.H.; de Hartog, J.; Harssema, H.; Knape, M.; van Vliet, P.
(1997).  Air pollution from truck traffic and lung function in children living near roadways.
Epidemiology 8:298-303.
251 Kim, J.J.; Smorodinsky, S.; Lipsett, M.; Singer, B.C.; Hodgson, A.T.; Ostro, B (2004).
Traffic-related air pollution near busy roads:  The East Bay children's respiratory health study.
Am. J. Respir. Crit. Care Med.  170:  520-526.
252
   U.S. EPA. 2006.  Regulatory Impact Analysis for the final PM NAAQS. This document is
available in Docket EPA-HQ-OAR-2005-0036.
253 National Research Council, 1993. Protecting Visibility in National Parks and Wilderness
Areas. National Academy of Sciences Committee on Haze in National Parks and Wilderness
Areas. National Academy Press, Washington, DC.  This book can be viewed on the National
Academy Press Website at http://www.nap.edu/books/0309048443/html/ and is available in
Docket EPA-HQ-OAR-2005-0036.
254 U.S. EPA (2004) Air Quality Criteria for Particulate Matter (Oct 2004), Volume I Document
No. EPA600/P-99/002aF and Volume II Document No. EPA600/P-99/002bF. This document is
available in Docket EPA-HQ-OAR-2005-0036.
255 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.
256 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.

                                         3-188

-------
257 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.
258
   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.
259 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.
260 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.
261 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.
262
   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.
263 U.S. EPA (2004) National Coastal Condition Report II. Office of Research and Development/
Office of Water. EPA-620/R-03/002.
264 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.
265 Kim, G., N. Hussain, J.R. Scudlark, and T.M. Church. 2000.  Factors influencing the
atmospheric depositional fluxes of stable Pb, 210Pb, and 7Be into Chesapeake Bay.  J. Atmos.
Chem. 36: 65-79.
266 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.
267 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.
268 Smith, W.H. 1991. Air pollution and forest damage." Chemical Engineering News, 69(45):
30-43.
269 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.
270 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.
271 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.
272 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")

                                        3-189

-------
273 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.
274 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.
275 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.
276 Rice, K.C. 1999. Trace element concentrations in streambed sediment across the
conterminous United States. Environmental Science & Technology, 33:  2499-2504.
277 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.
278 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.
279 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.
280 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.
281 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.
282 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.
283 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.
284 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.
285 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.
                                        3-190

-------
286 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.
287 Van Metre, P.C.; Mahler, B.J.; and Furlong, E.T. 2000. Urban sprawl leaves its PAH
signature." Environmental Science & Technology, 34: 4064-4070.
288 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.
289 Tuhackova, J., Cajthaml, T.; Novak, K.; et al. 2001. Hydrocarbon deposition and soil
microflora as affected by highway traffic. Environmental Pollution, 113: 255-262.
290 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.
291 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.
292 Finkelstein, M.M.; Jerrett, M.; Sears, M.R. (2004) Traffic air pollution and mortality rate
advancement periods. Am. J. Epidemiol.  160: 173-177.
293
   Finkelstein, M.M.; Jerrett, M.; Sears, M.R. (2005) Environmental inequality and circulatory
disease mortality gradients.  J. Epidemiol. Community Health 59: 481-486.
294 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.
295 Gehring, U.; Heinrich, J.; Kramer, U.; Grote, V.; Hochadel, M.; Dugiri, D.; Kraft, M.;
Rauchfuss, K.; Eberwein, H.G.; Wichmann, H-E. (2006) Long-term exposure to ambient air
pollution and cardiopulmonary mortality in women. Epidemiology 17: 545-551.
296 Lipfert, F.W.; Baty, J.D.; Miller, J.P.; Wyzga, R.E. PM2.5 constitutents and related air
quality variables as predictors of survival in a cohort of U.S. military veterans. Inhalation
Toxicol. 18: 645-657.
297 Maheswaran, M.; Elliott, P. (2003) Stroke mortality associated with living near main roads in
England and Wales. A geographical study. Stroke 34: 2776-2780.
298 Roemer, W.H.; van Wijnen, J.H. (2001) Daily mortality and air pollution along busy streets in
Amsterdam, 1987-1998. Epidemiology 12: 649-653.
299 Heinrich, J.; Wichmann, H-E. (2004) Traffic related pollutants in Europe and their effect on
allergic disease.  Current Opin. Clinical Epidemiol. 4: 341-348.
300 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.
                                         3-191

-------
301 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. J. Respir. Crit. Care Med.  170:
520-526.
302 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.
303 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.
304 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.
305 Heinrich, J.; Wichmann, H-E. (2004) Traffic related pollutant in Europe and their effect on
allergic disease. Current Opin. ClinicalEpidemiol. 4: 341-348.
306 McConnell,  R.; Berhane, K.; Yao. L., Jerrett, M.; Lurmann, F.; Gilliland, F.; Kunzli, N.;
Gauderman, J.;  Avol, E.; Thomas, D.; Peters, J. (2006) Traffic, susceptibility, and childhood
asthma. Environ. Health Perspect. 114:766-772.
307 Hoffmann, B.; Moebus, S.; Stang, A.; Beck, E.M.; Dragano, N.; Mohlenkamp, S.;
Schermund, A.; Memmesheimer, M.; Mann, K.; Erbel, R.; Jockel, K.H. (2006) Residence close
to high traffic and prevalence of coronary heart disease.  Eur. Heart J. 27: 2696-2702.
308 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.
309 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.
169:934-940.
310 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.
311
   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.
312
   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.
313 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.
314 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.
315 Ritz B; Yu F; Fruin  S; et al. (2002) Ambient air pollution and risk of birth defects in Southern
California. Am. J. Epidemiol. 155:17-25.
316 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.
                                          3-192

-------
317 Perera, P.P.; Rauh, V.; Whyatt, R.M.; Tsai, W.Y.; Tang, D.; Diaz, D.; Hoepner, L.; Barr, D.;
Tu, Y.H.; Camann, D.; Kinney, P. (2006) Effect of prenatal exposure to airborne polycyclic
aromatic hydrocarbons on neurodevelopment in the first 3 years of life among inner-city
children. Environ. Health Perspect. 114:  1287-1292.
318 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= 160003
319 U. S. EPA. 2002. Toxicological Review of Benzene (Noncancer effects). Report No. EPA
635/R-02/001-F.
320 Savitz, D.A.; Feingold, L. (1989) Association of childhood cancer with residential traffic
density. Scand. J. Work Environ. Health  15: 360-363.
321 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.
322 Feychting, M.; Svensson, D.; Ahlbom, A. (1998) Exposure to motor vehicle exhaust and
childhood cancer. Scan. J. Work Environ. Health 24(1): 8-11.
323
   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.
324
   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.
325 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.
326 Reynolds, P.; Von Behren, J.; Gunier, R.B.; et al. (2004) Residential exposure to traffic in
California and childhood cancer. Epidemiology 15: 6-12.
327 Knox, E.G. (2005) Oil combustion and childhood cancers. J. Epidemiol. Community Health
59: 755-760.
328 U.S. EPA. 2007.  The HAPEM6 User's Guide. Appendix B. Prepared for Ted Palma, Office
of Air Quality Planning and Standards, Research Triangle Park, NC, by Arlene Rosenbaum and
Michael Huang, ICF International, January 2007.
http://www.epa.gov/ttn/fera/human_hapem.html.  This document is available in Docket EPA-
HQ-OAR-2005-0036.


329United 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.
330This 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.
                                         3-193

-------
331Garshick, 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.
332Green, 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.
333Gunier, R.B.; Hertz, A.; Von Behren, I; Reynolds, P.  (2003) Traffic density in California:
socioeconomic and ethnic differences among potentially exposed children. J. Expos. Anal.
Environ. Epidemiol. 13: 240-246.
334 Emmerich, S.; Gorfain, J.C.; Huang, M.: Howard-Reed, C. (2003) Air and pollutant transport
from attached garages to residential living spaces. National Institute of Standards and
Technology report number NISTIR 7072.
335 Gordon, S.M.; Callahan, P.J.;  Nishioka, M.G; Brinkman, M.C.; O'Rourke, M.K.; Lebowitz,
M.D.; Moschandreas, DJ. (1999) Residential environmental measurements in the National
Human Exposure Assesment Survey (NHEXAS) pilot study in Arizona:  preliminary results for
pesticides and VOCs. J. Expos. Analysis. Environ. Epidemiol. 9: 456-470.
336 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 91st Annual Meeting & Exhibition. Paper no. 98-A504.
337 Batterman, S.; Hatzivasilis, G.; Jia, C. (2006) Concentrations and emissions of gasoline and
other vapors from residential vehicle garages. Atmos. Environ.  40: 1828-1844.
338 Memorandum from Tom Long, Ted Johnson, Jim Laurenson, and Arlene Rosenbaum to Ted
Palma. Subject: Development of penetration and proximity microenvironment factor
distributions for the HAPEM5 in support of the 1999 National-scale Air  Toxics Assessment
(NATA).  EPA contract no. GS-10F-0124J.
http://www.epa.gov/ttn/fera/hapem5/hapem5_me_factor_memo.pdf
339
   Palisade Corporation. Guide to Using @Risk. Risk Analysis and Simulation Add-In for
Microsoft Excel. Version 4.5. Newfield: Palisade Corporation. Pages i-iv.
340 Batterman, S.; Jia, C.; Hatzivasilis, G.; Godwin, C. (2006) Simultaneous measurement of
ventilation using tracer gas techniques and VOC concentrations in homes, garages and vehicles.
J Environ. Monit. 8: 249-256.
341 Data obtained from Dr. Clifford Weisel, EOHSI. E-mail correspondence dated November 3,
2006 from Dr. Clifford Weisel, EOHSI to Rich Cook and Chad Bailey, U.S. EPA. Docket
number Docket EPA-HQ-OAR-2005-0036.
342 J.L. Schoor, "Chapter 1: Chemical Fate and Transport in the Environment," in Fate of
Pesticides and Chemicals in the Environment (J.L. Schoor, Ed.). John Wiley & Sons, 1992, pp.
1-24.
343 Murray, D.M.; Burmaster, D.E. (1995) Residential air exchange rates in the United States:
empirical and estimated parametric distributions by season and climatic region. Risk Analysis
15: 459-465.
344 Fugler, D.; Grande, C.; Graham, L. (2002) Attached garages are likely path for pollutants.
ASHRAE IAQ Applications 3(3):  1-4.
                                         3-194

-------
345 Sheltersource, Inc. (2002) Evaluating Minnesota homes. Final report. Prepared for
Minnesota Department of Commerce. U.S. Department of Energy grant DE-FG45-96R530335.
346 Isbell, M.A.; Stolzberg, R.J.; Duffy, L.K. (2005) Indoor climate in interior Alaska:
simultaneous measurement of ventilation, benzene and toluene in residential air of two homes.
Sci. Total Environ. 345: 31-40.
347 Batterman, S.; Jia, C.; Hatzivasilis, G.; Godwin, C. (2006) Simultaneous measurement of
ventilation using tracer gas techniques and VOC concentrations in homes, garages and vehicles.
J. Environ. Monit. 8: 249-256.
348 Batterman, S.; Jia, C.; Hatzivasilis, G. (2006) Migration of volatile organic compounds from
attached garages to residences:  a major exposure source. Environ Res (In Press).
349 George, M.; Kaluza, P.; Maxwell, B.; Moore, G.; Wisdom, S. (2002) Indoor air quality &
ventilation strategies in new homes in Alaska. Final Report.  Alaska Building Science Network.
350 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.
351 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.
352 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.
353 Schlapia, A.; Morris, S.S. (1998) Architectural, behavioral, and environmental factors
associated with VOCs in Anchorage homes. Presented at the Air & Waste Management
Association's 91st Annual Meeting & Exhibition. June 14-18, 1998, San Diego, CA.
354 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.
355 Batterman, S.; Jia, C.; Hatzivasilis, G. (2006) Migration of volatile organic compounds from
attached garages to residences:  a major exposure source. Environ Res. (In  Press).
356 Data obtained from Dr. Clifford Weisel, EOHSI. E-mail correspondence dated November 3,
2006 from Dr. Clifford Weisel, EOHSI to Rich Cook and Chad Bailey, U.S. EPA. Docket
number Docket EPA-HQ-OAR-2005-0036.
357 U.S. Energy Information Administration. Residential Energy Consumption Survey, 2001.
Table HCl-4a. Housing Unit Characteristics by Type of Housing Unit, Million U.S. Households,
2001 [Online at http://www.eia.doe.gov/emeu/recs/recs2001/hcjdf/housunits/hcl-
4a_housingunits2001 .pdfl
358 Environmental Protection Agency. (1997) Exposure factors handbook.  National Center for
Environmental Assessment report EPA/600/P-95/002Fa-c [Online at
http://www.epa.gov/ncea/pdfs/efh/front.pdf]
                                         3-195

-------
359 Batterman, S.; Jia, C.; Hatzivasilis, G.; Godwin, C. (2006) Simultaneous measurement of
ventilation using tracer gas techniques and VOC concentrations in homes, garages and vehicles.
J. Environ. Monit. 8:  249-256.
360 Graham, L.A.; Noseworthy, L.; Fugler, D.; O'Leary, K.; Karman, D.; Grande, C. (2004)
Contribution of vehicle emission from an attached garage to residential indoor air pollution
levels. J. Air & Waste Manage. Assoc. 54: 563-584.
361 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.
                                         3-196

-------
Final Regulatory Impact Analysis
                              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	5
     4.1.3 Small Volume Manufacturers, Importers, and Alternative Fuel Vehicle Converters. 7
   4.2  Petroleum Refining Industry	7
     4.2.1 Gasoline Supply	8
     4.2.2 Gasoline Demand	8
     4.2.3 Industry Organization	9
     4.2.4 Gasoline Market Data	9
   4.3  Portable Fuel Container Industry	10
     4.3.1 Manufacture and Distribution	10
     4.3.2 Container Use	10
     4.3.3 Market Structure	11
     4.3.4 Market Entry	11
                                          4-1

-------
Final Regulatory Impact Analysis

                     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 fuel container 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 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 Table4.1.-l 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. Becoming 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
  EPA defines small volume manufacturers to be those with total U.S. sales of 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

-------
Final Regulatory Impact Analysis




        Previously, it has been relatively easy to characterize manufacturers as "domestic" or
                                             4-3

-------
Final Regulatory Impact Analysis
g
LJJ


D


O




LJJ
 S  
-------
Final Regulatory Impact Analysis

"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.
                                          4-5

-------
Final Regulatory Impact Analysis
                           Figure 4.1-1.
  LDV MANUFACTURER SALES
         2.4% -,
       2.9%
      4.0%
           2.4% -, r 1.4%
                          24.4%
    5.9%
   7.0%
   11.0%
                             13.5%
GENERAL MOTORS
FORD MOTOR CO.
DAIMLER CHRYSLER
TOYOTA MOTOR CO.
AMERICAN HONDA
NISSAN MOTOR CO.
HYUNDAI GROUP.
I | VW of AMERICA
• BMW GROUP
D MAZDA
• SUBARU
• MITSUBISHI
D ALL OTHERS
          13.7%
                      10.5%
                           Figure 4.1-2.
  LOT MANUFACTURER SALES
           0.3%
          2.5% -,
        4.9%
                0.8%
                 0.8%
  10.9%
                           30.1%
                                  GENERAL MOTORS
                                  FORD MOTOR CO.
                                  DAIMLER CHRYSLER
                  HYUNDAI GROUP.
                  VW of AMERICA
                  BMW GROUP
                                  TOYOTA MOTOR CO.   D MAZDA
                                  AMERICAN HONDA
                                  NISSAN MOTOR CO.
                  ALL OTHERS
    17.6%
                      24.8%

-------
Final Regulatory Impact Analysis

       For regulatory purposes, LDVs and LDTs 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 equally effective for both the
lighter and the heavier vehicles. Therefore, the Tier 2 emission standards now make no distinction
between weight categories.  In addition, Tier 2 applies to medium duty passenger vehicles
(MDPVs), i.e. passenger vehicles between 8,500 and 10,000 Ibs. GVW.  These are primarily the
very large SUVs, and passenger vans.

       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
  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).

                                           4-7

-------
Final Regulatory Impact Analysis

behavior.2  The information contained in the report is summarized below, supplemented by addi-
tional information found in this RIA and in other sources.

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

                                            4-8

-------
Final Regulatory Impact Analysis

is, for example, by purchasing a car with better fuel economy.

4.2.3   Industry Organization

       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.
High 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.
                                            4-9

-------
Final Regulatory Impact Analysis

       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 Fuel Container Industry

       EPA also contracted with RTI International for a characterization of the PFC 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.5'6 This report is also summarized below, and is again supplemented by additional information
found in this RIA and in other sources. PFCs include gasoline, kerosene, and diesel containers.

4.3.1   Manufacture and Distribution

       PFCs 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. PFCs
include utility jugs that are marketed for use with fuels, which are often used to refuel recreational
products such as personal watercraft and all-terrain vehicles.  PFCs range in capacity from a gallon
or less to over 6 gallons.  Standard PFCs 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 containers are made of high-density polyethylene (HOPE) 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, gas cans are colored red during the
manufacturing process. Diesel containers are colored yellow and kerosene containers are colored
blue to help consumers avoid misfueling of equipment.  Industry and other sources indicate that gas
cans and diesel and kerosene containers are distributed by manufacturers through their distribution
centers to major retail establishments.  Utility jugs are sold in several colors and are more often sold
through online retailers.

4.3.2   Container Use

       PFCs 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.7 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 commercial
applications as farming, logging, construction, lawn care, and automotive applications such as repair

                                           4-10

-------
Final Regulatory Impact Analysis

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 PFC will have a life expectancy of 3 to 5 years before
it needs to be replaced.

       The demand for fuel containers 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
equipment such as lawn  and garden equipment or recreational vehicles. Therefore, 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.

       Gasoline container sales for 2002, the latest year for which we were able to develop data,
were about 24.4 million units (including utility jug sales which were estimated to  be about 2.4
million units).  Diesel and kerosene container annual sales are estimated to be about 620,000 and 1
million units, respectively. Although the PFC manufacturing industry has become fairly
concentrated, with one firm accounting for more than half of all U.S. container 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 PFC 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). Data
for utility jug manufacturers was scarce, but we believe that there are likely about 5 manufacturers of
these containers, including Scribner Plastics.  There are other gasoline container 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 safety cans used in an industrial setting), which are not be covered by
the standards. These companies include Eagle Manufacturing and Protectoseal  Company. Table
4.3-1 provides relevant data about these firms.

4.3.4   Market Entry

       There are very few barriers to entering the PFC market. Only about 2 percent of the
containers 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 PFCs are in fact classified in the U.S. Economic

                                           4-11

-------
Final Regulatory Impact Analysis

Census as "All other plastics product manufacturing." Since manufacturing such PFCs 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, safety 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.
                                           4-12

-------
Final Regulatory Impact Analysis
                                              Table 4.3-1.  Manufacturers*
Ultimate Parent
Blitz USA
Eagle Manufacturing
Midwest Can
No-spill Research Inc.
Protectoseal Co.
Scepter Mfg., Ltd.
Scribner Plastics
The Plastics Group
Company name
Blitz USA
Eagle Manufacturing
Midwest Can
No-spill Research Inc.
Protectoseal Co.
Scepter Mfg., Ltd.
Scribner Plastics
Wedco Molded Prod.
Sales (Smillion)
20-50
50-100
20-50
2.5-5
20-50
10-20
5-10
20-50
Employment
200
100-249
45
5
100-249
200
20-49
600
Comments
Consumer market
Primarily Metal Safety Cans
Consumer market
Limited Distribution
Primarily Industrial
Canadian-Consumer
Specialty Containers
Consumer Market
* Businesses Engaged In NAICS Code 326119, All Other Plastic Product Manufacturing, Or NAICS Code 332431, Metal Can
Manufacturing

Source: Characterizing Gas Can Markets, a Profile," RTI International, Final Report, EPA Contract 68-D-99-024.
                                                          4-1

-------
Final Regulatory Impact Analysis

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 "Gas Can Industry Profile Updates", memorandum from RTI to EPA, January 4, 2007.

7 Memorandum from Terrance R. Karels, Consumer Product Safety Commission, January 3, 2003.
                                         4-1

-------
Final Regulatory Impact Analysis
                              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	4
           5.1.1.1.1  Idle Speed and Air Flow Control	4
           5.1.1.1.2  Spark Control	5
           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	11
     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 Programs	18
           5.1.2.2.1  2004 Chevrolet  Trailblazer Testing	19
           5.1.2.2.2  2006 Chrysler 3OOC Testing	22
   5.2    Feasibility of Evaporative Emissions Standards for Vehicles	25
                                           5-1

-------
Final Regulatory Impact Analysis
               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 National Low Emission
Vehicle (NLEV) program, contain stringent standards for light-duty vehicles that have 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 (ARB) originally projected that many vehicles
would require electrically-heated catalysts to meet their Low Emission Vehicle I (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 Low
Emission Vehicle II (LEV II) programs, currently being phased-in, have projected that some
additional emission control hardware and techniques may be required.  However, initial
                                          5-2

-------
Final Regulatory Impact Analysis
indications from the Tier 2 vehicles already certified indicate that increases in emission control
hardware have been kept to a minimum, likely to minimize cost.

       The Tier 2 program requires reductions in all regulated pollutants, but the largest
reductions are required for oxides of nitrogen (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  non-methane
hydrocarbons (NMHC) controls from the stringent NMOG standards (NMOG consists primarily
of NMHC) under the NLEV and LEV I programs. In fact, the NMOG standards for a Tier 2 Bin
5 package are the same for the passenger car and light-duty truck as those established under the
NLEV program. One of the largest challenges manufacturers have encountered under the Tier 2
program is 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 has
reached operational temperature 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 reaching operating temperature 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 accomplished 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
                                           5-3

-------
Final Regulatory Impact Analysis
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

       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 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, in which contribute to 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, 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 results in less
hydrocarbons from unburned fuel, 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
                                            5-4

-------
Final Regulatory Impact Analysis
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

       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.

       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 results in greater emission reductions than when either feature is used
independently.  Additionally, utilizing elevated idle speeds while retarding the  timing can offset
any engine vacuum level  concerns encountered when only retarded 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 generate 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  new. 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 during all modes of operation.  With the advent of the three-way catalyst, manufacturers
                                           5-5

-------
Final Regulatory Impact Analysis
began to use engine control modules to activate electric air pumps to only reduce start emissions
at 75° F, typically on vehicle packages with specific cold start emission challenges. For
example, vehicles that have 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 identical vehicle models that are sold in both Europe and the
U.S. are equipped with secondary air injection that does appear to be used  at 20° F on the U.S.
model, based on our analysis of the certification data. This is attributable to shared 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.
                                           5-6

-------
Final Regulatory Impact Analysis
       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),
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
                                           5-7

-------
Final Regulatory Impact Analysis
gasoline-fueled engine at stoichiometry provides the exhaust aftertreatment with oxygen required
for oxidation of HC and CO. Therefore, the amount of time requiring enrichment should be
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 engine control module (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.
                                           5-8

-------
Final Regulatory Impact Analysis
       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
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 indicates, 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
                                           5-9

-------
Final Regulatory Impact Analysis
followed by a closed-valve approach, described previously, once the intake valve is heated.
Many similar approaches which can be implemented at any cold start temperature are detailed in
past Society of Automotive Engineers (SAE) papers.2

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.  All of these concepts are equally
feasible at cold temperatures since they are not temperature dependent.

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.
                                           5-10

-------
Final Regulatory Impact Analysis


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
standards. We believe that manufacturers will use these same Tier 2 technologies in order to
meet the 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 20° F NMHC standard.

5.1.2   Data Supporting Cold NMHC Standard Technical Feasibility

       Data to support the feasibility of complying with the 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 standards we are finalizing today.

       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.  For example, manufacturers 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. Put another way, this design attempts to
account for maximum normal operating variability.

       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 or
tested by the EPA, were not subject to cold NMHC standards. The data represent the cold
NMHC emissions as tested, and suggest that a significant number of vehicles are within reach of
the standards we are finalizing today.

5.1.2.1        Certification Emission Level

       Currently, 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 and light-duty trucks.  (Cold CO requirements for medium-duty passenger vehicles
do not begin until model year 2008.4) 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)
                                          5-11

-------
Final Regulatory Impact Analysis
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 2004-2007 model year full useful life certification data for vehicles
certified to nationwide Tier 2 standards, interim non-Tier 2 standards, and California program
standards. Lists were compiled from certification data submissions that reported unrounded cold
THC results and for which an associated FTP full useful life deterioration factor (DF) was
available.  The DF was 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 through 5.1-6 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
2007
LDV/LLDTs
41
42
44
39
HLDT/MDPVs 
-------
Final Regulatory Impact Analysis
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.
             Table 5.1-3. 2004 Model Year Vehicles with Certification Data
             for Total Hydrocarbons at or below the 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
Kl 5 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

-------
Final Regulatory Impact Analysis
             Table 5.1-4.  2005 Model Year Vehicles with Certification Data
             for Total Hydrocarbons at or below the 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

-------
Final Regulatory Impact Analysis
             Table 5.1-5.  2006 Model Year Vehicles with Certification Data
             for Total Hydrocarbons at or below the 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

-------
Final Regulatory Impact Analysis
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
                                                5-16

-------
Final Regulatory Impact Analysis
             Table 5.1-6.  2007 Model Year Vehicles with Certification Data
             for Total Hydrocarbons at or below the Cold NMHC Standard
MANUFACTURER
CAR LINE
COLD TOTAL HC LEVEL
LDV/LLDTs
BMW
DAIMLER-CHRYSLER
GENERAL MOTORS
GENERAL MOTORS
GM/DAEWOO
HONDA
HONDA
HONDA
HONDA
HYUNDAI
LAMBORGHINI
LOTUS
MAZDA
MAZDA
MAZDA
MERCEDES-BENZ
MERCEDES-BENZ
MITSUBISHI
MITSUBISHI
PORSCHE
TOYOTA
TOYOTA
TOYOTA
VOLKSWAGEN
VOLVO
335X1
CROSSFIRE ROADSTER
IMPALA, UPLANDER FWD
VUE AWD
FORENZA WAGON
ACCORD HYBRID
ACCORD
CIVIC, CIVIC HYBRID, ELEMENT 4WD, S2000
CR-V 4WD, FIT, MDX 4WD, ODYSSEY 2WD, RDX 4WD, RL, TL
HD-4DR, MC(3/4DR)
GALLARDO COUPE, MURCIELAGO
ELISE/EXIGE
MAZDA 3
MAZDAS
MAZDA 6, MAZDA 6 SPORT WAGON
SL55 AMG
CLK550 (CABRIOLET)
GALANT
ECLIPSE SPYDER
CAYMAN S
CAMRY HYBRID
PRIUS
YARIS, RX 400H 4WD
PASSAT WAGON
S40
0.3
0.2
0.3
0.2
0
0.1
0.1-0.2
0.3
0.2
0.3
0.1
0.3
0.2-0.3
0.2
0.3
0.2
0.3
0.2-0.3
0.3
0.2
0.1
0.2
0.3
0.3
0.1
HLDT/MDPVs
BENTLEY MOTORS LTD.
DAIMLER-CHRYSLER
GENERAL MOTORS
GENERAL MOTORS
GENERAL MOTORS
HONDA
MERCEDES-BENZ
MERCEDES-BENZ
TOYOTA
TOYOTA
VOLKSWAGEN
VOLVO
CONTINENTAL GTC
RAM 1500 PICKUP 2WD
K1500 AVALANCHE 4WD
CHEVY C1500 CLASSIC PICKUP 2WD, ISUZU ASCENDER SUV
4WD, FULL SIZE CONVERSION VAN AWD
ESCALADE EXTAWD
RIDGELINE 4WD
R500 4MATIC
MAYBACH 62, GL450
GX470
LX 470, TOYOTA TUNDRA 4WD
TOUAREG
XC90
0.3
0.4-0.5
0.3-0.4
0.4
0.5
0.2
0.2
0.5
0.4
0.5
0.5
0.3, 0.5
                                         5-17

-------
Final Regulatory Impact Analysis
5.1.2.2
EPA Test Programs
       To further assess the feasibility of meeting the new cold temperature NMHC standards
through changes to engine and emission control system calibration, EPA performed a test
program involving two Tier 2 vehicles.  We considered several key aspects when selecting test
vehicles for a feasibility study. First, the vehicles currently produce 20° F NMHC levels that are
significantly higher than the standards we are finalizing today. Second, since higher vehicle
weight poses additional challenge, we considered heavier GVWR vehicles.  Finally, the
technological approach chosen by the manufacturer to meet stringent 75° F Tier 2 standards was
also considered. Specifically, we considered secondary air injection technology and close
coupled catalyst technology. Specifications for the vehicles included in the test program are
provided in Table 5.1-6.

                          Table 5.1-6. EPA Test Vehicle  Specifications
Vehicle
2004
Chevrolet
Trailblazer
2006
Chrysler
300C
Engine Family
4GMXT04.2185
6CRXV05.7VEO
Powertrain
4.2L 16
4-speed auto
Rear 2-WD
5.7LV8
5-speed auto
Rear 2-WD
GVWR
5550 Ibs.
5300 Ibs.
Emission
Class
Tier 2 Bin 5
Tier 2 Bin 5
Mileage
36,500
2,000
       The vehicles were tested at 20° F following EPA cold FTP test procedures established in
40 CFR 86.230-94.  In addition to regulated pollutant measurements, we also measured NMHC,
NOx, and direct 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
2006 Chrysler
300C
Test Weight
5000 Ibs.
4500 Ibs.
20° F Target
Coefficients
A=38.97
B=1.2526
C=.02769
A=61.09
B=.3105
C=.0247
A "Cold Chamber Sampling System Diagram," PDF file from test lab.
B Available at www.epa.gov/otaq/cert/dearmfr/dearmfr.htm.
                                          5-18

-------
Final Regulatory Impact Analysis
5.1.2.2.1      2004 Chevrolet Trailblazer Testing

       The 2004 Chevrolet Trailblazer was chosen as a test vehicle for several reasons. 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. 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 (i.e., the HLDT emission category).  Different Trailblazer models fall above and
below 6000 Ibs. GVWR, but have no 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 (HEGO) 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, and 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 on cold start temperatures at and below
freezing due to potential ice issues. However, this is 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.5

       A key element of the feasibility test program was  to imitate software initiated emission
control system behaviors which are observed at the currently regulated start temperatures of 75°
F and 50° F (California-only requirement).  These software features do not entail any new
hardware. In the case of the Trailblazer, while not all behaviors could be demonstrated, several
of the most important behaviors were replicated.  First, the secondary air injection  system was
                                          5-19

-------
Final Regulatory Impact Analysis


operated during the cold temperature test. Second, elevated idle speeds, similar to what the
Trailblazer currently uses after the start at the regulated start temperatures, were also used at cold
temperature.

       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.  We tested several delay periods from the start of the
engine until the secondary air system was activated in order to measure benefits of earlier
introduction of the air injection. The secondary air was always run until the 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
acceptable 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). An On-board
Diagnostics 2 (OBD2) dealership diagnostic tool  was capable of electronically elevating the idle
speed to the equivalent 50° F cold start idle  speed but internal tool or ECM software prevented
this idle speed and air flow for the full 30 to 60 second time period.  Hence it was determined
that for the purposes of this test program, a vacuum leak would accurately demonstrate the effect
of an acceptable idle speed and air flow following a start. 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, and also in the defroster tests.  The tests
with defroster operation were included to assess any  emission impacts of defroster-on, which is
required in the fuel economy rule.c

       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
 Fuel Economy Final Rule, 71 FR 77872, December 27, 2006, Defroster Operation Requirement for Cold FTP.


                                           5-20

-------
Final Regulatory Impact Analysis
ECM to control NMHC and NOx emissions simultaneously, as compared to this test program's
limitations.

       CO and direct PM measurements were also significantly reduced when NMHC controls
were activated. CO, the only currently regulated pollutant at 20° F, was consistently reduced
from baseline levels with each of the control combinations. PM was also generally reduced;
however, this 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
Standard < 6000 Ibs GVWR
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 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
lbs°.  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., 0.3 g/mi) with some additional minor calibration changes. We also recognize that this
feasibility study does not constitute a production calibration and that additional development
effort would be needed to achieve manufacturer functional objectives for cold starts.  This test
program simply demonstrates that in the case of this typical  secondary air injection equipped
vehicle, additional emission reduction opportunities exist without the requirement for additional
hardware.
D Tier 2 vehicles are tested (inertia setting) at loaded vehicle weight (LVW). LVW=curb weight + 300 Ibs.
                                           5-21

-------
Final Regulatory Impact Analysis
       Emissions results for the 20° F cold CO test are reported as a weighted three-bag average.
However, bag one (the first 505 seconds of the test) provides a better indication of emission
reductions achieved with controls, because 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.

       As observed below in Table 5.1-9, the results clearly show NMHC, CO and PM
reductions. NMHC and CO reductions occur with all the control attempts, but  control tests #6
and #8,  in which secondary air injection was activated immediately upon engine cranking,
achieve the best results. PM reductions follow similar behavior as NMHC, but they 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
2006 Chrysler 300C Testing
       A second vehicle, a 2006 Chrysler 300C, was chosen because it has specific engine
related challenges and a different method of controlling emissions than the first feasibility
                                           5-22

-------
Final Regulatory Impact Analysis
vehicle. Manufacturers have commented that the higher displacement engines typically used in
heavier vehicles pose additional challenges because they require more fuel to start and maintain
idle stability at cold temperatures than smaller displacement engines.6 The 300C is a light-duty
vehicle equipped with a large displacement V8 engine (5.7 liter). The same large displacement
V8 engine used in the 300C is also found in a broad range of vehicles, including full size trucks
and several sport utility vehicles above 6000 Ibs. GVWR. Unlike the Trailblazer, it does not use
secondary air injection, and instead relies more heavily on controlling fuel in the combustion
chamber. Like the Trailblazer, it is certified as a Tier 2 Bin 5 package. This represents what can
be considered the "typical" or average 75° F emission level once Tier 2 phase-in is complete.
The hardware used on the 300C to comply with Tier 2 Bin 5 standards also represents what we
might expect from many manufacturers and vehicle lines that do not have secondary air
injection.

       The 300C emission 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 (HEGO) before and after
the catalysts, electronic throttle control, and several other necessary supporting sensors. The
aftertreatment hardware consists of a dual catalysts (one  per bank) located in close proximity to
the engine exhaust manifold for optimum catalyst heating.  Additionally, the 300C contains
several new technologies, including cylinder deactivation and dual spark plugs per cylinder.

       The 300C is offered as both a US model (sold in North America)  and a European model
(sold in Europe) without any discernable differences in the emission control hardware (i.e.,
catalysts, oxygen sensors, etc.), with the exception  of the ECM.  Vehicles sold in the European
market are required to comply with the European Union  (EU) type VI test, which is a cold start
test performed at 20° F. In addition to CO emission standards, the EU cold temperature test also
requires stringent HC and NOx emission control. For this reason, any differences observed in
the emissions between a US model and a European model are likely due to software and
calibration differences targeted specifically at HC, CO and NOx emission reductions to meet the
European cold temperature emissions standards.

       These emission controls could be implemented in the US model 300C and other US
packages also sharing the same large displacement  V8 engine. While all the different
applications were not directly tested, the control techniques used in the 300C to reduce cold start
emissions for EU compliance should also be applicable to the other vehicle platforms that share
the same engine. In fact, some of the other vehicle platforms that share the same V8 engine are
also sold in Europe, which indicates that emission control techniques are likely already leveraged
across various European models sharing a common engine.

       In order to determine any emission control opportunities that exist in the European 300C,
an ECM containing the 2006 European software and calibration was purchased and installed in
the 300C. No other changes were made to the emission control system. The US and European
model configurations were each tested over the 20° F FTP. The tests were replicated to ensure
that the emissions levels were consistent. The testing was performed with defroster operation as
specified in the fuel economy ruleE to incorporate any potential emission impacts of defroster or
 '• Fuel Economy Final Rule, 71 FR 77872, December 27, 2006, Defroster Operation Requirement for Cold FTP.


                                          5-23

-------
Final Regulatory Impact Analysis


heater-on operation. The 300C utilizes an automatic interior climate control system which, as
expected, did not ramp up fan speed until the coolant temperature exceeded a threshold
established for driver comfort.  Since the methods used to control cold start NMHC emissions
are primarily used only in the first minute of operation prior to the ramp-up of the automatically
controlled interior fan, the emission results are likely not impacted by defroster operation.
However, as with the first feasibility vehicle test program, it was logical that the testing be
performed with the defroster usage as specified in the fuel economy rule to include any potential
emissions impact.
       Table 5.1-10 below contains the weighted test total (3 bags) emission results of the two
different test configurations attempted on the 300C.  All emissions were significantly reduced
when tested in the European configuration.  CO, the only currently regulated pollutant at 20° F in
the US, was consistently reduced below baseline levels with each test of the European
calibration. In the European model, NMHC, was reduced 32% reduced when compared to the
US model. The European models NMHC levels approached the new fleet average standard of 0.3
g/mi for lighter vehicles under 6000 Ibs. GVWR. The 300C is considered a worst-case model in
the LDV category due to its large displacement V8 engine and high test weight.  With the fleet
averaging provision provided in the final rule, this vehicle would likely be certified to a higher
PEL. However, we fully expect that with the lead time provided, further calibration
improvements to reduce NMHC levels to the 0.3 g/mi level are possible but we did not attempt
this capability in the limited test program.

       Table 5.1-10. 300C Test Configuration and 20° F FTP Weighted Test Total Results
Test Number
Standard < 6000 Ibs GVWR
Standard > 6000 Ibs GVWR

1-US model
2-US model
Average

1-EU model
2-EU model
Average

% Change
NMHC
g/mi
0.3
0.5

0.574
0.540
0.557

0.398
0.384
0.379

-32%
CO
g/mi



2.25
1.99
2.12

1.26
1.36
1.23

-42%
NOx
g/mi



0.06
0.07
0.07

0.04
0.05
0.04

-43%
PM
g/mi



0.011
0.009
0.010

0.005
0.006
0.006

-40%
Fuel
Economy
mi/gallon



14.90
14.89
14.90

14.24
14.93
14.64

-2%
       While this vehicle was tested at 4500 Ibs test weight representing the lower GVWR class,
as noted above, the same engine configuration is also sold in several over 6000 Ibs. GVWR
models. Vehicles over 6000 Ibs. GVWR are required to test at slightly higher weights (i.e., 5000
                                          5-24

-------
Final Regulatory Impact Analysis


Ibs. to 6000 lbs.)F. We expect that if tested at the higher test weight required for vehicles over
6000 Ibs. GVWR, the emission results would likely increase but would maintain an acceptable
margin below the 0.5 g/mi standard for heavier vehicles. As indicated previously, we believe
this vehicle can likely be brought into compliance with the LDV/LLDT fleetwide standard (i.e.,
0.3 g/mi) with some additional calibration changes, given the lead time provided in the program.

       Since almost all of the emissions at 20° F are emitted in the first few minutes of operation
and mostNMHC controls are attempted only during the first few minutes of operation, Table
5.1-11 presents  only the bag  1 emission results.
            Table 5.1-11.  300C Test Configuration and 20° F FTP Bag 1 Only Results
Test Number

1-US model
2-US model
Average

1-EU model
2-EU model
Average

% Change
NMHC
g/mi

2.75
2.49
2.62

1.897
1.830
1.864

-39%
CO
g/mi

10.4
9.1
9.8

5.6
6.0
5.8

-41%
NOx
g/mi

0.14
0.20
0.17

0.10
0.10
0.10

-41%
PM
g/mi

0.037
0.033
0.035

0.016
0.021
0.019

-54%
Fuel Economy
mi/gallon

13.26
13.41
13.34

12.42
12.74
12.58

-6%
       As observed above in Table 5.1-11, all measured emissions with the European
configuration were significantly reduced in bag 1 of the 20° F FTP test.  NMHC, CO, NOx and
PM reductions can be clearly seen from the bag 1 results. As described  earlier, these reductions
are attributed entirely to calibration and software changes in the European model as all other
emission control hardware remained the same with the US model.  These calibration and
software changes could easily be adapted to the US model to achieve significant emissions
reductions. Additional emission reduction opportunities may exist with further software and
calibration refinement.  We believe this test program demonstrates that significant emissions
reductions are available through only calibration on the most challenging Tier 2 vehicles (i.e.,
300C) and that most Tier 2 vehicles can achieve these standards. While the standards will be
challenging for some manufacturers across their product lines, we believe the lead time and other
program flexibilities provided will allow compliance with the new standards.

5.2    Feasibility of Evaporative Emissions Standards for Vehicles
 Tier 2 vehicles are tested (inertia setting) at loaded vehicle weight (LVW). LVW=curb weight + 300 Ibs.
                                          5-25

-------
Final Regulatory Impact Analysis


       The standards for evaporative emissions, which are equivalent to the California LEV II
standards, are technologically feasible now.  As discussed in Section V of the preamble for
today's rulemaking, the California LEV II program contains numerically more stringent
evaporative emissions standards compared to existing EPA Tier 2 standards, but because of
differences in testing requirements, we believe the programs are essentially equivalent.  This
view is supported by manufacturers and current industry practices.  (See Section V.C.5 of the
preamble for further discussion of such test differences (e.g., test temperatures and fuel
volatilities).) A review of recent model year certification results indicates that essentially all
manufacturers certify 50-state evaporative emission systems.7 Therefore, harmonizing with
California's LEV-II evaporative emission  standards will codify the approach manufacturers have
already indicated they are taking for 50-state evaporative systems.
                                           5-26

-------
Final Regulatory Impact Analysis


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 40 CFR 86.1811-04, paragraphs (g) and (k).

5 Memo to docket "Discussions Regarding Secondary Air System Usage at 20° F with European
  Automotive Manufacturers and Suppliers of Secondary Air Systems," December 2005.

6 Comments submitted by Alliance of Automobile Manufacturers (Alliance), OAR-2005-0036-
0881.1

7 Update for FRM: U.S. EPA, Evaporative Emission Certification Results for Model Years 2004
  to 2007, Memorandum to Docket EPA-HQ-OAR-2005-0036 from Bryan Manning, January 4,
  2007.
                                        5-27

-------
Final Regulatory Impact Analysis

                              Chapter 6: Table of Contents

Chapter 6: Feasibility of the Benzene Control Program	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    Pre-Fractionation to Reroute Benzene Precursors	22
        6.3.2.2    Benzene Saturation via Isomerization	24
        6.3.2.3    Reformate Post-Fractionation with Benzene Saturation	25
        6.3.2.4    Benzene Extraction	27
        6.3.2.5    Low-Pressure Reformer Operation	29
        6.3.2.6    Pre-fractionation Combined with Low-Pressure Reformer Operation	30
   6.4   Experience Using Benzene Control Technologies	30
     6.4.1      Benzene Levels Achievable through Reformate Benzene Control	31
     6.4.2      Other Benzene Controls	33
   6.5   Averaging, Banking, and Trading (ABT) Program	40
     6.5.1      Starting Gasoline Benzene Levels	40
     6.5.2      Model-Predicted Refinery Compliance Strategies	41
     6.5.3      Predicted Reductions in Gasoline Benzene	43
        6.5.3.1   Early Operational Changes in Benzene Control Technology	45
        6.5.3.2   Early Small Capital Investments in Benzene Control Technology	45
        6.5.3.3   Compliance with the 1.3 vol% Maximum Average Standard	46
        6.5.3.4   Small Refiner Compliance with the Benzene Standards	47
        6.5.3.5   Full Program Implementation / Ultimate Compliance with the 0.62 vol%
               Standard	47
     6.5.4      Credit Generation/Use Calculations & Considerations	48
        6.5.4.1   What factors  impact refiners' decisions to make early process changes?	48
        6.5.4.2   How are early credits calculated?	49
        6.5.4.3   How many early credits do we predict will be generated?	50
        6.5.4.4   How many early credits will be demanded?	50
        6.5.4.5   How are standard credits calculated?	52
        6.5.4.6   How much additional lead time would be generated by standard credits
               generated during the early credit "lag"?	52
        6.5.4.7   How do we estimate ongoing standard credit generation/demand?	53
        6.5.4.8   What are the  credit use provisions?	54
        6.5.4.9   Are there any geographic restrictions on credit trading?	54
        6.5.4.10     What are the credit life provisions?	55
        6.5.4.11     Consideration of credit availability	55
        6.5.4.12     What is the economic value of the ABT program?	60
   6.6   Feasibility for Recovering Octane	62
   6.7   Will the Benzene Standard Result in Any New Challenges to the Fuel Distribution
         System or End-Users?	65
   6.8   Impacts on the Engineering and Construction Industry	66
     6.8.1      Design and Construction Resources Related to Benzene Reduction Equipment67

                                          6-1

-------
Final Regulatory Impact Analysis

     6.8.2    Number and Timing of Benzene Reduction Units	67
     6.8.3    Timing of Projects Starting Up in the Same Year	68
     6.8.4    Timing of Design and Construction Resources Within a Project	68
     6.8.5    Projected Levels of Design and Construction Resources	70
   6.9   Time Needed to Comply with a Benzene Standard	74
   6.10  Will the Benzene Standards Be More Protective Than Current Programs?	76
     6.10.1     Modeling Approach	77
        6.10.1.1      Choice of Analysis Cases and Data Sources	78
        6.10.1.2     Adjustment of Fuel Parameters for Future Years	80
        6.10.1.3      Conversion of Production Properties to In-Use Properties	82
        6.10.1.4     Running the MOBILE Model	86
     6.10.2     Interpretation of Results	89
     6.10.3     Conclusions	90
   6.11  MS AT Fuel Effects Test Program	90
     6.11.1    Overview of Test Program	90
     6.11.2    Key Findings and Next Steps	94
   6.12  Analysis of Future Need for RFG Surveys of Toxics and NOX Performance under
         MSAT2	96
     6.12.1    Total Toxics Reduction	96
     6.12.2    NOX Reduction	97
Appendix 6A: Additional Background on Refining and Gasoline	99
   6A. 1  Petroleum Refining	99
   6A.2 Crude Oil	100
     6A.2.1   Crude Desalting	101
     6A.2.2   Atmospheric Crude Unit	102
     6A.2.3   Preflash	102
     6A.2.4   Crude Unit	103
     6A.2.5   Atmospheric Tower Gasoil and Residuum;  Vacuum Unit	104
     6A.2.6   Naphtha Splitter	107
     6A.2.7   Hydrotreating	107
     6A.2.8   Fluid Catalytic Cracker	108
     6A.2.9   Alkylation	110
     6A.2.10     Thermal Processing	110
   6A.3  Gasoline	110
     6A.3.1   Gasoline as a Complex Mixture	Ill
     6A.3.2   Octane	114
   6A.4 Kerosene and Diesel	117
                                         6-2

-------
Final Regulatory Impact Analysis

         Chapter 6: Feasibility of the Benzene Control Program

       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 benzene control program. Next the lead time
to apply the various control technologies and to comply with the new standards is evaluated.
Finally, the energy and supply impacts of the program 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.  It's important to note that not all refineries have all of these units,
which is a key factor in both the variation in their baseline benzene levels as well as their cost of
benzene control.
                                          6-3

-------
Final Regulatory Impact Analysis
           Figure 6.1-1. Process Flow Diagram for a Typical Complex Refinery
    Natural
      Gas
     Crude
       Oil
* Fuel Gas
                                                                           LPG
                                                                           Gasoline
                                                                           Aromatics
          Crude Tower
                                                                           Kerosine
                                                                         ^ Jet Fuel
 On-Highway
 Diesel
                                                                         Off-roadDiesel
                                                                        * Fuel Oil
                                                                         »• Resid
                   Vacuum Tower
                                                                           Coke
       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

-------
Final Regulatory Impact Analysis

       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),
typically necessary to produce gasoline with sufficient octane. 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

                                           6-5

-------
Final Regulatory Impact Analysis

produce a high quality, heavy gasoline product. Alkylation uses sulfuric or hydrofluoric acid as
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 olefins 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 0.97 volume percent for gasoline produced in and imported
into the U.S. in 2004, 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 2004 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

-------
Final Regulatory Impact Analysis
   Table 6.2-1.  Summary of U.S. Benzene Levels by Gasoline Type and Season for 2004

CG Summer
CG Winter
Total CG
% total volume
RFG Summer
RFC Winter
Total RFG
% total volume
Summer CG & RFG
Winter CG & RFG
Total CG & RFG
% of total volume
U.S. Production
(excl. California)
1.132
1.076
1.103
64.3
0.587
0.622
0.606
20.3
1.006
0.964
0.984
84.6
Imports
0.949
0.756
0.828
1.9
0.677
0.696
0.688
2.1
0.800
0.725
0.754
4.0
Production +
Imports
1.128
1.065
1.095
66.2
0.594
0.629
0.613
22.4
0.998
0.952
0.973
88.6
California
0
0.620
0.620
0.620
11.4
0.620
0.620
0.620
11.4
All Gasoline
1.128
1.065
1.095
66.2
0.603
0.626
0.616
33.8
0.955
0.914
0.933
100.0
       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

-------
Final Regulatory Impact Analysis
         Figure 6.2-1. Benzene Content of RFG and Conventional Gasoline, 2004.
4.5 -i
4n

o
§on
N o.U
O
OQ
c Zb
0
o
o o n
0. -i.U
3 1 ^
O
i n
1 .U
0.5








y
^^
___x^
.—- --' — "" — "~~"^

pp.

RFG

. u 	 i 	 i 	 i 	 i 	 i 	 i 	 i 	 i 	
0 10 20 30 40 50 60 70 80 90 100
Cumulative Gasoline Supplied (Billion Gallons)
       Figure 6.2-1 shows that the annual average benzene levels of conventional gasoline
produced by individual refineries varies from 0.3 to 4.2 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 0.61 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.

-------
Final Regulatory Impact Analysis
       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
                                                ** x. x
                                                                      CG   X RFG
Apr-03
 Jun-03
Batch Date
Sep-03
Dec-03
                                           6-9

-------
Final Regulatory Impact Analysis
       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

-------
Final Regulatory Impact Analysis
  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)	
  o
    2.5
    2.0
    1.5
    1.0
    0.5
    0.0
     Jan-03
• CG
XRFG
                                                                        •

                                                                        X
Apr-03
 Jun-03
Batch Date
Sep-03
Dec-03
                                          6-11

-------
Final Regulatory Impact Analysis
 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

-------
Final Regulatory Impact Analysis
 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 1
ss
0
o
1


• Regular Grade A Mid-Grade x Premium Grade

A
A A A A
A A A
A*
A A * AA A^ 4«
A ^ A • A A
' V '. . /* .A A^AA A ^ A A .
' ' ' -. ' A-"- -"Jl -* A." «AV * '
• -. - - ' - .- r^. -.;>/••• .- • - ..
x x • ,
X v X * . X x XX
^x^xxxx xxx xxXxx x x ^x^^x^x * 3< ^ x x* x^
x x

Jan-03 Apr-03 Jun-03 Sep-03 Dec-03
Batch Date
                                        6-13

-------
Final Regulatory Impact Analysis
 Figure 6.2-7. Premium and Regular Grade Gasoline Batch-by-Batch Benzene Levels for
 	Refinery "F" (volume percent benzene in 2003 gasoline)	
    1
                                                     x Premium Grade   • Regular Grade
       n-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 and were sold as CG.  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

-------
Final Regulatory Impact Analysis

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 as RFG.
                                           6-15

-------
Final Regulatory Impact Analysis

       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 2004.8  The information is presented for both conventional gasoline and
reformulated gasoline.
 Table 6.2-2. 2004 Benzene Levels by Gasoline Type and by PADD as Supplied in the U.S

Conventional
Gasoline
Reformulated
Gasoline
Gasoline
Average
PADD1
0.84
0.63
0.72
PADD 2
1.33
0.81
1.24
PADD 3
0.94
0.54
0.87
PADD 4
1.55
N/A
1.55
PADD 5
1.75
N/A
1.75
CA
0.62
0.61
0.62
U.S.
1.10
0.63
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
                                          6-16

-------
Final Regulatory Impact Analysis

naphtha is reforming. In the process of increasing the octane of this straight run material, the
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
Alkylate
Isomerate
Hydrocrackate
Butane
Light Straight Run
MTBE/Ethanol
Natural Gasoline
Coker Naphtha
Benzene Level
(volume %)
3-11
0.5-2
0
0
1-5
0
0.3-3
0.05
0.3-3
3
Typical Volume in
Gasoline (volume %)
30
36
12
4
3
4
4
3
3
1
Estimated Contribution to Gasoline
Benzene Content (volume %)
77
15
-
-
4
-
2
-
1
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

-------
Final Regulatory Impact Analysis
       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 c