United States
Environmental Protection
Agency
Office of Air and Radiation
Washington, DC 20460
EPA-452/R-98-003
December 1998
REGULATORY IMPACT ANALYSIS
FOR THE NOx SIP CALL, FIP, AND
SECTION 126 PETITIONS
Volume 2: Health and Welfare Benefits
-------
-------
REGULATORY IMPACT ANALYSIS
FOR THE NOx SIP CALL, FIP, AND
SECTION 126 PETITIONS
Volume 2: Health and Welfare Benefits
Prepared by
Office of Air Quality Planning and Standards
Office of Atmospheric Programs
U.S. Environmental Protection Agency
December 1998
-------
This page intentionally left blank.
-------
ACKNOWLEDGMENTS
This Regulatory Impact Analysis was prepared as a joint effort of the Office of Air Quality
Planning and Standards (OAQPS), and the Office or Atmospheric Programs (OAP). The preparation of
this report was managed by Scott Mathias (EPA/OAQPS), Sam Napolitano (EPA/OAP), and Ravi
Srivastava (EPA/OAP). Both offices thank the many individuals and companies that made significant
contributions to this study.
OAQPS thanks John Bachmann (EPA/OAQPS), Tammy Croote (EPA/OAQPS), Robin Dennis
(EPA/ORD), Ron Evans (EPA/OAQPS), Richard Haynes (USDA), Bryan Hubbell (EPA/OAQPS), Norm
Possiel (EPA/OAQPS), Rosalina Rodriguez (EPA/OAQPS), Allyson Siwik (EPA/Region 6), Eric
Slaughter (Association of National Estuary Programs), Larry Sorrels (EPA/OAQPS), Abt Associates,
Inc., ASI, Dyntel, Mathtech, Inc., Pechan-Avanti Group, and Systems Applications International, Inc.
OAP thanks Kevin Culligan (EPA/OAP), Sarah Dunham (EPA/OAP), Gene Sun (EPA/OAP),
Peter Tsirigotis (EPA/OAP), and ICF Resources, Inc.
-------
This page intentionally left blank.
-------
Table of Contents
TABLE OF CONTENTS vii
LIST OF FIGURES x
LIST OF TABLES xi
LIST OF ACRONYMS AND ABBREVIATIONS xvii
EXECUTIVE SUMMARY ES-1
Chapter 1. INTRODUCTION AND BACKGROUND
1.1 Introduction
1.2 The Clean Air Act
1.2.1 Ozone Requirements
1.2.2 NOx Control and Ozone Reduction
1.2.3 Title IV NOx Requirements
1.2.4 New Source Performance Standards
1.2.5 Reasonably Available Control Technology Requirements
1.2.6 Northeast Ozone Transport Region
1.3 Overview of the NOx SIP Call Rulemaking
-1
-2
-3
-4
-6
-7
-9
1.4 Relationship Between NOx SIP Call, FIP, and Section 126 Petitions 1-10
1.5 Statement of Need fortheNOx SIP Call 1-11
1.5.1 Statutory Authority and Legislative Requirements 1-11
1.5.2 Health and Welfare Effects of NOx Emissions 1-11
1.5.3 Need for Regulatory Action 1-12
1.6 Requirements for this Regulatory Impact Analysis 1-13
1.6.1 Executive Order 12866 1-13
1.6.2 Regulatory Flexibility Act and Small Business Regulatory Enforcement Fairness
Act of 1996 1-13
1.6.3 Unfunded Mandates Reform Act 1-14
1.6.4 Paperwork Reduction Act 1-15
1.6.5 Executive Order 12898 1-15
1.6.6 Health Risks for Children 1-15
1.7 Structure and Organization of the Regulatory Impact Analysis 1-16
1.8 References 1-18
Chapter 2. REGULATORY ALTERNATIVES
2.1 Elements Considered in Developing Regulatory Alternatives 2-1
2.1.1 Type of Control 2-1
2.1.2 Geographic Scope 2-2
2.1.3 Potentially Affected Sources 2-4
2.1.4 Stringency of Control Level 2-4
2.1.5 Effective Dates 2-5
2.1.6 Emissions Budget Trading System Design 2-5
2.2 2007 Emissions Estimates for Air Quality Modeling 2-6
VI
-------
Table of Contents (continued)
2.2.1 Electricity Generating Unit (EGU) Point Source Emissions 2-7
2.2.2 Non-EGU Point Source Emissions 2-7
2.2.3 Area and Mobile Source Emissions 2-7
2.2.4 Natural Emissions 2-7
2.2.5 Summary of 2007 Emissions Projections 2-7
2.3 References 2-12
Chapter3. AIR QUALITY IMPACTS
3.1 Ozone Air Quality Estimates 3-1
3.1.1 Modeling Domain 3-2
3.1.2 Simulation Periods 3-2
3.1.3 UAM-V Model Output 3-2
3.1.4 Converting Episode Estimates to Full-Season Profiles 3-4
3.1.5 Extrapolating from Monitored to Unmonitored Grid Cells 3-5
3.1.6 Ozone Air Quality Results 3-6
3.2 PM Air Quality Estimates Using RPM 3-7
3.2.1 Modeling Domain 3-7
3.2.2 Simulation Periods 3-9
3.2.3 RPM Model Outputs 3-9
3.2.4 Development of Total PM Estimates 3-9
3.2.5 RPM PM Air Quality Results 3-11
3.3 PM Air Quality Estimates Using the S-R Matrix 3-11
3.3.1 Climatological Regional Dispersion Model 3-12
3.3.2 Development of the S-R Matrix 3-13
3.3.3 Fugitive Dust Adjustment Factor 3-14
3.3.4 Normalizing S-R Matrix Results to Measured Data 3-16
3.3.5 Development of Annual Median PM25 Concentrations 3-17
3.3.6 S-R Matrix PM Air Quality Results 3-17
3.4 Nitrogen Deposition Estimates 3-19
3.5 Visibility Degradation Estimates Using RPM and the S-R Matrix 3-20
3.6 References 3-23
Chapter 4. BENEFITS OF REGIONAL NOx REDUCTIONS
4.1 Overview of Benefits Estimation 4-1
4.2 Issues in Estimating Changes in Health Effects 4-10
4.2.1 Baseline Incidences 4-11
4.2.2 Thresholds 4-15
4.3 Ozone and PM Health-Related Benefits 4-16
4.3.1 Premature Mortality 4-17
4.3.2 Hospital Admissions 4-25
4.3.3 Bronchitis 4-27
4.3.4 Acute Respiratory Symptoms 4-29
4.3.5 Worker Productivity 4-32
4.3.6 Work loss days 4-34
4.3.7 Minor restricted activity days 4-35
4.4 Ozone- and PM-related Welfare Effects 4-36
vin
-------
Table of Contents (continued)
4.4.1 Commodity Agricultural Crops 4-37
4.4.2 Commercial Forests 4-39
4.4.3 Nitrogen Deposition 4-40
4.4.4 Household Soiling Damage 4-43
4.4.5 Visibility 4-44
4.5 Total Benefits 4-48
4.6 Limitations of the Analysis 4-56
4.6.1 Projected Income Growth 4-56
4.6.2 Unquantifiable Benefits 4-56
4.6.3 Potential Disbenefits 4-60
4.7 References 4-62
Chapters. BENEFIT-COST COMPARISONS
5.1 Summary of Cost Estimates 5-1
5.2 Summary of Benefits Estimates 5-2
5.3 Summary of Net Benefits 5-2
5.4 Limitations to the Benefit-Cost Comparison 5-3
5.5 References 5-5
Appendix A. QUANTIFIED UNCERTAINTY IN HEALTH AND WELFARE BENEFITS
A.I Overview A-l
A.2 Underlying Sources of Uncertainty A-2
A.3 Quantified Uncertainty for Ozone-related Benefits A-4
A.4 Quantified Uncertainty for PM-related Benefits A-4
A.5 Statistical Uncertainty and Plausible Ranges A-5
A.6 References A-5
Appendix B. SUMMARY OF POPULATION-WEIGHTED AIR QUALITY METRICS B-l
Appendix C. EMISSION SUMMARIES FOR BENEFITS-RELATED AIR QUALITY MODELING E-l
Appendix D. NONLINEAR CHEMISTRY AND FINE PARTICLE PRODUCTION
D.I Background D-l
D.2 What this means for NOx Emissions and Sulfate D-2
D.3 Reference D-2
IX
-------
List of Figures
Figure 1-1
Flowchart of Analytical Steps Page 1-16
Figure 2-1
States Included in EPA's NOx SIP Call Page 2-3
Figure 3-1
UAM-V Modeling Domain Page 3-3
Figure 3-2
RADM-RPM Modeling Domain Page 3-8
Figure 3-3
S-R Matrix Air Quality Modeling Domain Page 3-15
Figure 4-1
Example Methodology of a Benefits Analysis Page 4-3
x
-------
List of Tables
Table ES-1
Regulatory Alternatives by Source Category Groupings for the NOx SIP Call Page ES-2
Table ES-2
Estimate of Emissions Reductions, Total Annual Costs, and Cost-Effectiveness in
2007 of the EPA's Selected Approach to NOx SIP Call Page ES-3
Table ES-3
Ozone, NOx, and PM Benefits from the NOx SIP Call Page ES-5
Table ES-4
Summary of Benefits in 2007 by Major Category for the Selected Regulatory
Alternative (millions of 1990 dollars) Page ES-6
Table ES-5
Comparison of Annual Costs and Monetized Benefits in 2007 Associated with the
NOx SIP Call (millions of 1990 dollars) Page ES-7
Table 2-1
Emissions Inputs for Air Quality Models Page 2-6
Table 2-2
2007 Base Case Projection Control Requirements by Major Sector Page 2-9
Table 2-3
Summary of Regulatory Alternatives for the 2007 Air Quality Modeling:
NOx Control Requirements by Major Sector Page 2-10
Table 2-4
Percent Change from 2007 Base Case in 37-State NOx Emissions Page 2-11
Table 3-1
Summary of UAM-V Derived Hourly Ozone Air Quality for Daylight Hours (7am to 7pm)
During the Ozone Season Page 3-6
Table 3-2
Summary of RPM Derived PM Air Quality Page 3-12
Table 3-3
Summary of S-R Matrix Derived PM Air Quality Page 3-18
Table 3-4
Summary of 2007 Nitrogen Deposition in RADM Domain Page 3-19
XI
-------
List of Tables (continued)
Table 3-5
Summary of 2007 Visibility Degradation Estimates (deciviews) Page 3-22
Table 3-6
2007 Population-Weighted Sum of Annual Mean Extinction: RPM Results Page 3-23
Table 4-1
Examples of Potential Benefits of Air Quality Improvements Page 4-5
Table 4-2
Unqualified Benefit Categories Page 4-7
Table 4-3
Quantified and Monetized Health and Welfare Effects Page 4-10
Table 4-4
PM and Ozone Health and Welfare Concentration-Response Function Summary
Data Page 4-13
Table 4-5
Quantified Ozone-Related Health Effects Included in the Benefits Analysis Page 4-18
Table 4-6
Quantified PM-Related Health Effects Included in the Benefits Analysis Page 4-19
Table 4-7
Summary of Mortality Valuation Estimates Page 4-22
Table 4-8
Range of Avoided Ozone-related Mortality Incidences and Monetary Benefits Associated
with the NOx SIP Call Page 4-23
Table 4-9
Avoided Long Term PM-related Mortality Incidences and Monetary Benefits
Associated with the NOx SIP Call Page 4-25
Table 4-10
Sensitivity Analysis: Premature Mortality Benefits Using Avoided Short Term
PM-related Mortality Incidences Page 4-26
Table 4-11
Sensitivity Analysis: Avoided Post Neo-natal PM-related Mortality Incidences Page 4-27
xn
-------
List of Tables (continued)
Table 4-12
Avoided Ozone-related Hospital Admissions and Monetary Benefits Associated
with the NOx SIP Call Page 4-28
Table 4-13
Avoided PM-related Hospital Admissions and Monetary Benefits Associated
with the NOx SIP Call Page 4-29
Table 4-14
Avoided Incidences of PM-related Chronic and Acute Bronchitis and Monetary
Benefits Associated with the NOx SIP Call Page 4-31
Table 4-15
Avoided Incidences of Ozone-related Respiratory Symptoms and Monetary
Benefits Associated with the NOx SIP Call Page 4-33
Table 4-16
Avoided Incidences of PM-related Upper and Lower Respiratory Symptoms
and Monetary Benefits Associated with the NOx SIP Call Page 4-35
Table 4-17
Monetary Benefits from Ozone-related Avoided Lost Worker Productivity
Associated with the NOx SIP Call Page 4-36
Table 4-18
Avoided PM-related Work Loss Days and Monetary Benefits Associated with
the NOx SIP Call Page 4-37
Table 4-19
Avoided PM-related Minor Restricted Activity Days and Monetary Benefits
Associated with the NOx SIP Call Page 4-38
Table 4-20
Quantified Ozone- and PM- Related Welfare Effects Included in the Benefits
Analysis Page 4-39
Table 4-21
Changes in Production of Commodity Crops and Monetary Benefits Associated
with the NOx SIP Call Page 4-41
Table 4-22
Sensitivity Analysis: AGSIM Generated Monetary Benefits Due to Changes in Production
of Commodity Crops Associated with the NOx SIP Call Page 4-42
Table 4-23
Commercial Forest Monetary Benefits Associated with the NOx SIP Call Page 4-43
xm
-------
List of Tables (continued)
Table 4-24
Monetary Benefits Associated with the NOx SIP Call from Avoided Costs
of Nitrogen Removal in Eastern Estuaries Page 4-46
Table 4-25
Monetary Benefits from Reduced Household Soiling Associated with the NOx SIP
Call Page 4-47
Table 4-26
Monetary Benefits from Improved Residential Visibility Associated with the
NOx SIP Call Page 4-48
Table 4-27
Monetary Benefits from Improved Visibility in National Parks in the Southeast
Associated with the NOx SIP Call Page 4-50
Table 4-28
Monetary Benefits Associated with Visibility Changes in National Parks
Outside the Southeast in the NOx SIP Call Region Page 4-51
Table 4-29
Total Quantified Monetary Benefits Associated with the NOx SIP Call,
Incremental to the 2007 Base Case: 0.12 Trading Alternative Page 4-55
Table 4-30
Total Quantified Monetary Benefits Associated with the NOx SIP Call,
Incremental to the 2007 Base Case: 0.15 Trading Alternative Page 4-56
Table 4-31
Total Quantified Monetary Benefits Associated with the NOx SIP Call,
Incremental to the 2007 Base Case: Regionality 1 Alternative Page 4-57
Table 4-32
Total Quantified Monetary Benefits Associated with the NOx SIP Call,
Incremental to the 2007 Base Case: 0.20 Trading Alternative Page 4-58
Table 4-33
Total Quantified Monetary Benefits Associated with the NOx SIP Call,
Incremental to the 2007 Base Case: 0.25 Trading Alternative Page 4-59
Table 4-34
Sources of Uncertainty in the Benefit Analysis Page 4-61
Table 5-1
Estimated Total Annual Cost of NOx SIP Call Alternatives in 2007 Page 5-1
xiv
-------
List of Tables (continued)
Table 5-2
"Plausible Range" of Annual Quantified Benefits Estimates for NOx SIP
Call Alternatives in 2007 (million 1990$) Page 5-2
Table 5-3
Estimated Annual Quantified Net Benefits for NOx SIP Call Aternatives
in 2007 (million 1990$) Page 5-3
Table A-l
Point Estimates and Assumed Distributions of MWTP for Health and Welfare
Endpoints in the NOx SIP Call Analysis Page A-3
Table B-l
2007 Population-Weighted Sum of 1-Hour Ozone Predictions Above 124 ppb:
Adjusted and Extrapolated UAM-V Results Page B-l
Table B-2
2007 Population-Weighted Sum of 8-Hour Average Ozone Predictions
Above 84 ppb: Adjusted and Extrapolated UAM-V Results Page B-2
Table B-3
2007 Population-Weighted Sum of Annual Mean PM25 Predictions
Above 15.04 (ig/m3: RPM Results Page B-3
Table B-4
2007 Population-Weighted Sum of Annual Mean PM10 Predictions
Above 50.4 (ig/m3: RPM Results Page B-4
Table C-l
Ozone Season Daily NOx Emissions by Major Sector and Regulatory Alternative:
UAM-V Inputs Page C-2
Table C-2
Warm Season Annualized NOx Emissions by Major Sector and Regulatory Alternative:
RADM-RPM Inputs Page C-2
Table C-3
Warm Season Annualized SO2 Emissions by Major Sector and Regulatory Alternative:
RADM-RPM Inputs Page C-2
Table C-4
Cold Season Annualized NOx Emissions by Major Sector and Regulatory Alternative:
RADM-RPM Inputs Page C-3
xv
-------
List of Tables (continued)
Table C-5
Cold Season Annualized SO2 Emissions by Major Sector and Regulatory Alternative:
RADM-RPM Inputs Page C-3
Table C-6
Annual NOx Emissions by Major Sector and Regulatory Alternative:
S-R Matrix Inputs Page C-4
Table C-7
Annual SO2 Emissions by Major Sector and Regulatory Alternative:
S-R Matrix Inputs Page C-4
xvi
-------
Select List of Acronyms and Abbreviations
AFS - AIRS Facility System
AIRS - Aerometric Information Retrieval System
ANPR - Advanced Notice of Proposed Rulemaking
bext - Total Atmospheric Light Extinction Coefficient
BACT - Best Available Control Technology
BEIS - Biogenic Emissions Inventory System
CAA - Clean Air Act
CAAA - Clean Air Act Amendments of 1990
CAPMS - Criteria Air Pollutant Modeling System
CASAC - Clean Air Scientific Advisory Committee
CASTNet - Clean Air Status and Trends Network
CB - Chronic Bronchitis
CO - Carbon Dioxide
COI-Cost of Illness
COPD - Chronic Obstructive Pulmonary Disease
C-R - Concentration-Response
CRDM - Climatalogical Regional Dispersion Model
CV - Contingent Valuation
dv - Deciview
ECOS - Environmental Council of States
EGUs - Electricity Generating Units
EO - Executive Order
EPA - Environmental Protection Agency
FIP - Federal Implementation Plan
H+ - Hydrogen Ion
H2O2- Hydrogen Peroxide
HNO3 - Nitric Acid
hr - Hour
IMPROVE - Interagency Monitoring for Protection of Visual Environments
IPM - Integrated Planning Model
Kg/ha - Kilograms Per Hectare
km2 - Square Kilometer
kWh - Kilowatt Hour
LAER - Lowest Achievable Emissions Rates
Ib - Pound
LDs - Loss Days
LRS - Lower Respiratory Symptoms
MCL - Maximum Contaminant Level
mills/kWh - Mills Per Kilowatt Hour
MM4 - Mesoscale Model, version 4
mmBtu - Millions of British Thermal Units
Mm- Megameter
MOU - Memorandum of Understanding
MRAD - Minor Restricted Activity Days
MRRAD - Minor Respiratory Restricted Activity Days
MW - Megawatts
xvii
-------
Acronyms and Abbreviations (continued)
MWh - Megawatt Hours
MWTP - Marginal Willingness to Pay
NAA - Nonattainment Area
NAAQS - National Ambient Air Quality Standards
NAPAP - National Acid Precipitation Assessment Program
NCLAN - National Crop Loss Assessment Network
NET - National Emission Trends
NH3 - Ammonia
NLEV - National Low Emission Vehicle
NMSCs - Nonmelanoma Skin Cancers
NOAA - National Oceanic and Atmospheric Administration
NOx - Oxides of Nitrogen
NO3 - Nitrate
NPR - Notice of Proposed Rulemaking
NPS - Non-Point Source
NSA - Nitrate, Sulfate, and Ammonium Components
NSPS - New Source Performance Standards
NSR - New Source Review
O3 - Ozone
OMB - Office of Management and Budget
QMS - Office of Mobile Sources
O&M - Operation and Maintenance
OTAG - Ozone Transport Assessment Group
OTC - Ozone Transport Commission
OTR - Ozone Transport Region
PM - Particulate Matter
ppm - Parts Per Million
PRA - Paperwork Reduction Act of 1995
PSD - Prevention of Significant Deterioration
RIA - Regulatory Impact Analysis
RACT- Reasonably Available Control Technology
RADM - Regional Acid Deposition Model
RFA - Regulatory Flexibility Act
RMF - Regional Model Farm
RPM - Regional Particulate Model
SAB - Science Advisory Board
SB A - Small Business Administration
SBREFA - Small Business Regulatory Enforcement Fairness Act of 1996
SNPR - Supplemental Notice of Proposed Rulemaking
SO2 - Sulfur Dioxide
SO42' - Sulfate Ion
SOA - Secondary Organic Aerosols
S-R - Source-Receptor
TAMM - Timber Assessment Market Model
tpd - Tons Per Day
tpy - Tons Per Year
TSP - Total Suspended Particulate s
xviii
-------
Acronyms and Abbreviations (continued)
(ig/m3 - Micrograms Per Meter Cubed
UAM-V - Urban Airshed Model - Variable Scale
UMRA - Unfunded Mandates Reform Act
URS - Upper Respiratory Symptoms
USDA - United States Department of Agriculture
UV-B - Ultraviolet-B Radiation
VOCs - Volatile Organic Compounds
WLDs - Work Loss Days
WTP - Willingness to Pay
xix
-------
-------
EXECUTIVE SUMMARY
EPA has finalized the nitrogen oxides (NOx) State -notation P^"«? "£ '""
EPA has modeled an illustrative State implementaton s«" '°. 'se a benefits
benefits of the TNOx SIP call.
At the same time that EPA promulgates the NOx SIP call, EPA is proposing
implementation p.ans (FIPs) that may be needed ***^^^
EPA is also proposing a response to Sect.on 126 ^Uons which we, e 1 lied Dy fc
askine FPA to address air pollution transported from upwind States, Pursuant to ^uuve
?ht RIA pVelts the potential costs, economic impacts and benefits of these nilemaktngs.
The existing 1 -hour and new 8-hour national ambient air quality
promote attainment with the current and new ozone standards.
in th, NOx SIP call EPA is setting ozone season NOx budgets for States that are in the SIP call
controls on the units covered,
!„ tlii. rule EPA has offered to administer an emissions trading program for the States. _
-
approach on which it based the NOx budgets.
Page ES-l
-------
Table ES-1
Regulatory Alternatives by Source Category Groupings
for the NOx SIP Call
Electricity Generating Units
(EGUs)--Emissions Budgets Based
on a NOx Limit of:
0.25 Ib/mmBtu"
0.20 Ib/mmBtu
0.15 Ib/mmBtu in Northeast;
0.20 Ib/mmBtu in Midwest &
Southeast
0.12 ib/mmBtu in Northeast;
0.15 Ib/mmBtu in Midwest;
0,20 Ib/mmBtu in Southeast
O.J 5 Ib/mmBtu
0.1 2 Ib/mmBtu
Non-Electricity Generating Units (non-EGUs)
Industrial Boilers and Combustion
Turbines— Emissions Budgets
Based on a Reduction from
Uncontrolled Levels of:
40%
50%
' :' • • 60%. . v . • y • '.'
70%
All Other Stationary Sources--
Emissions Budgets Based on
Highest Ozone Season NOx
Reduction Achievable without a
Source Paying More Than:
$l,500/ton
$2,000/ton
$3,000/ton
$4,000/ton
. ' , , . $5»000/to» . •
; Sec Cliapler 1 for a breakdown of the States covered in (he NOx SIP call
Emission Reductions, Costs, and Cost-effectiveness
Table ES-2 summarizes EPA estimates of the emission reductions, costs, and cost-effectiveness
for the regulatory approach that EPA selected as the basis for the NOx SIP Call's NOx budgets. (Please
note: Since these estimates were calculated EPA has fine-tuned its estimates of the NOx budgets and an
addendum to this executive summary provides revised emission reduction, cost and cost-effectiveness
information). Overall, 82% of the emission reductions expected under this regulatory alternative are
expected to come from the electric power industry, at an average ozone season cost-effectiveness of
$1,468 per ton. The table indicates the estimates of direct control costs for sources including costs
associated with emissions monitoring and reporting. The table also indicates the total administrative costs
to State governments and EPA. In EPA's analysis to support this rule, the Agency has shown that for
the electric power industry, the largest source of emissions for which it considered controls, a single
trading program across the SIP call region can provide a similar reduction to what direct command-and-
control requirements would accomplish, but do the job at lower cost. For this reason, the Agency is
encouraging States to participate in the trading program that it plans to administer.
Page ES-2
-------
Table ES-2
Estimate of Emission Reductions, Total Annual Costs,
and Cost-Effectiveness in 2007 of the EPA's Selected Approach to NOx SIP Call
Sector
Electricity Generating Units a
Industrial Boilers and Turbines b
Internal Combustion Engines c
Cement Manufacturing c
Administrative Costs for EGUs
Administrative Costs to States and
EPA
Total
Ozone Season
NOx Emission
Reductions
(1,000 tons)
938
104
83
16
if " , - " *
1,141
Total Annual Cost
(millions 1990$)
$1,378
$153
$100
$24
$6
$2
$1,660"
Average Ozone
Season Cost-
Effectiveness
($ per ozone
season ton)
$1,468
$1,467
$1,215
$1,458
.'. , " -•.,.•."
* Does not include additional monitoring costs (see later row),
b Includes additional monitoring and other administrative costs associated with participating in the NOx emissions trading program.
* Includes additional monitoring and other administrative costs associated with the SIP call rule
t Numbers do not add due to rounding
Economic Impacts
EPA considered what the economic impacts could be, if States implemented the regulatory
approach that EPA used to calculate the NOx SIP call budgets for electricity generating units.
Electricity prices could potentially rise in the NOx SIP call region by as much as 1.6 percent in 2007, if
the power industry is pricing its power on the basis of marginal costs in a fully competitive environment.
The price increase will be less, if these assumptions regarding the nature of the competitive environment
do not hold. There will be more new electric generation capacity built in response to the rule than will
retire early (there will be little generation capacity that closes). On net, EPA expects this NOx SIP call to
create more new jobs (from pollution control operations and increased natural gas use) than it reduces
(due to a small decline in forecasted coal demand).
The analysis of non-EGU sources indicates that fewer than 5% of potentially affected firms
experience costs in excess of 1% of revenues, and just over 2% of potentially affected firms experience
costs in excess of 3% of revenues. EPA also examined the potential affect of the NOx SIP call on small
entities that meet the Small Business Administration's definition of "small." The Agency adopted several
ways of minimizing potential impacts on small entities for the final NOx SIP call rulemaking. Of the
nearly 1,200 small entities (both ECU and non-EGU) in the NOx SIP call region that have large NOx
emissions sources, only 150 are potentially affected by the SIP call rule, and only 41 have potential
Page ES-3
-------
compliance costs in excess of 1% of total revenues, EPA expects States to use these results to help
them design control strategies that will reduce or eliminate adverse impacts on small entities.
Benefits
The estimated change in the incidence of health and welfare effects is assessed for several
modeled air quality scenarios. The estimated changes in incidence are then monetized by multiplying the
estimated change in incidence of each endpoint by its associated dollar value of avoiding an occurrence of
an adverse effect. These endpoint-specific benefits are then summed across all affected areas to derive
an estimate of total benefits. Because there are potentially significant categories for which health and
welfare benefits are not quantified or monetized due to a lack of modeling and economic data, the benefit
estimates presented in this analysis should be considered incomplete.
There are several possible sets of assumptions underlying the benefits estimates in this analysis.
The set of assumptions underlying the low end estimate represent a conservative approach which
assumes that both human health and the environment are less responsive to changes in air pollutants. For
example, the low end assumes that no health effects are associated with changes in PM2 s concentrations
below a threshold of 15 jig/m3. In addition, the low end assumes that reductions in NOx emissions have
less positive impacts on PM air quality relative to the high end case. The high end estimate is based on a
set of assumptions that represent potential benefits which would result if human health and the
environment are more responsive to reduced pollution levels than estimated in the low end. In addition,
the high end assumes that reductions in NOx emissions have greater positive impacts on PM air quality
relative to the low end. It should be emphasized that the high and low ends of the plausible range are not
the same as upper and lower bounds. For many of the quantitative assumptions involved in the analysis,
credible, though less plausible, arguments could be made for an even higher or lower choice, which could
lead to an even greater spread between the high end and low end estimates.
Table ES-3 lists the anticipated health and welfare benefits categories that are reasonably
associated with reducing the regional transport of NOx, specifying those for which sufficient quantitative
information exists to permit benefit calculations. Because of the inability to monetize some existing
benefit categories, such as changes in pulmonary function and altered host defense mechanisms, some
categories are not included in the calculation of the monetized benefits.
Page ES-4
-------
Table ES-3
Ozone, NOx, and PM Benefits from the NOx SIP Call
Benefits of Ozone and NOx Reductions-
Reductions in;
Benefits of PM Reciuctions-
Rcductions in:
Quantified
Health
Mortality (short-term exposures)
Hospital admissions for all respiratory illnesses
Acute respiratory symptoms
Mortality (long- and short-term exposures)
Hospital admissions for:
all respiratory illnesses
congestive heart failure
ischemic heart disease
Acute and chronic bronchitis
Lower and upper respiratory symptoms
Minor restricted activity days
Work loss clays
Welfare
Commodity crop yield losses
Commercial forest yield losses
Worker productivity losses
Household soiling
Impaired visibility
Nitrogen deposition to esluarinc and coastal waters
Unqualified
Health
Airway responsiveness
Pulmonary inflammation
Increased susceptibility lo respiratory
infection
Acule inflammation and respiratory cell damage
Chronic respiratory damage/Premature aging of lungs
UV-B (disbenefit)
Changes in pulmonary function
Morphological changes
Altered host defense mechanisms
Other chronic respiratory disease
Cancer
Welfare
Ecosystem and vegetation effects in Class I
areas (e.g., national parks)
Damages to urban ornamentals (e.g.,grass,
flowers, shrubs, and Irees in urban areas)
Fruit and vegetable crop losses
Reduced yields of tree seedlings and non-
commercial forests
Damage to ecosystems
Materials damage (other than consumer
cleaning cost savings)
Nitrates in drinking water
Brown clouds
Passive fertilization (disbenefit)
Materials damage (other than consumer cleaning
cost savings)
Damage to ecosystems (e.g., acid sulfate
deposition)
Nitrates in drinking water
Brown clouds
Table ES-4 shows the range of total monetized benefits for the main health and welfare endpoint
groups associated with the emissions changes that most closely approximate the NOx SIP call budget.
The range of values in this table is intended to represent the plausible range of the benefits that may result
from this rule after accounting for important uncertainties.
Page ES-5
-------
Table ES-4
Summary of Benefits in 2007 by Major Category for the
Selected Regulatory Alternative"
(millions of 1990 dollars)
Category
Ozone Health and Welfare
Agriculture & Forestry
Nitrogen Deposition
PM Health and Welfare
Total
"Low" Assumption Set
$27
$260
$238
$575
$1,100
"High" Assumption Set
$1,353
$574
$238
$2,005
$4,170
1 Potential benefit categories dial have no! been quantified and monetized are listed in Table ES-3.
Comparison of Costs and Monetized Benefits
Cost-benefit analysis provides a valuable framework for organizing and evaluating information on
the effects of environmental programs. This benefit-cost comparison is intended to generally inform the
public about the potential costs and benefits that may result when control strategies to limit NOx emissions
for mitigating regional ozone transport are implemented by the States, Table ES-5 presents a comparison
of monetized benefits and total annual costs for the selected alternative for setting State budgets in the
NOx SIP call. From EPA's examination of five major alternatives for setting the NOx budget, there are
several major conclusions that can be drawn from this RIA:
» For the "High" assumption set, monetized net benefits are positive and substantial for al!
regulatory alternatives,
» As modeled, Regionally 1 is an inferior alternative, i.e., even though 0.20 Trading is less stringent
it achieves greater benefits at lower total costs.
• Net benefits are greatest at the most stringent regulatory alternative evaluated, i.e., 0.12 Trading.
For the "High" assumption set, net benefits are approximately 33 percent higher for the 0.12
Trading relative to the 0.15 Trading alternative. For the "Low" assumption set, net benefits are
positive only for the 0.12 Trading alternative.
« While net benefits are negative for the "Low" assumption set for all but the 0.12 Trading
alternative, it is important to remember that while all of the costs are included, many benefit
categories could not be quantified. In addition, the "Low" assumption set estimate assumes that
there are no reductions in premature mortality associated with ozone reductions. Relaxing this
one assumption would result in positive net benefits for all alternatives for the "Low" assumption
set.
Page ES-6
-------
Table ES-S
Comparison of Annual Costs and Monetized Benefits in 2007 Associated with the NOx SIP Call
(millions of 1990 dollars)
Benefits
Case
"Low" Assumption Set
"High" Assumption Set
Total Annual
Costs
$1,660
$1,660
Annual Monetized
Benefits*
$1,100
$4,170
Annual Net
Benefits
($560)
$2,510
""There are many benefits of the NOx SIP call that EPA was not able to quantify or monetise.
Limitations
Comparing the benefits and the costs provides one framework for policy makers and the public to
assess policy alternatives. Not all the potential costs and benefits can be captured in any analysis.
However, EPA is generally able to estimate reasonably well the costs of pollution controls based on
today's control technology and assess the important impacts when it has sufficient information for its
analysis, EPA compiled through the OTAG process and from many other sources sufficient information
for this ruiemaking. There are, however, important limitations in the RIA analysis:
« EPA is increasingly able to estimate benefits from pollution controls, but EPA believes
that there are many important benefits that it can not quantify or monetize that are
associated with the NOx SIP call, including many health and welfare effects. There are
also potential disbenefits that are not quantified, including passive nitrogen fertilization and
UV-B screening,
• EPA must employ different pollutant models to characterize the effects of
alternative policies on relevant pollutants. Not all atmospheric models have been
widely validated against actual ambient data. The Agency has chosen the best
available models for its application needs in this RIA and tried to make the most
reasonable assumptions possible in using them for predicting air quality changes.
• There are some data limitations in some aspects of the RIA, despite the Agency's
extensive efforts to compile information for this ruiemaking. While they exist, EPA
believes that it has used the models and assumptions that are made to conduct its analysis
in a reasonable way based on the available evidence, but this should be kept in mind when
reviewing various aspects of the RIA's results.
» Another factor that adds to the uncertainty of the results is the potential for pollution
control innovations that can occur over time It is impossible to estimate how much of an
impact, if any, new technologies that are just now emerging may have in lowering the
compliance costs for the NOx SIP call, which goes into effect in 2003. We can only
recognize their possible influence.
Page ES-7
-------
There is the uncertainty regarding future costs that exists due to the flexibility that occurs
under the emissions cap-and-trade program that EPA is encouraging the States to set up.
The analysis that EPA has done to date has been fairly conservative in considering the
electric power industry and large industrial boilers and combustion turbines operating
separately under their own trading programs. In reality, they should enter the same
trading pool and there should be greater efficiency and lower costs that result.
Qualitative and more detailed discussions of the above and other uncertainties and
limitations are included in the analysis. Where information and data exists, quantitative
characterizations of these uncertainties are included. However, data limitations prevent
an overall quantitative estimate of the uncertainty associated with final estimates.
Nevertheless, the reader should keep all of these uncertainties and limitations in mind
when reviewing and interpreting the results.
Despite the- above limitations, EPA believes that the RIA provides evidence that the
benefits resulting from the NOx SIP call will be up to two and one-half times the costs.
Addendum to Executive Summary
In response to comments, EPA has revised the State NOx budgets that it set for the electric
power industry on the basis of .15 Ibs/mmBtus in the final days of the rulemaking process. The SIP call
region budget was lowered from 564 thousand tons of NOx during the ozone season to 544 thousand tons
of NOx, The Agency also decided to create a "compliance supplement pool" for use in 2003 and allow
banking with flow controls in the trading program that EPA is encouraging States to undertake.
For the- adjustment of the NOx budget to 544 thousand tons for the electricity generating units, the
Agency estimates that there will be a reduction of ozone season NOx emissions by 958 thousand tons in
2007 at an annual cost of $1,440 million. This is an average cost-effectiveness of $1,503 per ton of NOx
reductions during the ozone season. The total ozone season NOx emission reductions from the NOx SIP
call if the States implement the program the way EPA used to set the budget is 1,161 thousand tons.
An adjustment to the emissions inventory for the non-EGU sources was also made as a result of
public comments. The reanalysis following these emission inventory adjustments indicated only minor
changes in the costs. Benefit estimates are not recalculated for either adjustments to the NOx budget or
emission inventory, but the conclusions drawn from the RIA are not expected to differ significantly,
Page ES-8
-------
Chapter 1. INTRODUCTION AND BACKGROUND
1.1 Introduction
This document presents a Regulatory Impact Analysis for ttie final NOx SIP call rule, which
addresses regional transport issues related to ozone attainment1. This rule requires certain States to take
action to reduce emissions of nitrogen oxides (NOx) that contribute to nonattainment of ozone standards
in downwind States2. This R1A also satisfies the analytical requirements for the proposed NOx Federal
Implementation Plan (FIP) and Clean Air Act (CAA) section 126 petition actions. The proposed FIP
may be needed if any State fails to revise its SIP to comply with the final NOx SIP call. The proposed
action under CAA section 126 responds to petitions filed with EPA by eight Northeastern States
requesting that EPA provide relief from emissions sources in several upwind States that may be
contributing to ozone nonattainment in the petitioning States3.
The Clean Air Act (CAA) requires States to demonstrate attainment of the National Ambient Air
Quality Standards (NAAQS) for ozone. Many States have found it difficult to demonstrate attainment of
the ozone NAAQS due to the widespread regional transport of ozone and its precursors, NOx and
volatile organic compounds (VOCs). The Ozone Transport Assessment Group (OTAG) was established
in 1995 to undertake an assessment of the regional transport problem in the Eastern half of the United
States. OTAG was a collaborative process among 37 affected States, the District of Columbia, the U.S.
Environmental Protection Agency (EPA), and interested members of the public, including environmental
groups and industry representatives.
OTAG concluded that regional reductions in NOx emissions are needed to reduce the transport
of ozone and its precursors. OTAG recommended that major sources of NOx emissions (utility and other
stationary sources) be controlled under State NOx budgets, and also recommended development of an
emissions trading program.
After a review of OTAG's analysis, findings, and recommendations, EPA proposed a rule to limit
summer season NOx emissions in a group of States that the Agency believes are significant contributors
to ozone in downwind areas." In a November 7, 1997 Notice of Proposed Rulemaking (NPR), EPA made
1 This document is the second volume a two volume set. Volume 1 covers the estimated costs and
economic impacts of the final NOx SIP call rule,
2 Ground level (or tropospheric) ozone is an air pollutant that forms when its two primary components,
oxides of nitrogen and volatile organic compounds, combine in the presence of certain meteorological conditions.
Ozone is associated with a variety of adverse effects both to human health and to the environment. For more
information on these adverse effects refer to Chapter 4 of this RIA.
3 Unless necessary to provide specific emphasis, the term "NOx SIP call" will be used (rather than "FIP" or
"section 126 petitions") throughout this report when referring to the regulatory framework that is analyzed and
reported in this RIA, See section 1.4 for additional detail on the analytical relationship between these three
regulatory actions.
4 NOx emissions reductions were proposed for 22 Stales and the District of Columbia.
Page 1-i
-------
a determination that transport of ozone from certain States in the OTAG region5 makes a significant
contribution to nonattainment, or interferes with the maintenance of attainment, with the ozone NAAQS
in downwind States (FR 1997a). EPA proposed a summer season NOx budget (in tons of NOx) for each
of these States. These States will be required to amend their State Implementation Plans (SIPs) through a
call-in procedure established in Section 110 of the Clean Air Act Amendments of 1990 (CAAA), In a
May 1998 Supplemental Notice of Proposed Rulemaking (SNPR), EPA made technical corrections to
the State NOx budgets, and developed a proposed trading rule to provide for emissions trading (FR
1998a), The SNPR also included an analysis of the air quality impacts of the proposed rule. The State
NOx emissions budgets, trading rule, and related provisions are now being promulgated as a final rule,
A technical background support document prepared for the November 7, 1997 NPR estimated
costs and emissions reductions associated with an assumed strategy that States might take to achieving
the proposed budgets (EPA, 1997a). These analyses were updated to reflect technical corrections to the
population of sources and growth estimates on which the State-specific budgets were based and assess
the effects of the proposed trading system, in an analysis supporting the April 1998 SNPR (EPA 1998a).
This document provides the supporting Regulatory Impact Analysis (RIA) for the final rule. This
analysis expands and updates the previous analyses, to reflect the provisions of the final rule and to
provide analysis of the potential benefits and economic impacts as well as the costs, emissions reductions
and air quality impacts associated with the rule.
The remaining sections of this chapter address the following topics:
1.2 Relevant requirements of the Clean Air Act;
1.3 Overview of the NOx SIP call rulemaking;
1.4 Relationship between the NOx SIP call, FIP, and section 126 actions;
1.5 Statement of need for the NOx SIP call;
1.6 Administrative requirements addressed by this RIA;
1.7 Structure of the RIA and organization of this document; and
1.8 References for Chapter 1.
1.2 The Clean Air Act
The 1970 Clean Air Act Amendments required EPA to issue, periodically review, and, if
necessary, revise, NAAQS for ubiquitous air pollutants (Sections 108 and 109). States are required to
submit SIPs to attain those NAAQS, and Section 110 of the CAA lists minimum requirements that SIPs
must meet. Congress anticipated that all areas would attain the NAAQS by 1975. In 1977, the CAA was
amended to provide additional time for areas to reach the NAAQS and included the requirement that
5 The OTAG region consists of 37 States east of 104° W longitude.
Page 1-2
-------
States reach the NAAQS for ozone by 1982 or 1987. In addition, the 1977 amendments included
provisions that required SIPs to consider adverse downwind effects and allowed downwind States to
petition for tighter controls on upwind States that contribute to their NAAQS nonattainment status,
In 1990, the Clean Air Act was again amended . This section outlines requirements of the 1990
Clean Air Act Amendments (CAAA) related to NOx reductions and the NOx SIP call. The discussion
includes the ozone and NOx requirements and a review of the guidelines for new or advanced air
emissions
control technologies.
1,2,1 Ozone Requirements
The CAAA included provisions designed to address the continued nonattainment of the existing
ozone NAAQS, specified requirements that would apply if EPA revised the existing standard, and
addressed transport of air pollutants across State boundaries.
In 1991 and 1992, areas not in attainment with the 1-hour ozone NAAQS were placed in one of
five classifications, based on the degree of nonattainment. Requirements for moving toward attainment,
including definitions of "major source" for VOCs and NOx, attainment dates and new source offset
ratios, were established for each of the five classifications. Within an area known as the Northeast
Ozone Transport Region (OTR), all sources emitting 50 tons or more of ozone forming pollutants a year
are defined as "major sources," regardless of their current attainment classification. Certain emissions
limits apply to major sources, and even more stringent requirements apply for new major sources in
nonattainment areas.
Since passage of the 1990 CAAA, EPA has revised the NAAQS for ozone, EPA is required to
review the NAAQS at least eveiy five years to determine whether, based on new research, revisions to
the standards are necessary to continue to protect human health and the environment. As a result of the
most recent review, EPA revised the NAAQS for both particulate matter and ozone. The previous
ambient air quality standard for ozone was 0.12 ppm based on 1-hour averaging of monitoring results.
The revised standard was set at 0.08 ppm based on an 8- hour averaging period. The 1-hr standard
remains in effect until EPA determines that a given area has air quality meeting its 1-hour standard. This
is necessary to ensure continued progress in those areas and a smooth transition between the two
standards.
On July 16, 1997, President Clinton issued a directive to EPA on the implementation strategy for
the new ozone and particulate NAAQS. The goal of the implementation strategy is to provide flexible,
common-sense, and cost-effective means for communities and businesses to comply with the new
standard. The EPA has issued proposed guidance for public comment on implementation of the revised
standards (August 24, 1998, 63 PR 45060). Additional guidance will be proposed in October 1998. The
August and October guidance will be combined and issued as one document in December 1998, The
implementation strategy includes:
Endorsement of a Regional Approach; Citing EPA's work with the OTAG, the implementation
strategy notes that ozone needs to be addressed as a regional problem. The Directive indicates
that, based on OTAG recommendations, EPA will propose a rule to provide a flexible, common-
sense, and cost effective means for communities and businesses to comply with the new
Page 1-3
-------
standards. The strategy states that EPA will encourage and assist the States to develop a regional
emissions cap-and-trade system, modeled on the current acid rain program, as a way to achieve
reduction in NOx emissions at lower cost.
Transitional Classifications: Areas that attain the 1-hour standard but that do not attain the new
8-hour standard will be eligible for a specific "transitional" classification, if they participate in a
regional strategy and/or submit early plans addressing the new standard, EPA will revise its
rules for new source review (NSR) and conformity so that States will be able to comply with the
new standards with only minor revisions to the existing programs in such transitional areas.
Areas which will achieve attainment as a result of the regional strategy need not implement any
additional local controls. Areas that will not achieve the 8-hour standard even with the regional
strategy are eligible for transitional status if they submit revised SIPs in the year 2000
demonstrating attainment of the 8-hour standard on the same schedule as the regional transport
requirements,
Cost-Effective Implementation Strategies: EPA will encourage States to design strategies for
both the PM and ozone standards that focus on getting low cost reductions and that limit the cost
of control to under $10,000 per ton for all sources. EPA will encourage market-based strategies
to lower the cost of attainment and stimulate technology innovation.
The NOx SIP call, therefore, plays an important role in the implementation strategy for the new ozone
NAAQS, by instituting a regional strategy that will encourage cost-effective attainment of the new
standard,
1,2.2 NOx Control and Ozone Reduction
To address the CAAA provisions regarding continued nonattainment of the existing ozone
NAAQS, EPA's post-1994 attainment strategy guidance for the 1-hour ozone standard called for
continued emissions reductions within ozone nonattainment areas together with a national assessment of
the ozone transport phenomenon. Recognizing that no individual state or jurisdiction can effectively
assess or resolve all of the issues relevant to ozone transport, the Environmental Council of States
(ECOS) formed a national work group to address ozone pollution,6 OTAG was established to assist
states east of the Mississippi River to attain federal ozone standards and to develop regional strategies to
address regional transport problems,7 The multi-state, multi-stakeholder OTAG process included input
from State and local governments, industry, environmental groups, and the Federal government. The
stated goal of OTAG was to:
Identify and recommend a strategy to reduce transported ozone and its precursors which, in
combination with other measures, will enable attainment and maintenance of the national
ambient ozone standard in the OTAG region, A number of criteria will be used to select the
6 ECOS is a national organization of environmental commissioners with members from the 50 States and
territories,
7 Information on OTAG and copies of documents produces by the group can be accessed on-line at
http://www.epa.gov/ttn/otag.
Page 1-4
-------
strategy including, but not limited to, cost effectiveness, feasibility, and impacts on ozone levels
(OTAG, 1995).
OTAQ's work included development of a comprehensive base-year (1990) emissions inventory
for use in all OTAG analyses. The inventory contained information provided by the States and reviewed
by OTAG for point, area, and mobile sources. State-specific growth factors were used to project
emissions for the years 1999 and 2007, which represent the CAAA attainment dates for certain
nonattainment areas, Baseline 2007 emissions were also adjusted to reflect the effect of various controls
required under existing regulatory programs or expected from future programs.
OTAG then conducted modeling of NOx and ozone across the OTAG region for several
scenarios using geographic and atmospheric models:
Strategy Modeling. OTAG Strategy Modeling was done in several phases, and included analysis
of more than 25 emission control strategies. OTAG found that domain wide emissions of NOx in the
2007 baseline are approximately 12 percent lower than 1990 and emissions of VOC are approximately 20
percent lower. Thus, existing CAA programs are expected to produce a reduction in ozone
concentrations in many nonattainment areas. However, the analysis showed that some areas currently in
nonattainment will likely remain so in the future and that new 8-hour nonattainment and/or maintenance
problem areas may develop as a result of economic growth in some areas.
Geographic Modeling. OTAG conducted geographic modeling to isolate the effects of NOx
reductions on specific subregions. Among other results, OTAG found that a regional strategy focusing
on NOx reductions across a broad portion of the region will help mitigate the ozone problem in many
areas of the East. Further, a regional NOx emissions reduction strategy coupled with local NOx and/or
VOC reductions may be needed to achieve attainment and maintenance of theNAAQS in the region.
This analyses conducted by OTAG (OTAG 1997), as well as EPA's analyses in support of the
new ozone NAAQS (EPA 1997b)5 showed the important role that reducing NOx emissions plays in the
reduction of ozone levels. The extensive air quality modeling performed by OTAG indicated that both
ozone and NOx can be transported long distances, up to 500 miles. While reductions in either NOx and
VOCs may reduce ozone in localized urban areas, only NOx reductions would result in lower ozone
levels across the region. The OTAG analyses showed a correlation between the magnitude and location
of NOx reductions and the magnitude of reductions in ozone levels in downwind areas. OTAG,
therefore, reached the following conclusion:
Regional NOx reductions are effective in producing ozone benefits; the more NOx
reduced, the greater the benefit. Ozone benefits are greatest where emission reductions
are made and diminish with distance. Elevated and low level NOx reductions are both
effective (OTAG 1997, pp. 51-52).
Based on the evidence of the relationship between NOx emissions and regional ozone levels,
OTAG recommended that a range of NOx controls be applied in certain areas of the OTAG region. A
wide variety of sources are responsible for NOx emissions, including electricity generating units, other
(non-utility) stationary sources, area sources, non-road mobile sources, and highway vehicle sources.
OTAG did not suggest any one "right" approach to reducing major source NOx emissions. However,
Page 1-5
-------
OTAG developed a number of specific recommendations for EPA pertinent to the NOx SIP call,
including the following;8
• OTAG-related controls should be implemented in the "fine grid" states,9
* The range of utility NOx controls should fall between Clean Air Act controls and the less
stringent of 85% reduction from the 1990 rate (Ib/mmBtu) or 0,15 Ibs, of NOx /mmBtu summer
heat input,
* The stringency of controls for individual large non-utility point sources should be established in
a manner equitably with utility controls, and RACT should be considered for individual medium
non-utility point sources where appropriate,10 OTAG recommended that EPA calculate statewide
NOx tonnage budgets based on a specified relationship between control levels for coal-fired
power plants and control targets (emission reduction percentages) for large and medium non-
utility point sources.
• OTAG stated that market-based approaches are recognized as having a number of benefits in
relation to traditional command and control regulations, and that States have the option to select
market systems that best suit their needs. They described two basic approaches that States might
use to implement NOx emissions market systems, and recommended that a joint State/EPA
Workgroup be formed to develop design features and implementation provisions for market
systems that could be selected by the States.
OTAG also made recommendations that EPA develop and adopt a variety of specific national regulations
that were assumed for the modeling to result in reduced emissions of VOCs and/or NOx, and to reach
closure on the Tier 2 Motor Vehicle Study.
The recommendations resulting from the extensive analysis and air quality modeling conducted
by OTAG have played a major role in the design of the NOx SIP call,
8 Summaries of the OTAG findings and recommendations are provided in OTAG 1997.
9 The fine grid states include those modeled using UAM-V at a grid resolution of 12 km2. All other areas
constitute the coarse grid which is modeled at a grid resolution of 36 km2, Coarse grid states are Florida, Louisiana,
Texas, Arkansas, Oklahoma, Kansas, Nebraska, North Dakota, South Dakota, and Minnesota.
10 OTAG provided specific definitions of large and medium point sources, for purposes of their
recommendations.
Page 1-6
-------
1.2.3 Title IV NOx Requirements
Title IV of the CAAA requires annual reductions in NOx emissions. The Acid Rain NOx
Program under Title IV incorporates a two-phased strategy to reduce NOx emissions. In the first phase,
starting January 1, 1996, some Group 1 boilers (i.e., dry bottom wall-fired boilers and tangentially fired
Page 1-7
-------
boilers) are required to comply with specific NOx emission limitations." In the second phase, starting
January 1, 2000, the remaining Group 1 boilers must comply with more stringent NOx emission limits.12
Further, Group 2 boilers (i.e., wet bottom wall-fired boilers, cyclones, boilers using cell-burner
technology, and vertically fired boilers) must comply with recently established emission limits.13
Compliance results for 1996 show that, from 1990 to 1996, the Phase I affected population's
average NOx emission rate declined by 40 percent. Overall NOx emission reductions between 1990 and
1996 for the affected boilers totaled about 340,000 tons, i.e. a reduction of 33 percent (EPA, 1997c). In
Phase II, about 1.17 million tons per year of NOx reductions are projected to result from the Acid Rain
NOx Program requirements (EPA, 1996).
In developing State budgets for the NOx SIP call, EPA considered the NOx reductions
committed to by Title IV NOx Program requirements.
1.2.4 New Source Performance Standards
The EPA is under court order to promulgate a new source performance standard (NSPS) on
fossil-fuel-fired utility and industrial boilers in September 1998, and subpart GG of Part 60 regulates
NOx emissions from combustion turbines. The final standards revise the NOx emission limits for steam
generating units in subpart Da (Electric Utility Steam Generating Units) and subpart Db
(Industrial-Commereial-Institutional Steam Generating Units). Only those electricity generating units
and industrial steam generating units for which construction, modification, or reconstruction is
commenced after July 9, 1997 would be affected by these revisions.
The NOx emission limit in the final rule for new subpart Da units is 201 nanograms per joule
(ng/J) [1.6 Ib/megawatt-hour (MWh)] gross energy output regardless of fuel type. For existing sources
that become subject to subpart Da through modification or reconstruction, the NOx emission limit is 0.15
Ib/million Btu heat input. For subpart Db units, the NOx emission limit being proposed is 87 ng/J (0.20
Ib/million Btu) heat input from the combustion of any gaseous fuel, liquid fuel, or solid fuel; however,
for low heat release rate units firing natural gas or distillate oil, the current NOx emission limit of 43 ng/J
(0.10 Ib/million Btu) heat input is unchanged.
In developing the State budgets for the NOx SIP call, EPA considered the potential NOx
reductions attributable to this NSPS.
11 The affected dry-bottom wall-fired boilers must meet a limitation of 0,50 Ibs of NOx per mmBtu
averaged over the year, and tangentially fired boilers must achieve a limitation of 0,45 Ibs of NOx per mmBtu, again
averaged over the year (FR 1995).
12 Annual averages of 0.46 Ib/mmBtu for dry-bottom wall-fired boilers and 0.40 ib/mmBtu for tangentially
fired boilers.
" The limits are 0.68 Ib/mmBtu for cell burners, 0.86 Ib/mmBtu for cyclones greater than 155 MWe, 0.84
Ib/mmBtu for wet bottom boilers greater than 65 MWe, and 0.80 Ib/mmBtu for vertically fired boilers (FR 1996a).
Page 1-8
-------
1.2.5 Reasonably Available Control Technology Requirements
In the 1977 amendments to the CAA Congress required that all SIPs for nonattainment areas
contain reasonably available control measures (RACM) or reasonably available control technology
(RACT). In the 1990 Amendments to the Act, Congress created RACT requirements specifically for
ozone nonattainment areas under the 1-hour standard (see subpart 2 of part D of title I). Since 1977, EPA
has defined RACT for ozone as the lowest emission limitation that a particular source is capable of
meeting by the application of control technology that is reasonably available considering technological
and economic feasibility. The EPA historically has interpreted the RACT requirement in ozone
nonattainment areas to apply independent of a State's ability to demonstrate that an area will attain the
ozone standard, with certain exceptions.
In the ozone-specific RACT requirement enacted in 1990, States were required to correct all
existing deficiencies in RACT rules in marginal nonattainment areas to ensure the rules were adopted
consistently on a national basis. In addition, all nonattainment areas classified moderate and above were
required to adopt RACT for each source category for which EPA issued a Control Techniques Guideline
(CTG). Over the years, EPA has issued CTG documents to assist the States in determining RACT for
VOCs, Each CTG contains information on available air pollution control techniques and provides a
"presumptive norm" for RACT for a specific source category. Finally, RACT for controlling NOx was
also required in certain nonattainment areas classified moderate and above.
In developing implementation guidance for the revised 8-hour NAAQS, EPA is addressing the
RACM/RACT requirement under subpart 1 of part D of title I, rather than subpart 2, The EPA has
proposed implementation guidance for the revised ozone NAAQS which addresses several issues,
including RACM/RACT. The proposed policy states that "For the 8-hour ozone NAAQS, if the
[nonattainment] area is able to demonstrate attainment of the standard as expeditious!}' as practicable
with emission control measures in the SIP, then RACM/RACT will be met and additional measures
would not be required as being reasonably available," (August 24, 1998, 63 FR 45060) The policy will
be finalized by December 31, 1998.
1.2.6 Northeast Ozone Transport Region
Section 184 of the CAAA delineated a multi state ozone transport region (OTR) in the Northeast
and required specific additional NOx and VOC controls for all areas in this region (not only
nonattainment areas), Section 184 also established the Ozone Transport Commission (OTC) for the
purpose of assessing the degree of ozone transport in the OTR and recommending strategies to mitigate
the interstate transport of pollution. The OTR consists of the States of Connecticut, Delaware, Maine,
Maryland, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island,
Vermont, parts of northern Virginia, and the District of Columbia. The OTC was first convened in 1991,
and began analysis and evaluation of ozone reduction strategies for the region. They concluded that
regional reductions of NOx emissions are particularly important in reducing ozone. The OTR States
confirmed that they would implement RACT on major stationary sources of NOx, and agreed to a phased
approach for additional controls, beyond RACT, for power plants and other large fuel combustion
sources.
Page 1-9
-------
This agreement, known as the OTC Memorandum of Understanding (MOU) for stationary source
NGx controls was approved on September 27, 1994, All OTC States, except Virginia, are signatories to
the OTC NOx MOU, The OTC NOx MOU establishes an emissions trading system to reduce the costs
of compliance with the control requirements.
In developing State budgets for the NOx SIP call, EPA considered the NOx reductions
committed to by the OTR states in the OTC NOx MOU, along with the OTAG recommendations
discussed above.
1.3 Overview of the NOx SIP Call Rulemaking
EPA relied extensively on the OTAG analyses and recommendations in developing the NOx SIP
call. As recommended by OTAG, the rule establishes ozone season14 NOx emission budgets for 22
States and the District of Columbia,15 The 23 jurisdictions will be required to amend their SIPs by the
year 2000, to allocate emissions control requirements among sources and to develop compliance
programs for each affected source category to ensure that the NOx budget is met. These compliance
programs should include: necessary pollution control measures; monitoring, reporting, and accounting
procedures to ensure source emissions are not exceeding the State's NOx budget; and enforcement
requirements,
Consistent with OTAG's recommendation that NOx emissions reductions be achieved primarily
from large stationary sources in a trading program, EPA is encouraging States to consider additional
controls on electricity generating units and other large stationary sources as a strategy for meeting
statewide budgets. State budgets were developed using assumptions consistent with such a strategy. The
budget for each State was developed for components of major source categories. For non-road and
highway vehicle sources, budgets are based on estimates of the effectiveness in each State of national
measures that EPA is taking to control emissions from mobile sources. For electricity generating units
and other stationary sources, the budgets are based on applying further reasonable controls. A major
factor in determining controls is the cost-effectiveness of control measures,
EPA also followed OTAG's recommendation in urging States to consider implementing market
based systems to reduce the costs of complying with the new limits on NOx emissions. EPA is
encouraging the States and the District of Columbia to join a trading program administrated by EPA,
which is reflected in a model NOx Budget Trading Rule. This trading system would place a collective
cap on NOx emissions from electricity generating units and other large boilers and combustion turbines,
and provide for trading of allowances similar to the CAAA Title IV SO2 Allowance Trading Program
already in place.
Chapter 2 of Volume 1 of this report describes a number of regulatory alternatives that EPA
considered in the development of this final rule, Chapter 2 of Volume 2 describes the regulatory
H The ozone season for this rule is the period May 1 - September 30.
15 The States covered by the rule include: Alabama, Connecticut, Delaware, Georgia, Illinois, Indiana,
Kentucky, Maryland, Massachusetts, Michigan, Missouri, New Jersey, New York, North Carolina, Ohio,
Pennsylvania, Rhode Island, South Carolina, Tennessee, Virginia, West Virginia, and Wisconsin.
PageI-10
-------
alternatives that EPA used to model air quality effects for the purpose of estimating benefits of the final
rule.
1.4 Relationship Between NOx SIP Call, FIP, and Section 126 Petitions
In conjunction with promulgating the NOx SIP call, EPA has begun efforts to respond to
petitions filed by eight northeastern States (FR 1998b). These petitions were filed under section 126 of
the CAA, which authorizes States to petition EPA to address air pollution transported from upwind
States. The petitions request that EPA make a finding that NOx emissions from certain major stationary
sources significantly contribute to ozone nonattainment problems in the petitioning States. If EPA makes
such a finding, the Agency would be authorized to establish Federal emissions limits for these sources.
The petitions recommend control levels for EPA to consider. In an April 30, 1998 Advanced Notice of
Proposed Rulemaking (ANPR) (63 FR 24058), EPA presented a schedule for taking actions on the
petitions, made a preliminary identification of upwind sources that may significantly contribute to 1-hour
and 8-hour ozone nonattainment problems in the petitioning States (using information developed for the
NOx SIP call NPR), and requested comment on legal and policy issues raised by section 126 of the CAA.
In responding to the section 126 petitions, EPA intends to be consistent with the approaches taken in the
NOx SIP call.
At the same time that EPA promulgates the NOx SIP call rule, EPA is proposing NOx Federal
Implementation Plans (FIPs) that may be needed if any State fails to comply with the final NOx SIP call
rule. The FIP requirements are intended to be consistent with the approaches taken in the final NOx SIP
call, including a proposed federal NOx Budget Trading Program for electric utility sources and other
large industrial boilers and combustion turbines.
Since the final NOx SIP call and the proposed FIP and section 126 petition actions are generally
consistent in the manner in which they assess affected emissions sources, EPA is preparing only a single
RIA for all three actions. Even though the facts of the analysis contained in this report do not differ
significantly for any of the three actions, the results have slightly different interpretations. In the case of
the final NOx SIP call, the results in this report are illustrative of potential benefits that may result from
the SIP call. The NOx SIP call itself does not directly impose regulatory requirements on emissions
sources. Instead, the SIP call requires States to develop strategies to meet the State NOx budgets
contained in the final NOx SIP call rule. States have discretion on which emissions sources to control to
realize the required reductions.
However, the FIPs, if needed, and the section 126 petition responses will directly impose
regulatory requirements on emissions sources. EPA is proposing to regulate sources under the FIP and
section 126 petition actions with strategies that are modeled in this RIA. In these cases the results
presented in the RIA reflect potential outcomes from direct federal regulation, and, depending on the
outcome of the final actions, have a higher probability of reflecting the actual outcome of the rules.
The proposed section 126 actions will potentially affect only a subset of the sources potentially
affected by the broader NOx SIP call. Sources in Georgia, South Carolina, and Wisconsin are not
affected by the proposed section 126 rule. Therefore, the benefits associated with the proposed section
126 rule are likely to be smaller than for the final NOx SIP call. Since Georgia, South Carolina, and
Wisconsin are affected under the final NOx SIP call, and would be subject to a final FIP if they fail to
Page 1-11
-------
comply with the provisions of the final NOx SIP call, EPA did not see the need to separately address the
potentially smaller benefits for the proposed section 126 rule.
1.5 Statement of Need for the NOx SIP Call
The following sections discuss the statutory authority and legislative requirements of the NOx
SIP call, health and welfare effects of NOx emissions, and the basis for the regulatory actions of the NOx
SIP call.
1,5A Statutory Authority and Legislative Requirements
Section 110(a)(2)(D) provides that a SIP must contain provisions preventing its sources from
contributing significantly to nonattainment or interfering with maintenance of theNAAQS in a
downwind State. This section applies to all pollutants covered by NAAQS and all areas regardless of
their attainment designation. Section 110(k)(5) authorizes EPA to find that a SIP is substantially
inadequate to meet any CAA requirement, as well as being inadequate to mitigate interstate transport as
described in Sections 184 and 176A, Such a finding would require States to submit a SIP revision to
correct the inadequacy within a specified period of time.
1.5,2 Health and Welfare Effects of NOx Emissions1*
NOx emissions contribute to the formation of ozone during the summer season. Ozone is a
major component of smog and is harmful to both human health and the environment. Research has
shown the following health effects of ozone:
* Exposure to ambient ozone concentrations has been linked to increased hospital admissions for
respiratory ailments, such as asthma. Repeated exposure to ozone can make people more
susceptible to respiratory infection and lung inflammation, and can aggravate preexisting
respiratory diseases,
* Children are at risk for the effects of ozone because they are active outside during the summer
months when ozone levels are at their highest. Adults who are outdoors and moderately active
during the summer months are also at risk. These individuals can experience a reduction in lung
function and increased respiratory symptoms, such as chest pain and cough, when exposed to
relatively low ozone levels during periods of moderate exertion.
Long-term exposures to ozone can cause repeated inflammation of the lung, impairment of lung
defense mechanisms, and irreversible changes in lung structure, which could lead to premature
aging of the lungs and/or chronic respiratory illnesses such as emphysema and chronic
bronchitis.
'* A comprehensive discussion of health and environmental issues related to NOx appears in EPA, 1997d.
Page 1-12
-------
• Several peer reviewed epidemiology studies recently published suggest a possible association
between ozone exposure and mortality, though several other studies find no significant
association.
Ozone has also been shown to adversely affect vegetation, including reductions in agricultural and
commercial forest yields, reduced growth and decreased survivabUity of tree seedlings, and increased
tree and plant susceptibility to disease, pests and other environmental stresses.
NOx emissions also contribute to fine particle matter formation (PM). Exposure to airborne PM
has a wide range of adverse health effects. The key health effects associated with PM include: 1)
premature mortality; 2) aggravation of respiratory and cardiovascular disease (as indicated by increased
hospital admissions and emergency room visits, school absences, work loss days, and restricted activity
days); 3) changes in lung function and increased respiratory symptoms; 4) changes to lung tissues and
structure; 5) altered respiratory defense mechanisms; and 6) chronic bronchitis. Most of these effects
have been consistently associated with ambient PM concentrations, which have been used as a measure
of population exposure, in a number of community epidemiological studies. Although mechanisms by
which particles cause effects have not been elucidated, there is general agreement that the cardio-
respiratory system is the major target of PM effects. Participate matter also is associated with welfare
effects, which include visibility impairment, soiling, and materials damage.
Based on its review of the scientific evidence, EPA established standards for PM25 and retained
the standards for PM10. The EPA revised the secondary (welfare-based) PM NAAQS by making them
identical to the primary standards.
Finally, NOx emissions contribute to a wide range of health and environmental problems
independent of their contribution to ozone or PM formation. Among these problems are acid deposition,
nitrates in the drinking water, and nutrient loading in waterways, particularly in sensitive coastal
estuaries where air deposition is a major portion of nitrogen loadings.
1,5.3 Need for Regulatory Action
The existing and revised ambient air quality standards for ozone set levels necessary for the
protection of human health and the environment. Under the CAA, attainment of these standards depends
on the implementation of State-specific pollution control strategies contained in SIPs to reduce NOx and
volatile organic compound emissions, in conjunction with EPA promulgation of national controls for
some sources of pollution.
It is clear that, even with planned national measures in place, several States cannot bring existing
nonattainment areas into compliance with the current ozone standard, or avoid the application of very
costly local control measures, unless the transport of ozone from other upwind areas is reduced.
Furthermore, many States will find it hard, if not impossible, to avoid nonattainment with the revised
ozone NAAQS, or come into attainment with it in the future, unless mitigation of the ozone transport
problem occurs. This dilemma has raised concerns over the fairness of downwind areas having to cope
with the pollution coming from areas upwind. The current regulatory framework requires States to
develop SIPs that demonstrate air quality improvements sufficient to reach specific attainment levels.
States have no control over neighboring States' actions, and may be unable to meet their air quality goals
Page 1-13
-------
due to pollutants transported across State lines. The contribution of upwind sources outside of
nonattainment areas creates a dilemma for States seeking to reach air quality goals.
States could develop local ozone mitigation strategies to address the impact of transported ozone.
However, local efforts could lead to undesirable outcomes. Some States might develop SIPs that do not
achieve compliance in some serious and severe ozone nonattainment areas, because the States would
deem local measures needed to achieve attainment as too draconian.
The NOx SIP call is designed to mitigate these problems through a coordinated Federal and State
effort to address regional ozone transport. This rule will create a more effective, efficient and equitable
approach for EPA and the States to promote attainment with the current and new ozone NAAQS.
1.6 Requirements for this Regulatory Impact Analysis
This section describes various legislative and executive requirements that govern the analytical
requirements for Federal rulemakings, and describes how each analytical requirement is addressed in this
RIA.
1,6.1 Executive Order 12866
Executive Order 12866, "Regulatory Planning and Review" (FR, 1993), requires EPA to provide
the Office of Information and Regulatory Affairs of the Office of Management and Budget with an
assessment of the costs and benefits of significant regulatory actions. A "significant regulatory action" is
defined as "any regulatory action that is likely to result in a rule that may:
• Have an annual effect on the economy of $100 million or more or adversely affect in a material
way the economy, a sector of the economy, productivity, competition, jobs, the environment,
public health or safety, or State, local, or tribal governments or communities;
* Create a serious inconsistency or otherwise interfere with an action taken or planned by another
agency;
» Materially alter the budgetary impact of entitlements, grants, user fees, or loan programs or the
rights and obligations of recipients thereof; or
• Raise novel legal or policy issues arising out of legal mandates, the President's priorities, or the
principles set forth in the Executive Order" (FR, 1993).
For any such regulatory action, the Agency must provide a statement of the need for the proposed action,
must examine alternative approaches, and must estimate social benefits and costs.
EPA has determined that the NOx SIP call is a significant regulatory action because its effect on
the economy is expected to exceed $100 million per year. This volume of the RIA provides the benefits
information required by E.O. 12866 for a significant regulatory action; Volume 1 fulfills the associated
cost and economic impact requirements.
Page 1-14
-------
1.6.2 Regulatory Flexibility Act and Small Business Regulatory Enforcement Fairness Act of
1996
The Regulatory Flexibility Act (RFA) of 1980 (PL 96-354) requires that agencies conduct a
screening analysis to determine whether a regulation will have a significant impact on a substantial
number of small entities, including small businesses, governments and organizations. If a regulation will
have such an impact, agencies must prepare a Regulatory Flexibility Analysis, and comply with a number
of procedural requirements to solicit and consider flexible regulatory options that minimize adverse
economic impacts on small entities. The RFA's analytical and procedural requirements were
strengthened by the Small Business Regulatory Enforcement Fairness Act (SBREFA) of 1996.
For reasons explained more fully in the Federal Register notice for the final NOx SIP call, it is
EPA's position that the RFA as amended by SBREFA does not apply to the final NOx SIP call, because
the rule does not impose direct requirements on emissions sources. States will ultimately decide what
emissions limits are imposed for specific sources. However, the EPA has determined that the RFA as
amended by SBREFA does apply to both the proposed FIP and section 126 actions. Therefore, EPA has
examined the potential for small entity impacts to provide policy makers and States with additional
decision information.
The RFA and SBREFA require use of definitions of "small entities", including small businesses,
governments and non-profits, published by the Small Business Administration (SBA).17 Screening
analyses of economic impacts presented in Volume 1 of the RIA examine potential impacts on small
entities.
1.6.3 Unfunded Mandates Reform Act
The Unfunded Mandates Reform Act (UMRA) of 1995 (PL 104-4) was enacted to focus
attention on federal mandates that require other governments and private parties to expend resources
without federal funding, to ensure that Congress considers those costs before imposing mandates, and to
encourage federal financial assistance for intergovernmental mandates. The Act establishes a number of
procedural requirements. The Congressional Budget Office is required to inform Congressional
committees about the presence of federal mandates in legislation, and must estimate the total direct costs
of mandates in a bill in any of the first five years of a mandate, if the total exceeds $50 million for
intergovernmental mandates and $100 million for private-sector mandates.
Section 202 of UMRA directs agencies to provide a qualitative and quantitative assessment of
the anticipated costs and benefits of a Federal mandate that results in annual expenditures of $100
million or more. The assessment should include costs and benefits to State, local, and tribal governments
and the private sector, and identify any disproportionate budgetary impacts. Section 205 of the Act
requires agencies to identify and consider alternatives, including the least costly, most cost-effective, or
least burdensome alternative that achieves the objectives of the rule.
17 Where appropriate, agencies can propose and justify alternative definitions of "small entity." This RIA
relies on the SBA definitions.
Pace 1-15
-------
EPA has not reached a final conclusion as to the applicability of the requirements of UMRA to
the NOx SIP call rule, EPA has determined that UMRA does affirmatively apply to both the proposed
FIP and proposed section 126 rules. Volume 1 of this RIA presents a summary of analyses of the
potential impacts of the NOx SIP call on State and local governments, to support compliance with section
202 of UMRA, This analysis includes administrative requirements of State and local governments
associated with revising SIPs and collecting and reporting data to EPA. It also includes the compliance
and administrative costs to emissions sources owned by government entities. In addition, EPA has
prepared a more detailed written statement consistent with the requirements of section 202 and section
205of the UMRA and placed that statement in the docket for this ruiemaking.
1.6.4 Paperwork Reduction Act
The Paperwork Reduction Act of 1995 (PRA) requires Federal agencies to be responsible and
publicly accountable for reducing the burden of Federal paperwork on the public, EPA has submitted an
Information Collection Request (ICR) to the Office of Management and Budget (OMB) in compliance
with the PRA. The ICR explains the need for additional information collection requirements and provides
respondent burden estimates for additional paperwork requirements to State and local governments
associated with the NOx SIP call.
1.6,5 Executive Order 12898
Executive Order 12898, "Federal Actions to Address Environmental Justice in Minority
Populations and Low-Income Populations," requires federal agencies to consider the impact of
programs, policies, and activities on minority populations and low-income populations. Disproportionate
adverse impacts on these populations should be avoided. According to EPA guidance, agencies are to
assess whether minority or low-income populations face risk or a rate of exposure to hazards that is
significant (as defined by the National Environmental Policy Act) and that "appreciably exceeds or is
likely to appreciably exceed the risk or rate to the general population or other appropriate comparison
group." (EPA, 1996b) This guidance outlines EPA's Environmental Justice Strategy and discusses
environmental justice issues, concerns, and goals identified by EPA and environmental justice advocates
in relation to regulatory actions.
The NOx SIP call is expected to provide health and welfare benefits to eastern U.S. populations,
regardless of race or income. Chapter 3 of this RIA presents information on the changes in potential
ozone and PM exposure for white and non-white populations and low income populations, and compares
these relative changes to the general populations.
1.6.6 Health Risks for Children
Executive Order 13045, "Protection of Children from Environmental Health Risks and Safety
Risks," directs Federal agencies developing health and safety standards to include an evaluation of the
health and safety effects of the regulations on children. Regulatory actions covered under the Executive
Order include rulemakings that are economically significant under Executive Order 12866, and that
concern an environmental health risk or safety risk that the agency has reason to believe may
Page 1-16
-------
disproportionately affect children. EPA has developed internal guidelines for implementing the E.O.
13045. (EPA, 1998b)
The NOx SIP call is a "significant economic action," because the annual costs are expected to
exceed $100 million. Both NOx and ozone formed by NOx are known to affect the health of children
and other sensitive populations, which were addressed in the development of the new ozone NAAQS.
However, the NOx SIP call is not expected to have a disproportionate impact on children. Chapter 3 of
this RIA presents information on the changes in potential ozone and PM exposure for persons under the
age of 18.
1.7 Structure and Organization of the Regulatory Impact Analysis
The potential costs, economic impacts and benefits have been estimated for this ruiemaking. The
flow chart in Figure 1-1 summarizes the analytical steps taken in developing the results presented in this
RIA.
Figure 1-1
Flowchart of Analytical Steps
Mode! Air Quality
i
Select Control Strategies
COSTS
BENEFITS
Estimate Control Cost
Estimate Post-Control Air Quality
Estimate Small Business and Other
Economic Impacts
Estimate Human Health
and Welfare Effects
Estimate Monetized Value of
Health and Welfare Effects
The assessment of costs, economic impacts, and benefits consists of multiple analytical
components, dependent upon emissions and air quality modeling. In order to estimate baseline air quality
in the year 2007, emission inventories are developed for 1995 and then projected to 2007, based upon
estimated national growth in industry earnings and other factors. Current CAAA-mandated controls
Page 1-17
-------
(e.g., Title I reasonably available control measures, Title II mobile source controls, Title III air toxics
controls, Title IV acid rain sulfur dioxide (SO2) controls) are applied to these emissions to take account
of emission reductions that should be achieved in 2007 as a result of implementation of the current PM
and ozone requirements. These 2007 CAA emissions in turn are input to several air quality models that
relate emission sources to area-specific pollutant concentrations. This modeled air quality is used as the
base against which several alternative control options are measured and cost estimates developed. Given
the estimated costs of the alternative regulatory control options, the potential economic impacts of these
estimated costs on potentially affected industry sectors is subsequently analyzed. Potential health and
welfare benefits are also estimated from modeled changes in air quality as a result of control strategies
applied in the cost analysis. Finally, benefits and costs are compared.
The RIA analyses have been constructed such that benefits and costs are estimated incremental to
those derived from the effects of implementing the CAAA in the year 2007. These analyses provide a
"snapshot" of potential benefits and costs of this rulemaking in the context of implementation of CAA
requirements between now and 2007 and the air quality effects that derive from economic and population
growth.
States have discretion in how they achieve their NOx budgets, and different States may choose
different strategies. The RIA must, therefore, be based on assumptions about how the States will choose
to implement the NOx SIP call requirements. Consistent with EPA's recommendation that States focus
on major stationary sources, this RIA assumes that States impose additional controls — incremental to
those already required by other national programs that address NOx emissions — only for major
stationary sources, and that States implement the cap-and-trade system for electricity generating units and
industrial boiler and turbine sources. This assumption is illustrative of one cost-effective approach States
could take to meeting the NOx SIP call budgets. States may choose different allocations of controls
across major stationary sources than assumed here, or may choose to impose additional controls on area
or mobile sources as well. Costs and economic impacts would differ from those estimated in this RIA to
the extent that States' compliance strategies differ from the RIA assumptions.
Analysis of costs, changes in emissions, and economic impacts is conducted separately for two
groups of sources: electricity generating units and other stationary sources. The Integrated Planning
Model (IPM) allows analysis of trading and industry-level adjustments for electricity generating unit
sources. Other stationary sources are analyzed separately, using assumptions about baseline conditions
and control costs that are generally consistent with the IPM modeling assumptions used for electricity
generating units.
Predicted changes in emissions due to the additional controls for electricity generating units and
other stationary sources are then combined to estimate changes in air quality and to calculate the benefits
of the NOx SIP call. The estimation of benefits from environmental regulations poses special challenges.
These include the difficulty of quantify ing the incidence of health, welfare, and environmental endpoints
of concern, and the difficulty of assigning monetized values to these endpoints. As a result, many
categories of potential benefits have not been monetized at all, and those that have been are given in
ranges. Specifically, this RIA has adopted the approach of presenting a "plausible range" of monetized
benefits to reflect these uncertainties by selecting alternative values for each of several key assumptions.
Taken together, these alternative sets of assumptions define the range for the monetized benefits
categories.
Page 1-18
-------
The remainder of the RIA is organized in the following chapters and appendices:
* Chapter 2 presents a discussion of the regulatory alternatives modeled for the benefits analysis of
this rulemaking;
* Chapter 3 describes the methodology used to estimate baseline and post-control air quality
impacts;
» Chapter 4 presents the analysis of the benefits of the rule; and
Chapter 5 provides an integrated summary of costs and benefits, and compares costs and benefits
for the NOx SIP call as a whole.
Where appropriate, each chapter includes a discussion of limitations of the analysis, A series of
appendices follow Chapter 5 and provide more detailed descriptions of specific methodologies and
results.
1,8 References
Federal Register, 1993. Executive Order 12866, Regulatory Planning and Review, Vol. 58, October 4,
1993, pg 51735,
Federal Register, 1995. Acid Ram Program; Nitrogen Oxides Emission Reduction Program, Vol. 60,
No, 71, April 13, 1995,
Federal Register, 1996. Acid Rain Program; Nitrogen Oxides Emission Reduction Program: Final Rule.
Vol. 61, No. 245, December 19, 1996.
Federal Register, 1997a. Finding of Significant Contribution and Rulemaking for Certain Stales in the
Ozone Transport Assessment Group Region for Purposes of Reducing Regional Transport of Ozone;
Proposed Rule. Vol. 62, No. 216, November 1, 1997, pg 60318.
Federal Register, 1997b. Memorandum of July 16, 1997 from William J. Clinton, President of the
United States, to the Administrator of the Environmental Protection Agency. Subject: "Implementation
of the Revised Air Quality Standards for Ozone and Particulate Matter," Vol. 62, July 18, 1997, pg
38421.
Federal Register, 1997c. National Ambient Air Quality Standards for Ozone and Particulate Matter;
Notice of Proposed Rulemaking. Vol. 62, No. 241, December 13, 1997.
Federal Register, 1998a. Supplemental Notice for the Finding of Significant Contribution and
Rulemaking for Certain States in the Ozone Transport Assessment Group Region for Purposes of
Reducing Regional Transport of Ozone, Vol. 63, No, 90, May 11, 1998.
Federal Register, 1998b. Finding of Significant Contribution and Rulemaking on Section 126 Petitions
for Purposes of Reducing Interstate Ozone Transport. Vol. 63, No. 83, April 30, 1998,
Page 1-19
-------
Ozone Transport Assessment Group, 1995. Policy Paper approved by the Policy Group on December 4,
1995.
Ozone Transport Assessment Group, 1997. Executive Report 1997.
U.S. Environmental Protection Agency, 1996a. Fact Sheet on Nitrogen Oxides Emission Reduction
Program. Office of Air and Radiation, Washington, D.C., December 1996.
U.S. Environmental Protection Agency ,1996b. Guidance for Incorporating Environmental Justice
Concerns in EPA "s NEPA Compliance Analyses (Review Draft). Office of Federal Activities,
Washington, D.C., July 12, 1996.
U.S. Environmental Protection Agency, 1997a. Proposed Ozone Transport Rulemaking Regulatory
Analysis, Office of Air and Radiation, Washington, D.C., September 1997.
U.S. Environmental Protection Agency, I997b, Regulatory Impact Analysis for the Paniculate Matter
and Ozone National Ambient Air Quality Standards and Proposed Regional Haze Rule. Office of Air
Quality Planning and Standards, Research Triangle Park, NCS July 1997,
U.S. Environmental Protection Agency, 1997c. 1996 Compliance Report. EPA Report 43O/R-97-025,
Office of Air and Radiation, Washington D.C., June 1997.
U.S. Environmental Protection Agency, 1997d. Nitrogen Oxides Impacts on Public Health and the
Environment. EPA Report 452/R-97-003, Office of Air and Radiation, Washington, D.C., August 1997.
U.S. Environmental Protection Agency, 1997e. EPA Interim Guidance for Implementing the Small
Business Regulatory Enforcement Fairness Act and Related Provisions of the Regulatory Flexibility Act.
Februarys, 1997.
U.S. Environmental Protection Agency, 1998a. Supplemental Ozone Transport Rulemaking Regulatory
Analysis. Office of Air and Radiation, Washington, D.C., April 7, 1998.
U.S. Environmental Protection Agency, 1998b. Memorandum from Trovato and Kelly to Assistant
Administrators. Subject: "Implementation of Executive Order 13045, Protection of Children from
Environmental Health and Safety Risks". April 21, 1998.
Page 1-20
-------
Chapter 2. REGULATORY ALTERNATIVES AND EMISSIONS IMPACTS
This chapter explains the various regulatory alternatives considered in the benefits analyses.
Section 2.1 provides background on the elements that differentiate the alternatives that were considered,
and Section 2.2 provides information on the emissions impacts used to generate air quality changes for the
benefits analyses,
2,1 Elements Considered in Developing Regulatory Alternatives
EPA's NOx SIP call sets summer NOx emissions budgets for eastern States that the Agency
has found significantly contribute to the nonattainment by other States of the pre-existing ozone standard
(1-hour) and will contribute in the future to nonattainment by other States with the revised ozone standard
(8-hour), EPA relied heavily on its estimation of the NOx reductions that the electric power industry and
other stationary sources could provide cost-effectively in setting the State budgets. Other factors, such as
the feasibility of implementing controls in a reasonable time frame, also influenced the Agency's final
decisions. To estimate the cost-effectiveness of controls for various sources, the Agency considered
several ways that controls could be implemented in the SIP call region. However, States can place
controls on their sources of NOx emissions differently than the approach that EPA used in the budget
setting process, if they can show that control strategy will provide the same level of NOx reduction in the
SIP call region.
This section describes the elements that make up the various regulatory alternatives considered
for this analysis. The regulatory alternatives used in the benefits analysis (described in Section 2.2)
represent various combinations of these elements. Some elements of the rule remain the same for all the
options considered. Other elements are considered in varying combinations, including stringency of
controls, geographic scope, affected sources and design of the trading system. For all options analyzed,
the timing of regulatory requirements was also considered, as this issue is critical in terms of feasibility of
compliance and attainment of both the pre-existing ant the revised ozone standard.
2.1.1 Type of Control
EPA had to decide on the types of regulatory approaches that the Agency wanted States to
consider in their efforts to lower NOx emissions from various source categories. EPA used those
approaches in estimating the cost-effectiveness of ozone season NOx controls at various levels for
different types of sources. OTAG recommended that the Agency consider controls that allow for
emissions trading, rather than traditional command-and-control regulation. OTAG's analysis of trading
programs had shown that there could be considerable savings from this type of approach for the electric
power industry (OTAG, 1997).
EPA also demonstrated the potential savings from a NOx emissions trading program that could
result in its regulatory analysis for the proposed NOx SIP call (EPA, 1997a). That analysis showed that
in 2005 a command-and-control program for the electric power industry would cost about 30 percent
more than a trading program in the NOx SIP call region. For that reason, the Agency has focused heavily
Page 2-1
-------
on developing regulatory approaches that States can use collectively that are based on allowance-based
NOx emissions trading. It was also clear from OTAG analysis and EPA's own work that further
savings and flexibility could be gained from allowing banking as part of a trading program. EPA's
regulatory analysis over the last year has also considered banking options for inclusion in the Model
Trading Rule for States (EPA, I997b).
2.1.2 Geographic Scope
After considering OTAG's recommendations and other relevant information, EPA identified 22
States plus the District of Columbia (i.e., 23 jurisdictions) as significantly contributing to nonattainment
with, or interfering with maintenance of, air quality standards in a downwind State. The SIP call region is
shown in Figure 2-1 and consists of Alabama, Connecticut, Delaware, District of Columbia, Georgia,
Illinois, Indiana, Kentucky, Massachusetts, Maryland, Michigan, Missouri, North Carolina, New Jersey,
New York, Ohio, Pennsylvania, Rhode Island, South Carolina, Tennessee, Virginia, West Virginia, and
Wisconsin,
The final rule reflects State NOx budgets that are developed using the same region-wide
stringency targets and region-wide analyses of cost-effectiveness for all 23 jurisdictions. EPA also
considered dividing the SIP call region into two or three subregions in an effort to make a distinction
among the States that may contribute the most to the ozone transport problem and those where the wind
patterns may be less likely to affect air quality in the other States. The SIP call region was divided into
two regions—Northeast and Southeast, or into three regions—Northeast., Midwest, and Southeast.
Different levels of stringency are then applied in the different regions, as described below.
The two region area consists of Connecticut, Delaware, District of Columbia, Massachusetts,
Maryland, New Jersey, New York, Ohio, Pennsylvania, Rhode Island, Virginia, and West Virginia in the
Northeast; and Alabama, Georgia, Illinois, Indiana, Kentucky, Michigan, Missouri, North Carolina, South
Carolina, Tennessee, and Wisconsin in the Southeast.
The three region area consists of Connecticut, Delaware, District of Columbia, Maryland,
Massachusetts, New Jersey, New York, Pennsylvania, and Rhode Island in the Northeast; Illinois,
Indiana, Kentucky, Michigan, Missouri, Ohio, Virginia, West Virginia, and Wisconsin in the Midwest; and
Alabama, Georgia, North Carolina, South Carolina, and Tennessee in the Southeast.
Page 2-2
-------
Figure 2-1
States Included in EPA's NOx SIP call
Pace 2-3
-------
L_3] Ozone Transport Region States in the NOx SIP Call
[.U Other States in the NOx SIP Call
2.1.3 Potentially-Affected Sources
EPA has developed State budgets based on the effects of additional controls (beyond those
already required by the CAAA-related or reflected in existing SJPs) only for major stationary sources of
NOx emissions. These sources include: (1) electricity generating utility boilers; (2) industrial, commercial
and institutional boilers; (3) combustion turbines; (4) reciprocating internal combustion engines; (4) cement
manufacturing operations; and (5) other industrial processes that emit NOx. Only existing or planned
CAAA-related controls are considered in calculating budgets for other sectors (area and mobile sources)
that contribute to NOx emissions. States ultimately have discretion to determine which sources to
regulate to achieve the budget level.
The analyses of benefits in this RIA are based on a range of assumptions about which major
stationary sources will actually be targeted for additional controls by the States in implementing the NOx
SIP call. The primary assumption in this analysis is that States will allocate NOx emissions reduction
Page 2-4
-------
requirements to the largest electricity generating utility boilers; industrial, commercial and institutional
boilers; combustion turbines; cement manufacturing units; and internal combustion engines.
Large electricity generating units are defined as those generating more than 25 megawatts (MW),
Large industrial boilers, combustion turbines, reciprocating internal combustion engines, and other
industrial NOx sources are those capable of firing greater than 250 mmBtu/hour, or that emit greater than
one ton of NOx per summer day.
2.1,4 Stringency of Control Level
In order to develop a cost-effective NOx reduction strategy as a basis for establishing State
budgets, EPA considered various emission reduction levels for the affected sources for the summer ozone
season defined as May 1 through September 30. For the electricity generating units (EGUs), EPA
considered emissions budgets based on emission limits of 0,12 Ib/mmBtu, 0,15 lb/mmBtus 0,20 Ib/mmBtu,
and 0.25 Ib/mmBtu.1 For the large industrial boilers and combustion turbines, EPA considered a uniform
percent emission reduction from uncontrolled projected 2007 emission levels ranging from 40 percent to
70 percent. For the remaining large nonutility sources, EPA considered source category-specific control
levels corresponding to cost-effectiveness cut-offs ranging from $l,500/ton to $5,000/ton.
Taking into consideration the emission reductions and associated costs projected under each of
the above scenarios, EPA identified cost-effective NOx reduction strategies. Based on the reduced
emissions achieved by this strategy, EPA then established State-specific budgets for ozone season NOx
emissions. Alternative budgets are calculated for the different stringency levels considered for EGUs.
"The details of NOx budget development can be found in the budget technical support document (EPA,
1998b).
2.1.5 Effective Dates
States subject to the NOx SIP call must submit revised SIPs by September 1999. The affected
sources in the States must implement NOx controls by May 2003, and EPA will assess how each State's
SIP has actually performed in 2007,
2.1.6 Emissions Budget Trading System Design
To allow for use of the most cost-effective emission reduction alternatives, an emissions budget
trading program is an optional component of the NOx SIP call. Each of the States subject to the NOx
SIP call are encouraged to participate in this model NOx Budget Trading Program and thereby provide a
mechanism for sources to achieve cost-effective NOx reductions. The trading unit is a NOx Allowance,
Limits for each electricity generating unit are expressed as a specific NOx limit of pounds of NOx per
mmBtu of summer heat input projected for 2007, the year which was the focus of OTAG's analysis (the year for
which air quality modeling was done).
Page 2-5
-------
equal to one ton of emitted NOx. Details of the trading program are described in the Federal Register
notice accompanying the final rule.
Under the NOx Budget Trading Program, each of the participating States would determine how
its seasonal State trading program budget is allocated among its sources. Each source would be given a
certain quantity of NOx allowances. If a source's actual NOx emissions exceed its allocated NOx
allowances, the source may purchase additional allowances. Conversely, if a source's actual NOx
emissions are below its allocated NOx allowances, then it may sell the additional NOx allowances. Such
a program creates a competitive market for NOx allowances that encourages use of the most efficient
means for reducing NOx emissions.
For purposes of this analysis, trading may occur among any of the sources within the entire SIP
call region or within each of the subregions. Where subregions are developed for the SIP call region, only
infra-regional (within the region) trading is allowed.
Banking would allow sources that do not use all of their NOx allowances for a given year to save
them for later use. If banking is allowed, however, mechanisms such as flow controls can be put in place
to limit the level of exceedance of the emissions cap. Flow controls restrict the use of the banked NOx
allowances by restricting their use at certain times or within certain areas. For example, a restriction may
be placed on the banked allowances that allows only a set amount to be used during a defined time period.
For this RIA, EPA analyzed a variety of trading options, and trading with banking only for the .15
trading option, where banking begins after the start of the program in 2003. Banking of "early" reductions
was not modeled for the 0.15 option because earlier IPM analysis suggested that owners of electricity
generating units would want to use it to a very limited degree to lower the costs of future compliance
(EPA, 1997). The following considerations were part of the 1997 analysis:
• Beginning in 2003 (and each year thereafter), the fossil fuel-fired electricity generating units over
25 MW in the SIP call region are assumed to hold NOx allowances during the summer ozone
season equal to 489 thousand tons.
« Electricity generating units could trade allowances without restrictions or bank them for later use
or sale to another generation unit. Trading could occur within the entire SIP call region,
* Analysis with and without flow controls.
EPA's analysis in 1997 was conducted using the 1996 version of the Integrated Planning Model (IPM).
This model is described in EPA, 1996. EPA's analysis shows that on strict economic grounds, (i.e., under
minimization of the total direct operating costs over the simulation period) limited banking was forecasted
by the IPM based on the scenarios described above. However, EPA believes that some banking, which
the IPM could not estimate, should occur when some power plants overcontrol their NOx emissions in
order to bank allowances for use in years in which units experience utilization greater than forecasted.
More discussion of this issue can be found in Chapter 6 of Volume 1 of this RIA.
Page 2-6
-------
2,2 2007 Emissions Estimates for Air Quality Modeling
The initial step in the assessment of changes in air quality attributable to each regulatory
alternative is the development of future year 2007 emissions estimates, These estimates generally start
off with 1995 emissions data, which are then grown to 2007. Table 2-1 identifies the emissions inputs
used for the air quality models. These include nitrogen oxides (NOx), volatile organic compounds (VOC),
sulfur dioxide (SO2), directly emitted paniculate matter (primary PM10 and PM2 5)? carbon dioxide (CO),
and ammonia (NH3). RADM and RPM work in tandem and therefore require the same emissions inputs.
Emissions are estimated only for the geographic area covered by each air quality modeling domain, which
in each case is roughly equivalent to the 37 easternmost states. Air quality estimation is not restricted to
the smaller SIP call region because the SIP call regulatory alternatives may result in shifts in power
generation, and hence shifts in emissions, among utility sources located inside and outside the SIP call
domain. The broad 37 state area modeled for air quality purposes more clearly captures the effects of
any modeled shifts in power generation. All emissions estimates were developed using information that
was accurate as of March 1998, before final NOx SIP call emissions inventories and control alternatives
were established. The emissions inventories and control alternatives proposed in the original November 7,
1997 and supplemental May 11, 1998 Federal Register notices have changed as a result of public
comments received on both proposal notices. These changes are not reflected in the air quality modeling
results used in the benefits analysis.
The subsections that follow briefly describe emissions development for each emissions sector,
including electricity generating utility point sources, other stationary point sources, area and non-road
mobile sources, highway mobile sources, and non-anthropogenic sources. A final subsection identifies
differences between the air quality modeling emissions inventory and the final emissions inventory, and
discusses the implications for interpreting the air quality results used in this analysis.
Table 2-1
Emissions Inputs for Air Quality Models
Air Quality Model
UAM-V
RADM-RPM
S-R Matrix
Emissions Inputs
typical summer day hourly VOC and NOx
warm season (May - September) and cold season (October - April) NOx, SO2,
VOC, and NH3
annual NOx, SO2, CO, VOC, NOx, primary PM10, primary PM2 s, and Mi,
S04, CO,
2.2,1 Electricity Generating Unit Point Source Emissions
EPA developed projections of 2007 NOx and S02 emissions from electricity generating units
(EGUs) using the latest version of the Integrated Planning Model (U.S. EPA, 1998a). The CO and VOC
profiles for each EGU are added based on data from EPA's National Emission Trends (NET) inventory
projections. Primary PM10 and primary PM2 5 are derived using IPM-generated ash content data and the
Page 2-7
-------
latest AP-42 emission factors (U.S. EPA, 1998c). AP-42 emission factors are also used to derive NH3
emissions. These emissions estimates are made for the 2007 base case and each control alternative.
2.2.2 Non-EGU Point Source Emissions
EPA developed projections of 2007 NOx and VOC emissions using information gathered by the
Ozone Transport Assessment Group in 1997, These projections were later revised based on more recent
information, including public comments submitted on the proposed NOx SIP call. Emissions of SO2,
primary PM10, primary PM2J, and NH3 are taken from EPA's NET projections2. These emissions
estimates are made for the 2007 base case and each control alternative.
2.2.3 Area and Mobile Source Emissions
All area source, non-road mobile source, and highway mobile source emissions are taken directly
from EPA's NET projections. Emissions are developed for all counties and then allocated to UAM-V
and RADM grid cells, Additional reductions in area and mobile source emissions are not part of the
control alternatives, therefore emissions estimates are made for these source categories only for the 2007
base case.
2.2.4 Natural Emissions
Natural emissions come from geogenic, biogenic, and wild fire sources. For some pollutants,
natural emissions comprise a significant fraction of total emissions. For example, man-made emissions of
ammonia are a small component of total ammonia emissions. The majority of the ammonia that enters the
atmosphere is produced by the biological decomposition of organic material in soils, plant residues, and
wastes from animals and humans (NAPAP, 1991), Biogenic VOC emissions are developed based on
EPA's Biogenic Emissions Inventory System (BEIS) (Pierce et al.s 1990), Natural sources of PM
emissions (i.e., wind erosion) are taken from the NET. Additional reductions in natural emissions are not
part of the control alternatives, therefore emissions estimates for these sources are made only for the
2007 base case.
2,2,5 Summary of 2007 Emissions Projections
Table 3-2 summarizes the major control requirements that are accounted for in the 2007 Base
Case emissions projection. These include all federal motor vehicle controls and nonattainment area
(NAA)-related controls required by the Clean Air Act, additional reductions from large stationary NOx
sources required by the Ozone Transport Commission in the northeast U.S., and a national low emission
vehicle (NLEV) standard starting in model year 1999.
EPA's NET does not currently include estimates of primary PM2S. Primary PM2i emissions are estimated
from primary PMIO emissions using regionally derived relationships between PMIO and PM2 5.
Page 2-8
-------
The control requirements included in the alternative policy scenarios modeled for air quality
purposes, and hence for the purpose of estimating benefits, are shown in Table 2-3. The tables in
Appendix B provide additional information on the emissions associated with the scenarios that are
modeled for the benefits analyses. Since the scenarios modeled differ slightly from the regulatory
alternatives evaluated in the cost analyses (0.15 trading for EGUs, and 60%/$5,000 for non-EGUs) in
Volume 1 of this RIA, Table 2-3 also indicates the controls associated with preferred regulatory
alternative upon which EPA based the final emissions budgets. The air quality changes associated with
the set of control requirements in the actual final emissions budgets have not been estimated in this
analysis.
Table 2-4 provides some perspective on the potential differences between the emissions
reductions associated with the air quality modeling alternatives and the cost analysis alternatives. The
table indicates the percent change in NOx emissions associated with the air quality scenarios that are
modeled for the benefits analyses, and the percent changes associated with similar alternatives modeled in
the cost analyses. The RADM-RPM reductions for the warm season are generally higher than the
annual reductions for the S-R Matrix input. The seasonal RADM-RPM reductions and the annual S-R
Matrix reductions can be reconciled by considering the simple linear combination of RADM-RPM warm
and cold season emissions. For instance, for the 0.15 trading alternative, the linear combination of the
warm and cold season reductions is 24.7% (37.1%*5/12 + 15.8%*7/12), which is very close to the annual
S-R Matrix reduction of 24.0%.
Overall, the emission reductions associated with the air quality modeling inputs are larger than the
emission reductions associated with the cost analyses. For most of the regulatory alternatives, the
difference in modeled point source NOx reductions across the 37 states is about 5%. The difference for
EGU sources is small, around 2%, but the difference for non-EGU sources is much larger at about 10%,
As indicated in Table 2-3, this is because the air quality modeling inputs for non-EGU sources are based
on the proposed emissions budget level; the final emissions budget for non-EGU sources is less stringent
than proposed budget. The actual benefits associated with the preferred regulatory alternative and the
final emissions budgets will depend exactly on the specific seasonal and geographic distribution of
emissions changes.
Page 2-9
-------
Table 2-2
2007 Base Case Projection Control Requirements by Major Sector
Major Sector
Major Base Case Requirements
EGU Point
Sources
Title IV Phase I and Phase II NOx and SO2 limits for all boiler types,
250 ton Prevention of Significant Deterioration (PSD) and New Source Performance Standards
(NSPS) for NOx, NO2! VOC, CO, and SO,,
RACT and New Source Review (NSR) NOx limits for all non-waived NAAs.
Phase I of the Ozone Transport Commission (OTC) NOx memorandum of understanding (MOU).
:Non-EGU Point
Sources
VOC and NOX RACT for ail NAAs (except NOx waivers),
New control technique guidelines (CTGs) for VOC.
OTAG Level 2 NOx controls across OTAG States.
MACT standards (primarily affects VOC).
Area Sources
VOC and NOx RACT requirements.
New CTGs for VOC.
MACT Standards (primarily affects VOC).
PM10 NAA controls.
Onboard vapor recovery (vehicle refueling—VOC).
Stage II vapor recovery systems (VOC).
Federal rules (consumer/commercial product limits, architectural and industrial maintenance
(AIM) coating limits) (VOC).
Nonroad Mobile
Sources
Federal Tier 2 and 3 a 50 hp compression ignition (CI) engine standards (VOC, NOx, PM).
Federal Tier 1 and 2 < 50 hp CI engine standards.
Federal Phase 1 and 2 small (
-------
Table 2-3
Summary of Regulatory Alternatives for the 2007 Air Quality Modeling:
NOx Control Requirements by Major Sector
Major Sector
EGU Point Sources3
Non-EGU Point
Sources
Area
Nonroad Mobile
Sources
Highway Mobile
Sources
Control Requirements by Alternative
0,25 Trading
Emissions budget
based on 0.25
Ib/mmBTU limit
60% (UAM-V) or
70% (RPM/S-R)
reduction for >250
mmBtu/hour input
capacity; RACT for
all other sources
>1 tpdc
0.20 Trading
Emissions budget
based on 0.20
Ib/mmBTU limit
70% reduction for
>250 mmBtu/hour
input capacity;
RACT for all other
sources > I tpd
Regionality 1
Emissions budget
based on 0.1 5
Ib/mmBTU limit in
NE and 0.20
Ib/mmBTU in
MW/SEb
70% reduction for
>250 mmBtu/hour
input capacity;
RACT for all other
sources >1 tpd
0.15 Trading
Emissions budget
based on 0. 1 5
Ib/mmBTU limit
70% reduction for
>250 mmBtu/hour
input capacity;
RACT for all other
sources > 1 tpd
0.12 Trading
Emissions budget
based on 0.12
Ib/mmBTU limit
70% reduction for
>250 mmBtu/hour
input capacity;
RACT for all other
sources >I tpd
Preferred
Alternative
Emissions budget
based on 0.1 5
Ib/mmBTU limit
60% reduction for
boilers and turbines,
90% reduction for
reciprocating 1C
engines, 30%
reduction for cement
kilns'1
No additional control requirements.
No additional control requirements.
No additional control requirements.
' All EGU control requirements apply to fossil-fuel fired units greater than 25 MW.
b NE states include Connecticut, Delaware, District of Columbia, Massachasetts, Maryland. New Jersey, New York, Ohio, Pennsylvania, Rhode Island Virginia, and West Virginia.
MW/SE states include Alabama, Georgia, Illinois, Indiana, Kentucky, Michigan, Missouri, North Carolina, South Carolina, Tennessee, and Wisconsin. Emissions trading is allowed
within each region but not across regions.
e A 60% reduction strategy was used in the UAM-V and RADM modeling. To save additional time for the S-R Matrix modeling, a 70% reduction strategy was used to be consistent with
the remaining regulatory' alternatives.
a All non-EGU point source requirements for the preferred alternative apply to sources > 250 mrnBtuAoar input capacity, or where such data does not apply, > 1 ton per day NOx
emissions. For 1C engines and cement manufacturing, the emission reduction requirements correspond to the highest reduction achievable for the source category at tess than $5,000 per
ozone season ton.
Page 2-11
-------
Table 2-4
Percent Change from 2007 Base Case in 37-State NOx Emissions
Major Sector
0.25 Trading
0.20 Trading
Region ality 1
0.1 5 Trading
0.12 Trading
Ozone Season NOx Corresponding to UAM-V Inputs
EGU Sources
Non-EGU Sources
Total Point Sources
-26.1
-21.9
-24.7
-35.3
-27.7
-32.7
-39.0
-27.7
-35.2
-44.5
-27.7
-38.8
-49.2
-27.7
-41.8
Warm Season Annualized NOx Corresponding to RADM-RPM Inputs "
EGU Sources
Non-EGU Sources
Total Point Sources
-25.7
-15.8
-22.0
-34.7
-26.0
-31.5
-38.4
-25.9
-33.8
-43.7
-25.9
-37.1
-48.5
-25.8
-40.1
Cold Season Annualized NOx Corresponding to RADM-RPM Inputs a
EGU Sources
Non-EGU Sources
Total Point Sources
-9.1
-16.0
-11.8
-9.1
-26.3
-15.8
-9.1
-26.3
-15.8
-9.2
-26.3
-15.8
-9.4
-26.3
-15.9
Annual NOx Corresponding to S-R Matrix Inputs
EGU Sources
Non-EGU Sources
Total Point Sources
-15.9
-24.5
-19.3
-19.8
-24.5
-21.6
-21.3
-24.5
-22.6
-23.6
-24.5
-24.0
-25.8
-24.5
-25.3
Ozone Season NOx Corresponding to Cost Analyses
EGU Sources
Non-EGU Sources b
Total Point Sources
-24.9
-17.1
-22.1
-33.7
-17.1
-27.7
-37.2
-17.1
-30.0
-42.3
-17.1
-33.3
-47.0
-17.1
-36.2
' The RADM-RPM uses seasonal emissions inputs. The warm season is defined as the 5 months May - September, and the cold season is defined as the ? months October - Aprii. A
simple linear combination of the 5-month and 7-month NOx reductions will yield an approximation of the annual NOx reductions.
* The NOx reduction estimates for the non-EGU sources is based on the 60%/$5,000 per ton alternative, which is described as the "preferred alternative" in Table 2-3.
Page 2-12
-------
f>
-------
2.3 References
OTAG, 1997a, Draft of Costs o/NOx Control Strategies on Electric Power Generation Using (he
Integrated Planning Model, For incorporation into the OTAG Final Report, June 1997.
OTAG, 1997b. Trading and Incentives Work Group Report - Draft of OTAG Final Report. Chapter
7, June 1997.
U.S. Environmental Protection Agency, 1996, Analyzing Electric Power Generation under the CAAA.
Office of Air and Radiation, Washington, D.C., July 1996.
U.S. Environmental Protection Agency, 1997a. Proposed Ozone Transport Rulemaking Regulatory
Analysis, Office of Air and Radiation, Washington, D.C., September 1997.
U.S. Environmental Protection Agency, 1997b. Model NOx Cap and Trade Rule Workshop Working
Papers, Office of Atmospheric Programs, Washington, D.C., December 1997.
U.S. Environmental Protection Agency, 1998a. Analyzing Electric Power Generation under the
CAAA. Office of Air and Radiation, Washington, D.C., March 1998.
U.S. Environmental Protection Agency, 1998b. Development of Modeling Inventory and Budgets for
Regional NOx SIP Call, Office of Air Quality Planning and Standards, Research Triangle Park,
September 1998.
Page 2-14
-------
Chapter 3, AIR QUALITY IMPACTS
This chapter describes the methods for estimating air quality for the 2007 base case and several
alternative policy scenarios, EPA has estimated the air quality changes that have been linked to health,
welfare, and ecological benefits, including changes in ambient paniculate matter (PMIO and PM2S),
ambient ozone, nitrogen deposition, and visibility degradation. First, emission levels corresponding to the
base case and alternative policy scenarios are estimated, and this data is used to run air quality models.
Using the methods identified and described in Chapter 4, air quality changes are then associated with
human populations and ecosystems to estimate changes in health and welfare effects,
EPA has used a regional-scale version of the Urban Airshed Model (UAM-V) to estimate ozone
air quality. The Regional Acid Deposition Model (RADM) is used to estimate nitrogen deposition. To
estimate PM air quality and visibility degradation, EPA has used both the Regional Paniculate Model
(RPM) and a Source-Receptor Matrix (S-R Matrix) based on the Climatological Regional Dispersion
Model (CRDM). Two different PM models are used because the results of each model can be
interpreted to represent the range of atmospheric chemistry that may exist in the year 2007. The results
of RPM are used to represent a future eastern U.S. atmosphere where acid sulfate levels are still high
enough to control atmospheric chemistry, and more specifically ammonium nitrate particle formation. In
this circumstance, reductions in NOx emissions may result in non-linear responses in total fine particle
levels, involving both decreases and increases'. The results of the S-R Matrix are used to represent a
future eastern U.S. atmosphere where acid sulfate levels do not dominate particle formation chemistry.
In this case, reductions in NOx emissions would be expected to result more directly in reductions in PM.
Section 3.1 covers the estimation of ozone air quality using UAM-V. Section 3.2 covers the
estimation of particulate matter air quality using RPM, and section 3.3 discusses the S-R Matrix. Section
3.4 discusses the estimation of nitrogen deposition using RADM. Finally, section 3.5 covers the
estimation of visibility degradation using both RPM and the S-R Matrix.
3.1 Ozone Air Quality Estimates
The EPA has used the emissions inputs discussed in Chapter 2 with a regional-scale version of
the Urban Airshed Model (UAM-V) to estimate ozone air quality. UAM-V was the primary modeling
tool relied on by the OTAG process that provided the foundation for the NOx SIP call. Because it
accounts for spatial and temporal variations as well as differences in the reactivity of emissions, the
UAM-V is ideal for evaluating the air-quality effects of emission control scenarios.
Model inputs are prepared from observed meteorological, emissions, and air quality data for the
episode days using prognostic meteorological modeling and/or diagnostic and interpolative modeling
techniques. The model is then applied with these inputs, and the results are evaluated to determine model
performance. Once the model results have been evaluated and determined to perform within prescribed
1 See Appendix D for a more thorough discussion of non-linear chemistry and fine particle formation.
Page 3-1
-------
levels, the same base-case meteorological inputs are combined with modified or projected emission
inventories to simulate possible alternative future emission scenarios.
For this study, EPA uses the UAM-V modeling system for the eastern U.S. The modeling
system is applied for a base-year of 1995 and for several future-year scenarios, including a 2007 baseline
and several control strategy options. The UAM-V modeling system requires a variety of input Files that
contain information pertaining to the modeling domain and simulation period. These include gridded, day-
specific emissions estimates and meteorological fields; initial and boundary conditions; and land-use
information.
3.1.1 Modeling Domain
The modeling domain for this application is identical to that used by OTAG in their modeling
analyses. The domain encompasses most of the eastern U.S. and consists of two grids, as illustrated in
Figure 3-1. The horizontal resolution for the inner grid, or fine grid, which is the shaded area of Figure 3-
I, is approximately 12 km; this grid consists of seven vertical layers. The horizontal resolution for the
outer grid, which is the unshaded area of Figure 3-1, is approximately 36 km; this grid consists of five
vertical layers.The top of the modeling domain is 4000 meters above ground level.
3,1,2 Simulation Periods
The four OTAG multi-day simulation periods are used in this study to prepare the future-year
ozone profiles. These include the following 31 days: 4-11 July 1988, 16-21 July 1991, 22-29 July 1993,
and 4-18 July 1995 (the start-up days for each simulation period are not listed here). All four simulation
periods are characterized by high ozone concentrations in one or more portions of the eastern U.S.;
numerous exceedances of the 1-hour National Ambient Air Quality Standard for ozone were recorded
during each of the periods.
3.1.3 UAM-V Model Output
Standard output from the UAM-V modeling system includes: (1) hourly, gridded surface-layer
ozone concentrations (provided as hourly averages); (2) instantaneous ozone values for all grid cells and
layers for each hour of the simulation; and (3) detailed information on the sub-grid-scale plume-in-grid
treatment. For this study, hourly, gridded, surface-layer ozone concentrations were extracted from the
file containing hourly average ozone values. This information was used in the calculation of adjustment
factors as described in the following section.
Page 3-2
-------
Figure 3-1
UAM-V Modeling Domain
Page 3-3
-------
3.1.4 Converting Episode Estimates to Full-Season Profiles
The UAM-V runs generate surface layer hourly average ozone values for each of the 31 episode
days for a total of 744 hourly predictions for each model grid cell. These predictions are used in
conjunction with actual 1995 observations to generate ozone values for the entire ozone season.2
This procedure uses observed hourly ozone data obtained from EPA's AIRS monitor database
for 1995. Individual monitors are mapped onto the gridded UAM-V output, and the concentrations of the
corresponding grid cells are used to calculate an adjustment factor for each monitor-year-scenario. The
adjustment factor at monitor,/' is calculated as:
A ,. t , , , Mem(C°nCj,2Ml policy option"
Adjustment factor. = J- ——
' Mean(ConcJ>Bageyear-QM)
where Conc^,, is the estimated hourly ozone concentration at monitor j, in ppm, for the indicated year
and policy option. The adjustment factor is calculated only for concentrations above continental
background, which is assumed to be 0.04 ppm.
The estimated 2007 hourly ozone concentrations are obtained as follows:
Cone, , - ,„„„ ,. ,. = (Com. , . 100, - 0.04) x (Adjustment Factor) + 0,04
j, hour i, 2007 policy option ^ j> hour i, 1995 ' v J f
policy opfion =
Concj. hour i, 1995 * °'04
0.04
Therefore, values less than 0.04 ppm are assumed insensitive to control strategies and are not adjusted.
These air quality modeling procedures result in a set of estimated hourly ozone levels for each
existing monitor for every hour in 2007 for the baseline and for each policy option. Certain hourly
observations are identical in all scenarios (specifically hours outside the ozone season and 1995
observations below 0.04 ppm). Hours with ozone levels above 0.04 in the base year have new predicted
values based on the average change in the UAM-V modeling results for the four modeled July periods.
2 The 5-month ozone season for this analysis is defined as May to September for health benefits. For
agricultural benefits for some crops, the relevant growing season extends into April and into October and November.
In this analysis, no changes in ozone concentrations are assumed to occur outside the 5-month ozone season.
However, the ozone metric (SUM06) used to estimate certain crop yield benefits requires that the baseline level of
ozone concentrations be estimated for months outside the 5-month ozone season.
Page 3-4
-------
3.1.5 Extrapolating from Monitored to Unmonitored Grid Cells
Given monitor-specific hourly ozone data, ozone measures (e.g., daily average) for each grid-cell
are obtained in two steps: (1) hourly data are converted to an ozone measure of interest, such as the daily
average, and (2) monitor-specific ozone measures are used to estimate ozone measures at each grid cell.
The conversion from hourly data to ozone measures of interest is straightforward. The estimation of
ozone measures at each grid cell uses a convex polygon interpolation procedure method.
The convex polygon method interpolates air
quality estimates from the monitors to the center of
each population grid cell. The convex polygon is a
generalization of the planar interpolation method.
Rather than limit the selection of monitors to, say, three,
the convex polygon method identifies the set of monitors
that best "surrounds" the center of each grid cell. The
result of the convex polygon method is illustrated in
Diagram 1. The set of monitors that best surround the
grid cell are determined by identifying which monitor is
closest (considering both angular direction and horizontal
distance) in each direction from the grid cell center.
The set of monitors found using this approach will form
a convex polygon around the grid cell center. The
agricultural and forest effects analysis also uses air
quality interpolated using the convex polygon method. However, calculating the benefits for these
welfare categories is best accomplished by using air quality data at the county level, rather than the grid
cell level. Therefore, the convex polygon approach is used to estimate the ozone levels at the county
centroid locations for these analyses.
In an effort to avoid extrapolating from monitored grid cells that are too far removed from the
unmonitored cell to be representative, yet still maintain relatively complete geographic coverage of the
UAM-V modeling domain, monitors are dropped from the set of defining monitors for a population grid
cell if the monitor is greater than a 200 km. Grid-cells that have no monitors within 200 km are dropped
from the health effects analysis. An exception to this rule is made for the analysis of agricultural and
forest effects which uses the SUM06 index at the county level.3 Since ozone monitors are generally
located in areas of greatest population density, rather than in rural areas where crop and forest land are
more dense, the 200 km rule does not provide sufficient coverage in many rural counties in the OTAG
region. Therefore, for the SUM06 data set a 1,000 km range is used for any county that is not within 200
km of a monitor. This affects 214 counties, primarily those on the western edge of the region, including
all of South Dakota. After determining the final set of surrounding monitors, the location's air quality level
is calculated as an inverse-distance weighted average of the selected monitors.
Diagram 1: C
/'
* ,
onvex Pol]
X
*
4
fgon
3 The SUM06 is a crop exposure index which sums the ozone concentration for every hour that exceeds 0.06
ppm, within a 12-hour period from 8:00 A.M. to 8:00 P.M.
Page 3-5
-------
3.1.6 Ozone Air Quality Results
A summary of the ozone air quality profiles used to assess the benefits of alternative policy
scenarios is presented in Table 3-1. The average change in daytime hourly ozone values across the entire
UAM-V model domain ranges from -0,0005 ppm to -0.0008 ppm across alternatives. The population-
weighted average change across the same model domain ranges from -0.0004 person-ppm to -0.0007
person-ppm. The SUM06 index used in the agriculture and forest benefits analysis indicates a range of
changes from -9 to -16 percent.
Population-weighted changes in predicted 1-hour average and 8-hour average concentrations
above the level of each ambient air quality standard are presented in Appendix, B, These changes are
estimated for the total exposed population and for various subpopulations, including minority groups,
children, the elderly, and the impoverished. In the SIP call states, the predicted decline in total population
exposure above the 1-hour ozone standard level ranges from 51% to 74%. The predicted decline in total
population exposure above the 8-hour ozone standard level ranges from 31% to 51%.
Table 3-1
Summary of UAM-V Derived Hourly Ozone Air Quality for Daylight Hours (7am to 7pm)
During the Ozone Season
Statistic
Minimum (ppm) b
Maximum (ppm) b
Spatial Average (ppm)
Population-Weighted
Average (person-ppm) c
SUM06 d
2007 Base
Case
0.0230
0,0532
0.0426
0.0411
25.35
Change Relative To 2007 Base Case"
0.25
Trading
-0.0020
0.0003
-0.0005
-0.0004
-2.37
0.20
Trading
-0.0027
0,0008
-0.0006
-0.0005
-3.12
Reg, 1
-0.0029
0.0004
-0.0007
-0.0006
-3.28
0.15
Trading
-0.0029
0,0004
-0.0007
-0.0006
-3.67
0.12
Trading
-0.0036
0,0004
-0.0008
-0.0007
-4.14
* The change is defined as the control case value minus She base case value.
* The base case minimum (maximum) is the value for the county with the lowest (highest) seasonal Turn to 7 pm average, where the
season is defined as May through September. The change relative to the base case picks the minimum (maximum) from the set of
changes in all counties.
c Calculated by summing the product of (he projected 200? county population and the estimated 2007 county seasonal ozone
concentration, and then dividing by the total population.
d SUM06 is defined as the cumulative sum of hourly ozone concentrations over 0.06 ppm that occur from Sam to 8pm in the months
of May through September. (As noted, the other statistics in this Table are for the hours of 7am to 7pm.)
Page 3-6
-------
Note that only ozone concentrations during daylight hours are assumed to be affected by the
control strategies.4 UAM-V predictions of changes in nighttime values have not been validated in this
analysis, and therefore EPA has assumed no changes in nighttime values. Benefit values are therefore
potentially understated to the extent that the policy alternatives evaluated in this analysis result in
improvements in nighttime values, and to the extent that nighttime values are associated with health and
welfare effects.
The air quality technical support document for this RIA (Abt Associates, 1998) contains maps
showing the base case ozone concentrations and ozone concentration changes for each of five regulatory
alternatives (0,25 Trading, 0.20 Trading, Regionally 1, 0.15 Trading, and 0.12 Trading). These maps only
convey information about the 5-month ozone season used for the health benefits analysis.
3,2 PM Air Quality Estimates Using RPM
Ambient concentrations of PM are composed of directly emitted particles and of secondary
aerosols of sulfate, nitrate, ammonium, and organics. The EPA has used the emissions inputs discussed in
section 3.1 with both the Regional Paniculate Model (RPM) and the Source-Receptor Matrix (S-R
Matrix) to estimate PM air quality. Relative to the S-R Matrix, RPM is designed to more realistically
account for the complex chemical interactions that take place in the atmosphere in the secondary
formation of PM. This section discusses the use of RPM and section 3,4 discusses the S-R Matrix.
RPM is an "Eulerian" model that functions in tandem with the Regional Acid Deposition Model
(RADM). RPM predicts the chemistry, transport, and dynamics of the secondary aerosols of sulfate,
nitrate, ammonium, and organics. An evaluation of RPM predictions of sulfate, nitrate, and ammonium
against the Clean Air Status and Trends Network (CASTNet) was performed and RPM's predictions
were deemed reasonable for policy assessment purposes. RPM predictions of organic aerosol were not
used because RPM has a large under prediction for organics, the sources of which have not yet been
thoroughly characterized. Inputs to RPM include RADM-processed fields of water and RADM-
predicted fields of oxidants, nitric acid (for total nitrate), and ammonia (for total ammonia). The field of
total nitrate is partitioned between particulate nitrate and nitric acid, depending on numerous factors,
including the availability of ammonia, relative acid sulfate levels, and ambient temperature. RPM also
uses a subset of the RADM emissions inputs and the RADM meteorological fields as inputs.
3.2.1 Modeling Domain
The RPM domain is also the same as RADM, covering the eastern U.S. and southern Canada
from James Bay to the Florida Keys (see Figure 3-2). The RPM grid size and vertical resolution is the
same as RADM's, 80-km and 15 layers up to the top of the free troposphere, respectively, with the
lowest layer having a nominal height of 40 meters. This relatively large grid areas (e.g., compared to
UAM-V) is though reasonable for estimating large scale deposition, but introduces uncertainty in
estimating the location of PM changes that are related to estimating health effects in the benefits analysis.
4 In the UAM-V analysis, daytime hours are 7:00 A.M. to 7:00 P.M.
Page 3-7
-------
Page 3-8
-------
Figure 3-2
RADM-RPM Modeling Domain*
*Note: The dots are located at the center of the 80km grid squares that cover the eastern U.S. and southern Canada,
Page 3-9
-------
3.2.2 Simulation Periods
To develop annual estimates with seasonal controls, an aggregation set of 30 meteorological cases
is separated into a warm season set and a cold season set. Because RADM predicts chemistry on a
synoptic, or daily, time scale (chemical meteorology) an aggregation technique developed during NAPAP
is used to calculate annual estimates of acidic deposition. The development and evaluation of the
aggregation simulation set is described by Brooks et al, 1995. Meteorological cases with similar 850-mb
wind flow patterns were grouped by applying cluster analysis to classify the wind flow patterns from 1982
to 1985, resulting in 19 sampling groups, or strata. Meteorological cases were randomly selected from
each stratum; the number selected was based on the number of wind flow patterns in that stratum relative
to the number of patterns in each of the other strata, to approximate proportionate sampling. A total of
thirty cases are used in the current aggregation approach. Each case is run for 5 days, using a separate
initial condition for each season specific to the scenario being run, Outputs from only the last 3 days are
used to avoid the influence of initial conditions. For each emissions scenario modeled, seasonal initial
conditions are developed by running for 10 days with those emissions after starting with ambient
concentrations representative of clean continental conditions. Results for the 30 selected 3-day cases are
weighted according to the strata sampling frequencies to form annual averages. Application of the
aggregation technique is described in Dennis et al, 1990. Note, the aggregation method results in an
annual average produced by meteorology that is representative of many years of meteorology, a decade
or more, rather than for a single, given year.
While the aggregation method was developed for acidic deposition, it has been extended to daily
average particulate concentrations to calculate the annual mean, median and 90"' percentile of the
distribution. The applicability of the aggregation method to particulate matter was studied by Eder, et al.,
1996 using an extinction coefficient (bcxt) for mid-day estimate from human observations of visible range
at airports (Husar and Wilson, 1993). The thirty RADM aggregation cases were found to be very
representative from an extinction coefficient (inferred fine particuiate matter) perspective and sufficient
to derive annual estimates of fine particulate matter.
3.2.3 RPM Model Outputs
RPM outputs used in this analysis include ambient concentrations (measured in units of
micrograms per cubic meter, ug/m3) of particulate SO4", NO3", and NH/~ The outputs produced by the
simulation period aggregation method described in section 3.3.2 include the annual mean of daily average
ambient concentrations, and each decile of the distribution of daily average ambient concentrations.
However, the health effect concentration-response functions that are used to estimate changes in health
effects for each policy scenario require estimates of total PM. Section 3.3.4 discusses the procedures
used to estimate the remaining fraction of total PM at each location.
3.2.4 Development of Total PM Estimates
RPM provides the mean, the median, and the peak (90th percentile value) of daily concentrations,
but only for the major portion of PM that will change as a result of the NOx SIP call — i.e., the nitrate,
sulfate, and ammonium components (NSA). According to the latest assessment of PM data for the
Page 3-10
-------
NAAQS review, NSA comprise 48.2% of total fine particulate in the eastern U.S. (U.S. EPA, 1996a).
To proceed with the RPM results, it is necessary to first use the NSA concentrations to estimate total PM
concentrations at each location under the baseline and each control scenario for the NOx SIP call5.
A first step in using information supplied by RPM is to estimate distributional statistics for total
PM, using ambient air quality data currently being developed for the CAAA §812 analysis. The location-
specific inputs available from RPM and the upcoming §812 analysis are as follows:
» mean and peak PM in the §812 data (PMmoailj8i2, PM090j8!2);
• mean and peak NSA in the §812 data (NSAm(.ai]> 812, NSA0 90> gi2);
• mean, median, and peak NSA in the NOx SIP call baseline (NSAmcjin]bwdi,lc, NSAmediailib)lsei;ne,
« mean, median, and peak NSA in each control scenario of the NOx SIP call (NSAmci)ni mnlto],
NSAmKjtalli (jomroi, NSA0 90_ COJ1(ro|).
Subtracting the mean NSA from mean total PM, one obtains the "other" component of PM
(which includes such components as soil and elemental carbon):
Othermean;g|2 = I M,,,,,.,,,^^ - NSAm<,aUj812
It is assumed that the mean of this (location-specific) "other" component is the same in the NOx
SIP call baseline as it is in the §812 data:
O"lei'racan, baseline " O^'mean, 812 '
Total PM is estimated in the baseline as:
* mean, baseline ^^ ^ mean, baseline ^•^•meim, baseline •
To obtain an estimate of the 90th percentile level of PM in the baseline, it is assumed that the
proportion (p) of NSA]neanibasolillc to PMlllcanibas(!liTO is constant across days at the same location:
p * '^^nican, baseline ' *"*niean, baseline '^"•'"H, baseline ' i, baseline »
where i denotes the fth day. Given a constant ratio, p, the peak (90th percentile) day for baseline PM is
the same day as the peak day for baseline NSA. This implies:
* "^Q.90, baseline = ( 1 'P) " * A0 90_ baseline
In the last step, given the mean and 90th percentile point of the distribution of daily PM
concentrations (PMlw,an bBSC|jn<, and PM0!)0 ta^ij^), and assuming that the distribution can be fit well by a
5 The RPM estimates of NSA are anhydrous estimates. However, ambient measurements of NSA do
contain water, which can increase the total NSA mass by as much as 10% to 50%. Therefore, the estimates of total
NSA mass changes derived from RPM understate total NSA mass changes.
Page 3-11
-------
Gamma distribution, the analysis uses a maximum likelihood estimation to estimate the parameters of the
Gamma distribution that are most consistent with the estimated mean and peak values, A distribution of
daily PM values based on NSA components predicted by RPM may then be generated for the baseline
scenario. A similar procedure is used to estimate PM values in the control scenarios. Additional detail on
these procedures can be found in Abt, 1998.
3.2.5 RPM PM Air Quality Results
Table 3-2 provides a summary of the predicted ambient PMi0 and PM25 concentrations used in
this study. Since only the NSA fraction of total PM changes, the estimates of changes for PM!0 and
PM25 are identical. The concentration changes are generally very small. For the 0.15 option, annual
mean PM changes range from an increase of 0.29 jig/m3 to a decrease of-0.49 |ig/m3, with an average
annual mean change across the RPM domain of-0.06 ug/m3.
Population-weighted changes in RPM predicted annual mean PM2|5 and PM10 concentrations
above the level of each ambient air quality standard are presented in Appendix B. These changes are
estimated for the total exposed population and for various subpopulations, including minority groups,
children, the elderly, and the impoverished. In the SIP call states, the predicted decline in total population
exposure above the PM2|5 annual standard level ranges from 5% to 15%. There is no predicted change in
total population exposure above the PM)0 annual standard level because there is no predicted baseline
exposure above the standard.
The air quality technical support document for this RIA (Abt Associates, 1998) contains maps
showing the base case PM concentrations and PM concentration changes generated using RPM for each
of five regulatory alternatives (0.25 Trading, 0.20 Trading, Regionality I, 0.15 Trading, and 0.12 Trading).
3.3 PM Air Quality Estimates Using the S-R Matrix
The Source-Receptor Matrix (S-R Matrix) reflects the relationship between annual average PM
concentration values at a single receptor in each county (a hypothetical design value monitor sited at the
county population centroid) and the contribution by PM species to this concentration from each emission
source (E.H. Pechan, 1996). The receptors that are modeled include all U.S. county centroids plus
receptors in 10 Canadian provinces and 29 Mexican cities/states. The methodology used in this RIA for
estimating PM air quality concentrations using the S-R Matrix is similar to the method used in the July
1997 PM and Ozone NAAQS RIA (U.S. EPA, I997a). The S-R Matrix was developed using the
Climatalogical Regional Dispersion Model (CRDM), and has been calibrated using 1993 -1995 PM10 and
PM2 5 monitoring data. These calibration factors, referred to as "normalization factors," are applied to all
S-R Matrix predictions.
Page 3-12
-------
Table 3-2
Summary of RPM Derived PM Air Quality
Statistic
Minimum Annual Mean
PMlo(ng/m3)b
Maximum Annual Mean
PM10(ng/m3)»
Average Annual Mean
PM,0(re/m3)
Population-Weighted
Average Annual Mean
PM10 (person-j,ig/m3) c
Minimum Annual Mean
PM23 (ng/m3) b
Maximum Annual Mean
PM^dig/m')11
Average Annual Mean
PMj., (ng/mj)
Population-Weighted
Average Annual Mean
PM2 , (person- ng/m3) c
2007 Base
Case
15.45
35.91
26.76
26.46
6.65
22.63
14.96
14.53
Change Relative to 2007 Base Case"
0.25
Trading
-0.49
0,24
.0.04
-0.03
-0.49
0.24
-0.04
-0.03
0,20
Trading
-0,46
0.24
-0.05
-0.05
-0.46
0.24
-0.05
-0.05
Reg. 1
-0.45
0.26
-0.05
-0.05
-0.45
0.26
-0.05
-0.05
0.15
Trading
-0.49
0.29
-0.06
-0.05
-0.49
0.29
-0.06
-0,05
0.12
Trading
-0,52
0.18
-0.12
-0.13
-0,52
0.18
-0.12
-0.13
• The change is defined as ihe control case value minus the base case value. Note (hat there is no difference between the changes in
PM2 s and PM,n because RADM/RPM only estimates the change in nitrates and sulfates which are both in the PM2 5 fraction.
b The base case minimum (maximum) is the value for the county with the lowest (highest) annual average. The change relative to the
base case picks the minimum (maximum) from ihe set of changes in all counties.
c Calculated by summing the product of the projected 2007 county population and the estimated 2007 county PM concentration, and
then dividing by the iota! population.
3,3,1 Climatological Regional Dispersion Model
The CRDM uses assumptions similar to the Industrial Source Complex Short Term (ISCST3), an
EPA-reeommended short range Gaussian dispersion model. CRDM incorporates terms for wet and dry
deposition and chemical conversion of SO2 and NOx, and uses climatological summaries (annual average
mixing heights and joint frequency distributions of wind speed and direction) from 100 upper air
meteorological sites throughout North America. Meterological data for 1990 coupled with emissions data
from version 2.0 of the 1990 National Particulate Inventory were used to develop the S-R Matrix,
In order to evaluate the performance of the Phase II CRDM, model-predicted PM concentrations
and measured ambient PM concentrations were compared. Measured annual average PM
Page 3-13
-------
concentrations by chemical species from the Interagency Monitoring for Protection of Visual
Environments (IMPROVE) network were examined for the three-year period March 1988 - February
1991. This period was chosen because it relates closely to 1990 emissions and meteorological data used
in the CRDM, Since the IMPROVE network monitors are primarily concerned with evaluating visibility
impairment in predominantly rural Class I areas, these comparisons are incomplete due to the lack of
coverage in urban areas. With the exception of the fugitive dust component of PM2.5 and PM10, modeled
and measured concentrations of sulfate, nitrate and organics are comparable (Latimer, 1996).
The CRDM has also been benchmarked against the RADM-RPM for the Eastern U.S. using
1990 emissions and meteorology (U.S. EPA, 1997b). RADM-RPM incorporates more comprehensive
physics and chemistry to enable better characterization of secondarily-formed pollutants than Lagrangian-
based methods. In general, the CRDM results show a similar trend in sulfate and nitrate concentrations
within the same modeling region. Also, the CRDM-predicted annual average concentrations of sulfate
are within the range of RADM-RPM base-case predictions. Relative to RADM-RPM base case results,
CRDM appears to overpredict nitrate concentrations in the Midwest and underpredict nitrate
concentrations in the Mid-Atlantic states.
3.3.2 Development of the S-R Matrix
To develop the S-R Matrix, a nationwide total of 5,944 sources (i.e., industrial point, utility, area,
nonroad, and motor vehicle) of primary and precursor emissions were modeled with CRDM. In addition,
secondary organic aerosols formed from anthropogenic and biogenic VOC emissions were modeled.
Natural sources of PMIO and PM2J (i.e., wind erosion and wild fires) were also included. Emissions of
SO2, NOx, and ammonia were modeled in order to calculate ammonium sulfate and ammonium nitrate
concentrations, the primary particulate forms of sulfate and nitrate. The CRDM produced a matrix of
transfer coefficients for each of these primary and particulate precursor pollutants. These coefficients
can be applied to the emissions of any unit (area source or individual point source) to calculate a particular
source's contribution to a county receptor's total annual average PM10 or PM2J concentration. Each
individual unit in the inventory is associated with one of the modeled source types (i.e., area, point sources
with effective stack height of 0 to 250 m> 250 m to 500 m, and individual point sources with effective
stack height above 500 m) for each county.
The S-R Matrix transfer coefficients were adjusted to reflect concentrations of secondarily-
formed particulates (Latimer, 1996). First, the transfer coefficients for S02, NOx, and ammonia were
multiplied by the ratios of the molecular weights of su!fate/SO2, nitrate/nitrogen dioxide and
ammonium/ammonia to obtain concentrations of sulfate, nitrate and ammonium,6 The relative
concentrations in the atmosphere of ammonium sulfate and ammonium nitrate depend on complex
chemical reactions. In the presence of sulfate and nitric acid (the gas phase oxidation product of NOx),
ammonia reacts preferentially with sulfate to form particulate ammonium sulfate rather than react with
nitric acid to form particulate ammonium nitrate. Under conditions of excess ammonium and low
temperatures, ammonium nitrate forms. For each county receptor, the suifate-nitrate-ammonium
equilibrium is estimated based on the following simplifing assumptions:
6 Ratio of molecular weights; SuIfate/SO2= 1,50; nitrate/nitrogen dioxide = 1.35; ammonium/ammonia = 1.06.
Page 3-14
-------
» All sulfate is neutralized by ammonium;
* • Ammonium nitrate forms only when there is excess ammonium;
• Because ammonium nitrate forms only under relatively low temperatures, annual average particle
nitrate concentrations are divided by four assuming that sufficiently low temperatures are present
only one-quarter of the year.
Finally, the total particle mass of ammonium sulfate and ammonium nitrate is calculated.7
For application to the NOx SIP call, emissions data for only those counties located in the 37
OTAG states plus the District of Columbia are used. Because nationwide emissions are not used, the S-
R Matrix results are incomplete for air quality predictions in the counties located in states along the
western border of the OTAG domain. For example, emissions from New Mexico are expected to have a
significant downwind impact on ambient PM concentrations in neighboring counties in Texas. However,
New Mexico emissions are not estimated in this analysis. Incomplete air quality predictions for the six
western border states make unreliable any analysis that imposes a threshold for health effects (see
Chapter 11). As shown in Figure 3-3, EPA has chosen not to include the air quality results from the six
"buffer" states in the benefits analyses that are performed using the S-R Matrix results. Since the 31
remaining states are generally located more than 525 km (approximately 330 miles) from the states for
which emissions information is not available, the air quality results for the 31 states is believed to be more
reliable.
3,3,3 Fugitive Dust Adjustment Factor
As indicated in subsection 3.4.1, the 1990 CRDM predictions for fugitive dust are not consistent
with measured ambient data. The CRDM-predicted average fugitive dust contribution to total PM2 5 mass
is 31% in the East and 32% in the West (E.H. Pechan, 1997b). Speciated monitoring data from the
IMPROVE network show that minerals (i.e., crustal material) comprise approximately 5% of PM25 mass
in the East and approximately 15% of PM25 mass in the West (U.S. EPA, 1996a). These disparate
results suggest a systematic overbias in the fugitive dust contribution to total PM. This overestimate is
further complicated by the recognition that the 1990 NPI significantly overestimates fugitive dust
emissions. The most recent National Emissions Trends inventory indicates that the NPI overestimates
fugitive dust PM10 and PM25 emissions by 40% and 73% respectively8 (U.S. EPA, 1997c).
7 To calculate total particle mass of ammonium sulfate and ammonium nitrate, the anion concentrations of
sulfate and nitrate are multiplied by 1.375 and 1.290 respectively.
8 Natural and man-made fugitive dust emissions account for 86% of PM10 emissions and 59% of PM25
emissions in the most recent 1990 estimates in the National Emission Trends Inventory.
Page 3-15
-------
Figure 3-3
S-R Matrix
Air Quality Modeling Domain
Page 3-16
-------
*Buffer states are shaded-emissions are included for these states, but air quality concentrations are not calculated for these states.
Pase3-17
-------
To address this bias, a multiplicative factor of 0,25 is applied nationally to fugitive dust emissions
as a reasonable first-order attempt to reconcile differences between modeled predictions of PM25 and
actual ambient data. This is the same adjustment that was used in the 1997 PM NAAQS RIA. A 0.25
multiplicative adjustment results in a fugitive dust contribution to modeled ambient PM2.5 concentrations of
10% to 17%.' Even after this adjustment the fugitive dust fraction of total eastern PM2.5 mass is 10,4%,
which is still greater than the 5% indicated by IMPROVE monitors. However, given that the 0.25
multiplicative factor appears to bring the modeled fugitive dust contribution to PM25 mass more within the
range of values reported from speciated monitoring data, the fugitive dust contribution to total PM that is
estimated by the S-R Matrix is adjusted by this factor. Since this factor still may result in an
overprediction of the fugitive dust contribution, the S-R Matrix may tend to underpredict the effectiveness
of strategies that affect NSA.
3.3.4 Normalizing S-R Matrix Results to Measured Data
In an attempt to further ensure comparability between S-R Matrix results and measured annual
average PM values, the S-R results are calibrated using factors developed for the PM and Ozone
NAAQS RIA (U.S. EPA, 1997a). For the NAAQS RIA, a "normalization factor" was developed for
each Tier 1 to Tier 3 monitored county10. Nonmonitored counties were calibrated using the appropriate
regional normalization factor calculated as the average of Tier 1 normalization factors across a given
modeling region. The- normalization factor was calculated as the monitored value divided by the modeled
value.
All S-R Matrix predictions were normalized regardless of overprediction or underprediction
relative to monitored values. This factor was applied equally across all particle species contributing to the
annual average PM value at a county-level receptor.
The calibration procedure was conducted employing 1993 - 1995 PMIO ambient monitoring data
from the AIRS database following the air quality tier data completeness parameters discussed above.
The PM,0 data represent the annual average of design value monitors averaged over three years (U.S.
* See U.S. EPA, 1997b, page 6-5 for a map delineating modeling regions. Using 0.25 multiplicative factor,
fugitive dust as percentage of PMZ5 mass for: Central U.S. = 17.2%; Eastern U.S.= 10.4%; Western U.S.= 10.6%. By
comparison, without using a multiplicative factor, fugitive dust as a percentage of PM25 mass for: Central U.S. =
44.6%; Eastern U.S. = 30.9%; Western U.S. = 31.5%.
10 The normalization procedure was conducted for county-level modeled PM!0 and PM2 5 estimates falling
into one of four air quality data tiers. The tiering scheme reflects increasing relaxation of data completeness criteria
and therefore increasing uncertainty for the annual design value (U.S. EPA, 1997f). Nationwide, Tier 1 monitored
counties cover the 504 counties with at least 50% data completeness and therefore have the highest ievel of
certainty associated with the annual design value. Tier 2 monitored counties cover 100 additional counties with at
least one data point (i.e., one 24-hour value) for each of the three years during the period 1993 -1995. Tier 3
monitored counties cover 107 additional counties with missing monitoring data for one or two of the three years 1993
-1995, In total., Tiers I, 2 and 3 cover 711 counties currently monitored for PMIO in the 48 contiguous states. In 1997
the PM10 monitoring network consisted of approximately 1600 individual monitors with a coverage of approximately
711 counties in the 48 contiguous states. Tier 4 covers the remaining 2369 nonmonitored counties.
Page 3-18
-------
EPA, 1997d), The standardization for temperature and pressure was eliminated from this concentration
data based upon proposed revisions to the reference method for PM!0."
Because there is little PM25 monitoring data available, a general linear model was developed to
predict PM2.S concentrations directly from the 1993 - 1995 PM10 values (U.S. EPA, 1996b). A SAS™
general linear model (i.e., GLM) procedure was used to predict PM25 values (dependent variable) as a
function of independent variables for season, region, and measured PM10 value. These derived PM25
data were used to calibrate model predictions of annual average PM2S,
3.3.5 Development of Annual Median PM2S Concentrations
The CRDM procedure does not directly produce estimates of daily 24-hour average PM
concentrations or annual median PM concentrations. Some health benefits have concentration-response
functions that rely on estimates of either the daily 24-hour average or annual median concentrations.
Using historical data, EPA developed 24-hour average estimates corresponding to the 99th percentile
value for PM,0 and the 98th percentile value for PM2 5 reflecting forms of PM!0 and PM2 5 daily
standards.
Peak-to-mean ratios (i.e., ratio of the 24-hour averge value to annual average value) are
established from actual PM10 monitor data for 1993 to 1995 from Tier 1 through Tier 3 monitored
counties. For PM)0, the peak value is defined exactly the way it is for the new PM!0 NAAQS, i.e., the
value corresponding to the 99th percentile value of the distribution of actual daily 24-hour average PMi0
values. For PM2.5, the peak value is also defined exactly the way it is for the new PM2 5 NAAQS, i.e.,
the value corresponding to the 98th percentile value of the distribution of estimated daily 24-hour average
PM2_5 values. These historical peak-to-mean ratios for each monitored county are assumed to hold for
the 2007 model year in this analysis and are applied to the annual average PM estimates generated by the
S-R Matrix, Peak values in nonmonitored counties are estimated using the regional average peak-to-
mean ratios in Tier 1 monitored counties.
Starting with the annual mean and peak values developed from the S-R Matrix, maximum
likelihood is then used to estimate the parameters of a Gamma distribution that are most consistent with
the S-R Matrix results. The parameters of the Gamma distribution are then used to estimate the annual
median concentration and the concentration corresponding to each decile of the distribution.
" See Appendix J - Reference Method for PM|0, Final Rule for National Ambient Air Quality Standards for
Participate Matter (Federal Register, Vol. 62, No. 138, p. 41, July 18, 1997).
Page 3-19
-------
3.3.6 S-R Matrix PM Air Quality Results
Table 3-3 provides a summary of the predicted ambient PM!0 and PM2iS concentrations used in
this study. Similar to the results using the RPM approach, the concentration changes are generally very
small. For the 0.15 option, annual mean PM10 changes range from an increase of 0.26 ng/m3 to a
decrease of -0.57 ug/m3, with an average annual mean change across the 31 state domain of -0.04 ng/m3
Therefore, the absolute changes in PM occur within a slightly wider band than with the RPM, and the
average annual mean change is slightly lower (-0.04 jug/m3 versus -0.06 ug/m3).
Table 3-3
Summary of S-R Matrix Derived PM Air Quality
Statistic
Minimum Annual Mean
PM,0(u-g/m3)b
Maximum Annual Mean
PM,0(ng/mV
Average Annual Mean
ipM,0(ng/m3)
Population-Weighted
Average Annual Mean
PM]0 (person-ng/m3) c
Minimum Annual Mean
PMwCug/mY
Maximum Annual Mean
PM2S(ng/m3)"
Average Annual Mean
PMj.5 (Ug/m3)
Population-Weighted
Average Annual Mean
PMj , (person-jig/m3) c
2007 Base
Case
5.39
66.37
22,62
25.96
3.49
27.63
10,74
12.62
Change Relative To 2007 Base Case "
0.25
Trading
-0.56
0.20
-0.03
-0.03
-0.52
0,20
-0,03
-0.03
0,20
Trading
-0.56
0.2!
-0.03
-0.04
-0,52
0.21
-0.03
-0.04
Reg. 1
-0.53
0.26
-0.02
-0.03
-0.49
0.25
-0.02
-0.03
0.15
Trading
-0.57
0.26
-0.04
-0.04
-0.53
0,26
-0.04
-0.04
0.12
Trading
-0.67
0.16
-0.06
-0.07
-0.67
0,17
-0,05
-0.06
" The change is defined as She control case value minus the base case value.
11 The base case minimum (maximum) is the value for the county with the lowest (highest) annual average, The change relative to the
base case picks the minimum (maximum) from the set of changes in all counties.
c Calculated by summing the product of the projected 2007 county population and (he estimated 2007 county PM concentration, and
then dividing by the total population in ihe 31 slates modeled using the S-R Matrix1
Page 3-20
-------
Population-weighted air quality changes were not estimated using the S-R Matrix results. The
results generated from the RPM modeling shown in Appendix B should be generally representative of the
direction and magnitude of changes that would be estimated using the S-R Matrix results.
The air quality technical support document for this RIA (Abt Associates, 1998) contains maps
showing the base case PM concentrations and PM concentration changes generated using the S-R Matrix
for each of five regulatory alternatives (0,25 Trading, 0.20 Trading, Regionally 1, 0.15 Trading, and 0.12
Trading), Similar maps can also be found in Pechan, 1998.
3.4 Nitrogen Deposition Estimates
Nitrogen deposition estimates are generated using RADM, The RADM was developed over a
ten year period, 1984 - 1993, under the auspices of the National Acid Precipitation Assessment Program
to address policy and technical issues associated with acidic deposition. The model is designed to provide
a scientific basis for predicting changes in deposition and air quality resulting from changes in precursor
emissions and to predict the levels of acidic deposition in certain sensitive receptor regions. To do so
requires that RADM be a multipollutant model that predicts the oxidizing capacity of the atmosphere,
including the prediction of ozone, and chemical transformations involving oxides of sulfur and nitrogen.
The development, application, and evaluation of the RADM has been documented extensively by
NAPAP (Chang, et al. 1987 & 1990; Dennis et al, 1990). RADM has been used in several recent
studies of acidic deposition, including BPA's 1995 Acid Deposition Standard Feasibility Study Report
to Congress (U.S. EPA, 1995), EPA's 1997 Deposition of Air Pollutants to the Great Waters Report
to Congress (U.S. EPA, 1997e), and in work estmating the nitrogen deposition airshed of the Chesapeake
Bay watershed (Dennis, 1997),
RADM estimates deposition in units of kilograms per hectare (kg/ha). Wet deposition is
estimated in the form of SO42", NCy, NH3> H+, Dry deposition is estimated in the form of SO2, SO4 as
aerosol, O3, HNO3, NO2, H2O2, The deposition estimates are mapped to specific East Coast and Gulf
Coast estuaries and their watersheds12. Land deposited nitrogen in each watershed is multiplied by a
factor of 10% to obtain the nitrogen load delivered via export (pass-through) to the corresponding estuary.
Table 3-4 provides a summary of the nitrogen deposition estimates for each cell in the RADM
domain. The changes range from 0.01 kg/ha to 3.66 kg/ha. The results for the 0.15 option represent an
11% reduction in the average annual deposition across the entire domain. The air quality technical
support document for this RIA (Abt Associates, 1998) contains maps showing the nitrogen deposition
changes generated using RADM for each of five regulatory alternatives (0,25 Trading, 0,20 Trading,
Regionality 1, 0.15 Trading, and 0.12 Trading). Another technical support document for this RIA (EPA,
1998d) contains additional information on the reduction in nitrogen loads to 12 study set estuaries.
12 EPA has developed a methodology to assess nitrogen deposition benefits directly for 12 different
estuaries: Albemarle/ Pamlico Sounds, Cape Cod Bay, Chesapeake Bay, Delaware Bay, Delaware Inland Bays,
Gardiners Bay, Hudson R,/ Raritan Bay, Long Island Sound, Massachusetts Bay, Narragansett Bay, Sarasota Bay,
and Tampa Bay.
Page 3-21
-------
Table 3-4
Summary of 2007 Nitrogen Deposition in RADM Domain
Statistic
Minimum Annual
Deposition (kg/ha)
Maximum Annual
Deposition (kg/ha)
Average Annual
Deposition (kg/ha)
2007 Base
Case
0.74
31,91
4.83
Change Relative To 2007 Base Case
0.2S
Trading
0.00
-2.62
-0.33
0.20
Trading
0.00
-3.19
-0.47
Reg. 1
-0.01
-3.22
-0.50
0.15
Trading
-0,01
-3.58
-0.55
0,12
Trading
-0.01
-3.66
-0,57
3.5 Visibility Degradation Estimates Using RPM and tlie S-R Matrix
Visibility degradation is often directly proportional to decreases in light transmittance in the
atmosphere. Light transmittance is attenuated by scattering and absorption by both gases and particles.
The light-extinction coefficient is a measure of the total fraction of light that is attenuated per unit distance
(Sisler, 1996):
where:
b,
b,,
ag
total light extinction coefficient (1/Mm),
light extinction coefficient due to natural Rayleigh scatter (I/Mm),
light extinction coefficient due to scattering by particles (I/Mm),
light extinction coefficient due to absorption by gases (I/Mm), and
light extinction coefficient due to absorption by particles (I/Mm).
The light extinction coefficient is calculated by multiplying the concentration of aerosol species and
particles species by their corresponding light-extinction efficiency, and summing over all species.
The term Z> can be broken into the various species of fine and coarse particles that scatter light.
Because fine particles are much more efficient at light scattering than coarse particles, several fine
particle species are specified, whereas coarse particles are kept as one category. Fine particles with
significant light-extinction efficiencies include sulfates, nitrates, organic carbon, elemental carbon (soot),
and soil (Sisler, 1996).
Page 3-22
-------
Once the light-extinction coefficient is determined, the visibility index called deciview (dv) can be
calculated (Sisler, 1996):
dv * 10
where:
10"3 = constant to convert M m"1 to km"1.
A change of one deciview represents a change of approximately ten percent in b^,, "which is a small but
perceptible scenic change under many circumstances" (Sisler, 1996, p. I-7),
Visibility degradation estimates in "recreational" (e.g., federally designated Class I areas such as
national parks and recreation areas) and "residential" (non-Class I areas) areas are generated using the
results of both RPM and the S-R Matrix. RPM computes the light extinction due to the combined
scattering of light by sulfates, nitrates, and organics which are part of the bv term of the extinction
equation. RPM does not include the soil part ofbsp, but it does indirectly account for the soot part through
a slight modification of the index of refraction for particles. RPM does not include gas absorption (by
NO2) because it is quite small. Rayleigh scattering is accounted for in RPM's calculation of deciview,
because Rayleigh is taken as the "floor" for deciview; however, Rayleigh scattering is not in RPM's
calculation ofbexl. Therefore, RPM estimates underpredicl total light extinction. A correction for this
underprediction is discussed in Chapter 4, section 4.4.5.
The S-R Matrix results are also used to estimate county-level light extinction. The recreational
visibility calculations use the full set of parameters in the extinction equation, due to the availability of data
from IMPROVE monitors, while the residential visibility calculations use county-level sulfate, nitrate and
coarse particle estimates. Using less than the full set of terms in the residential visibility calculations leads
to an underestimate of light extinction, and as noted for RPM a correction for this underprediction is
presented in Chapter 4, section 4.4.5.
The visibility benefits analysis (see Chapter 4) distinguishes between general regional visibility
degradation and visibility degradation in certain Federally-designated Class I areas (i.e., national parks,
forests, recreation areas, wilderness areas, etc.). Therefore visibility degradation estimates are separated
into "residential" and "recreational" categories depending upon the geographic area covered by the
estimate.
Table 3-5 provides a summary of the visibility degradation estimates derived both from the RPM
and the S-R Matrix results in terms of deciviews. The valuation methodology for recreational visibility
requires separate treatment of visibility changes in Class I areas in the Southeast region versus Class I
areas in the Central and Northeast regions. Table 3-5 provides visibility degradation estimates for both
regions. All predicted visibility changes are small, with the largest residential changes occuring the the
Southeast region. The S-R Matrix predictions are similar to the RPM predictions in the Southeast for
each alternative except 0,12 trading, but are slightly smaller than the RPM predictions in the Central and
Northeast regions.
Page 3-23
-------
Table 3-5
Summary of 2007 Visibility Degradation Estimates
(dcciviews)
Visibility Degradation
2007 Base
Case"
Change Relative To 2007 Base Case "
0.25
Trading
0.20
Trading
Reg. 1
0,15
Trading
0.12
Trading
Southeast *
RPM Annual Average-
Residential
3RPM Annual Average-
Recreational c
S-R Matrix Annual
Average—Residential
S-R Matrix Annual
Average— Recreational c
21.87
20.27
20.13
23.06
-0.07
-0.12
-0.08
-0.07
-0.08
-0.16
-0.08
-0.07
-0.07
-0.16
-0.07
-0.06
-0.11
-0.15
-0.10
-0.08
-0.22
-0.24
-0.12
-0.09
Central & Northeast *
RPM Annual Average-
Residential
RPM Annual Average--
Recreational d
S-R Matrix Annual
Average— Residential
S-R Matrix Annual
Average— Recreational d
17,12
15.87
18.77
19.39
-0.02
-0.01
-0.02
0.02
-0.04
-0.03
-0.01
0.02
-0.03
-0.03
0.00
0.02
-0.03
-0.06
-0.02
0.02
-0.09
-0.10
-0,05
-0.01
" The RPM only accounts for visibility degradation due to suifates and nitrates, but not other variables such as coarse PM and organic
matter. This leads RPM to underestimate total residential and recreational visibility degradation. However, the absolute change from
lie base case to Hie control options is essentially correct since the change in suifates and nitrates dominates the change in oilier
variables affecting visibility. The S-R Matrix recreational estimates arc correct since they include all visibility variables, but the S-R
Matrix residential estimates only account for suifates, nitrates, and coarse PM, and thus underestimate total residential visibility. As
with RPM, the S-R matrix correctly estimates the change in visibility from the base case to the control options. The effect of
underestimating visibility on the dollar value of visibility change is discussed in Chapter 1 .
b The change is defined as the control case deciview level minus the base case deciview level.
e The four Southeastern national parks are: Shenandoah, Mammoth Cave, Great Smoky Mountains, and the Everlglades.
** The three Central and Northeastern national parks are: Voyageurs, Isle Royale, and Acadia.
5 The Southeast region includes the following 12 states plus Washington D.C.: AL, DE, FL, GA, KY, MD, MS, NC, SC, TN, VA, WV.
The remaining states in the 37 state OTAG region comprise the Central and Northeast regions.
Page 3-24
-------
Table 3-6 presents changes in the population-weighted visibility degradation estimates from the
RPM in terms of annual mean extinction. As shown, the change is much less than 1% for each
alternative except 0.12 trading.
Table 3-6
2007 Population-Weighted Sum of Annual Mean Extinction;
RPM Results
Percent Change from Base Case
0.25
Trading
-0.3
0,20
Trading
-0,4
Rcgionality
I
-0.4
0.15
Trading
-0.3
0.12
Trading
-1.3
The air quality technical support document for this RIA (Abt Associates, 1998) contains maps
showing the base case visibility degradation and visibility degradation changes generated using both RPM
and the S-R Matrix for each of five regulatory alternatives (0.25 Trading, 0.20 Trading, Regionally 1,
0.15 Trading, and 0.12 Trading). Maps showing visibility degradation changes for selected Class I areas
can also be found in Pechan, 1998.
3.6 References
Abt Associates, Inc, 1998. Air Qualify Estimation for the NOx SIP Call RIA. Prepared for the U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park,
NC. September 1998.
Brooks, J.R., P.J. Samson and S. Sillman, 1995a. "Aggregation of selected three-day periods to estimate
annual and seasonal wet deposition totals for sulfate, nitrate, and acidity. Part I: A synoptic and chemical
climatology for eastern North America." Journal of Applied Meteorology, 34(2), 297-325.
Brooks, J.R., P.J. Samson and S. Sillman, 1995b, "Aggregation of selected three-day periods to estimate
annual and seasonal wet deposition totals for suffate, nitrate, and acidity. Part II: Selection of events,
deposition totals, and source-receptor relationships." Journal of Applied Meteorology, 34(2), 326-339.
Chang, J.S., R.A. Brest, I.S.A. Isaksen, S. Madronich, P. Middleton, W.R. Stockwell, and C.J. Walcek,
1987. "A Three-Dimensional Eulerian Acid Deposition Model: Physical Concepts and Formulation." J.
of Geophysical Research, 92, 14,681-14,700.
Chang, J.S., P.B. Middleton, W.R. Stockwell, C.J. Walcek, I.E. Pleim, H.H. Lansford, F.S. Binkowski, S.
Madronich, N.L. Seaman, and D.R. Stauffer, 1990. "The Regional Acid Deposition Model and
Engineering Model. NAPAP SOS/T Report 4." In: Acidic Deposition: State of Science and
Page 3-25
-------
Technology, Volume /. National Acid Precipitation Assessment Program, Washington D.C., December
1990,
Dennis, R.L. 1997. "Using the Regional Acid Deposition Model to Determine the Nitrogen Deposition
Airshed of the Chesapeake Bay Watershed," In: Joel E, Baker, editor. Atmospheric Deposition to the
Great Lakes and Coastal Waters, Society of Environmental Toxicology and Chemistry, Permsacola, FL,
pp 393-413.
Dennis, R.L., W.R. Barchet, T.L. Clark, and S.K. Seilkop, 1990. "Evaluation of Regional Acid
Deposition Models (Part I). NAPAP SOS/T Report 5." In: Acidic Deposition: State of Science and
Technology, Volume 1. National Acid Precipitation Assessment Program, Washington, D.C., September
1990.
Dennis, R.L., F.S. Binkowski, T.L. Clark, J.N. McHenry, SJ. Reynolds, and S.K. Seilkop, 1990.
"Selected Applications of the Regional Acid Deposition Model and Engineering Model, Appendix 5F (Part
2) of NAPAP SOS/T Report 5." In: National Acid Precipitation Assessment Program; Slate of
Science and Technology, Volume /, National Acid Precipitation Assessment Program, Washington,
DC., September 1990.
E.H. Pechan and Associates, 1996. Regional Paniculate Control Strategies Phase II. Prepared for
the U.S. Environmental Protection Agency, Office of Policy, Planning, and Evaluation, Washington, DC,
September, 1996.
E. H, Pechan and Associates, 1997. Control Measure Analysis of Ozone and PM Alternatives:
Methodology and Results. Prepared for Innovative Strategies and Economics Group, Office of Air
Quality Planning and Standards, U.S. EPA. Research Triangle Park, NC. July 1997.
Eder, B.K., S.K. LeDuc and F.D. Vestal, 1996. "Aggregation of selected RADM simulations to estimate
annual ambient air concentrations of fine particulate matter." In: Preprints, Ninth Joint conference of
Applications of Air Pollution Meteorology with A&WMA, January 28-February 2, 1996, Atlanta, Georgia.
American Meteorological Society, Boston, MA, pp 390-392.
Eder, B.K. and S.K. LeDuc, 1996. "Can selected RADM simulations be aggregated to estimate annual
concentrations of fine particulate matter?" In: Measurement of toxic and related air pollutants. VIP-
64. Proceedings of an International Specialty Conference, Research Triangle Park, NC, May 7-9, 1996.
Air & Waste Management Association, Pittsburgh, PA, pp 732-739.
Husar, R.B. and W.E, Wilson, 1993. "Haze and sulfur emission trends in the eastern United States."
Environmental Science and Technology, 29, pp 13-16.
Latimer and Associates, 1996. Paniculate Matter Source - Receptor Relationships Between All Point
and Area Sources in the United States and PSD Class/Area Receptors. Prepared for Bruce
Polkowsky, Office of Air Quality Planning and Standards, U.S. EPA. Research Triangle Park, NC.
September 1996.
Page 3-26
-------
National Acid Precipitation Program, 1991. NAPAP 1990 Interim Assessment Report, Volume II.
Emissions and Controls, Washington, DC. November 1991.
Pechan-A vanti Group, 1998. PM and RH A ir Quality Analysis for the NOx SIP Call RIA, Prepared
for the U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research
Triangle Park, NC. September 9, 1998,
Pierce, T., B. Lamb, and A. VartMeter, 1990, "Development of a Biogenic Emissions Inventory System
for Regional Scale Air Pollution Models." Proceedings of the 83rd Air and Waste Mangement
Association Annual Meeting, Pittsburgh, PA. 1990.
Sisler, J.F., 1996. Spatial ana'Seasonal Patterns and Long Term Variability of the Composition of
the Haze in the United States: An Analysis of Data from the IMPROVE Network, Cooperative
Institute for Research in the Atmosphere, Colorado State University; Fort Collins, CO, July, 1996.
U.S. Environmental Protection Agency, 1995. Add Deposition Standard Feasibility Study Report to
Congress, EPA Report 430-R-95-00 la, Office of Air and Radiation, Acid Rain Division, Washington,
D.C.'
U.S. Environmental Protection Agency, 1996a, Review of the National Ambient Air Quality Standards
for Paniculate Matter: Policy Assessment of Scientific and Technical Information, OAQPS Staff
Paper. EPA-452NR-96-OI3. Office of Air Quality Planning and Standards, Research Triangle Park, NC,
July, 1996.
U.S. Environmental Protection Agency, 1996b, Proposed Methodology for Predicting PM2J from
PMlo Values to Assess (he Impact of Alternative Forms and Levels of the PMNAAQS. Prepared by
Terence Fitz-Simons, David Mintz and Miki Wayland (U.S. Environmental Protection Agency, Office of
Air Quality Planning and Standard, Air Quality Trends Analysis Group), June 26, 1996.
U.S. Environmental Protection Agency, 1997a. Regulatory Impact Analyses for the Paniculate Matter
and Ozone National Ambient Air Quality Standards and Proposed Regional Haze Rule, Office of
Air Quality Planning and Standards, Research Triangle Park, NC, July 1997.
U.S. Environmental Protection Agency, 1997b. "The Effects of SOx and NOX Emission Reductions on
Sulfate and Nitrate Paniculate Concentrations" Thomas Braverman, Air Quality Modeling Group, Office
of Air Quality Planning and Standards. Research Triangle Park, NC. May 1997.
U.S. Environmental Protection Agency, I997c. National Air Pollutant Emission Trends Procedures
Document 1990-1996, Section 4:0: National Criteria Pollutant Estimates, 1985-1996. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. Draft Document. June 1997.
U.S. Environmental Protection Agency, 1997d. Memorandum to the docket from Terence Fitz-Simons
(Office of Air Quality Planning and Standards, Air Quality Trends Analysis Group), Subject: "Response to
Comments Made by AISI on EPA Methodology for Predicting PM2iS from PM10," February 6, 1997.
Page 3-27
-------
U.S. Environmental Protection Agency, 1997e, Deposition of Air Pollutants to the Great Waters, EPA
Report 453/R-97-011, Office of Air Quality Planning and Standards, Research Triangle Park, NC.
U.S. Environmental Protection Agency, 1997f. Memorandum from David Mintz, Air Quality Trends
Analysis Group, Office of Air Quality Planning and Standards to Allyson Siwik Innovative Strategies and
Economics Group, Office of Air Quality Planning and Standards. Subject: "Methodology Used to Create
PMIO and PM25 Air Quality Databases for RIA Work." July 15, 1997.
U.S. Environmental Protection Agency, 1998a. Analyzing Electric Power Generation under the Clean
Air Act. Office of Air and Radiation, Washington, D.C., May 1998.
U.S. Environmental Protection Agency, 1998b, Development of Modeling Inventory and Budgets for
Regional NOx SIP Call. Office of Air Quality Planning and Standards, September 1998
U.S. Environmental Protection Agency, 1998c, Compilation of Air Pollutant Emission Factors, Volume
I: Stationary Point and Area Sources, Fifth Edition (AP-42). Office of Air Quality Planning and
Standards, Research Triangle Park, NC. Supplement D, August 1998.
U.S. Environmental Protection Agency, 1998d, The Regional NOx SIP Call and Reduced Atmospheric
Deposition of Nitrogen: Benefits to Selected Estuaries, Washington, DC, September 22, 1998.
Page 3-28
-------
Chapter 4. BENEFITS OF REGIONAL NOx REDUCTIONS
The changes in ozone and PM ambient concentrations described in Chapter 3 will result in
changes in the physical damages associated with elevated ambient concentrations of these pollutants.
The damages include changes in both human health and welfare effects categories.
This chapter presents the methods used to estimate the physical and monetary benefits of the
modeled NOx and S02 emissions changes from implementing the revised SIPs, the estimates of the
avoided physical damages (e.g., incidence reductions), and the results of the benefits analysis for a range
of regulatory alternatives considered for the SIP call. EPA decided to analyze the benefits of the most
significant alternatives that it considered for determining state NOx budgets for the electric power
industry and other stationary sources. The five alternatives are described in Table 2-3 in Chapter 2 of
Volume 2 of the RIA. In order to conserve analytical resources, the benefits of Regionally 2 are not
analyzed. Regionality 2 achieves emission reductions and air quality improvements that are similar,
though not identical, to Regionality 1 and 0.15 Trading. It is likely that total benefits for Regionality 2
would fall somewhere in between the benefits estimates for Regionality 1 and 0.15 Trading.
The remainder of this chapter is laid out as follows. Section 4.1 provides an overview of the
benefits methodology. Section 4.2 discusses issues in estimating health effects. Sections 4.3 discusses
methods and provides estimated values for avoided incidences and monetary benefits for ozone and PM
related health effects. Section 4.4 discusses methods and provides estimated values for ozone and PM
related welfare effects. Section 4.5 provides estimates of total health and welfare benefits associated
with alternative NOx emission limit policies. Finally, Section 4.6 discusses potential benefit categories that
are not quantified due to data and/or methodological limitations, and provides a list of analytical
uncertainties, limitations, and biases.
4.1 Overview of Benefits Estimation
Most of the specific methods and information used in this benefit analysis are similar to those
used in the §812 Retrospective of the Benefits and Costs of the Clean Air Act and forthcoming §812
Prospective EPA Reports to Congress, which were reviewed by EPA's Science Advisory Board (EPA,
1997c),as well as the approach used by EPA in support of revising the ozone and PM NAAQS in 1997
(EPA,1997aand 1997b).
Prior to describing the details of the approach for the benefits analysis, it is useful to provide an
overview of the approach. The overview is intended to help the reader better identify the role of each
issue described later in this chapter.
The general term "benefits" refers to any and all outcomes of the regulation that are considered
positive; that is, that contribute to an enhanced level of social welfare. The economist's meaning of
"benefits" refers to the dollar value associated with all the expected positive impacts of the regulation;
that is, all regulatory outcomes that lead to higher social welfare. If the benefits are associated with
market goods and services, the monetary value of the benefits is approximated by the sum of the
predicted changes in "consumer (and producer) surplus." These "surplus" measures are standard and
Page 4-1
-------
widely accepted measures in the field of applied welfare economics, and reflect the degree of well being
enjoyed by people given different levels of goods and prices. If the benefits are non-market benefits
(such as the risk reductions associated with environmental quality improvements), however, other methods
of measuring benefits must be used. In contrast to market goods, non-market goods such as
environmental quality improvements are public goods, whose benefits are shared by many people. The
total value of such a good is the sum of the dollar amounts that all those who benefit are willing to pay,
In addition to benefits, regulatory actions may also lead to potential disbenefits, i.e. outcomes that
have a negative impact on social welfare. In general these disbenefits will be incidental to the stated
goals of the regulation, otherwise (in an efficient regulatory environment) the regulation would not have
been promulgated. In order to fully quantify the benefits and costs of a regulatory action, both the
benefits and disbenefits should be calculated, so that net benefits (equal to benefits minus disbenefits
minus costs) will not be biased upwards. In many cases, however, disbenefits are difficult to quantify, as
it is often unclear where and how disbenefits will occur. Benefits may also be difficult to quantify, since
many benefits are not measurable using market based measures,
This conceptual economic foundation raises several relevant issues and potential limitations for
the benefits analysis of the regulation. First, the standard economic approach to estimating environmental
benefits is anthropocentric -- all benefits values arise from how environmental changes are perceived and
valued by people in present-day values. Thus, all near-term as well as temporally distant future physical
outcomes associated with reduced pollutant loadings need to be predicted and then translated into the
framework of present-day human activities and concerns. Second, as noted below, it is not possible to
quantify or to value all of the benefits resulting from environmental quality improvements.
Conducting a benefits analysis for anticipated changes in air emissions is a challenging exercise,
Assessing the benefits of a regulatory action requires a chain of events to be specified and understood.
As shown in Figure 4-1, illustrating the causality for air quality related benefits, the estimation of benefits
requires information about: (1) institutional relationships and policy-making; (2) the technical feasibility of
pollution abatement; (3) the physical-chemical properties of air pollutants and their consequent linkages to
biological or ecological responses in the environment, and (4) human responses and values associated with
these changes.
The first two steps of Figure 4-1 reflect the institutional and technical aspects of implementing the
NOx SIP call regulation (the improved process changes or pollutant abatement). The estimated changes
in ambient PM or ozone concentrations are directly linked to the estimated changes in precursor pollutant
emission reductions through the use of air quality modeling, as described in Chapter 10.
Page 4-2
-------
Figure 4-1 Example Methodology of a Benefits Analysis
Specify SIP call Policy
I
Evaluate Changes in Production Processes
and/or Treatment
Estimate Reductions in
Pollutant Emissions
I
Model Reductions in Ambient Ozone and
PM Air Quality
Estimate Changes in Plant Damage,
Crop Yields, and Other Welfare Effects
1
Estimate Changes in Supply and Value
of Crops, Vegetables, and Other
Welfare Effects
\
Estimate Changes in Adverse Human
Health Symptoms and Risk
I
Estimate Value of Averted Adverse
Human Health Symptoms and Risk
Page 4-3
-------
A "damage function" approach is used to estimate the adverse physical effects from air pollution
that will be avoided in the eastern United States due to implementation of the emission reductions required
by the NOx SIP call (the exception to this is the estimation of nitrogen deposition benefits, which uses an
avoided cost approach). An "economic unit value" approach is used (for most effect categories, e.g.,
premature mortality or chronic bronchitis) to estimate society's aggregate demand (i.e., willingness to
pay) for avoiding each type of physical effect on a per incidence level. Total value for a given physical
effect is simply the product of the number of incidences avoided and the value per incidence avoided. All
dollar estimates of monetary benefits presented in this chapter are in 1990 dollars.
The valuation of avoided incidences of health effects and avoided degradation of welfare effects
relies on benefits transfer. The benefits transfer approach takes values or value functions generated by
previous research and transfers them from the study to the policy of interest. For example, the value of
reduced mortality is obtained from a distribution of values of statistical life based on 26 wage-risk and
contingent valuation studies. None of the values for the health and welfare categories valued in this
benefit analysis were generated specifically in the context of the NOx SIP call.
The first step in a benefits analysis using this approach is the identification of the types or
categories of benefits associated with the anticipated changes in ambient air quality conditions. The
second step is the identification of relevant studies examining the relationships between air quality and
these benefit categories and studies estimating the value of avoiding damages. Table 4-1 provides an
example of the types of benefits potentially observed as a result of changes in air quality. The types of
benefits identified in both the health and welfare categories can generally be classified as use benefits or
non-use benefits.
Use benefits are the values associated with an individual's desire to avoid exposure to an
environmental risk. Use benefits include both direct and indirect uses of affected ambient air, and
embrace both consumptive and non-consumptive activities. In most applications to air pollution scenarios,
the most prominent use benefits are those related to human health risk reductions, effects on crops and
plant life, visibility, and materials damage.
Non-use (intrinsic) benefits are values an individual may have for lowering air pollution
concentrations or the level of risk unrelated to his or her own exposure. Individuals apart from any past,
present, or anticipated future use of the resource in question can value improved environmental quality.
Such non-use values may comprise a significant portion of the total monetary benefits. However, the
dollar amount to assign to these non-use values often is a matter of considerable debate. While human
uses of a resource can be observed directly and valued with a range of technical economic techniques,
non-use values must be ascertained through indirect methods, such as asking survey respondents to reveal
their values.
Page 4-4
-------
Table 4-1
Examples of Potential Benefits of Air Quality Improvements
USE BENEFITS
Direct
Indirect
! Option Value
Aesthetic
NON-USE BENEFITS
Bequest
Existence
EXAMPLES
Human Health Improvements (e.g., less incidences of coughing)
Increased Crop Yields
Non-Consumptive Use (e.g., improved visibility for recreational activities)
Risk Premium for Uncertain Future Demand
Risk Premium for Uncertain Future Supply (e.g., treating as insurance, the protection
of a forest just in case a new use for a forest product will be discovered in
the future)
Residing, working, traveling, and/or owning property in reduced smog locations
intergenerational Equity (e.g., an older generation wanting a younger generation to
inherit a protected environment)
Stewardship/Preservation/Altruistic Values (e.g., individuals wanting to protect a
forest even if they know that they will never use the forest)
Ecological Benefits
Non-use values may be related to the desire that a clean environment be available for the use of
others now and in the future, or may be related to the desire to know that the resource is being preserved
for its own sake, regardless of human use. The component of non-use value that is related to the use of
the resource by others in the future is referred to as the bequest value. This value is typically thought of
as altruistic in nature. For example, the value that an individual places on reducing the general
population's risk of PM and/or ozone exposure either now or in the future is referred to as the bequest
value. Another potential component of non-use value is the value that is related to preservation of the
resource for its own sake, even if there is no human use of the resource. This component of non-use
value is sometimes referred to as existence value. An example of an existence value is the value placed
on protecting the habitats of endangered species from the effects of air pollution, even if the species have
no direct use to humans.
The majority of health and welfare benefits categories included in this analysis can be classified
as direct use benefits. These benefits are discussed in greater detail than other benefits categories
presented in Table 4-1 because more scientific and economic information has been gathered for the direct
use benefits category. Detailed scientific and economic information is not as readily available for the
remainder of the potential benefits categories listed in Table 4-1. Information pertaining to indirect use,
option value, aesthetic, bequest, and existence benefits is often more difficult to collect. For example,
lowering ambient ozone concentrations in an area is expected to reduce physical damage to ornamental
plants in the area. Homeowners living in the affected area with ornamental plants in their yards are
expected to benefit from the reduced damage to their plants, with the plants possibly exhibiting an
improved appearance or experiencing an extended life. Although scientific information can help identify
the benefits category of decreased damage to urban ornamentals, lack of more detailed scientific and
Page 4-5
-------
economic information (e.g., exposure-response relationships for urban ornamentals and values associated
with specific types of injuries and mitigation) currently prevents quantification of this benefits category.
It is also difficult to identify ail the types of benefits that might result from environmental
regulation and to value those benefits that are identified. A cost analysis is expected to provide a more
comprehensive estimate of the cost of an environmental regulation because technical information is
available for identifying the technologies that would be necessary to achieve the desired pollution
reduction. In addition, market or economic information is available for the many components of a cost
analysis (e.g., energy prices, pollution control equipment, etc.). A similar situation typically does not exist
for estimating the benefits of environmental regulation. This problem is due to the non-market nature of
many benefits categories. Since many pollution effects (e.g., adverse health or ecological effects)
traditionally have not been traded as market commodities, economists and analysts cannot look to changes
in market prices and quantities to estimate the value of these effects. This lack of observable markets
may lead to the omission of significant benefits categories from an environmental benefits analysis.
Likewise, difficulties in measuring disbenefits may lead to a positive bias in net benefits. The net result of
underestimating benefits and disbenefits will depend on how completely each category is measured.
Because of the inability to quantify many of the benefits categories listed in Table 4-1, as well as
the omission of unknown but relevant environmental benefits categories, the quantified benefits presented
in this report may underestimate total benefits. It is not possible to quantify the magnitude of this
underestimation. The more important of these omitted effect categories are shown in Table 4-2.
Underestimation of total benefits may be mitigated to some extent if there are also relevant disbenefit
categories that are omitted or unqualified.
Within each effect category, there may be several possible estimates of health and welfare
effects or monetary benefit values. Each of these possibilities represents a health or welfare "endpoint."
The basic structure of the method used to conduct the benefits analysis is to create a set of benefit
estimates reflecting different key assumptions concerning environmental conditions and the
responsiveness of human health and the environment to changes in air quality. Total benefits are
presented as a plausible range representing the sensitivity of benefits over the set of maintained
assumptions. The upper and lower ends of the plausible range of total benefits are constructed using
estimates of non-overlapping endpoints for each effect category, selected to avoid double counting.
Double counting occurs when two endpoints contain values for the same thing. For example, an endpoint
measuring avoided incidences of all hospital admissions would incorporate avoided incidences of hospital
admissions just for heart disease. Thus including values for avoiding both types of hospital admissions
would double count the value of avoided hospital admissions for heart disease. The upper and lower ends
of the plausible range do not necessarily represent the sum of the highest values for each endpoint.
Instead, they represent the points associated with the combinations of assumptions that are expected to
generate the lowest and highest benefit estimates for the majority of regulatory alternatives. The plausible
range does not provide information on the likelihood of any set of assumptions being the correct one.
Thus, while the plausible range indicates the sensitivity of benefits to the various assumptions, it requires a
subjective determination of which assumption set most closely represents reality.
Page 4-6
-------
Table 4-2
Unqualified Benefit Categories*
Unqualified Benefit Categories Associated
with Ozone and Nitrogen Oxides
Unqualified Benefit Categories
Associated with PM
Health
Categories
Airway responsiveness
Pulmonary inflammation
Increased susceptibility to respiratory
infection
Acute inflammation and respiratory cell
damage
Chronic respiratory damage/Premature aging
of lungs
Uitraviolet-B radiation (disbenefit)
Changes in pulmonary function
Morphological changes
Altered host defense mechanisms
Cancel-
Other chronic respiratory disease
Welfare
Categories
Ecosystem and vegetation effects in Class I
areas (e.g., national parks)
Damage to urban ornamentals (e.g.,grass,
flowers, shrubs, and trees in urban
areas)
Fruit and vegetable crops
Reduced yields of tree seedlings, commercial
and non-commercial forests
Damage to ecosystems
Materials damage (other than consumer
cleaning cost savings)
Nitrates in drinking water
Brown Clouds
Materials damage (other than consumer
cleaning cost savings)
Damage to ecosystems (e.g., acid sulfate
deposition)
Nitrates in drinking water
Brown Clouds
* Note that there are other pollutants that are reduced in conjunction with strategies implemented to reduce NOx emissions for the
SIP call. These include carbon {a pollutant associated with global climate change) and mercury (a toxic pollutant). These emission
reductions are also not considered in this benefits analysis.
There are many subjective judgements that must be made in order to select the set of
relationships and values for the benefits analysis. The specific selections used to develop the plausible
range are designed to reflect the EPA's best current judgement on each issue, considering the state of
current scientific knowledge, previous Agency analyses, and the most recent advice provided by EPA's
Science Advisory Board on performing benefits analysis for criteria air pollution control programs. There
are, however, defensible alternatives to virtually every decision about the makeup of the plausible range.
In order better to inform the reader of important alternative assumptions that could have been made, and
to provide an understanding of the impact of each alternative on the overall assessment of the monetary
benefits, the benefits analysis includes a number of quantitative sensitivity analyses. Individual sensitivity
analyses examine the effects of using alternative assumptions about individual choices incorporated in the
benefits analysis, such as the impact of using short-term (daily) mortality functions instead of a long-term
(chronic exposure) function.
Sensitivity analyses are also used to explore the impacts of including other endpoints, such as PM-
related infant mortality, that are not as well understood as the effects included in the benefits analysis.
Page 4-7
-------
Sensitivity analyses will also be used to explore the effects of alternative valuation approaches, such as
the use of alternative agricultural market simulation models.
A very important component in estimating PM-related health and welfare benefits is the
characterization of air quality changes. Several models developed in recent years are capable of
estimating PMIO concentrations, but have not been rigorously tested for estimating ambient concentrations
of PM25 because there is currently sparse monitoring of the data necessary to benchmark model
performance. As indicated in Chapter 3, two air quality models, RADM-RPM and the Source-Receptor
(S-R) Matrix, are used in this analysis to predict changes in ambient PM levels given changes in NOx and
SO2 emissions. The defining characteristics of each of these models are laid out in Chapter 3. It is not
clear which of these models will better predict the Eastern U.S. atmosphere in the 2007 policy year. In
order to reflect this uncertainty, the plausible range for individual PM-related health and welfare endpoints
will incorporate estimates of avoided incidences and monetary benefits generated under each modeling
framework.
Because of the nonlinear chemistry used by RADM-RPM to predict changes in PM2j, estimates
of avoided incidences and associated monetary benefits may not follow a predictable pattern across
regulatory alternatives. Previous experience with PM air quality models would suggest that benefits will
increase as controls become more stringent, however, these models lacked or used incomplete
characterizations of the role of NOx emissions in the production of oxidant fields that convert SO2 to acid
sulfates, leading to a simplified characterization of the interactions between acid sulfates, nitrates, and
ammonium. The role of NOx emissions in photochemistry can introduce non-linearities which, given the
right set of atmospheric conditions, can result in situations where decreases in NOx can lead to increases
in PM in some regions'. This can lead to smaller benefits for a more stringent regulatory alternative
relative to a less stringent regulatory alternative (Dennis, 1998). This seems to be occurring in the 0.15
trading alternative, especially in the Northwest, Upper Midwest and Upper New England regions of the
OTAG domain. In addition, implementation of specific control strategies, such as shifting of power
generation and emissions trading under the Acid Rain Trading program can result in increases in SO2
emissions in states outside the NOx SIP call region. In the 0.15 and 0.12 trading alternatives, significant
shifts in power generation seem to be occuring between SIP call states and Gulf Coast states, leading to
increases in both NOx and SO2 emissions along the Gulf Coast, relative to the analysis baseline.
Increases in PM in the Northeast and northwestern regions of the SIP call region seem to be caused by a
combination of atmospheric chemistry and emissions trading, as well as transport of pollutants, especially
in Pennsylvania and the upper New England area.
Because the modeled distributions of PM concentrations are non-normally distributed, the
ordering effect is dependent on whether health effects are calculated using the median versus mean
PM2S concentration. The non-linearities in the air-chemistry change with movement from the lower tail to
the upper tail of the distribution of PM concentrations. Non-linearities are more pronounced at the 50th
percentile of the distribution than at the 90th percentile (Dennis, 1998), This can result in a greater
degree of non-linearity in benefits results that are dependent on the median versus the mean. Implications
of this for the SIP call benefits analysis are that relative to other endpoints, estimates of PM-related long-
1 See Chapter 3 Sections 3.3 and 3.4 for a more detailed discussion of the air quality models. See Appendix
E for a more thorough discussion of the affect of non-linear chemistry on particle formation.
Page 4-8
-------
term mortalities, which are based on median PM2.5 concentrations, are more sensitive to the non-linear
chemistry effects between alternatives.
Throughout this benefits analysis, sensitivity analyses for assumptions affecting only a single
endpoint and with no expected directional effect will be presented directly following the plausible range.
These sensitivity analyses include short-term PM-related mortality, PM-related neo-natal mortality,
alternative agricultural models, and an analysis of the effect of using only ozone mortality studies with a
significant ozone coefficient to generate avoided ozone mortality incidences.
Table 4-3 lists the specific health and welfare effects that are included in the benefits analysis,
indicating the specific effect categories that are included in the plausible range of benefits, as well as
effects that are presented (or explored in greater detail) as quantified sensitivity analyses. Also included
in Table 4-3 are the estimates of mean Willingness to Pay (WTP), or "unit values" used to monetize the
benefits for each endpoint.
Page 4-9
-------
Table 4-3
Quantified and Monetized Health and Welfare Effects
Endpoint
Health Effects in the Benefits Analysis
Mortality, Long-term Exposure - Over age 30
Mortality, Short-term Exposure
Chronic Bronchitis - AH Ages
Hospital Admissions - All Respiratory, All Ages
Hospital Admissions - Congestive heart failure
Hospital Admissions - Ischemic heart disease
Any of 1 9 Acute Respiratory Symptoms -Adult
Acute Bronchitis - Children
Lower Respiratory Symptoms - Children
Upper Respiratory Symptoms - Children
Work Loss Days - Adult
Minor Restricted Activity Days (MRAD) - Adult
Welfare Effects in the Benefits Analysis
Agriculture - Select Commodity Crops
Household Soiling
Nitrogen Deposition in Estuarine and Coastal Waters
Decreased Worker Productivity
Visibility - Residential
Visibility - Select Class I areas
Pollutant
PM2.5
Ozone
PM,o
Ozone &
PMi0/PM15
PMIO
PM10
Ozone
PMIO/PM2.S
PM,0
PMIO
PM2J
PMj.5
Ozone
PMlfl
NOx
Ozone
Light
Extinction'1
Light
Extinction"
Mean WTP per incident
($1990)
$4,800,000
$4,800,000
$260,000
$6,7 12 (Ozone)
$6,344 (PM)
$8,280
$10,308
$18
$45
$12
$19
$83
€i$
4 Jo
Direct Valuation
$2.S2/household/|ig/m;i
change in PMIO
$105/kg of nitrogen
$! /worker/ 10% change in Ozone
$ 14/household/deciview
$4/household/deciview
Page 4-10
-------
Endpoint
Pollutant
Mean WTP per incident
($1990)
Endpoints Presented as Sensitivity Analyses
Mortality, Short-term Exp, - Only significant studies
Mortality, Short-term Exp. - Over age 65
Post-Neonatal Mortality
Commercial Crops - AGSIM model
Ozone
PM2,5
PMIO
NOx
$4,800,000
$4,800,000
quantified but not monetized
$1 OS/kg
" Measured in terms of cteciview change.
4.2 Issues in Estimating Changes in Health Effects
This benefits analysis relies on concentration-response (C-R) functions estimated in published
epidemiological studies relating adverse health and welfare effects to ambient air quality. The specific C-
R functions used are included in Table 4-4.
When a single published study is selected as the basis of the C-R relationship between a pollutant
and a given health endpoint, applying the C-R function is straightforward. This is the case for most of the
endpoints selected for inclusion in the benefits analysis. A single C-R function may be chosen over other
potential functions because the underlying epidemiological study used superior methods, data or
techniques, or because the C-R function is more generalized and comprehensive. For example, the study
that estimated the effects of PM on hospital admissions for all ages and all respiratory diseases is selected
over studies limited to the over 65 year old population or specific categories of respiratory diseases.
The exceptions to the "single study" selection in the benefits analysis are mortality associated
with exposure to ozone and chronic bronchitis associated with exposure to PM, which have multiple
studies selected for use. Mortality associated with short-term exposure to PM25, presented as a
sensitivity analysis, is based on six C-R functions estimated in a single study. When several estimated C-
R relationships between a pollutant and a given health endpoint have been selected, they are combined or
pooled to derive a single estimate of the relationship. The details of the procedures used to combine
multiple C-R functions are presented in a separate technical support document (Abt Associates, 1998a).
Whether the concentration-response relationship between a pollutant and a given health endpoint
is estimated by a single function from a single study or by a pooled function of concentration-response
functions from several studies, that same concentration-response relationship is applied everywhere in the
benefits analysis. Although the concentration-response relationship may in fact vary somewhat from one
location to another (for example, due to differences in population susceptibilities or differences in the
composition of PM), location-specific concentration-response functions are generally not available. While
a single function applied everywhere may result in overestimates of incidence changes in some locations
and underestimates of incidence changes in other locations, these location-specific biases will to some
extent cancel each other out when the total incidence change is calculated. It is not possible to know the
extent or direction of the bias in the total incidence change based on application of a single C-R function
everywhere.
Page 4-II
-------
The remainder of this section discusses two key issues involving the use of C-R functions to
estimate the benefits of the NOx SIP call: baseline incidences and health effect thresholds, i.e. levels of
pollution below which changes in air quality have no impacts on health.
4.2.1 Baseline Incidences
The epidemiological studies of the association between pollution levels and adverse health effects
generally provide a direct estimate of the relationship of air quality changes to the relative risk of a health
effect, rather than an estimate of the absolute number of avoided cases. For example, a typical result
might be that a 10 \ig/m3 decrease in daily PM2 s levels might decrease hospital admissions by three
percent". The baseline incidence of the health effect is necessary to convert this relative change into a
number of eases.
United States county-level baseline mortality rates for 1990 were obtained from the National
Center for Health Statistics (US Department of Health and Human Services, 1994). Because most PM
and ozone studies that estimate C-R functions for mortality considered only non-accidental mortality,
county-specific baseline mortality rates used in the estimation of PM- and ozone-related mortality were
adjusted to provide a better estimate of county-specific non-accidental mortality. Each county-specific
mortality rate was multiplied by the ratio of national non-accidental mortality to national total mortality
(0.93).
Although total mortality incidences (over all ages) are available for counties, age-specific
mortality incidences is not generally available at the county level. Therefore, county-specific baseline
mortality incidences among individuals aged 30 and over (necessary for PM2.5-related long-term exposure
mortality, estimated by Pope et al., 1995) are estimated by applying national age-specific death rates to
county-specific age distributions, and adjusting the resulting estimated age-specific incidences so that the
estimated total incidences (including all ages) equals the actual county-specific total incidences.
Page 4-12
-------
Table 4-4
PM and Ozone Health and Welfare Concentration-Response Function Summary Data
Endpoint
Pollutant
Concentration-Response Function
Source
Functional
Form
Averaging Time
Studied
Applied
Population4
Pollutant
Coefficient1'
Mortality
Mortality (long-term
exposure) -PM2 5
Mortality (short-term
exposure) -PM, s
Mortality (short-term
exposure)
PM23
PM2.5
Ozone
Pope et at, 1995
Schwartz et al.,
I996a(6cities)
Kirmey, at al.,
1995 (Los
Angeles)
Ito and
Thurston, 1996
(Chicago)
Moolgavkar, et
al., 1995
(Philadelphia)
Samet,et al.,
1997
(Philadelphia)
log-linear
log-linear
log-linear
log-linear
log-linear
log-linear
annual median
2-day average
daily 1-hour
max
I -day average
1 -day average
1-day average
annual median5
1 -day average
daily 1-hour
max
1 -day average
1 -day average
1 -day average
ages 30+
all
all
all
all
all
0.006408
0,000782
0.000000
0.000677
0.000611
0.000936
Page 4-13
-------
Endpoint
Hospital Admissions
All respiratory illnesses
Congestive heart failure
Ischemic heart disease
All respiratory illnesses
Pollutant
PM2S/PM,0
PMIO
PMIO
Ozone
Concentration-Response Function
Source
Functional
Form
Averaging Time
Studied
Applied
Population4
Pollutant
Coefficient15
Thurston et a!.,
1994
Schwartz &
Morris, 1995
Schwartz &
Morris, J995
Thurston et al.,
1992
linear
log-linear
log-linear
linear
1-day average
2-day average
1 -day average
daily 1-hour
max
1 -day average
1 -day average
1 -day average
daily I -hour
max
all
age 65+
age 65+
all
3.45 X 10'8
0.00098
0.00056
0,00137
Respiratory Symptoms/Illnesses not requiring hospitalization
Development of chronic
bronchitis
Acute bronchitis
Upper respiratory
symptoms (URS)
Lower respiratory
symptoms (LRS)
Any of 1 9 acute
respiratory symptoms
MRAD
Work loss days (WLDs)
PM10
PM15/PMio
PM10
PM10
Ozone
PM,,5
PM2.5
Schwartz, 1993
Dockery et al.,
1989
Pope etal., 1991
Schwartz et at.,
1994
Krupnick et a!.,
1990
Ostro and
Rothschild,
1989
Ostro, 1987
logistic
log-iinear
logistic
logistic
log-linear
log-linear
annual mean
annual mean
1 -day average
1-day average
daily 3 -hour
max
2-week average
2-week average
annual mean
annual mean"5
1-day average
1-day average
daily 1-hour
max
1-day average
1 -day average
all
ages 10-12
asthmatics, ages
9-11
ages 8- 12
ages 18-65
ages 18-65
ages 18-65
0.012
0.0298
0.0036
0.01823
0.00014
0.00741
0.0046
Page 4-14
-------
Endpoint
Pollutant
Concentration-Response Function
Source
Functional
Form
Averaging Time
Studied
Applied
Welfare Endpoints
Decreased worker
productivity
Household soiling and
damage
Ozone
PM!0
Crocker & Horst,
1981 and EPA,
1994
ESEERCO,
1994
percent change
linear
1-day average
annual mean
1-day average
annual mean
Population5
laborers
all households
Pollutant
Coefficient
n/a
2.52 (dollars per
jig/ m3 PM10 per
household)
The population examined in the study and to which this analysis applies the reported concentration-response relationship. In general, epidemiologies! studies analyzed the
concentration-response relationship for a specific age group (e.g., ages 65+) fa a specific geographical area. This analysis applies the reported pollutant coefficient to all individuals in
the age group nationwide.
b A single pollutant coefficient reported for several studies indicates a pooled analysis; see text for discussion of pooling concentration-response relationships across studies.
c All 1-day averages are 24-hour averages, 2-day averages are 48-hour averages, etc.
Paae4-15
-------
4.2.2 Thresholds
A very important issue in applied modeling of changes in ozone and PM is whether to apply the
concentration-response functions to all predicted changes in ambient concentrations, even small changes
occurring at levels approaching "anthropogenic background". Different assumptions about how to model
thresholds can have a major effect on the resulting benefits estimates.
The Criteria Documents, Staff Papers and the Federal Register Notices promulgating the new
ozone and PM standards conclude that there are no known threshold levels for any of the health effects
to be included in the SIP call analysis (EPA, 1996c-1996f). For example, the Federal Register Notice
promulgating the ozone standard included the following;
The Administrator's consideration of an appropriate level for an 8-hour standard to protect public
health with an adequate margin of safety necessarily reflects a recognition, as emphasized by
CASAC, that it is likely that "03 may elicit a continuum of biological responses down to
background concentrations" (Wolff, 1995b). Thus, in the absence of any discernible threshold, it
is not possible to select a level below which absolutely no effects are likely to occur. Nor does it
seem possible, in the Administrator's judgment, to identify a level at which it can be concluded
with confidence that no "adverse" effects are likely to occur (Federal Register, Vol. 62, No. 138, p.
38863, July 18, 1997).
Similarly, the Administrator did not identify a PM threshold for the same reasons.
The underlying epidemiological functions used in most of the analysis are in fact continuous down
to zero levels. However, in order to remain consistent with the available scientific information., the
analysis does not model effects below certain levels. The approach used in the benefits analysis is
identical to the approach used in the ozone and PM NAAQS RIA; individual concentration-response
functions will not be applied to ambient concentrations occurring below the lowest observed levels
reported by the authors of the underlying epidemiological studies. Where no lowest observed level was
reported, the functions will be applied down to the "anthropogenic background" level. Theoretically, C-R
functions should be reestimated when a threshold is assumed to insure consistency with the observed
correlation between mortality incidences and the pollutant. If no threshold is assumed in the
epidemiological study, then the slope of the C-R function will be flatter than for a function with a
threshold. This reflects the fact that all of the observed changes in mortality would have to be associated
with changes above the threshold, rather than being associated with changes along the full spectrum of
pollutant concentrations. Unadjusted C-R functions are used in this benefits analysis due to a lack of
availability of the underlying data used to estimate the C-R functions. These data are necessary to
develop threshold adjusted C-R functions. Use of an unadjusted C-R function will result in an
underestimate of total avoided incidences.
Because the issue of possible thresholds can have a major effect on the benefits estimation,
estimates for individual benefit endpoints will be generated using alternative assumptions of thresholds for
PM. Following advice from EPA's Science Advisory Board, both high and low threshold assumptions will
be used to generate benefits estimates. The low threshold assumption will assume a threshold equal to
anthropogenic background concentrations and the high threshold assumption will assume a threshold equal
to the PM standard of 15 u,g/m3.
Page 4-16
-------
4,3 Ozone and PM Health-Related Benefits
While a broad range of adverse health effects have been associated with exposure to elevated
ozone and PM levels, only subsets of health effects are selected for inclusion in the quantified benefit
analysis. Effects are excluded from the current analysis (1) in order to prevent double counting (such as
hospital admissions for specific respiratory diseases); (2) due to uncertainties in applying effect
relationships based on clinical studies (where human subjects are exposed to various levels of air pollution
in a carefully controlled and monitored laboratory situation) to the NOx SIP call affected population; or
(3) due to a lack of an established concentration-response relationship.
The general format for the following sections detailing benefits for each endpoint is to begin with
a discussion of the method and studies used for economic valuation, then present the studies used to
obtain the concentration-response function for estimation of avoided incidences. Following these
discussions, tables of avoided incidences and associated monetary benefits for ozone-related effects and
PM-related effects are presented. Benefits estimates are presented for a subset of the regulatory
alternatives presented in chapters 6, 7 and 9, Air quality changes used to generate the benefits estimates
are not based on the final NOx SIP call control requirements. For additional information on air quality
modeling scenarios for the benefits analysis, see Section 10.1,5. Numbers presented in the tables
represent changes in the number of incidences and associated monetary benefits given the illustrative
implementation of particular NOx control strategies relative to the 2007 baseline air quality. For endpoints
which are affected by both ozone and PM, ozone-related benefits are presented first, followed by PM-
related benefits.
A preliminary explanatory note on the calculation of the point estimates presented in the tables
below is warranted. Each point estimate of monetary benefits in the tables below is the mean of a
distribution of monetary benefits derived through a Monte Carlo procedure2. The estimate derived by this
method approaches the simple product of the mean of the unit dollar distribution and the mean of the
incidence change distribution, but for a finite number of iterations may be slightly off. For an illustrative
example of the procedure and for further details, see Appendix A and the technical support document for
this RIA (Abt Associates, 1998a).
For ozone, three health effects are selected for inclusion: mortality associated with short-term
exposure, hospital admissions for all respiratory diseases, and acute respiratory symptoms. One other
human health-related effect, decreased worker productivity, is included as a welfare effect rather than a
health effect (see Section 4.3.4). The ozone-related effect categories that are included in the NOx SIP
call analysis are shown in Table 4-5. Premature mortality is the only ozone-related endpoint for which a
range of benefits is presented.
2 Each point estimate of avoided incidences presented in the tables below is the mean of a Latin Hypercube
approximation of a distribution of avoided incidences reflecting the uncertainty in the pollutant coefficient in the C-R function.
In the Latin Hypercube method 100 perccntile points (in this ease, the (n-O.S)th percentile points of the distribution, for n = 1,2,
..,, 100) are selected to represent the distribution. This reduces the computational burden associated with preserving the full
distribution.
Page 4-17
-------
Table 4-5
Quantified Ozone-Related Health Effects Included in the Benefits Analysis
Health Effect
Affected Population
Study
Mortality
Ozone-related short-term exposure
mortality
all ages
pooled analysis of 4 U.S. studies
Hospital Admissions
"All Respiratory"
ail ages
Thurstonetal., 1992
Respiratory Symptoms/Illnesses Not Requiring Hospitalizafion
Acute respiratory symptoms (any of 1 9)
ages 18-65
Krupnick et al., 1990
Although the primary environmental purpose of the NOx SIP call is to help achieve attainment of
the ozone NAAQS in the eastern United States, significant monetary benefits will also be associated with
changes in ambient levels of PM. Several PM health endpoints are included in the quantified benefits
estimation. The PM-related effect categories that are included in this analysis are shown in Table 4-6,
For all of the PM-related endpoints, benefits are estimated using both the RADM-RPM and S-R Matrix
generated PM concentrations. In addition, for health endpoints, benefits are estimated under both a
background threshold assumption and an assumed threshold of 15 fxg/m3.
4.3.1 Premature Mortality
Both ozone and particulate matter have been associated with increased risk of premature
mortality in adult populations. Avoided mortality is a very important health endpoint in this economic
analysis due to the high monetary value associated with risks to life.
The benefits analysis uses the "statistical lives lost" approach to value avoided premature
mortality. The mean value of avoiding one statistical death is estimated to be $4.8 million. This
represents an intermediate value from a variety of estimates that appear in the economics literature, and is
a value that EPA has frequently used in RIAs for other rules. This estimate is the mean of a fitted
Weibull distribution of the estimates from 26 value-of-iife studies identified in the §812 study as
"applicable to policy analysis." The approach and set of selected studies mirrors that of Viscusi (1992)
(with the addition of two studies), and uses the same criteria used by Viscusi in his review of value-of-Iife
studies. The $4.8 million estimate is consistent with Viscusi's conclusion that "most of the reasonable
estimates of the value of life are clustered in the $3 to $7 million range." Five of the 26 studies are
contingent valuation (CV) studies, which directly solicit WTP information from subjects; the rest are
wage-risk studies, which base WTP estimates on estimates of the additional compensation demanded in
the labor market for riskier jobs. The 26 studies used to form the distribution of the value of a statistical
life are listed in Table 4-7.
Page 4-18
-------
Table 4-6
Quantified PM-Reiated Health Effects Included in the Benefits Analysis
Endpoint
Population to Which Applied | Study
Mortality
PM, ,-related long-term exposure
mortality
ages 30+
Popeetal., 1995
Hospital Admissions
"all respiratory"
Congestive heart failure
Ischemic heart disease
all ages
age 65+
age 65+
Thurston et al., 1994
Schwartz and Morris, 1995
Schwartz and Morris, 1 995
Chronic Bronchitis
Development of chronic bronchitis
all
Schwartz, 1993
Respiratory Symptoms/Illnesses Not Requiring Hospitalization
Acute bronchitis
'Mjj-related lower respiratory
symptoms (LRS)
Upper respiratory symptoms (URS)
MRADs
Work loss days (WLDs)
ages 10-12
ages 8-12
asthmatics, age 9-1 1
ages 18-65
ages 18-65
Dockeryet al,, 1989
Schwartz etal,, 1994
Popeetal., 1991
Ostro and Rothschild, 1989
Ostro, 1987
There are two types of exposure to elevated levels of air polluation that may result in premature
mortality. Acute (short-term) exposure (e.g., exposure on a given day) to peak pollutant concentrations
may result in excess mortality on the same day or within a few days of the elevated exposure. Chronic
(long-term) exposure (e.g., exposure over a period of a year or more) to levels of pollution that are
generally higher may result in mortality in excess of what it would be if pollution levels were generally
lower. The excess mortality that occurs will not necessarily be associated with any particular episode of
elevated air pollution levels. Both types of effects are biologically plausible, and there is an increasing
body of consistent corroborating evidence from animal toxicity studies indicating that both types of effects
exist.
Page 4-19
-------
Table 4-7
Summary of Mortality Valuation Estimates*
Study
Kneisner and Leeth ( 1 991 ) (U.S.)
Smith and Gilbert (1984)
Dillingham(I985)
Butler (1983)
Miller and Guria ( 1991)
Moore and Viscusi (1988a)
Viscusi, Magat, and Huber (1991b)
Oegaxetal. (1985)
Marin and Psacharopoulos (1982)
Kneisner and Leeth (1991) (Australia)
Gerking, de Haan, and Schulze (1988)
Cousineau, Lacroix, and Oirard (1 988)
Jones-Lee (1989)
Dillingham(l985)
Viscusi (1978, 1979)
R.S Smith (1976)
V.K. Smith (1976)
Olson (1981)
Viscusi (1981)
R.S. Smith (1974)
Vloore and Viscusi (1988a)
Kneisner and Leeth (1991) (Japan)
Herzog and Schlottman (1987)
Leigh and Poison ( 1984)
Leigh (1987)
Galen (1988)
Type of Estimate
Labor Market
Labor Market
Labor Market
Labor Market
Contingent Valuation
Labor Market
Contingent Valuation
Contingent Valuation
Labor Market
Labor Market
Contingent Valuation
Labor Market
Contingent Valuation
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Valuation per Statistical
Life (millions of 1990 S)
0.6
0,7
0.9
1.1
1.2
2,5
2.7
3.3
2,8
3.3
3.4
3,6
3,8
3.9
4.1
4.6
4.7
5.2
6.5
7.2
7.3
7,6
9,1
9.7
10.4
13.5
'Source: Viscusi, 1992
Page 4-20
-------
There are, similarly, two basic types of epidemiological studies of the relationship between
mortality and exposure to pollutants. Long-term studies (e.g., Pope et al,, 1995) estimate the association
between long-term (chronic) exposure to air pollution and the survival of members of a large study
population over an extended period of time. Such studies examine the health endpoint of concern in
relation to the general long-term level of the pollutant of concern — for example, relating annual mortality
to some measure of annual pollutant level Daily peak concentrations would impact the results only
insofar as they affect the measure of long-term (e.g., annual) pollutant concentration. In contrast, short-
term studies relate daily levels of the pollutant to daily mortality. By their basic design, daily studies can
detect acute effects but cannot detect the effects of long-term exposures. A chronic exposure study
design (a prospective cohort study, such as the Pope study) is best able to identify the long-term exposure
effects, and will likely detect some of the short-term exposure effects as well. Because a long-term
exposure study may detect some of the same short-term exposure effects detected by short-term studies,
including both types of study in a benefit analysis would likely result in some degree of double counting of
benefits.
Another major advantage of the long-term study design concerns the issue of the degree of
prematurity of mortality associated with air pollution. It is possible that the short-term studies are
detecting an association between air pollution and mortality that is primarily occurring among terminally ill
people. Critics of the use of short-term studies for policy analysis purposes correctly point out that an
added risk factor that results in a terminally ill person dying a few days or weeks earlier than they
otherwise would have (known as "short-term harvesting") is potentially included in the measured air
pollutant mortality "signal" detected in such a study. As the short-term study design does not examine
individual people (it examines daily mortality rates in large populations, typically a large city population ), it
is impossible to know anything about the overall health status of the specific population that is detected as
dying early. While some of the detected excess deaths may have resulted in a substantial loss of life
(measuring loss of life in terms of lost years of remaining life), others may have lost a relatively short
amount of lifespan.
While the long-term study design is preferred, these types of studies are expensive to conduct and
consequently there are relatively few well designed long-term studies. For PM, there has only been one
high quality study accepted by the Science Advisory Board, and for ozone, no acceptable long-term
studies have been published. For this reason, short-term ozone mortality is used as the basis for
determining ozone-related mortality benefits for the NOx SIP call.
The next two sections provide details on the measurement and valuation of changes in incidences
of premature mortality associated with changes in ozone and PM arising from implementation of the NOx
SIP call.
Ozone-related Mortality
The literature on the possible relationship between exposure to ambient ozone and premature
mortality has been evolving rapidly. Of the 28 time-series epidemiology studies identified in the literature
that report results on a possible association between daily ozone concentrations and daily mortality (see
EPA, 1997a, Appendix J), 21 were published or presented since 1995. In particular, a series of studies
published in 1995 through 1997 (after closure on the ozone Criteria Document) from multiple cities in
western Europe has significantly increased the body of studies finding a positive association. Fifteen of
Page 4-21
-------
the 28 studies report a statistically significant relationship between ozone and mortality, with the more
recent studies tending to find statistical significance more often than the earlier studies. The ozone-
mortality datasets have also tended to become larger in more recent studies as longer series of air quality
monitoring data have become available over time. This suggests that it may take many years of data
before the ozone effect can be separated from the daily weather and seasonal patterns with which it
tends to be correlated.
In 1997, as a part of the ozone NAAQS promulgation RIA, EPA staff reviewed this recent
literature. They identified 9 studies that met a defined set of selection criteria, and conducted a meta-
analysis of the results of the 9 studies. The result of this work was included as Appendix J in the
NAAQS RIA, "Assessment and Synthesis of Available Epidemiological Evidence of Mortality Associated
with Ambient Ozone from Daily Time-series Analyses" (EPA, 1997a).
The NOx SIP call related benefits analysis implements the same basic meta-analysis approach to
quantifying ozone mortality as the NAAQS RIA, with the exception that a subset of 4 of the 9 studies is
used, representing only U.S. based analyses. In a post-NAAQS RIA review of the methodology for
assessing ozone mortality effects, it was determined that the relationships between ambient ozone and
mortality in the non-U.S. study locations included in the original NAAQS-related meta-analysis may not
be representative of the range of ozone-mortality concentration-response relationships in the United
States. Although ozone is the same everywhere (in contrast to PM), its effects on mortality may depend
on its interactions with other pollutants and with meteorological variables. In addition, there are population
and societal differences (air conditioning incidence, building construction, human activity patterns, etc.)
across locations that could affect the relationship between ambient ozone levels and mortality. To reduce
the potential for applying inappropriate concentration-response functions in analysis of the ozone mortality
benefits from the NOx SIP call, only U.S. studies are included, based on the assumption that demographic
and environmental conditions on average would be more similar between the study and policy sites.
However, the full body of peer-reviewed ozone mortality studies should be considered when evaluating
the weight of evidence regarding the presence of an association between ambient ozone concentrations
and premature mortality.
Because of differences in the averaging times used in the underlying studies (some use daily
average ozone levels, while others use 1-hour daily maximum values), it is not possible to conduct a
meaningful meta-analysis directly on the coefficients of the C-R functions. Instead, for each pair of air
quality modeling results (for the baseline and a given regulatory alternative) for the NOx SIP call, each C-
R function is translated into a set of predicted mortality incidence changes that would be estimated by that
C-R function, given the set of air quality changes. The meta-analysis approach is then applied to the
predicted mortality incidence changes that would be estimated by each of the studies. Additional details
of the approach are described in the technical support document for the NOx SIP call (Abt Associates,
1998a).
Table 4-8 presents the range of estimates of avoided incidences of ozone-related mortality and
monetary benefits associated with five regulatory alternatives for the NOx SIP call. Note that the lower
estimate for this endpoint is zero to reflect both the number of peer-reviewed studies finding no significant
relationship between ozone and premature mortality and the lack of a directly established biological
mechanism linking ozone and premature mortality. In its review of the epidemiological ozone-mortality
literature, EPA has determined that there is a reasonable probability that increased ozone concentrations
Page 4-22
-------
are associated with incidences of premature mortality. In Table 4-8 the higher estimate allows for the
existence of an ozone-mortality relationship, but assumes there is some probability that for any specific
location within the SIP call region that the effect of ozone on premature mortality is zero. This probability
is embedded in the previously discussed meta-analysis approach, which includes studies both with and
without findings of a statistically significant relationship between ozone concentrations and premature
mortality.
Note that the high estimate presented in Table 4-8 does not represent the limiting upper end of
potential ozone-related mortality benefits. The high estimate in Table 4-8 represents the mean estimate of
benefits derived from the distribution of concentration-response functions generated from the
meta-analysis described above. The limiting upper end for ozone-related mortality is generated by
assuming that the highest concentration-response function applies everywhere. This would generate a
benefits estimate considerably higher than the high estimate in Table 4-8.
Table 4-8
Range of Avoided Ozone-related Mortality Incidences and Monetary Benefits
Associated with the NOx SIP Calf
Regulatory Alternative
0.1 2 Trading
0,1 5 Trading
Regionality 1
0.20 Trading
0,25 Trading
Avoided Incidences
(cases/year)
Low
0
0
0
0
0
High
315
279
251
234
174
Monetary Benefits
(millions 1990$)
Low
$0
$0
$0
$0
so
High
$1,496
$1,326
$1,191
$1,108
$824
"Annual baseline incidence for non-accidental deaths in the general population for ail ages is 803/100,000. Total annual baseline
incidence for the NOx SIP call region is 1,768,014 non-accidental deaths.
PM-rdated Mortality
PM-associated mortality in the benefits analysis is estimated using the PM2 5 relationship from
Pope et al., 1995. This decision reflects the Science Advisory Board's explicit recommendation for
modeling the mortality effects of PM in both the completed §812 Retrospective Report to Congress and
the ongoing §812 Prospective Study. The Pope study estimates the association between long-term
(chronic) exposure to PM2S and the survival of members of a large study population. This relationship is
selected for use in the benefits analysis instead of short-term (daily pollution) studies for a number of
reasons.
The Pope long-term study is selected as providing the best available estimate of the relationship
between PM and mortality. It is used alone, rather than considering the total effect to be the sum of
Page 4-23
-------
estimated short-term and long-term effects, because summing creates the possibility of double-counting a
portion of total mortality. The Pope study is selected in preference to other available long-term studies
because it uses the best methods (i.e., a prospective cohort method with a Cox proportional hazard
model), and has a much larger cohort population, the longest exposure interval, and more locations (51
cities) in the United States, than other studies. It is unlikely that the Pope study contains any significant
amount of short-term harvesting. First, the health status of each individual tracked in the study is known
at the beginning of the study period. Persons with known pre-existing serious illnesses were excluded
from the study population. Second, the Cox proportional hazard statistical model used in the Pope study
examines the question of survivability throughout the study period (10 years). Deaths that are premature
by only a few days or weeks within the 10-year study period (for example, the deaths of terminally ill
patients, triggered by a short duration PM episode) are likely to have little impact on the calculation of the
average probability of surviving the entire 10 year interval. In relation to the other prospective cohort
study (Dockery, et al,, 1992., the "Six-cities" cohort study), the Pope study found a smaller increase in
excess mortality for a given PM air quality change.
Table 4-9 presents point estimates of avoided incidences of long-term PM-related mortality and
monetary benefits associated with the five regulatory alternatives for the NOx SIP call. As noted earlier,
non-linearities inherent in the RADM-RPM air quality model lead to an inconsistent ranking of results
between the RADM-RPM and S-R Matrix results, With the exception of the 0.12 trading alternative,
estimated premature mortality incidences are higher for S-R Matrix generated PM changes than for
RADM-RPM generated PM changes.
Table 4-9
Avoided Long Term PM-rclated Mortality Incidences and Monetary Benefits
Associated with the NOx SIP calf
Regulatory
Alternative
0.1 2 Trading
0.15 Trading
Regional ity 1
0.20 Trading
0.25 Trading
Avoided Incidences (cases/year)
RADM-RPM
15 ng/m3
310
53
67
78
44
Back
ground
657
101
94
149
75
S-R Matrix
15 ng/m3
306
231
190
216
202
Back
ground
561
370
278
315
294
Monetary Benefits (millions 1990$)
RADM-RPM
IS ng/m3
$1,468
$251
$317
$370
$208
Back
ground
$3,173
$482
$459
$715
$358
S-R Matrix
15 fig/ra3
$1,459
$1,099
$904
$1,028
$962
Back
ground
$2,672
$1,763
$1,326
$1,499
$1,400
" Annual baseline incidence for non-accidental deaths in the general population aged over 30 is 759/100,000. Total annual
baseline incidence for the NOx SIP call region is 929,557 non-accidental deaths.
The estimates of excess mortality from the short-term studies are presented as an important
sensitivity analysis. Because there is only one short-term study (presenting results from 6 separate U.S.
Page 4-24
-------
cities) that uses PM15 as the metric of PM (Schwartz et al., 1996), an estimate based on the pooled city-
specific, short-term PM2.5 results will be presented.
Table 4-10 presents the results of a sensitivity analysis using mortality associated with short-term
exposure to PM25. In some cases, the avoided incidences of mortality (and corresponding monetary
benefits) predicted using the short-term function are higher than those predicted using the long-term
function, and in other cases the reverse is true. For the RADM-RPM background threshold results, the
magnitude of the difference between the value of avoided incidences of short- and long-term mortality
ranges from $-1,539 million for the 0.12 trading alternative to $207 million for the Regionality 1 alternative.
For the S-R Matrix background threshold results, the magnitude of the difference ranges from $-1,703
million for the 0.12 trading alternative to $-709 million for the Regionality 1 alternative. As with long-
term mortality, the relationship between the RADM-RPM and S-R Matrix generated results is not
consistent across alternatives, due to the differences in air chemistry modeling between the two models.
In addition, the rank ordering across threshold levels is not consistent across alternatives.
Table 4-10
Sensitivity Analysis; Premature Mortality Benefits
Using Avoided Short Term PM-related Mortality Incidences"
Regulatory
Alternative
0.1 2 Trading
0,1 5 Trading
Regionality 1
0.20 Trading
0.25 Trading
Avoided Incidences (cases/year)
RADM-RPM
ISng/m1
293
130
129
113
76
Back
ground
339
128
136
129
91
S-R Matrix
IS (ig/m3
136
110
99
103
99
Back
ground
203
152
129
138
132
Monetary Benefits (millions 1990S)
RADM-RPM
15 ng/m3
$1,393
$619
$614
$536
$360
Back
ground
$1,634
$615
$666
$619
$434
S-R Matrix
15 [ig/m3
$649
$523
$473
$493
$473
Back
ground
$969
$726
$617
$658
$630
• Annual baseline incidence for non-accidental deaths in the general population is 803/100,000. Total annual baseline incidence
for the NOx SIP call region is 1,768,014 non-accidental deaths for the population aged over 30.
A new study (Woodruff et al, 1997) finds a significant association between annual PM,0 levels
and post-neonatal (infants aged 28-51 weeks) mortality. Conceptually any additional mortality from this
function would be additive to the Pope results (because the Pope function covers only the population over
30 years old), although not additive to the daily mortality studies (which cover all ages). The SAB
recently advised the §812 Prospective project to not include this in the §812 primary analysis at this time,
primarily because the study is of a new endpoint and the results have not been replicated in other studies
in the U.S. The coherence and consistency arguments which support the use of the Pope study are not
present with this study at this time. For the SIP call analysis, this endpoint is presented as a sensitivity
analysis. PM25 changes associated with the NOx SIP call are used with this PM10 C-R function. This
will produce a conservative estimate of infant mortality for two reasons. First, there may be some
Page 4-25
-------
reductions in the coarse fraction (PM between 2,5 and 10 microns in diameter) that result from the NOx
reductions which will be omitted from the analysis. Perhaps more importantly, estimating infant mortality
using the estimated change in PM25 levels in a PM]0 function implicitly assumes that the fine fraction of
PM is no more toxic than the coarse fraction. EPA's decision in 1997 to set an additional NAAQS using
PM2 5 in addition to a PMIO standard, is based in part on a growing scientific consensus that the fine
fraction of the total PM,0 mass is likely to be most associated with adverse health effects. If in fact the
toxicity of PM25 is greater than the toxicity of PM10> then using changes in PM23 in a C-R function based
on PM|0 will underestimate the total effect on infant mortality.
Table 4-11 presents a sensitivity analysis using neo-natal mortality. Monetary benefits associated
with the avoided incidences are not presented due to a lack of information about the value of avoided neo-
natal mortality. It is likely that avoided infant mortalities will be valued higher than mortalities for adults.
However, at present, no studies have been conducted to determine this value. For this reason, only
avoided incidences of neo-natal mortality are presented in Table 4-11.
Table 4-11
Sensitivity Analysis: Avoided Post Neo-natal PM-related Mortality Incidences
Regulatory Alternative
1
! 0.1 2 Trading
0. 1 5 Trading
Regionally 1
0.20 Trading
0,25 Trading
Avoided Incidences (cases/year)
RADM-RPM
15 jig/m3
5
2
2
2
I
Back
ground
5
2
2
2
1
S-R Matrix
15 ng/m3
2
2
1
1
1
Back
ground
2
2
1
1
1
4.3,2 Hospital Admissions
An individual's WTP to avoid a hospital admission will include, at a minimum, the amount of
money they pay for medical expenses (i.e., what they pay towards the hospital charge and the associated
physician charge) and the loss in earnings. In addition, however, an individual is likely to be willing to pay
some amount to avoid the pain and suffering associated with the illness itself. That is, even if they
incurred no medical expenses and no loss in earnings, most individuals would still be willing to pay
something to avoid the illness.
Because medical expenditures are to a significant extent shared by society, via medical insurance,
Medicare, etc., the medical expenditures actually incurred by the individual are likely to be less than the
total medical cost to society. The total value to society of an individual's avoidance of hospital admission,
then, might be thought of as having two components: (1) the cost of illness (CO1) to society, including the
Page 4-26
-------
total medical costs plus the value of the lost productivity, as well as (2) the individual's WTP to avoid the
disutility of the illness itself (e.g., the pain and suffering associated with the illness).
In the absence of estimates of social WTP to avoid hospital admissions for specific illnesses
(components 1 plus 2 above), estimates of total COI (component 1) are typically used as conservative
(lower bound) estimates. Because these estimates do not include the value of avoiding the disutility of the
illness itself (component 2), they are biased downward. Some analyses adjust COI estimates upward by
multiplying by an estimate of the ratio of WTP to COI, to better approximate total WTP. Other analyses
have avoided making this adjustment because of the possibility of over adjusting - that is, possibly
replacing a known downward bias with an upward bias. The previous RIAs for PM and ozone, as well as
the revised RIA for ozone and PM NAAQS, did adjust the COI estimate upward, The COI values used
in the benefits analysis for the SIP call benefits will not be adjusted to better reflect the total WTP. This
is consistent with the guidance offered by the §812 SAB committee.
The COI estimates used in this RIA include the estimated hospital and physician charges, based
on the average length of a hospital stay for the illness, and the estimated opportunity cost of time spent in
the hospital.
Ozone-related Hospital Admissions
The benefits analysis includes a single ozone-related effect category for hospital admissions; all
respiratory diseases. The study that estimated the C-R function (Thurston et al,, 1992) examined hospital
admissions for all ages in the population. Because of the comprehensiveness of the Thurston study, it is
selected over other available studies that are restricted to limited age ranges (e.g., the population aged 65
year and older), and/or specific diagnoses (e.g., hospital admissions for pneumonia). The age- and
disease-specific effect categories are subsets of the all-age, all-respiratory disease hospital admission
category. Therefore, the benefits of avoided hospital admissions for respiratory illnesses for all ages
should be larger than the benefits for more restricted categories. However, that is not true for the
estimated benefits, based on the available studies. The Thurston estimated relationship produces fewer
benefits than either of the two available alternatives: all respiratory disease admissions for the population
over 65; or the sum of pneumonia and chronic obstructive pulmonary disease (COPD) admissions for the
population over 65. Clearly adding the results for these study types would involve a serious amount of
double counting. Therefore, selecting the Thurston study may underestimate the total benefits of hospital
admissions.
Table 4-12 presents point estimates of avoided incidences of hospital admissions for all ozone-
related respiratory symptoms and monetary benefits associated with the five regulatory alternatives for
the NOx SIP call.
Page 4-27
-------
Table 4-12
Avoided Ozone-related Hospital Admissions and Monetary Benefits
Associated with tiie NOx SIP Call"
Regulatory Alternative
0,1 2 Trading
0, ! 5 Trading
Regional ity 1
0,20 Trading
0.25 Trading
Avoided Incidences
(cases/year)
719
637
571
533
396
Monetary Benefits
(millions 1990$)
$5
$4
$4
$4
$3
" Annual baseline incidence for all respiratory-related hospital admissions (not just ozone related) in the general population is
504/100,000. Total annual baseline incidence for the NOx SIP call region is 1,109,687 admissions.
PM-related Hospital A dmissions
The benefits analysis includes three PM-related hospital admissions, due to all respiratory
illnesses, congestive heart failure, and ischemic heart disease. As with ozone-induced hospital admissions,
the benefits analysis relies on a study of all respiratory hospital admissions for all age groups, rather than
studies examining the population over 65. Table 4-13 presents point estimates of avoided incidences of
PM-related hospital admissions and monetary benefits associated with the five regulatory alternatives for
the NOx SIP call.
Page 4-28
-------
Table 4-13
Avoided PM-related Hospital Admissions and Monetary Benefits
Associated with the NOx SIP Call"
Regulatory
Alternative
0.1 2 Trading ARb
CHF
! IHD
0,1 5 Trading AR
CHF
IHD
Regionally AR
1 CHF
IHD
0,20 Trading AR
CHF
IHD
0.25 Trading AR
CHF
IHD
Avoided Incidences (cases/year)
RADM-RPM
IS |ig/m3
305
6!
67
135
23
25
133
24
27
117
22
24
78
14
15
Back
ground
354
63
70
133
24
27
139
26
28
133
24
27
94
17
19
S-R Matrix
15 jAg/m3
518
26
29
491
17
19
479
13
14
484
14
16
480
13
15
Back
ground
589
28
31
535
18
20
511
14
15
520
15
17
515
14
16
Monetary Benefits (millions 1990$)
RADM-RPM
15 ng/m3
$1.9
$0.5
$0,7
$0.9
$0.2
$0,3
$0.8
$0.2
$0.3
$0.7
$0.2
$0,2
$0.5
$0.1
$0.2
Back
ground
$2.2
$0.5
$0.7
S0.8
S0.2
S0.3
$0,9
$0.2
$0.3
$0.8
$0.2
$0.3
$0.6
$0.1
$0.2
S-R Matrix
ISng/m3
$3.3
$0.2
$0.3
$3.1
$0.1
$0,2
$3.0
$0.1
$0.1
$3.1
$0.1
$0.2
$3,0
$0.1
$0,2
Back
ground
$3,7
$0.2
$0.3
$3.4
$0.1
$0.2
$3.2
$0.1
$0.2
$3.3
$0.1
$0.2
$3.3
$0.1
$0.2
"Annual baseline incidence in the general population is 504/100,000 for all respiratory, 231/100,000 for congestive heart failure,
and 450/100,000 for isehemic heart disease. Total annual baseline incidence for the NOx SIP call region is 1,109,687 admissions
for ail respiratory, 508,607 admissions for congestive heart failure, and 990,792 admissions for isehemic heart disease.
b AR refers to all respiratory, CHF refers to congestive heart failure, and IHD refers to isehemic heart disease.
4,3,3 Bronchitis
Chronic bronchitis is the only measured morbidity endpoint that may be expected to last from the
initial onset of the illness throughout the rest of the individual's life. WTP to avoid chronic bronchitis
would therefore be expected to incorporate the present discounted value of a potentially long stream of
costs (e.g., medical expenditures and lost earnings) and pain and suffering associated with the illness.
Two studies, Viscusi et al, (1991) and Krupnick and Cropper (1992) provide estimates of WTP to avoid a
case of chronic bronchitis. The study by Viscusi et al., however, uses a sample that is larger and more
representative of the general population than the study by Krupnick and Cropper (which selects people
who have a relative with the disease). The valuation of chronic bronchitis in this analysis is therefore
based on the distribution of WTP responses from Viscusi et al. (1991),
Both Viscusi et al. (1991) and Krupnick and Cropper (1992), however, defined a case of severe
chronic bronchitis. It is unclear what proportion of the cases of chronic bronchitis predicted to be
Page 4-29
-------
associated with exposure to pollution would turn out to be severe cases. The estimated incidence of
pollution-related chronic bronchitis related to the SIP call emission reductions is based on two studies
(Abbey et al., 1993 and Schwartz, 1993), which consider only new cases of the illness. While a new case
may not start out being severe, chronic bronchitis is an illness that may progress in severity from onset
throughout the rest of the individual's life. It is the chronic illness that is being valued, rather than the
illness at onset.
The WTP to avoid a case of pollution-related chronic bronchitis (CB) is derived by starting with
the WTP to avoid a severe case of chronic bronchitis, as described by Viseusi et al, (1991), and adjusting
it downward to reflect (1) the decrease in severity of a case of pollution-related CB relative to the severe
case described in the Viseusi study, and (2) the elasticity of WTP with respect to severity reported in the
Krupnick and Cropper study. The adjustment procedure is described in more detail in the technical
support document (Abt Associates, 1998a), The mean value of the adjusted distribution is $260,000.
Estimating WTP to avoid a case of acute bronchitis is difficult for several reasons. First, WTP to
avoid acute bronchitis itself has not been estimated. Estimation of WTP to avoid this health endpoint
therefore must be based on estimates of WTP to avoid symptoms that occur with this illness. Second, a
case of acute bronchitis may last more than one day, whereas it is a day of avoided symptoms that is
typically valued. Finally, the concentration-response function used in the benefit analysis for acute
bronchitis was estimated for children, whereas WTP estimates for those symptoms associated with acute
bronchitis were obtained from adults.
With these caveats in mind, a rough estimate of WTP to avoid a case of acute bronchitis was
derived as the midpoint of a low and a high estimate. The low estimate ($13.29) is the sum of the
midrange values recommended by IBc (JEc, 1994) for two symptoms believed to be associated with acute
bronchitis: coughing ($6.29) and chest tightness ($7,00), The high estimate was taken to be twice the
value of a minor respiratory restricted activity day ($38,37)s or $76.74. The midpoint between the low
and high estimates is $45.00. This value was used as the point estimate of MWTP to avoid a case of
acute bronchitis in the benefit analysis.
Table 4-14 presents point estimates of avoided incidences of PM-related chronic and acute
bronchitis and monetary benefits associated with the five regulatory alternatives for the NOx SJP call.
Page 4-30
-------
Table 4-14
Avoided Incidences of PM-related Chronic and Acute Bronchitis and Monetary Benefits
Associated with the NOx SIP Call
Regulatory
Alternative
0,1 2 Trading
Chronic
Acute
0.1 5 Trading
Chronic
Acute
I Regionally 1
Chronic
Acute
0.20 Trading
Chronic
Acute
0.25 Trading
Chronic
Acute
Avoided Incidences (eases/year)
RADM-RPM
15 |ig/m3
2,060
1,148
787
535
825
517
785
472
550
352
Back
ground
2,054
2,171
782
827
823
859
784
809
548
574
S-R Matrix
15 (ig/m3
891
424
587
313
449
257
499
294
469
276
Back
ground
909
917
595
609
455
453
500
519
471
490
Monetary Benefits (millions 1990$)
RADM-RPM
15 [Ag/m3
$589
$0.1
$225
$0.0
$236
$0.0
$216
SO.O
$148
$0.0
Back
ground
$574
$0.1
$213
SO.O
$223
$0.0
$221
$0.0
$150
$0.0
S-R Matrix
1 5 Hg/m3
$240
$0.0
$158
$0.0
$121
$0.0
$134
$0.0
$126
$0.0
Back
ground
$245
$0.0
$160
$0.0
$122
$0.0
$135
$0.0
$127
$0,0
4.3,4 Acute Respiratory Symptoms
There are three sources of uncertainty in the valuation of upper or lower respiratory symptoms:
(1) an occurrence of URS or of LRS may be comprised of one or more of a variety of symptoms (i.e.,
URS and LRS are each potentially a "complex of symptoms"), so that what is being valued may vary
from one occurrence to another; (2) for a given symptom, there is uncertainty about the mean WTP to
avoid the symptom; and (3) the WTP to avoid an occurrence of multiple symptoms may be greater or less
than the sum of the WTPs to avoid the individual symptoms.
Ozone-Related Respiratory Symptoms
"Presence of any of 19 acute respiratory symptoms" is a somewhat subjective "health endpoint"
used by Krupnick et al. (1990), Moreover, not all 19 symptoms are listed in the Krupnick study. It is
Page 4-31
-------
therefore not clear exactly what symptoms were included in the study. Even if all 19 symptoms were
known, it is unlikely that WTP estimates could be obtained for all of the symptoms. Finally, even if all 19
symptoms were known and WTP estimates could be obtained for all 19 symptoms, the assumption of
additivity of WTPs becomes tenuous with such a large number of symptoms. The likelihood that all 19
symptoms would occur simultaneously, moreover, is very small.
Acute respiratory symptoms must be either upper respiratory symptoms or lower respiratory
symptoms. In the absence of further knowledge about which of the two types of symptoms is more likely
to occur among the "any of 19 acute respiratory symptoms," it was assumed that they occur with equal
probability. Because this health endpoint may also consist of combinations of symptoms, it was also
assumed that there is some (smaller) probability that upper and lower respiratory symptoms occur
together.
To value avoidance of a day of "the presence of any of 19 acute respiratory symptoms" it was
therefore assumed that this health endpoint consists either of URS, or LRS, or both. It was also assumed
that it is as likely to be URS as LRS and that it is half as likely to be both together. That is, it was
assumed that "the presence of any of 19 acute respiratory symptoms" is a day of URS with 40%
probability, a day of LRS with 40% probability, and a day of both URS and LRS with 20% probability.
Using the point estimates of WTP to avoid a day of URS and LRS derived above, the point estimate of
WTP to avoid a day of "the presence of any of 19 acute respiratory symptoms" is
(0.40XS18.70) + (0.40X$11.82) + (0.20XS18.70 + $11,82) = $18.31
(1)
Because this health endpoint is only vaguely defined, and because of the lack in information on the
relative frequencies of the different combinations of acute respiratory symptoms that might qualify as
"any of 19 acute respiratory symptoms," the unit dollar value derived for this health endpoint must be
considered only a rough approximation.
Table 4-15 presents point estimates of avoided incidences of ozone-related respiratory symptoms
and monetary benefits associated with the five regulatory alternatives for the NOx SIP call.
Table 4-15
Avoided Incidences of Ozone-related Respiratory Symptoms and Monetary Benefits
Associated with the NOx SIP Call
Regulatory Alternative
0.1 2 Trading
0.15 Trading
Regional ity I
0,20 Trading
0.25 Trading
Avoided Incidences
(cases/year)
68,919
61,015
54,757
51,053
37,908
Monetary Benefits
(millions 1990S)
$1.3
$1.2
$1,0
$1,0
$0,7
Page 4-32
-------
PM Related Upper Respiratory Symptoms
The concentration-response function for URS is taken from Pope et ai. (1991). Pope et al,
describe URS as consisting of one or more of the following symptoms: runny or stuffy nose; wet cough;
and burning, aching, or red eyes. The children in the Pope study were asked to record respiratory
symptoms in a daily diary, and the daily occurrences of URS and LRS, as defined above, were related to
daily PM-10 concentrations. Estimates of WTP to avoid a day of symptoms are therefore appropriate
measures of benefit.
Willingness to pay to avoid a day of URS is based on symptom-specific WTPs to avoid those
symptoms identified by Pope et al. as part of the URS complex of symptoms. Three contingent valuation
(CV) studies have estimated WTP to avoid various morbidity symptoms that are either within the URS
symptom complex defined by Pope et al. or are similar to those symptoms identified by Pope et al. In
each CV study, participants were asked their WTP to avoid a day of each of several symptoms. The
three individual symptoms that were identified as most closely matching those listed by Pope et al. for
URS are cough, head/sinus congestion, and eye irritation. A day of URS could consist of any one of
seven possible "symptom complexes" consisting of at least one of these symptoms. It is assumed that
each of the seven types of URS is equally likely. The ex ante MWTP to avoid a day of URS is therefore
the average of the MWTPs to avoid each type of URS, or $18.70. This is the point estimate for the dollar
value for URS used in the benefit analysis. Finally, it is worth emphasizing that what is being valued here
is URS as defined by Pope et al,, 1991, While other definitions of URS are certainly possible, this
definition of URS is used in this benefit analysis because it is the incidence of this specific definition of
URS that has been related to PM exposure by Pope et al., 1991.
PM Related Lower Respiratory Symptoms
Schwartz et al. (1994) estimated the relationship between LRS and PM-10 concentrations. The
method for deriving a point estimate of MWTP to avoid a day of LRS is the same as for URS. Schwartz
et al. (1994) define LRS as at least two of the following symptoms; cough, chest pain, phlegm, and
wheeze. The symptoms for which WTP estimates are available that reasonably match those listed by
Schwartz et al. for LRS are cough (C), chest tightness (CT), coughing up phlegm (CP), and wheeze (W).
A day of LRS, as defined by Schwartz et al., could consist of any one of the 11 combinations of at least
two of these four symptoms.
It is assumed that each of the eleven types of LRS is equally likely. The ex ante MWTP to avoid
a day of LRS as defined by Schwartz is therefore the average of the MWTPs to avoid each type of LRS,
or $11.82. This is the point estimate used in the benefit analysis for the dollar value for LRS as defined by
Schwartz et al. The WTP estimates are based on studies which considered the value of a day of avoided
symptoms, whereas the Schwartz study used as its measure a case of LRS. Because a case of LRS
usually lasts at least one day, and often more, WTP to avoid a day of LRS should be a conservative
estimate of WTP to avoid a case of LRS.
Finally, as with URS, it is worth emphasizing that what is being valued here is LRS as defined by
Schwartz et al, 1994, While other definitions of LRS are certainly possible, this definition of LRS is
Page 4-33
-------
used in this benefit analysis because it is the incidence of this specific definition of LRS that has been
related to PM exposure by Schwartz et al., 1994.
The point estimates derived for MWTP to avoid a day of URS and a case of LRS are based on
the assumption that WTPs are additive. For example, if WTP to avoid a day of cough is $7.00, and WTP
to avoid a day of shortness of breath is $5.00, then WTP to avoid a day of both cough and shortness of
breath is $12.00. If there are no synergistic effects among symptoms, then it is likely that the marginal
utility of avoiding symptoms decreases with the number of symptoms being avoided. If this is the case,
adding WTPs would tend to overestimate WTP for avoidance of multiple symptoms. However, there
may be synergistic effects — that is, the discomfort from two or more simultaneous symptoms may
exceed the sum of the discomforts associated with each of the individual symptoms. If this is the case,
adding WTPs would tend to underestimate WTP for avoidance of multiple symptoms. It is also possible
that people may experience additional symptoms for which WTPs are not available, again leading to an
underestimate of the correct WTP. However, for small numbers of symptoms, the assumption of
additivity of WTPs is unlikely to result in substantive bias.
Table 4-16 presents point estimates of avoided incidences of PM-related upper and lower
respiratory symptoms and monetary benefits associated with the five regulatory alternatives for the NOx
SIP call. Note that the magnitude of incidences and the magnitude of monetary benefits are very
different. This is due to the small value per avoided incidence for upper and lower respiratory symptoms.
4,3.5 Worker Productivity
The valuation used to monetize benefits associated with increased worker productivity resulting
from improved ozone air quality is based on information reported in Crocker and Horst, 1981 and
summarized in EPA, 1994, Crocker and Horst (1981) examined the impacts of ozone exposure on the
productivity of outdoor citrus workers. Productivity impacts were measured as the change in income
associated with a change in ozone exposure, given as the elasticity of income with respect to ozone
concentration (-0.1427). The reported elasticity, which is used as the central estimate in this analysis,
translates a 10 percent reduction in ozone to a 1.4 percent increase in income. Given the average daily
income for outdoor workers engaged in strenuous activity reported by the 1990 U.S. census, $73 per day,
the 10 percent reduction in ozone yields approximately $1 in increased daily wages.
Table 4-17 presents estimates of monetary benefits arising from ozone-related avoided lost
worker productivity associated with the five regulatory alternatives for the NOx SIP call.
Page 4-34
-------
Table 4-16
Avoided Incidences of PM-related Upper and Lower Respiratory Symptoms
and Monetary Benefits Associated with the NOx SIP Call"
Regulatory
Alternative
0,1 2 Trading
URSb
LRS
0,15 Trading
URS
LRS
Regionality 1
URS
LRS
0,20 Trading
URS
LRS
0.25 Trading
URS
LRS
Avoided Incidences (cases/year)
RADM-RPM
15 ng/m3
1,639
16,051
623
7,162
642
6,853
585
6,190
392
4,107
Back
ground
1,683
18,675
655
7,151
667
7,313
650
7,057
462
4,934
S-R Matrix
15 ng/m3
682
4,604
459
3,208
343
2,567
389
2,858
372
2,677
Back
ground
748
8,328
494
5,572
368
4,189
414
4,771
396
4,521
Monetary Benefits (millions 1990S)
RADM-RPM
ISfAg/m3
$0.0
$0,2
$0.0
$0.1
$0.0
$0,1
$0.0
$0,1
$0.0
$0.0
Back
ground
$0.0
$0.2
$0.0
$0.1
$0.0
$0,1
$0,0
$0,1
$0.0
$0,1
S-R Matrix
IS (ig/m3
$0,0
$0.1
$0.0
$0.0
$0.0
$0,0
$0,0
$0,0
$0.0
$0.0
Back
ground
$0.0
$0.1
$0.0
$0.1
$0,0
$0,0
$0,0
$0.1
$0.0
$0.1
1 Annual baseline incidence in the applied population (asthmatics, ages 9-1 1) is 38,187 for upper respiratory symptoms.
Information on baseline incidence is not available for lower respiratory symptoms.
b URS=upper respiratory symptom, LRS=iower respiratory symptom.
Page 4-35
-------
Table 4-17
Monetary Benefits from Ozone-related Avoided Lost Worker Productivity
Associated with the NOx SIP Call
Regulatory
Alternative
O.I 2 Trading
0.1 5 Trading
' Regional ity 1
0,20 Trading
0.25 Trading
Monetary Benefits
(millions 1990$)
$25
$22
$20
$19
$14
4,3.6 Work loss days
Willingness to pay to avoid the loss of one day of work was estimated by dividing the median
weekly wage for 1990 (U.S. Department of Commerce, 1992) by 5 (to get the median daily wage). This
values the loss of a day of work at the median wage for the day lost. Valuing the loss of a day's work at
the wages lost is consistent with economic theory, which assumes that an individual is paid exactly the
value of his labor,
The use of the median rather than the mean, however, requires some comment. If ail individuals
in society were equally likely to be affected by air pollution to the extent that they lose a day of work
because of it, then the appropriate measure of the value of a work loss day would be the mean daily
wage. It is highly likely, however, that the loss of work days due to pollution exposure does not occur
with equal probability among all individuals, but instead is more likely to occur among lower income
individuals than among high income individuals. It is probable, for example, that individuals who are
vulnerable enough to the negative effects of air pollution to lose a day of work as a result of exposure
tend to be those with generally poorer health care. Individuals with poorer health care have, on average,
lower incomes. To estimate the average lost wages of individuals who lose a day of work because of
exposure to PM pollution, then, would require a weighted average of all daily wages, with higher weights
on the low end of the wage scale and lower weights on the high end of the wage scale. Because the
appropriate weights are not known, however, the median wage was used rather than the mean wage.
The median is more likely to approximate the correct value than the mean because means are highly
susceptible to the influence of large values in the tail of a distribution (in this case, the small percentage of
very large incomes in the United States), whereas the median is not susceptible to these large values,
The median daily wage in 1990 was $83.00.
Table 4-18 presents point estimates of avoided PM-related work loss days and monetary benefits
associated with the five regulatory alternatives for the NOx SIP call.
Page 4-36
-------
Table 4-18
Avoided PM-related Work Loss Days and Monetary Benefits Associated with the NOx SIP Calf
Regulatory
Alternative
0.12 Trading
0,1 5 Trading
Regionally 1
0.20 Trading
0.25 Trading
Avoided Incidences (cases/year)
RADM-RPM
IS ng/m3
167,124
74,345
73,703
(54,608
43,136
Back
ground
194,481
73,325
77,953
74,176
52,333
S-R Matrix
15ng/m3
53,315
38,262
32,149
34,696
32,391
Back
ground
91,557
62,300
49,258
54,163
50,801
Monetary Benefits (millions 1990$)
RADM-RPM
ISng/m3
$14
$6
$6
$5
S4
Back
ground
$16
$6
$6
$6
$4
S-R Matrix
lS>ig/mJ
$4
$3
$3
$3
$3
Back
ground
$8
$5
$4
$4
$4
1 Annual baseline incidence is 150,750 days/year per 100,000 workers between the ages of 18 and 65.
4.3.7 Minor restricted activity days
No studies are reported to have estimated WTP to avoid a minor restricted activity day (MRAD),
However, lEe (1993) has derived an estimate of WTP to avoid a minor respiratory restricted activity day
(MRRAD), using WTP estimates from Toiley et al. (1986) for avoiding a three symptom combination of
coughing, throat congestion, and sinusitis. This estimate of WTP to avoid a MRRAD, so defined, is
$38.37. Although Ostro and Rothschild (1989) estimated the relationship between PM-2,5 and MRADs,
rather than MRRADs (a component of MRADs), it is likely that most of the MRADs associated with
exposure to PM-2.5 are in fact MRRADs. For the purpose of valuing this health endpoint, then, it is
assumed that MRADs associated with PM exposure may be more specifically defined as MRRADs, and
the estimate of MWTP to avoid a MRRAD is used.
Any estimate of MWTP to avoid a MRRAD (or any other type of restricted activity day other
than WLD) will be somewhat arbitrary because the endpoint itself is not precisely defined. Many
different combinations of symptoms could presumably result in some minor or less minor restriction in
activity. It has been argued (Krupnick and Kopp, 1988) that mild symptoms will not be sufficient to result
in a MRRAD, so that WTP to avoid a MRRAD should exceed WTP to avoid any single mild symptom.
A single severe symptom or a combination of symptoms could, however, be- sufficient to restrict activity.
Therefore WTP to avoid a MRRAD should, these authors argue, not necessarily exceed WTP to avoid a
single severe symptom or a combination of symptoms. The "severity" of a symptom, however, is similarly
not precisely defined; moreover, one level of severity of a symptom could induce restriction of activity for
one individual while not doing so for another. The same is true for any particular combination of
symptoms.
Given that there is inherently a substantial degree of arbitrariness in any point estimate of WTP to
avoid a MRRAD (or other kinds of restricted activity days), the reasonable bounds on such an estimate
are considered. By definition, a MRRAD does not result in loss of work. WTP to avoid a MRRAD
Page 4-37
-------
should therefore be less than WTP to avoid a WLD, At the other extreme, WTP to avoid a MRRAD
should exceed WTP to avoid a single mild symptom. The highest lEc midrange estimate of WTP to avoid
a single symptom is $15.72, for eye irritation. The point estimate of WTP to avoid a WLD in the benefit
analysis is $83. If all the single symptoms evaluated by the studies listed in Exhibit 4.5 are not severe,
then the estimate of WTP to avoid a MRRAD should be somewhere between $15.72 and $83,00.
Because the lEc estimate of $38.37 falls within this range (and acknowledging the degree of arbitrariness
associated with any estimate within this range), the IBc estimate is used as the point estimate of MWTP
to avoid a MRRAD.
Table 4-19 presents point estimates of avoided incidences of PM-related Minor Restricted
Activity Days and monetary benefits associated with the five regulatory alternatives for the NOx SIP call.
Table 4-19
Avoided PM-related Minor Restricted Activity Days and Monetary Benefits
Associated with the NOx SIP Call"
Regulatory
Alternative
0,1 2 Trad ing
0.15 Trading
Regionality 1
0.20 Trading
0.25 Trading
Avoided Incidences (cases/year)
RADM-RPM
IS (ig/m3
1,394,423
620,193
614,892
537,468
358,285
Back
ground
1,621,039
611,497
648,058
617,173
434,783
S-R Matrix
IS^ig/m3
423,788
298,208
247,195
268,453
249,224
Back
ground
742,977
498,840
389,994
430,935
402,875
Monetary Benefits (millions 1990$)
RADM-RPM
15 ng/rn*
$53
$24
$24
$21
$14
Back
ground
$62
$24
$25
$24
$17
S-R Matrix
15 |ig/m3
$16
$11
$10
$10
$10
Back
ground
$29
$19
$15
$17
$16
" Annual baseline incidence is 780,000 days/year per 100,000 population between the ages of 18 and 65. Total annual baseline
incidence for the NOx SIP call region is 1,060,597,322 days/year.
4,4
Ozone- and PM-reiatcd Welfare Effects
In addition to the effects on human health described above, reducing NOx emissions in the
eastern United States will also have welfare (i.e., non-health) effects. Welfare effects cover a
potentially broad range of adverse effects, including adverse impacts on plants, animals, structural
materials, visibility, and ecosystem functions. Like health effects, in order to be included in a quantified
monetary benefits analysis, all of the analytical links between changes in emissions and the monetary
value of the effects must be available. While the required analytical components are available for certain
welfare endpoints, many other likely or possible welfare categories are omitted from the analysis. The
availability of information on each analytical step limits the total coverage of the welfare effects, All of
the welfare benefits that are quantified and included in the benefits analysis were included in the PM and
ozone NAAQS RIA. However, there have been some changes in the quantification of certain welfare
Page 4-38
-------
effects, which are described in this section. Table 4-20 lists the welfare categories that are included in
the benefits analysis.
The welfare categories included in the SIP call analysis that use the identical procedures
previously used are described in the technical support document for this RIA (Abt Associates, 1998a).
The remainder of this section describes aspects of the welfare analysis that are different than the ozone
and PM NAAQS RIA.
Table 4-20
Quantified Ozone- and PM- Related Welfare Effects
Included in the Benefits Analysis
Welfare Effect [ Pollutant
Agriculture - Commodity Crops
Nitrogen Deposition in Estuarine and
Coastal Waters
Decreased Worker Productivity
Visibility-Class 1 Areas (SE only)
Visibility-Residential
Household Soiling
Ozone
NOx
Ozone
PM
PM
PM
Study
Mathtech, 1998
EPA, 1998
Crocker and Horst, 3 98 1 and
EPA, 1994
Chestnut etal., 1997
McClelland etal., 1991
ESEERCO, 1994
4.4.1 Commodity Agricultural Crops
The economic value associated with varying levels of yield loss for ozone-sensitive commodity
crops is analyzed using a revised and updated Regional Model Farm (RMF) agricultural benefits model
(Mathtech, 1998a). The RMF is an agricultural benefits model for commodity crops that account for
about 75 percent of all U.S. sales of agricultural crops. The RMF explicitly incorporates exposure-
response functions into microeconomic models of agricultural producer behavior. The model uses the
theory of applied welfare economics to value changes in ambient ozone concentrations brought about by
particular policy actions such as the NOx SIP call.
The measure of benefits calculated by the model is the net change in consumers' and producers'
surplus from baseline ozone concentrations to the ozone concentrations resulting from attainment of
alternative standards. Using the baseline and post-control equilibria, the model calculates the change in
net consumers' and producers' surplus on a crop-by-crop basis3. Dollar values are aggregated across
crops for each standard. The total dollar value represents a measure of the change in social welfare
3 Agricultural benefits differ from other health and welfare endpoints in the length of the assumed ozone
season. For agriculture, the ozone season is assumed to extend from April to September. This assumption is made
to ensure proper calculation of the ozone statistic used in the exposure-response functions. The only crop affected
by changes in ozone during April is winter wheat.
Page 4-39
-------
associated with the regulatory alternative. Although the model calculates benefits under three alternative
welfare measures (perfect competition, price supports, and modified agricultural policy), results presented
here are based on the "perfect competition" measure to reflect recent changes in agricultural subsidy
programs. Under the recently revised 1996 Farm Bill, most eligible farmers have enrolled in the program
to phase out government crop price supports for the RMF-relevant crops: wheat, corn, sorghum, and
cotton.
For the purpose of this analysis, the six most economically significant crops are analyzed: corn,
cotton, peanuts, sorghum, soybean, and winter wheat. In the 37-state region modeled in this analysis,
these crops were valued at over $70 billion in 1997. The model employs biological exposure-response
information derived from controlled experiments conducted by the National Crop Loss Assessment
Network (NCLAN) (Lee et al., 1996). Four main areas of the RMF have been updated to reflect the
1996 Farm Bill and USDA data projections to 2005 (the year farthest into the future for which projections
are available). These four areas are yield per acre, acres harvested, production costs, and model farms.
Documentation outlining the 2005 update is provided in EPA, 1997a.
Table 4-21 presents estimates of monetary benefits due to changes in the production of all six
commodity crops associated with five regulatory alternatives for the NOx SIP call. Estimates for both
most and least ozone sensitive crops are presented in Table 4-21. The highest benefit estimate of $415
million (assuming relatively sensitive cuitivars for the 0.12 Trading alternative) is a relatively small 0.6%
of the total 1997 crop value. This suggests that individual farmers are not likely to identify ozone
sensitivity as a major factor in observed yield changes in the presence of other more obvious factors, such
as meterology, fertilization, and pest resistance. Likewise, given the relative importance of other yield
enhancing crop traits, such as pest insistence, it is unlikely that seed developers will focus on development
of ozone tolerant varieties. Nonetheless, to the extent that ozone resistant cuitivars are available and
farmers respond to increased ozone levels by subtituting towards more ozone resistant cuitivars, crop
losses will be reduced.
Table 4-21
Changes in Production of Commodity Crops and Monetary
Benefits Associated with the NOx SIP Call
Regulatory Alternative
0.1 2 Trading
0.15 Trading
Regional ity 1
0.20 Trading
0,25 Trading
Monetary Benefits
(millions 1990$)
Least Sensitive Cuitivars
$53
$47
$43
$42
$34
Most Sensitive Cuitivars
$415
$361
$318
$312
$242
Page 4-40
-------
AGSIM is an alternative agricultural sector model which has gained popularity in the agricultural
economics field. It has been extensively peer-reviewed and it estimates a more complete set of
responses to yield changes than RMF. The primary difference is that AGSIM models planted acreage as
a behavioral response to yield and relative price changes, while RMF treats planted acres as a fixed
factor. As a sensitivity analysis, AGSIM was run for the five regulatory alternatives to determine how
AGSIM performs relative to RMF. RMF was chosen for the primary analysis because it has been
extensively tested and used in previous regulatory impact analyses. For a complete description of
AGSIM, see the AGSIM technical support document (Abt Associates, 1998b).
Table 4-22 presents AGSIM generated estimates of monetary benefits due to changes in the
production of all six commodity crops associated with the five regulatory alternatives for the NOx SIP
call. Estimates for both most and least ozone sensitive crops are presented in Table 4-22. As might be
expected, differences between the results of the models are relatively small when the least sensitive
cultivars are used, while differences are much larger when the most sensitive cultivars are used. Values
for AGSIM exceed those from RMF by 40 to 44 percent, depending on the regulatory alternative.
Table 4-22
Sensitivity Analysis: AGSIM Generated Monetary Benefits Due to Changes in
Production of Commodity Crops Associated with the NOx SIP Call
Regulatory Alternative
0.12 Trading
0,1 5 Trading
Regionally 1
0.20 Trading
0,25 Trading
Monetary Benefits
(millions 1990S)
Least Sensitive Cultivars
$5!
$44
$38
$37
$29
Most Sensitive Cultivars
$595
$521
$451
$440
$338
4.4,2 Commercial Forests
Any attempt to estimate economic benefits for commercial forests associated with reductions in
ozone arising from implementation of the NOx SIP call is constrained by a lack of exposure-response
functions for the commercially important mature trees. Although exposure-response functions have been
developed for seedlings for a number of important tree species, these seedling functions cannot be
extrapolated to mature trees based on current knowledge. Recognizing this limitation, a study (Pye, 1988
and deSteiger & Pye, 1990) involving expert judgment about the effect of ozone levels on percent growth
change is used to develop estimates of ozone-related economic losses for commercial forest products.
Page 4-41
-------
An analysis by Mathtech in conjunction with the USDA Forest Service (Mathtech, 1998b) of
forestry sector benefits quantifies the effect of ozone on tree growth and the demand and supply
characteristics of the timber market. The estimates do not include possible non-market benefits such as
aesthetic effects. Forest aesthetics is discussed qualitatively later in this chapter.
The economic value of yield changes for commercial forests was estimated using the 1993 timber
assessment market model (TAMM). TAMM is a U.S. Forest Service (Adams and Haynes, 1996) spatial
model of the solidwood and timber inventory elements of the U.S. forest products sector. The model
provides projections of timber markets by geographic region and wood type through the year 2040. Nine
regions covering the continental U.S. are included in the analysis. TAMM simulates the effects of
reduced O3 concentrations on timber markets by changing the annual growth rates of commercial forest
growing-stock inventories. The model uses applied welfare economics to value changes in ambient O3
concentrations. Specifically, TAMM calculates benefits as the net change in consumer and producer
surplus from baseline O3 concentrations to the O3 concentrations resulting from implementation of the
NOx SIP call policy.
Table 4-23 presents estimates of monetary benefits of yield changes of commercial forests
associated with the five policy alternatives for the NOx SIP call EPA did not estimate monetary benefits
for ail policy alternatives. Benefits for excluded alternatives can be easily estimated using a ratio of
estimated benefits to a similar benefit category, such as commodity crops. Benefits for the 0.25 trading
and Regionality 1 alternatives are estimated by applying the ratio of forestry to agricultural benefits for the
0.15 trading alternative, equal to 0.59, to the agricultural benefits for these two alternatives.
Because of the long harvesting cycle of commercial forests and the cumulative effects of higher
growth rates, the benefits to the future economy will be much larger than the estimates reported in Table
4-23. For example, the 0.12 trading policy alternative would result in about $8.0 billion additional forest
inventories by 2040. The estimated annualized benefits for this alternative., $233 million, are much lower
because of smaller benefits in earlier years (i.e., the 2010 and 2020 decades) and because the higher
benefits realized in later years are heavily discounted.
Table 4-23
Commercial Forest Monetary Benefits Associated with the NOx SIP Call
Regulatory
Alternative
0.12 Trading
0.15 Trading
Regionality 1
0.20 Trading
0.25 Trading
Monetary Benefits
(millions 1990$)
$233
$213
$188
$185
$143
4.4,3 Nitrogen Deposition
Page 4-42
-------
Excess nutrient loads, especially that of nitrogen, are responsible for a variety of adverse
consequences to the health of estuarine and coastal waters. These effects include toxic and/or noxious
algal blooms such as brown and red tides, low (hypoxic) or zero (anoxic) concentrations of dissolved
oxygen in bottom waters, the loss of submerged aquatic vegetation due to the light-filtering effect of thick
alga! mats, and fundamental shifts in phytoplankton community structure. Direct concentration-response
functions relating deposited nitrogen and reductions in estuarine benefits are not available. The preferred
willingness-to-pay based measure of benefits depends on the availability of these concentration-response
functions and on estimates of the value of environmental responses. Because neither appropriate
concentration-response functions nor sufficient information to estimate the marginal value of changes in
water quality exist at present, an avoided cost approach is used instead of willingness-to-pay to generate
estuary related benefits of the NOx SIP call.
The benefits to surrounding communities of reduced nitrogen loadings resulting from various
control strategies for atmospheric NOx emissions are calculated for 10 East and 2 Gulf Coast case study
estuaries, and extrapolated to all 43 Eastern U.S. estuaries. The 10 East Coast case study estuaries
represent approximately half of the estuarine watershed area in square miles along the East Coast. The
12 case study estuaries are chosen because of the availability of necessary data and their potential
representativeness. This analysis uses the following data for each estuary; (1) total nitrogen load from all
sources; (2) direct nitrogen load from atmospheric deposition to the estuary surface; (3) indirect nitrogen
load from atmospheric deposition to the estuary watershed and subsequent pass-through to the estuary
itself; (4) established nitrogen thresholds and reduction goals adopted by the community; and (5) costs
associated with using agreed upon non-point water pollution control technologies.
Atmospheric nitrogen reductions are valued in this analysis on the basis of avoided costs
associated with agreed upon controls of nonpoint water pollution sources. Benefits are estimated using an
average, locally-based cost for nitrogen removal from water pollution (EPA, 1998). Valuation reflects
water pollution control cost avoidance based on average cost/pound of current non-point source water
pollution controls for nitrogen in three case study estuaries; Albemarle/Pamlico Sounds, Chesapeake Bay,
and Tampa Bay. Taking the weighted cost/pound of these available controls assumes States will combine
low cost and high cost controls, which could inflate avoided cost estimates.
In a recent advisory statement, the EPA's Science Advisory Board (SAB), charged with
reviewing the benefits methodology for the §812 Prospective report on the benefits and costs of the Clean
Air Act Amendments, raised concerns about the use of the avoided cost approach to value reduced
ecosystem damages. Specifically, they identified a key requirement which should be met in order for
avoided costs to approximate environmental benefits. This requirement is that there is a direct link
between implementation of the air pollution regulation and the abandonment of a separate costly
regulatory program by some other agency, i.e. a state environmental agency. Reductions in nitrogen
deposition from the NOx SIP call are expected to impact estuaries all along the eastern seaboard and the
Gulf Coast. Many of the estuaries in these areas are currently being targeted by nitrogen reduction
programs due to current impairment of estuarine water quality by excess nutrients. Some of the largest of
these estuaries, including the Chesapeake Bay, have established goals for nitrogen reduction and target
dates by which these goals should be achieved. Using the best and most easily implemented existing
technologies, many of the estuaries will not be able to achieve the stated goals by the target dates. For
example, the Chesapeake Bay needs an additional 9,000 tons of nitrogen reductions per year and Long
Island Sound needs an additional 3,500 tons of reductions per year. Meeting these additional reductions
Page 4-43
-------
will require development of new technologies, implementation of costly existing technologies (such as
stormwater controls), or use of technologies with significant implementation difficulties, such as
agricultural best management practices (BMPs). Reductions in nitrogen deposition from the atmosphere
due to the NOx SIP call will directly reduce the need for these additional costly controls. Thus while the
NOx SIP call does not eliminate the need for nutrient management programs already in place, it may
substitute for some of the incremental costs and programs (such as an agricultural BMP program)
necessary to meet the nutrient reduction goals for each estuary. This then meets the SAB requirement
since the NOx SIP call will directly reduce the need for elements of separate costly reduction actions.
EPA believes that the use of an avoided cost approach in this RIA is consistent with the SAB
advice for appropriate use of avoided costs. The SAB did not provide direct guidance on alternative
approaches to measuring the benefits of reduced nitrogen deposition to estuaries. However, EPA
recognizes the fact that avoided costs do not directly measure the benefits of reduced ecological impacts
due to nitrogen deposition. Thus, while avoided cost is only a proxy for benefits, and should be viewed as
inferior to willingness-to-pay based measures, it is preferred to excluding any quantitative estimate of
benefits for this category. Current research is underway to develop other approaches for valuing
estuarine benefits, including contingent valuation and hedonic property studies. However, this research is
still sparse, and does not contain sufficient information on the marginal willingness-to-pay for changes in
concentrations of nitrogen (or changes in water quality or water resources as a result of changes in
nitrogen concentrations). As more studies become available, more complete estimates of the commercial
and ecological benefits of reduced atmospheric deposition of nitrogen can be incorporated into regulatory
analyses.
The fixed capital costs for non-point controls in the case study estuaries is ranged from $0.61 to
$45.27 per pound for agricultural and other rural best management practices and from $35 to $142.64 per
pound for urban nonpoint source controls (stormwater controls, reservoir management, onsite disposal
system changes, onsite BMPs). Using these as a base, the total fixed capital cost per pound (weighted on
the basis of fractional relationship of nitrogen load controlled for the estuary goal) is calculated for each of
the case-study estuaries and applied in the valuation of their avoided nitrogen load controlled. The
weighted capital costs per pound for the case-study estuaries are $32.88 for Albemarle-Pamlico Sounds,
$22.31 for Chesapeake Bay, and $88.25 for Tampa Bay4. For the purposes of this analysis, EPA
assumes that estuaries that have not yet established nutrient reduction goals will utilize the same types of
nutrient management programs as projected for the case study estuaries. For the other nine estuaries, an
average capital cost per pound of nitrogen (from the three case-estuaries) of $47.65/lb ($105/kg) is
calculated and applied; this cost may understate or overstate the costs associated with reductions in these
other estuaries. The other nine estuaries generally represent smaller, more urban estuaries (like Tampa
Bay), which typically have fewer technical and financial options available to control nitrogen loadings
from nonpoint sources. This may result in higher control costs more similar to the Tampa Bay case. On
the other hand, these estuaries may have opportunities to achieve additional point source controls at a
lower costs. Also, increased public awareness of nutrification issues and technological innovation may, in
the future, result in States finding lower cost solutions to nitrogen removal.
4 The value for Tampa Bay is not a true weighted cost per pound, but a midpoint of a range of $58,54 to
$117.65 developed by Apogee Research for the control possibilities (mostly urban BMPs) in the Tampa Bay estuary.
Page 4-44
-------
The 12 estuaries directly analyzed represent approximately 48% of the estuarine watershed area
along the East Coast (there are 43 East Coast estuaries of which 10 were in the sample, and 31 Gulf of
Mexico estuaries of which 2 are in the sample). Because NOAA data indicate that approximately 89%
(92.6% by watershed area plus surface area) of East Coast estuaries are highly or moderately nutrient
sensitive, it is reasonable to expect that estuaries not included in this analysis would also benefit from
reduced deposition of atmospheric nitrogen. Total benefits from the 12 representative estuaries are
scaled-up to include the remainder of the nutrient sensitive estuaries along the East Coast (92.6% of all
East Coast estuaries) on the basis of estuary watershed plus water surface area. Since the 12
representative estuaries account for 48 percent of total eastern estuarine area, estimates are scaled up by
multiplying the estimate for the 12 estuaries by 2.083 and then taking 92,6 percent of this estimate to
adjust for nutrient sensitivity.
All capital cost estimates are then annualized based on a 7% discount rale and a typical
implementation horizon for control strategies. Based on information from the three case study estuaries,
this typically ranges from 5 to 10 years. EPA has used the midpoint of 7.5 years for annualization, which
yields an annualization factor of 0.1759, Non-capital installation costs and annual operating and
maintenance costs are not included in these annual cost estimates. Depending upon the control strategy,
these costs can be significant. Reports on the Albemarle-Pamlieo Sounds indicate, for instance, that
planning costs associated with control measures comprises approximately 15% of capital costs.
Information received from the Association of National Estuary Programs indicates that operating and
maintenance costs are about 30% of capital costs, and that permitting, monitoring, and inspections costs
are about 1 to 2% of capital costs. For these reasons, the annual cost estimates may be understated.
Table 4-24 presents estimates of monetary benefits arising from the avoided costs of nitrogen
removal for the 12 estuaries with directly modeled nitrogen deposition changes and for the full set of 43
East Coast estuaries including extrapolated benefits associated with five regulatory alternatives for the
NOx SIP call. Estimates in Table 4-24 assume that 10 percent of nitrogen deposited over the watershed
reaches the estuary 5 costs for non-study estuaries are equal to the average of the costs for the three case
studies, and benefits are applied only to nutrient-sensitive estuaries.
Table 4-24
Monetary Benefits Associated with the NOx SIP Call from Avoided Costs
of Nitrogen Removal in Eastern Estuaries
Regulatory Alternative
0.12 Trading
0.15 Trading
Regionality 1
0.20 Trad ing
0.25 Trading
Monetary Benefits (millions 1990$)
12 Modeled Eastern Estuaries
$129
$123
$115
$109
S79
Extrapolation to 43 Eastern Estuaries
$248
$238
$221
$210
$152
Page 4-45
-------
4.4.4 Household Soiling Damage
Welfare benefits also accrue from avoided air pollution damage, both aesthetic and structural, to
architectural materials and to culturally important articles. At this time, data limitations preclude the ability
to quantify benefits for all materials whose deterioration may be promoted and accelerated by air pollution
exposure. However, this analysis addresses one small effect in this category, the soiling of households by
particulate matter.
Assumptions regarding the air quality indicator are necessary to evaluate the concentration-
response function. PM]0 and PM2S are both components of TSP. However, it is not clear which
components of TSP cause household soiling damage. The Criteria Document cites some evidence that
smaller particles may be primarily responsible, in which case these estimates are conservative.
Several studies have provided estimates of the cost to households of PM soiling. The study that is
cited by ESEERCO (1994) as one of the most sophisticated and is relied upon by EPA in its 1988
Regulatory Impact Analysis for SO2 is Manuel et al, (1982). Using a household production function
approach and household expenditure data from the 1972-73 Bureau of Labor Statistics Consumer
Expenditure Survey for over twenty cities in the United States, Manuel et al. estimate the annual cost of
cleaning per ^g/m3 PM per household as $1.26 ($0.48 per person times 2.63 persons per household). This
estimate is low compared with others (e.g., estimates provided by Cummings et al., 1981, and Watson and
Jaksch, 1982, are about eight times and five times greater, respectively). The ESEERCO report notes,
however, that the Manuel estimate is probably downward biased because it does not include the time cost
of do-it-yourselfers. Estimating that these costs may comprise at least half the cost of PM-related
cleaning costs, they double the Manuel estimate to obtain a point estimate of $2.52 (reported by
ESEERCO in 1992 dollars as $2.70).
Table 4-25 presents estimates of monetary benefits arising from the avoided household soiling
associated with the five regulatory alternatives for the NOx SIP call. Household soiling benefits are not
affected by the threshold assumption.
Page 4-46
-------
Table 4-2S
Monetary Benefits from Reduced Household Soiling Associated with the NOx SIP Call
Regulatory Alternative
0.1 2 Trading
0,15 Trading
Regional ity 1
0.20 Trading
0,25 Trading
Monetary Benefits (millions 1990$)
RADM-RPM
$26
$10
$10
$10
$7
S-R Matrix
$11
$7
$6
$6
$6
4.4.5 Visibility
Visibility effects are measured in terms of changes in deciview, a measure useful for comparing
the effects of air quality on visibility across a range of geographic locations. This measure is directly
related to two other common visibility measures: visual range (measured in km) and light extinction
(measured in knr1). The deciview measure characterizes visibility in terms of perceptible changes in
haziness independent of baseline conditions. Based on the deciview measure, two types of valuation
estimates are applied to the expected visibility changes: residential visibility and recreational visibility.
Visibility is a function of the ability of gases and aerosols to scatter and absorb light. RPM only
computes the loss of visibility due to sulfates, nitrates, organic matter, and elemental carbon, but not other
variables, such as coarse PM and fine soil. By not including these other terms, the resulting estimates of
WTP for residential and recreational visibility improvement are overestimated. Based on the full suite of
variables available at IMPROVE sites, the WTP estimates should, on average, be multiplied by 0.82 to
correct for this bias. The range of correction factors is from 0.40 to LOO, depending on the site and to a
lesser extent the policy alternative. Similarly, when calculating residential visibility, the S-R matrix
estimate includes terms for sulfates, nitrates and coarse PM, but does not include organic matter and
other variables. The results from the IMPROVE monitors suggest that to correct this bias, the WTP
estimates should be multiplied by 0.65 at the mean, with the correction factors ranging from 0.28 to 1.00.
Note that the S-R matrix recreational visibility estimates include the full suite of visibility variables, so no
correction is necessary,
Residential Visibility
The residential visibility valuation estimate is derived from the results of an extensive visibility
study (McClelland et al,, 1991). A household WTP value is derived by dividing the value reported in
McClelland et al, by the corresponding hypothesized change in deciview, yielding an estimate of $14 per
unit change in deciview. This WTP value is applied to all households in any area estimated to experience
a change in visibility.
Page 4-47
-------
Table 4-26 presents estimates of monetary benefits arising from improvements in residential
visibility due to reductions in PM associated with the five regulatory alternatives for the NOx SIP call.
Table 4-26 includes both unadjusted visibility values and values adjusted based on the average adjustment
factor Of 0.82 for the RADM-RPM set and 0.65 for the S-R Matrix set.
Table 4-26
Monetary Benefits from Improved Residential Visibility Associated with the NOx SIP Call
1
Regulatory
Alternative
0.1 2 Trading
0.1 5 Trading
Regionality I
0.20 Trading
0.25 Trading
Monetary Benefits (millions 1990S)
RADM-RPM
Unadjusted
$144
$34
$42
$46
$30
Adjusted
$118
$28
$34
$38
$25
S-R Matrix
Unadjusted
$92
$59
$41
$48
$46
Adjusted
$60
$38
$27
$31
$30
Recreational Visibility
The value of visibility improvements in certain National Parks in the Southeast is based on the
results of a 1990 Cooperative Agreement project jointly funded by the EPA and the National Park
Service, "Preservation Values For Visibility Protection at the National Parks". Based on that contingent
valuation study of visibility improvements, Chestnut (1997) calculates a household willingness to pay
(WTP) for visibility improvements, capturing both use and non-use recreational values, and accounts for
geographic variations in the willingness to pay. This method was used in the PM and ozone NAAQS RIA
analysis, and is adopted for the SIP call benefits analysis.
The Preservation Values study examined the demand for visibility in three broad regions of the
country, but only the Southeast region is directly relevant for the SIP call. Respondents both inside and
outside the Southeast region were asked their willingness to pay to protect visibility at four National Parks
in the region; Shenandoah, Mammoth Cave, Great Smoky Mountains, and Everglades National Parks.
Photos from Shenandoah (the "indicator park" in the Southeast region) were provided as part of the
survey instrument. Respondents were first asked for their value for preserving "only visibility at National
Parks in the Southeast". They were later asked to state what portion of their stated total value was for
visibility at the indicator park alone. Prior to providing their values, respondents were instructed that
"These questions concern only visibility at national parks in the Southeast and assume there wiil be no
change in visibility at national parks in other regions. Other households are being asked about visibility,
human health and vegetation protection in urban areas and at national parks in other regions". Therefore,
the estimated valuation functions for the Southeastern National Parks are specifically designed to be in
addition to any value for urban visibility. Note that the total value of recreational visibility improvements in
Southeastern National Parks is the sum of the value for indicator and non-indicator parks. The high
Page 4-48
-------
Southeast recreational visibility estimate applies the "in-region" value for Southeastern visibility changes to
the total population inside the Southeastern region, and the "out-of-region" value for Southeastern visibility
changes to all other populations in the U.S. The total in-region WTP per household is $6.50 per deciview
change, while the total out-of-region WTP per household is $4 per deciview change.
To take into account the possibility that the study did not fully account for double-counting, the
low Southeast recreational visibility estimate will apply values of non-Southeast residents for Southeastern
National Parks to populations both in and out of the Southeast region. The out-of-region value should not
include any value for improved residential visibility, because non-Southeast residents, by definition, live
outside the region, and thus are not included in the Southeast residential visibility calculation.
Table 4-27 presents estimates of monetary benefits arising improvements in recreational visibility
due to reductions in PM associated with the five regulatory alternatives for the NOx SIP call. Table 4-27
includes both unadjusted visibility values and values adjusted based on the average adjustment factor of
0.82 for the RADM-RPM set. As described in the beginning of this section, recreational visibility results
generated using the S-R Matrix do not need to be adjusted.
Table 4-27
Monetary Benefits from Improved Visibility in National Parks in the Southeast
Associated with the NOx SIP Call
Regulatory Alternative
0.1 2 Trading
0,1 5 Trading
Regionally 1
0.20 Trading
0.25 Trading
Monetary Benefits (millions 1990$)
RADM-RPM
Unadjusted
Low
$64
$36
$40
$36
$28
High
$77
$43
$49
$43
$34
Adjusted
Low
$52
$30
$33
$30
$23
High
$63
$35
$40
$35
$28
S-R Matrix
Low
$21
$15
$11
$14
$12
High
$22
$15
$10
$14
$11
The SIP call will impact visibility at other national parks than the specific parks examined in the
Preservation Values Study's Southeast region. Visibility conditions will improve at additional national
parks and recreation areas in the Southeast, as well as parks in the Northeast and Midwest. The air
quality model (RADM-RPM and S-R Matrix) used to estimate visibility improvements produces estimates
of the improvements at all locations throughout the SIP call region. However, there are no direct
valuation studies available for these other areas.
To explore the potential magnitude of the value of improved visibility outside the Preservation
Values Study's Southeast region, valuation information about the demand for visibility in the Southeast
Page 4-49
-------
from the Preservation Values study is used to approximate the value of visibility improvements at national
parks outside the Southeast, In order to account for geographic variability in WTP, the Preservation
Values valuation method divided the recreational areas of the United States into three regions. Separate
values were estimated for households living in each region, as well as for households living in other parts
of the United States for visibility improvements in each region. In-region respondents placed higher
values on visibility improvements at a local recreational area than out-of-region respondents. The lowest
resident and non-resident values for any national parks examined in the Preservation Values study was
for visibility in the Southeast. The out-of-region values in the Southeast will be used as an approximation
of the value of national parks in the Central and Northeastern U.S. that are impacted by the NOx SIP
call.
For the low Central and Northeastern recreational visibility estimates, out-of-region values per
household for the Southeastern non-indicator parks (equal to 60 percent of the total value per deciview, or
$2.40) are used to approximate the value to populations both outside and inside the Central and
Northeastern U.S. of visibility at Central and Northeastern national parks. Out-of-region values are used
for both sets of populations to avoid the possibility of double-counting benefits already accounted for in the
calculation of residential visibility benefits. Non-indicator park values are used to account for the fact that
indicator parks in a region may have unique values relative to non-indicator parks and therefore values for
these indicator parks are not appropriate to transfer to non-indicator parks. For the high Central and
Northeastern recreational visibility estimates, out-of-region values per household for the Southeastern
non-indicator parks are used to approximate the value to populations outside the Central and Northeastern
U.S. of visibility at Central and Northeastern national parks, and in-region values per household for the
Southeastern non-indicator parks are used to approximate the value to populations within the Central and
Northeastern U.S. The sum of monetary benefits for Southeast and Central and Northeastern visibility
benefits will be used in the calculation of total benefits.
Table 4-28 presents the recreational visibility values for national parks outside the Southeast.
Table 4-28 includes both unadjusted visibility values and values adjusted based on the average adjustment
factor of 0.82 for the RADM-RPM set. As described in the beginning of this section, recreational
visibility results generated using the S-R Matrix do not need to be adjusted. Recreational visibility benefits
are predicted to be reduced when using the S-R Matrix generated visibility changes. This is due to
predicted increases in PM in Minnesota and Maine, where two of the major parks outside the Southeast
are located.
Total low-end recreational visibility benefits (Southeast plus Northeast) using RADM-RPM
generated visibility changes range from $24 million for the 0.25 trading alternative to $71 million for the
0.12 trading alternative. Total RADM-RPM based high-end recreational visibility benefits range from
$29 million for the 0.25 trading alternative to $85 million for the 0.12 trading alternative. Total S-R Matrix
based low-end recreational visibility benefits range from $5 million for the Regionally 1 alternative to $22
million for the 0.12 trading alternative. Total S-R Matrix based high-end recreational visibility benefits
range from $2 million for the Regionality I alternative to $21 million for the 0.12 trading alternative.
Page 4-50
-------
Table 4-28
Monetary Benefits Associated with Visibility Changes in National Parks
Outside the Southeast in the NOx SIP Call Region
Regulatory Alternative
0.12 Trading
0.15 Trading
Regionally 1
0.20 Trading
0,25 Trading
Monetary Benefits (millions 1990$)
RADM-RPM
Unadjusted
Low
$22.8
$13.4
$7. 1
$8.3
$1.3
High
$27.0
$15,9
$8.5
$9.8
$1.6
Adjusted
Low
$18.7
$11.0
$5.8
$6.8
$1,1
High
$22.1
$13,0
$7.0
$8,0
$1.3
S-R Matrix
Low
$0.5
$-5,4
$-6.5
S.C O
-JrQ
$*% 0
-J.Z,
High
$0.6
$-6.4
$-7.7
$-6,9
$-6.2
4.5 Total Benefits
The dollar benefits from reducing ozone and PM levels resulting from implementing the SIP call
NOx reductions is the sum of dollar benefits from the reductions in incidence of all non-overlapping health
and welfare endpoints associated with PM and ozone for a given set of assumptions. If two endpoints are
overlapping, then adding the benefits associated with each will result in double counting of some benefits.
Although study-specific point estimates of dollar benefits associated with specific, possibly overlapping
endpoints are presented separately, estimation of total benefits requires that the benefits from only non-
overlapping endpoints be included in the total. Four non-overlapping broad categories of health and
welfare endpoints will be included in the estimation of total dollar benefits for the SIP call: (1) mortality,
(2) hospital admissions, (3) respiratory symptoms/illnesses not requiring hospital admission, and (4)
welfare endpoints. When considering only point estimates, aggregation of the benefits from different
endpoints is relatively straightforward. Once a set of non-overlapping categories is determined, the point
estimate of the total benefits associated with the health and welfare endpoints in the set is just the sum of
the endpoint-specific point estimates. If each endpoint-specific point estimate is the mean of a distribution
of dollar benefits associated with that endpoint, then the point estimate of total dollar benefits is just the
sum of those means.
There is uncertainty about the magnitude of the total monetized benefits associated with any of
the SIP call regulatory alternatives examined in the benefits analysis. The benefits are uncertain because
there is uncertainty surrounding each of the factors that affect these benefits: the changes in ambient
pollutant concentrations that will result from the SIP call implementation; the relationship between these
changes in pollutant concentrations and each of the associated health and welfare endpoints; and the
value of each adverse health and welfare effect avoided by the reduction in pollutant concentrations.
Page 4-51
-------
Much of the uncertainty derives from uncertainty about the true values of analysis components,
such as the value of the ozone coefficient in a concentration-response function relating ozone to a
particular health endpoint, or the true dollar value of an avoided hospital admission for congestive heart
failure. The analysis relies on estimates of these parameters, but the true values being estimated are
unknown. This type of uncertainty can often be quantified. For example, the uncertainty about pollutant
coefficients is typically quantified by reported standard errors of the estimates of the coefficients in the
concentration-response functions estimated by epidemiologicai studies. Appendix A presents a formal
quantitative analysis of the statistical uncertainty imparted to the benefits estimates by the variability in the
underlying concentration-response and valuation functions,
Some of the uncertainty surrounding the results of a benefits analysis, however, involves basically
discrete choices and is less easily quantified. For example, the decision of which air quality model to use
to generate changes in ambient PM concentrations is a choice between two models, embodying discrete
sets of air chemistry and mathematical assumptions. Decisions and assumptions must be made at many
points in an analysis in the absence of complete information. The estimate of total benefits is sensitive to
the decisions and assumptions made. Among the most critical of these are the following:
• Ozone mortality: There is some uncertainty surrounding the existence of a relationship between
tropospheric ozone exposure and premature mortality. The two possible assumptions are: (1) that
there is no relationship between ozone and mortality; and (2) that there is a potential relationship
between ozone and mortality, which we can quantify based on the meta-anaiysis of current U.S.
ozone mortality studies,
• Ozone agriculture effects: The existing set of exposure-response functions relating crop yields
to changes in ozone exposure include both ozone-sensitive and ozone-insensitive cultivars.
Possible assumptions are: (1) plantings of commodity crop cultivars are primarily composed of
sensitive varieties; (2) plantings of commodity crop cultivars are primarily composed of non-
sensitive varieties.
* PM2,S concentration threshold; Health effects are measured only down to the assumed
ambient concentration threshold. Changes in air quality below the threshold will have no impact
on estimated benefits. EPA's Science Advisory Board has recommended examining alternative
thresholds, including background and 15 u,g/m3.
« Sulfate Dominance: There are two possible interpretations of PM-related health and welfare
benefits depending on the model used to assess air quality changes: (1) results generated with
RADM-RPM are indicative of a future eastern U.S. atmosphere where acid sulfate levels are
still high enough to control atmospheric chemistry, and more specifically ammonium nitrate
particle formation. In this circumstance, reductions in NOx emissions may result in non-linear
responses in total fine particle levels, involving both decreases and increases; and (2) results
generated with the Source-Receptor Matrix are indicative of a future eastern U.S. atmosphere
where acid sulfate levels do not dominate particle formation chemistry. Jn this case, reductions in
NOx emissions would be expected to result more directly in linear reductions in PM.
Page 4-52
-------
* Recreational visibility; Recreational visibility benefits for residents of the Southeast may
overlap with "residential" visibility benefits. Two alternative assumptions may be considered for
in-region residents: (1) recreational visibility benefits overlap with residential visibility benefits, and
to avoid this overlap, the recreational visibility value of $4 per deciview for out-of-region residents
is used for in-region residents ($2.40 for non-indicator parks, and $1.60 for the indicator park); or
(2) recreational visibility benefits are in addition to residential visibility benefits, and the in-region
value of $6.50 is used ($3,25 for non-indicator parks, and $3.25 for the indicator park).
Benefits from visibility improvements may also occur in NOx SIP call states outside of the
Southeast. The current literature on the value of recreational visibility in national parks is limited to
studies of values in California, the Southwest, and the Southeast, and thus excludes the Central
and Northeast (CNE) portion of the NOx SIP call region. Three alternative assumptions may be
considered when valuing visibility changes in the CNE: (1) recreational visibility values in the
CNE are much less than that in the Southeast and therefore to insure benefits are not overstated,
no value should be associated with visibility changes in the CNE; (2) recreational visibility values
in the CNE are similar to the values for non-indicator parks in the Southeast, and recreational and
residential benefits overlap: people in and out of the CNE region value CNE recreational visibility
at $2.40 per deciview; or (3) recreational visibility values in the CNE are similar to the values for
non-indicator parks in the Southeast, and there is no overlap of recreational and residential
benefits: the in-region CNE value is based on the Southeast in-region value of $3.25 per deciview,
and the out-of-region CNE value is based on the Southeast out-of-region value of $2.40 per
deciview.
Tables 4-29 through 4-33 present summaries of the endpoint specific monetary values and the
estimate of total benefits for each of the five regulatory alternatives. Aggregate results are presented for
two assumption sets: 1) a "low" assumption set reflecting the assumptions that human health and the
environment have low responsiveness to changes in ambient air quality, and 2) a "high" assumption set
reflecting the assumptions that human health and the environment are highly responsive to changes in
ambient air quality. The "low" assumption set includes the following assumptions: 1) there are no PM-
related health effects occurring below a threshold of 15 (ig/m3, 2) changes in PM concentrations are
more accurately represented by the RADM-RPM air quality model, 3) there is no relationship between
ozone and premature mortality, 4) agricultural commodity crops are less sensitive to ozone, 5)
Southeastern recreational visibility values are not transferable to changes in recreational visibility in the
Northeast and Central U.S., and 6) the low-end recreational visibility valuation method is correct, The
"high" assumption set includes the following assumptions: 1) PM-related health effects occur down to the
anthropogenic background threshold, 2) changes in PM concentrations are more accurately represented
by the S-R Matrix air quality model, 3) the relationship between ozone and premature mortality is
characterized by the distribution of avoided incidences derived from the ozone mortality meta-analysis, 4)
agricultural commodity crops are more sensitive to ozone, 5) Southeastern recreational visibility values are
transferable to the Northeastern and Central U.S., and 6) the high-end recreational visibility method is
correct.
Page 4-53
-------
Table 4-29
Total Qiiantified Monetary Benefits Associated with the NOx SIP Call,
Incremental to the 2007 Base Case: 0.12 Trading Regulatory Alternative*
Endpoint
Monetary Benefits (million 1990$)
"Low" Assumption Set
"High" Assumption Set
Ozone-related Enclpjjints
Short-term mortality
Hospital admissions
Acute respiratory symptoms
Worker productivity
Commodity crops
Commercial forests
$0
$5
$1
$25
$53
$233
$1,496
$5
$1
$25
$415
$233
PM-related Endpoints
Long-term mortality
Hospital admissions
Chronic bronchitis
Acute bronchitis
Acute respiratory symptoms
Work loss days
MRADs
Household soiling
Residential visibility
Recreational visibility
Nitrogen deposition
TOTAL
$1,468
$3
$589
$0
$0
$14
$53
$26
$118
$52
$248
$2,888
$2,672
$4
$245
$0
$0
$8
$29
$11
$60
$21
$248
$5,473
* Not all possible benefits are quantified and monetized in this analysis. Potential benefit categories that have not been quantified
and monetized are listed in Table 4-2.
Page 4-54
-------
Table 4-30
Total Quantified Monetary Benefits Associated with the NOx SIP Call,
Incremental to the 2007 Base Case: (US Trading Regulatory Alternative"
Endpoint
Monetary Benefits (million 1990$)
"Low" Assumption Set
"High" Assumption Set
Ozone-related End points
Short-term mortality
Hospital admissions
Acute respiratory symptoms
Worker productivity
Commodity crops
Commercial forests
$0
$4
$1
$22
$47
$213
$1,326
$4
$1
$22
$361
$213
PM-related Endpoints
Long-term mortality
Hospital admissions
Chronic bronchitis
Acute bronchitis
Acute respiratory symptoms
Work loss days
MRADs
Household soiling
Residential visibility
Recreational visibility
Nitrogen deposition
TOTAL
$251
$1
$225
$0
$0
$6
$24
$10
$28
$30
$238
$1,100
$1,763
$4
$160
$0
$0
$5
$19
$7
$38
$9
$238
$4,170
a Not all possible benefits are quantified and monetized in this analysis, Potential benefit categories that have not been quantified
and monetized are listed in Table 4-2.
Page 4-55
-------
Table 4-31
Total Quantified Monetary Benefits Associated with the NOx SIP Call,
Incremental to the 2007 Base Case: Regionally 1 Regulatory Alternative"
End point
Monetary Benefits (million 1990$)
"Low" Assumption Set
"High" Assumption Set
Ozone-related Endpoints
Short-term mortality
Hospital admissions
Acute respiratory symptoms
Worker productivity
Commodity crops
Commercial forests
$0
$4
$1
$20
$43
$188
$1,191
$4
$1
$20
$318
$188
PM-related Endpoints
Long-term mortality
Hospital admissions
Chronic bronchitis
Acute bronchitis
Acute respiratory symptoms
Work loss days
MRADs
Household soiling
Residential visibility
Recreational visibility
Nitrogen deposition
TOTAL
$317
$1
$236
$0
$0
$6
$24
$10
$34
$33
$221
$1,138
$1,326
$4
$122
$0
$0
$4
$15
$6
$27
$10
$221
$3,457
" Nol all possible benefits are quantified and monetized in this analysis. Potential benefit categories that have not been quantified
and monetized are listed in Table 4-2.
Page 4-56
-------
Table 4-32
Total Quantified Monetary Benefits Associated with the NOx SIP Call,
Incremental to the 2007 Base Case: 0.20 Trading Regulatory Alternative*
End point
Monetary Benefits (million 1990S)
"Low" Assumption Set
"High" Assumption Set
Ozone-related Endpoints
Short-term mortality
Hospital admissions
Acute respiratory symptoms
Worker productivity
Commodity crops
Commercial forests
$0
$4
$1
$20
$42
$185
$1,108
$4
$1
$20
$312
$185
PM-rclated Endpoints
Long-term mortality
Hospital admissions
Chronic bronchitis
Acute bronchitis
Acute respiratory symptoms
Work loss days
MRADs
Household soiling
Residential visibility
Recreational visibility
Nitrogen deposition
TOTAL
$370
$1
$216
$0
$0
$5
$24
$10
$38
$30
$210
$1,156
$1,499
$4
$135
$0
$0
$4
$17
$6
$31
$7
$210
$3,543
a Not all possible benefits are quantified and monetized in this analysis. Potential benefit categories that have not been quantified
and monetized are listed in Table 4-2.
Pace 4-57
-------
Table 4-33
Total Quantified Monetary Benefits Associated with the NOx SIP Call,
Incremental to the 2007 Base Case: 0,25 Trading Regulatory Alternative"
Endpoint
Monetary Benefits {million 1990$)
"Low" Assumption Set
"High" Assumption Set
Ozone-related End points
Short-term mortality
Hospital admissions
Acute respiratory symptoms
Worker productivity
Commodity crops
Commercial forests
$0
$3
$1
$14
$34
$143
$824
$3
$1
$14
$242
$143
PM-rcIated Endpoints
Long-term mortality
Hospital admissions
Chronic bronchitis
Acute bronchitis
Acute respiratory symptoms
Work loss days
MRADS
Household soiling
Residential visibility
Recreational visibility
Nitrogen deposition
TOTAL
$208
$1
$148
$0
$0
$4
$14
$7
$25
$23
$152
$777
$1,400
$4
$127
$0
$0
$4
$16
$6
$30
$5
$152
$2,971
" Not alt possible benefits are quantified and monetized in this analysis, Potential benefit categories thai have not been quantified
and monetized are listed in Table 4-2,
Page 4-5 8
-------
4.6 Limitations of the Analysis
Given incomplete information, this national benefits analysis yields approximate results because of
the uncertainty associated with any estimate. Potentially important sources of uncertainty exist and many
of these are summarized in Table 4-34, In most cases, there is no apparent bias associated with the
uncertainty. For those cases for which the nature of the uncertainty suggests a direction of possible bias,
this direction is noted in the table.
4.6,1 Projected Income Growth
This analysis does not attempt to adjust benefits estimates to reflect expected growth in real
income, "Economic theory argues, however, that WTP for most goods (such as environmental protection)
will increase if real incomes increase. The degree to which WTP may increase for the specific health
and welfare benefits provided by the NOx SIP call cannot be estimated due to insufficient income
elasticity information. Thus, all else being equal, the benefit estimates presented in this analysis are likely
to be understated.
4.6,2 Unquantifiable Benefits
In considering the monetized benefits estimates, the reader should be aware that many limitations
for conducting these analyses are mentioned throughout this RIA, One significant limitation of both the
health and welfare benefits analyses is the inability to quantify many PM and ozone-induced adverse
effects. Table 4-2 lists the categories of benefits that this analysis is able to quantify and those discussed
only in a qualitative manner. In general, if it were possible to include the unquantified benefits categories
in the total monetized benefits, the benefits estimates presented in this RIA would increase. Specific
examples of unquantified benefits explored in more detail below include other human health effects, urban
ornamentals, aesthetic injury to forests, nitrogen in drinking water, and brown clouds.
The benefits of reductions in a number of ozone- and PM-induced health effects have not been
quantified due to the unavailability of concentration-response and/or economic valuation data. These
effects include; reduced pulmonary function, morphological changes, altered host defense mechanisms,
cancer, other chronic respiratory diseases, infant mortality, airway responsiveness, increased susceptibility
to respiratory infection, pulmonary inflammation, acute inflammation and respiratory cell damage, and
premature aging of the lungs.
Page 4-59
-------
Table 4-34
Sources of Uncertainty in the Benefit Analysts
1. Uncertainties Associated With Concentration-Response Functions
There is uncertainty surrounding the ozone or PM coefficient in each C-R function,
There is uncertainty about applying a single C-R function to pollutant changes and populations in all locations.
It is uncertain how similar future year C-R relationships will be to current concentration-response relationships.
The correct functional form of each C-R relationship is uncertain. For example, it is uncertain whether there are
thresholds and, if so, what they are.
There is uncertainty associated with extrapolation of C-R relationships beyond the range of ozone or PM
concentrations observed in the study.
2. Uncertainties Associated With Daily Ozone and PM Concentrations
jThere is uncertainty surrounding the projected hourly ozone and daily PM concentrations.
The changes in ozone and PM concentrations resulting from the SIP call provisions are uncertain.
3. Uncertainties Associated With Possible Lagged Effects
It is uncertain what portion of the PM-related long-term exposure mortality effects associated with changes in
annual PM levels would occur in a single year, and what portion might occur in subsequent years.
4. Uncertainties Associated With Baseline Incidence Rates
Some baseline incidence rates are not location-specific (e.g., those taken from studies) and may therefore not
accurately represent the actual location-specific rates.
It is uncertain how well current baseline incidence rates approximate what baseline incidence rates will be in the
year 2007, given either "as is" ozone and PM concentrations or any alternative SIP call scenario.
It is uncertain how well the projected population and demographics; used to derive incidences, approximate what
the actual population and demographics will be in the year 2007.
5. Uncertainties Associated With Economic Valuation
Unit dollar values associated with health and welfare endpoints are only estimates of MWTP and therefore have
uncertainty surrounding them. Possible directions of bias are discussed in the technical support document (Abt
Associates, I998a),
Even using constant dollars (e.g., 1990 dollars), it is uncertain whether MWTP for each type of risk reduction will
be the same in the year 2007 as the current MWTP,
There is uncertainty about the appropriate discount rate for benefits achieved in the future (2007),
6. Uncertainties Associated With Aggregation of Monetized Benefits
Because benefit estimation is limited to those health and welfare endpoints for which concentration-response
functions have been estimated, there may be components of total benefit omitted. This would lead to a downward
bias in the estimated total monetized benefit,
Page 4-60
-------
In addition to the above non-monetized health benefits, there are a number of non-monetized
welfare benefits of NOx emission controls from reduced adverse effects on vegetation, forests, and other
natural ecosystems. The CAA and other statutes, through requirements to protect natural and ecological
systems, indicate that these are scarce and highly valued resources. Lack of comprehensive information,
insufficient valuation tools, and significant uncertainties result in understated welfare benefits estimates in
this RIA. However, a number of expert biologists, ecologists, and economists (Costanza, 1997) argue that
the benefits of protecting natural resources are enormous and increasing as ecosystems become more
stressed and scarce in the future. Additionally, agricultural, forest and ecological scientists (Heck, 1997)
believe that vegetation appears to be more sensitive to ozone than humans and consequently, that damage
is occurring to vegetation and natural resources at concentrations below the ozone NAAQS, Experts also
believe that the effect of ozone on plants is both cumulative and long-term. The specific non-monetized
benefits from reductions in ambient ozone concentrations would accrue from: decreased foliar injury;
averted growth reduction of trees in natural forests; maintained integrity of forest ecosystems (including
habitat for native animal species); and the aesthetics and utility of urban ornamentals (e.g., grass, flowers,
shrubs and trees). Other welfare categories for which there is incomplete information to estimate the
economic value of reduced adverse effects include; existence value of Class I areas; materials damage;
reduced sulfate deposition to aquatic and terrestrial ecosystems; and visibility impairment due to "brown
clouds" (i.e., distinct brown layers of trapped air pollutants close to the ground).
Other Human Health Effects
Human exposure to PM and ozone is known to cause health effects such as: airway
responsiveness, increased susceptibility to respiratory infection, acute inflammation and respiratory cell
damage, premature aging of the lungs and chronic respiratory damage. An improvement in ambient PM
and ozone air quality is expected to reduce the- number of incidences within each effect category that the
U.S. population would experience. Although these health effects are known to be PM or ozone-induced,
concentration-response data is not available for quantifying the benefits associated with reducing these
effects. The inability to quantify these effects leads to an underestimation of the monetized benefits
presented in this analysis.
Urban Ornamentals
Urban ornamentals represent an additional vegetation category likely to experience some degree
of effects associated with exposure to ambient ozone levels and likely to impact large economic sectors.
In the absence of adequate exposure-response functions and economic damage functions for the potential
range of effects relevant to these types of vegetation, no direct quantitative economic benefits analysis
has been conducted. Ornamentals used in the urban and suburban landscape include shrubs, trees,
grasses, and flowers. The types of economic losses that could potentially result from effects that have
been associated with ozone exposure include: 1) reduction in aesthetic services over the realized lifetime
of a plant; 2) the loss of aesthetic services resulting from the premature death (or early replacement) of
an injured plant; 3) the cost associated with removing the injured plant and replacing it with a new plant;
4) increased soil erosion, 5) increased energy costs from loss of shade in the urban environment; 6)
reduced seedling survivability; and 7) any additional costs incurred over the lifetime of the injured plant to
mitigate the effects of ozone-induced injury. It is estimated that more than $20 billion (1990 dollars) are
spent annually on landscaping using ornamentals (Abt Associates, 1995), both by private property
owners/tenants and by governmental units responsible for public areas, making this a potentially important
Page 4-61
-------
welfare effects category. However, information and valuation 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,
Aesthetic Injury to Forests
Ozone is a regionally dispersed air pollutant that has been shown conclusively to cause discernible
injury to forest trees (Fox, 1995). One of the welfare benefits expected to accrue as a result of
reductions in ambient ozone concentrations in the United States is the economic value the public receives
from reduced aesthetic injury to forests. There is sufficient scientific information available that ambient
ozone levels cause visible injury to foliage and impair the growth of some sensitive plant species. Ozone
inhibits photosynthesis and interferes with nutrient uptake, causing a loss in vigor that affects the ability of
trees to compete for resources and makes them more susceptible to a variety of stresses (EPA, 1996a, p.
5-251). Extended or repeated exposures may result in decline and eventual elimination of sensitive
species. Ozone concentrations of 0.06 ppm or higher are capable of causing injury to forest ecosystems.
The most notable effects of ozone on forest aesthetics and ecosystem function have been
documented in the San Bernardino Mountains in California. Visible ozone-related injury, but not
necessarily ecosystem effects, have also been observed in the Sierra Nevada in California, the
Appalachian Mountains from Georgia to Maine, the Blue Ridge Mountains in Virginia, the Great Smoky
Mountains in North Carolina and Tennessee, and the Green Mountains in Vermont (EPA, 1996a, pp. 5-
250 to 5-251). These are all locations where there is substantial recreation use and where scenic quality
of the forests is an important characteristic of the resource. Economic valuation studies of lost aesthetic
value of forests attributed to plant injuries caused by ozone are limited to two studies conducted in
Southern California (Crocker, 1985; Peterson et al., 1987), Both included contingent valuation surveys
that asked respondents what they would be willing to pay for reductions in (or preventions of increases in)
visible ozone injuries to plants. Crocker found that individuals are willing to pay a few dollars more per
day to gain access to recreation areas with only slight ozone injury instead of areas with moderate to
severe injury. Peterson et al. estimated that a one-step change (on a 5 point scale) in visible ozone injury
in the San Bernardino and Angeles National Forests would be valued at an aggregate amount of between
$27 million and $144 million for all residents of Los Angeles, Orange, and San Bernardino counties, A
reassessment of the survey design, in light of current standards for contingent valuation research, suggests
that it is plausible that concerns for forest ecosystems and human health could have been embedded into
these reported values. The extent of this possible bias is uncertain.
Present analytic tools and resources preclude EPA from quantifying the benefits of improved
forest aesthetics in the eastern U.S. expected to occur from the NOx SIP call. This is due to limitations
in our ability to quantify the relationship between ozone concentrations and visible injury, and limited
quantitative information about the value to the public of specific changes in visible aesthetic quality of
forests. However, there is sufficient supporting evidence in the physical sciences and economic literature
to support the finding that the proposed NOx SIP call can be expected to reduce injury to forests, and that
reductions in these injuries will likely have a significant economic value to the public.
Pase 4-62
-------
Nitrates in Drinking Water
Nitrates in drinking water are currently regulated by a maximum contaminant level (MCL) of 10
mg/L on the basis of the risk to infants of methemoglobinemia, a condition which adversely affects the
blood's oxygen carrying capacity. In an analysis of pre-1991 data, Raucher, et al, (1993) found that
approximately 2 million people were consuming public drinking water supplies which exceed the MCL.
Supplementing these findings, the National Research Council concluded that 42 percent of the public
drinking water users in the U.S. (approximately 105 million people) are either not exposed to nitrates or
are exposed to concentrations below 1.3 mg/L (National Research Council, 1995).
In a recent epidemiological study by the National Cancer Institute, a statistically significant
relationship between nitrates in drinking water and incidence of non-Hodgkin's lymphoma were reported
(Ward, et al., 1996), Though it is generally acknowledged that traditional water pollution sources such as
agricultural runoff are mostly responsible for violations of the MCL, other more diffuse sources of nitrate
to drinking water supplies, such as that from atmospheric deposition, may also become an important health
concern should the cancer link to nitrates be found valid upon further study.
Brown Clouds
NOx emissions, especially gaseous NO2 and NOx aerosols, can cause a brownish color to appear
in the air (EPA, I996c). In higher elevation western cities where wintertime temperature inversions
frequently trap air pollutants in atmospheric layers close to the ground, this can result in distinct brown
layers, In the eastern U.S., a layered look is not as common, but the ubiquitous haze sometimes takes on
a brownish hue. To date, economic valuation studies concerning visual air quality have focused primarily
on the clarity of the air, and have not addressed the question of how the color of the haze might be related
to aesthetic degradation. It may be reasonable to presume that brown haze is likely to be perceived as
dirty air and is more likely to be associated with air pollution in people's minds. It has not, however, been
established that the public would have a greater value for reducing brown haze than for a neutral colored
haze. Results of economic valuation studies of visibility aesthetics conducted in Denver and in the eastern
U.S. (McClelland et al., 1991) are not directly comparable because changes in visibility conditions are not
defined in the same units of measure. However, the WTP estimates for improvements in visibility
conditions presented in this assessment are based on estimates of changes in clarity of the air (measured
as deciview) and do not take into account any change in color that may occur. It is possible that there
may be some additional value for reductions in brownish color that may also occur when NOx emissions
are reduced.
Other Unquantiflable Benefits Categories
There are other welfare benefits categories for which there is incomplete information to permit a
quantitative assessment for this analysis. For some endpoints, gaps exist in the scientific literature or key
analytical components and thus do not support an estimation of incidence. In other cases, there is
insufficient economic information to allow estimation of the economic value of adverse effects.
Potentially significant, but unquantified welfare benefits categories include: existence and user values
related to the protection of Class I areas (e.g., Shenendoah National Park), damage to tree seedlings of
more than 10 sensitive species (e.g., black cherry, aspen, ponderosa pine), non-commercial forests,
ecosystems, materials damage, and reduced sulfate deposition to aquatic and terrestrial ecosystems.
Page 4-63
-------
Although scientific and economic data are not available to allow quantification of the effect of ozone in
these categories, the expectation is that, if quantified, each of these categories would lead to an increase
in the monetized benefits presented in this RIA,
4.6,3 Potential Disbenefits
In this discussion of unqualified benefits, a discussion of potential disbenefits must also be
mentioned. Several of these disbenefit categories are related to nitrogen deposition while one category is
related to the issue of ultraviolet light.
Passive Fertilization
Several disbenefit categories are related to nitrogen deposition. Nutrients deposited on crops
from atmospheric sources are often referred to as passive fertilization. Nitrogen is a fundamental nutrient
for primary production in both managed and unmanaged ecosystems. Most productive agricultural
systems require external sources of nitrogen in order to satisfy nutrient requirements. Nitrogen uptake by
crops varies, but typical requirements for wheat and corn are approximately 150 kg/ha/yr and 300
kg/ha/yr, respectively (NAPAP, 1990). These rates compare to estimated rates of passive nitrogen
fertilization in the range of 0 to 5.5 kg/ha/yr (NAPAP, 1991). So, for these crops, deposited nitrogen
could account for as much as 2 to 4 percent of nitrogen needs. Holding all other factors constant,
farmers' use of purchased fertilizers or manure may increase as deposited nitrogen is reduced. EPA has
not estimated the potential value of this possible increase in the use of purchased fertilizers, but a
qualitative assessment of several factors suggests that the overall value is very small relative to the value
of other health and welfare endpoints presented in this analysis. First, reductions in NOx emissions affect
only a fraction of total nitrogen deposition. Approximately 70 to 80 percent of nitrogen deposition is in the
form of nitrates (and thus can be traced to NOx emissions) while most of the remainder is due to
ammonia emissions (personal communication with Robin Dennis, NOAA Atmospheric Research Lab,
1997). Table 3-4 in Chapter 3 indicates the annual average change in nitrogen deposition attributable to
the 0.15 Trading alternative of the NOx SIP call is about 11 percent of baseline levels, suggesting a
relatively small potential change in passive fertilization. Second, some sources of nitrogen, such as animal
manure, are available at no cost or at a much lower cost than purchased nitrogen. In addition, in certain
areas nitrogen is currently applied at rates which exceed crop uptake rates, usually due to an
overabundance of available nutrients from animal waste. Small reductions in passive fertilization in these
areas is not likely to have any consequence to fertilizer application. The combination of these factors
suggests that the cost associated with compensating for reductions in passive fertilization is relatively
minor.
Information on the effects of changes in passive nitrogen deposition on forestlands and other
terrestrial ecosystems is very limited. The multiplicity of factors affecting forests, including other potential
stressors such as ozone, and limiting factors such as moisture and other nutrients, confound assessments
of marginal changes in any one stressor or nutrient in forest ecosystems. However, reductions in
deposition of nitrogen could have negative effects on forest and vegetation growth in ecosystems where
nitrogen is a limiting factor (EPA, 1993).
Page 4-64
-------
On the other hand, there is evidence that forest ecosystems in some areas of the United States
are nitrogen saturated (EPA, 1993). Once saturation is reached, adverse effects of additional nitrogen
begin to occur such as soil acidification which can lead to leaching of nutrients needed for plant growth
and mobilization of harmful elements such as aluminum. Increased soil acidification is also linked to
higher amounts of acidic runoff to streams and lakes and leaching of harmful elements into aquatic
ecosystems.
Ultraviolet Light
A reduction of tropospheric ozone is likely to increase the penetration of ultraviolet light,
specifically UV-b, to ground level. UV-b is an issue of concern because depletion of the stratospheric
ozone layer (i.e., ozone in the upper atmosphere) due to chlorofluorocarbons and other ozone-depleting
chemicals is associated with increased skin cancer and cataract rates. Currently, EPA is not able to
adequately quantify these effects for the purpose of valuing benefits for this policy.
Other EPA programs exist to address the risks posed by changes in UV-b associated with
changes in total column ozone. As presented in the Stratospheric Ozone RIA (EPA, 1992), stratospheric
ozone levels are expected to significantly improve over the next century as the major ozone depleting
substances are phased out globally. This expected improvement in stratospheric ozone levels is estimated
to reduce the number of nonmelanoma skin cancers (NMSC's) by millions of cases in the U.S. by 2075.
4.7 References
Abt Associates, Inc. 1992. The Medical Costs of Five Illnesses Related to Exposure to Pollutants,
Prepared for the U.S. Environmental Protection Agency, Office of Pollution Prevention and Toxics,
Washington, DC.
Abt Associates, Inc. 1995. Urban Ornamental Flams: Sensitivity to Ozone and Potential Economic
Losses. Prepared for the U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards; Research Triangle Park, TN.C,; July.
Abt Associates, Inc. 1998a. Selected Health and Welfare Benefits Methods for the NOx SIP Call RIA,
Prepared for the U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards; Research Triangle Park, N.C., September.
Abt Associates, Inc. 1998b. Agricultural Benefits Using AGSIMfor the NOx SIP Call Draft Report
Prepared for the U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards;
Research Triangle Park, N.C., September.
Chestnut, L. 1997. Draft Memorandum: Methodology for Estimating Values for Changes in
Visibility at National Parks. (April 15)
Costanza, R.; d'Arge, R.; de Groot, R.; Farber, S.; Grasso, M.; Hannon, B.; Limburg, K; Naeem, S.;
O'Neill, R.V.; Paruelo, J.; Raskin, R.G.; Sutton, P.; and van den Belt, M. (1997), The Value of the
World'sEcosystem Services and Natural Capital. Nature, Vol. 387: 253-259.
Page 4-65
-------
Crocker T.D., Horst R.L., Jr. 1981. Hours of Work, Labor Productivity, and Environmental Conditions:
a Case Study. The Review of Economics and Statistics 63:361-368.
Crocker, T.D. (1985), On the Value of the Condition of a Forest Stock. Land Economics 61(3):244-254.
Cropper, M.L, and A.J, Krupnick. 1990. The Social Costs of Chronic Heart and Lung Disease.
Resources for the Future Discussion Paper QE 89-16-REV.
Cummings, R., H. Burness and R. Norton, 1981. Methods Development for Environmental Control
Benefits Assessment, Volume V. Measuring Household Soiling Damages from Suspended Air
Particulates, A Methodological Inquiry, Report prepared for the U.S. Environmental Protection
Agency, Washington, D.C.
Dennis, R. 1998, Personal communication, NOAA Atmospheric Research Lab; Research Triangle Park,
N.C.; August 31.
DeSteiger, I.E.; Pye, J.M.; Love, C.S. (1990), Air Pollution Damage to U.S. Forests. Journal of
Forestry, 88-8: 17-22.
Dickie, M. et al. 1987. Reconciling Averting Behavior and Contingent Valuation Benefit Estimates of
Reducing Symptoms of Ozone Exposure (draft), as cited in Neumann, J.E., Dickie, M.T., and R.E.
Unsworth. 1994, Industrial Economics, Incorporated. Memorandum to Jim DeMocker, U.S. EPA,
Office of Air and Radiation. March 31.
Dockery, D.W., F.E. Speizer, D.O. Stram, J.H. Ware, J.D, Spengler, and E.G. Ferris, Jr. 1989. Effects
of Inhaiable Particles on Respiratory Health of Children. Am. Rev. Respir. Dis. 139: 587-594.
Empire State Electric Energy Research Corporation (ESEERCO). 1994. New York State
Environmental Externalities Cost Study. Report 2; Methodology. Prepared by: RCG/Hagler, Bailly,
Inc., November.
Fox, S, and Miekler, R.A, (1995), Impact of Air Pollutants on Southern Pine Forests, Ecological Studies
118; Springer-Verlag; New York.
Heck, W.W. and Cowling, E.B. (1997), The Need for a Long Term Cumulative Secondary Ozone
Standard—An Ecological Perspective. EM, January 1997; 23-33.
Industrial Economics, Incorporated (IBc). 1992. Memorandum to Jim DeMocker, Office of Air and
Radiation, Office of Policy Analysis and Review, U.S. Environmental Protection Agency. November 6,
Industrial Economics, Incorporated (lEc). 1993. Memorandum to Jim DeMocker, Office of Air and
Radiation, Office of Policy Analysis and Review, U.S. Environmental Protection Agency, September 30,
1993.
Industrial Economics, Incorporated (lEc). 1994. Memorandum to Jim DeMocker, Office of Air and
Radiation, Office of Policy Analysis and Review, U.S. Environmental Protection Agency, March 31.
Page 4-66
-------
Ito, K, and Thurston, G.D., 1996, Daily PMlO/Mortality Associations: An Investigation of At-Risk
Subpopulations. Journal of Exposure Analysis and Environmental Epidemiology 6(1): 79-225.
Kinney et al,, 1995. A Sensitivity Analysis of Mortality/PM 10 Associations in Los Angeles. Inhalation
Toxicology 7: 59-69.
Krupnick, A.J. and M.L. Cropper. 1992. The Effect of Information on Health Risk Valuations. Journal
of Risk and Uncertainty 5(2): 29-48.
Krupnick, A.J. and RJ. Kopp. 1988. The Health and Agricultural Benefits of Reductions in Ambient
Ozone in the United States. Resources for the Future Discussion Paper QE88-10, Washington, DC.
August.
Krupnick A.J., Harrington W., Ostro B. 1990. Ambient Ozone and Acute Health Effects: Evidence
from Daily Data. Journal of Environmental Economics and Management 18:1-18.
Loehman, E.T., S.V. Berg, A,A. Arroyo, R.A. Hedinger, J.M. Schwartz, ME, Shaw, R.W. Fahien, V.H.
De, R.P. Fishe, D.E. Rio, W.F. Rossley, and A.E.S. Green. 1979. Distributional Analysis of Regional
Benefits and Cost of Air Quality Control. Journal of Environmental Economics and Management 6;
222-243.
Manuel, E.H., R.L. Horst, K.M. Brennan, W.N. Lanen, M.C. Duff and J.K. Tapiero. 1982. Benefits
Analysis of Alternative Secondary National Ambient Air Quality Standards for Sulfur Dioxide and
Total Suspended Particulates, Volumes I-IV. Prepared for U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards, Research Triangle Park, NC. [Cited in ESEERCO, 1994].
Mathtech, Inc.. 1998. Regional Model Farm Benefit Estimation of Alternative Emission Controls for
the NOXSIP Call Prepared for Science Applications International Corporation. August, 1998.
McClelland, G., W. Schulze, D. Waldman, J. Irwin, D. Schenk, T. Stewart, L, Deck, and M. Thayer.
1991. Valuing Eastern Visibility: A Field Test of the Contingent Valuation Method. Prepared for
Office of Policy, Planning and Evaluation, U.S. Environmental Protection Agency. June.
Moolgavkar etal., 1995. Air Pollution and Daily Mortality in Philadelphia. Epidemiology 6(5): 476-484.
Moore, M.J., and W,K. Viscusi. 1988. "The Quantity-Adjusted Value of Life". Economic Inquiry
26(3): 369-388.
National Research Council (1995), Nitrate and Nitrite in Drinking Water, Subcommittee on Nitrate and
Nitrite in Drinking Water, National Academy Press; Washington, DC.
Ostro, B.D. 1987. Air Pollution and Morbidity Revisited: a Specification Test. J, Environ, Econ.
Manage, 14:87-98.
Ostro, 1995, Fine Participate Air Pollution and Mortality in Two Southern California Counties.
Environmental Research 70: 98-104.
Page 4-67
-------
Ostro B.D, and S. Rothschild. 1989. Air Pollution and Acute Respiratory Morbidity: An Observational
Study of Multiple Pollutants. Environmental Research 50:238-247.
Ostro, B.D., M.J. Lipsett, M.B. Wiener, and J.C, Seiner. 1991. Asthmatic Responses to Airborne Acid
Aerosols, American Journal of Public Health 81:694-702.
Ostro, B.D., M.J. Lipsett, J.K. Mann, H. Braxton-Owens, and M.C. White. 1995. Air Pollution and
Asthma Exacerbations Among African American Children in Los Angeles. Inhalation Toxicology,
Peterson, D.C.; Rowe, R.D.; Sehulze, W.D.; Russell, G.W.; Boyce, R.R.; Elliott, S.R.; Hurd, B. (1987),
Improving Accuracy and Reducing Costs of Environmental Benefit Assessments: Valuation of Visual
Forest Damages from Ozone. Prepared for the U.S. Environmental Protection Agency, Office of Air and
Radiation; Washington, D.C.; Cooperative Agreement #CR812054-02.
Pope, C.A., III, D.W. Dockery, J.D. Spengler, and M.E. Raizenne. 1991. Respiratory Health and PMi0
Pollution: a Daily Time Series Analysis. Am. Rev. Respir. Dis. 144: 668-674.
Pope, C.A., III, M.J. Thun, M.M. Namboodiri, D.W. Dockery, J.S. Evans, F.E. Speizer, and C.W.
Heath, Jr. 1995. Partieulate Air Pollution as a Predictor of Mortality in a Prospective Study of U.S.
Adults. Am. J, Respir, Crit, Care Med. 151: 669-674.
Portney, P.R. and J. Mullahy. 1990. Urban Air Quality and Chronic Respiratory Disease. Regional
Science and Urban Economics 20: 407-418.
Pye, J.M.; deSteiguer, J.E.; Love, C. (1988), Expert Opinion Survey on the Impact of Air Pollutants on
Forests of the USA. Proceedings of Air Pollution and Forest Decline; Interlaken, Switzerland; October.
Samet et al., 1996. Air Pollution and Mortality in Philadelphia, 1974 - 1988. Report to the Health Effects
Institute on Phase IB: Particle Epidemiology Evaluation Project, March 25, 1996 (draft, accepted for
publication).
Samet et al,, 1997. Partieulate Air Pollution and Daily Mortality: Analysis of the Effects of Weather and
Multiple Air Pollutants. The Phase IB Report of the Particle Epidemiology Evaluation Project. Health
Effects Institute, March 1997.
Schwartz, J. 1993. Partieulate Air Pollution and Chronic Respiratory Disease. Environmental
Research 62: 7-13.
Schwartz, J. 1994a, Air Pollution and Hospital Admissions in Elderly Patients in Birmingham, Alabama.
American Journal of Epidemiology 139:589-98.
Schwartz, J. 1994b. Air Pollution and Hospital Admissions for the Elderly in Detroit, Michigan.
American Journal of Respiratory Care Med 150:648-55.
Schwartz, J. 1994c. PMIO, Ozone and Hospital Admissions for the Elderly in Minneapoiis-St. Paul,
Minnesota. Archives of Environmental Health 49(5): 366-374.
Page 4-68
-------
Schwartz, J. 1995. Short Term Fluctuations in Air Pollution and Hospital Admissions of the Elderly for
Respiratory Disease, Thorax 50:531-538.
Schwartz, J. 1996. Air Pollution and Hospital Admissions for Respiratory Disease. Epidemiology17(1):
20-28.
Schwartz, J,, and D.W. Dockery. 1992. Increased mortality in Philadelphia associated with daily air
pollution concentrations. Am. Rev, Respir. Dis. 145: 600-604.
Schwartz, J. and R. Morris. 1995. Air Pollution and Hospital Admissions for Cardiovascular Disease in
Detroit, Michigan. Am. J, Epidemiol. 142: 23-35.
Schwartz, J., Dockery, D.W., Neas, L.M, Wypij, D., Ware, J.H., Spengler, J.D., Koutrakis, P., Speizer,
F.E., and Ferris, Jr., E.G. 1994. Acute Effects of Summer Air Pollution on Respiratory Symptom
Reporting in Children. Am.,/, Respir. Crit, Care Med, 150: 1234-1242,
Schwartz, J., Doekery, D., and L. Neas. 1996, Is Daily Mortality Specifically Associated With Fine
Particles? J. Air & Waste Man, Assoc. 46: 927-939,
Thurston, G.D. K. Ito, P.L. Kinneym, and M. Lippman. 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. Journal of Exposure Analysis and Environmental Epidemiology. 2 (4):429-450,
Thurston, G. K. Ito, C. Hayes, D. Bates, and M. Lippmann. 1994. Respiratory Hospital Admission and
Summertime Haze Air Pollution in Toronto, Ontario: Consideration of the Role of Acid Aerosols.
Environmental Research 65: 271 -290.
Tolley, G.S. et al. 1986. Valuation of Reductions in Human Health Symptoms and Risks. University
of Chicago. Final Report for the U.S. Environmental Protection Agency. January.
U.S. Department of Commerce, Economics and Statistics Administration. 1992. Statistical Abstract of
the United States, 1992: The National Data Book, 112th Edition, Washington, D,C,
U.S. Department of Commerce, Bureau of Economic Analysis. BEA Regional Projections to 2045: Vol.
1, States. Washington, D.C. U.S. Govt. Printing Office, July 1995,
U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National
Center for Health Statistics. 1994. Vital Statistics of the United States, 1990, Volume II-Mortality.
Hyattsville, MD.
U.S. Environmental Protection Agency, 1994. Documentation for Oz-One Computer Model (Version
2.0). Office of Air Quality Planning and Standards. Prepared by: Mathtech, Inc., under EPA Contract
No. 68D30Q30, WA 1-29. August.
Page 4-69
-------
U.S. Environmental Protection Agency, 1996a. Regulatory Impact Analysis for Proposed Paniculate
Matter National Ambient Air Quality Standard, Prepared by: Innovative Strategies and Economics
Group, Office of Air Quality Planning and Standards, Research Triangle Park, N.C. December.
U.S. Environmental Protection Agency, 1996b, Regulatory Impact Analysis for Proposed Ozone
National Ambient Air Quality Standard. Prepared by: Innovative Strategies and Economics Group,
Office of Air Quality Planning and Standards, Research Triangle Park, N.C. December.
U.S. Environmental Protection Agency, 1996c. Air Quality Criteria for Ozone and Related Photochemical
Oxidants, Office of Research and Development; Office of Health and Environmental Assessment;
Research Triangle Park, N.C.; EPA report nos. EPA/600/P-93/G04aF-eF.
U.S. Environmental Protection Agency, 1996d. Air Quality Criteria for Particulate Matter. Office of
Research and Development; Office of Health and Environmental Assessment; Research Triangle Park,
N.C.; EPA report nos. EPA/600/P-95/001aF; April.
U.S. Environmental Protection Agency, 1996e, Review of the National Ambient Air Quality Standards for
Ozone: Assessment of Scientific and Technical Information. Office of Air Quality Planning and
Standards, Research Triangle Park, N.C.; EPA report no. EPA/4521R-96-007.
U.S. Environmental Protection Agency, 1996f. Review of the National Ambient Air Quality Standards for
Particulate Matter: Assessment of Scientific and Technical Information. Office of Air Quality Planning
and Standards, Research Triangle Park, N.C,; EPA report no. EPA/4521R-96-013.
U.S. Environmental Protection Agency, 1997a. Regulatory Impact Analysis for Particulate Matter
and Ozone National Ambient Air Quality Standards and Proposed Regional Haze Rule. Prepared
by: Innovative Strategies and Economics Group, Office of Air Quality Planning and Standards, Research
Triangle Park, N.C. July.
U.S. Environmental Protection Agency, 1997b. Technical Support Document for the Regulatory
Impact Analysis for Paniculate Matter and Ozone National Ambient Air Quality Standards and
Proposed Regional Haze Rule. Prepared by: Innovative Strategies and Economics Group, Office of Air
Quality Planning and Standards, Research Triangle Park, N.C. July.
U.S. Environmental Protection Agency, 1997c. The Benefits and Costs of the Clean Air Act, 1970 to
1990, Prepared for U.S. Congress by U.S. EPA, Office of Air and Radiation/Office of Policy Analysis
and Review, Washington, D.C. (April, 1997 - Draft)
U.S. Environmental Protection Agency, 1998. The Regional NOx SIP Call & Reduced Atmospheric
Deposition of Nitrogen: Benefits to Selected Estuaries, September, 1998.
Viscusi, W.K. 1992. Fatal Tradeoffs: Public and Private Responsibilities for Risk, (New York:
Oxford University Press).
Page 4-70
-------
Ward, M. H.; Mark, S. D,; Cantor, K.P.; Weisenburger, D.D.; Correa-Villasenor, A.; Zahm, S.H. (1996),
Drinking Water Nitrate and the Risk of Non-Hodgkin's Lymphoma. Epidemiology 7:465-471;
September.
Watson, W. and J. Jaksch. 1982. Air Pollution; Household Soiling and Consumer Welfare Losses.
Journal of Environmental Economics and Management. 9: 248-262.
Weisel, C.P., R.P. Cody., and P.J. Lioy. 1995, Relationship Between Summertime Ambient Ozone
Levels and emergency Department Visits for Asthma in Central New Jersey. Environmental Health
Perspectives, 103(Suppl. 7): 97-102.
Whittemore AS, Korn EL. 1980. Asthma and Air Pollution in the Los Angeles Area. American
Journal of Public Health 70:687-696.
Page 4-71
-------
-------
Chapter 5. BENEFIT-COST COMPARISONS
This Regulatory Impact Analysis (RIA) provides cost, economic impact, and benefit estimates
that are potentially useful for evaluating alternative policy options for the NOx SIP call. Benefit-cost
analysis provides a systematic framework for assessing and comparing such alternatives. According to
economic theory, the efficient alternative maximizes net benefits to society (i.e., social benefits minus
social costs). However, there are practical limitations for the comparison of benefits to costs in this
analysis. This chapter also discusses the key limitations and uncertainties associated with the benefit and
cost estimates. Nonetheless, if one is mindful of these limitations, the relative ordering and magnitude of
the benefit-cost comparisons presented here can be useful policy information.
5,1 Summary of Cost Estimates
This section provides a summary of cost results presented in Volume 1 of this RIA. Table 5-1
summarizes the total annual control cost estimates developed in this analysis for the year 2007 for a
selected set of regulatory alternatives that closely approximate the alternatives analyzed in the benefits
analysis. These costs include potential changes in the nationwide costs of electricity generation for the
network of electricity generating sources (EQUs), direct control costs to potentially affected non-EGU
sources, emissions monitoring costs associated with the administrative costs associated with monitoring
The majority of the total annual cost is due to control on electricity generating units (EGUs) for each
alternative.
Table S-l
Estimated Total Annual Cost of NOx SIP Call Alternatives in 2007
Regulatory Alternative
0.1 2 Trading
60%/S5,000
0.1 5 Trading
6G%/$5,000
Regionally 1
60%/$5,OGO
0,20 Trading
60%/$5,000
0.25 Trading
60%/$5,000
Total Annual Costs
(million 1990S)
$2,128
11,660
$1,400
$1,230
$925
Page 5-1
-------
5.2 Summary of Benefits Estimates
Table 5-2 summarizes the total annual benefits developed in this analysis for the year 2007 for the
"plausible range" of assumptions. Not all possible benefits are quantified and monetized in this analysis.
Potential benefit categories that have not been quantified and monetized are listed in Table 4-2 in Chapter
4 of Volume 2 of this R1A.
Table 5-2
"Plausible Range" of Annual Quantified Benefits Estimates
for NOx SIP Call Alternatives in 2007
(million 1990S)8
Regulatory Alternative
0.1 2 Trading
60%/$5,000
0.15 Trading
60%/$5,000
Regionally 1
6Q%/$5,000
0,20 Trading
fiO%/$5,000
0.25 Trading
60%/$5,000
Annual Quantified Benefits-"Low"
Assumption Set
$2,888
$1,100
$1,138
$1,156
$777
Annual Quantified Benefits--
"High" Assumption Set
$5,473
$4,170
$3,457
$3,543
$2,971
" Not all possible benefits are quantified and monetized in lliis analysis. Potential benefit categories that have not been quantified and
monetized are listed in Table 4-2 in Chapter 4 of this R1A.
5,3 Summary of Net Benefits
Table 5-3 summarizes the total annual quantifiable net benefits for NOx SIP call regulatory
alternatives for the year 2007. There are several conclusions that can be drawn from Table 5-3.
« For the "High" assumption set, monetized net benefits are positive and substantial for all
regulatory alternatives.
• As modeled, Regionally 1 is an inferior alternative, i.e., even though 0.20 Trading is less stringent
it achieves greater benefits at lower total costs.
* Net benefits are greatest at the most stringent regulatory alternative evaluated, i.e., 0.12 Trading.
For the "High" assumption set, net benefits are approximately 33 percent higher for the 0.12
Trading relative to the 0.15 Trading alternative. For the "Low" assumption set, net benefits are
positive only for the 0.12 Trading alternative.
Page 5-2
-------
While net benefits are negative for the "Low" assumption set for all but the 0.12 Trading
alternative, it is important to remember that while all of the costs are included, many benefit
categories could not be quantified. In addition, the "Low" assumption set estimate assumes that
there are no reductions in premature mortality associated with ozone reductions. Relaxing this
one assumption would result in positive net benefits for alt alternatives at the low end.
Table 5-3
Estimated Annual Quantified Net Benefits"
for NOx SIP Call Alternatives in 2007
(million 1990$)
Regulatory
Alternative
0.1 2 Trading
60%/$5,000
0,15 Trading
60%/$5;GOO
Regionally 1
60%/$5,OG0
0.20 Trading
60%/$5,000
0.25 Trading
60%/$5,OQO
Quantified Net Benefits—
"Low" Assumption Set
$760
($560)
($262)
($74)
($148)
Quantified Net Benefits—
"High" Assumption Set
$3,345
$2,510
$2,057
$2,313
$2,046
* Calculated as quantified benefits minus costs. Not all possible benefits are quantified and monetized in this analysis. Potential benefit
categories that have not been quantified and monetized are listed in Table 4-2 in Chapter 4 of this volume of the RIA.
5.4 Limitations to the Benefit-Cost Comparison
Cost-benefit analysis provides a valuable framework for organizing and evaluating information on
the effects of environmental programs. When used properly, cost-benefit analysis helps illuminate
important potential effects of alternative policies. However, not all potential costs and benefits can be
captured in any analysis, and there always the issue of how much technological changes will lower future
pollution abatement costs, or change the nature of compliance actions by the regulated community over
time, EPA is generally able to estimate reasonably well the costs of pollution controls based on today's
control technology and assess the important impacts when it has sufficient information for its analysis.
EPA is developing an increasing ability to estimate benefits associated with changes in emissions, but
EPA believes that there are many important benefits that it can not quantify or monetize that are
associated with the NOx SIP call, including many health and welfare effects. Potential benefit
categories that have not been quantified and monetized are listed in Table 4-2 in Chapter 4 of this volume
of the RIA and should be remembered in comparing the above quantitative benefits.
Several other important limitations deserve to be mentioned:
Page 5-3
-------
The state of atmospheric modeling is not sufficiently advanced to provide a workable
"one atmosphere" model capable of characterizing ground-level pollutant exposure for all
pollutants of interest (e.g., ozone, particulate matter, carbon monoxide, nitrogen
deposition, etc). Therefore, EPA must employ several different pollutant models to
characterize the effects of alternative policies on relevant pollutants. Also, not all
atmospheric models have been widely validated against actual ambient data. In
particular, since a broad-scale monitoring network does not yet exist for fine particulate
matter (PM25), atmospheric models designed to capture the effects of alternative policies
on PM2 5 are not fully validated. The Agency has chosen the best available models for
the application needs of this RIA and tried to make the most reasonable assumptions
possible in using them for predicting air quality changes. Limitations are noted in
appropriate areas of the RIA.
There are limitations in some aspects of the data that are available to perform these
analyses. These limitations have been identified along the way in this RIA. While they
exist, EPA believes that it has used all models and assumptions in this analysis in a
reasonable way based on the available evidence. Qualitative and more detailed
discussions of the above and other uncertainties and limitations are included in the
analysis. Where information and data exists, quantitative characterizations of these
uncertainties are included. An illustrative example of how one aspect of uncertainty can
be quantified is provided in Appendix A of Volume 2. However, data limitations prevent
an overall quantitative estimate of the uncertainty associated with final estimates.
Nevertheless, the reader should keep all of these uncertainties and limitations in mind
when reviewing and interpreting the results.
Another dimension adding to the uncertainty of the results is the potential for pollution
control innovations that can occur over time. For the NOx SIP call, EPA expects that
the most significant costs of this regulation (i.e., the costs associated with installation and
operation of NOx pollution control equipment at coal-fired electricity generating units
throughout the SIP call region) will occur by May 2003, The Agency is aware of some
innovations that equipment vendors are considering now for application at some units
before 2003 that are not part of the cost analyses presented in Volume 1, These
innovations include the possible use of SCR and SNCR in hybrid technologies, or
improved combustion controls beyond what vendors have installed in the past that could
be used with and without the addition of post-combustion control technology. It is
impossible to anticipate exactly how much of an impact, if any, these new technologies
may have in lowering the compliance costs for the NOx SIP call in the future. Their
possible influence can only be recognized.
There is also the uncertainty over future costs due to the flexibility afforded by an
emissions cap-and-trade program that EPA is encouraging the States to set up under this
rule. The analysis that EPA has completed to date has been fairly conservative-the
analysis of the electric power industry and large industrial boilers and combustion turbines
assumes these sources operate under separate trading programs. In reality, they should
enter the same trading pool and there should be greater efficiency resulting from their
ability to trade NOx emissions allowances with each other. There is also the possibility of
Page 5-4
-------
unforeseen innovation, which a cap-and-trade program fosters, since it allows the
regulated community to work out the best approaches to future compliance. Therefore,
the Agency believes that its cost analysis is a reasonably conservative estimate of the
future compliance costs that will occur if States enter into the trading program that the
Agency has described in the Model NOx Budget Trading Rule as part of the NOx SIP
call rule. If some States do not enter the program, any inefficiencies that result for the
regulated community in those States should not be viewed as a cost of this rule.
Despite the above limitations and uncertainties, EPA believes that the analysis provided in this
report provides the Agency with a basis for believing that in the year 2007, benefits resulting from the
regulatory alternatives that EPA analyzed for NOx SIP call will be up to two and one-half times costs.
5.5 References
U.S. Environmental Protection Agency, 1997. Proposed Ozone Transport Rulemaking Regulatory
Analysis, Office of Air and Radiation, Washington, D.C. September, 1997.
Page 5-5
-------
-------
APPENDICES
to the
REGULATORY IMPACT ANALYSIS
FOR THE NOx SIP CALL, FIP, AND
126 PETITIONS
Volume 2: Health and Welfare Benefits
-------
This page intentionally left blank
-------
Appendix A. QUANTIFIED UNCERTAINTY IN HEALTH AND WELFARE BENEFITS
A.1 Overview
Chapter 4 presents point estimates of the monetary benefits associated with each health and
welfare endpoint. For most endpoints, an estimate of the statistical uncertainty range based on measured
variability in the underlying health effects and valuation components of the analysis can also be computed.
Uncertainty regarding other aspects of the analysis (such as emissions and resulting air quality) is not
included in the uncertainty analysis, resulting in a likely underestimate of the overall uncertainty of the
monetized benefits for each category.
The two sources of uncertainty that are quantified in the NOx SIP call benefits analysis are the
uncertainty about the concentration-response functions (and thus about changes in incidence) and the
uncertainty about mean willingness to pay for each unit change in incidence (i.e., unit dollar values). The
total dollar benefit associated with a given endpoint depends on how much the endpoint will change if a
given NOx SIP call alternative is implemented (e.g., how many premature deaths will be avoided) and
how much each unit of change is worth (e.g., how much a premature death avoided is worth). Not all
endpoints have quantified uncertainty for both the concentration-response function and valuation function
The uncertainty about each component is characterized by a distribution of values that the
component might have. This distribution is essentially a Bayesian posterior distribution, based on the
available information. A distribution of possible incidence changes and a distribution of possible unit dollar
values for each endpoint is constructed from available information whenever possible. The uncertainty
about the true incidence change (or the true unit dollar value) for a given endpoint is expressed as a 90
percent credible interval. This is the interval from the fifth percentile point to the ninety-fifth percentile
point of the Bayesian posterior distribution of incidence changes (or unit dollar values) for that endpoint.
The 90 percent credible interval is a "credible range" within which, according to the available information
(embodied in the Bayesian posterior distribution of possible values), the true value lies with 90 percent
probability.
The uncertainty surrounding estimates of total monetary benefits for each endpoint is similarly
characterized by a distribution of possible values, the fifth and ninety-fifth percentile points of which
comprise the 90 percent credible interval of total monetary benefits for the endpoint. The distribution of
total monetary benefits for an endpoint is generated from the distribution of incidence changes and the
distribution of unit dollar values for that endpoint, using Monte Carlo techniques. In this procedure, on
each of many iterations, a value is randomly drawn from the incidence distribution and a value is randomly
drawn from the unit dollar value distribution, and the total dollar benefit for that iteration is the product of
the two.1 If this is repeated for many (e.g., thousands of) iterations, an estimate of the distribution of total
! This method assumes that the incidence change and the unit dollar value for an endpoint are stochastically
independent.
Page A-l
-------
dollar benefits associated with the endpoint is generated2. The mean of this Monte Carlo-generated
distribution is presented as the point estimate of total monetary benefits for the endpoint. As the number
of Monte Carlo draws gets larger and larger, the Monte Carlo-generated distribution becomes a better
and better approximation to the underlying Bayesian distribution of total monetary benefits. In the limit, it
is identical to the underlying distribution, and its mean, presented as the point estimate, is identical to the
mean of the underlying distribution of total monetary benefits for the endpoint.
The distributions of unit dollar values for those health and welfare endpoints considered in this
uncertainty analysis, and the means of those distributions, are given in Table A-l below. In addition, the
means and 90 percent credible intervals (the fifth and ninety-fifth percentile points of the distributions of
possible values) of avoided incidences and the corresponding means and 90 percent credible intervals of
total monetary benefits associated with these endpoints can be found in the results technical support
document (Abt Associates, 1998a).
A.2 Underlying Sources of Uncertainty
For most health endpoints (with the exception of ozone-related mortality and short-term PM-
related mortality), the concentration-response function is obtained from a single epidemiological study.
For all of these studies, the uncertainty about the unknown parameter in the concentration-response
function is characterized by a normal distribution with mean equal to the estimate of the parameter value
reported in the study and standard deviation equal to the standard error of the estimate reported in the
study. To the extent that avoided incidence is a linear function of this concentration-response function
parameter, the distribution of avoided incidence will also be normally distributed.3
For ozone-related mortality and short-term PM-related mortality, the distribution of incidence
changes is based on a pooling of the information in several concentration-response functions. In the case
of short-term PM-related mortality, the input components to the concentration-response (C-R) functions
estimated in the studies (e.g., functional forms, pollutant averaging times, study populations) are all the
same or very similar, so that a pooled, "central tendency" C-R function can be derived from multiple
study-specific C-R functions. For ozone-related mortality, however, the pollutant averaging time is not the
same across all studies. Some of the four studies measured daily 1-hour maximum ozone concentrations
while others measured daily (or some other) average ozone concentrations. It is therefore not possible to
pool the C-R functions to derive a central tendency "pooled" C-R function for ozone-related mortality.
Instead, using the ozone data appropriate to each study (either one-hour daily maxima or daily averages),
national ozone mortality incidence distributions are derived corresponding to the C-R function from each
study, and these study-specific national incidence distributions are then pooled. That is, the pooling of
results is done in "national incidence space" rather than in "ozone coefficient space," For a more detailed
discussion of the methodology of pooling results from studies, see the report titled Selected Health and
Welfare Benefits Methods for the NOx SIP Call R1A (Abt, Associates, September 1998b).
2 To improve computer efficiency, a Latin Hypercube technique is actually used for the incidence change
distribution when implementing all phases of the Monte Carlo analysis,
5 The concentration-response functions are almost linear functions of the parameter.
Page A-2
-------
Construction of distributions of unit dollar values, or mean willingness to pay (MWTP) for a case
avoided, is often not as straightforward to describe. Estimates of MWTP can be complicated functions of
estimated parameters, for which information about the statistical distributions are not available in published
studies. The assumed distributions for MWTP for each endpoint are listed in Table A-l. For a more
complete description of the underlying studies and derivation of the distributions of MWTP, see Chapter 4
and the benefits technical support document (Abt Associates, 1998b). For some endpoints, while
uncertainty is recognized and a range of possible values is available, there is insufficient information to
construct
Page A-3
-------
Table A-l
Point Estimates and Assumed Distributions of MWTP for Health and
Welfare Endpoints in the NOx SIP Call Analysis
Endpoirtt
Mortality
Chronic bronchitis
URS (as defined by Pope et al.,
1991)
LRS (as defined by Schwartz et al.,
1994)
"Presence of any of 1 9 acute
respiratory symptoms"
Acute Bronchitis
Minor Restricted Activity Days
(MRADs)
Work Loss Days
Worker Productivity
Visibility - Residential
Visibility - Recreational
Consumer Cleaning Cost Savings
Point Estimate of MWTP
$4,800,000
$260,000
$19
$12
$18
$45
$38
S83
$ I/ worker/ 10% change in O3
$14 per unit change in dv
$6,50 per unit change in dv
(in-region)
$4 per unit change in dv
(out-of-region)
$2.52 per jig/m3 change in
PM10 per household
Derived Distribution of the Estimate of
MWTP (1990S)
A Weibull distribution, Std, Dev, = $3,24
million
A Monte Carlo-generated distribution,
based on three underlying distributions,
as described in the technical support
document.
Continuous uniform distribution over the
interval [$7,00, $32,72]
Continuous uniform distribution over the
interval [$5.27, $18.57]
Continuous uniform distribution over the
interval [$0,00, $36,62]
Continuous uniform distribution over the
interval [$13.29, $76.74]
triangular distribution centered at $38.37
on the interval [$ 1 5.72, $6 1 .02]
N.A."
N.A.
Triangular distribution centered at $14 on
the interval [$8, $21]
Normal distribution with std. dev. equal
to 0.42 (in-region) and 0.10 (out-of-region)
Beta distribution with, std, dev,=$1.00 on
the interval [$1 .26, $1 0.08]. The shape
parameters of this distribution are «=1 ,2
and(J=7.3.
"N.A." indicates that a distribution is not available.
Page A-4
-------
anything other than a uniform distribution of unit dollar values, which assumes that each point in the range
is equally likely.
For agricultural, forestry, and nitrogen deposition benefits, there are no distributions or ranges of
unit values available. Variation in the endpoints occurs due to changes in the underlying assumptions in
the models generating the benefit estimates, i.e. sensitivity of cultivars to ozone for agricultural benefits.
Thus, for these endpoints, uncertainty ranges are not reported. Sensitivity to changes in the underlying
assumptions can be examined through the plausible range approach, where assumptions can be grouped to
form a range with low and high estimates.
A.3 Quantified Uncertainty for Ozone-related Benefits
The quantification of the uncertainty about the magnitude of the ozone mortality relationship is a
very important issue in the economic benefits estimation. While the growing body of epidemiological
studies suggests that there in a positive relationship between ozone and premature mortality, it is still
unclear whether the apparent ozone effect on mortality is real. There is a diversity of published results
and substantial measured uncertainty within each study. This high degree of uncertainty has lead to some
counterintuitive "results." Based on the meta-analysis generated distribution of avoided ozone-related
cases of premature mortality corresponding to the NOx SIP call estimated future air quality, there is
approximately a 13 percent probability that there is a negative relationship between ozone exposure and
premature mortality (i.e., that elevated ozone prevents premature mortality). This "result" should be
interpreted with caution, however. It is biologically implausible that elevated ozone levels are beneficial to
human health. The portion of the estimated incidence change distribution in the negative range is most
likely the result of random error in the estimation of the ozone coefficients in concentration-response
functions and/or the result of modeling misspecification (the underlying models do not prevent negative
results a priori, and the estimated coefficients are asymptotically normal, which results in a negative
lower tail of the distribution). By construction, the meta-analysis distribution incorporates both between-
location variability and within-location sampling error. As more studies become available, and, in
particular, as newer studies incorporate information from longer periods of time and therefore have results
based on more observations, the sampling error component of the meta-analysis distribution will decrease.
As this occurs, the meta-analysis distributions will better approximate the underlying distributions of which
they are estimates, and the portions of the distributions in the negative range are likely to diminish
accordingly.
A.4 Quantified Uncertainty for PM-related Benefits
For household soiling damage and visibility, no point estimates or distributions are presented for
avoided incidences. This is because PM-related household soiling is directly valued on a per household
basis, rather than measuring some unit of incidence (such as hours lost) and multiplying by a value per
unit. Visibility is valued on a constant percentage deciview change per household, so there are no avoided
incidences. The correct unit measure is percent change in deciview, which is then input into a valuation
function to get value per household, which is then summed over all households in the NOx SIP call region.
Thus, uncertainty is measured in the valuation stage, but not in the generation of changes in visibility.
Page A-5
-------
A.5 Statistical Uncertainty and Plausible Ranges
The tables of benefit estimates presented in Chapter 4 represent monetary benefits estimates for
five NOx SIP call regulatory alternatives under the "Low" and "High" sets of assumptions regarding the
PM threshold level, ozone mortality, agricultural benefits, and the PM air quality model. Benefits
estimates associated with the "High" and "Low" assumption sets have corresponding statistical
uncertainty ranges as well, with 5th percentile, mean, and 95"' percentile estimates of benefits. As
discussed in Chapter 4, the range of values from the mean of the "Low" assumption set to the mean of
the "High" assumption set represents a "plausible range" across the different assumptions. However, this
range does not provide information on the likelihood of any set of assumptions being the correct one.
Thus, while the plausible range indicates the sensitivity of benefits to the various assumptions, it does not
express the uncertainty associated with any particular benefits estimate. To understand the uncertainty
associated with a particular estimate, it is necessary to know both the underlying assumption set and the
statistical distribution around the estimate determined by the variance of the underlying concentration-
response functions and valuation functions.
A.6 References
Abt Associates, Inc. 1998a, Benefit Analysis Results of Selected Health and Welfare Endpointsfor
the NOx SIP Call RIA, Prepared for the U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards; Research Triangle Park, N.C, September.
Abt Associates, Inc. 1998b. Selected Health and Welfare Benefits Methods for the NOx SIP Call RIA,
Prepared for the U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards; Research Triangle Park, N.C,, September.
Page A-6
-------
Appendix B, SUMMARY OF POPULATION-WEIGHTED AIR QUALITY METRICS
This appendix summarizes the predicted air quality changes used in the benefits analyses for this
RIA, weighted by population. The population-weighted air quality changes are calculated for several
"metrics," or measures of air quality based on results of the air quality modeling described in Chapter 3.
For ozone, these metrics include 1-hour and 8-hour average concentration predictions above the level of
the respective health standards. For PM, the metrics include annual mean PM2 s and annual mean PM!0
predictions above the level of the respective health standards. The metrics are calculated for the total
population and for various subpopulations, including minority groups (represented by the Census Bureau's
"non-white" category), the elderly (65 years of age and older), children (under 18 years old), and the low
income group (1990 annual income under $13,359 for a family of 4). The air quality changes cover the
entire area of each modeling domain; data is presented for the entire modeling domain (37 States & D.C.)
and for the SIP Call region states. The population-weigh ted metric for modeled visibility degradation is
presented in Chapter 2, Section 2,6, rather than in this appendix. A population-weighted metric is not
calculated for modeled nitrogen deposition changes. Additional detail on all population-weighted metrics
can be found in the report titled Air Quality Estimation for the NOx SIP Call RIA (Abt, Associates,
September 1998).
PageB-l
-------
Table B-l
2007 Population-Weighted Sum of 1-Hour Ozone Predictions Above 124 ppb*:
Adjusted and Extrapolated UAM-V Results
Population
Percent Change from Base Case
0.25
Trading
0.20
Trading
Regionally
I
0,15
Trading
0,12
Trading
SIP Call States
All Populations
Non-White
Under 18
65 and over
Low Income
-50,9
-48,9
-50.7
.49,2
-58.5
-63.0
-61,6
-62.8
-61.3
-69.0
-66,6
-64.0
-66.2
-65,5
-71.4
-69.4
-67.8
-69.2
-68.4
-73.7
-73.9
-73.7
-73.8
-73.1
-78.0
37 States & D,C.
AH Populations
Non-White
Under 18
65 and over
Low Income
-19.2
-14.7
-17.3
-25,1
-18.3
-24.1
-18,9
-21.7
-31.5
-22.1
-25,3
-19,5
-22.8
-33.5
-22.7
-26.0
-14.7
-17.3
-25.1
-18.3
-27.3
-21.4
-24.5
-36.6
-23,6
The 1-hour ozone standard allows an average of 1 excecdancc above 120 ppb (rounded to the nearest ppb) over a 3 year period.
This analysis does not predict three years worth of air quality, and is therefore no! directly comparable to the official standard.
Page B-2
-------
Table B-2
2007 Population-Weighted Sum of 8-Hour Average Ozone Predictions Above 84 ppb":
Adjusted and Extrapolated UAM-V Results
Population
Percent Change from Base Case
0.25
Trading
0.20
Trading
Regionality
1
0,15
Trading
0,12
Trading
SIP Call States
All Populations
Non-White
Under 18
65 and over
Low Income
-31.0
-26.8
-31.4
-30.8
-38,8
-39.9
-35.4
-40,3
-39.7
-48,3
-44.0
-39.5
-44.3
-43.9
-54.3
-46.8
•42.3
-47.2
-46.7
-56,7
-51.0
-46.8
-51.5
-50.9
-60.4
37 States & D.C
All Populations
Non- White
Under 18
65 and over
Low Income
-24.1
-20.0
-23,6
-26.0
-18.3
-31.2
-26,4
-30.4
-33.6
-22.1
-34.3
-29.3
-33.3
-37.2
-22.7
-36.5
-31.4
-35.5
-39,5
-22,5
-39.6
-34.6
-38.5
-42.9
-23.6
a The 8-hour ozone standard of 80 ppb (rounded lo lite nearest ppb} is based on each year's 4lli highest daily mazimum 8-hour average
ozone concentration averaged over a 3 year period. This analysis does not predict three years worth of air quality, and is therefore
not directly comparable lo the official standard.
Page B-3
-------
Table B-3
2007 Population-Weighted Sum of Annual Mean PM,,5 Predictions Above 15.04 ng/m5":
RPM Results
Population
Percent Change from 2007 Base Case
0,25
Trading
0.20
Trading
Regionally
1
0.15
Trading
0.12
Trading
SIP Call States
All Populations
Non-White
Under 1 8
65 and over
Low Income
-4.6
•4,5
-4,6
-4.5
-5,4
-6.0
-6.0
-6.0
-5.9
-8.7
-6.8
-6.5
-6.8
-6.8
-11.4
-6.7
-6.3
-6.8
-6.5
-6.2
-14.5
-14.8
-14,5
-14,3
-19.1
37 States & D.C.
All Populations
Non-White
Under 1 8
65 and over
Low Income
-3.0
-3.8
-3,0
-2.8
-4.4
-4,1
-5.1
-4.1
-3.9
-7.2
-4.5
-5.5
-4.5
-4.4
-9.3
-4.6
-5.4
-4.6
-4.3
-5.2
-10.0
-12.8
-9.9
-9.5
-15.9
* The PM2,5 annual mean standard of 15 ng/m3 (rounded to (he nearest l/10th) averaged over a 3 year period. This analysis is not
based on an extensive network of actual PM2 5 observations and does not predict three years worth of air quality, and is therefore not
directly comparable to the offlctai standard.
Page B-4
-------
Table B-4
2007 Population-Weighted Sum of Annual Mean PM,d Predictions Above 50,4 jig/m3 *:
RPM Results b
Population
All Populations
Non-White
Under 1 8
65 and over
Low Income
Percent Change from Base Case
0.25
Trading
-0.0
-0.0
-0,0
-0.0
-0.0
0.20
Trading
-0.0
-0.0
-0.0
-0.0
-0.0
Regionality
1
-0.0
-0.0
-0.0
-0.0
-0.0
0.15
Trading
-0.0
-0.0
-0.0
-0.0
-0.0
0.12
Trading
-0.0
-0.0
-0.0
-0.0
-0.0
• The PMjK, annual mean standard of SO ng/m3 (rounded to the nearest ug) averaged over a 3 year period. This analysis does not
predict three years worth of" air quality, and is therefore not directly comparable to the official standard,
b All locations have predictions below the PM10 standard, hence (here are no reductions in exposures relative a 50 ug/nr1 annual mean
threshold.
Page B-5
-------
-------
Appendix D. NONLINEAR CHEMISTRY AND FINE PARTICLE PRODUCTION
Evaluating the effect of major reductions in emissions of nitrogen oxides on particles in made
more difficult because NOx plays an important role in the atmospheric chemistry, including formation of
nitrate, organic, and sulfate particles. The extent to which NOx reductions reduce fine particle levels is
dependent on a number of factors that vary with time and location, including the concentration of key
reactive gases and particles, as well as emissions and meteorology. The following discussion provides
some background on the nature of these chemical reactions as a basis for understanding the relevance of
non-linear modeling results..
D.I Background
The hydroxyl radical, OH, is a major oxidizing specie in the atmosphere. The largest single
source of OH is the breakdown of ozone, O3, by ultraviolet light in the presence of water. Clearly,
reduction of NOx can affect the production of ozone and OH radical in particular locations. Sulfate, a
major component of fine particles, is formed by the oxidation of SO2 through two main pathways; gas-
phase oxidation and aqueous-phase oxidation (in water droplets). Sulfate is formed in the gas phase by
oxidation of SO2 by OH. This process does not use up OH, by and large, so a change in SO2 produces
approximately an equivalent change in SO4. Also, an increase/decrease in OH will cause an
increase/decrease in gas-phase produced SO4,
Sulfate (SO4) is also formed in the aqueous phase in cloud droplets. This occurs by conversion of
SO2 by hydrogen peroxide, H2O2, and O3. The vast proportion of the aqueous-phase SO4 is produced by
H2O2 for eastern North American conditions. This process does use up oxidants. If SO2 is very high,
then there can be oxidant-limited conditions, i.e., the nonlinearity that was of concern in NAPAP with
respect to the effectiveness of acid rain controls. Under oxidant-limited conditions, the production of SO,,
will be mostly determined by the availability of the oxidant, not by the availability of SO2. Modeling
analysis with RADM estimates that a majority of the eastern U.S. sulfate comes from the aqueous-phase
oxidation of sulfur dioxide, the second pathway.
Nitric acid, HNO3, the precursor to aerosol nitrate, in formed by NO2 combining with OH. The
formation of HNO3 is a termination reaction and uses up both OH and N02. This reaction is part of the
photochemical process that accounts for the production of O3.
The radicals that are produced during any one day all disappear or terminate during that day in a
matter of tens of minutes. Production equals termination in the photochemical process. There are two
important pathways of termination that are in constant competition:
a) OH combines with NO2, taking out one radical. If there is a lot of NOx around, this pathway
out competes and inhibits the availability of radicals by taking so many OH's out of the action so
quickly. If NOx is scarce, then this pathway cannot compete as well and a second termination
pathway becomes most important.
b) In the second main termination pathway, OH combines with itself, to form H2O2; hydroxy
radicals combine with peroxy radicals to form organic peroxides. Two radicals are taken out.
PageD-l
-------
The relative fraction of nitrate that exists in the particle phase (as opposed to vapor phase nitric
acid or ammonium nitrate) depends in turn upon the relative concentration of acid sulfate species and
ammonia. In areas with high acid sulfate concentrations (e.g. the eastern U.S. in the summer), nitrate
tends to occur in the vapor phase and reductions of NOx emissions could not result in a large reductions in
fine particle nitrates. Where sulfate levels are much then nitrate particle levels would be higher.
D.2 What this means for NOx Emissions and Sulfate
In urban areas with relatively high NOx emissions radical formation and propagation is inhibited.
If NOx emissions are reduced by small to moderate amounts, O3 and OH will increase. Note, the
response for average O3 can be different than the response for peak O3. If SO2 is available, then where
the OH increases, an increase in the amount of SO2 will be oxidized to S04 in the gas phase, This
produces a "nonlinear" response in SO,, (increase) to a reduction in NOx.
In rural areas with relatively models NOx levels, O3 and OH will decrease when NOx is reduced.
Percentage-wise, the decrease in 03 and OH will less than NOx and NO2. The ratio of OH to NO2 will
therefore increase and more of the OH budget will terminate as H2O2. This may or may not lead to an
increase in H2O2. In the RADM simulations there was an increase in H2O2 in certain areas, especially
over Ohio and western Pennsylvania. If there is excess SO2, then where H202 increases, in increased
amount of SO2 will be converted to SO4 in the aqueous phase. This produces a "nonlinear" response in
SO4. The larger and more pervasive source of the nonlinearity in these model results seems to be the
change in H2O2.
The projected relative abundance of sulfate vs. nitrate particles in future years depends upon
assumptions about the effectiveness of acid rain and as yet undecided strategies to implement regional
haze and fine particle standards. Under the scenarios examined for this RIA, the sulfate levels were high
enough to limit the amount of nitrate aerosol reduction that would accompany a regional NOx reduction.
If, however, future strategies further reduced SOx emissions, the expected reductions in fine particles
from NOx reductions would be decidedly larger, and more linear. Because of the uncertainties in the
atmospheric chemistry for these future years, the RIA relies on modeling tools that estimate both linear
and non-linear responses for PM reductions.
D.3 Reference
Dennis, R.L. Memorandum to Scott Mathias and John Bachmann, U.S. EPA Office of Air Quality
Planning and Standards from Robin Dennis, U.S. EPA, Office of Research and Development,, September
17, 1998.
Page D-2
-------
TECHNICAL REPORT DATA
(Please read Instructions on reverse before completing)
1. REPORT NO.
EPA-452/R-98-003
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Regulatory Impact Analysis for the NOx SIP Call, FIP, and
Section 126 Petitions
Volume 1: Costs and Economic Impacts
Volume 2: Health and Welfare Benefits
5. REPORT DATE
Volume 1: September 1998
Volume 2: December 1998
6. PERFORMING ORGANIZATION CODE
OAQPS/AQSSD
7. AUTHOR(S)
Office of Air Quality Planning and Standards
Office of Atmospheric Programs
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards &
Office of Atmospheric Programs
Research Triangle Park, NC 27711
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
Director
Office of Air Quality Planning and Standards
Office of Air and Radiation
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/200/04
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This report contains EPA's estimates of the annual costs and benefits of the final NOx SIP call and the
proposed NOx FIP and CAA section 126 petition actions. The report also contains a brief profile of
potentially affected sources and potential economic impacts.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Regulatory impact analysis; benefits-cost
comparison
Air Pollution control
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (Report)
Unclassified
21. NO. OF PAGES
20. SECURITY CLASS (Page)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION IS OBSOLETE
-------
------- |