&EPA
United Sfcfes
Environirwntal Protection
Agency
Final Regulatory Impact Analysis (RIA) for the SO2
National Ambient Air Quality Standards (NAAQS)
June 2010
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Health and Environmental Impact Division
Air Benefit-Cost Group
Research Triangle Park, North Carolina
-------
Table of Contents
Page
Executive Summary
ES.l Overview ES-1
ES.2 Summary of Analytic Approach ES-2
ES.3. Results of Analysis ES-6
ES.4. Caveats and Limitations ES-10
ES.5. References ES-14
Chapter 1: Introduction and Background
1.1 Background 1-2
1.2 Role of the Regulatory Impact Analysis in the NAAQS Setting Process. 1-2
1.3 Overview and Design of the RIA 1-6
1.4 S02 Standard Alternatives Considered 1-8
1.5 References 1-8
Chapter 2: SO2 Emissions and Monitoring Data
2.1 Sources of S02 2-1
2.2 Air Quality Monitoring Data 2-2
Chapter 3: Air Quality Analysis
3.1 2005-2007 Design Values 3-1
3.2 Calculation of 2020 Projected Design Values 3-3
3.3 Results 3-12
3.4 Summary 3-23
3.5 References 3-31
Appendix 3a: 2005-2007 and 2020 Design Values
Chapter 4: Emissions Controls Analysis - Design and Analytical Results
4.1 Developing the Identified Control Strategy Analysis 4-3
4.2 S02 Emission Reductions Achieved with Identified Controls Analysis.. 4-7
4.3 Impacts Using Identified Controls 4-11
4.4 Emission Reductions Needed Beyond Identified Controls 4-12
4.5 Key Limitations 4-14
4.6 References 4-14
Chapter 5: Benefits Analysis Approach and Results
5.1 Introduction 5-2
5.2 Primary Benefits Approach 5-2
-------
5.3 Overview of analytical framework for benefits analysis 5-3
5.4 Estimating Avoided Health Effects from S02 Exposure 5-6
5.5 Valuation of Avoided Health Effects from S02 Exposure 5-18
5.6 Health Benefits of Reducing Exposure to S02 Results 5-20
5.7 PM2.5 Health Co-Benefits 5-22
5.8 Summary of Total Monetized Benefits (S02 and PM2.5) 5-33
5.9 Unquantified Welfare Benefits 5-36
5.10 Limitations and Uncertainties 5-55
5.11 Discussion 5-60
5.12 References 5-62
Chapter 6: Cost Analysis Approach and Results
6.1 Engineering Cost Estimates 6-2
6.2 Economic Impacts 6-21
6.3 Energy Impacts 6-22
6.4 Limitations and Uncertainties Associated with Engineering Cost
Estimates 6-23
Chapter 7: Estimates of Costs and Benefits
7.1 Benefits and Costs 7-1
7.2 Discussion of Uncertainties and Limitations 7-4
Chapter 8: Statutory and Executive Order Reviews
1.0 Executive Order 12866: Regulatory Planning and Review
2.0 Paperwork Reduction Act
3.0 Regulatory Flexibility Act
4.0 Unfunded Mandates Reform Act
5.0 Executive Order 13132: Federalism
6.0 Executive Order 13175: Consultation and Coordination with Indian Tribal
Governments
7.0 Executive Order 13045: Protection of Children from Environmental Health & Safety
Risks
8.0 Executive Order 13211: Actions that Significantly Affect Energy Supply, Distribution
or Use
9.0 National Technology Transfer and Advancement Act
10.0 Executive Order 12898: Federal Actions to Address Environmental Justice in
Minority Populations and Low-Income Population
-------
Executive Summary
ES.l Overview
This Regulatory Impact Analysis (RIA) provides illustrative estimates of the incremental
costs and monetized human health benefits of attaining a revised short-term Sulfur Dioxide
(S02) National Ambient Air Quality Standard (NAAQS) within the current monitoring network of
488 S02 monitors. Because this analysis only considers counties with an S02 monitor, the
possibility exists that there may be many more potential nonattainment areas than have been
analyzed in this RIA.
This RIA chiefly serves two purposes. First, it provides the public with an estimate of the
costs and benefits of attaining a new S02 NAAQS. Second, it fulfills the requirements of
Executive Order 12866 and the guidelines of OMB Circular A-4. 1 These documents present
guidelines for EPA to assess the benefits and costs of the selected regulatory option, as well as
one less stringent and one more stringent option. The RIA analyzes the new short-term S02
NAAQS of 75 parts per billion (ppb), based on the 3-year average of the 99th percentile of 1-
hour daily maximum concentrations. This RIA also analyzes alternative primary standards of 50
and 100 ppb.
This analysis does not estimate the projected attainment status of areas of the country
other than those counties currently served by one of the approximately 488 monitors in the
current network. It is important to note that the final rule requires a monitoring network
comprised of monitors sited at locations of expected maximum hourly concentrations, and also
provides for nonattainment designations using air quality modeling near large stationary
sources. Only about one third of the existing S02 network may be source-oriented and/or in
the locations of maximum concentration required by the final rule because the current network
is focused on population areas and community-wide ambient levels of S02. Actual monitored
levels using the new monitoring network and/or air quality modeling results near large
stationary sources may be higher than levels measured using the existing network. We
recognize that once the new requirements are put in place, more areas could find themselves
exceeding the new S02 NAAQS. However for this RIA analysis, we lack sufficient data to predict
which counties might exceed the new NAAQS after implementation of the new monitoring
network and modeling requirements. Therefore we lack a credible analytic path to estimating
costs and benefits for such a future scenario.
U.S. Office of Management and Budget. Circular A-4, September 17, 2003. Available at
http://www.whitehouse.gov/omb/circulars/a004/a-4.pdf.
-------
In setting primary ambient air quality standards, EPA's responsibility under the law is to
establish standards that protect public health, regardless of the costs of implementing a new
standard. The Clean Air Act requires EPA, for each criteria pollutant, to set a standard that
protects public health with "an adequate margin of safety." As interpreted by the Agency and
the courts, the Act requires EPA to create standards based on health considerations only.
The prohibition against the consideration of cost in the setting of the primary air quality
standard, however, does not mean that costs or other economic considerations are
unimportant or should be ignored. The Agency believes that consideration of costs and benefits
is essential to making efficient, cost effective decisions for implementation of these standards.
The impacts of cost and efficiency are considered by states during this process, as they decide
what timelines, strategies, and policies are most appropriate. This RIA is intended to inform the
public about the potential costs and benefits associated with a hypothetical scenario that may
result when a new S02 standard is implemented, but is not relevant to establishing the
standards themselves.
ES.2 Summary of Analytic Approach
This RIA includes several key elements, including specification of baseline S02 emissions
and concentrations; development of illustrative control strategies to attain the standard in
2020; and analyses of the control costs and health benefits of reaching the various alternative
standards. Additional information on the methods employed by the Agency for this RIA is
presented below.
Overview of Baseline Emissions Forecast and Baseline S02 Concentrations
The baseline emissions and concentrations for this RIA are emissions data from the 2005
National Emissions Inventory (NEI), and baseline S02 concentration values from 2005-2007
across the community-wide monitoring network. We used results from community multi-scale
air quality model (CMAQ) simulations to calculate the expected reduction in ambient S02
concentrations between the 2005 base year and 2020. More specifically, design values (i.e. air
quality concentrations at each monitor) were calculated for 2020 using monitored air quality
concentrations from 2005 and modeled air quality projections for 2020, countywide emissions
inventory data for 2005 and 2006-8, and emissions inventory projections for 2020. These data
were used to create ratios between emissions and air quality, and those ratios (relative
response factors, or RRFs) were used to estimate air quality monitor design values for 2020.
The 2020 baseline air quality estimates revealed that 27 monitors in 24 counties were projected
to exceed the 75 ppb NAAQS in 2020.
ES-2
-------
Development of Illustrative Control Strategies
For each alternative standard, we analyzed the impact that additional emissions
controls applied to numerous sectors would have on predicted ambient S02 concentrations,
incremental to the baseline set of controls. Thus the modeled analysis for a revised standard
focuses specifically on incremental improvements beyond the current standards, and uses
control options that might be available to states for application by 2020. The hypothetical
modeled control strategy presented in this RIA is one illustrative option for achieving emissions
reductions to move towards a national attainment of a tighter standard. It is not a
recommendation for how a tighter S02 standard should be implemented, and states will make
decisions regarding implementation strategies once a final NAAQS has been set.
The baseline for this analysis is complicated by the expected issuance of additional air
quality regulations. The S02 NAAQS is only one of several regulatory programs that are likely to
affect ECU emissions nationally in the next several years. We thus expect that EGUs will apply
controls in the coming years in response to multiple rules. These include the maximum
achievable control technology (MACT) rule for utility boilers, revisions to the Clean Air
Interstate Rule, and reconsideration of the Clean Air Mercury Rule. Therefore controls and
costs attributed solely to the S02 NAAQS in this analysis will likely be needed for compliance
with other future rules as well.
The 2020 baseline air quality estimates revealed that 27 monitors in 24 counties were
projected to exceed the 75 ppb NAAQS in 2020. We then developed hypothetical control
strategies that could be adopted to bring the current highest emitting monitor in each of those
counties into attainment with 75 ppb by 2020, as well as hypothetical control strategies for
counties exceeding the lower bound analytic target of 50 ppb, and the upper bound analytic
target of 100 ppb. Controls for three emissions sectors were included in the control analysis:
non-electricity generating unit point sources (nonEGU), area sources (area), and electricity
generating unit point sources (ECU). Finally, we note it was not possible, in this analysis, to
bring all areas into attainment with alternative standards in all areas using identified
engineering controls. For these monitor areas we estimated the cost of unspecified emission
reductions.
Analysis of Costs and Benefits
We estimated the benefits and costs for the final NAAQS of 75 ppb, as well as
alternative S02 NAAQS levels of 50 ppb and 100 ppb (99th percentile). These costs and benefits
ES-3
-------
are associated with an incremental difference in ambient concentrations between a baseline
scenario and a pollution control strategy. As indicated in Chapter 4, several areas of the
country may not be able to attain some alternative standard using known pollution control
methods. Because some areas require substantial emission reductions from unknown sources
to attain the various standards, the results are very sensitive to assumptions about the costs of
full attainment. For this reason, we provide the full attainment results and the partial
attainment results for both benefits and costs.
Benefits
Our benefits analysis estimates the human health benefits for each of the alternative
standard levels including benefits related to reducing S02 concentrations and the co-benefits of
reducing concentrations of fine particulate matter (PM2.5). For the S02 benefits analysis, we use
the Environmental Benefits Mapping and Analysis Program (BenMAP) to estimate the health
benefits occurring as a result of implementing alternative S02 NAAQS levels. BenMAP has been
used extensively in previous RIAs to estimate the health benefits of reducing exposure to
various pollutants.
The primary input to the benefits assessment for S02 effects is the estimated changes in
ambient air quality expected to result from a simulated control strategy or attainment of a
particular standard. CMAQ projects both design values at S02 monitors and air quality
concentrations at 12 km by 12 km grid cells nationwide. To estimate the benefits of fully
attaining the standards in all areas, EPA employed the "monitor rollback" approach to
approximate the air quality change resulting from just attaining alternative S02 NAAQS at each
design value monitor. Under this approach, we use data from the existing S02 monitoring
network and the inverse distance-squared variant of the Veronoi Neighborhood Averaging
(VNA) interpolation method to adjust the air quality modeled concentrations such that each
area just attains the target NAAQS levels.
We quantified S02-related health endpoints for which the S02 ISA provides the strongest
evidence of an effect. In this analysis, we only estimated the benefits for those endpoints with
sufficient evidence to support a quantified concentration-response relationship using the
information presented in the S02 ISA, which contains an extensive literature review for several
health endpoints related to S02 exposure. Based on our review of this information, we
quantified three short-term morbidity endpoints that the S02 ISA identified as "sufficient to
infer a likely causal relationship": asthma exacerbation, respiratory-related emergency
department visits, and respiratory-related hospitalizations. We then selected concentration-
response functions and valuation functions based on criteria detailed in chapter 5. The
ES-4
-------
valuation functions, ambient concentrations, and population data in the monitor areas are
combined in BenMAP to provide the benefits estimates for this analysis. In this analysis, we
decided not to quantify the premature mortality from S02 exposure in this analysis despite
evidence suggesting a positive association. As the literature continues to evolve, we may revisit
this decision in future benefits assessment for S02.
In addition, because S02 is also a precursor to PM2.5, reducing S02 emissions in the
projected non-attainment areas will also reduce PM2.5 formation, human exposure, and the
incidence of PM2.5-related health effects. In this analysis, we estimated the co-benefits of
reducing PM2.5 exposure for the alternative standards. Due to analytical limitations, it was not
possible to provide a comprehensive estimate of PM2.5-related benefits. Instead, we used the
"benefit-per-ton" method to estimate these benefits. The PM2.5 benefit-per-ton estimates
provide the total monetized human health benefits (the sum of premature mortality and
premature morbidity) of reducing one ton of PM2.5from a specified source. EPA has used these
estimates in previous RIAs, including the recent N02 NAAQS RIA.
These estimates reflect EPA's most current interpretation of the scientific literature and
are consistent with the methodology used for the proposal RIA. These benefits are incremental
to an air quality baseline that reflects attainment with the 2008 ozone and 2006 PM2.5 National
Ambient Air Quality Standards (NAAQS). More than 99% of the total dollar benefits are
attributable to reductions in PM2.5 exposure resulting from S02 emission controls. Higher or
lower estimates of benefits are possible using other assumptions; examples of this are provided
in Figure 5.1 for the selected standard of 75 ppb. Methodological limitations prevented EPA
from quantifying the impacts to, or monetizing the benefits from several important benefit
categories, including ecosystem effects from sulfur deposition, improvements in visibility, and
materials damage. Other direct benefits from reduced S02 exposure have not been quantified,
including reductions in premature mortality.
Costs
Consistent with our development of the illustrative control strategies described above,
our analysis of the costs associated with the range of alternative NAAQS focuses on S02
emission controls for electric generating units (ECU) and nonEGU stationary and area sources.
ECU, nonEGU and area source controls largely include measures from the Control Strategy Tool
(CoST), and the AirControlNET control technology database. For these sources, we estimated
costs based on the cost equations included in AirControlNET.
ES-5
-------
As indicated in the above discussion on illustrative control strategies, implementation of
the S02 control measures identified from AirControlNET and other sources does not result in
attainment with the selected NAAQS in several areas. In these areas, additional unspecified
emission reductions might be necessary to reach some alternative standard levels. In order to
bring these monitor areas into attainment, we calculated controls costs using a fixed cost per
ton approach similar to that used in the ozone RIA analysis. We recognize that a single fixed
cost of control of $15,000 per ton of emissions reductions does not account for the significant
emissions cuts that are necessary in some areas, and so its use provides an estimate that is
likely to differ from actual future costs.
ES.3 Results of Analysis
Air Quality
Table ES.l presents the number of monitors and counties exceeding the various target
NAAQS levels in 2020 prior to control, out of 229 monitors from which a full set of data were
available for this analysis.
Table ES.l. Number of monitors and counties projected to exceed 50, 75, and 100
ppb alternative NAAQS target levels in 2020.
Alternative standard (ppb) Number of monitors Number of counties
507156
752724
100119
Table ES.2 presents the emission reductions achieved through applying identical control
measures, both by sector and in total. As this table reveals, a majority of the emission
reductions would be achieved through ECU emission controls.
ES-6
-------
Table ES.2: Emission Reductions from Identified Controls in 2020 in Total and by Sector (Tons)
afor Each Alternative Standard
Total Emission
Reductions from
Identified Controls'5
EGUs
Non-EGUs
Area Sources
50 ppb
800,000
540,000
250,000
15,000
75ppb
370,000
260,000
110,000
200
100 ppb
190,000
110,000
79,000
100
All estimates rounded to two significant figures. As such, totals may not sum down columns.
bThese values represent emission reductions for the identified control strategy analysis. Then
able to attain the alternative standard being analyzed with identified controls only.
Table ES.3 shows the emission reductions needed beyond identified controls for
counties to attain the alternative standards being analyzed.
Table ES.3: Total Emission Reductions and those from Extrapolated Controls in 2020 in Total
and by Sector (Tons)a for Each Alternative Standard
50 ppb 75 ppb 100 ppb
Total Emission
Reductions from
920,000 350,000 170,000
Identified and
Unidentified Controls
Total Emission
Reductions from
Unidentified Controls
Unidentified Reductions
from EGUs
Unidentified Reductions
from non-EGUs
Unidentified Reductions
from Area Sources
110,000
33,000
54,000
19,000
33,000
5,000
22,000
6,400
18,000
~
15,000
3,000
All estimates rounded to two significant figures.
Benefit and Cost Estimates
When estimating the S02- and PM2.5-related human health benefits and compliance
costs in Table ES.4 below, EPA applied methods and assumptions consistent with the state-of-
the-science for human health impact assessment, economics and air quality analysis. EPA
applied its best professional judgment in performing this analysis and believes that these
estimates provide a reasonable indication of the expected benefits and costs to the nation of
the selected S02 standard and alternatives considered by the Agency. The Regulatory Impacts
Analysis (RIA) available in the docket describes in detail the empirical basis for EPA's
ES-7
-------
assumptions and characterizes the various sources of uncertainties affecting the estimates
below.
EPA's 2009 Integrated Science Assessment for Particulate Matter concluded, based on
the scientific literature, that a no-threshold log-linear model most adequately portrays the PM-
mortality concentration-response relationship. Nonetheless, consistent with historical practice
and our commitment to characterizing the uncertainty in our benefits estimates, EPA has
included a sensitivity analysis with an assumed threshold in the PM-mortality health impact
function in the RIA. EPA has included a sensitivity analysis in the RIA to help inform our
understanding of the health benefits which can be achieved at lower air quality concentration
levels. While the primary estimate and the sensitivity analysis are not directly comparable, due
to differences in population data and use of different analysis years, as well as the difference in
the assumption of a threshold in the sensitivity analysis, comparison of the two results provide
a rough sense of the proportion of the health benefits that occur at lower PM2.5 air quality
levels. Using a threshold of 10 u.g/m3 is an arbitrary choice (EPA could have assumed 6, 8, or 12
u.g/m3 for the sensitivity analysis). Assuming a threshold of 10 u.g/m3, the sensitivity analysis
shows that roughly one-third of the benefits occur at air quality levels below that threshold.
Because the primary estimates reflect EPA's current methods and data, EPA notes that caution
should be exercised when comparing the results of the primary and sensitivity analyses. EPA
appreciates the value of sensitivity analyses in highlighting the uncertainty in the benefits
estimates and will continue to work to refine these analyses, particularly in those instances in
which air quality modeling data are available.
Table ES.4 shows the results of the cost and benefits analysis for each standard
alternative. As indicated above, implementation of the S02 control measures identified from
AirControlNET and other sources does not result in attainment with the all target NAAQS levels
in several areas. In these areas, additional unspecified emission reductions might be necessary
to reach some alternative standard levels. The first part of the table, labeled Partial attainment
(identified controls), shows only those benefits and costs from control measures we were able
to identify. The second part of the table, labeled Unidentified Controls, shows only additional
benefits and costs resulting from unidentified controls. The third part of the table, labeled Full
attainment, shows total benefits and costs resulting from both identified and unidentified
controls. It is important to emphasize that we were able to identify control measures for a
significant portion of attainment for many of those counties that would not fully attain the
target NAAQS level with identified controls. Note also that in addition to separating full and
partial attainment, the table also separates the portion of benefits associated with reduced S02
exposure (i.e., S02 benefits) from the additional benefits associated with reducing S02
emissions, which are precursors to PM2.5 formation - (i.e., the PM2.5 co-benefits). For instance,
ES-8
-------
for the selected standard of 75 ppb, $2.2 million in benefits are associated with reduced S02
exposure while $15 billion to $37 billion are associated with reduced PM2.5 exposure.
Table ES.4: Monetized Benefits and Costs to Attain Alternate Standard Levels in 2020
(millions of 2006$)a
£ 1
Partial
Attainme
(identified con
Unidentified
Controls
Full Attainment
50 ppb
75 ppb
100 ppb
50 ppb
75 ppb
100 ppb
50 ppb
75 ppb
100 ppb
# Counties
Fully
Controlled
40
20
6
16
4
3
56
24
9
Discount
Rate
3%
7%
3%
7%
3%
7%
3%
7%
3%
7%
3%
7%
3%
7%
3%
7%
3%
7%
Monetized .. . .„..
Monetized PM2 5
„ I. Co-Benefits c'd
Benefits
b $30,000 to $74,000
$28,000 to $67,000
b $14,000 to $35,000
$13,000 to $31,000
b $6,900 to $17,000
$6,200 to $15,000
b $4,000 to $9,000
$3,000 to $8,000
b $1,000 to $3,000
$1,000 to $3,000
b $500 to $1,000
$500 to $1,000
, $34,000 to $83,000
? $31,000 to $75,000
§ $15,000 to $37,000
^ ' $14,000 to $34,000
$7,400 to $18,000
$6,700 to $16,000
Costs
$2,600
$960
$470
$1,800
$500
$260
$4,400
$1,500
$730
Net Benefits
$27,000 to $7 1,000
$25,000 to $64,000
$13,000 to $34,000
$12,000 to $30,000
$6,400 to $17,000
$5,700 to $15,000
$2,200 to $7,200
$1,200 to $6,200
$500 to $1,500
$500 to $2,500
$240 to $740
$240 to $740
$30,000 to $79,000
$27,000 to $71,000
$14,000 to $36,000
$13,000 to $33,000
$6,700 to $17,000
$6,000 to $15,000
Estimates have been rounded to two significant figures and therefore summation may not match table estimates.
bThe approach used to simulate air quality changes for SO2 did not provide the data needed to distinguish partial
attainment benefits from full attainment benefits from reduced SO2 exposure. Therefore, a portion of the SO2
benefits is attributable to the known controls and a portion of the SO2 benefits are attributable to the unidentified
controls. Because all SO2-related benefits are short-term effects, the results are identical for all discount rates.
c Benefits are shown as a range from Pope et al (2002) to Laden et al. (2006). Monetized benefits do not include
unquantified benefits, such as other health effects, reduced sulfur deposition, or improvements in visibility.
d These models assume that all fine particles, regardless of their chemical composition, are equally potent in
causing premature mortality because there is no clear scientific evidence that would support the development of
differential effects estimates by particle type. Reductions in SO2 emissions from multiple sectors to meet the SO2
NAAQS would primarily reduce the sulfate fraction of PM2.5. Because this rule targets a specific particle precursor
(i.e., SO2), this introduces some uncertainty into the results of the analysis.
ES-9
-------
ES.4. Caveats and Limitations
Air Quality, Emissions, and Control Strategies
The estimates of emission reductions associated with the control strategies described
above are subject to important limitations and uncertainties. We summarize these limitations
as follows:
• Actual State Implementation Plans May Differ from our Simulation: In order to reach
attainment with the proposed NAAQS, each state will develop its own implementation
plan implementing a combination of emissions controls that may differ from those
simulated in this analysis. This analysis therefore represents an approximation of the
emissions reductions that would be required to reach attainment and should not be
treated as a precise estimate.
• Use of Existing CMAQ Model Runs: This analysis represents a screening level analysis.
We did not conduct new regional scale modeling specifically targets to S02; instead we
relied upon impact ratios developed from model runs used in the analysis underlying
the PM2.5 NAAQS.
• Unidentified controls: We have limited information on available controls for some of
the monitor areas included in this analysis. For a number of small non-EGU and area
sources, there is little or no information available on S02 controls.
Costs
We do not have sufficient information for all of our known control measures to calculate
cost estimates that vary with an interest rate. We are able to calculate annualized costs
at an interest rate other than 7% (e.g., 3% interest rate) where there is sufficient
information—available capital cost data, and equipment life—to annualize the costs for
individual control measures. For the vast majority of nonEGU point source control
measures, we do have sufficient capital cost and equipment life data for individual
control measures to prepare annualized capital costs using the standard capital recovery
factor. Hence, we are able to provide annualized cost estimates at different interest
rates for the point source control measures.
ES-10
-------
• There are some unquantified costs that are not adequately captured in this illustrative
analysis. These costs include the costs of federal and State administration of control
programs, which we believe are less than the alternative of States developing
approvable SIPs, securing EPA approval of those SIPs, and Federal/State enforcement.
Additionally, control measure costs referred to as "no cost" may require limited
government agency resources for administration and oversight of the program not
included in this analysis; those costs are generally outweighed by the saving to the
industrial, commercial, or private sector. The Agency also did not consider transactional
costs and/or effects on labor supply in the illustrative analysis.
Benefits
Although we strive to incorporate as many quantitative assessments of uncertainty, there
are several aspects for which we are only able to address qualitatively. These aspects are
important factors to consider when evaluating the relative benefits of the attainment strategies
for each of the alternative standards:
• The 12 km CMAQ grid, which is the air quality modeling resolution, may be too coarse to
accurately estimate the potential near-field health benefits of reducing S02 emissions.
These uncertainties may under- or over-estimate benefits.
• The interpolation techniques used to estimate the full attainment benefits of the
alternative standards contributed some uncertainty to the analysis. The great majority
of benefits estimated for the various standard alternatives were derived through
interpolation. As noted previously in this chapter, these benefits are likely to be more
uncertain than if we had modeled the air quality scenario for both S02 and PM2.s. In
general, the VNA interpolation approach may under-estimate benefits because it does
not account for the broader spatial distribution of air quality changes that may occur
due to the implementation of a regional emission control program.
• There are many uncertainties associated with the health impact functions used in this
modeling effort. These include: within study variability (the precision with which a given
study estimates the relationship between air quality changes and health effects); across
study variation (different published studies of the same pollutant/health effect
relationship typically do not report identical findings and in some instances the
differences are substantial); the application of C-R functions nationwide (does not
account for any relationship between region and health effect, to the extent that such a
relationship exists); extrapolation of impact functions across population (we assumed
that certain health impact functions applied to age ranges broader than that considered
in the original epidemiological study); and various uncertainties in the C-R function,
ES-11
-------
including causality and thresholds. These uncertainties may under- or over-estimate
benefits.
Co-pollutants present in the ambient air may have contributed to the health effects
attributed to S02 in single pollutant models. Risks attributed to S02 might be
overestimated where concentration-response functions are based on single pollutant
models. If co-pollutants are highly correlated with S02, their inclusion in an S02 health
effects model can lead to misleading conclusions in identifying a specific causal
pollutant. Because this collinearity exists, many of the studies reported statistically
insignificant effect estimates for both S02 and the co-pollutants; this is due in part to the
loss of statistical power as these models control for co-pollutants. Where available, we
have selected multipollutant effect estimates to control for the potential confounding
effects of co-pollutants; these include NYDOH (2006), Schwartz et al. (1994) and
O'Connor et al. (2008). The remaining studies include single pollutant models.
This analysis is for the year 2020, and projecting key variables introduces uncertainty.
Inherent in any analysis of future regulatory programs are uncertainties in projecting
atmospheric conditions and source level emissions, as well as population, health
baselines, incomes, technology, and other factors.
This analysis omits certain unquantified effects due to lack of data, time and resources.
These unquantified endpoints include other health effects, ecosystem effects, and
visibility. EPA will continue to evaluate new methods and models and select those most
appropriate for estimating the benefits of reductions in air pollution. Enhanced
collaboration between air quality modelers, epidemiologists, toxicologists, ecologists,
and economists should result in a more tightly integrated analytical framework for
measuring benefits of air pollution policies.
PM2.5 co-benefits represent a substantial proportion of total monetized benefits (over
99% of total monetized benefits), and these estimates are subject to a number of
assumptions and uncertainties.
a. PM2.5 co-benefits were derived through benefit per-ton estimates, which do not
reflect local variability in population density, meteorology, exposure, baseline
health incidence rates, or other local factors that might lead to an over-estimate
or under-estimate of the actual benefits of controlling directly emitted fine
particulates.
b. We assume that all fine particles, regardless of their chemical composition, are
equally potent in causing premature mortality. This is an important assumption,
because PM2.5 produced via transported precursors emitted from EGUs may
differ significantly from direct PM2.5 released from diesel engines and other
industrial sources, but no clear scientific grounds exist for supporting differential
effects estimates by particle type.
ES-12
-------
c. We assume that the health impact function for fine particles is linear down to
the lowest air quality levels modeled in this analysis. Thus, the estimates include
health benefits from reducing fine particles in areas with varied concentrations
of PM2.5, including both regions that are in attainment with fine particle standard
and those that do not meet the standard down to the lowest modeled
concentrations.
d. To characterize the uncertainty in the relationship between PM2.5and premature
mortality (which typically accounts for 85% to 95% of total monetized benefits),
we include a set of twelve estimates based on results of the expert elicitation
study in addition to our core estimates. Even these multiple characterizations
omit the uncertainty in air quality estimates, baseline incidence rates,
populations exposed and transferability of the effect estimate to diverse
locations. As a result, the reported confidence intervals and range of estimates
give an incomplete picture about the overall uncertainty in the PM2.5 estimates.
This information should be interpreted within the context of the larger
uncertainty surrounding the entire analysis. For more information on the
uncertainties associated with PM2.5 co-benefits, please consult the PM2.5 NAAQS
RIA (Table 5.5).
While the monetized benefits of reduced S02 exposure appear small when compared to the
monetized benefits of reduced PM2.5 exposure, readers should not necessarily infer that the
total monetized benefits of attaining a new S02 standard are minimal. For this rule, the
monetized PM2.5 co-benefits represent over 99% of the total monetized benefits. This result is
consistent with other recent RIAs, where the PM2.5 co-benefits represent a large proportion of
total monetized benefits. This result is amplified in this RIA by the decision not to quantify S02-
related premature mortality and other morbidity endpoints due to the uncertainties associated
with estimating those endpoints. Studies have shown that there is a relationship between S02
exposure and premature mortality, but that relationship is limited by potential confounding.
Because premature mortality generally comprises over 90% of the total monetized benefits,
this decision may substantially underestimate the monetized health benefits of reduced S02
exposure.
In addition, we were unable to quantify the benefits from several welfare benefit
categories. We lacked the necessary air quality data to quantify the benefits from
improvements in visibility from reducing light-scattering particles. Previous RIAs for ozone (U.S.
EPA, 2008a) and PM2.5 (U.S. EPA, 2006a) indicate that visibility is an important benefit category,
and previous efforts to monetize those benefits have only included a subset of visibility
ES-13
-------
benefits, excluding benefits in urban areas and many national and state parks. Even this subset
accounted for up to 5% of total monetized benefits in the Ozone NAAQS RIA (U.S. EPA, 2008a).
We were also unable to quantify the ecosystem benefits of reduced sulfur deposition
because we lacked the necessary air quality data, and the methodology to estimate ecosystem
benefits is still being developed. Previous assessments (U.S. EPA, 1999; U.S. EPA, 2005; U.S.
EPA, 2009e) indicate that ecosystem benefits are also an important benefits category, but those
efforts were only able to monetize a tiny subset of ecosystem benefits in specific geographic
locations, such as recreational fishing effects from lake acidification in the Adirondacks. We
were also unable to quantify the benefits of decreased mercury methylation from sulfate
deposition. Quantifying the relationship between sulfate and mercury methylation in natural
settings is difficult, but some studies have shown that decreasing sulfate deposition can also
decrease methylmercury.
ES.5. References
Laden, F., J. Schwartz, F.E. Speizer, and D.W. Dockery. 2006. "Reduction in Fine Particulate Air
Pollution and Mortality." American Journal of Respiratory and Critical Care Medicine
173:667-672. Estimating the Public Health Benefits of Proposed Air Pollution Regulations.
Washington, DC: The National Academies Press.
Pope, C.A., III, R.T. Burnett, M.J. Thun, E.E. Calle, D. Krewski, K. Ito, and G.D. Thurston. 2002.
"Lung Cancer, Cardiopulmonary Mortality, and Long-term Exposure to Fine Particulate Air
Pollution." Journal of the American Medical Association 287:1132-1141.
U.S. Environmental Protection Agency (U.S. EPA). 1999a. The Benefits and Costs of the Clean
Air Act 1990 to 2010: EPA Report to Congress. EPA-410-R-99-001. Office of Air and
Radiation, Office of Policy, Washington, DC. November. Available on the Internet at <
http://www.epa.gov/air/sect812/1990-2010/fullrept.pdf>.
U.S. Environmental Protection Agency (U.S. EPA). 2005. Regulatory Impact Analysis for the
Final Clean Air Interstate Rule. Office of Air and Radiation. March. Available on the
Internet at < http://www.epa.gov/cair/pdfs/finaltech08.pdf>.
U.S. Environmental Protection Agency (U.S. EPA). 2006a. Regulatory Impact Analysis, 2006
National Ambient Air Quality Standards for Particulate Matter, Chapter 5. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. October. Available on the
Internet at .
U.S. Environmental Protection Agency (U.S. EPA). 2008a. Regulatory Impact Analysis, 2008
National Ambient Air Quality Standards for Ground-level Ozone, Chapter 6. Office of Air
ES-14
-------
Quality Planning and Standards, Research Triangle Park, NC. March. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2009e. Risk and Exposure Assessment for
Review of the Secondary National Ambient Air Quality Standards for Oxides of Nitrogen and
Oxides of Sulfur (Final). EPA-452/R-09-008a. Office of Air Quality Planning and Standards,
Research Triangle Park, NC. September. Available on the Internet at
-------
Chapter 1: Introduction and Background
Synopsis
This document estimates the incremental costs and monetized human health benefits of
attaining a revised primary sulfur dioxide (S02) National Ambient Air Quality Standard (NAAQS)
nationwide. This document contains illustrative analyses that consider limited emission control
scenarios that states, tribes and regional planning organizations might implement to achieve a
revised S02 NAAQS. EPA weighed the available empirical data and photochemical modeling to
make judgments regarding the proposed attainment status of certain urban areas in the future.
According to the Clean Air Act, EPA must use health-based criteria in setting the NAAQS and
cannot consider estimates of compliance cost. This Regulatory Impact Analysis (RIA) is intended
to provide the public a sense of the benefits and costs of meeting new alternative S02 NAAQS,
and to meet the requirements of Executive Order 12866 and OMB Circular A-4 (described
below in Section 1.2.2).
This RIA provides illustrative estimates of the incremental costs and monetized human
health benefits of attaining a revised primary S02 National Ambient Air Quality Standard
(NAAQS) in 2020 within the current monitoring network1. This proposal would add a new
short-term (1-hour exposure) standard, in addition to the current annual average standard.
This analysis does not estimate the projected attainment status of areas of the country
other than those counties currently served by one of the approximately 488 monitors in the
current network. It is important to note that the final rule requires a monitoring network
comprised of monitors sited at locations of expected maximum hourly concentrations, and also
provides for nonattainment designations using air quality modeling near large stationary
sources. Only about one third of the existing S02 network may be source-oriented and/or in
the locations of maximum concentration required by the final rule because the current network
is focused on population areas and community-wide ambient levels of S02. Actual monitored
levels using the new monitoring network and/or air quality modeling results near large
stationary sources may be higher than levels measured using the existing network. We
recognize that once the new requirements are put in place, more areas could find themselves
exceeding the new S02 NAAQS. However for this RIA analysis, we lack sufficient data to predict
which counties might exceed the new NAAQS after implementation of the new monitoring
network and modeling requirements. Therefore we lack a credible analytic path to estimating
costs and benefits for such a future scenario.
1 There are 488 monitors. Currently xx monitors (representing xx counties) exceed the final NAAQS in this analysis
(75 ppb, 99th percentile daily 1-hour maximum SO2 concentration).
1-1
-------
1.1 Background
Two sections of the Clean Air Act ("Act") govern the establishment and revision of
NAAQS. Section 108 (42 U.S.C. 7408) directs the Administrator to identify pollutants which
"may reasonably be anticipated to endanger public health or welfare," and to issue air quality
criteria for them. These air quality criteria are intended to "accurately reflect the latest
scientific knowledge useful in indicating the kind and extent of all identifiable effects on public
health or welfare which may be expected from the presence of [a] pollutant in the ambient air."
S02 is one of six pollutants for which EPA has developed air quality criteria.
Section 109 (42 U.S.C. 7409) directs the Administrator to propose and promulgate
"primary" and "secondary" NAAQS for pollutants identified under section 108. Section
109(b)(l) defines a primary standard as "the attainment and maintenance of which in the
judgment of the Administrator, based on [the] criteria and allowing an adequate margin of
safety, [are] requisite to protect the public health." A secondary standard, as defined in section
109(b)(2), must "specify a level of air quality the attainment and maintenance of which in the
judgment of the Administrator, based on [the] criteria, [are] requisite to protect the public
welfare from any known or anticipated adverse effects associated with the presence of [the]
pollutant in the ambient air." Welfare effects as defined in section 302(h) [42 U.S.C. 7602(h)]
include but are not limited to "effects on soils, water, crops, vegetation, manmade materials,
animals, wildlife, weather, visibility and climate, damage to and deterioration of property, and
hazards to transportation, as well as effects on economic values and on personal comfort and
well-being."
Section 109(d) of the Act directs the Administrator to review existing criteria and
standards at 5-year intervals. When warranted by such review, the Administrator is to retain or
revise the NAAQS. After promulgation or revision of the NAAQS, the standards are
implemented by the States.
1.2 Role of the Regulatory Impact Analysis in the NAAQS Setting Process
1.2.1 Legislative Roles
In setting primary ambient air quality standards, EPA's responsibility under the law is to
establish standards that protect public health, regardless of the costs of implementing a new
standard. The Clean Air Act requires EPA, for each criteria pollutant, to set a standard that
protects public health with "an adequate margin of safety." As interpreted by the Agency and
the courts, the Act requires EPA to create standards based on health considerations only.
1-2
-------
The prohibition against the consideration of cost in the setting of the primary air quality
standard, however, does not mean that costs or other economic considerations are
unimportant or should be ignored. The Agency believes that consideration of costs and benefits
are essential to making efficient, cost effective decisions for implementation of these
standards. The impact of cost and efficiency are considered by states during this process, as
they decide what timelines, strategies, and policies make the most sense. This RIA is intended
to inform the public about the potential costs and benefits that may result when a new S02
standard is implemented, but is not relevant to establishing the standards themselves.
1.2.2 Role of Statutory and Executive Orders
There are several statutory and executive orders that dictate the manner in which EPA
considers rulemaking and public documents. This document is separate from the NAAQS
decision making process, but there are several statutes and executive orders that still apply to
any public documentation. The analysis required by these statutes and executive orders is
presented in Chapter 8.
EPA presents this RIA pursuant to Executive Order 12866 and the guidelines of OMB
Circular A-4.2 These documents present guidelines for EPA to assess the benefits and costs of
the selected regulatory option, as well as one less stringent and one more stringent option.
OMB circular A-4 also requires both a benefit-cost, and a cost-effectiveness analysis for rules
where health is the primary effect. Within this RIA we provide a benefit-cost analysis.
Methodological and data limitations prevent us from performing a cost-effectiveness analysis
and a meaningful more formal uncertainty analysis for this RIA.
The proposal would set a new short-term S02 standard based on the 3-year average of
the 99th percentile of 1-hour daily maximum concentrations, establishing a new standard within
the range of 75 parts per billion (ppb). This RIA analyzes alternative primary standards of 50
ppb, and 100 ppb.
1.2.3 Market Failure or Other Social Purpose
OMB Circular A-4 indicates that one of the reasons a regulation such as the NAAQS may
be issued is to address market failure. The major types of market failure include: externality,
market power, and inadequate or asymmetric information. Correcting market failures is one
reason for regulation, but it is not the only reason. Other possible justifications include
U.S. Office of Management and Budget. Circular A-4, September 17, 2003, available at
.
1-3
-------
improving the function of government, removing distributional unfairness, or promoting
privacy and personal freedom.
An externality occurs when one party's actions impose uncompensated benefits or costs
on another party. Environmental problems are a classic case of externality. For example, the
smoke from a factory may adversely affect the health of local residents while soiling the
property in nearby neighborhoods. If bargaining was costless and all property rights were well
defined, people would eliminate externalities through bargaining without the need for
government regulation. From this perspective, externalities arise from high transaction costs
and/or poorly defined property rights that prevent people from reaching efficient outcomes
through market transactions.
Firms exercise market power when they reduce output below what would be offered in
a competitive industry in order to obtain higher prices. They may exercise market power
collectively or unilaterally. Government action can be a source of market power, such as when
regulatory actions exclude low-cost imports. Generally, regulations that increase market power
for selected entities should be avoided. However, there are some circumstances in which
government may choose to validate a monopoly. If a market can be served at lowest cost only
when production is limited to a single producer of local gas and electricity distribution services,
a natural monopoly is said to exist. In such cases, the government may choose to approve the
monopoly and to regulate its prices and/or production decisions. Nevertheless, it should be
noted that technological advances often affect economies of scale. This can, in turn, transform
what was once considered a natural monopoly into a market where competition can flourish.
Market failures may also result from inadequate or asymmetric information. Because
information, like other goods, is costly to produce and disseminate, an evaluation will need to
do more than demonstrate the possible existence of incomplete or asymmetric information.
Even though the market may supply less than the full amount of information, the amount it
does supply may be reasonably adequate and therefore not require government regulation.
Sellers have an incentive to provide information through advertising that can increase sales by
highlighting distinctive characteristics of their products. Buyers may also obtain reasonably
adequate information about product characteristics through other channels, such as a seller
offering a warranty or a third party providing information.
There are justifications for regulations in addition to correcting market failures. A
regulation may be appropriate when there are clearly identified measures that can make
government operate more efficiently. In addition, Congress establishes some regulatory
programs to redistribute resources to select groups. Such regulations should be examined to
ensure that they are both effective and cost-effective. Congress also authorizes some
1-4
-------
regulations to prohibit discrimination that conflicts with generally accepted norms within our
society. Rulemaking may also be appropriate to protect privacy, permit more personal freedom
or promote other democratic aspirations.
From an economics perspective, setting an air quality standard is a straightforward case
of addressing an externality, in this case where entities are emitting pollutants, which cause
health and environmental problems without compensation for those suffering the problems.
Setting a standard with a reasonable margin of safety attempts to place the cost of control on
those who emit the pollutants and lessens the impact on those who suffer the health and
environmental problems from higher levels of pollution.
1.2.4 Illustrative Nature of the Analysis
This S02 NAAQS RIA is an illustrative analysis that provides useful insights into a limited
number of emissions control scenarios that states might implement to achieve a revised S02
NAAQS. Because states are ultimately responsible for implementing strategies to meet any
revised standard, the control scenarios in this RIA are necessarily hypothetical in nature. They
are not forecasts of expected future outcomes. Important uncertainties and limitations are
documented in the relevant portions of the analysis.
The illustrative goals of this RIA are somewhat different from other EPA analyses of
national rules, or the implementation plans states develop, and the distinctions are worth brief
mention. This RIA does not assess the regulatory impact of an EPA-prescribed national or
regional rule, nor does it attempt to model the specific actions that any state would take to
implement a revised S02 standard. This analysis attempts to estimate the costs and human and
welfare benefits of cost-effective implementation strategies which might be undertaken to
achieve national attainment of new standards. These hypothetical strategies represent a
scenario where states use one set of cost-effective controls to attain a revised S02 NAAQS.
Because states—not EPA—will implement any revised NAAQS, they will ultimately determine
appropriate emissions control scenarios. State implementation plans would likely vary from
EPA's estimates due to differences in the data and assumptions that states use to develop these
plans.
The illustrative attainment scenarios presented in this RIA were constructed with the
understanding that there are inherent uncertainties in projecting emissions and controls.
Furthermore, certain emissions inventory, control, modeling and monitoring limitations and
uncertainties inhibit EPA's ability to model full attainment in all areas. Despite these limitations,
EPA has used the best available data and methods to produce this RIA.
1-5
-------
1.3 Overview and Design of the RIA
This Regulatory Impact Analysis evaluates the costs and benefits of hypothetical
national strategies to attain several potential revised primary S02 standards. The document is
intended to be straightforward and written for the lay person with a minimal background in
chemistry, economics, and/or epidemiology. Figure 1-1 provides an illustration of the process
used to create this RIA.
Figure 1-1: The Process Used to Create this RIA
Use air quality monitoring
data to determine number
areas exceeding alternative
S02 NAAQS
Determine sources of
SOX emissions in areas
exceeding alternative
S02 NAAQS
Determine baseline: estimated
emission reductions to meet
other federal regulations & the
current S02 NAAQS
Determine emission reductions &
engineering costs incremental to baseline
to meet alternative S02 NAAQS using
known & if appropriate extrapolated
Determine energy and
economic impacts
Estimate S02 and where
appropriate particulate
benefits associated with air
quality changes from
application of simulated
emission reductions
Present benefit-cost
results
Identify uncertainties and
limitations, providing
appropriate context for the
RIA results
1.3.1 Baseline and Years of Analysis
The analysis year for this regulatory impact analysis is 2020, which approximates the
required attainment year under the Clean Air Act. Many areas will reach attainment of any
alternative S02 standard before 2020. For purposes of this analysis, we assess attainment by
2020 for all areas. Some areas for which we assume 2020 attainment may in fact need more
time to meet one or more of the analyzed standards, while others will need less time. This
analysis does not prejudge the attainment dates that will ultimately be assigned to individual
areas under the Clean Air Act.
The methodology first estimates what baseline S02 levels might look like in 2020 with
existing Clean Air Act programs, including application of controls to meet the current S02
NAAQS, various maximum achievable control technology (MACT) standards, and then predicts
1-6
-------
the change in S02 levels following the application of additional controls to reach tighter
alternative standards. This allows for an analysis of the incremental change between the
current standard and alternative standards.
1.3.2 Control Scenarios Considered in this RIA
In this RIA we analyzed the final NAAQS of 75 ppb, as well as hypothetical target NAAQS
levels of 50 and 100 ppb. Hypothetical control strategies were developed for each NAAQS
level. First, we used outputs from CMAQ model runs to estimate air quality changes that would
result from the application of emissions control options that are known to be available to
different types of sources in areas with monitoring levels currently exceeding the alternative
standards. However, given and the amount of improvement in air quality needed to reach the
some standards in some areas, as well as circumstances specific to those areas, it was also
expected that applying these known controls would not reduce S02 concentrations sufficiently
to allow these two areas to reach some standards. In order to bring these monitor areas into
attainment, we calculated the cost of unspecified emission reductions by extrapolating from a
range of fixed costs per ton of emission control that are generally identified nationally.
1.3.3 Evaluating Costs and Benefits
We applied a two step methodology for estimating emission reductions needed to reach
full attainment. First, we quantified the costs associated with applying known controls. Second,
we estimated costs of the additional tons of extrapolated emission reductions estimated which
were needed to reach full attainment. This methodology enabled us to evaluate nationwide
costs and benefits of attaining a tighter S02 standard using hypothetical strategies, albeit with
substantial additional uncertainty regarding the second step estimates.3
To streamline this RIA, this document refers to several previously published documents,
including two technical documents EPA produced to prepare for promulgation of the S02
NAAQS. The first was the Integrated Science Assessment (ISA) created by EPA's Office of
Research and Development (U.S. EPA, 2008), which presented the latest available pertinent
information on atmospheric science, air quality, exposure, health effects, and environmental
effects of S02. The second was the Risk and Exposure Assessment (REA) (U.S. EPA, 2009) for
various standard levels. The REA also includes staff conclusions and recommendations to the
Administrator regarding potential revisions to the standards.
3 Because the secondary SO2 NAAQS is under development in a separate regulatory process, no additional
costs and benefits were calculated in this RIA.
1-7
-------
1.4 SO2 Standard Alternatives Considered
EPA has performed an illustrative analysis of the potential costs and human health and
visibility benefits of nationally attaining S02 NAAQS of 50, 75, and 100 ppb, assuming a baseline
of no additional control beyond the controls expected from rules that are already in place
(including the current PM2.5 NAAQS), and solely within the bounds of the existing monitoring
network. The benefit and cost estimates below are calculated incremental to a 2020 baseline
that incorporates air quality improvements achieved through the projected implementation of
existing regulations and attainment of the existing PM National Ambient Air Quality Standards
(NAAQS). The baseline also includes the MACT program, the clean air interstate rule (CAIR), and
implementation of current consent decrees, all of which would help many areas move toward
attainment of the S02 standard.
1.5 References
U.S. Environmental Protection Agency (U.S. EPA). 1970. Clean Air Act. 40 CFR 50.
U.S. Environmental Protection Agency (U.S. EPA). 2008. Integrated Science Assessment for
Sulfur Oxides - Health Criteria (Final Report). National Center for Environmental
Assessment, Research Triangle Park, NC. September. Available on the Internet at <
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=198843>.
U.S. Environmental Protection Agency (U.S. EPA). 2009. Risk and Exposure Assessment to
Support the Review of the S02 Primary National Ambient Air Quality Standards: Final
Report. Office of Air Quality Planning and Standards, Research Triangle Park, NC. August.
Available on the Internet at
.
1-8
-------
Chapter 2: SO2 Emissions and Monitoring Data
Synopsis
This chapter describes the available S02 emissions and air quality data used to
inform and develop the control strategies outlined in this RIA. We first describe data on
S02 emission sources contained in available EPA emission inventories. We then provide
an overview of data sources for air quality measurement. For a more in-depth discussion
of S02 emissions and air quality data, see the Integrated Science Assessment for the S02
NAAQS.1
2.1 Sources of SO2
In order to estimate risks associated with S02 exposure, principal sources of the
pollutant must first be characterized because the majority of human exposures are likely
to result from the release of emissions from these sources. Anthropogenic S02
emissions originate chiefly from point sources, with fossil fuel combustion at electric
utilities (~66%) and other industrial facilities (~29%) accounting for the majority of total
emissions (ISA, section 2.1). Other anthropogenic sources of S02 include both the
extraction of metal from ore as well as the burning of high sulfur containing fuels by
locomotives, large ships, and non-road diesel equipment. Notably, almost the entire
sulfur content of fuel is released as S02 or S03 during combustion. Thus, based on the
sulfur content in fuel stocks, oxides of sulfur emissions can be calculated to a higher
degree of accuracy than can emissions for other pollutants such as PM and N02 (ISA,
section 2.1).
The largest natural sources of S02 are volcanoes and wildfires. Although S02
constitutes a relatively minor fraction (0.005% by volume) of total volcanic emissions,
concentrations in volcanic plumes can be in the range of several to tens of ppm
(thousands of ppb). Volcanic sources of S02 in the U.S. are limited to the Pacific
Northwest, Alaska, and Hawaii. Emissions of S02 can also result from burning
vegetation. The amount of S02 released from burning vegetation is generally in the
range of 1 to 2% of the biomass burned and is the result of sulfur from amino acids
being released as S02 during combustion.
1 U.S. Environmental Protection Agency (2007c), Review of the National Ambient Air Quality
Standards for SO2: Policy Assessment of Scientific and Technical Information, Integrated Science
Assessment, Chapter 2, EPA-452/R-08-xxx, Office of Air Quality Planning and Standards, RTP, NC.
-------
Emissions inventory inputs representing the year 2005 for the sources above
were developed to provide a base year for the air quality analysis presented in Chapter
3. The 2005 National Emissions Inventory (NEI), version 2 from October 6, 2008 was the
starting point for the U.S. inventories used for the air quality analysis. This inventory
includes 2005-specific data for most point and mobile sources, while most nonpoint and
other data were carried forward from version of the 2002 NEI. For more information on
the 2005 NEI, upon which significant portions of the 2005 modeling platform are based,
see http://www.epa.gov/ttn/chief/net/2005inventory.html.
2.2 Air Quality Monitoring Data
2.2.1 Background on S02 monitoring network
The following section provides general background on the S02 monitoring
network. A more detailed description of this network can be found in Watkins (2009).
The S02 monitoring network was originally deployed to support implementation of the
S02 NAAQS established in 1971. Despite the establishment of an S02 standard, uniform
minimum monitoring requirements for S02 monitoring did not appear until May 1979.
From the time of the implementation of the 1979 monitoring rule through 2008, the S02
network has steadily decreased in size from approximately 1496 sites in 1980 to the
approximately 488 sites operating in 2008.
The 1979 monitoring rule established two categories of S02 monitoring sites:
State and Local Ambient Monitoring Stations (SLAMS) and the smaller set of National
Ambient Monitoring Stations (NAMS). No minimum requirements were established for
SLAMS. Minimum requirements (described below) were established for NAMS. The
1979 rule also required that S02 only be monitored using Federal Reference Methods
(FRMs) or Federal Equivalent Methods (FEMs). The 1979 monitoring rule called fora
range of number of sites in a metropolitan statistical area (MSA) based both on
population size and known concentrations relative to the NAAQS (at that point in time;
see Watkins, 2009).
In October 2006, EPA revised the monitoring requirements for S02 in light of the
fact that there was not an S02 non-attainment problem (Watkins, 2009). The 2006 rule
eliminated the minimum requirements for the number of S02 monitoring sites. The
current S02 monitoring rule, 40 CFR Part 58, Appendix D, section 4.4 states:
-------
Sulfur Dioxide (SO2) Design Criteria:
(a) There are no minimum requirements for the number of S02 monitoring sites.
Continued operation of existing SLAMS S02 sites using FRM or FEM is required
until discontinuation is approved by the EPA Regional Administrator. Where
SLAMS S02 monitoring is ongoing, at least one of the SLAMS S02 sites must be a
maximum concentration site for that specific area.
(b) The appropriate spatial scales for S02 SLAMS monitoring are the microscale,
middle, and possibly neighborhood scales. The multi-pollutant NCore sites can
provide for metropolitan area trends analyses and general control strategy
progress tracking. Other SLAMS sties are expected to provide data that are
useful in specific compliance actions, for maintenance plan agreements, or for
measuring near specific stationary sources of S02.
(1) Micro and middle scale - Some data uses associated with microscale
and middle scale measurements for S02 include assessing the effects of control
strategies to reduce concentrations (especially for the 3-hour and 24-hour
averaging times) and monitoring air pollution episodes.
(2) Neighborhood scale -This scale applies where there is a need to
collect air quality data as part of an ongoing S02 stationary source impact
investigation. Typical locations might include suburban areas adjacent to S02
stationary sources for example, or for determining background concentrations as
part of these studies of population responses to exposure to S02.
(c) Technical guidance in reference 1 of this appendix should be used to evaluate
the adequacy of each existing S02 site, to relocate an existing site, or to locate
new sites.
To ascertain what the current S02 network is addressing or characterizing, and in
light of the relatively recent removal of a specific S02 monitoring requirement, EPA
reviewed some of the S02 network meta-data (Watkins, 2009). The data reviewed are
those available from AQS for calendar year 2008, for any monitors reporting data at any
point during the year. In 2008, there were 488 S02 monitors reporting data to AQS at
some point during the year.
2.2.2 Ambient concentrations ofS02
Since the integrated exposure to a pollutant is the sum of the exposures over all
time intervals for all environments in which the individual spends time, understanding
the temporal and spatial patterns of S02 levels across the U.S is an important
component of conducting air quality, exposure, and risk analyses. S02 emissions and
-------
ambient concentrations follow a strong east to west gradient due to the large numbers
of coal-fired electric generating units in the Ohio River Valley and upper Southeast
regions. In the 12 CMSAs that had at least 4 S02 regulatory monitors from 2003-2005,
24-hour average concentrations in the continental U.S. ranged from a reported low of
~1 ppb in Riverside, CA and San Francisco, CA to a high of ~12 ppb in Pittsburgh, PA and
Steubenville, OH (ISA, section 2.4.4). In addition, inside CMSAs from 2003-2005, the
annual average S02 concentration was 4 ppb (ISA, Table 2-8). However, spikes in hourly
concentrations occurred; the mean 1-hour maximum concentration was 130 ppb, with a
maximum value of greater than 700 ppb (ISA, Table 2-8).
In addition to considering 1-hour, 24-hour, and annual S02 levels, examining the
temporal and spatial patterns of 5-minute peaks of S02 is also important given that
human clinical studies have demonstrated exposure to these peaks can result in adverse
respiratory effects in exercising asthmatics (see REA, Chapter 4). Although the total
number of S02 monitors across the continuous U.S. can vary from year to year, in 2006
there were approximately 500 S02 monitors in the NAAQS monitoring network (ISA,
section 2.5.2). State and local agencies responsible for these monitors are required to
report 1-hour average S02 concentrations to the EPA Air Quality System (AQS).
However, a small number of sites, only 98 total from 1997 to 2007, and not the same
sites in all years, voluntarily reported 5-minute block average data to AQS (ISA, section
2.5.2). Of these, 16 reported all twelve 5-minute averages in each hour for at least part
of the time between 1997 and 2007. The remainder reported only the maximum 5-
minute average in each hour. When maximum 5-minute concentrations were reported,
the absolute highest concentration over the ten-year period exceeded 4000 ppb, but for
all individual monitors, the 99th percentile was below 200 ppb (ISA, section 2.5.2).
Medians from these monitors reporting data ranged from 1 ppb to 8 ppb, and the
average for each maximum 5-minute level ranged from 3 ppb to 17 ppb. Delaware,
Pennsylvania, Louisiana, and West Virginia had mean values for maximum 5-minute
data exceeding 10 ppb (ISA, section 2.5.2). Among aggregated within-state data for the
16 monitors from which all 5-minute average intervals were reported, the median
values ranged from 1 ppb to 5 ppb, and the means ranged from 3 ppb to 11 ppb (ISA,
section 2.5.2). The highest reported concentration was 921 ppb, but the 99th percentile
values for aggregated within-state data were all below 90 ppb (ISA, section 2.5.2).
-------
Chapter 3: Air Quality Analysis
Synopsis
This chapter describes the approach used to calculate 2020 baseline S02 design values
and the amount of emissions reductions needed to attain the alternative 1-hour S02 NAAQS.
The NAAQS being analyzed are 50, 75, and 100 ppb based on design values calculated using the
3-year average of the 99th percentile 1-hour daily maximum concentrations based on the
monitoring network described in Chapter 2. The projected 2020 baseline S02 design values are
used to identify 2020 nonattainment counties and to calculate, for each such county, the
amount of reduction in S02 concentration necessary to attain the alternative NAAQS. This
chapter also describes the approach for calculating "ppb S02 concentration per ton S02
emissions" ratios that are used to estimate the amount of S02 emissions reductions that may
be needed to provide for attainment of the alternative S02 standards. As described below, the
air quality analysis relies on S02 emissions from simulations of the Community Multiscale Air
Quality (CMAQ) model coupled with ambient 2005-2007 design values and emissions data to
project 2020 S02 design value concentrations and the "ppb per ton" ratios. A description of
CMAQ is provided in the Ozone NAAQS RIA Air Quality Modeling Platform Document (EPA,
2008).
3.1 2005-2007 Design Values
The proposed standard is based on the 3-year average of the 99th percentile
concentration of the daily 1-hour maximum concentration for a year. The design value for each
percentile is calculated as:
• Identify daily 1-hour maximum concentration for each day for each year
• Calculate 99th percentile values of the daily 1-hour maximum concentrations for each
year
• Average the 99th percentile values for the three years.
Monitors that had valid measurements for at least 75% of the day, 75% of the days in a
quarter and all 4 quarters for all three years were included in the analysis1. The resulting 3-year
averaged 99th percentile daily 1-hour maximum concentrations are shown in Figure 3.1 for 229
monitored counties. Counties in blue, green, and dark red would exceed the lowest alternative
standard considered in the RIA, 50 ppb. Monitors with design values of 50.0 to 50.4 ppb would
not exceed the standard 50 ppb as those concentrations would round to 50 ppb.
1 Email from Rhonda Thompson to James Thurman, January 22, 2009.
3-1
-------
Concentrations 50.5 ppb and higher are considered exceeding the lowest alternative standard.
Similar rounding is done for the 75, and 100 ppb alternative standards (75.4 and 100.4 are the
cut-offs for nonattainment). A summary of the number of counties exceeding the alternative
standards for 2005-2007 is shown in Table 3.1. Appendix 3 contains the complete list of 2005-
2007 design values used in calculation of the 2020 design values. Table 3.2 lists the top ten
counties for the 99th percentile design values for 2005-2007.
Figure 3.1. 2005-2007 3-year averaged design values (ppb) for 99th percentile daily 1-hour
maximum SO2 concentrations. Values shown are county maxima.
3-2
-------
Table 3.1. Number of monitors and counties exceeding 50, 75, and 100 ppb alternative
standards for the 99th percentile design values for 2005-07.
Alternative standard Number of monitors Number of counties
(ppb)
50 169 119
759570
1005946
Table 3.2. Top 10 2005-07 counties 99th percentile design values.
State
MO
AZ
IL
PA
TN
PA
IN
OH
Wl
IN
County
Jefferson
Gila
Tazewell
Warren
Blount
Northampton
Fountain
Lake
Oneida
Floyd
Design value (ppb)
350.6
286.0
222.3
214.0
196.3
187.0
183.0
180.3
179.0
176.3
3.2 Calculation of 2020 Projected Design Values
The 2020 baseline design values were determined using CMAQ gridded emissions for
2005 and 2020. Gridded emissions were utilized instead of county emissions because of the
influence of stationary sources on S02 concentrations. For monitors near county boundaries,
stationary sources in a neighboring county may have more influence over the monitor than a
stationary source in the monitor's home county. The S02 emissions in the CMAQ runs reflect
reductions from the following controls and programs shown in Table 3.3.
3-3
-------
Table 3.3. Controls in the 2020 SO2 inventory.
Control Strategies
Approach or
Reference:
Non-EGU Point Controls
Consent decrees apportioned to several plants
DOJ Settlements: plant SCC controls
Alcoa, TX 1
Premcor (formerly MOTIVA), DE
Refinery Consent Decrees: plant/SCC controls 2
Closures, pre-2007: plant control of 100%
Auto plants
Pulp and Paper
Large Municipal Waste Combustors
Small Municipal Waste Combustors
Plants closed in preparation for 2005 inventory
Small Municipal Waste Combustors (SMWC) 4
Solid Waste Rules (Section 129d/llld)
Hospital/Medical/Infectious Waste Incinerator Regulations EPA' 2005
MACT rules, plant-level, PM & SO2: Lime Manufacturing 5
Stationary Area Assumptions
Residential Wood Combustion Growth and Changeouts to year 2020 6
EGU Point Controls
Clean Air Interstate Rule 7; EPA, 2005
Onroad Mobile and Nonroad Mobile Controls (list includes all key mobile control strategies but is not
exhaustive)
Tier2Rule EPA, 1999
2007 Onroad Heavy-Duty Rule EPA, 2000
Final Mobile Source Air Toxics Rule (MSAT2) EPA, 2007
Renewable Fuel Standard EPA, 2010
Clean Air Nonroad Diesel Final Rule-Tier 4 g EpA 2Q04
Control of Emissions from Nonroad Large-Spark Ignition Engines and Recreational Engines
(Marine and Land Based): "Pentathalon Rule"
Clean Bus USA Program 8,9,10
Control of Emissions of Air Pollution from Locomotives and Marine Compression-Ignition
Engines Less than 30 Liters per Cylinder
Aircraft, Locomotives, and Commercial Marine Assumptions
Aircraft:
Itinerant (ITN) operations at airports to year 2020
Locomotives:
Energy Information Administration (EIA) fuel consumption projections for freight rail
Clean Air Nonroad Diesel Final Rule-Tier 4 EPA, 2009; 12; 9
Locomotive Emissions Final Rulemaking, December 17,1997
Control of Emissions of Air Pollution from Locomotives and Marine
3-4
-------
Control Strategies
Approach or
Reference:
Commercial Marine:
EIA fuel consumption projections for diesel-fueled vessels
OTAQ EGA C3 Base 2020 inventory for residual-fueled vessels
Clean Air Nonroad Diesel Final Rule-Tier 4 ' '
Emissions Standards for Commercial Marine Diesel Engines, December 29,1999
Tier 1 Marine Diesel Engines, February 28, 2003
1. For ALCOA consent decree, used http:// cfpub.epa.gov/compliance/cases/index.cfm;
for MOTIVA: used information sent by State of Delaware
2. Used data provided by Brenda Shine, EPA, OAQPS
3. Closures obtained from EPA sector leads; most verified using the world wide web.
4. Used data provided by Walt Stevenson, EPA, OAQPS
5. Percent reductions recommended are determined from the existing plant estimated
baselines and estimated reductions as shown in the Federal Register Notice for the
rule. SO2 % reduction will therefore be 6147/30,783 = 20% and PM10 and PM2.5
reductions will both be 3786/13588 = 28%
6. Expected benefits of woodstoves change-out program:
http://www.epa.gov/woodstoves/index.html
7. http://www.epa.gov/airmarkets/progsregs/epa-ipm/docs/summary2006.pdf
8. http://www.epa.gov/nonroad-diesel/2004fr.htm
9. http://www.epa.gov/cleanschoolbus/
10. http://www.epa.gov/otaq/marinesi.htm
11. Federal Aviation Administration (FAA) Terminal Area Forecast (TAF) System,
December 2007: http://www.apo.data.faa.gov/main/taf.asp
12. http://www.epa.gov/nonroad-diesel/2004fr.htm
In brief, these CMAQ emissions were at 12 km horizontal resolution for two modeling
domains which, collectively, cover the lower 48 States and adjacent portions of Canada and
Mexico. The boundaries of these two domains are shown in Figure 3.2. The spatial
distribution of the emissions for 2005 and 2020 can be seen in Figures 3.3 and 3.4 respectively.
In both figures, the lines radiating from the coast are the commercial marine vessel emissions.
Figure 3.5 shows the reduction in emissions between 2005 (16.3 million tons) and 2020 (9.6
million tons) by source sector (ECU, non-EGU point, commercial marine vessel, and other
sources) with the decrease from 2005 to 2020 due mostly to decreases in ECU emissions.
3.2.1 2020 Design Value Calculation Methodology
Ambient monitored data were assigned to CMAQ grid cells using ArcGIS. Since there
were areas of the country where the eastern and western domains overlapped, monitors in
these overlapping areas were assigned to the eastern or western grid cells by using a
"combined grid." This combined grid was a mesh of the eastern and western domains, with
overlapping areas assigned eastern grid cells or western grid cells based on the location relative
to the dividing line shown in Figure 3.2. Figure 3.2 shows the assignment of monitors to the
3-5
-------
two domains. An example of monitors in both domains was the El Paso County monitors.
These monitors were assigned to the western domain. The gridded 2006 and 2020 emissions
were also assigned to the combined grid based on the same grid assignments as the monitors.
Figure 3.2. Monitor domain assignments. Western domain is outlined in blue and eastern
domain outlined in red. Black vertical line denotes dividing line between eastern and
western domains for monitor assignments. Monitors in blue were assigned to the western
domain and monitors in red were assigned to the eastern domain.
3-6
-------
Figure 3.3. 2005 annual 12 km gridded SO2 emissions (tons).
Legend
o
1-7
8-26
27-93
94-336
337- 1210
| 1211 -4351
| 4352- 15639
[ 15640-56294
I 56205-201989
3-7
-------
Figure 3.4. 2020 annual 12 km gridded SO2 emissions (tons).
3-8
-------
Figure 3.5. 2005 and 2020 SO2 emissions (tons) by source sector.
1 Q -
ID
tit ie
C 16
0
+*
**~ 1A -
o 14
n 12 -
O \f-
in-
,_, 1 U
en
o 8 -
O o
en
en
E 6 -
HI
c
C , _
< 2
16.3
3.4
.5
2.1
in 4
2005
Year
9 a
.0
2 9
0.2
1.9
4E
,v
2020
D Other
Df^rt m m^r^* ial Marino
DNon-EGU point
DEGU
Once the monitors and emissions were assigned to the combined grid, for each monitor,
a 9x9 matrix of grid cells was selected, centered on the monitor's grid cell. An example is
shown in Figure 3.6. The 9x9 matrix represented an approximate domain of emissions
extending out 50 km from the monitor, the upper range of near-field dispersion. Since the
design values were based on hourly concentrations, extending the radius of influential
emissions on the monitor grid cell to 50 km was considered appropriate.
3-9
-------
Figure 3.6. 9x9 matrix of 12km grid cells centered on CMAQ cell containing an SO2 monitor
(star).
10
20
: Kilometers
Once the matrices of grid cells were created for each monitor, the 2005 and 2020
gridded emissions were summed for each year across the 81 grid cells to result in total 2005
and 2020 emissions for each monitor. The summed 2020 emissions were then divided by the
2005 emissions to get an emissions change ratio:
f,
- rat/o
-2020
-2005
(3.1)
Where £2020 are the summed 81 grid cell emissions for 2020, E2oos are the summed 81
grid cell emissions for 2005 and Eratio is the ratio of 2020 emissions to 2005 emissions.
,th
The 2005-2007 99 percentile design value concentrations were then multiplied by the
emissions ratio to calculate the 2020 design values.
DV.
2020"99
" ^^2005-2007:99 X *- ratio
(3.2)
3-10
-------
Where Eratj0 is as defined above, DV20o5-2oo7:99 is the 2005-2007 3-year averaged design
value for the 99th percentile, and DV202o:99 is the projected 2020 design value for the 99th
percentile.
After calculating the 2020 design values, a ppb/ton estimate was calculated by:
(n\/ — n\/ } /
Ppb I tOngg = \ 20205:99 u " 2005-2007:99 )/, , (33)
/ (p 2020 ~t 2005 /
Where E202o and E20os are the summed emissions as defined for Equation 3.1, DV20os-
2007:99 and DV202o:99 are as defined above and ppb/ton99 is the ppb/ton estimate for the 99th
percentile.
Residual nonattainment estimates for the three alternative standards of 50, 75, and 100
ppb were calculated by subtracting the alternative standard from the 2020 design value. The
absolute values of the alternative standards (50, 75, or 100 ppb) were not subtracted but rather
the highest value that would meet the standards (50.4, 75.4, and 100.4 ppb) if design values
were rounded to the nearest whole ppb. Once residual nonattainment was calculated for each
alternative standard, for monitors exceeding the standards, tons needed for control were
calculated by dividing residual nonattainment by the ppb/ton estimate:
A/A
ppb I ton
99
Where ppb/ton99 is as defined above, NA99:As is the residual nonattainment for
alternative standard AS (50, 75, or 100 ppb) for the 99th percentile, and Tons99:Asare the tons
needed to reach attainment for alternative standard AS for the 99th percentile.
3.2.2 Methodology Limitations
While the approach described in Section 3.2.1 is reasonable for a national analysis, there
are limitations to the approach that may be better addressed by other methods such as near-
field dispersion modeling on a case by case basis or fine scale CMAQ modeling. Given the
number of monitors in the analysis, dispersion modeling for all monitors would not be feasible.
Also, given that the CMAQ concentrations associated with the emissions used in this analysis
are at 12 km horizontal resolution and that S02 is affected by nearby stationary sources, the
CMAQ results may not be reasonable for this analysis, due to allocation of individual emission
points within the grid cell. Limitations of this analysis include:
• Distance from source to monitor is not factored in the emissions sums used in Equation
3.1. All emission sources, regardless of distance and tonnage, are weighted equally.
3-11
-------
Using Figure 3.6 as an example, a source may be located in the most northwestern grid
cell and a source may be located in the same grid cell that contains the monitor. No
distance weighting is applied to either source, based on its proximity to the monitor.
They are both added to the emissions sum as is. Some monitors' emission sums may
include large emission sources that are farther away from the monitor than smaller
emission sources but the large emissions sources dominate the emissions used to
calculate the ratio in Equation 3.1. These large sources, may have large changes in
emissions from 2005 to 2020 and these changes could drastically affect the emissions
ratio. Given the nature of the projection approach described in Section 3.2.1, these
large emission changes may overestimate or underestimate the concentration change at
the monitor given the distance from the source to the monitor and the factors
mentioned in the points below, meteorology and terrain.
• Meteorology and terrain influences are not factored into the analysis. A source may not
have a significant impact on a monitor because the prevailing wind direction is not from
the source to the monitor, or the terrain between the source and monitor is configured
such that the source does not have a significant impact on the monitor. This would also
depend on building downwash effects and stack parameters such as stack height, exit
temperature, stack diameter, and exit velocity.
3.3 Results
3.3.1. Nonattainment results
Table 3.4 lists the number of monitors and counties exceeding the three alternative standards
for the 99th percentile 2020 design values. The number of counties exceeding each of the
alternative standards decreased from 2005-2007 to 2020. Figure 3.7 shows the maximum 2020
design value for monitored counties for the 99th percentile design values. Counties in blue,
green, and scarlet exceed the 50 ppb alternative standard. Table 3.5 lists the top 10 counties in
2020 for the 99th percentile design value along with residual nonattainment and tons needed
for control to meet attainment. A complete list of 2020 design values for all monitors can be
found in Appendix 3.
3-12
-------
Table 3.4. Number of monitors and counties exceeding 50, 75, and 100 ppb alternative
standards for the 99th percentile design values for 2020.
Alternative standard
(ppb)
Number of monitors
Number of counties
50
71
56
75
27
24
100
11
Figure 3.7. 2020 design values (ppb) for 99th percentile daily 1-hour maximum SO2
concentrations. Values shown are county maxima.
3-13
-------
,th
Table 3.5. Top 10 2020 counties 99 percentile design values (ppb).
State
Alternative standards (ppb)
50
75
100
Cou nty
2020 DV
Residual
nonattainment
Tons for
control
Residual
nonattainment
Tons for
control
Residual
nonattainment
Tons for
control
MO
Jefferson
285.5
235.1
139,033
210.1
124,249
185.1
109,464
AZ
Gila
284.8
234.4
21,930
209.4
19,591
184.4
17,252
PA
Warren
217.2
166.8
10,379
141.8
8,824
116.8
7,268
Wl
Oneida
175.3
124.9
6,866
99.9
5,491
74.9
4,117
TN
Montgomery
144.3
93.9
19,764
68.9
14,502
43.9
9,240
IN
Wayne
134.3
83.9
24,088
58.9
16,911
33.9
9,733
IA
Muscatine
126.2
75.8
27,365
50.8
18,340
25.8
9,314
OK
Muskogee
104.9
54.5
45,542
29.5
24,651
4.5
3,760
OH
Summit
103.9
53.5
26,690
28.5
14,218
3.5
1,746
PA Northampton 100.4
50.0
20,652
25.0
10.326
3-14
-------
3.3.2 Example monitors
This section describes the emissions changes for two monitors' 99th percentile design
values shown in Figures 3.8 and 3.9. One monitor's design value, Tazewell County, IL decreased
from 2005-2007 to 2020 (Figure 3.8) and the other monitor's (Montgomery County, TN) design
value increased from 2005-2007 to 2020 (Figure 3.9). Emissions summaries in the 81 cell
matrices for both monitors are shown in Figure 3.10.
Figure 3.8. Location of monitor in Tazewell County, IL.
2005-07 DV: 222.3
2020 DV: 89.3
3-15
-------
Figure 3.9. Location of monitor in Montgomery County, TN.
2005-07 DV: 115.6
2020 DV: 144.3
3-16
-------
Figure 3.10. Tazewell County, IL and Montgomery County, TN monitors emissions (tons) for
2005 and 2020.
2,218
Tazewell 2005 emissions: 94,030 tons
1,013
Tazewell 2020 emissions: 37,799 tons
Montgomery 2005 emissions: 24,284 tons
Montgomery 2020 emissions: 30,325 tons
3.3.2.1
Tazewell County
Emissions affecting the Tazewell County monitor decreased from approximately 94,000
tons in 2005 to approximately 38,000 tons in 2020 (Figure 3.10 a and b). The decrease was
mostly due to decreases in ECU emissions. The decrease caused the ECU sector drop from
about 75% of the emissions to around 40% of the emissions. Figure 3.11 shows the spatial
distribution of 2005 total emissions (all sources) within 50 km of the monitor and Figure 3.12
shows the spatial distribution of 2020 total emissions within 50 km of the monitor. The
decrease in emissions can be seen as the emissions become more uniform outside of the
"hotspot" grid cells.
3-17
-------
Figure 3.11. 2005 12 km grid cell SO2 total emissions (tons) for Tazewell County monitor. The
red star represents the monitor location.
5- 12
13-15
16-22
23-49
50- 145
146-430
481 - 1662
1663-5327
5S2B - 20496
20497-7217
3-18
-------
Figure 3.12. 2020 12 km grid cell SO2 total emissions (tons) for Tazewell County monitor. The
red star represents the monitor location.
0- 12
13-15
16-22
23-49
50- 145
146-430
481 - 1662
1663-5327
5S2B - 20496
20497-7217
3-19
-------
3.3.2.2 Montgomery County
The design value for Montgomery County increased from 2005-07 to 2020 due to an
increase in ECU emissions (Figure 3.10 c and d). Figures analogous to Figure 3.11 and Figure
3.12 are shown in Figure 3.13 and Figure 3.14. While emissions decrease outside the "hotspot"
grid cells, the emissions within those hotspots increase from 2005 to 2020, as these are the
locations of ECU facilities and the emissions increase from 2005 to 2020.
3-20
-------
Figure 3.13. 2005 12 km grid cell SO2 total emissions (tons) for Montgomery County monitor.
The red star represents the monitor location.
\
3-21
-------
Figure 3.14. 2020 12 km grid cell SO2 total emissions (tons) for Montgomery County monitor.
The red star represents the monitor location.
3-22
-------
3.4 Summary
In summary, 2020 baseline N02 design value concentrations were projected from 2005-
2007 observed design values using CMAQ emissions output from 2005 and 2020. Results of the
projections showed that, in 2020, nonattainment occurred for all three alternative standards
(50, 75, and 100 ppb). However, the number of counties exceeding the standards dropped from
the 2005-2007 period.
3.5 References
U.S. Environmental Protection Agency (EPA). 1999. Air Pollution from New Motor Vehicles: Tier
2 Motor Vehicle Emissions Standards and Gasoline Sulfur Control, U.S. Environmental
Protection Agency, Office of Transportation and Air Quality, Assessment and Standards
Division, Ann Arbor, Ml 48105, EPA420-R-99-023, December 1999. Available at
http://www.epa.gov/tier2/frm/ria/r99023.pdf.
U.S. Environmental Protection Agency (EPA). 2000. Regulatory Impact Analysis: Heavy-Duty
Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements, U.S.
Environmental Protection Agency, Office of Transportation and Air Quality, Assessment and
Standards Division, Ann Arbor, Ml 48105, EPA420-R-00-026, December 2000. Available at
http://www.epa.gov/otaq/highway-diesel/regs/exec-sum.pdf
U.S. Environmental Protection Agency (EPA). 2004. Final Regulatory Analysis: Control of
Emissions from Nonroad Diesel Engines, U.S. Environmental Protection Agency, Office of
Transportation and Air Quality, Assessment and Standards Division, Ann Arbor, Ml 48105,
EPA420-R-04-007, May 2004. Available at http://www.epa.gov/nonroaddiesel/
2004fr/420r04007.pdf.
U.S. Environmental Protection Agency (EPA). 2005. Clean Air Interstate Rule Emissions Inventory
Technical Support Document, U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, March 2005. Available at
http://www.epa.gov/cair/pdfs/finaltech01.pdf.
U.S Environmental Protection Agency (EPA), 2007. Control of Hazardous Air Pollutants from
Mobile Sources; Final Rule. 40 CFR Parts 59, 80,85, 86, et al. Available at
http://www.epa.gov/fedrgstr/EPA-AIR/2007/February/Day-26/a2667a.pdf
3-23
-------
U.S. Environmental Protection Agency (EPA). 2008. Air Quality Modeling Platform for the Ozone
National Ambient Air Quality Standard Final Rule Regulatory Impact Analysis. North
Carolina. EPA-454/R-08-003.
U.S. Environmental Protection Agency (EPA), 2009. Regulatory Impact Analysis: Control of
Emissions of Air Pollution from Locomotive Engines and Marine Compression Ignition Engines
Less than 30 Liters Per Cylinder. U.S. Environmental Protection Agency Office of
Transportation and Air Quality, Assessment and Standards Division, Ann Arbor, Ml 48105,
EPA420-R-08-001a, May 2009. Available at:
http://www.epa.gov/otaq/regs/nonroad/420r08001a.pdf
U.S. Environmental Protection Agency (EPA), 2010. RFS2 Emissions Inventory for Air Quality
Modeling Technical Support Document, February, 2010. Available at:
http://www.epa.gov/otaq/renewablefuels/420rl0005.pdf.
3-24
-------
Appendix 3a: 2005-2007 and 2020 Design Values
Table 3a-l lists the 2005-2007 design values used in projecting 2020 design values for all
monitors meeting the completeness criteria described in Section 3.1 of Chapter 3. Design
values in black are below the 50 ppb alternative standard. Design values in blue exceed the 50
ppb alternative standard but are below 75 ppb. Design values in green exceed the 75 ppb
alternative standard but are below 100 ppb. Values in red exceed 100 ppb. Exceedances of the
alternative standards are based on the criteria discussed in Section 3.1 of Chapter 3.
Table 3a-l. SO2 2005-2007 and 2020 projected 99th percentile design values (ppb).
State
AL
AZ
AZ
AZ
AZ
AZ
AR
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
County
Jefferson
Gila
Gila
Maricopa
Maricopa
Pima
Pulaski
Contra Costa
Contra Costa
Contra Costa
Contra Costa
Contra Costa
Contra Costa
Imperial
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Orange
Sacramento
Sacramento
San Bernardino
San Bernardino
San Bernardino
San Diego
San Francisco
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Monitor
1003
9
1001
3002
3003
1011
7
2
6
1002
1004
2001
3001
5
1002
1103
4002
5005
1003
2
6
306
1234
2002
1
5
8
1013
1020
1025
2004
2011
2005-07
63.3
131.6
286.0
14.0
9.3
14.0
10.0
18.6
18.0
12.3
14.6
22.6
25.6
20.9
6.6
10.6
27.6
19.6
9.3
5.0
5.6
10.0
11.3
8.0
9.6
15.3
4.0
4.6
44.3
8.0
5.6
3.3
2020
19.3
131.2
284.8
4.1
2.8
16.5
12.5
12.5
11.6
8.1
9.4
14.8
17.2
20.4
4.0
6.3
15.6
11.6
5.4
4.5
5.1
8.2
19.6
7.2
8.6
9.9
0.6
2.0
6.7
1.3
1.6
0.5
3-1
-------
State
CA
CA
CO
CT
CT
CT
CT
CT
DE
DE
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
GA
GA
GA
GA
GA
ID
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
County
Santa Barbara
Solano
Denver
Fairfield
Fairfield
Fairfield
New Haven
New Haven
New Castle
New Castle
Broward
Duval
Duval
Duval
Escambia
Hamilton
Hillsborough
Hillsborough
Hillsborough
Hillsborough
Orange
Pinellas
Pinellas
Pinellas
Pinellas
Putnam
Chatham
Chatham
Floyd
Fulton
Fulton
Bannock
Cook
Cook
Cook
Cook
Cook
Macon
Macoupin
Madison
Madison
Madison
Peoria
Randolph
St. Clair
Sangamon
Tazewell
Wabash
Wabash
Monitor
4003
4
2
12
1123
9003
27
2123
1008
2004
10
80
81
97
4
15
81
95
109
1035
2002
23
3002
5002
5003
1008
21
1002
3
48
55
4
50
63
76
1601
4002
13
2
1010
3007
3009
24
1
10
6
4
1
1001
2005-07
2.6
10.0
32.6
35.6
25.3
27.6
60.6
27.8
125.0
49.6
64.6
21.3
69.0
42.0
76.3
31.6
47.3
42.6
119.0
71.3
11.3
96.3
42.0
77,6
83.3
51.6
62.3
94,6
110.0
73.0
60.0
69.6
37.0
40.6
45.6
104.0
68.3
47.0
27.0
83.6
59.0
142.0
73.6
29.6
91.3
110.6
222.3
152.3
125.3
2020
1.3
6.5
66.8
46.4
24.2
29.4
60.9
22.8
48.7
23.0
35.4
17.6
57.0
34.5
26.7
24.5
20.6
19.1
53.5
32.1
4.7
36.4
15.8
27.8
43.2
11.7
57.5
87,4
10.2
10.2
22.7
61.7
27.7
29.2
33.3
63.7
48.9
48.6
13.8
52.6
37.1
89,4
31.1
20.9
59.4
99,3
89,3
40.5
33.3
3-2
-------
State
IL
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IA
IA
IA
IA
IA
IA
IA
IA
IA
KS
KS
KS
KS
KY
KY
KY
KY
KY
KY
LA
LA
LA
County
Will
Daviess
Dearborn
Floyd
Floyd
Floyd
Fountain
Gibson
Hendricks
Jasper
Lake
Lake
La Porte
Marion
Marion
Marion
Morgan
Pike
Porter
Spencer
Vanderburgh
Vanderburgh
Vigo
Vigo
Warrick
Wayne
Wayne
Cerro Gordo
Clinton
Linn
Linn
Muscatine
Muscatine
Muscatine
Scott
Van Buren
Montgomery
Sumner
Trego
Wyandotte
Boyd
Daviess
Greenup
Jefferson
Livingston
McCracken
Bossier
Calcasieu
East Baton Rouge
Monitor
13
2
4
4
7
1004
1
1
2
2
22
2008
5
42
57
73
1001
5
11
10
12
1002
18
1014
2
6
7
18
19
29
31
16
17
20
15
6
6
2
1
21
17
5
7
1041
4
1024
8
8
9
2005-07
64.6
112.6
109.6
140.3
159.6
176.3
183.0
108.6
41.0
57.0
92.0
42.6
27.3
92.3
117.3
62.0
129.6
19.3
63.6
60.0
67.3
35.0
93.6
125.0
148.3
106.7
84.1
13.2
48.3
46.0
88.6
122.1
65.5
165.1
27.6
6.9
16.6
8.6
4.3
50.0
60.3
71.0
46.0
150.6
53.3
26.3
20.6
42.3
65.3
2020
32.0
36.5
36.4
52.7
59.9
66.2
56.0
28.8
19.5
56.9
81.8
32.8
27.0
36.2
45.5
24.4
52.5
6.2
59.6
15.9
18.9
9.1
28.4
31.8
38.3
134.3
105.9
12.3
41.3
48.8
94.0
91.7
50.0
126.2
21.0
6.8
15.0
4.7
2.1
33.2
19.1
20.0
13.3
73.4
53.5
26.2
16.7
36.1
54.6
3-3
-------
State
LA
ME
MD
MA
MA
MA
MA
MA
MA
MA
MA
MA
MN
MN
MN
MN
MN
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MT
MT
MT
NE
NE
NV
NH
NH
NH
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NM
County
Ouachita
Hancock
Baltimore
Bristol
Hampden
Hampshire
Suffolk
Suffolk
Suffolk
Suffolk
Suffolk
Worcester
An oka
Dakota
Dakota
Dakota
Dakota
Greene
Greene
Greene
Greene
Greene
Jackson
Jefferson
St. Louis
St. Louis city
St. Louis city
Yellowstone
Yellowstone
Yellowstone
Douglas
Douglas
Clark
Hillsborough
Merrimack
Rockingham
Atlantic
Bergen
Burlington
Camden
Camden
Cumberland
Gloucester
Hudson
Hudson
Middlesex
Morris
Union
Eddy
Monitor
4
103
3001
1004
16
4002
2
20
21
40
42
23
1002
20
423
441
442
26
32
37
40
41
34
4
3001
7
86
16
1065
2005
53
55
539
20
1006
14
5
5001
1001
3
1001
7
2
6
1002
2003
3001
4
1004
2005-07
22.3
6.3
99.3
64.3
39.0
17.0
26.6
23.0
32.3
40.3
27.3
20.6
21.3
18.0
14.0
7.0
8.0
67.6
25.0
90.6
81.3
25.6
156.3
350.6
49.6
56.6
67.6
40.0
68.0
54.6
f
18.6
8.0
58.3
157.0
59.6
19.0
29.3
27.6
38.0
26.6
23.0
32.6
42.0
47.6
29.3
36.0
51.0
4.6
2020
20.4
5.4
43.3
21.5
29.7
13.0
17.1
14.7
20.6
25.9
17.5
17.7
10.4
7.2
5.6
2.8
3.2
48.0
17.7
65.0
58.3
18.3
97,4
285.5
34.6
40.3
47.2
46.3
73.3
58.8
87,6
18.2
6.3
20.6
51.8
28.3
11.7
21.6
12.8
16.7
13.3
8.6
13.9
33.7
38.2
12.1
14.4
23.2
4.6
3-4
-------
State
NM
NM
NM
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NC
NC
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
County
Grant
San Juan
San Juan
Albany
Chautauqua
Chautauqua
Chemung
Erie
Erie
Essex
Franklin
Hamilton
Herkimer
Madison
Monroe
New York
Niagara
Onondaga
Putnam
Queens
Schenectady
Suffolk
Ulster
Beaufort
New Hanover
Billings
Burke
Cass
Dunn
McKenzie
McKenzie
McKenzie
Mercer
Mercer
Mercer
Mercer
Mercer
Oliver
Williams
Adams
Allen
Ashtabula
Butler
Butler
Clark
Columbiana
Cuyahoga
Cuyahoga
Cuyahoga
Monitor
1003
9
1005
12
6
11
3
5
4002
3
4
5
5
6
1007
56
2008
1015
5
124
3
9
1005
6
6
2
4
1004
3
2
104
111
4
102
118
123
124
2
103
1
2
1001
4
1004
3
22
45
60
65
2005-07
4.0
12.6
77.0
22.0
61.4
32.1
24.6
30.6
118.6
9.9
9.1
10.3
9.8
20.0
52.0
62.6
21.7
17.0
21.9
44.0
23.0
56.0
15.5
47.3
87.6
6.3
29.4
5.5
11.6
11.0
17.6
25.6
35.0
35.3
34.3
39.0
37.3
56.3
44.3
88.3
22.3
36.6
72.0
57.3
40.0
121.3
65.0
84.3
87.0
2020
2.1
5.3
33.0
21.0
41.5
28.7
24.8
16.4
75.9
9.2
8.3
9.2
8.8
27.2
58.6
44.3
13.8
39.8
20.0
33.4
21.9
75.6
15.2
45.9
58.4
3.1
29.2
4.1
8.8
5.6
12.3
16.9
18.8
19.0
18.5
21.0
21.7
30.4
37.3
21.8
19.6
30.3
29.0
23.6
62.8
42.7
35.2
45.7
47.2
3-5
-------
State
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OK
OK
OK
OK
OK
OK
OK
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
County
Franklin
Hamilton
Jefferson
Lake
Lake
Lawrence
Lucas
Lucas
Mahoning
Meigs
Scioto
Scioto
Summit
Summit
Tuscarawas
Kay
Kay
Muskogee
Oklahoma
Tulsa
Tulsa
Tulsa
Allegheny
Allegheny
Allegheny
Allegheny
Beaver
Beaver
Blair
Bucks
Cambria
Centre
Dauphin
Erie
Indiana
Lackawanna
Lancaster
Lawrence
Lehigh
Lycoming
Mercer
Montgomery
Northampton
Northampton
Perry
Philadelphia
Schuylkill
Warren
Warren
Monitor
34
10
17
3
3002
6
8
24
13
1001
13
20
17
22
6
602
9010
167
1037
175
235
501
10
21
64
67
2
14
801
12
11
100
401
3
4
2006
7
15
4
100
100
13
25
8000
301
55
3
3
4
2005-07
41.6
123.6
175.6
53.3
180.3
53.3
68.3
53.3
63.0
98.6
36.6
51.8
108.0
62.0
71.0
40.3
14.6
65.6
6.6
65.3
61.3
48.6
71.3
73.0
142.0
67.0
140.0
69.0
58.6
37.3
86.3
31.0
64.6
54.0
111.3
40.6
66.0
95.0
52.6
50.3
45.3
32.3
46.6
187.0
33.6
40.0
55.3
63.0
214.0
2020
14.9
49.9
52.6
27.1
94.7
15.4
32.4
25.3
48.4
25.3
20.6
17.4
103.9
59.6
15.8
67.8
24.3
104.9
4.8
51.3
48.2
38.2
18.4
31.5
60.0
22.5
48.1
34.2
57.2
17.3
34.4
25.8
15.7
30.4
47.0
20.5
19.5
44.0
30.1
7.0
30.6
16.4
26.3
100.4
6.4
17.4
10.1
63.9
217.2
3-6
-------
State
PA
PA
PA
PA
PA
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SD
SD
SD
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
UT
UT
VT
VA
County
Washington
Washington
Washington
Westmoreland
York
Barnwell
Charleston
Charleston
Georgetown
Greenville
Greenville
Lexington
Oconee
Richland
Richland
Custer
Jackson
Minnehaha
Blount
Blount
Bradley
Davidson
Montgomery
Montgomery
Shelby
Shelby
Sullivan
Sullivan
Dallas
El Paso
El Paso
Galveston
Gregg
Harris
Harris
Harris
Harris
Harris
Harris
Jefferson
Jefferson
Kaufman
Nueces
Nueces
Nueces
Davis
Salt Lake
Rutland
Charles City
Monitor
5
200
5001
8
8
1
3
46
6
8
9
8
1
7
1003
132
1
7
2
6
102
11
6
106
46
1034
7
9
69
37
53
5
1
46
51
62
70
1035
1050
9
11
5
25
26
32
4
1001
2
2
2005-07
79,6
79,6
90,0
76,6
104.0
17.0
37.3
23.6
55.0
27.0
25.0
96,3
20.0
28.6
36.3
4.3
3.6
18.0
196.3
84.9
85,3
23.6
53.0
115.6
65.3
81.3
170.6
141.8
11.6
9.3
12.6
59.0
78.3
34.0
31.0
55.3
68.6
74.6
17.3
123.0
94,6
15.3
24.0
8.0
36.0
22.6
32.0
48.2
88,6
2020
32.8
20.0
29.6
30.3
30.7
19.1
24.4
9.6
14.2
15.8
14.6
68.9
17.7
20.1
25.5
3.2
1.5
15.2
60.0
25.6
80,2
26.1
66.1
144.3
49.0
56.5
88,2
73.3
10.3
9.1
12.4
42.9
38.9
27.4
24.9
43.7
54.3
58.9
12.7
98,9
74.9
13.4
12.4
4.1
18.7
24.1
34.5
45.5
24.9
3-7
-------
State
VA
VA
VA
VA
VA
VA
VA
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
Wl
Wl
County
Fairfax
Fairfax
Fairfax
Rockingham
Alexandria city
Hampton city
Richmond city
Brooke
Brooke
Brooke
Cabell
Hancock
Hancock
Hancock
Hancock
Hancock
Hancock
Kanawha
Marshall
Monongalia
Wood
Brown
Oneida
Monitor
5
1005
5001
3
9
4
24
5
7
11
6
5
7
8
9
15
1004
10
1002
3
1002
5
996
2005-07
25.6
37.0
37.3
14.6
55.3
64.0
62.0
150.3
164.6
155.3
41.6
164.0
132.0
115.3
136.6
121.3
135.6
88.0
155.0
171.3
130.6
74.3
179.0
2020
6.8
8.2
14.6
13.0
12.2
46.3
15.2
45.0
49.3
46.5
7.4
56.3
42.4
40.6
43.9
42.7
43.6
22.4
41.8
41.5
37.8
64.7
175.3
3-8
-------
Chapter 4: Emissions Controls Analysis - Design and Analytical Results
Synopsis
This chapter documents the illustrative emission control strategy we applied to simulate
attainment with the alternative standards being analyzed for the final S02 NAAQS. Section 4.1
describes the approach we followed to select emissions controls to simulate attainment in each
geographic area of analysis. Section 4.2 summarizes the emission reductions we simulated in
each area based on current knowledge of identified emission controls, while Section 4.3
presents the air quality impacts of these emissions reductions. Section 4.4 discusses the
application of additional controls, beyond the level of control already assumed to be in place
for the analysis year1, that we estimate will be necessary to reach attainment in certain monitor
areas. Section 4.5 discusses key limitations in the approach we used to estimate the optimal
control strategies for each alternative standard.
The final rule will set a new short-term S02 primary standard based on the average of
the 99th percentile of 1-hour daily maximum concentrations from three consecutive years of
data. This new standard will be set at 75 parts per billion (ppb). OMB Circular A-4 requires the
RIA to contain, in addition to analysis of the impacts of the final NAAQS, analysis of a level more
stringent and a level less stringent than the final NAAQS. For a more stringent standard level,
we chose an alternative primary standard of 50 parts per billion (ppb). We also include
analyses for a less stringent standard, 100 ppb.
For the range of alternative standards, we analyzed the impact that additional emissions
controls applied to numerous sectors would have on predicted ambient S02 concentrations,
incremental to the baseline set of controls. Thus the analysis for a revised standard focuses
specifically on incremental improvements beyond the current standards, and uses control
options that might be available to states for application by 2020. The hypothetical control
strategy presented in this RIA is one illustrative option for achieving emissions reductions to
move towards a national attainment of a tighter standard. It is not a recommendation for how
a tighter S02 standard should be implemented, and states will make all final decisions regarding
implementation strategies once a final NAAQS has been set.
Generally, we expect that the nation will be able to make significant progress towards
attainment of a tighter S02 NAAQS without the addition of new controls beyond those already
being planned for the attainment of existing PM2.5 standards by the year 2020. As States
1 Note that the baseline or starting point for this analysis includes rules that are already "on the books" and will
take affect prior to the analysis year, as well as control strategies applied in the recent PM and Ozone NAAQS RIAs.
4-1
-------
develop their plans for attaining these existing standards, they are likely to consider adding
controls to reduce sulfur dioxide, as S02 is a precursor to both PM2.5. In addition, proposed
standards such as the Portland cement NESHAP, the ICI boilers NESHAPs, and the eventual rule
to replace the existing CAIR may also yield in total considerable additional reductions of S02
emissions if they are implemented as proposed. These controls will also directly help areas
meet a tighter S02 standard.
As part of our economic analysis of the tighter S02 standard, our 2020 analysis baseline
assumes that States will put in place the necessary control strategies to attain the current PM2.5
standards. The cost of these control strategies was included in the RIAs for those rulemakings.
We do not include the cost of those controls in this analysis, in order to prevent counting the
cost of installing and operating the controls twice. Of course, the health and environmental
benefits resulting from installation of those controls were attributed to attaining those
standards, and are not counted again for the analysis of this S02 standard.
In addition, we include the S02 control requirements for Category 3 (C3) marine vessels
that will be affected by a new mobile source rule promulgated by EPA in December 2009.2
These requirements call for changes in the diesel fuel program to allow for use of lower sulfur
fuel (1,000 ppm sulfur content) in U.S.-flagged C3 marine vessels beginning in 2011. Reductions
of S02 associated with this final rule are included in our 2020 analysis baseline. Thus, we
estimate no costs or benefits associated with these reductions.
It is important to note also that this analysis does not attempt to estimate attainment or
nonattainment for any areas of the country other than those counties currently served by one
of the 349 monitors in the current network. Chapter 3 explains that the current network is
focused on longer terms indicators that that included in this final rule.
Finally, we note that because it was not possible, in this analysis, to bring all areas into
attainment with the alternative standards in all areas using only identified (or known) controls,
EPA conducted a second step in the analysis, and estimated the cost of further tons of emission
reductions needed to attain the alternative primary NAAQS. It is uncertain what controls States
would put in place to attain a tighter standard, since additional abatement strategies are not
currently recognized as being commercially available. We should also note that because of data
and resource limitations, we are not able to adequately represent in this analysis the impacts of
some local emission control programs such as discussed in Chapter 3.
2 Control of Emissions from New Marine Compression-Ignition Engines at or Above 30 Liters per Cylinder. Signed
on December 18, 2009. For more information on this final rule and its RIA, please refer to
http://www.epa.gov/otaq/oceanvessels.htm.
4-2
-------
4.1 Developing the Identified Control Strategy Analysis
The 2020 baseline air quality estimates revealed that 27 monitors in 24 counties had
projected design values exceeding 75 ppb. We then developed a hypothetical control strategy
that could be adopted to bring the current highest emitting monitor in each of those counties
into attainment with a primary standard of 75 ppb, as well as additional target levels of 50 ppb
and 100 ppb, by 2020. (For more information on the development of the air quality estimates
for this analysis see Chapter 3.) Controls for three emissions sectors were included in the
control analysis: Non-Electricity Generating Unit Point Sources (nonEGU), Non-Point Area
Sources (Area), and Electricity Generating Unit Point Sources (ECU). Each of these sectors is
defined below for clarity.
• NonEGU point sources as defined in the National Emissions Inventory (NEI) are
stationary sources that emit 100 tons per year or more of at least one criteria
pollutant. NonEGU point sources are found across a wide variety of industries, such
as chemical manufacturing, cement manufacturing, petroleum refineries, and iron
and steel mills.
• Area Sources3 are stationary sources that are too numerous or whose emissions are
too small to be individually included in a stationary source emissions inventory.
Area sources are the activities where aggregated source emissions information is
maintained for the entire source category instead of each point source, and are
reported at the county level.
• Electricity Generating Unit Point Sources are stationary sources of 25 megawatts
(MW) capacity or greater producing and selling electricity to the grid, such as fossil-
fuel-fired boilers and combustion turbines.
It should be noted that no additional S02 controls beyond our baseline are applied to
onroad and nonroad mobile sources because mobile source measures to reduce sulfur content
from diesel engine rules will be well-applied in onroad and nonroad mobile source fleets by
2020, and thus there is little capability to achieve further reductions for this analysis beyond
those described in this report.
We began the control strategy analysis by applying controls to EGUs first before
applying controls to other sources. We applied controls in this sequence for the following
reasons: 1) there are many more S02 emissions from EGUs than from non-EGU sources in the
areas included in this analysis, and 2) S02 reductions from EGUs are less costly than from other
3 Area Sources include the nonpoint emissions sector only.
4-3
-------
source categories included in this analysis. Chapter 6 provides a table showing that the ECU
control costs for S02 as estimated for this analysis have a lower annual cost/ton compared to
those from the non-EGU point and area source categories.
The air quality impact of the needed emissions reductions was calculated using impact
ratios as discussed further in Chapter 3. The results of analyzing the control strategy indicate
that there were four areas projected not to attain 75 ppb in 2020 using all identified control
measures. To complete the analysis, EPA then extrapolated the additional emission reductions
required to reach attainment. The methodology used to develop those estimates and those
calculations are presented in Section 4.4.
4.1.1 Controls Applied for ECU Sector
The baseline in this RIA for EGUs accounts for extensive reductions in S02 emissions
from EGUs as implemented in the Clean Air Interstate Rule (CAIR).4 While the US District Court
for District of Columbia has remanded the CAIR, it still is in full effect. The Agency is working at
this time on a proposal to replace the CAIR, but that proposal is not yet complete. No
additional controls for S02 from EGUs are implemented in the baseline.
The Integrated Planning Model (IPM) was used to develop the baseline emissions for the
control strategy applied for the alternative standards. Historically, EPA has used the IPM model
to assess the cost and effectiveness of additional ECU controls for a large number of
rulemakings (e.g., CAIR, NOx SIP call, Ozone NAAQS, etc.). For this RIA, we applied controls on
a unit by unit basis to obtain reductions from units that contribute to nonattainment at
violating monitors in 2020. The end result of this approach mimics an approach which could be
used by individual states as they try to apply targeted controls on EGUs which affect attainment
in a specific area.
In this analysis, ECU controls were applied to uncontrolled coal-fired units of size 25
MW and larger within the 50 km radius of violating monitors. Each unit was retrofitted with a
Wet Flue Gas Desulfurization (FGD) scrubber with 95 percent S02 reduction efficiency. This
control measure is applicable to coal-fired EGUs with unit capacities above 25 MW.5 More
4 For more information on the CAIR rule, please refer to http://www.epa.gov/airmarkt/progsregs/cair/.
5 Costs of FGD scrubber applications increase progressively as EGU capacity approaches 25 MW. At an capital cost
of more than $1000/kW, it is typically more economical to retire a unit than to operate it with a scrubber. It is
possible to duct emissions from more than one EGU to a single scrubber, but that approach is not included in this
analysis.
4-4
-------
information on ECU S02 measures, particularly for EGUs with 100 MW or larger capacity, can be
found in the documentation for the IPM version used for this RIA.6
4.1.2 Controls Applied for the NonEGU Point and Area Sectors
NonEGU point and Area control measures were identified using AirControlNET 4.2 as
well as the Control Strategy Tool7 (CoST). To reduce nonEGU point S02 emissions, least cost
control measures were identified for emission sources within 50 km of the violating monitor
(see Chapter 3 for rationale). Area source emissions data are generated at the county level,
and therefore controls for this emission sector were applied to the county containing the
violating monitor.
The S02 emission control measures used in this analysis are similar to those used in the
PM2.5 RIA prepared about three years ago. FGD scrubbers can achieve 95% control of S02 for
non-EGU point sources and for utility boilers. Spray dryer absorbers (SDA) are another
commonly employed technology, and SDA can achieve up to 90% control of S02. For specific
source categories, other types of control technologies are available that are more specific to
the sources controlled. The following table lists these technologies. For more information on
these technologies, please refer to the AirControlNET 4.2 control measures documentation
report.8
6 Documentation on the version of IPM used for this RIA can be found at
http://www.epa.gov/airmarkt/progsregs/epa-ipm/index.html.
7 See http://www.epa.gov/ttn/ecas/cost.htm for a description of CoST.
8 For a complete description of AirControlNET control technologies see AirControlNET 4.2 control measures
documentation report, prepared by E.H. Pechan and Associates. May 2008. More information on AirControlNET
(in this case, version 4.1) and the control technologies included in the tool are available at
http://www.epa.gov/ttn/ecas/AirControlNET.htm.
4-5
-------
Table 4-1: Example SO2 Control Measures for Non-EGU Point Sources Applied in Identified
Control Measures Control Strategy Analyses3
Control Measure
Wet and Dry FGD
scrubbers and SDA
Sectors to which These Control
Measures Can Be Applied
ICI boilers— all fuel types, kraft
pulp mills, Mineral Products (e.g.,
Portland cement plants (all fuel
types), primary metal plants,
petroleum refineries
Control
Efficiency
(percent)
95-FGD
scrubbers,
90- for SDA
Average Annualized
Cost/ton (2006$)
$800-$8,000-FGD
$900-7,000-5 DA
Increase percentage sulfur
conversion to meet
sulfuric acid NSPS (99.7%
reduction)
Sulfur recovery plants
75 to 95
$4,000
Sulfur recovery and/or tail
gas treatment
Sulfuric Acid Plants
95-98
$1,000-4,000
Cesium promoted catalyst
Sulfuric Acid Plants with Double-
Absorption process
50%
$1,000
Sources: AirControlNET4.2 control measures documentation report, May 2008, NESCAUM Report on
Applicability of NOx, SO2, and PM Control Measures to Industrial Boilers, November 2008 available at
http://www.nescaum.org/documents/ici-boilers-20081118-final.pdf, and Comprehensive Industry Document on
Sulphuric Acid Plant, Govt. of India Central Pollution Control Board, May 2007. The estimates for these control
measures reflect applications of control where there is no SO2 control measure currently operating except for
the Cesium promoted catalyst.
In applying these S02 controls, we employ a decision rule in which we do not apply
controls to any non-EGU source with 50 tons/year of emissions or less. This decision rule is the
same one we employed for such sources in the PM2.5 RIA completed four years ago.9 The
reason for applying this decision rule is based on a finding that most point sources with
emissions of this level or less had S02 controls already on them. This decision rule aids in gap
filling for a lack of information regarding existing controls on nonEGU sources. In addition, we
also apply the decision rule that we do not apply S02 nonEGU point source controls that yield
emission reductions of 50 tons/year or less. We apply this decision rule in order to reduce the
number the sources affected our non-EGU control strategies to those sources whose reductions
are relatively more cost-effective.
The analysis for non-EGUs mostly applied controls to the following source categories:
industrial boilers, commercial and institutional boilers, sulfuric acid plants (both standalone and
at other facilities such as copper and lead smelters), primary metal plants (iron and steel mills,
PM2.5 RIA, Chapter 3, p. 3-10. This RIA was completed in October, 2006 and is available at
http://www.epa.gov/ttn/ecas/ria.html.
4-6
-------
lead smelters), mineral products (primarily cement kilns) and petroleum refineries. These
source categories are the most prevalent S02 emitters in the areas included in this analysis.
4.1.3 Data Quality for this Analysis
The estimates of emission reductions associated with our control strategies above are
subject to important limitations and uncertainties. EPA's analysis is based on its best judgment
for various input assumptions that are uncertain. As a general matter, the Agency selects the
best available information from available engineering studies of air pollution controls and has
set up what it believes is the most reasonable framework for analyzing the cost, emission
changes, and other impacts of regulatory control.
4-7
-------
4.2 SO2 Emission Reductions Achieved with Identified Controls Analysis
We identified illustrative control strategies that might be employed to reduce emissions
to bring air quality into compliance with the alternative standard being analyzed. As part of this
exercise, we considered the cost-effectiveness of various control options and selected the
lowest cost controls, based on available cost information. Applying identified control measures,
we were able to illustrate attainment for most, but not all of the areas.10
Table 4.2 presents the emission reductions achieved through applying identical control
measures, both by sector and in total. As this table reveals, a majority of the emission
reductions were achieved through ECU emission controls. As indicated in this table, the
estimate emission reductions from the identified controls applied in this analysis under the 75
ppb alternative standard in 2020 are 372,000 tons. About 260,000 tons of the reductions are
from EGUs, and 112,000 are from non-EGU point sources. For the other alternative standards,
the total emission reductions in 2020 are estimated to range from 186,000 tons for the 100 ppb
standard to 803,000 tons for the 50 ppb standard. For all of these standards, this analysis
shows that roughly 60 to 70 percent of these reductions are from EGUs. Most of the remaining
reductions obtained come from non-EGU point sources. Reductions from area sources are
generally a very small portion of those estimated except for the 50 ppb alternative standard,
where 1.8 percent of reductions come from this sector.
Table 4.2: Emission Reductions from Identified Controls in 2020 in Total and by Sector (Tons)a
for Each Alternative Standard
Total Emission Reductions
from Identified Controls:
EGUs
Non-EGUs
Area Sources
50 ppb
800,000
540,000
250,000
15,000
75 ppb
370,000
260,000
110,000
200
100 ppb
190,000
110,000
79,000
100
All estimates rounded to two significant figures. As such, totals may not sum down columns.
bThese values represent emission reductions for the identified control strategy analysis. There were locations not
able to attain the alternative standard being analyzed with identified controls only.
Table 4.3 presents the emission reductions by individual non-EGU point source category
in 2020. As this table shows, the majority of reductions are from industrial boilers for all
alternative standards except for 100 ppb. The percentage of non-EGU point source reductions
from industrial boilers ranges from 50 (50 ppb) to 33 (100 ppb). Reductions from primary metal
10 As will be discussed below, the application of identified controls was insufficient to bring all monitor areas into
compliance with the alternative standards.
4-8
-------
units provide most of the reductions at 100 ppb (59 percent) and this source category has the
next highest percent of reductions for the other alternative standards (21 percent at 50 ppb, 43
percent at 75 ppb).
Table 4.3: Emission Reductions from Identified Controls By Non-EGU Point Source Category in
2020 in Total (Tons)3 for Each Alternative Standard
Total Non-EGU Emission
Reductions from Identified
Controls:b
Industrial Boilers
Sulfuric Acid Plants
Commercial/Institutional
Boilers
Primary Metal Products
Petroleum Refineries
Mineral Products
50 ppb
246,000
124,000
3,000
20,000
52,000
23,000
22,000
75 ppb
112,000
49,000
2,000
4,000
48,000
6,000
5,000
100 ppb
79,000
26,000
1,000
4,000
47,000
1,000
600
All estimates rounded to two significant figures. As such, totals may not sum down columns.
bThese values represent emission reductions for the identified control strategy analysis. There were locations not
able to attain the alternative standard being analyzed with identified controls only.
Table 4.4 presents the S02 emissions reductions realized in each geographic area under
the control strategies applied for the final standard of 75 ppb and also for the other two
alternative standards.
Table 4.4: Emission Reductions by County in 2020 for Each Alternative Standard Analyzed a
State
Arizona
Colorado
Connecticut
Florida
Florida
Georgia
Idaho
Illinois
Illinois
Illinois
Illinois
Illinois
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
County
Gila Co
Denver Co
New Haven Co
Duval Co
Hillsborough Co
Chatham Co
Bannock Co
Cook Co
Madison Co
StClairCo
Sangamon Co
Tazewell Co
Floyd Co
Fountain Co
Jasper Co
Lake Co
Morgan Co
Porter Co
50 ppb
9,000
10,000
8,000
5,100
1,300
19,000
590
39,000
29,000
82,000
22,000
17,000
15,000
9,000
21,000
65,000
3,300
50,000
75 ppb
9,000
-
-
-
-
5,400
-
-
14,000
-
11,000
6,700
-
-
-
20,000
-
-
100 ppb
9,000
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
4-9
-------
Indiana
Iowa
Iowa
Kentucky
Kentucky
Louisiana
Missouri
Missouri
Missouri
Montana
Nebraska
New Hampshire
New York
New York
New York
North Carolina
Ohio
Ohio
Ohio
Ohio
Oklahoma
Oklahoma
Oklahoma
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
South Carolina
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Texas
Texas
West Virginia
Wisconsin
Wisconsin
Wayne Co
Linn Co
Muscatine Co
Jefferson Co
Livingston Co
East Baton Rouge Par
Greene Co
Jackson Co
Jefferson Co
Yellowstone Co
Douglas Co
MerrimackCo
Erie Co
Monroe Co
Suffolk Co
New Hanover Co
Clark Co
Jefferson Co
Lake Co
Summit Co
Kay Co
Muskogee Co
Tulsa Co
Allegheny Co
Blair Co
Northampton Co
Warren Co
Lexington Co
Blount Co
Bradley Co
Montgomery Co
Shelby Co
Sullivan Co
Harris Co
Jefferson Co
Hancock Co
Brown Co
Oneida Co
10,000
9,200
27,000
16,000
4,900
12,000
3,000
25,000
130,000
6,100
24,000
2,700
8,200
12,000
11,000
6,200
6,000
12,000
34,000
22,000
18,000
52,000
15,000
8,800
4,300
21,000
6,100
7,800
4,000
11,000
1,000
4,900
24,000
28,000
12,000
25,000
11,000
7,000
10,000
4,700
21,000
-
-
-
-
13,000
130,000
-
24,000
-
3,200
-
4,400
-
-
-
15,000
15,000
-
35,000
-
-
-
12,000
6,100
-
-
1,200
1,000
-
8,400
-
7,000
-
-
7,000
9,800
-
11,000
-
-
-
-
-
120,000
-
-
-
-
-
-
-
-
-
-
3,100
-
17,000
-
-
-
-
6,100
-
-
-
1,000
-
-
-
-
-
-
7,000
All estimates rounded to two significant figures.
4-10
-------
4.3 Impacts Using Identified Controls
As discussed in Chapter 3, we estimated the overall change in ambient air quality
achieved as a result of each of the control strategies identified above using an impact ratio of
emission reductions to air quality improvement. Table 4.5 presents a detailed breakdown of
the estimated ambient S02 concentrations in 2020 at each of the counties that do not reach
attainment under one or more of the alternative standards.
According to the data presented in Table 4.5, 20 of the 24 monitor areas are expected to
reach attainment with a standard of 75 ppb following implementation of the identified control
strategy. For four areas, identified controls are not sufficient to reach attainment with the
standard of 75 ppb. For the areas projected to violate the NAAQS with the application of
identified controls, we assume that emission reductions beyond identified controls will be
applied, as discussed further below.
Table 4.5: 2020 SO2 Design Values after Application of Identified Controls for Alternative
Standards
State
Arizona
Colorado
Connecticut
Florida
Florida
Georgia
Idaho
Illinois
Illinois
Illinois
Illinois
Illinois
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Iowa
Iowa
Kentucky
Kentucky
Louisiana
Missouri
Missouri
County
Gila Co
Denver Co
New Haven Co
Duval Co
Hillsborough Co
Chatham Co
Bannock Co
Cook Co
Madison Co
StClairCo
Sangamon Co
Tazewell Co
Floyd Co
Fountain Co
Jasper Co
Lake Co
Morgan Co
Porter Co
Wayne Co
Linn Co
Muscatine Co
Jefferson Co
Livingston Co
East Baton Rouge Par
Greene Co
Jackson Co
50 ppb
188.9
50.3
46.9
50.4
52.5
34.4
41.2
39.6
57.0
20.1
35.9
47.9
53.2
46.3
33.6
49.1
47.8
37.4
98.1
50.8
50.0
54.6
50.2
48.6
44.5
47.3
75 ppb
188.9
72.1
74.0
67.5
73.5
71.5
98.1
71.7
68.3
71.9
100 ppb
188.9
100.2
96.9
4-11
-------
Missouri
Montana
Nebraska
New Hampshire
New York
New York
New York
North Carolina
Ohio
Ohio
Ohio
Ohio
Oklahoma
Oklahoma
Oklahoma
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
South Carolina
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Texas
Texas
West Virginia
Wisconsin
Wisconsin
Jefferson Co
Yellowstone Co
Douglas Co
MerrimackCo
Erie Co
Monroe Co
Suffolk Co
New Hanover Co
Clark Co
Jefferson Co
Lake Co
Summit Co
Kay Co
Muskogee Co
Tulsa Co
Allegheny Co
Blair Co
Northampton Co
Warren Co
Lexington Co
Blount Co
Bradley Co
Montgomery Co
Shelby Co
Sullivan Co
Harris Co
Jefferson Co
Hancock Co
Brown Co
Oneida Co
66.4
45.8
47.2
42.6
51.5
46.5
66.4
44.7
50.7
46.0
37.3
59.2
41.2
42.2
28.3
57.0
50.1
49.8
118.8
39.2
52.9
33.2
139.5
46.0
45.2
42.4
49.6
42.7
47.2
47.1
73.8
47.2
66.4
72.0
70.4
74.6
63.2
70.4
118.8
75.2
139.5
73.3
69.3
47.1
78.7
97.6
84.2
118.8
139.5
47.1
Table 4.6 Number of Areas Projected to be in Nonattainment for Each Alternative Standard
After Application of Identified Controls in 2020a
50 ppb
75ppb
100 ppb
Number of Areas Needing Emission
Reductions Beyond Identified Controls
16
There are 56 areas included in this analysis.
4.4 Emission Reductions Needed Beyond Identified Controls
As shown through the identified control strategy analysis, there were not enough
identified controls for every area in the analysis to achieve attainment with neither the 75 ppb
final standard nor the other alternative standards in 2020. Therefore additional emission
reductions will be needed for these areas to attain these alternative standards. Table 4.7
shows the emission reductions needed beyond identified controls for counties to attain the
alternative standards being analyzed. The total emission reductions for full attainment of each
4-12
-------
alternative standard are also included in this table. Table 4.8 presents the emission reductions
needed for each area beyond identified controls for each alternative standard. Chapter 6
presents the discussion of extrapolated costs associated with the emission reductions needed
beyond identified controls.
Table 4.7: Total Emission Reductions and those from Extrapolated Controls in 2020 in Total
and by Sector (Tons)3 for Each Alternative Standard
Total Emission Reductions
from Identified and
Unidentified Controls
Total Emission Reductions
from Unidentified Controls
Unidentified Reductions
from EGUs
Unidentified Reductions
from non-EGUs
Unidentified Reductions
from Area Sources
3 All estimates rounded to two s
50 ppb
920,000
110,000
33,000
54,000
19,000
ignificant figures.
75 ppb
350,000
33,000
5,000
22,000
6,400
100 ppb
170,000
18,000
-
15,000
3,000
Table 4.8: Emission Reductions Needed Beyond Identified Controls in 2020
State
Arizona
Colorado
Connecticut
Florida
Florida
Georgia
Idaho
Illinois
Illinois
Illinois
Illinois
Illinois
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Iowa
Iowa
Kentucky
Kentucky
Louisiana
Cou nty
Gila Co
Denver Co
New Haven Co
Duval Co
Hillsborough Co
Chatham Co
Bannock Co
Cook Co
Madison Co
StClairCo
Sangamon Co
Tazewell Co
Floyd Co
Fountain Co
Jasper Co
Lake Co
Morgan Co
Porter Co
Wayne Co
Linn Co
Muscatine Co
Jefferson Co
Livingston Co
East Baton Rouge Par
50 ppb 75 ppb
13,000 11,000
-
-
-
2,800
-
-
-
5,800
-
-
-
3,200
-
-
-
-
-
14,000 6,500
84
-
3,500
-
-
100 ppb
8,300
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
4-13
-------
Missouri
Missouri
Missouri
Montana
Nebraska
New Hampshire
New York
New York
New York
North Carolina
Ohio
Ohio
Ohio
Ohio
Oklahoma
Oklahoma
Oklahoma
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
South Carolina
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Texas
Texas
West Virginia
Wisconsin
Wisconsin
Greene Co
Jackson Co
Jefferson Co
Yellowstone Co
Douglas Co
MerrimackCo
Erie Co
Monroe Co
Suffolk Co
New Hanover Co
Clark Co
Jefferson Co
Lake Co
Summit Co
Kay Co
Muskogee Co
Tulsa Co
Allegheny Co
Blair Co
Northampton Co
Warren Co
Lexington Co
Blount Co
Bradley Co
Montgomery Co
Shelby Co
Sullivan Co
Harris Co
Jefferson Co
Hancock Co
Brown Co
Oneida Co
-
-
9,500
-
-
-
360
-
19,000
-
130
-
-
4,400
-
-
-
20,000
-
-
4,300 2,700
-
1,400
-
19,000 13,000
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
1,100
-
-
-
8,200
-
-
-
-
-
-
-
All estimates rounded to two significant figures.
4.5 Key Limitations
The estimates of emission reductions associated with the control strategies described
above are subject to important limitations and uncertainties. We summarize these limitations
as follows:
• Actual State Implementation Plans May Differ from our Simulation: In order to reach
attainment with the final NAAQS, each state will develop its own implementation
plan implementing a combination of emissions controls that may differ from those
simulated in this analysis. This analysis therefore represents an approximation of
the emissions reductions that would be required to reach attainment and should not
be treated as a precise estimate.
4-14
-------
• Use of Existing CMAQ Model Runs: This analysis represents a screening level
analysis. We did not conduct new regional scale modeling specifically targeting S02.
More explanation on the screening level analysis done for this RIA can be found in
Chapter 3.
• Analysis Year of 2020: Data limitations necessitated the choice of an analysis year of
2020, as opposed to the presumptive implementation year of 2017. Emission
inventory projections are available for 5-year increments; i.e. we have inventories
for 2015 and 2020, but not 2017. In addition, the CMAQ model runs upon which we
relied were also based on an analysis year of 2020.
• Unidentified controls: We have limited information on available controls for some of
the monitor areas included in this analysis. For a number of small non-EGU and
area sources, there is little or no information available on S02 controls.
4-15
-------
Chapter 5: Benefits Analysis Approach and Results
Synopsis
EPA estimated the monetized human health benefits of reducing cases of morbidity
among populations exposed to S02 and cases of morbidity and premature mortality among
populations exposed to PM2.5 in 2020 for the selected standard and alternative standard levels
in 2006$. Because S02 is also a precursor to PM2.5, reducing S02 emissions in the projected
non-attainment areas will also reduce PM2.5 formation, human exposure and the incidence of
PM2.5-related health effects. For the selected S02 standard at 75 ppb (99th percentile, daily 1-
hour maximum), the total monetized benefits would be $15 to $37 billion at a 3% discount rate
and $14 to $33 billion at a 7% discount rate. For an S02 standard at 50 ppb, the total
monetized benefits would be $34 to $83 billion at a 3% discount rate and $31 to $75 billion at a
7% discount rate. For an S02 standard at 100 ppb, the total monetized benefits would be $7.4
to $18 billion at a 3% discount rate and $6.7 to $16 billion at a 7% discount rate.
These estimates reflect EPA's most current interpretation of the scientific literature and
are consistent with the methodology used for the proposal RIA. These benefits are incremental
to an air quality baseline that reflects attainment with the 2008 ozone and 2006 PM2.5 National
Ambient Air Quality Standards (NAAQS). More than 99% of the total dollar benefits are
attributable to reductions in PM2.5 exposure resulting from S02 emission controls. Higher or
lower estimates of benefits are possible using other assumptions; examples of this are provided
in Figure 5.1 for the selected standard of 75 ppb. Methodological limitations prevented EPA
from quantifying the impacts to, or monetizing the benefits from several important benefit
categories, including ecosystem effects from sulfur deposition, improvements in visibility, and
materials damage. Other direct benefits from reduced S02 exposure have not been quantified,
including reductions in premature mortality.
5-1
-------
Figure 5.1: Total Monetized Benefits (SO2and PM2.5) of Attaining 75 ppb in 2020*
$50
3%DR
7%DR
Laden et al
$40
X $30
5 $20
Pope eta I
$10
$0
PM2.s mortality benefits estimates derived from 2 epidemiology functions and 12 expert functions
This graph shows the estimated total monetized benefits in 2020 for the selected standard of 75 ppb using the no-
threshold model at discount rates of 3% and 7% using effect coefficients derived from the Pope et al. study and the
Laden et al. study, as well as 12 effect coefficients derived from EPA's expert elicitation on PM mortality. The results
shown are not the direct results from the studies or expert elicitation; rather, the estimates are based in part on the
concentration-response function provided in those studies. Graphs for alternative standards would show a similar
pattern.
5.1 Introduction
This chapter documents our analysis of health benefits expected to result from
achieving alternative levels of the S02 NAAQS in 2020, relative to baseline ambient
concentrations that represent attainment with previously promulgated regulations, including
the 2008 ozone and 2006 PM2.5 NAAQS. We first describe our approach for estimating and
monetizing the health benefits associated with reductions of S02. Next, we provide a summary
of our results, including an analysis of the sensitivity of several assumptions in our model. We
then estimate the PM2.5 co-benefits from controlling S02 emissions. Finally, we discuss the key
results of the benefits analysis and indicate limitations and areas of uncertainty in our
approach.
5.2 Primary Benefits Approach
This section presents our approach for estimating avoided adverse health effects due to
S02 exposure in humans resulting from achieving alternative levels of the S02 NAAQS, relative
5-2
-------
to a baseline concentration of ambient S02. First, we summarize the scientific evidence
concerning potential health effects of S02 exposure, and then we present the health endpoints
we selected for our primary benefits estimate. Next, we describe our benefits model, including
the key input data and assumptions. Finally, we describe our approach for assigning an
economic value to the S02 health benefits. The approach for estimating the benefits associated
with exposure to PM is described in section 5.7.
We estimated the economic benefits from annual avoided health effects expected to
result from achieving alternative levels of the S02NAAQS (the "control scenarios") in the year
2020. We estimated benefits in the control scenarios relative to the incidence of health effects
consistent with the ambient S02 concentration expected in 2020 (the "baseline"). Note that
this "baseline" reflects emissions reductions and ambient air quality improvements that we
anticipate will result from implementation of other air quality rules, including compliance with
previously promulgated regulations, including the 2008 ozone and 2006 PM2.5 NAAQS.1
We compare benefits across three alternative S02 NAAQS levels: 50 ppb, 75 ppb, and
100 ppb (99th percentile). Consistent with EPA's approach for RIA benefits assessments, we
estimate the health effects associated with an incremental difference in ambient
concentrations between a baseline scenario and a pollution control strategy. As indicated in
Chapter 4, several areas of the country may not be able to attain the alternative standard levels
using known pollution control methods. For this reason, we provide an estimate of the benefits
associated with partially attaining the standard using known controls as well as the full
attainment results in Table 5.13 of this chapter. Because some areas require emission
reductions from unknown sources to attain the various standards, the results are sensitive to
assuming full attainment. All of the other results tables in this chapter assume full attainment
with the various standard levels. The full attainment results include extrapolated tons from
unknown controls, which were spread across the sectors in proportion to the emissions in the
county.2
5.3 Overview of analytical framework for benefits analysis
5.3.1 Benefits Model
For the S02 benefits analysis, we use the Environmental Benefits Mapping and Analysis
Program (BenMAP, version 3) (Abt Associates, 2008) to estimate the health benefits occurring
as a result of implementing alternative S02 NAAQS levels. Although EPA has used BenMAP
1 See Chapter 2 of this RIA for more information on the rules incorporated into the baseline.
2 See Chapter 4 of this RIA for more information on the extrapolated tons estimated to reach full attainment.
5-3
-------
extensively to estimate the health benefits of reducing exposure to PM2.5 and ozone in previous
RIAs, the proposal RIA was the first RIA in which EPA used BenMAP to estimate the health
benefits directly attributable to reducing exposure to S02to support a change in the NAAQS.
Figure 5.2 below shows the major components of, and data inputs to, the BenMAP model.
Figure 5.2: Diagram of Inputs to BenMAP model for SO2 Analysis
Census
Population Data
2020
Population
Projections
Modeled
Baseline 2020
Ambient SCh
Concentrations
SO2 Health
Functions
Woods &
Poole
Population
Projections
Monitor Rollback
of Design Values to
Full Attainment
SCh Incremental
Air Quality Change
SO2-Related Health
Impacts
Background
Incidence and
Prevalence Rates
Valuation
Functions
Monetized Benefits
5.3.2 Air Quality Estimates
As Figure 5.2 shows, the primary input to any benefits assessment is the estimated
changes in ambient air quality expected to result from a simulated control strategy or
attainment of a particular standard. EPA typically relies upon air quality modeling to generate
these data, but time and technical limitations described in Chapter 3 prevented us from
generating new air quality modeling to simulate the changes in ambient S02 resulting from
each control strategy. Instead, we utilize the ambient S02concentrations modeled by CMAQas
part of the upcoming PM NAAQS RIA as our baseline.3
See Chapter 3 for more detail regarding the air quality data used in this analysis.
5-4
-------
The CMAQ air quality model provides projects both design values at S02 monitors and
air quality concentrations at 12 km by 12 km grid cells nationwide. To estimate the benefits of
fully attaining the standards in all areas, EPA employed the "monitor rollback" approach to
approximate the air quality change resulting from just attaining alternative S02 NAAQS at each
design value monitor. Figure 5.3 depicts the rollback process, which differs from the technique
described in Chapter 3. The emission control strategy estimated the level of emission
reductions necessary to attain each alternate NAAQS standard, whereas the approach
described here aims to estimate the change in population exposure associated with attaining an
alternate NAAQS. This approach relies on data from the existing S02 monitoring network and
the inverse distance squared variant of the Veronoi Neighborhood Averaging (VNA)
interpolation method to adjust the CMAQ-modeled S02 concentrations such that each area just
attains the standard alternatives. We believe that the interpolation method using inverse
distance squared most appropriately reflects the exposure gradient for S02 around each
monitor (EPA, 2008c). A sensitivity analysis in Table 5.6 shows that the results are not
particularly sensitive to the interpolation method.
Step 1. Receive 12 km
CMAQ baseline air quality
modeling
Figure 5.3: Diagram of Rollback Method
Use modeled air quality data
to establish ratios between
99th percentile SO2 design
value and SO2 air quality
metric at each monitor.*
Step 2. Rollback SO2 monitor design^
values to just attain each standard
alternative
Alternative 1: 50 ppb
Alternative 2: 75 ppb
Alternative 3: 100 ppb
Step 3. Interpolate
incremental reduction
in design value change
to 12 km grid using
VNA in BenMAP and
calculate benefits
Convert interpolated
DV change to
equivalent change in
SO-, metric and adjust
Calculate
benefits for
each
standard
*Metrics used in the epidemiology studies include the 24-hr mean, 3-hr mean, 8-hr max, and 1-hr max.
5-5
-------
Because the VNA rollback approach interpolates monitor values, it is most reliable in
areas with a denser monitoring network. In areas with a sparser monitoring network, there is
less observed monitoring data to support the VNA interpolation and we have less confidence in
the predicted air quality values further away from the monitors. For this reason, we
interpolated air quality values—and estimated health impacts—within the CMAQ grid cells that
are located within 50 km of the monitor, assuming that emission changes within this radius
would affect the S02 concentration at each monitor. Limiting the interpolation to this radius
attempts to account for the limitations of the VNA approach, the air quality data limitations
identified in Chapter 3 and ensures that the benefits and costs analyses consider a consistent
geographic area.4 Therefore, the primary benefits analysis assesses health impacts occurring to
populations living in the CMAQ grid cells located within the 50 km buffer for the specific
geographic areas assumed to not attain the alternate standard levels. We test the sensitivity of
this assumption relative to other exposure buffers in Table 5.6.
5.4 Estimating Avoided Health Effects from SO2 Exposure
Following an extensive evaluation of health evidence from epidemiologic and laboratory
studies, the U.S. EPA has concluded that there is a causal relationship between respiratory
health effects and short-term exposure to S02(U.S. EPA, 2008c). The immediate effect of S02
on the respiratory system in humans is bronchoconstriction. This response is mediated by
chemosensitive receptors in the tracheobronchial tree, which trigger reflexes at the central
nervous system level resulting in bronchoconstriction, mucus secretion, mucosal vasodilation,
cough, and apnea followed by rapid shallow breathing. In some cases, local nervous system
reflexes also may be involved. Asthmatics are more sensitive to the effects of S02 likely
resulting from preexisting inflammation associated with this disease. This inflammation may
lead to enhanced release of mediators, alterations in the autonomic nervous system and/or
sensitization of the chemosensitive receptors. These biological processes are likely to underlie
the bronchoconstriction and decreased lung function observed in response to S02 exposure. A
clear concentration-response relationship has been demonstrated in laboratory studies
following exposures to S02 at concentrations between 20 and 100 ppb, both in terms of
increasing severity of effect and percentage of asthmatics adversely affected.
5.4.1 Selection of Health EndpointsforS02
Epidemiological researchers have associated S02exposure with adverse health effects in
numerous toxicological, clinical and epidemiological studies, as described in the Integrated
Please see Chapter 3 for more information regarding the technical basis for the 50 km assumption.
5-6
-------
Science Assessment for Oxides of Sulfur - Health Criteria (Final Report) (U.S. EPA, 2008c);
hereafter, "S02 ISA"). The S02 ISA provides a comprehensive review of the current evidence of
health and environmental effects of S02.
Previous reviews of the S02 primary NAAQS, most recently in 1996, did not include a
quantitative benefits assessment for S02 exposure. As the first health benefits assessment for
S02 exposure, we build on the methodology and lessons learned from the S02 risk and exposure
assessment (U.S. EPA, 2009c) and the benefits assessments for the recent PM2.5; 03; and N02
NAAQS (U.S. EPA, 2006a; U.S. EPA, 2008a; U.S. EPA, 2010a; U.S. EPA, 2010b).
We quantified S02-related health endpoints for which the S02 ISA provides the strongest
evidence of an effect. In general, we follow a weight of evidence approach, based on the
biological plausibility of effects, availability of concentration-response functions from well
conducted peer-reviewed epidemiological studies, cohesiveness of results across studies, and a
focus on endpoints reflecting public health impacts (like hospital admissions) rather than
physiological responses (such as changes in clinical measures like Forced Expiratory Volume
(FEV1)). The differing evidence and associated strength of the evidence for these different
effects is described in detail in the S02 ISA.
Although a number of adverse health effects have been found to be associated with S02
exposure, this benefits analysis only includes a subset due to limitations in understanding and
quantifying the dose-response relationship for some of these health endpoints. In this analysis,
we only estimated the benefits for those endpoints with sufficient evidence to support a
quantified concentration-response relationship using the information presented in the S02 ISA,
which contains an extensive literature review for several health endpoints related to S02
exposure. Because the ISA only included studies published or accepted for publication through
April 2008, we also performed supplemental literature searches in the online search engine
PubMed® to identify relevant studies published between January 2008, and the present.5
Based on our review of this information, we quantified four short-term respiratory morbidity
endpoints that the S02 ISA identified as a "causal relationship": acute respiratory symptoms,
asthma exacerbation, respiratory-related emergency department visits, and respiratory-related
hospitalizations.
Table 5.1 presents the health effects related to S02 exposure quantified in this benefits
analysis. In addition, the table includes other endpoints potentially linked to S02 exposure, but
which we are not yet ready to quantify with dose-response functions. Fora list of the health
5 The O'Connor et al. study (2008) is the only study included in this analysis that was published after the cut-off
date for inclusion in the SO2 ISA.
5-7
-------
effects related to PM2.5 exposure that we quantify in this analysis, please see Table 5.6 in
section 5.7.
The S02 ISA concluded that the relationship between short-term S02 exposure and
premature mortality was "suggestive of a causal relationship" because it is difficult to attribute
the mortality risk effects to S02 alone. Therefore, we decided not to quantify premature
mortality from S02 exposure in this analysis despite evidence suggesting a positive association
(U.S. EPA, 2008c). Although the S02 ISA stated that studies are generally consistent in reporting
a relationship between S02 exposure and mortality, there was a lack of robustness of the
observed associations to adjustment for co-pollutants. As the literature continues to evolve,
we may revisit this decision in future benefits assessment for S02.
As noted in Table 5.1, we are not able to quantify several welfare benefit categories in
this analysis because we are limited by the available data or resources. Although we cannot
quantify the ecosystem benefits of reducing sulfur deposition or visibility improvements in this
analysis, we provide a qualitative analysis in section 5.9.
Table 5.1: Human Health and Welfare Effects of SO2
Pollutant/ Quantified and Monetized in Primary Unquantified Effects 'c
Effect Estimates3 Changes in:
SO2/Health Respiratory Hospital Admissions Premature mortality
Asthma ER visits Pulmonary function
Asthma exacerbation Other respiratory emergency department visits
Acute Respiratory symptoms Other respiratory hospital admissions
SO2/Welfare Visibility improvements
Commercial fishing and forestry from acidic deposition
Recreation in terrestrial and aquatic ecosystems from
acid deposition
Increased mercury methylation
Primary quantified and monetized effects are those included when determining the primary estimate of total
monetized benefits of the alternative standards.
bThe categorization of unquantified toxic health and welfare effects is not exhaustive.
c Health endpoints in the unquantified benefits column include both a) those for which there is not consensus on
causality and those for which causality has been determined but empirical data are not available to allow
calculation of benefits.
5.4.2 Selection of Concentration-Response Functions
After identifying the health endpoints to quantify in this analysis, we then selected
concentration-response functions drawn from the epidemiological literature identified in the
S02 ISA. We considered several factors, in the order below, in selecting the appropriate
epidemiological studies and concentration-response functions for this benefits assessment.
5-8
-------
1. We considered ambient S02 studies that were identified as key studies in the S02
ISA (or a more recent study), excluding those affected by the general additive
model (GAM) S-Plus issue.6
2. We judged that studies conducted in the United States are preferable to those
conducted outside the United States, given the potential for effect estimates to
be affected by factors such as the ambient pollutant mix, the placement of
monitors, activity patterns of the population, and characteristics of the
healthcare system especially for hospital admissions and emergency department
visits. We include Canadian studies in sensitivity analyses, when available.
3. We only incorporated concentration-response functions for which there was a
corresponding valuation function. Currently, we only have a valuation function
for asthma-related emergency department visits, but we do not have a valuation
function for all-respiratory-related emergency department visits.
4. We preferred concentration-response functions that correspond to the age
ranges most relevant to the specific health endpoint, with non-overlapping ICD-9
codes. We preferred completeness when selecting functions that correspond to
particular age ranges and ICD codes. Age ranges and ICD codes associated with
the selected functions are identified in Table 5.2.
5. We preferred multi-city studies or combined multiple single city studies, when
available.
6. When available, we judged that effect estimates with distributed or cumulative
lag structures were most appropriate for this analysis.
7. When available, we selected S02 concentration-response functions based on
multi-pollutant models. Studies with multi-pollutant models are identified in
Table 5.2.
These criteria reflect our preferences for study selection, and it was possible to satisfy
many of these, but not all. There are trade-offs inherent in selecting among a range of studies,
as not all studies met all criteria outlined above. At minimum, we ensured that none of the
studies were GAM affected, we selected only U.S. based studies, and we quantified health
endpoints for which there was a corresponding valuation function.
We believe that U.S.-based studies are most appropriate studies to use in this analysis
to estimate the number of hospital admissions associated with S02 exposure because of the
6 The S-Plus statistical software is widely used for nonlinear regression analysis in time-series research of health
effects. However, in 2002, a problem was discovered with the software's default conversion criteria in the
general additive model (GAM), which resulted in biased relative risk estimates in many studies. This analysis
does not include any studies that encountered this problem. For more information on this issue, please see U.S.
EPA (2002).
5-9
-------
characteristics of the ambient air, population, and healthcare system. Using only U.S.-based
studies, we are limited to one epidemiology study for hospital admissions (Schwartz, 1996).
However, there are several Canada-based epidemiology studies that also estimate respiratory
hospital admissions (Fung, 2006; Luginaah, 2005; Yang, 2003). Table 5.12 provides the
sensitivity of the S02 benefits using the effect estimates from the Canadian studies. Compared
to the U.S. based study, the Canadian studies produce a substantially larger estimate of hospital
admissions associated with S02exposure.
When selecting concentration-response functions to use in this analysis, we reviewed
the scientific evidence regarding the presence of thresholds in the concentration-response
functions for S02 -related health effects to determine whether the function is approximately
linear across the relevant concentration range. The S02 ISA concluded that, "The overall limited
evidence from epidemiologic studies examining the concentration-response function of S02
health effects is inconclusive regarding the presence of an effect threshold at current ambient
levels." For this reason, we have not incorporated thresholds in the concentration-response
functions for S02 -related health effects in this analysis.
Table 5.2 shows the studies and health endpoints that we selected for this analysis.
Table 5.3 shows the baseline health data used in combination with these health functions.
Following these tables is a description of each of the epidemiology studies used in this analysis.
5-10
-------
Table 5.2: SO2-Related Health Endpoints Quantified, Studies Used to Develop Health Impact
Functions and Sub-Populations to which They Apply
Endpoint
Study
Study
Population
Hospital Admissions
All respiratory
Emergency Department Visits
Schwartz et al., 1996 - ICD-9 460-519
65-99
Asthma
Other Health Endpoints
Pooled Estimate:
Itoetal. (2007)-ICD-9 493
Michaud (2004) - ICD-9 493
NYDOH (2006)b-ICD-9 493
Peeletal. (2005)-ICD-9 493
Wilson (2005) - ICD-9 493
All ages
Asthma exacerbations
Pooled estimate:
Mortimer et al. (2002) (one or more symptoms)3
O'Connor et al. (2008) (slow play, missed school days0,
nighttime asthma)3'b
Schildcrout et al. (2006) (one or more symptoms)3
4-12
Acute Respiratory
Symptoms
Schwartz et al. (1994)
7-14
The original study populations were 4 to 9 for the Mortimer et al. (2002) study and 5 to 12 for the O'Connor et al.
(2008) study and the Schildcrout et al. (2006) study. We extended the applied population to facilitate the pooling
process, recognizing the common biological basis for the effect in children in the broader age group. See: National
Research Council (NRC). 2002. Estimating the Public Health Benefits of Proposed Air Pollution Regulations.
Washington, DC: The National Academies Press, pg 117.
b Study specifies a multipollutant model.
c The form of this one function was not clear from the study. For this analysis, we assumed that it was log-linear,
but we have subsequently determined that it is logistic. This adds a small amount to uncertainty regarding the
asthmas incidence estimates, but this uncertainty is obscured by the rounding of the monetized estimates.
5-11
-------
Table 5.3: National Average Baseline Incidence Rates used to Calculate SO2 -Related Health
Impacts
Endpoint
Respiratory
Hospital
Admissions
Asthma ER
visits
Minor
Restricted
Activity Days
(MRADs)
Asthma
Exacerbations
Source
1999 NHDS
public use data
files b
2000 NHAMCS
public use data
files c; 1999
NHDS public use
data files b
Schwartz (1994,
table 2)
Mortimer et al.
(2002)
O'Connor et al.
(2008)
Schildcrout et
al. (2006)
Notes Rate per
100 people per year by Age Group
<18 18-24 25-34 35-44 45-54 55-64 65+
incidence 0.043
incidence 1.011
incidence 0.416
Incidence (and
prevalence) among
asthmatic children
Incidence (and
prevalence) among
asthmatic children
Incidence (and
prevalence) among
asthmatic children
0.084 0.206 0.678 1,
1.087 0.751 0.438 0,
— — —
Any morning symptom
Missed school
One or more symptoms
Slow play
Nighttime asthma
One or more symptoms
,926 4.389 11.629
,352 0.425 0.232
— — —
0.116 (0.0567) d
0.057 (0.0567) d
0.207 (0.0567) d
0.157 (0.0567) d
0.121 (0.0567) d
0.52 (0.0567) d
The following abbreviations are used to describe the national surveys conducted by the National Center for Health
Statistics: HIS refers to the National Health Interview Survey; NHDS—National Hospital Discharge Survey; NHAMCS—
National Hospital Ambulatory Medical Care Survey.
b See ftp://ftp.cdc.gov/pub/Health Statistics/NCHS/Datasets/NHDS/
c See ftp://ftp.cdc.gov/pub/Health Statistics/NCHS/Datasets/NHAMCS/
d We assume that this prevalence rate for ages 5 to 9 is also applicable down to age 4.
Schwartz et al. (1996)
Schwartz et al. (1996) is a review paper with an example drawn from hospital
admissions of the elderly in Cleveland, Ohio from 1988-1990. The authors argued that the
central issue is control for seasonality. They illustrated the use of categorical variables for
weather and sinusoidal terms for filtering season in the Cleveland example. After controlling
for season, weather, and day of the week effects, hospital admissions of persons aged 65 and
older in Cleveland for respiratory illness was associated with ozone (RR = 1.09, 95% Cl 1.02,
1.16) and PM10 (RR = 1.12, 95% Cl 1.01,1.24), and marginally associated with S02 (RR = 1.03,
95% Cl = 0.99, 1.06). All of the relative risks are for a 100 micrograms/m3 increase in the
pollutant.
5-12
-------
Fung et al. (2006) - Sensitivity Analysis
Fung et al. (2006) assessed the impact of ambient gaseous pollutants (S02, N02, CO, and
03) and particulate matters (PMi0, PM2.5, and PMi0-2.5) as well as the coefficient of haze (COM)
on recurrent respiratory hospital admissions (ICD-9 codes 460-519) among the elderly in
Vancouver, Canada, for the period of June 1,1995, to March 31, 1999, using a new method
proposed by Dewanji and Moolgavkar (2000; 2002). The authors found significant associations
between respiratory hospital admissions and 3-day, 5-day, and 7-day moving averages of the
ambient S02 concentrations, with the strongest association observed at the 7-day lag (RR =
1.044, 95% Cl: 1.018-1.070). The authors also found PMi0-2.5 for 3-day and 5-day lag to be
significant, with the strongest association at 5-day lag (RR = 1.020, 95% Cl: 1.001-1.039). No
significant associations with admission were found with current day exposure.
Luginaah et al. (2005) - Sensitivity analysis
Luginaah et al. (2005) assessed the association between air pollution and daily
respiratory hospitalization (ICD-9 codes 460-519) for different age and sex groups from 1995 to
2000. The pollutants included were N02, S02, CO, 03, PMio, coefficient of haze (COM), and total
reduced sulfur (TRS). The authors estimated relative risks (RR) using both time-series and case-
crossover methods after controlling for appropriate confounders (temperature, humidity, and
change in barometric pressure). The results of both analyses were consistent. They found
associations between N02, S02, CO, COM, or PMi0 and daily hospital admission of respiratory
diseases especially among females. For females 0-14 years of age, there was 1-day delayed
effect of N02 (RR = 1.19, case-crossover method), a current-day S02 (RR = 1.11, time series),
and current-day and 1- and 2-day delayed effects for CO by case crossover (RR = 1.15, 1.19,
1.22, respectively). Time-series analysis showed that 1-day delayed effect of PMi0 on
respiratory admissions of adult males (15-64 years of age), with an RR of 1.18. COM had
significant effects on female respiratory hospitalization, especially for 2-day delayed effects on
adult females, with RRs of 1.15 and 1.29 using time-series and case-crossover analysis,
respectively. There were no significant associations between 03 and TRS with respiratory
admissions.
Yang et al. (2003) - Sensitivity analysis
Yang et al. (2003) examined the impact of ozone, nitrogen dioxide, sulfur dioxide,
carbon monoxide, and coefficient of haze on daily respiratory admissions (ICD-9 codes 460-519)
in both young children (<3 years of age) and the elderly (65-99 years of age) in greater
5-13
-------
Vancouver, British Columbia during the 13-yr period 1986-1998. Bidirectional case-crossover
analysis was used to investigate associations and odds ratios were reported for single-pollutant,
two-pollutant and multiple-pollutant models. Sulfur dioxide was found marginally significant in
all models for elderly.
Ito et al. (2007)
Ito et al. (2007) assessed associations between air pollution and asthma emergency
department visits in New York City for all ages. Specifically they examined the temporal
relationships among air pollution and weather variables in the context of air pollution health
effects models. The authors compiled daily data for PM2.5, 03, N02, S02, CO, temperature, dew
point, relative humidity, wind speed, and barometric pressure for New York City for the years
1999-2002.The authors evaluated the relationship between the various pollutants' risk
estimates and their respective concurvities, and discuss the limitations that the results imply
about the interpretability of multi-pollutant health effects models.
Michaud et al. (2004)
Michaud et al. (2004) examined the association of emergency department (ED) visits in
Hilo, Hawai'i, from January 1997 to May 2001 with volcanic fog, or "vog", measured as sulfur
dioxide (S02) and submicrometer particulate matter (PMi). Log-linear regression models were
used with robust standard errors. The authors studied four diagnostic groups: asthma/COPD;
cardiac; flu, cold, and pneumonia; and gastroenteritis. Before adjustments, highly significant
associations with vog-related air quality were seen for all diagnostic groups except
gastroenteritis. After adjusting for month, year, and day of the week, only asthma/COPD had
consistently positive associations with air quality. They found that the strongest associations
were for S02 with a 3-day lag (6.8% per 10 ppb; P=0.001) and PMi, with a 1-day lag (13.8% per
10 ug/m3; P=0.011).
NYDOH (2006)
New York State Department of Health (NYDOH) investigated whether day-to-day
variations in air pollution were associated with asthma emergency department (ED) visits in
Manhattan and Bronx, NYC and compared the magnitude of the air pollution effect between
the two communities. NYDOH (2006) used Poisson regression to test for effects of 14 key air
contaminants on daily ED visits, with control for temporal cycles, temperature, and day-of-week
effects. The core analysis utilized the average exposure for the 0- to 4-day lags. Mean daily S02
was found significantly associated with asthma ED visits in Bronx but not Manhattan. Their
5-14
-------
findings of more significant air pollution effects in the Bronx are likely to relate in part to
greater statistical power for identifying effects in the Bronx where baseline ED visits were
greater, but they may also reflect greater sensitivity to air pollution effects in the Bronx.
Peel et al. (2005)
Peel et al. (2005) examined the associations between air pollution and respiratory
emergency department visits (i.e., asthma (ICD-9 code 493, 786.09), COPD (491,492,496), URI
(460-466, 477), pneumonia (480-486), and an all respiratory-disease group) in Atlanta, GAfrom
1 January 1993 to 31 August 2000. They used 3-Day Moving Average (Lags of 0, 1, and 2 Days)
and unconstrained distributed lag (Lags of 0 to 13 Days) in the Poisson regression analyses. In
single-pollutant models, positive associations persisted beyond 3 days for several outcomes,
and over a week for asthma. The effects of N02, CO or PMio on asthma ED visits were found
significant but S02 or 03 were not significantly associated with asthma ED visits.
Wilson et al. (2005)
Daily emergency room (ER) visits for all respiratory (ICD-9 codes 460-519) and asthma
(ICD-9 code 493) were compared with daily S02, 03, and weather variables over the period
1998-2000 in Portland, Maine and 1996-2000 in Manchester, New Hampshire. Seasonal
variability was removed from all variables using nonparametric smoothed function (LOESS).
Wilson et al.(2005) used generalized additive models to estimate the effect of elevated levels of
pollutants on ER visits. Relative risks of pollutants were reported over their inter-quartile range
(IQR, the 75th -25th percentile pollutant values). In Portland, an IQR increase in S02 was
associated with a 5% (95% Cl 2-7%) increase in all respiratory ER visits and a 6% (95% Cl 1-12%)
increase in asthma visits. An IQR increase in 03 was associated with a 5% (95% Cl 1-10%)
increase in Portland asthmatic ER visits. No significant associations were found in Manchester,
New Hampshire, possibly due to statistical limitations of analyzing a smaller population. The
absence of statistical evidence for a relationship should not be used as evidence of no
relationship. This analysis reveals that, on a daily basis, elevated S02 and 03 have a significant
impact on public health in Portland, Maine.
Villeneuve et al. (2007) - Sensitivity Analysis
Villeneuve et al. (2007) examined the associations between air pollution and emergency
department (ED) visits for asthma among individuals two years of age and older in the census
metropolitan area of Edmonton, Canada between April 1, 1992 and March 31, 2002 using a
time stratified case-crossover design. Daily air pollution levels for the entire region were
5-15
-------
estimated from three fixed-site monitoring stations. Odds ratios and their corresponding 95%
confidence intervals were estimated using conditional logistic regression with adjustment for
temperature, relative humidity and seasonal epidemic of viral related respiratory disease.
Villeneuve et al.(2007) found positive associations for asthma ED visits with outdoor air
pollution levels between April and September, but such associations were absent during the
remainder of the year. Effects were strongest among young children (2-4 years of age) and
elderly (>75 years of age). Air pollution risk estimates were largely unchanged after adjustment
for aeroallergen levels. This study is not included in the S02 ISA only because it was published
after the cut-off date, but it met all of the other criteria for inclusion in this analysis.
Mortimer et al. (2002)
Mortimer et al. (2002) examined the effect of daily ambient air pollution within a cohort
of 846 asthmatic children residing in eight urban areas of the USA between June 1 to August 31,
1993, using data from the National Cooperative Inner-City Asthma Study. Daily air pollution
concentrations were extracted from the Aerometric Information Retrieval System database
from the Environment Protection Agency in the USA. Logistic models were used to evaluate the
effects of several air pollutants (03, N02, S02 and PMi0) on peak expiratory flow rate (PEFR) and
symptoms in 846 children (ages 4-9 yrs) with a history of asthma. In single pollutant models,
each pollutant was associated with an increased incidence of morning symptoms: (odds ratio
(OR) = 1.16 (95% Cl 1.02-1.30) per IQR increase in 4-day average 03, OR = 1.32 (95% Cl 1.03-
1.70) per IQR increase in 2-day average S02, OR = 1.48 (95% Cl 1.02-2.16) per IQR increase in 6-
day average N02 and OR = 1.26 (95% Cl 1.0-1.59) per IQR increase in 2-day average PMi0. This
longitudinal analysis supports previous time-series findings that at levels below current USA air-
quality standards, summer-air pollution is significantly related to symptoms and decreased
pulmonary function among children with asthma.
O'Connor et al. (2008)
O'Connor et al. (2008) investigated the association between fluctuations in outdoor air
pollution and asthma exacerbation (wheeze-cough, nighttime asthma, slow play and school
absence) among 861 inner-city children (5-12 years of age) with asthma in seven US urban
communities. Asthma symptom data were collected every 2 months during the 2-year study
period. Daily pollution measurements were obtained from the Aerometric Information
Retrieval System between August 1998 and July 2001. The relationship of symptoms to
fluctuations in pollutant concentrations was examined by using logistic models. In single-
pollutant models, significant or nearly significant positive associations were observed between
higher N02 concentrations and each of the health outcomes. The 03, PM2.5, and S02
5-16
-------
concentrations did not appear significantly associated with symptoms or school absence except
for a significant association between PM2.5 and school absence. This study is not included in the
S02 ISA only because it was published after the cut-off date, but it met all of the other criteria
for inclusion in this analysis.
Schildcrout et al. (2006)
Schildcrout et al. (2006) investigated the relation between ambient concentrations of
the five criteria pollutants (PMio, 03, N02, S02, and CO) and asthma exacerbations (daily
symptoms and use of rescue inhalers) among 990 children in eight North American cities during
the 22-month prerandomization phase (November 1993-September 1995) of the Childhood
Asthma Management Program. Short-term effects of CO, N02, PMi0, S02, and warm-season 03
were examined in both one-pollutant and two-pollutant models, using lags of up to 2 days in
logistic and Poisson regressions. Lags in CO and N02 were positively associated with both
measures of asthma exacerbation, and the 3-day moving sum of S02 levels was marginally
related to asthma symptoms. PMi0 and 03were unrelated to exacerbations. The strongest
effects tended to be seen with 2-day lags, where a 1-parts-per-million change in CO and a 20-
parts-per-billion change in N02 were associated with symptom odds ratios of 1.08 (95%
confidence interval (Cl): 1.02, 1.15) and 1.09 (95% Cl: 1.03, 1.15), respectively.
Schwartz et al. (1994)
Schwartz et al. (1994) studied the association between ambient air pollution exposures
and respiratory illness among 1,844 school children (7-14 years of age) in six U.S. cities during
five warm season months between April and August. Daily measurements of ambient sulfur
dioxide (S02), nitrogen dioxide (N02), ozone (03), inhalable particles (PMi0), respirable particles
(PM2.5), light scattering, and sulfate particles were made, along with integrated 24-h measures
of aerosol strong acidity. Significant associations in single pollutant models were found
between S02, N02, or PM2.5 and incidence of cough, and between sulfur dioxide and incidence
of lower respiratory symptoms. Significant associations were also found between incidence of
coughing symptoms and incidence of lower respiratory symptoms and PMio, and a marginally
significant association between upper respiratory symptoms and PMi0.
5-17
-------
Delfino et al. (2003)-SensitivityAnalysis
Delfino et al. (2003) conducted a panel study of 22 Hispanic children with asthma who
were 10-16 years old and living in a Los Angeles community with high traffic density. Subjects
filled out symptom diaries daily for up to 3 months (November 1999 through January 2000).
Pollutants included ambient hourly values of ozone (03), nitrogen dioxide (N02), sulfur dioxide
(S02), and carbon monoxide (CO) and 24-hr values of volatile organic compounds (VOCs),
particulate matter with aerodynamic diameter < 10 micro (PMio), and elemental carbon (EC)
and organic carbon (OC) PMio fractions. Asthma symptom severity was regressed on pollutants
using logistic models. The authors found positive associations of symptoms with criteria air
pollutants (03, N02, S02, and PMio). Selected adjusted odds ratio for more severe asthma
symptoms from interquartile range increases in pollutants was, for 2.5 ppb 8-hr max S02,1.36
[95% confidence interval (Cl), 1.08-1.71]. Their findings support the view that air toxins in the
pollutant mix from traffic and industrial sources may have adverse effects on asthma in
children.
5.4.3 Pooling Multiple Health Studies
After selecting which health endpoints to analyze and which epidemiology studies
provide appropriate effect estimates, we then selected a method to combine the multiple
health studies to provide a single benefits estimate for each health endpoint. The purpose of
pooling multiple studies together is to generate a more robust estimate by combining the
evidence across multiple studies and cities. Because we used a single study for acute
respiratory symptoms and a single study for hospital admission for asthma, there was no
pooling necessary for those endpoints.
See Table 5.2 for more information on how the asthma studies were adjusted. Because
asthma represents the largest benefits category in this analysis, we tested the sensitivity of the
S02 benefits to alternate pooling choices in Table 5.6.
5.5 Valuation of Avoided Health Effects from SO2 Exposure
The selection of valuation functions very similar to the N02 NAAQS RIA (U.S. EPA, 2010b)
and the PM2.5 NAAQS RIA (U.S. EPA, 2006a) with a couple exceptions. First, in this analysis, we
estimated changes in all respiratory hospital admissions. This is consistent with the PM2.5
NAAQS RIA, but inconsistent with the N02 NAAQS RIA, which estimated changes for only a
subset of respiratory hospital admissions (i.e., chronic lung disease and asthma) because
concentration-response functions were only available for the subset. Second, in this analysis,
5-18
-------
we used the any-of-19 symptoms valuation function for acute respiratory symptoms. This is
consistent with the N02 NAAQS RIA, but inconsistent with the PM2.5 NAAQS RIA, which used the
valuation function for "minor-restricted activity day" (MRADs). The valuation for any-of-19-
symptoms is approximately 50% of the valuation for MRADs. Consistent with economic theory,
these valuation functions include adjustments for inflation (2006$) and income growth over
time (2020 income levels). Table 5.4 provides the unit values used to monetize the benefits of
reduced exposure to S02.
Table 5.4: Central Unit Values SO2 Health Endpoints (2006$)*
Health Endpoint
Central Unit Value Per
Statistical Incidence
(2020 income level)
Derivation of Distributions of Estimates
Hospital Admissions and ER Visits
Respiratory Hospital
Admissions
$24,000
No distributional information available. The COI point
estimates (lost earnings plus direct medical costs) are based
on ICD-9 code level information (e.g., average hospital care
costs, average length of hospital stay, and weighted share of
total COPD category illnesses) reported in Agency for
Healthcare Research and Quality, 2000 (www.ahrq.gov).
Asthma Emergency
Room Visits
$370
No distributional information available. Simple average of
two unit COI values:
(1) $400 (2006$), from Smith et al. (1997) and
(2) $340 (2006$), from Stanford et al. (1999).
Respiratory Ailments Not Requiring Hospitalization
Asthma Exacerbation
$53
Asthma exacerbations are valued at $49 (2006$) per
incidence, based on the mean of average WTP estimates for
the four severity definitions of a "bad asthma day," described
in Rowe and Chestnut (1986). This study surveyed asthmatics
to estimate WTP for avoidance of a "bad asthma day," as
defined by the subjects. For purposes of valuation, an
asthma exacerbation is assumed to be equivalent to a day in
which asthma is moderate or worse as reported in the Rowe
and Chestnut (1986) study. The value is assumed have a
uniform distribution between $19 and $83 (2006$).
Acute Respiratory
Symptoms
$30
The valuation estimate for "any of 19 acute respiratory
symptoms" is derived from Krupnick et al. (1990) assuming
that this health endpoint consists either of upper respiratory
symptoms (URS) or lower respiratory symptoms (LRS), or
both. We assumed the following probabilities for a day of
"any of 19 acute respiratory symptoms": URS with 40 percent
probability, LRS with 40 percent probability, and both with 20
percent probability. The point estimate of WTP to avoid a
day of "the presence of any of 19 acute respiratory
symptoms" is $28 (2006$). The value is assumed have a
uniform distribution between $0 and $56 (2006$).
*AII estimates rounded to two significant figures. All values have been inflated to reflect values in 2006 dollars and
income levels in 2020.
5-19
-------
5.6 Health Benefits of Reducing Exposure to SO2 Results
EPA estimated the monetized human health benefits of reducing cases of morbidity
among populations exposed to S02 in 2020 for the selected standard and the alternative
standard levels in 2006$. For the selected S02 standard at 75 ppb, the monetized benefits from
reduced S02 exposure would be $2.2 million in 2020. Figure 5.4 shows the breakdown of the
monetized S02 benefits by health endpoint. Table 5.5 shows the incidences of health effects
and monetized benefits of attaining the alternative standard levels by health endpoint.
Because all health effects from S02 exposure are expected to occur within the analysis year, the
monetized benefits for S02 do not need to be discounted. Please note that these benefits do
not include any of the benefits listed as "unqualified" in Table 5.1, nor do they include the PM
co-benefits, which are presented in the section 5.7.
Figure 5.4: Breakdown of Monetized SO2 Health Benefits by Endpoint
ER Visits
4%
Hospital
Admissions
50%
5-20
-------
Table 5.5: SO2 Health Benefits of Attaining Alternate Standard Levels in 2020 in 2006$
(95th percentile confidence interval)
Incidence
J2
Q.
Q.
O
J2
Q.
Q.
Q.
Q.
8
Acute Respiratory Symptoms
Hospital Admissions, Respiratory
Asthma Exacerbation
Emergency Room Visits, Respiratory
Acute Respiratory Symptoms
Hospital Admissions, Respiratory
Asthma Exacerbation
Emergency Room Visits, Respiratory
Acute Respiratory Symptoms
Hospital Admissions, Respiratory
Asthma Exacerbation
Emergency Room Visits, Respiratory
38,000
170
55,000
930
9,400
46
14,000
260
2,600
13
3,800
74
(-21,000-
(-10 --
(7,800 -
(-230 --
(-5,200 --
(-3-
(1,900 -
(-65 --
(-1,500-
(-1-
(530 -
(-19 --
- 97,000)
360)
130,000)
2,600)
24,000)
•95)
33,000)
720)
- 6,700)
•27)
9,200)
200)
Valuation
$1,100,000
$4,100,000
$2,900,000
$340,000
Total $8,500,000
$280,000
$1,100,000
$720,000
$95,000
Total $2,200,000
$80,000
$310,000
$200,000
$27,000
Total $620,000
(-$730,000 -- $4,200,000)
($120,000 --$8, 100,000)
($440,000 - $8,800,000)
(-$53,000 - $940,000)
(-$210,000 - $22,000,000)
(-$180,000 --$1,100,000)
($33,000 -- $2,100,000)
($110,000 --$2,200,000)
(-$15,000 - $260,000)
(-$52,000 -- $5,600,000)
(-$50,000 - $290,000)
($9,500 -- $620,000)
($30,000 --$6 10,000)
(-$4,400 - $74,000)
(-$15,000 -- $1,600,000)
*AII estimates are rounded to two significant figures. The negative 5th percentile incidence estimates for acute
respiratory symptoms are a result of the weak statistical power of the study and should not be inferred to indicate
that decreased S02 exposure may cause an increase in this health endpoint.
In Table 5.6, we present the results of sensitivity analyses for the S02 benefits. We
indicate each input parameter, the value used as the default, and the values for the sensitivity
analyses, and then we provide the total monetary benefits for each input and the percent
change from the default value.
Table 5.6: Sensitivity Analyses for SO2 Health Benefits to Fully Attain 50 ppb Standard
Exposure Estimation Method
Location of Hospital Admission
Studies
Asthma Pooling Method
Interpolation Method
50km radius
75km radius
100km radius
150km radius
w/US-based studies only
w/Canada-based studies only
Pool all endpoints together
One or more symptoms only
Inverse distance squared
Inverse distance
Total SO2 Benefits
(millions of 2006$)
$2.2
$2.7
$3.1
$3.7
$2.2
$12
$2.2
$2.2
$2.2
$2.5
% Change
from Default
N/A
25%
42%
71%
N/A
438%
N/A
-0.2%
N/A
12%
5-21
-------
5.7 PM2.5 Co-Benefits
Because S02 is also a precursor to PM2.5, reducing S02 emissions in the projected non-
attainment areas will also reduce PM2.5 formation, human exposure and the incidence of PM2.5-
related health effects. In this analysis, we estimated the co-benefits of reducing PM2.5 exposure
for the alternative standards. Due to analytical limitations, it was not possible to provide a
comprehensive estimate of PM2.5-related benefits. Instead, we used the "benefit-per-ton"
method to estimate these benefits (Fann et al, 2009). Please see Chapter 4 for more
information on the tons of emission reductions calculated for the control strategy.7
The PM2.5 benefit-per-ton methodology incorporates key assumptions described in
detail below. These PM2.5 benefit-per-ton estimates provide the total monetized human health
benefits (the sum of premature mortality and premature morbidity) of reducing one ton of
PM2.5from a specified source. EPA has used the benefit per-ton technique in previous RIAs,
including the recent Ozone NAAQS RIA (U.S. EPA, 2010a) and N02 NAAQS RIA (U.S. EPA, 2010b).
Table 5.7 shows the quantified and unquantified benefits captured in those benefit-per-ton
estimates.
Table 5.7: Human Health and Welfare Effects of PM
2.5
Pollutant /
Effect
Quantified and Monetized
in Primary Estimates
Unquantified Effects
Changes in:
PM,
Adult premature mortality
Bronchitis: chronic and acute
Hospital admissions: respiratory and
cardiovascular
Emergency room visits for asthma
Nonfatal heart attacks (myocardial infarction)
Lower and upper respiratory illness
Minor restricted-activity days
Work loss days
Asthma exacerbations (asthmatic population)
Infant mortality
Subchronic bronchitis cases
Low birth weight
Pulmonary function
Chronic respiratory diseases other than chronic
bronchitis
Non-asthma respiratory emergency room visits
Visibility
Household soiling
Consistent with the Portland Cement NESHAP, the benefits estimates utilize the
concentration-response functions as reported in the epidemiology literature, as well as the 12
functions obtained in EPA's expert elicitation study as a sensitivity analysis.
Pollution controls installed to comply with this standard would also reduce ambient PM2.s concentrations. This
illustrative analysis is incremental to the 2006 PM NAAQS, so these benefits are in addition to those estimates
for that rule. Furthermore, the controls installed to comply with this standard might also help states attain a
more stringent PM NAAQS if one is promulgated in 2011.
5-22
-------
• One estimate is based on the concentration-response (C-R) function developed
from the extended analysis of American Cancer Society (ACS) cohort, as reported in
Pope et al. (2002), a study that EPA has previously used to generate its primary
benefits estimate. When calculating the estimate, EPA applied the effect coefficient
as reported in the study without an adjustment for assumed concentration
threshold of 10 u.g/m3 as was done in recent (2006-2009) Office of Air and Radiation
RIAs.
• One estimate is based on the C-R function developed from the extended analysis
of the Harvard Six Cities cohort, as reported by Laden et al. (2006). This study,
published after the completion of the Staff Paper for the 2006 PM2.5 NAAQS, has
been used as an alternative estimate in the PM2.5 NAAQS RIA and PM2.5 co-benefits
estimates in RIAs completed since the PM2.5 NAAQS. When calculating the estimate,
EPA applied the effect coefficient as reported in the study without an adjustment for
assumed concentration threshold of 10 u.g/m3 as was done in recent (2006-2009)
RIAs.
• Twelve estimates are based on the C-R functions from EPA's expert elicitation
study (lEc, 2006; Roman et al., 2008) on the PM2.5 -mortality relationship and
interpreted for benefits analysis in EPA's final RIA for the PM2.5 NAAQS. For that
study, twelve experts (labeled A through L) provided independent estimates of the
PM2.5 -mortality concentration-response function. EPA practice has been to develop
independent estimates of PM2.5 -mortality estimates corresponding to the
concentration-response function provided by each of the twelve experts, to better
characterize the degree of variability in the expert responses.
The effect coefficients are drawn from epidemiology studies examining two large
population cohorts: the American Cancer Society cohort (Pope et al., 2002) and the Harvard Six
Cities cohort (Laden et al., 2006).8 These are logical choices for anchor points in our
presentation because, while both studies are well designed and peer reviewed, there are
strengths and weaknesses inherent in each, which we believe argues for using both studies to
generate benefits estimates. Previously, EPA had calculated benefits based on these two
empirical studies, but derived the range of benefits, including the minimum and maximum
results, from an expert elicitation of the relationship between exposure to PM2.5and premature
mortality (Roman et al., 2008). Within this assessment, we include the benefits estimates
derived from the concentration-response function provided by each of the twelve experts to
better characterize the uncertainty in the concentration-response function for mortality and
the degree of variability in the expert responses. Because the experts used these cohort
studies to inform their concentration-response functions, benefits estimates using these
functions generally fall between results using these epidemiology studies (see Figure 5.1). In
These two studies specify multi-pollutant models that control for SO2, among other co-pollutants.
5-23
-------
general, the expert elicitation results support the conclusion that the benefits of PM2.5 control
are very likely to be substantial.
Readers interested in reviewing the general methodology for creating the benefit-per-
ton estimates used in this analysis should consult Fann et al. (2009) or the Technical Support
Document (TSD) accompanying the ozone NAAQS RIA (USEPA 2008a). As described in the
documentation for the benefit per-ton estimates cited above, national per-ton estimates are
developed for selected pollutant/source category combinations. The per-ton values calculated
therefore apply only to tons reduced from those specific pollutant/source combinations (e.g.,
S02 emitted from electric generating units; S02 emitted from area sources). Our estimate of
PM2.5 co-control benefits is therefore based on the total PM2.5 emissions controlled by sector
and multiplied by this per-ton value.
The benefit-per-ton coefficients in this analysis were derived using modified versions of
the health impact functions used in the PM NAAQS Regulatory Impact Analysis. Specifically,
this analysis uses the benefit-per-ton estimates first applied in the Portland Cement NESHAP
RIA (U.S. EPA, 2009a), which incorporated three updates: a new population dataset, an
expanded geographic scope of the benefit-per-ton calculation, and the functions directly from
the epidemiology studies without an adjustment for an assumed threshold.9 Removing the
threshold assumption is a key difference between the method used in this analysis of PM-co
benefits and the methods used in RIAs prior to Portland Cement, and we now calculate
incremental benefits down to the lowest modeled PM2.5 air quality levels.
EPA strives to use the best available science to support our benefits analyses, and we
recognize that interpretation of the science regarding air pollution and health is dynamic and
evolving. Based on our review of the body of scientific literature, EPA applied the no-threshold
model in this analysis. EPA's final Integrated Science Assessment (2009d), which was recently
reviewed by EPA's Clean Air Scientific Advisory Committee (U.S. EPA-SAB, 2009a; U.S. EPA-SAB,
2009b), concluded that the scientific literature consistently finds that a no-threshold log-linear
model most adequately portrays the PM-mortality concentration-response relationship while
recognizing potential uncertainty about the exact shape of the concentration-response
function. In Table 5-12, we include an estimate of the sensitivity of the results to an assumed
threshold at 10 u.g/m3.
As is the nature of Regulatory Impact Analyses (RIAs), the assumptions and methods
used to estimate air quality benefits evolve over time to reflect the Agency's most current
9 The benefit-per-ton estimates have also been updated since the Cement RIA to incorporate a revised VSL, as
discussed on the next page.
5-24
-------
interpretation of the scientific and economic literature. Fora period of time (2004-2008), the
Office of Air and Radiation (OAR) valued mortality risk reductions using a value of statistical life
(VSL) estimate derived from a limited analysis of some of the available studies. OAR arrived at a
VSL using a range of $1 million to $10 million (2000$) consistent with two meta-analyses of the
wage-risk literature. The $1 million value represented the lower end of the interquartile range
from the Mrozek and Taylor (2002) meta-analysis of 33 studies. The $10 million value
represented the upper end of the interquartile range from the Viscusi and Aldy (2003) meta-
analysis of 43 studies. The mean estimate of $5.5 million (2000$)10 was also consistent with the
mean VSL of $5.4 million estimated in the Kochi et al. (2006) meta-analysis. However, the
Agency neither changed its official guidance on the use of VSL in rule-makings nor subjected the
interim estimate to a scientific peer-review process through the Science Advisory Board (SAB)
or other peer-review group.
During this time, the Agency continued work to update its guidance on valuing mortality
risk reductions, including commissioning a report from meta-analytic experts to evaluate
methodological questions raised by EPA and the SAB on combining estimates from the various
data sources. In addition, the Agency consulted several times with the Science Advisory Board
Environmental Economics Advisory Committee (SAB-EEAC) on the issue. With input from the
meta-analytic experts, the SAB-EEAC advised the Agency to update its guidance using specific,
appropriate meta-analytic techniques to combine estimates from unique data sources and
different studies, including those using different methodologies (i.e., wage-risk and stated
preference) (U.S. EPA-SAB, 2007).
Until updated guidance is available, the Agency determined that a single, peer-reviewed
estimate applied consistently best reflects the SAB-EEAC advice it has received. Therefore, the
Agency has decided to apply the VSL that was vetted and endorsed by the SAB in the Guidelines
for Preparing Economic Analyses (U.S. EPA, 2000)11 while the Agency continues its efforts to
update its guidance on this issue. This approach calculates a mean value across VSL estimates
derived from 26 labor market and contingent valuation studies published between 1974 and
1991. The mean VSL across these studies is $6.3 million (2000$).12 The Agency is committed to
using scientifically sound, appropriately reviewed evidence in valuing mortality risk reductions
10 After adjusting the VSL to account for a different currency year (2006$) and to account for income growth to
2020, the $5.5 million VSL is $7.7m.
11 In the (draft) update of the Economic Guidelines (U.S. EPA, 2008d), EPA retained the VSL endorsed by the SAB
with the understanding that further updates to the mortality risk valuation guidance would be forthcoming in
the near future. Therefore, this report does not represent final agency policy.
In this analysis, we adjust the VSL to account for a different currency year (2006$) and to account for income
growth to 2020. After applying these adjustments to the $6.3 million value, the VSL is $8.9m.
5-25
-------
and has made significant progress in responding to the SAB-EEAC's specific recommendations.
The Agency anticipates presenting results from this effort to the SAB-EEAC in Spring 2010 and
that draft guidance will be available shortly thereafter.
Table 5.8 provides the unit values used to monetize the benefits of reduced exposure to
PM2.5. Figure 5.5 illustrates the relative breakdown of the monetized PM25 health benefits.
Table 5.8: Unit Values used for Economic Valuation of PM2.5 Health Endpoints (2006$)*
Health Endpoint
Central Estimate
of Value Per
Statistical
Incidence (2020
income level)
Derivation of Distributions of Estimates
Premature
Mortality
(Value of a
Statistical Life)
$8,900,000
EPA currently recommends a central VSL of $6.3m (2000$) based on
a Weibull distribution fitted to 26 published VSL estimates (5
contingent valuation and 21 labor market studies). The underlying
studies, the distribution parameters, and other useful information
are available in Appendix B of EPA's current Guidelines for Preparing
Economic Analyses (U.S. EPA, 2000).
Chronic Bronchitis
(CB)
$490,000
The WTP to avoid a case of pollution-related CB is calculated as WTPX
= WTP13 * e"p*(13"x>, where x is the severity of an average CB case,
WTP13 is the WTP for a severe case of CB, and $ is the parameter
relating WTP to severity, based on the regression results reported in
Krupnick and Cropper (1992). The distribution of WTP for an average
severity-level case of CB was generated by Monte Carlo methods,
drawing from each of three distributions: (1) WTP to avoid a severe
case of CB is assigned a 1/9 probability of being each of the first nine
deciles of the distribution of WTP responses in Viscusi et al. (1991); 2)
the severity of a pollution-related case of CB (relative to the case
described in the Viscusi study) is assumed to have a triangular
distribution, with the most likely value at severity level 6.5 and
endpoints at 1.0 and 12.0; and (3) the constant in the elasticity of
WTP with respect to severity is normally distributed with mean =
0.18 and standard deviation = 0.0669 (from Krupnick and Cropper
[1992]). This process and the rationale for choosing it is described in
detail in the Costs and Benefits of the Clean Air Act, 1990 to 2010
(U.S. EPA, 1999b).
Nonfatal Myocardial Infarction
(heart attack)
3% discount rate
Age 0-24
Age 25-44
Age 45-54
Age 55-65
Age 66 and over
$80,000
$96,000
$100,000
$180,000
$80,000
No distributional information available. Age-specific cost-of-illness
values reflect lost earnings and direct medical costs over a 5-year on
period following a nonfatal Ml. Lost earnings estimates are based
Cropper and Krupnick (1990). Direct medical costs are based on
simple average of estimates from Russell et al. (1998) and Wittels et
al. (1990).
Lost earnings: Cropper and Krupnick (1990). Present discounted
value of 5 years of lost earnings in (2006$):
age of onset: at 3%, at 7%
25-44: $11,000, $10,000
45-54: $17,000, $15,000
55-65: $96,000, $86,000
5-26
-------
7% discount rate
Age 0-24
Age 25-44
Age 45-54
Age 55-65
Age 66 and over
$80,000
$88,000
$92,000
$160,000
$78,000
Direct medical expenses: An average of:
1. Wittels et al. (1990) ($130,000—no discounting)
2. Russell et al. (1998), 5-year period ($29,000 at 3%, $27,000 at
Hospital Admissions and ER Visits
Chronic
Obstructive
Pulmonary Disease
(COPD)
$17,000
No distributional information available. The COI estimates (lost
earnings plus direct medical costs) are based on ICD-9 code-level
information (e.g., average hospital care costs, average length of
hospital stay, and weighted share of total COPD category illnesses)
reported in Agency for Healthcare Research and Quality (2000)
(www.ahrq.gov).
Asthma
Admissions
$8,900
No distributional information available. The COI estimates (lost
earnings plus direct medical costs) are based on ICD-9 code-level
information (e.g., average hospital care costs, average length of
hospital stay, and weighted share of total asthma category illnesses)
reported in Agency for Healthcare Research and Quality (2000)
(www.ahrq.gov).
All Cardiovascular
$25,000
No distributional information available. The COI estimates (lost
earnings plus direct medical costs) are based on ICD-9 code-level
information (e.g., average hospital care costs, average length of
hospital stay, and weighted share of total cardiovascular category
illnesses) reported in Agency for Healthcare Research and Quality
(2000) (www.ahrq.gov).
All respiratory
(ages 65+)
All respiratory
(ages 0-2)
No distributions available. The COI point estimates (lost earnings
plus direct medical costs) are based on ICD-9 code level information
$25,000 (e.g., average hospital care costs, average length of hospital stay, and
weighted share of total COPD category illnesses) reported in Agency
for Healthcare Research and Quality, 2000 (www.ahrq.gov).
No distributions available. The COI point estimates (lost earnings
plus direct medical costs) are based on ICD-9 code level information
$10,000 (e.g., average hospital care costs, average length of hospital stay, and
weighted share of total COPD category illnesses) reported in Agency
for Healthcare Research and Quality, 2000 (www.ahrq.gov).
Emergency Room
Visits for Asthma
$370
No distributional information available. Simple average of two unit
COI values:
(1) $400 (2006$), from Smith et al. (1997) and
(2) $340 (2006$), from Stanford et al. (1999).
Respiratory Ailments Not Requiring Hospitalization
5-27
-------
Upper Respiratory
Symptoms
(URS)
$31
Combinations of the three symptoms for which WTP estimates are
available that closely match those listed by Pope et al. result in seven
different "symptom clusters," each describing a "type" of URS. A
dollar value was derived for each type of URS, using mid-range
estimates of WTP (lEc, 1994) to avoid each symptom in the cluster
and assuming additivity of WTPs. In the absence of information
surrounding the frequency with which each of the seven types of URS
occurs within the URS symptom complex, we assumed a uniform
distribution between $11 and $50 (2006$).
Lower Respiratory
Symptoms
(LRS)
$19
Combinations of the four symptoms for which WTP estimates are
available that closely match those listed by Schwartz et al. result in
11 different "symptom clusters," each describing a "type" of LRS. A
dollar value was derived for each type of LRS, using mid-range
estimates of WTP (lEc, 1994) to avoid each symptom in the cluster
and assuming additivity of WTPs. The dollar value for LRS is the
average of the dollar values for the 11 different types of LRS. In the
absence of information surrounding the frequency with which each
of the 11 types of LRS occurs within the LRS symptom complex, we
assumed a uniform distribution between $8 and $29 (2006$).
Asthma
Exacerbations
$53
Asthma exacerbations are valued at $49 (2006$) per incidence, based
on the mean of average WTP estimates for the four severity
definitions of a "bad asthma day," described in Rowe and Chestnut
(1986). This study surveyed asthmatics to estimate WTP for
avoidance of a "bad asthma day," as defined by the subjects. For
purposes of valuation, an asthma exacerbation is assumed to be
equivalent to a day in which asthma is moderate or worse as
reported in the Rowe and Chestnut (1986) study. The value is
assumed have a uniform distribution between $19 and $83 (2006$).
Acute Bronchitis
$440
Assumes a 6-day episode, with the distribution of the daily value
specified as uniform with the low and high values based on those
recommended for related respiratory symptoms in Neumann et al.
(1994). The low daily estimate of $12 (2006$) is the sum of the mid-
range values recommended by lEc for two symptoms believed to be
associated with acute bronchitis: coughing and chest tightness. The
high daily estimate was taken to be twice the value of a minor
respiratory restricted-activity day, or $130 (2006$).
Work Loss Days
(WLDs)
Variable
No distribution available. Point estimate is based on county-specific
median annual wages divided by 50 (assuming 2 weeks of vacation)
and then by 5—to get median daily wage. U.S. Year 2000 Census,
compiled by Geolytics, Inc.
Minor Restricted
Activity Days
(MRADs)
$63
Median WTP estimate to avoid one MRAD from Tolley et al. (1986).
Distribution is assumed to be triangular with a minimum of $26 and a
maximum of $97 (2006$). Range is based on assumption that value
should exceed WTP for a single mild symptom (the highest estimate
for a single symptom—for eye irritation—is $19 (2006$)) and be less
than that for a WLD. The triangular distribution acknowledges that
the actual value is likely to be closer to the point estimate than either
extreme.
*AII estimates rounded to two significant figures. All values have been inflated to reflect values in 2006 dollars.
5-28
-------
Figure 5.5: Breakdown of Monetized PM2.s Health Benefits using Mortality Function from
Popeetal.*
Adult Mortality - Pope et
al.93%
Hospital Admissions, Resp
0.04%
Asthma Exacerbation 0.01%
Acute Bronchitis 0.01%
U pper Resp Symp 0.00%
Lower Resp Symp 0.00%
ER Visits, Resp 0.00%
*This pie chart is an illustrative breakdown of the monetized PM co-benefits, using the results based on Pope et al.
(2002) as an example. Using the Laden et al. (2006) function for premature mortality, the percentage of total
monetized benefits due to adult mortality would be 97%. This chart shows the breakdown using a 3% discount
rate, and the results would be similar if a 7% discount rate was used.
Because epidemiology studies have indicated that there is a lag between exposure to
PM2.5 and premature mortality, the discount rate has a substantial effect on the final monetized
benefits.13 We provide the PM co-benefit results using discount rates of 3% and 7% in Table
5.11 and the total monetized benefits (i.e., S02 and PM2.5) results using both discount rates in
Table 5.13. We test the sensitivity of the PM results to discount rates of 3% and 7% in Table
5.12.
To comply with Circular A-4, EPA provides monetized benefits using discount rates of 3% and 7% (OMB, 2003).
These benefits are estimated for a specific analysis year (i.e., 2020), and most of the PM benefits occur within
that year with two exceptions: acute myocardial infarctions (AMIs) and premature mortality. For AMIs, we
assume 5 years of follow-up medical costs and lost wages. For premature mortality, we assume that there is a
"cessation" lag between PM exposures and the total realization of changes in health effects. Although the
structure of the lag is uncertain, EPA follows the advice of the SAB-HES to assume a segmented lag structure
characterized by 30% of mortality reductions in the first year, 50% over years 2 to 5, and 20% over the years 6 to
20 after the reduction in PM2.5 (U.S. EPA-SAB, 2004). Changes in the lag assumptions do not change the total
number of estimated deaths but rather the timing of those deaths. Therefore, discounting only affects the AMI
costs after the analysis year and the valuation of premature mortalities that occur after the analysis year. As
such, the monetized benefits using a 7% discount rate are only approximately 10% less than the monetized
benefits using a 3% discount rate.
5-29
-------
The benefit-per-ton estimates are provided in Table 5.9 and the health incidences are
provided in Table 5.10. Table 5.11 shows the monetized results using the two epidemiology-
based estimates as well as the 12 expert-based estimates. Figure 5.6 provides a graphical
breakdown of the PM2.5 co-benefits by sector. Figure 5.7 provides a graphical representation of
all 14 of the PM2.5 co-benefits, at both a 3 percent and 7 percent discount rate.
Table 5.9: PM2.5 Co-benefits associated with reducing SO2 emissions (2006$)*
PM2.5 Precursor
SO2EGU:
SO2non-EGU:
SO2 area:
Benefit per Ton Estimate
(Pope)
$42,000
$30,000
$19,000
Benefit per Ton Estimate
(Laden)
$100,000
$74,000
$47,000
* Estimates have been rounded to two significant figures. Confidence intervals are not available for benefit per-ton
estimates. Estimates shown use a 3% discount rate. Estimates at a 7% discount rate would be approximately 9%
lower.
Table 5.10: Summary of Reductions in Health Incidences from PM2.5 Co-Benefits to Attain
Alternate Standard Levels in 2020*
Avoided Premature Mortality
Pope
Laden
Woodruff (Infant Mortality)
Avoided Morbidity
Chronic Bronchitis
Acute Myocardial Infarction
Hospital Admissions, Respiratory
Hospital Admissions, Cardiovascular
Emergency Room Visits, Respiratory
Acute Bronchitis
Work Loss Days
Asthma Exacerbation
Acute Respiratory Symptoms
Lower Respiratory Symptoms
Upper Respiratory Symptoms
50 ppb
5,100
13,000
20
3,500
8,600
1,300
2,800
4,900
8,200
650,000
90,000
3,900,000
98,000
74,000
75 ppb
2,300
5,900
9
1,600
3,900
570
1,300
2,200
3,700
290,000
41,000
1,700,000
44,000
33,000
100 ppb
1,100
2,900
5
780
1,900
280
620
1,100
1,800
150,000
20,000
870,000
22,000
17,000
*AII estimates are for the analysis year (2020) and are rounded to two significant figures. All fine particles are
assumed to have equivalent health effects, but each PM2.5 precursor pollutant has a different propensity to form
PM2.5. These results reflect full attainment with the various standard levels, including extrapolated tons, which
were spread across the sectors in proportion to the emissions in the county.
5-30
-------
Table 5.11: All PM2.5 Co-Benefits Estimates to Attain Alternate Standard Levels in 2020 at
discount rates of 3% and 7% (in millions of 2006$)*
Benefit-per-ton
Pope et al.
Laden et al.
Benefit-per-ton
Expert A
Expert B
Expert C
Expert D
Expert E
Expert F
Expert G
Expert H
Expert 1
Expert J
Expert K
Expert L
50 ppb
3%
Coefficients Derived
$34,000
$83,000
Coefficients Derived
$88,000
$67,000
$67,000
$47,000
$110,000
$61,000
$40,000
$50,000
$66,000
$54,000
$13,000
$49,000
7%
75 ppb
3%
7%
100
3%
Ppb
7%
from Epidemiology Literature
$31,000
$75,000
from Expert
$79,000
$61,000
$60,000
$43,000
$98,000
$55,000
$36,000
$46,000
$60,000
$49,000
$12,000
$44,000
$15,000
$37,000
Elicitation
$40,000
$30,000
$30,000
$21,000
$49,000
$27,000
$18,000
$23,000
$30,000
$24,000
$5,900
$22,000
$14,000
$34,000
$36,000
$27,000
$27,000
$19,000
$44,000
$25,000
$16,000
$21,000
$27,000
$22,000
$5,400
$20,000
$7,400
$18,000
$19,000
$15,000
$15,000
$10,000
$24,000
$13,000
$8,700
$11,000
$14,000
$12,000
$2,900
$11,000
$6,700
$16,000
$17,000
$13,000
$13,000
$9,400
$21,000
$12,000
$7,900
$9,900
$13,000
$11,000
$2,600
$9,600
All estimates are rounded to two significant figures. Estimates do not include confidence intervals because they
were derived through the benefit-per-ton technique described above. The benefits estimates from the Expert
Elicitation are provided as a reasonable characterization of the uncertainty in the mortality estimates associated
with the concentration-response function. These results reflect full attainment with the various standard levels,
including extrapolated tons, which were spread across the sectors in proportion to the emissions in the county.
In Table 5.12, we present the results of sensitivity analyses for the PM co-benefits. We
indicate each input parameter, the value used as the default, and the values for the sensitivity
analyses, and then we provide the total monetary benefits for each input and the percent
change from the default value.
5-31
-------
Table 5.12: Sensitivity Analyses for PM2.5 Health Co-Benefits to Fully Attain 75 ppb
Threshold Assumption (with
Epidemiology Study)
Discount Rate (with
Epidemiology Study)
Simulated Attainment
(using Pope)
No Threshold (Pope)
No Threshold (Laden)
Threshold (Pope)*
Threshold (Laden)*
3% (Pope)
3% (Laden)
7% (Pope)
7% (Laden)
Full attainment
Partial Attainment
Total PM2.5 Co-Benefits
(billions of 2006$)
$15
$37
$10
$22
$15
$37
$14
$34
$15
$14
% Change from
Default
N/A
N/A
-33%
-41%
N/A
N/A
-8%
-9%
N/A
-7%
*The Threshold model is not directly comparable to the no-threshold model. The threshold model estimates do
not include two technical updates, and they are based on data for 2015, instead of 2020. Directly comparable
estimates are not available.
Figure 5.6: Monetized PM2.s Co-Benefits of Fully Attaining 75 ppb by PM2.s Precursor
$40
$30
$20
$10
Popeetal Laden etal
PM2.5 Mortality Function
S02area S02non-EGU S02EGU
* All estimates are for the analysis year (2020). All fine particles are assumed to have equivalent health effects, but
each PM2.5 precursor pollutant has a different propensity to form PM2.5. Results using a 7% discount rate would
show a similar breakdown. These results reflect full attainment with the various standard levels, including
extrapolated tons, which were spread across the sectors in proportion to the emissions in the county.
5-32
-------
Figure 5.7: Monetized PM2.s Co-Benefits of Fully Attaining 75 ppb*
$50
3%DR
7%DR
Laden et al
$40
Pope eta I
X $30
5 $20
$10
$0
PM2.s mortality benefits estimates derived from 2 epidemiology functions and 12 expert functions
* This graph shows the estimated co- benefits in 2020 for the selected standard of 75 ppb using the no-threshold model
at discount rates of 3% and 7% using effect coefficients derived from the Pope et al. study and the Laden et al. study, as
well as 12 effect coefficients derived from EPA's expert elicitation on PM mortality. The results shown are not the direct
results from the studies or expert elicitation; rather, the estimates are based in part on the concentration-response
function provided in those studies. Graphs for alternative standards would show a similar pattern. These results reflect
full attainment with the various standard levels, including extrapolated tons, which were spread across the sectors
in proportion to the emissions in the county.
5.8 Summary of Total Monetized Benefits (SO2and PM2.s)
EPA estimated the monetized human health benefits of reducing cases of morbidity and
premature mortality among populations exposed to S02 and PM2.5 in 2020 for each of the
alternative standard levels in 2006$. For the selected S02 standard at 75 ppb, the total
monetized benefits would be $15 to $37 billion at a 3% discount rate and $14 to $34 billion at a
7% discount rate.
All of the results in this chapter present benefits estimates that assume full attainment
with the alternative standard levels. Partial attainment only incorporates the emission
reductions from identified controls without the extrapolated emission reductions.14 These
results are shown in Table 5.13 along with the full attainment at discount rates of 3% and 7%.
Table 5.14 shows the total incidences of avoided health effects. Figure 5.8 provides a graphical
14 See Chapter 4 for more information regarding the control strategy, including the identified and extrapolated
emission reductions.
5-33
-------
representation of all 14 total monetized benefits estimates, at both a 3 percent and 7 percent
discount rate, for the selected standard of 75 ppb, respectively.
Table 5.13: Total Monetized Benefits to attain Alternate Standard Levels at Discount Rates of
3% and 7% for Full and Partial Attainment (millions of 2006$)a'c
-Q
Q.
Q. -
O
U1
£2
a.
a. -
in
r^
£2
a.
a.
8
iH
3% Full Attainment
7% Full Attainment
3% Partial Attainment
7% Partial Attainment
3% Full Attainment
7% Full Attainment
3% Partial Attainment
7% Partial Attainment
3% Full Attainment
7% Full Attainment
3% Partial Attainment
7% Partial Attainment
S02
$8.5
$8.5
b
b
$2.2
$2.2
b
b
$0.62
$0.62
b
b
PM2.5
(Pope)
$34,000
$31,000
$30,000
$28,000
$15,000
$14,000
$14,000
$13,000
$7,400
$6,700
$6,900
$6,200
PM2.5
(Laden)
$83,000
$75,000
$74,000
$67,000
$37,000
$34,000
$35,000
$31,000
$18,000
$16,000
$17,000
$15,000
TOTAL
(with Pope)
$34,000
$31,000
$30,000
$28,000
$15,000
$14,000
$14,000
$13,000
$7,400
$6,700
$6,900
$6,200
TOTAL
(with Laden )
$83,000
$75,000
$74,000
$67,000
$37,000
$34,000
$35,000
$31,000
$18,000
$16,000
$17,000
$15,000
Estimates have been rounded to two significant figures and therefore summation may not match table estimates.
b The approach used to simulate air quality changes for SCh did not provide the data needed to distinguish partial
attainment benefits from full attainment benefits from reduced SCh exposure. Therefore, a portion of the SCh
benefits is attributable to the known controls and a portion of the SCh benefits are attributable to the extrapolated
controls.
c These models assume that all fine particles, regardless of their chemical composition, are equally potent in
causing premature mortality because there is no clear scientific evidence that would support the development of
differential effects estimates by particle type. Reductions in SO2 emissions from multiple sectors to meet the SO2
NAAQS would primarily reduce the sulfate fraction of PM2.5. Because this rule targets a specific particle precursor
(i.e., SO2), this introduces some uncertainty into the results of the analysis.
5-34
-------
Table 5.14: Summary of Reductions in Health Incidences from SO2 and PM2.5 to attain
Alternate Standard Levels*
Avoided Premature Mortality
Pope
Laden
Woodruff (Infant Mortality)
Avoided Morbidity
Chronic Bronchitis
Acute Myocardial Infarction
Hospital Admissions, Respiratory
Hospital Admissions, Cardiovascular
Emergency Room Visits, Respiratory
Acute Bronchitis
Work Loss Days
Asthma Exacerbation
Acute Respiratory Symptoms
Lower Respiratory Symptoms
Upper Respiratory Symptoms
SOppb
5,100
13,000
20
3,500
8,600
1,400
2,800
5,800
8,200
650,000
150,000
3,900,000
98,000
74,000
75ppb
2,300
5,900
9
1,600
3,900
570
1,300
2,500
3,700
290,000
54,000
1,700,000
44,000
33,000
100 ppb
1,100
2,900
5
780
1,900
280
620
1,200
1,800
150,000
24,000
870,000
22,000
17,000
*AII estimates are for the analysis year (2020) and are rounded to two significant figures. All fine particles are
assumed to have equivalent health effects, but each PM2.5 precursor pollutant has a different propensity to form
PM2.5. These results reflect full attainment with the various standard levels, including extrapolated tons, which
were spread across the sectors in proportion to the emissions in the county.
5-35
-------
Figure 5.8: Total Monetized Benefits (SO2and PM2.5) of Fully Attaining 75 ppb in 2020*
$50
3%DR
7%DR
Laden et al
$40
X $30
5 $20
Pope eta I
$10
$0
PM2.5 mortality benefits estimates derived from 2 epidemiology functions and 12 expert functions
* This graphs shows the estimated total monetized benefits in 2020 for the selected standard of 75 ppb using the no-
threshold model at discount rates of 3% and 7% using effect coefficients derived from the Pope et al. study and the
Laden et al. study, as well as 12 effect coefficients derived from EPA's expert elicitation on PM mortality. The results
shown are not the direct results from the studies or expert elicitation; rather, the estimates are based in part on the
concentration-response function provided in those studies. Graphs for alternative standards would show a similar
pattern.
5.9 Unquantified Welfare Benefits
The monetized benefits estimated in this RIA only reflect the portion of benefits
attributable to the health effect reductions associated with ambient fine particles and direct
exposure to S02. Data, resource, and methodological limitations prevented EPA from
quantifying or monetizing the benefits from several important benefit categories, including
benefits from reducing ecosystem effects and visibility impairment. In this section, we provide
a qualitative assessment of two welfare benefit categories: ecosystem benefits of reducing
sulfur deposition and visibility improvements.
5.9.1 Ecosystem Benefits of Reduced Sulfur Deposition
Ecosystem services can be generally defined as the benefits that individuals and
organizations obtain from ecosystems. EPA has defined ecological goods and services as the
"outputs of ecological functions or processes that directly or indirectly contribute to social
welfare or have the potential to do so in the future. Some outputs may be bought and sold, but
5-36
-------
most are not marketed" (U.S. EPA, 2006c). Figure 5.9 provides the Millennium Ecosystem
Assessment's schematic demonstrating the connections between the categories of ecosystem
services and human well-being. The interrelatedness of these categories means that any one
ecosystem may provide multiple services. Changes in these services can affect human well-
being by affecting security, health, social relationships, and access to basic material goods
(MEA, 2005).
In the Millennium Ecosystem Assessment (MEA, 2005), ecosystem services are classified
into four main categories:
1. Provisioning: Products obtained from ecosystems, such as the production of food and
water
2. Regulating: Benefits obtained from the regulation of ecosystem processes, such as the
control of climate and disease
3. Cultural: Nonmaterial benefits that people obtain from ecosystems through spiritual
enrichment, cognitive development, reflection, recreation, and aesthetic experiences
4. Supporting: Services necessary for the production of all other ecosystem services, such
as nutrient cycles and crop pollination
Figure 5.9. Linkages between categories of ecosystem services and components of human
well-being from Millennium Ecosystem Assessment (MEA, 2005)
CONSTITUENTS OFWELL-BEING
Security
PERSONAL SAFETY
SECURE RESOURCEACCESS
SECURITY FROM DISASTERS
ECOSYSTEM SERVICES
Supporting
NUTRIENT CYCLING
SOIL FORMATION
PRIMARY PRODUCTION
Provisioning
FOOD ^
FRESHWATER •
WOOD AND FIBER
FUEL
Regulating
CLIMATE REGULATION
FLOOD REGULATION
DISEASE REGULATION M
WATER PURIFICATION
Cultural
AESTHETIC *
SPIRITUAL
EDUCATIONAL
RECREATIONAL
LIFE ON EARTH - BIODIVERSITY
Basic material
for good life
ADEQUATE LIVEUHOODS
SUFFICIENT NUTRITIOUS FOOD
SHELTER
ACCESS TO GOODS
Health
STRENGTH
FEELING WELL
ACCESS TO CLEAN AIR
AND WATER
Good social relations
SOCIAL COHESION
MUTUAL RESPECT
ABILITY TO HELP OTHERS
Freedom
of choice
and action
OPPORTUNITY TO BE
ABLE TO ACHIEVE
WHAT AN INDIVIDUAL
VALUES DOING
AND BEING
Source: Millennium Ecosystem Assessment
The monetization of ecosystem services generally involves estimating the value of
ecological goods and services based on what people are willing to pay (WTP) to increase
ecological services or by what people are willing to accept (WTA) in compensation for
5-37
-------
reductions in them (U.S. EPA, 2006c). There are three primary approaches for estimating the
monetary value of ecosystem services: market-based approaches, revealed preference
methods, and stated preference methods (U.S. EPA, 2006c). Because economic valuation of
ecosystem services can be difficult, nonmonetary valuation using biophysical measurements
and concepts also can be used. An example of a nonmonetary valuation method is the use of
relative-value indicators (e.g., a flow chart indicating uses of a water body, such as beatable,
fishable, swimmable, etc.). It is necessary to recognize that in the analysis of the environmental
responses associated with any particular policy or environmental management action, only a
subset of the ecosystem services likely to be affected are readily identified. Of those ecosystem
services that are identified, only a subset of the changes can be quantified. Within those
services whose changes can be quantified, only a few will likely be monetized, and many will
remain nonmonetized. The stepwise concept leading up to the valuation of ecosystems
services is graphically depicted in Figure 5.10.
Figure 5.10: Schematic of the benefits assessment process (U.S. EPA, 2006c]
EPA action
Ecological goods and sen/ices
affected by the policy
Planning and problem formulation
[
Goods and sen/ices
identified
Ecological analysis
Goods and services
quantified
Goods and
services not
identified
Identified
goods and
services not
quantified
Quantified
goods and
services not
monetized
Science of Sulfur Deposition
Sulfur emissions occur over large regions of North America. Once these pollutants are
lofted to the middle and upper troposphere, they typically have a much longer lifetime and,
with the generally stronger winds at these altitudes, can be transported long distances from
their source regions. The length scale of this transport is highly variable owing to differing
chemical and meteorological conditions encountered along the transport path (U.S. EPA,
2008f). Sulfur is primarily emitted as S02, and secondary particles are formed from SOX gaseous
5-38
-------
emissions and associated chemical reactions in the atmosphere. Deposition can occur in either
a wet (i.e., rain, snow, sleet, hail, clouds, or fog) or dry form (i.e., gases or particles). Together
these emissions are deposited onto terrestrial and aquatic ecosystems across the U.S.,
contributing to the problems of acidification, nutrient enrichment, and methylmercury
production as represented in Figure 5-11.
Figure 5-11: Schematic of Ecological Effects of Sulfur Deposition
SO2 Atmospheric
Fate and Transpor:
Deposition
Processes
Acidification
MeHg Production
Aquatic
Terrestrial
Aquatic
Terrestrial
The lifetimes of particles vary with particle size. Accumulation-mode particles such as
sulfates are kept in suspension by normal air motions and have a lower deposition velocity than
coarse-mode particles; they can be transported thousands of kilometers and remain in the
atmosphere for a number of days. They are removed from the atmosphere primarily by cloud
processes. Particulates affect acid deposition by serving as cloud condensation nuclei and
contribute directly to the acidification of rain. In addition, the gas-phase species that lead to
the dry deposition of acidity are also precursors of particles. Therefore, reductions in S02
emissions will decrease both acid deposition and PM concentrations, but not necessarily in a
linear fashion (U.S. EPA, 2008f). Sulfuric acid is also deposited on surfaces by dry deposition
and can contribute to environmental effects (U.S. EPA, 2008f).
Ecological Effects of Acidification
Deposition of sulfur can cause acidification, which alters biogeochemistry and affects
animal and plant life in terrestrial and aquatic ecosystems across the U.S. Soil acidification is a
natural process, but is often accelerated by acidifying deposition, which can decrease
concentrations of exchangeable base cations in soils (U.S. EPA, 2008f). Major terrestrial effects
5-39
-------
include a decline in sensitive tree species, such as red spruce (Picea rubens) and sugar maple
(Acer saccharum) (U.S. EPA, 2008f). Biological effects of acidification in terrestrial ecosystems
are generally linked to aluminum toxicity and decreased ability of plant roots to take up base
cations (U.S. EPA, 2008f). Decreases in the acid neutralizing capacity and increases in inorganic
aluminum concentration contribute to declines in zooplankton, macro invertebrates, and fish
species richness in aquatic ecosystems (U.S. EPA, 2008f).
Geology (particularly surficial geology) is the principal factor governing the sensitivity of
terrestrial and aquatic ecosystems to acidification from sulfur deposition (U.S. EPA, 2008f).
Geologic formations having low base cation supply generally underlie the watersheds of acid-
sensitive lakes and streams. Other factors contribute to the sensitivity of soils and surface
waters to acidifying deposition, including topography, soil chemistry, land use, and hydrologic
flow path (U.S. EPA, 2008f).
Aquatic Ecosystems
Aquatic effects of acidification have been well studied in the U.S. and elsewhere at
various trophic levels. These studies indicate that aquatic biota have been affected by
acidification at virtually all levels of the food web in acid sensitive aquatic ecosystems. Effects
have been most clearly documented for fish, aquatic insects, other invertebrates, and algae.
Biological effects are primarily attributable to a combination of low pH and high inorganic
aluminum concentrations. Such conditions occur more frequently during rainfall and snowmelt
that cause high flows of water and less commonly during low-flow conditions, except where
chronic acidity conditions are severe. Biological effects of episodes include reduced fish
condition factor15, changes in species composition and declines in aquatic species richness
across multiple taxa, ecosystems and regions. These conditions may also result in direct fish
mortality (Van Sickle et al., 1996). Biological effects in aquatic ecosystems can be divided into
two major categories: effects on health, vigor, and reproductive success; and effects on
biodiversity. Surface water with ANC values greater than 50 u.eq/L generally provides moderate
protection for most fish (i.e., brook trout, others) and other aquatic organisms (U.S. EPA,
2009c). Table 5-15 provides a summary of the biological effects experienced at various ANC
levels.
15 Condition factor is an index that describes the relationship between fish weight and length, and is one measure
of sublethal acidification stress that has been used to quantify effects of acidification on an individual fish
(U.S.EPA, 2008f).
5-40
-------
Table 5-15: Aquatic Status Categories
Category Label ANC Levels Expected Ecological Effects
Acute
Concern
Severe
Concern
Elevated
Concern
Moderate
Concern
Low
Concern
<0 micro
equivalent per
Liter (u.eq/L)
0-20 u.eq/L
20-50 u.eq/L
50-100 u.eq/L
>100 u.eq/L
Near complete loss offish populations is expected. Planktonic communities
have extremely low diversity and are dominated by acidophilic forms. The
number of individuals in plankton species that are present is greatly reduced.
Highly sensitive to episodic acidification. During episodes of high acidifying
deposition, brook trout populations may experience lethal effects. Diversity and
distribution of zooplankton communities decline sharply.
Fish species richness is greatly reduced (i.e., more than half of expected species
can be missing). On average, brook trout populations experience sublethal
effects, including loss of health, reproduction capacity, and fitness. Diversity
and distribution of zooplankton communities decline.
Fish species richness begins to decline (i.e., sensitive species are lost from
lakes). Brook trout populations are sensitive and variable, with possible
sublethal effects. Diversity and distribution of zooplankton communities also
begin to decline as species that are sensitive to acidifying deposition are
affected.
Fish species richness may be unaffected. Reproducing brook trout populations
are expected where habitat is suitable. Zooplankton communities are
unaffected and exhibit expected diversity and distribution.
A number of national and regional assessments have been conducted to estimate the
distribution and extent of surface water acidity in the U.S (U.S. EPA, 2008f). As a result, several
regions of the U.S. have been identified as containing a large number of lakes and streams that
are seriously impacted by acidification. Figure 5-12 illustrates those areas of the U.S. where
aquatic ecosystems are at risk from acidification.
5-41
-------
Figure 5-12: Areas Potentially Sensitive to Aquatic Acidification (U.S. EPA, 2008f]
| Hgn Potential Sensitivity
Acid Sensitive Wteters :USC5 .
I Slatts.
750 1 ,CK>0
Because acidification primarily affects the diversity and abundance of aquatic biota, it
also affects the ecosystem services that are derived from the fish and other aquatic life found in
these surface waters.
While acidification is unlikely to have serious negative effects on, for example, water
supplies, it can limit the productivity of surface waters as a source of food (i.e., fish). In the
northeastern United States, the surface waters affected by acidification are not a major source
of commercially raised or caught fish; however, they are a source of food for some recreational
and subsistence fishermen and for other consumers. For example, there is evidence that
certain population subgroups in the northeastern United States, such as the Hmong and
Chippewa ethnic groups, have particularly high rates of self-caught fish consumption (Hutchison
and Kraft, 1994; Peterson et al., 1994). However, it is not known if and how their consumption
patterns are affected by the reductions in available fish populations caused by surface water
acidification.
Inland surface waters support several cultural services, including aesthetic and
educational services and recreational fishing. Recreational fishing in lakes and streams is
among the most popular outdoor recreational activities in the northeastern United States.
5-42
-------
Based on studies conducted in the northeastern United States, Kaval and Loomis (2003)
estimated average consumer surplus values per day of $36 for recreational fishing (in 2007
dollars); therefore, the implied total annual value of freshwater fishing in the northeastern
United States was $5.1 billion in 2006.16 For recreation days, consumer surplus value is most
commonly measured using recreation demand, travel cost models.
Another estimate of the overarching ecological benefits associated with reducing lake
acidification levels in Adirondacks National Park can be derived from the contingent valuation
(CV) survey (Banzhaf et al., 2006), which elicited values for specific improvements in
acidification-related water quality and ecological conditions in Adirondack lakes. The survey
described a base version with minor improvements said to result from the program, and a
scope version with large improvements due to the program and a gradually worsening status
quo. After adapting and transferring the results of this study and converting the 10-year annual
payments to permanent annual payments using discount rates of 3% and 5%, the WTP
estimates ranged from $48 to $107 per year per household (in 2004 dollars) for the base
version and $54 to $154 for the scope version. Using these estimates, the aggregate annual
benefits of eliminating all anthropogenic sources of NOX and SOX emissions were estimated to
range from $291 million to $829 million (U.S. EPA, 2009c).17
In addition, inland surface waters provide a number of regulating services associated
with hydrological and climate regulation by providing environments that sustain aquatic food
webs. These services are disrupted by the toxic effects of acidification on fish and other aquatic
life. Although it is difficult to quantify these services and how they are affected by acidification,
some of these services may be captured through measures of provisioning and cultural services.
Terrestrial Ecosystems
Acidifying deposition has altered major biogeochemical processes in the U.S. by
increasing the nitrogen and sulfur content of soils, accelerating nitrate and sulfate leaching
from soil to drainage waters, depleting base cations (especially calcium and magnesium) from
soils, and increasing the mobility of aluminum. Inorganic aluminum is toxic to some tree roots.
Plants affected by high levels of aluminum from the soil often have reduced root growth, which
restricts the ability of the plant to take up water and nutrients, especially calcium (U. S. EPA,
2008f). These direct effects can, in turn, influence the response of these plants to climatic
16 These estimates reflect the total value of the service, not the marginal change in the value of the service as a
17 These estimates reflect the total value of the service, not the marginal change in the value of the service as a
result of the emission reductions achieved by this rule.
These estimates reflect the total value of the service, nc
result of the emission reductions achieved by this rule.
5-43
-------
stresses such as droughts and cold temperatures. They can also influence the sensitivity of
plants to other stresses, including insect pests and disease (Joslin et al., 1992) leading to
increased mortality of canopy trees. In the U.S., terrestrial effects of acidification are best
described for forested ecosystems (especially red spruce and sugar maple ecosystems) with
additional information on other plant communities, including shrubs and lichen (U.S. EPA,
2008f).
Certain ecosystems in the continental U.S. are potentially sensitive to terrestrial
acidification, which is the greatest concern regarding sulfur deposition U.S. EPA (2008f). Figure
5-13 depicts the areas across the U.S. that are potentially sensitive to terrestrial acidification.
Figure 5-13: Areas Potentially Sensitive to Terrestrial Acidification (U.S. EPA, 2008f]
I Araa of Hige51 Potential Sensitivity
| Top Quartil* N
I Top Quartite S
1.CTOD
] km
Both coniferous and deciduous forests throughout the eastern U.S. are experiencing
gradual losses of base cation nutrients from the soil due to accelerated leaching for acidifying
deposition. This change in nutrient availability may reduce the quality of forest nutrition over
the long term. Evidence suggests that red spruce and sugar maple in some areas in the eastern
U.S. have experienced declining health because of this deposition. For red spruce, (Picea
rubens) dieback or decline has been observed across high elevation landscapes of the
northeastern U.S., and to a lesser extent, the southeastern U.S., and acidifying deposition has
been implicated as a causal factor (DeHayes et al., 1999). Figure 5-14 shows the distribution of
red spruce (brown) and sugar maple (green) in the eastern U.S.
5-44
-------
Figure 5-14: Distribution of Red Spruce (pink) and Sugar Maple (green) in the Eastern U.S.
(U.S. EPA, 2008f)
Terrestrial acidification affects several important ecological endpoints, including
declines in habitat for threatened and endangered species (cultural), declines in forest
aesthetics (cultural), declines in forest productivity (provisioning), and increases in forest soil
erosion and reductions in water retention (cultural and regulating).
Forests in the northeastern United States provide several important and valuable
provisioning services in the form of tree products. Sugar maples are a particularly important
commercial hardwood tree species, providing timber and maple syrup. In the United States,
sugar maple saw timber was nearly 900 million board feet in 2006 (USFS, 2006), and annual
production of maple syrup was nearly 1.4 million gallons, accounting for approximately 19% of
worldwide production. The total annual value of U.S. production in these years was
approximately $160 million (NASS, 2008). Red spruce is also used in a variety of products
including lumber, pulpwood, poles, plywood, and musical instruments. The total removal of
red spruce saw timber from timberland in the United States was over 300 million board feet in
2006 (USFS, 2006).
Forests in the northeastern United States are also an important source of cultural
ecosystem services—nonuse (i.e., existence value for threatened and endangered species),
5-45
-------
recreational, and aesthetic services. Red spruce forests are home to two federally listed species
and one delisted species:
1. Spruce-fir moss spider (Microhexura montivaga)—endangered
2. Rock gnome lichen (Gymnoderma lineare)—endangered
3. Virginia northern flying squirrel (Glaucomys sabrinusfuscus)—de\\sted, but important
Forestlands support a wide variety of outdoor recreational activities, including fishing,
hiking, camping, off-road driving, hunting, and wildlife viewing. Regional statistics on
recreational activities that are specifically forest based are not available; however, more
general data on outdoor recreation provide some insights into the overall level of recreational
services provided by forests. More than 30% of the U.S. adult population visited a wilderness
or primitive area during the previous year and engaged in day hiking (Cordell et al., 2005).
From 1999 to 2004, 16% of adults in the northeastern United States participated in off-road
vehicle recreation, for an average of 27 days per year (Cordell et al., 2005). The average
consumer surplus value per day of off-road driving in the United States was $25 (in 2007
dollars), and the implied total annual value of off-road driving recreation in the northeastern
United States was more than $9 billion (Kaval and Loomis, 2003). More than 5% of adults in the
northeastern United States participated in nearly 84 million hunting days (U.S. FWS and U.S.
Census Bureau, 2007). Ten percent of adults in northeastern states participated in wildlife
viewing away from home on 122 million days in 2006. For these recreational activities in the
northeastern United States, Kaval and Loomis (2003) estimated average consumer surplus
values per day of $52 for hunting and $34 for wildlife viewing (in 2007 dollars). The implied
total annual value of hunting and wildlife viewing in the northeastern United States was,
therefore, $4.4 billion and $4.2 billion, respectively, in 2006.
As previously mentioned, it is difficult to estimate the portion of these recreational
services that are specifically attributable to forests and to the health of specific tree species.
However, one recreational activity that is directly dependent on forest conditions is fall color
viewing. Sugar maple trees, in particular, are known for their bright colors and are, therefore,
an essential aesthetic component of most fall color landscapes. A survey of residents in the
Great Lakes area found that roughly 30% of residents reported at least one trip in the previous
year involving fall color viewing (Spencer and Holecek, 2007). In a separate study conducted in
Vermont, Brown (2002) reported that more than 22% of households visiting Vermont in 2001
made the trip primarily for viewing fall colors.
Two studies estimated values for protecting high-elevation spruce forests in the
southern Appalachian Mountains. Kramer et al. (2003) conducted a contingent valuation study
estimating households' WTP for programs to protect remaining high-elevation spruce forests
5-46
-------
from damages associated with air pollution and insect infestation. Median household WTP was
estimated to be roughly $29 (in 2007 dollars) for a smaller program, and $44 for the more
extensive program. Jenkins et al. (2002) conducted a very similar study in seven Southern
Appalachian states on a potential program to maintain forest conditions at status quo levels.
The overall mean annual WTP for the forest protection programs was $208 (in 2007 dollars).
Multiplying the average WTP estimate from these studies by the total number of households in
the seven-state Appalachian region results in an aggregate annual range of $470 million to $3.4
billion for avoiding a significant decline in the health of high-elevation spruce forests in the
Southern Appalachian region.
Forests in the northeastern United States also support and provide a wide variety of
valuable regulating services, including soil stabilization and erosion control, water regulation,
and climate regulation. The total value of these ecosystem services is very difficult to quantify
in a meaningful way, as is the reduction in the value of these services associated with total
sulfur deposition. As terrestrial acidification contributes to root damages, reduced biomass
growth, and tree mortality, all of these services are likely to be affected; however, the
magnitude of these impacts is currently very uncertain.
Ecological Effects of Associated with Sulfate in the Mercury Methylation Process
Mercury is a highly neurotoxic contaminant that enters the food web as a methylated
compound, methylmercury (U.S. EPA, 2008f). The contaminant is concentrated in higher
trophic levels, including fish eaten by humans. Experimental evidence has established that only
inconsequential amounts of methylmercury can be produced in the absence of sulfate (U.S.
EPA, 2008f). Many variables influence how much mercury accumulates in fish, but elevated
mercury levels in fish can only occur where substantial amounts of methylmercury are present
(U.S. EPA, 2008f). Current evidence indicates that in watersheds where mercury is present,
increased sulfate deposition very likely results in methylmercury accumulation in fish (Drevnick
et al., 2007; Munthe et al., 2007). The ISA for Oxides of Nitrogen and Sulfur: Ecological Criteria
ISA concluded that evidence is sufficient to infer a casual relationship between sulfur deposition
and increased mercury methylation in wetlands and aquatic environments (U.S. EPA, 2008f).
Establishing the quantitative relationship between sulfate and mercury methylation in
natural settings is difficult because of the presence of multiple interacting factors in aquatic and
terrestrial environments, including wetlands, aquatic environments where sulfate, sulfur-
reducing bacteria (SRB), and inorganic mercury are present (U.S. EPA, 2008f). These are the
three primary requirements for bacterially-mediated conversion to methylmercury. Additional
factors affecting conversion include the presence of anoxic conditions, temperature, the
5-47
-------
presence and types of organic matter, the presence and types of mercury-binding species, and
watershed effects (e.g., watershed type, land cover, water body limnology, and runoff loading).
With regard to methylmercury, the highest concentrations in the environment generally occur
at or near the sedimentary surface, below the oxic-anoxic boundary. Although mercury
methylation can occur within the water column, there is generally a far greater contribution of
mercury methylation from sediments because of anoxia and of greater concentrations of SRB,
substrate, and sulfate. Figure 5-15 depicts the mercury cycle.
Figure 5-15: The mercury cycle in an ecosystem (USGS, 2006)
ATMOSPHERIC DEPOSITION
\
DRY DEPOSITION
VOLATILIZATION and
RE-DEPOSITION
OVERLAND RUNOFF ST°RMWATER °ISCHARGE ^ °OUTFLOW
DE-METHYLATION
>METHYLATION
METHYLMERCURY FOOD-CHAIN MAGNIFICATION
DIFFUSION and
RE-SUSPENSION
.- _y^5
GROUND-WATER DISCHARGE' '^.Vt-^ O « DE-METHYLATION
/ I* ' *,»"_.%.• . f T " Wsi—k—jfc
^e"'
Figure 5-16 illustrates a map of mercury-sensitive watersheds based on sulfate
concentrations, ANC, levels of dissolved organic carbon and pH, mercury species
concentrations, and soil types to gauge the methylation sensitivity (Myers et al., 2007).
5-48
-------
Figure 5.16: Preliminary USGS map of mercury methylation-sensitive watersheds
(Myers et al., 2007)
Mercury sensitivity
itorc (unitlessJ
Decreases in sulfate deposition/emissions have already shown reductions in
methylmercury (U.S. EPA, 2008f). Observed decreases in methylmercury fish tissue
concentrations have been linked to decreased acidification and declining sulfate and mercury
deposition (Hrabik and Watras, 2002; Drevnick et al., 2007).
In the U.S., consumption offish and shellfish are the main sources of methylmercury
exposure to humans. Methylmercury builds up more in some types offish and shellfish than in
others. The levels of methylmercury in high and shellfish vary widely depending on what they
eat, how long they live, and how high they are in the food chain. Most fish, including ocean
species and local freshwater fish, contain some methylmercury. For example, in recent studies
by EPA and the U.S. Geological Survey (USGS) of fish tissues, every fish samples contained some
methylmercury.
State-level fish consumption advisories for mercury are based on state criteria, many of
which are based on EPA's fish tissue criterion for methylmercury (U.S. EPA, 2001) or on U.S.
Food and Drug Administration's action levels (U.S. FDA, 2001). In 2008, there were 3,361 fish
advisories issued at least in part for mercury contamination (80% of all fish advisories), covering
5-49
-------
16.8 million lake acres (40% of total lake acreage) and 1.3 million river miles (35% of total river
miles) over all 50 states, one U.S. territory, and 3 tribes (U.S. EPA, 2009f). Recently, the U.S.
Geological Survey (USGS) examined mercury levels in top-predator fish, bed sediment, and
water from 291 streams across the U.S. (Scudder et al., 2009). USGS detected mercury
contamination in every fish sampled, and the concentration of mercury in fish exceeded EPA's
criterion in 27% of the sites sampled.
The ecosystem service most directly affected by sulfate-mediated mercury methylation
is the provision offish for consumption as a food source. This service is of particular
importance to groups engaged in subsistence fishing, pregnant women and young children.
While it is not possible to quantify the reduction in fish consumption due to the presence of
methylmercury in fish from sulfur deposition, it is likely, given the number of state advisories
and the EPA/FDA guidelines (U.S. EPA/FDA, 2004) on consumption for pregnant women and
young children, that this service is negatively affected.
Research shows that most people's fish consumption does not cause a mercury-related
health concern. However, certain people may be at higher risk because of their routinely high
consumption of fish (e.g., tribal and other subsistence fishers and their families who rely heavily
on fish fora substantial part of their diet). It has been demonstrated that high levels of
methylmercury in the bloodstream of unborn babies and young children may harm the
developing nervous system, making the child less able to think and learn. Moreover, mercury
exposure at high levels can harm the brain, heart, kidneys, lungs, and immune system of people
of all ages. The majority of fish consumed in the U.S. are ocean species. The methylmercury
concentrations in ocean fish species are primarily influences by the global mercury pool.
However, the methylmercury found in local fish can be due, at least partly, to mercury
emissions from local sources.
Several studies suggest that the methylmercury content of fish may reduce these
cardio-protective effects offish consumption. Some of these studies also suggest that
methylmercury may cause adverse effects to the cardiovascular system. For example, the NRC
(2000) review of the literature concerning methylmercury health effects took note of two
epidemiological studies that found an association between dietary exposure to methylmercury
and adverse cardiovascular effects.18 Moreover, in a study of 1,833 males in Finland aged 42 to
60 years, Solonen et al. (1995) observed a relationship between methylmercury exposure via
18 National Research Council (NRC). 2000. Toxicological Effects of Methylmercury. Committee on the Toxicological
Effects of Methylmercury, Board on Environmental Studies and Toxicology. National Academies Press.
Washington, DC. pp.168-173.
5-50
-------
fish consumption and acute myocardial infarction (AMI or heart attacks), coronary heart
disease, cardiovascular disease, and all-cause mortality.19 The NRCalso noted a study of 917
seven year old children in the Faroe Islands, whose initial exposure to methylmercury was in
utero although post natal exposures may have occurred as well. At seven years of age, these
children exhibited an increase in blood pressure and a decrease in heart rate variability.20 Based
on these and other studies, NRC concluded in 2000 that, while "the data base is not as
extensive for cardiovascular effects as it is for other end points (i.e. neurologic effects) the
cardiovascular system appears to be a target for methylmercury toxicity."21
Since publication of the NRC report there have been some 30 published papers
presenting the findings of studies that have examined the possible cardiovascular effects of
methylmercury exposure. These studies include epidemiological, toxicological, and
toxicokinetic investigations. Over a dozen review papers have also been published. If there is
a causal relationship between methylmercury exposure and adverse cardiovascular effects,
then reducing exposure to methylmercury would result in public health benefits from reduced
cardiovascular effects.
In early 2010, EPA sponsored a workshop in which a group of experts were asked to
assess the plausibility of a causal relationship between methylmercury exposure and
cardiovascular health effects and to advise EPA on methodologies for estimating population
level cardiovascular health impacts of reduced methylmercury exposure. The report from that
workshop is in preparation.
Because establishing the quantitative relationship between sulfate and mercury
methylation in natural settings is difficult, we were unable to model the changes in the
methylation process, bioaccumulation in fish tissue, and human consumption of mercury-
contaminated fish that would be needed in order to estimate the human health benefits from
reducing sulfate emissions in this rule.
19Salonen, J.T., Seppanen, K. Nyyssonen et al. 1995. "Intake of mercury from fish lipid peroxidation, and the risk of
myocardial infarction and coronary, cardiovascular and any death in Eastern Finnish men." Circulation, 91
(3):645-655.
20Sorensen, N, K. Murata, E. Budtz-Jorgensen, P. Weihe, and Grandjean, P., 1999. "Prenatal Methylmercury
Exposure As A Cardiovascular Risk Factor At Seven Years of Age", Epidemiology, pp370-375.
21National Research Council (NRC). 2000. Toxicological Effects of Methylmercury. Committee on the Toxicological
Effects of Methylmercury, Board on Environmental Studies and Toxicology. National Academies Press.
Washington, DC. p. 229.
5-51
-------
Ecological Effects Associated with Gaseous Sulfur Dioxide
Uptake of gaseous sulfur dioxide in a plant canopy is a complex process involving
adsorption to surfaces (leaves, stems, and soil) and absorption into leaves. S02 penetrates into
leaves through to the stomata, although there is evidence for limited pathways via the cuticle.
Pollutants must be transported from the bulk air to the leaf boundary layer in order to get to
the stomata. When the stomata are closed, as occurs under dark or drought conditions,
resistance to gas uptake is very high and the plant has a very low degree of susceptibility to
injury. In contrast, mosses and lichens do not have a protective cuticle barrier to gaseous
pollutants or stomates and are generally more sensitive to gaseous sulfur than vascular plants
(U.S. EPA, 2008f). Acute foliar injury usually happens within hours of exposure, involves a rapid
absorption of a toxic dose, and involves collapse or necrosis of plant tissues. Another type of
visible injury is termed chronic injury and is usually a result of variable S02 exposures over the
growing season. Besides foliar injury, chronic exposure to low S02 concentrations can result in
reduced photosynthesis, growth, and yield of plants (U.S. EPA, 2008f). These effects are
cumulative over the season and are often not associated with visible foliar injury. As with foliar
injury, these effects vary among species and growing environment. S02 is also considered the
primary factor causing the death of lichens in many urban and industrial areas (Hutchinson et
al., 1996).
5.9.2 Visibility Improvements
Reductions in S02 emissions and secondary formation of PM2.5due to the alternative
standards will improve the level of visibility throughout the United States. These suspended
particles and gases degrade visibility by scattering and absorbing light. Visibility directly affects
people's enjoyment of a variety of daily activities. Individuals value visibility both in the places
they live and work, in the places they travel to for recreational purposes, and at sites of unique
public value, such as the Great Smokey Mountains National Park. Without the necessary air
quality data, we were unable to calculate the predicted change in visibility due to control
strategy to attain various alternate standard levels. However, in this section, we describe the
process by which S02 emissions impair visibility and how this impairment affects the public.
Visual air quality (VAQ) is commonly measured as either light extinction, which is defined
as the loss of light per unit of distance in terms of inverse mega meters (Mm"1) or the deciview
(dv) metric (Pitchford and Malm, 1993), which is a logarithmic function of extinction. Extinction
and deciviews are physical measures of the amount of visibility impairment (e.g., the amount of
"haze"), with both extinction and deciview increasing as the amount of haze increases.
Pitchford and Malm characterize a change of one deciview as "a small but perceptible scenic
5-52
-------
change under many circumstances." Light extinction is the optical characteristic of the
atmosphere that occurs when light is either scattered or absorbed, which converts the light to
heat. Particulate matter and gases can both scatter and absorb light. Fine particles with
significant light-extinction efficiencies include sulfates, nitrates, organic carbon, elemental
carbon, and soil (Sisler, 1996). The extent to which any amount of light extinction affects a
person's ability to view a scene depends on both scene and light characteristics. For example,
the appearance of a nearby object (i.e. a building) is generally less sensitive to a change in light
extinction than the appearance of a similar object at a greater distance. See Figure 5-17 for an
illustration of the important factors affecting visibility.
Figure 5-17: Important factors involved in seeing a scenic vista (Malm, 1999)
Light from clouds
scattered into
mage-forming
light scattered
out of sight path
L|gM
from ground
scattered Into
sight path
Image-forming
light absorbed
In conjunction with the U.S. National Park Service, the U.S. Forest Service, other Federal
land managers, and State organizations in the U.S., the U.S. EPA has supported visibility
monitoring in national parks and wilderness areas since 1988. The monitoring network known
as IMPROVE (Interagency Monitoring of Protected Visual Environments) now includes 150 sites
that represent almost all of the Class I areas across the country (see Figure 5-18) (U.S. EPA,
2009d).
5-53
-------
Figure 5-18: Mandatory Class I Areas in the U.S.
Produced by IMPS Air Resources Division
* Rainbow Lake, WI and Brad well Bay, FL are Class 1 Areas
where visibility is not an important air quality related value
Annual average visibility conditions (reflecting light extinction due to both
anthropogenic and non-anthropogenic sources) vary regionally across the U.S. (U.S. EPA,
2009d). The rural East generally has higher levels of impairment than remote sites in the West,
with the exception of urban-influenced sites such as San Gorgonio Wilderness (CA) and Point
Reyes National Seashore (CA), which have annual average levels comparable to certain sites in
the Northeast (U.S. EPA, 2004). Higher visibility impairment levels in the East are due to
generally higher concentrations of fine particles, particularly sulfates, and higher average
relative humidity levels. While visibility trends have improved in most Class I areas, the recent
data show that these areas continue to suffer from visibility impairment. In eastern parks,
average visual range has decreased from 90 miles to 15-25 miles, and in the West, visual range
has decreased from 140 miles to 35-90 miles (U.S. EPA, 2004; U.S. EPA, 1999b).
Visibility has direct significance to people's enjoyment of daily activities and their overall
sense of wellbeing (U.S. EPA, 2009d). Good visibility increases the quality of life where
individuals live and work, and where they engage in recreational activities. When the necessary
AQ data is available, EPA generally considers benefits from these two categories of visibility
changes: residential visibility (i.e., the visibility in and around the locations where people live)
and recreational visibility (i.e., visibility at Class I national parks and wilderness areas.) In both
5-54
-------
cases, economic benefits are believed to consist of use values and nonuse values. Use values
include the aesthetic benefits of better visibility, improved road and air safety, and enhanced
recreation in activities like hunting and bird watching. Nonuse values are based on people's
beliefs that the environment ought to exist free of human-induced haze. Nonuse values may be
more important for recreational areas, particularly national parks and monuments. In addition,
evidence suggests that an individual's WTP for improvements in visibility at a Class I area is
influenced by whether it is in the region in which the individual lives, or whether it is
somewhere else (Chestnut and Rowe, 1990). In general, people appear to be willing to pay
more for visibility improvements at parks and wilderness areas that are "in-region" than at
those that are "out-of-region." This is plausible, because people are more likely to visit, be
familiar with, and care about parks and wilderness areas in their own part of the country. EPA
generally uses a contingent valuation study as the basis for monetary estimates of the benefits
of visibility changes in recreational areas (Chestnut and Rowe, 1990). To estimate the
monetized value of visibility changes, an analyst would multiply the willingness-to-pay
estimates by the amount of visibility impairment, but this information in unavailable for this
analysis.
5.10 Limitations and Uncertainties
The National Research Council (NRC) (2002) concluded that EPA's general methodology
for calculating the benefits of reducing air pollution is reasonable and informative in spite of
inherent uncertainties. To address these inherent uncertainties, NRC highlighted the need to
conduct rigorous quantitative analysis of uncertainty and to present benefits estimates to
decisionmakers in ways that foster an appropriate appreciation of their inherent uncertainty.
In response to these comments, EPA's Office of Air and Radiation (OAR) is developing a
comprehensive strategy for characterizing the aggregate impact of uncertainty in key modeling
elements on both health incidence and benefits estimates. Components of that strategy
include emissions modeling, air quality modeling, health effects incidence estimation, and
valuation.
In this analysis, we use three methods to assess uncertainty quantitatively: Monte Carlo
analysis, sensitivity analysis, and alternate concentration-response functions for PM mortality.
We also provide a qualitative assessment for those aspects that we are unable to address
quantitatively in this analysis. Each of these analyses is described in detail in the following
sections.
This analysis includes many data sources as inputs, including emission inventories, air
quality data from models (with their associated parameters and inputs), population data, health
5-55
-------
effect estimates from epidemiology studies, and economic data for monetizing benefits. Each
of these inputs may be uncertain and would affect the benefits estimate. When the
uncertainties from each stage of the analysis are compounded, small uncertainties can have
large effects on the total quantified benefits. In this analysis, we are unable to quantify the
cumulative effect of all of these uncertainties, but we provide the following analyses to
characterize many of the largest sources of uncertainty.
5.10.1 Monte Carlo analysis
Similar to other recent RIAs, we used Monte Carlo methods for characterizing random
sampling error associated with the concentration response functions and economic valuation
functions. Monte Carlo simulation uses random sampling from distributions of parameters to
characterize the effects of uncertainty on output variables, such as incidence of morbidity.
Specifically, we used Monte Carlo methods to generate confidence intervals around the
estimated health impact and dollar benefits. The reported standard errors in the
epidemiological studies determined the distributions for individual effect estimates, as shown
in Table 5.6 for S02 benefits. Unfortunately, the associated confidence intervals are not
available for the PM2.s co-benefits due to limitations in the benefit-per-ton methodology.
5.10.2 Sensitivity analyses
We performed a variety of sensitivity analyses on the benefits results to assess the
sensitivity of the primary results to various data inputs and assumptions. We then changed
each default input one at a time and recalculated the total monetized benefits to assess the
percent change from the default. In Tables 5.6 and 5.12, we provided the results of this
sensitivity analysis. We indicate each input parameter, the value used as the default, and the
values for the sensitivity analyses, and then we provide the total monetary benefits for each
input and the percent change from the default value. This sensitivity analysis indicates that the
results are most sensitive to assumptions regarding the attainment status and the threshold
assumption in the PM-mortality relationship, and the results are less sensitive to alternate
assumptions regarding the interpolation method, discount rate, and various assumptions
regarding S02 exposure. To account for the large difference in magnitude between benefits
from reduced S02 exposure and PM2.s exposure, we provide separate sensitivity analyses. We
show the sensitivity analysis for selected standard (75 ppb), but other standard levels would
show similar sensitivity to these perturbations, albeit with smaller magnitudes. Descriptions of
the sensitivity analyses are provided in the relevant sections of this chapter.
5-56
-------
5.10.3 Alternate concentration-response functions for PM mortality
PM2.5 mortality co-benefits are the largest benefit category that we monetized in this
analysis. To better understand the concentration-response relationship between PM2.5
exposure and premature mortality, EPA conducted an expert elicitation in 2006 (Roman et al.,
2008; lEc, 2006). In general, the results of the expert elicitation support the conclusion that the
benefits of PM2.5 control are very likely to be substantial. In previous RIAs, EPA presented
benefits estimates using concentration response functions derived from the PM2.5 Expert
Elicitation as a range from the lowest expert value (Expert K) to the highest expert value (Expert
E). However, this approach did not indicate the agency's judgment on what the best estimate
of PM benefits may be, and EPA's Science Advisory Board described this presentation as
misleading. Therefore, we began to present the cohort-based studies (Pope et al, 2002; and
Laden et al., 2006) as our core estimates in the Portland Cement RIA (U.S. EPA, 2009a). Using
alternate relationships between PM2.5and premature mortality supplied by experts, higher and
lower benefits estimates are plausible, but most of the expert-based estimates fall between the
two epidemiology-based estimates (Roman et al., 2008).
In this analysis, we present the results derived from the expert elicitation as indicative of
the uncertainty associated with a major component of the health impact functions, and we
provide the independent estimates derived from each of the twelve experts to better
characterize the degree of variability in the expert responses. In this chapter, we provide the
results using the concentration-response functions derived from the expert elicitation in both
tabular (Table 5.11) and graphical form (Figure 5.1). Please note that these results are not the
direct results from the studies or expert elicitation; rather, the estimates are based in part on
the concentration-response function provided in those studies. Because in this RIA we estimate
PM co-benefits using benefit-per-ton estimates, technical limitations prevent us from providing
the associated credible intervals with the expert functions.
5.10.4 Qualitative assessment of uncertainty and other analysis limitations
Although we strive to incorporate as many quantitative assessments of uncertainty,
there are several aspects for which we are only able to address qualitatively. These aspects are
important factors to consider when evaluating the relative benefits of the attainment strategies
for each of the alternative standards:
5-57
-------
1. The 12 km by 12 km resolution of the air quality modeling grid may be too coarse to
accurately estimate the potential near-field health benefits of reducing S02 emissions.
These uncertainties likely result in an underestimate of the S02-related benefits.
2. The interpolation techniques used to estimate the full attainment benefits from reduced
S02 exposure of the alternative standards contributed some uncertainty to the analysis.
The great majority of benefits estimated for the various standard levels were derived
through interpolation. As noted previously in this chapter, these benefits are likely to
be more uncertain than if we had modeled the air quality scenario for both S02and
PM2.5. In general, the VNA interpolation approach will underestimate benefits because
it does not account for the broader spatial distribution of air quality changes that may
occur due to the implementation of a regional emission control program.
3. There are many uncertainties associated with the health impact functions used in this
modeling effort. These include: within study variability (the precision with which a given
study estimates the relationship between air quality changes and health effects); across
study variation (different published studies of the same pollutant/health effect
relationship typically do not report identical findings and in some instances the
differences are substantial); the application of C-R functions nationwide (does not
account for any relationship between region and health effect, to the extent that such a
relationship exists); extrapolation of impact functions across population (we assumed
that certain health impact functions applied to age ranges broader than that considered
in the original epidemiological study); and various uncertainties in the C-R function,
including causality and thresholds. These uncertainties may under- or over-estimate
benefits.
4. Co-pollutants present in the ambient air may have contributed to the health effects
attributed to S02 in single pollutant models. Risks attributed to S02 might be
overestimated where concentration-response functions are based on single pollutant
models. If co-pollutants are highly correlated with S02, their inclusion in an S02 health
effects model can lead to misleading conclusions in identifying a specific causal
pollutant. Because this collinearity exists, many of the studies reported statistically
insignificant effect estimates for both S02 and the co-pollutants; this is due in part to the
loss of statistical power as these models control for co-pollutants. Where available, we
have selected multipollutant effect estimates to control for the potential confounding
effects of co-pollutants; these include NYDOH (2006), Schwartz et al. (1994) and
O'Connor et al. (2008). The remaining studies include single pollutant models.
5. This analysis is for the year 2020, and projecting key variables introduces uncertainty.
Inherent in any analysis of future regulatory programs are uncertainties in projecting
5-58
-------
atmospheric conditions and source level emissions, as well as population, health
baselines, incomes, technology, and other factors.
6. This analysis omits certain unquantified effects due to lack of data, time and resources.
These unquantified endpoints include other health effects, ecosystem effects, and
visibility. EPA will continue to evaluate new methods and models and select those most
appropriate for estimating the benefits of reductions in air pollution. Enhanced
collaboration between air quality modelers, epidemiologists, toxicologists, ecologists,
and economists should result in a more tightly integrated analytical framework for
measuring benefits of air pollution policies.
7. PM2.5 co-benefits represent a substantial proportion of total monetized benefits (over
99% of total monetized benefits), and these estimates are subject to a number of
assumptions and uncertainties.
a. PM2.5 co-benefits were derived through benefit per-ton estimates, which do not
reflect local variability in population density, meteorology, exposure, baseline
health incidence rates, or other local factors that might lead to an over-estimate
or under-estimate of the actual benefits of controlling directly emitted fine
particulates.
b. We assume that all fine particles, regardless of their chemical composition, are
equally potent in causing premature mortality. This is an important assumption,
because PM2.5 produced via transported precursors emitted from EGUs may
differ significantly from direct PM2.5 released from diesel engines and other
industrial sources, but no clear scientific grounds exist for supporting differential
effects estimates by particle type.
c. We assume that the health impact function for fine particles is linear down to
the lowest air quality levels modeled in this analysis. Thus, the estimates include
health benefits from reducing fine particles in areas with varied concentrations
of PM2.5; including both regions that are in attainment with fine particle standard
and those that do not meet the standard down to the lowest modeled
concentrations.
d. To characterize the uncertainty in the relationship between PM2.5and premature
mortality, we include a set of twelve estimates based on results of the expert
elicitation study in addition to our core estimates. Even these multiple
characterizations omit the uncertainty in air quality estimates, baseline incidence
rates, populations exposed and transferability of the effect estimate to diverse
locations. As a result, the reported confidence intervals and range of estimates
give an incomplete picture about the overall uncertainty in the PM2.5 estimates.
5-59
-------
This information should be interpreted within the context of the larger
uncertainty surrounding the entire analysis. For more information on the
uncertainties associated with PM2.5 co-benefits, please consult the PM2.5 NAAQS
RIA (Table 5.5).
5.11 Discussion
The results of this benefits analysis suggest that fully attaining the selected S02 standard
of 75 ppb would produce important health benefits from reduced S02 exposure in the form of
fewer respiratory hospitalizations, respiratory emergency department visits and cases of acute
respiratory symptoms. In addition, attaining the selected S02 standard standards would also
produce substantial health co-benefits from reducing PM2.5 exposure in the form of avoided
premature mortality and other morbidity effects.
The proposal version of this analysis was the first time that EPA has estimated the
monetized human health benefits of reducing exposure to S02to support a change in the
NAAQS. In contrast to recent PM2.5 and ozone-related benefits assessments, there was far less
analytical precedent on which to base this assessment. For this reason, we developed entirely
new components of the health impact analysis, including the identification of health endpoints
to be quantified and the selection of relevant effect estimates within the epidemiology
literature. Because we did not receive any substantive comments on this approach during the
comment period, we duplicated this methodology using the updated air quality estimates for
the final RIA. As the S02 health literature continues to evolve, EPA will reassess the health
endpoints and risk estimates used in this analysis.
While the monetized benefits of reduced S02 exposure appear small when compared to
the monetized benefits of reduced PM25 exposure, readers should not necessarily infer that the
total monetized benefits of attaining a new S02 standard are minimal. As shown in Table 5.13,
the monetized PM2.5 co-benefits represent over 99% of the total monetized benefits. This
result is consistent with other recent RIAs, where the PM2.5 co-benefits represent a large
proportion of total monetized benefits. This result is amplified in this RIA by the decision not to
quantify S02-related premature mortality and other morbidity endpoints due to the
uncertainties associated with estimating those endpoints. Studies have shown that there is a
relationship between S02exposure and premature mortality, but that relationship is limited by
potential confounding. Because premature mortality generally comprises over 90% of the total
monetized benefits, this decision may substantially underestimate the monetized health
benefits of reduced S02 exposure.
5-60
-------
We were unable to quantify the benefits from several welfare benefit categories. We
lacked the necessary air quality data to quantify the benefits from improvements in visibility
from reducing light-scattering particles. Previous RIAs for ozone (U.S. EPA, 2008a) and PM2.s
(U.S. EPA, 2006a) indicate that visibility is an important benefit category, and previous efforts to
monetize those benefits have only included a subset of visibility benefits, excluding benefits in
urban areas and many national and state parks. Even this subset accounted for up to 5% of
total monetized benefits in the Ozone NAAQS RIA (U.S. EPA, 2008a).
We were also unable to quantify the ecosystem benefits of reduced sulfur deposition
because we lacked the necessary air quality data and resources to run the ecosystem benefits
models. Previous assessments (U.S. EPA, 1999a; U.S. EPA, 2005; U.S. EPA, 2009e) indicate that
ecosystem benefits are also an important benefits category, but those efforts were only able to
monetize a tiny subset of ecosystem benefits in specific geographic locations, such as
recreational fishing effects from lake acidification in the Adirondacks. We were also unable to
quantify the benefits of decreased mercury methylation from sulfate deposition. Quantifying
the relationship between sulfate and mercury methylation in natural settings is difficult, but
some studies have shown that decreasing sulfate deposition can also decrease methylmercury.
In section 5.7 of this RIA, we discuss the revised presentation using benefits based on
Pope et al. and Laden et al. as the core estimates instead of using the range based on the low
and high end of the expert elicitation. This change was incorporated in direct response to
recommendations from EPA's Science Advisory Board (U.S.EPA-SAB, 2008). Although using
benefit-per-ton estimates limited our ability to incorporate all of their suggestions fully, we
have incorporated the following recommendations into this analysis:
• Added "bottom line" statements where appropriate
• Clarified that the benefits results shown are not the actual judgments of the experts
• Acknowledged uncertainties exist at each stage of the analytic process, although
difficult to quantify when using benefit-per-ton estimates
• Did not use the expert elicitation range to characterize the uncertainty as it focuses on
the most extreme judgments with zero weight to all the others,
• Described the rationale for using expert elicitation in the context of the regulatory
process (to characterize uncertainty)
• Identified results based on epidemiology studies and expert elicitation separately
• Showed central mass of expert opinion using graphs
• Presented the quantitative results using diverse tables and more graphics
5-61
-------
5.12 References
Abt Associates. 2008. Environmental Benefits and Mapping Program (Version 3.0). Bethesda,
MD. Prepared for U.S. Environmental Protection Agency Office of Air Quality Planning and
Standards. Research Triangle Park, NC. Available on the Internet at
.
Agency for Healthcare Research and Quality (AHRQ). 2000. HCUPnet, Healthcare Cost and
Utilization Project.
Banzhaf, S., D. Burtraw, D. Evans, and A. Krupnick. 2006. "Valuation of Natural Resource
Improvements in the Adirondacks." Land Economics 82:445-464.
Brown, L.H. 2002. Profile of the Annual Fall Foliage Tourist in Vermont: Travel Year 2001.
Report prepared for the Vermont Department of Tourism and Marketing and the Vermont
Tourism Data Center in association with the University of Vermont, Burlington, VT.
Chestnut, L.G., and R.D. Rowe. 1990. A New National Park Visibility Value Estimates. In
Visibility and Fine Particles, Transactions of an AWMA/EPA International Specialty
Conference, C.V. Mathai, ed. Air and Waste Management Association, Pittsburgh.
Cordell, H.K., C.J. Betz, G. Green, and M. Owens. 2005. Off-Highway Vehicle Recreation in the
United States, Regions and States: A National Report from the National Survey on
Recreation and the Environment (NSRE). Prepared for the U.S. Department of Agriculture
Forest Service, Southern Research Station, National OHV Policy and Implementation Teams,
Athens, GA. Available on the Internet at
.
Cropper, M. L. and A. J. Krupnick. 1990. The Social Costs of Chronic Heart and Lung Disease.
Resources for the Future. Washington, DC. Discussion Paper QE 89-16-REV.
DeHayes, D.H., P.G. Schaberg, G.J. Hawley, and G.R. Strimbeck. 1999. Acid rain impacts on
calcium nutrition and forest health. Bioscience 49(10):789-800.
Delfino, R. J., H. Gong, Jr., W. S. Linn, E. D. Pellizzari and Y. Hu. 2003. Asthma symptoms in
Hispanic children and daily ambient exposures to toxic and criteria air pollutants. Environ
Health Perspect. Vol. Ill (4): 647-56.
Dewanji, A. and S. H. Moolgavkar. 2000. A Poisson process approach for recurrent event data
with environmental covariates. Environmetrics. Vol. 11: 665-673.
Dewanji, A. and S. H. Moolgavkar. 2002. Choices of stratification in Poisson process analysis of
recurrent event data with environmental covariates. Stat. Med. Vol. 21: 3383-3393.
5-62
-------
Drevnick, P.E., D.E. Canfield, P.R. Gorski, A.L.C. Shinneman, D.R. Engstrom, D.C.G. Muir, G.R.
Smith, P.J. Garrison, L.B. Cleckner, J.P. Hurley, R.B. Noble, R.R. Otter, and J.T.Oris. 2007.
Deposition and cycling of sulfur controls mercury accumulation in Isle Royale fish.
Environmental Science and Technology 41(21):7266-7272.
Fann, N., C.M. Fulcher, B.J. Hubbell. 2009. The influence of location, source, and emission type
in estimates of the human health benefits of reducing a ton of air pollution. Air Qual Atmos
Health (2009) 2:169-176.
Fung, K. Y., S. Khan, D. Krewski and Y. Chen. 2006. Association between air pollution and
multiple respiratory hospitalizations among the elderly in Vancouver, Canada. Inhal Toxicol.
Vol. 18 (13): 1005-11.
Guallar, et. al., 2002. "Mercury, Fish Oils, and the Risk of Myocardial Infarction." New England
Journal of Medicine, Vol. 374, No. 22, November.
Hallgren et al. 2001. "Markers of high fish intake are associated with decreased risk of a first
myocardial infarction," British Journal of Nutrition, 86, 397-404.
Hrabik, T.R., and C.J. Watras. 2002. Recent declines in mercury concentration in a freshwater
fishery: isolating the effects of de-acidification and decreased atmospheric mercury
deposition in Little Rock Lake. Science of the Total Environment 297:229-237.
Hutchison, R., and C.E. Kraft. 1994. Hmong Fishing Activity and Fish Consumption. Journal of
Great Lakes Research 20(2):471-487.
Industrial Economics, Incorporated (lEc). March 31., 1994. Memorandum to Jim DeMocker,
Office of Air and Radiation, Office of Policy Analysis and Review, U.S. Environmental
Protection Agency.
Industrial Economics, Inc. 2006. Expanded Expert Judgment Assessment of the Concentration-
Response Relationship Between PM2.s Exposure and Mortality. Prepared for the U.S. EPA,
Office of Air Quality Planning and Standards, September. Available on the Internet at
.
Ito, K., G. D. Thurston and R. A. Silverman. 2007. Characterization of PM2.s, gaseous pollutants,
and meteorological interactions in the context of time-series health effects models. J Expo
Sci Environ Epidemiol. Vol. 17 Suppl 2: S45-60.
Jenkins, D.H., J. Sullivan, and G.S. Amacher. 2002. Valuing high altitude spruce-fir forest
improvements: Importance of forest condition and recreation activity. Journal of Forest
Economics 8:77-99.
5-63
-------
Joslin, J.D., Kelly, J.M., van Miegroet, H. 1992. Soil chemistry and nutrition of North American
spruce-fir stands: evidence for recent change. Journal of Environmental Quality, 21, 12-30.
Kaval, P., and J. Loomis. 2003. Updated Outdoor Recreation Use Values With Emphasis On
National Park Recreation. Final Report October 2003, under Cooperative Agreement CA
1200-99-009, Project number IMDE-02-0070.
Kochi, I., B. Hubbell, and R. Kramer. 2006. An Empirical Bayes Approach to Combining
Estimates of the Value of Statistical Life for Environmental Policy Analysis. Environmental
and Resource Economics. 34: 385-406.
Kramer, A., T. Holmes, and M. Haefel. 2003. Contingent valuation of forest ecosystem
protection. Pp. 303-320 in Forests in a Market Economy. Edited by E.O. Sills and K.L. Abt.
Dordrecht, The Netherlands: Kluwer Academic Publishers.
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.
Laden, F., J. Schwartz, F.E. Speizer, and D.W. Dockery. 2006. "Reduction in Fine Particulate Air
Pollution and Mortality." American Journal of Respiratory and Critical Care Medicine
173:667-672. Estimating the Public Health Benefits of Proposed Air Pollution Regulations.
Washington, DC: The National Academies Press.
Luginaah, I. N., K. Y. Fung, K. M. Gorey, G. Webster and C. Wills. 2005. Association of ambient
air pollution with respiratory hospitalization in a government-designated "area of concern":
the case of Windsor, Ontario. Environ Health Perspect. Vol. 113 (3): 290-6.
Malm, WC. 1999. Introduction to visibility. Colorado State University, Fort Collins, CO, USA,
1983, Revised edition, 1999. Available on the Internet at
.
Michaud, J. P., J. S. Grove and D. Krupitsky. 2004. Emergency department visits and "vog"-
related air quality in Hilo, Hawai'i. Environ Res. Vol. 95 (1): 11-9.
Millennium Ecosystem Assessment Board (MEA). 2005. Ecosystems and Human Well-being:
Synthesis. Washington, DC: World Resources Institute.
Mortimer, K. M., L. M. Neas, D. W. Dockery, S. Redline and I. B. Tager. 2002. The effect of air
pollution on inner-city children with asthma. Eur Respir J. Vol. 19 (4): 699-705.
Mrozek, J.R., and L.O. Taylor. 2002. "What Determines the Value of Life? A Meta-Analysis."
Journal of Policy Analysis and Management 21(2):253-270.
5-64
-------
Munthe, J., R.A. Bodaly, B.A. Branfireun, C.T. Driscoll, C.C. Gilmour, R. Harris, M. Horvat, M.
Lucotte, and 0. Malm. 2007. Recovery of mercury-contaminated fisheries. AMBIO:A
Journal of the Human Environment 36:33-44.
Myers, M.D., M.A. Ayers, J.S. Baron, P.R. Beauchemin, K.T. Gallagher, M.B. Goldhaber, D.R.
Hutchinson, J.W. LaBaugh, R.G. Sayre, and S.E. Schwarzbach. 2007. USGS goals for the
coming decade. Science 318:200-201.
National Agricultural Statistics Service (NASS). 2008. Maple Syrup -June 12, 2008: Maple Syrup
Production Up 30 Percent Nationwide. U.S. Department of Agriculture, National
Agricultural Statistics Service, New England Agricultural Statistics, Concord, NH.
National Center for Health Statistics (NCHS). 1999a. Health Interview Survey (HIS). National
Vital Statistics Reports. U.S. Department of Health and Human Services, Centers for Disease
Control and Prevention, National Center for Health Statistics. Washington, DC. Volume 47,
Number 19. June 30.
National Center for Health Statistics (NCHS). 1999b. National Hospital Discharge Survey (NHDS).
Available on the internet at
.
National Center for Health Statistics (NCHS). 2000. National Hospital Ambulatory Medical Care
Survey (NHAMCS). Available on the internet at
.
National Research Council (NRC). 2000. Toxicological Effects of Methylmercury. Committee on
the Toxicological Effects of Methylmercury, Board on Environmental Studies and Toxicology.
National Academies Press. Washington, DC.
National Research Council (NRC). 2002. Estimating the Public Health Benefits of Proposed Air
Pollution Regulations. Washington, DC: The National Academies Press.
Neumann, J.E., M.T. Dickie, and R.E. Unsworth. March 31,1994. "Linkage Between Health
Effects Estimation and Morbidity Valuation in the Section 812 Analysis—Draft Valuation
Document." Industrial Economics Incorporated (lEc) Memorandum to Jim DeMocker, U.S.
Environmental Protection Agency, Office of Air and Radiation, Office of Policy Analysis and
Review.
New York Department of Health (NYDOH). 2006. A Study of Ambient Air Contaminants and
Asthma in New York City. New York State Department of Health Center for Environmental
Health.
O'Connor, G. T., L. Neas, B. Vaughn, M. Kattan, H. Mitchell, E. F. Grain, R. Evans, 3rd, R.
Gruchalla, W. Morgan, J. Stout, G. K. Adams and M. Lippmann. 2008. Acute respiratory
5-65
-------
health effects of air pollution on children with asthma in US inner cities. J Allergy Clin
Immunol. Vol. 121 (5): 1133-1139 el.
Peel, J. L, P. E. Tolbert, M. Klein, K. B. Metzger, W. D. Flanders, K. Todd, J. A. Mulholland, P. B.
Ryan and H. Frumkin. 2005. Ambient air pollution and respiratory emergency department
visits. Epidemiology. Vol. 16 (2): 164-74.
Peterson, D.E., M.S. Kanarek, M.A. Kuykendall, J.M. Died rich, H.A. Anderson, P.L. Remington,
and T.B. Sheffy. 1994. Fish Consumption Patterns and Blood Mercury Levels in Wisconsin
Chippewa Indians. Archives of Environmental Health 49(l):53-58.
Pitchford, M. and W. Malm. 1993. Development and applications of a standard visual index.
Atmospheric Environment 28(5):1049-1054.
Pope, C.A., III, R.T. Burnett, M.J. Thun, E.E. Calle, D. Krewski, K. Ito, and G.D. Thurston. 2002.
"Lung Cancer, Cardiopulmonary Mortality, and Long-term Exposure to Fine Particulate Air
Pollution." Journal of the American Medical Association 287:1132-1141.
Rissanen, T., S Voutilainen, T.A. Lakka, and J.T. Salonen. 2000. "Fish oil-derived fatty acids,
docasahexaenoic acid and docosapentainoic acid, and the risk of acute coronary events: The
Kuopio ischaemic heart disease risk factor study", Circulation. 102(22), pp2677-2679.
Roman, Henry A., Katherine D. Walker, Tyra L. Walsh, Lisa Conner, Harvey M. Richmond, Bryan
J. Hubbell, and Patrick L. Kinney. 2008. Expert Judgment Assessment of the Mortality
Impact of Changes in Ambient Fine Particulate Matter in the U.S. Environ. Sci. Technol.,
42(7):2268-2274.
Rowe, R. D. and L. G. Chestnut. 1986. Oxidants and Asthmatics in Los Angeles: A Benefits
Analysis - Executive Summary. Prepared for U.S. Environmental Protection Agency, Office
of Policy Analysis. Prepared by Energy and Resource Consultants, Inc. Washington, DC.
EPA-230-09-86-018. March.
Russell, M.W., D.M. Huse, S. Drowns, E.G. Hamel, and S.C. Hartz. 1998. "Direct Medical Costs of
Coronary Artery Disease in the United States." American Journal of Cardiology 81(9):1110-
1115.
Salonen, J.T., Seppanen, K. Nyyssonen et al. 1995. "Intake of mercury from fish lipid
peroxidation, and the risk of myocardial infarction and coronary, cardiovascular and any
death in Eastern Finnish men." Circulation, 91 (3):645-655.
Salonen, J.T., Seppanen,T.A. Lakka, R. Salonen and G.A. Kaplan. 2000. "Mercury accumulation
and accelerated progression of carotid atherosclerosis: A population-based prospective 4-
year follow-up study in men in Eastern Finland. Atherosclerosis. 148: pp265-273.
5-66
-------
Schildcrout, J. S., L. Sheppard, T. Lumley, J. C. Slaughter, J. Q. Koenig and G. G. Shapiro. 2006.
Ambient air pollution and asthma exacerbations in children: an eight-city analysis. Am J
Epidemiol. Vol. 164 (6): 505-17.
Schwartz, J., D. W. Dockery, L. M. Neas, D. Wypij, J. H. Ware, J. D. Spengler, P. Koutrakis, F. E.
Speizer and B. G. Ferris. 1994. Acute Effects of Summer Air Pollution on Respiratory
Symptom Reporting in Children. Am J Respir Crit Care Med. Vol. 150 (5): 1234-1242.
Schwartz, J., C. Spix, G. Touloumi, L. Bacharova, T. Barumamdzadeh, A. le Tertre, T. Piekarksi, A.
Ponce de Leon, A. Ponka, G. Rossi, M. Saez and J. P. Schouten. 1996. Methodological issues
in studies of air pollution and daily counts of deaths or hospital admissions. J Epidemiol
Community Health. Vol. 50 Suppl 1: S3-11.
Scudder, B.C., Chasar, L.C., Wentz, D.A., Bauch, N.J., Brigham, M.E., Moran, P.W., and
Krabbenhoft, D.P. 2009. Mercury in fish, bed sediment, and water from streams across the
United States, 1998-2005: U.S. Geological Survey Scientific Investigations Report 2009-
5109, 74 p.
Sisler, J.F. 1996. Spatial and seasonal patterns and long-term variability of the composition of
the haze in the United States: an analysis of data from the IMPROVE network. CIRA Report,
ISSN 0737-5352-32, Colorado State University.
Smith, D. H., D. C. Malone, K. A. Lawson, L. J. Okamoto, C. Battista and W. B. Saunders. 1997. A
national estimate of the economic costs of asthma. Am J Respir Crit Care Med. Vol. 156 (3
Pt 1): 787-93.
Sorensen, N., K. Murata, E. Budtz-Jorgensen, P. Weihe, and Grandjean, P., 1999. "Prenatal
Methylmercury Exposure As A Cardiovascular Risk Factor At Seven Years of Age",
Epidemiology, pp370-375.
Spencer, D.M., and D.F. Holecek. 2007. Basic characteristics of the fall tourism market.
Tourism Management 28:491-504.
Stanford, R., T. Mclaughlin and L. J. Okamoto. 1999. The cost of asthma in the emergency
department and hospital. Am J Respir Crit Care Med. Vol. 160 (1): 211-5.
Stern, A. 2005, "A review of the studies of the cardiovascular health effects of methylmercury
with consideration of their suitability for risk assessment," Environmental Research, 98 133-
142.
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
5-67
-------
U.S. Environmental Protection Agency (U.S. EPA). 1999a. The Benefits and Costs of the Clean
Air Act 1990 to 2010: EPA Report to Congress. EPA-410-R-99-001. Office of Air and
Radiation, Office of Policy, Washington, DC. November. Available on the Internet at <
http://www.epa.gov/air/sect812/1990-2010/fullrept.pdf>.
U.S. Environmental Protection Agency (U.S. EPA). 1999b. Regional Haze Regulations, July 1,
1999. (64 FR 35714, July 1,1999).
U.S. Environmental Protection Agency (U.S. EPA). 2000. Guidelines for Preparing Economic
Analyses. EPA 240-R-00-003. National Center for Environmental Economics, Office of Policy
Economics and Innovation. Washington, DC. September. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2001. Water Quality Criterion for the
Protection of Human Health: Methylmercury (Final). Office of Science and Technology.
Office of Water. January. Washington, DC. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2002. Letter to Recipients of Third Draft EPA
PM Criteria Document. National Center for Environmental Assessment, Research Triangle
Park, NC. June 5, 2002. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2004. Air Quality Criteria for Particulate
Matter. Volume I EPA600/P-99/002aF and Volume II EPA600/P-99/002bF. Retrieved on
March 19, 2009 from Docket EPA-HQ-OAR-2003-0190 at http://www.regulations.gov/.
U.S. Environmental Protection Agency (U.S. EPA). 2005. Regulatory Impact Analysis for the
Final Clean Air Interstate Rule. Office of Air and Radiation. March. Available on the
Internet at < http://www.epa.gov/cair/pdfs/finaltech08.pdf>.
U.S. Environmental Protection Agency (U.S. EPA). 2006a. Regulatory Impact Analysis, 2006
National Ambient Air Quality Standards for Particulate Matter, Chapter 5. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. October. Available on the
Internet at .
U.S. Environmental Protection Agency (U.S. EPA). 2006c. Ecological Benefits Assessment
Strategic Plan. EPA-240-R-06-001. Office of the Administrator. Washington, DC. October.
Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2008a. Regulatory Impact Analysis, 2008
National Ambient Air Quality Standards for Ground-level Ozone, Chapter 6. Office of Air
5-68
-------
Quality Planning and Standards, Research Triangle Park, NC. March. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2008b. Technical Support Document:
Calculating Benefit Per-Ton estimates, Ozone NAAQS Docket #EPA-HQ-OAR-2007-0225-
0284. Office of Air Quality Planning and Standards, Research Triangle Park, NC. March.
Available on the Internet at .
U.S. Environmental Protection Agency (U.S. EPA). 2008c. Integrated Science Assessment for
Sulfur Oxides - Health Criteria (Final Report). National Center for Environmental
Assessment, Research Triangle Park, NC. September. Available on the Internet at <
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=198843>.
U.S. Environmental Protection Agency (U.S. EPA). 2008d. Guidelines for Preparing Economic
Analyses: External Review Draft. National Center for Environmental Economics, Office of
Policy Economics and Innovation. Washington, DC. September. Available on the Internet
at .
U.S. Environmental Protection Agency (U.S. EPA). 2009a. Regulatory Impact Analysis: National
Emission Standards for Hazardous Air Pollutants from the Portland Cement Manufacturing
Industry. Office of Air Quality Planning and Standards, Research Triangle Park, NC. April.
Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2009c. Risk and Exposure Assessment to
Support the Review of the S02 Primary National Ambient Air Quality Standards: Final
Report. Office of Air Quality Planning and Standards, Research Triangle Park, NC. August.
Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2009d. Integrated Science Assessment for
Particulate Matter (Final Report). EPA-600-R-08-139F. National Center for Environmental
Assessment-RTP Division. December. Available on the Internet at
.
5-69
-------
U.S. Environmental Protection Agency (U.S. EPA). 2009e. Risk and Exposure Assessment for
Review of the Secondary National Ambient Air Quality Standards for Oxides of Nitrogen and
Oxides of Sulfur (Final). EPA-452/R-09-008a. Office of Air Quality Planning and Standards,
Research Triangle Park, NC. September. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA) 2010a. Regulatory Impact Analysis, National
Ambient Air Quality Standards for Ozone, Office of Air Quality Planning and Standards,
Research Triangle Park, NC. October. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2010b. Final Regulatory Impact Analysis (RIA)
for the N02 National Ambient Air Quality Standards (NAAQS). Office of Air Quality Planning
and Standards, Research Triangle Park, NC. January. Available on the Internet at
.
U.S. Environmental Protection Agency - Science Advisory Board (U.S. EPA-SAB). 2007. SAB
Advisory on EPA's Issues in Valuing Mortality Risk Reduction. EPA-SAB-08-001. October.
Available on the Internet at
.
U.S. Environmental Protection Agency - Science Advisory Board (U.S. EPA-SAB). 2008.
Characterizing Uncertainty in Particulate Matter Benefits Using Expert Elicitation. EPA-
COUNCIL-08-002. July. Available on the Internet at <
http://yosemite.epa.gov/sab%5CSABPRODUCT.NSF/43B91173651AED9E85257487004EA6C
B/$File/EPA-COUNCIL-08-002-unsigned.pdf>.
U.S. Environmental Protection Agency - Science Advisory Board (U.S. EPA-SAB). 2009a. Review
of EPA's Integrated Science Assessment for Particulate Matter (First External Review Draft,
December 2008). EPA-COUNCIL-09-008. May. Available on the Internet at
.
U.S. Environmental Protection Agency - Science Advisory Board (U.S. EPA-SAB). 2009b.
Consultation on EPA's Particulate Matter National Ambient Air Quality Standards: Scope and
Methods Plan for Health Risk and Exposure Assessment. EPA-COUNCIL-09-009. May.
Available on the Internet at
5-70
-------
.
U.S. Environmental Protection Agency and U.S. Food and Drug Administration (U.S. EPA/FDA).
2004. What You Need to Know About Mercury in Fish and Shellfish: Advice for Women Who
Might Become Pregnant Women, Who are Pregnant, Nursing Mothers, Young Children.
EPA-823-F-04-009. Available on the Internet at
.
U.S. Fish and Wildlife Service (FWS) and U.S. Census Bureau. 2007. 2006 National Survey of
Fishing, Hunting, and Wildlife-Associated Recreation. FHW/06-NAT. U.S. Department of the
Interior, U.S. Fish and Wildlife Service, Washington, DC, and U.S. Department of Commerce,
U.S. Census Bureau, Washington, DC.
U.S. Food and Drug Administration (U.S. FDA). 2001. Fish and Fisheries Products Hazards and
Controls Guidance: Chapter 10: Methyl Mercury. Third Edition. June. Available on the Internet at <
http://www.fda.gov/Food/GuidanceComplianceRegulatorylnformation/GuidanceDocuments/Seafoo
d/ucm092041.htm>.
U.S. Forest Service (USFS). 2006. Forest Inventory and Analysis National Program: Forest
Inventory Data Online. Online database. U.S. Department of Agriculture Forest Service,
Forest Inventory and Analysis, Arlington, VA. Available on the Internet at
.
U.S. Geological Survey (USGS). 2006. Investigations and Monitoring of Mercury in Indiana by
the U.S. Geological Survey. U.S. Department of the Interior, U.S. Geological Survey, Indiana
Water Science Center, Indianapolis, IN. Available athttp://in.water.usgs.gov/mercury.
Van Sickle, J., Baker, J.P., Simonin, H.A., Baldigo, B.P., Kretser, W.A., Sharpe, W.E. 1996.
Episodic acidification of small streams in the northeastern United States: Fish mortality in
field bioassays. Ecological Applications, 6, 408-421.
Villeneuve, P. J., L. Chen, B. H. Rowe and F. Coates. 2007. Outdoor air pollution and emergency
department visits for asthma among children and adults: a case-crossover study in northern
Alberta, Canada. Environ Health. Vol. 6 (1): 40.
Viscusi, W.K., W.A. Magat, and J. Huber. 1991. "Pricing Environmental Health Risks: Survey
Assessments of Risk-Risk and Risk-Dollar Trade-Offs for Chronic Bronchitis." Journal of
Environmental Economics and Management 21:32-51.
Virtanen JK, S Voutilainen, T.H. Rissanen, J Mursu, T.P.Tuomainen, M.J. Korhonen,
V.P.Walkonen, K. Seppanen,J.A. Laukkanen, J.T.Salonen.Grandjean, P., K. Murata, E. Budtz-
Jorgensen, and P.Weihe. 2004. "Autonomic Activity in Methylmercury Neurotoxicity:14-Year
Follow-Up of a Faroese Birth Cohort", The Journal of Pediatrics, February, pp!69-176.
5-71
-------
Viscusi, V.K., and J.E. Aldy. 2003. "The Value of a Statistical Life: A Critical Review of Market
Estimates throughout the World." Journal of Risk and Uncertainty 27(l):5-76.
Wilson, A. M., C. P. Wake, T. Kelly and J. C. Salloway. 2005. Air pollution, weather, and
respiratory emergency room visits in two northern New England cities: an ecological time-
series study. Environ Res. Vol. 97 (3): 312-21.
Wittels, E.H., J.W. Hay, and A.M. Gotto, Jr. 1990. Medical Costs of Coronary Artery Disease in
the United States. American Journal of Cardiology 65(7):432-440.
Yang, Q., Y. Chen, D. Krewski, R. T. Burnett, Y. Shi and K. M. McGrail. 2005. Effect of short-term
exposure to low levels of gaseous pollutants on chronic obstructive pulmonary disease
hospitalizations. Environ Res. Vol. 99 (1): 99-105.
Yang, Q., Y. Chen, Y. Shi, R. T. Burnett, K. M. McGrail and D. Krewski. 2003. Association
between ozone and respiratory admissions among children and the elderly in Vancouver,
Canada. Inhal Toxicol. Vol. 15 (13): 1297-308.
Yoshizawa, K et. al. 2002. Mercury and the Risk of Coronary Heart Disease in Men, New England
Journal of Medicine, Vol. 374, No. 22, November.
5-72
-------
Chapter 6: Cost Analysis Approach and Results
Synopsis
This chapter describes our illustrative analysis of the engineering costs and monitoring
costs associated with attaining the final and alternative standards for the National Ambient Air
Quality Standard (NAAQS) for S02. We present our analysis of these costs in four separate
sections. Section 6.1 presents the cost estimates. Sections 6.2 and 6.3 summarize the
illustrative economic and energy impacts of these standards, respectively, while Section 6.4
outlines the main limitations of the analysis. As mentioned previously, the analysis is presented
here for the final standard of 75 ppb, and two alternative standards: 50 ppb and 100 ppb in the
year 2020.
Section 6.1 breaks out discussion of cost estimates into five subsections. The first
subsection summarizes the data and methods that we employed to estimate the costs
associated with the control strategies outlined in Chapter 4. The second subsection presents
county level estimates of the costs of identified controls associated with the regulatory
alternatives examined in this RIA. Following this discussion, the third subsection describes the
approach used to estimate the extrapolated costs of unspecified emission reductions that may
be needed to comply with the final and alternative standards. The fourth subsection provides a
brief discussion of the monitoring costs associated with the final NAAQS. The fifth subsection
provides the estimated total costs of the regulatory alternatives examined. This section
concludes with a discussion of technological innovation and how that affects regulatory cost
estimates.
This analysis does not estimate the projected attainment status of areas of the country
other than those counties currently served by one of the approximately 349 monitors with
complete data in the current network. It is important to note that the final rule will require a
monitoring network wholly comprised of monitors sited at locations of expected maximum
hourly concentrations. Only about one third of the existing S02 network may be source-
oriented and/or in the locations of maximum concentration required by the final rule because
the current network is focused on population areas and community-wide ambient levels of S02.
Actual monitored levels using the new monitoring network may be higher than levels measured
using the existing network. We recognize that once a network of monitors located at
maximum-concentration is put in place, more areas could find themselves exceeding the new
S02 NAAQS. However for this RIA analysis, we lack sufficient data to predict which counties
might exceed the new NAAQS after implementation of the new monitoring network. Therefore
we lack a credible analytic path to estimating costs and benefits for such a future scenario.
6-1
-------
In addition, this chapter presents cost estimates associated with both identified control
measures and unspecified emission reductions needed to reach attainment. Identified control
measures include known measures for known sources that may be implemented to attain the
alternative standard, whereas the achievement of unspecified emission reductions requires
implementation of hypothetical additional measures in areas that would not attain the selected
standard following the implementation of identified controls to known sources.
Note that the universe of sources achieving unspecified emission reductions beyond
identified controls is not completely understood; therefore we are not able to identify known
control devices, work practices, or other control measures to achieve these reductions. We
calculated extrapolated costs for unspecified emission reductions using a fixed cost per ton
approach. The analysis presents hypothetical costs of attaining the S02 NAAQS, subject to
States' abilities to find emission reductions whose costs are finite, although likely to be higher
than those of the identified control measures we believe to exist. Section 6.1 below describes
in more detail our approaches for estimating both the costs of identified controls and the
extrapolated costs of unspecified emission reductions needed beyond identified controls.
As is discussed throughout this RIA, the technologies and control strategies selected for
this analysis are illustrative of one approach that nonattainment areas may employ to comply
with the revised S02 standard. Potential control programs may be designed and implemented
in a number of ways, and EPA anticipates that State and Local governments will consider those
programs that are best suited for local conditions. As such, the costs described in this chapter
generally cover the annualized costs of purchasing, installing, and operating the referenced
technologies. We also present monitoring costs. Because we are uncertain of the specific
actions that State Agencies will take to design State Implementation Plans to meet the revised
standard, we do not estimate the costs that government agencies may incur to implement
these control strategies.
6.1 Engineering Cost Estimates
6.1.1 Data and Methods: Identified Control Costs
Consistent with the emissions control strategy analysis presented in Chapter 4, our
analysis of the costs associated with the final S02 NAAQS focuses S02 emission controls for
ECU sources first, then nonEGU point sources, and then area sources.
6-2
-------
6.1.1.1 ECU Sources
We used equations for wet FGD scrubber controls used in the Integrated Planning
Model (IPM) to estimate the control cost for S02 reductions from EGUs. Equations are available
for estimating capital and annual costs, and these equations are dependent on unit capacity
and capacity factor (fraction of hours in a year that an ECU operates). Annual costs for control
measures applied in IPM include those for fixed and variable operating and maintenance
(O&M) items and annualized capital costs calculated using a capital recovery factor and are
specifically applicable to EGUs.
6.1.1.2 NonEGU Point and Area Sources
After designing the hypothetical control strategy using the methodology discussed in
Chapter 4, EPA used the Control Strategy Tool (CoST) and AirControlNET to estimate
engineering control costs for nonEGU and Area sources. CoST calculates engineering costs
using three different methods: (1) by multiplying an average annualized cost per ton estimate
against the total tons of a pollutant reduced to derive a total cost estimate; (2) by calculating
cost using an equation that incorporates key plant information; or (3) by using both cost per ton
and cost equations. Most control cost information within CoST has been developed based on
the cost per ton approach. This is because estimating engineering costs using an equation
requires more data, and parameters used in other non-cost per ton methods may not be readily
available or broadly representative across sources within the emissions inventory. The costing
equations used in CoST require either plant capacity or stack flow to determine annual, capital
and/or operating and maintenance (O&M) costs. Capital costs are converted to annual costs
using the capital recovery factor (CRF)1. Where possible, cost calculations are used to calculate
total annual control cost (TACC) which is a function of the capital (CC) and O&M costs. The
capital recovery factor incorporates the interest rate and equipment life (in years) of the
control equipment. Operating costs are calculated as a function of annual O&M and other
variable costs. The resulting TACC equation is TACC = (CRF * CC) + O&M.
Engineering costs will differ based upon quantity of emissions reduced, plant capacity,
or stack flow which can vary by emissions inventory year. Engineering costs will also differ in a
nominal sense by the year the costs are calculated for (i.e., 1999$ versus 2006$).2 For capital
For more information on this cost methodology and the role of AirControlNET in control strategy analysis, see
Section 6 of the 2006 PM RIA, AirControlNET 4.1 Control Measures Documentation (Pechan, 2006b), or the EPA Air
Pollution Control Cost Manual, Section 1, Chapter 2, found at http://www.epa.gov/ttn/catc/products.htmlftcccinfo.
The engineering costs will not be any different in a real (inflation-adjusted) sense if calculated in 2006 versus
1999 dollars if properly escalated. For this analysis, all costs are reported in real 2006 dollars.
6-3
-------
investment, we do not assume early capital investment in order to attain standards by 2020.
For 2020, our estimate of annualized costs represents a "snapshot" of the annualized costs,
which include annualized capital and O&M costs, for those controls included in our identified
control strategy analysis. Our engineering cost analysis uses the equivalent uniform annual
costs (EUAC) method, in which annualized costs are calculated based on the equipment life for
the control measure along with the interest rate by use of the CRF as mentioned previously in
this chapter. Annualized costs are estimated as equal for each year the control is expected to
operate. Hence, our annualized costs for nonEGU point and area sources estimated for 2020
are the same whether the control measure is installed in 2019 or in 2010. We make no
presumption of additional capital investment in years beyond 2020. The EUAC method is
discussed in detail in the EPA Air Pollution Control Cost Manual3. Applied controls and their
respective engineering costs are provided in the S02 NAAQS docket.
6.1.2 Identified Control Strategy Analysis Engineering Costs
In this section, we provide engineering cost estimates of the control strategies identified
in Chapter 4 that include control measures applied to nonEGU sources, area sources, and EGUs.
Engineering costs generally refer to the expense of capital equipment installation, the site
preparation costs for the application, and annual operating and maintenance costs.
The total annualized cost of control in each geographic area of our analysis for the
hypothetical control scenario is provided in Table 6.1. These numbers reflect the engineering
costs across all sectors. Estimates are annualized at a discount rate of 7%.
Table 6.1 summarizes these costs in total and by sector nationwide. As indicated in the
table, the estimated annualized costs of these controls under the 75 ppb final standard in 2020
are $960 million per year (2006$). For the other 2 alternative standards examined, in 2020 the
annualized costs range from $470 million to $2,600 million. Consistent with Chapter 4's
summary of the air quality impacts associated with identified controls, the cost estimates in
Table 6.1 reflect partial attainment with the alternative standard being examined in this RIA.
Consistent with the identified control strategy analysis emission reductions presented in
Chapter 4, a majority of the costs are from controls applied to ECU sources, but a relatively
large share of costs is borne by nonEGU point sources.
The costs of the ECU strategy reflect application of controls (described in Chapter 4)
where needed to obtain as much reductions as possible to attain each alternative standard.
3 http://epa.gov/ttn/catc/products.htmlftcccinfo
6-4
-------
Table 6.2 presents the identified control costs in 2020 by county for each alternative
standard. These costs are shown for a 7 percent discount rate.
Table 6.1: Annual Control Costs of Identified Controls in 2020 in Total and by Sector
(Millions of 2006$) a'b
Total Costs for Identified Controls0' d
EGUs
nonEGUs
Area Sources
50 ppb
$ 2,600
$ 1,700
$ 900
$ 40
75 ppb
$ 960
$ 700
$ 260
$ 0.55
100 ppb
$ 470
$ 300
$ 170
$ 0.24
All estimates rounded to two significant figures. As such, totals will not sum down columns.
b All estimates provided reflect the engineering cost of the identified control strategy analysis, incremental to a
2020 baseline.
c Total annualized costs were calculated using a 7% discount rate
dThese values represent partial attainment costs for the identified control strategy analysis. There were locations
not able to attain the alternative standard being analyzed with identified controls only.
Table 6.2: Identified Controls - Total Annual Cost by County in 2020 (Millions of 2006$)a'b'c'd
state
Arizona
Colorado
Connecticut
Florida
Florida
Georgia
Idaho
Illinois
Illinois
Illinois
Illinois
Illinois
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Iowa
Iowa
Kentucky
Kentucky
Louisiana
county
Gila Co
Denver Co
New Haven Co
Duval Co
Hillsborough Co
Chatham Co
Bannock Co
Cook Co
Madison Co
StClairCo
Sangamon Co
Tazewell Co
Floyd Co
Fountain Co
Jasper Co
Lake Co
Morgan Co
Porter Co
Wayne Co
Linn Co
Muscatine Co
Jefferson Co
Livingston Co
East Baton Rouge Par
50 ppb
$8.8
$39.0
$8.2
$24.0
$3.2
$42.0
$0.6
$16.0
$65.0
$60.0
$120.0
$0.14
$19.0
$210.0
$10.0
$47.0
$26.0
$89.0
$85.0
$11.0
$29.0
75 ppb
$8.8
$12.0
$31.0
$30.0
$27.0
$49.0
$47.0
$18.0
$65.0
100 ppb
$8.8
$35.0
$31.0
6-5
-------
state
Missouri
Missouri
Missouri
Montana
Nebraska
New Hampshire
New York
New York
New York
North Carolina
Ohio
Ohio
Ohio
Ohio
Oklahoma
Oklahoma
Oklahoma
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
South Carolina
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Texas
Texas
West Virginia
Wisconsin
Wisconsin
county
Greene Co
Jackson Co
Jefferson Co
Yellowstone Co
Douglas Co
MerrimackCo
Erie Co
Monroe Co
Suffolk Co
New Hanover Co
Clark Co
Jefferson Co
Lake Co
Summit Co
Kay Co
Muskogee Co
Tulsa Co
Allegheny Co
Blair Co
Northampton Co
Warren Co
Lexington Co
Blount Co
Bradley Co
Montgomery Co
Shelby Co
Sullivan Co
Harris Co
Jefferson Co
Hancock Co
Brown Co
Oneida Co
SOppb
$16.0
$59.0
$310.0
$12.0
$17.0
$19.0
$38.0
$7.5
$50.0
$19.0
$19.0
$18.0
$110.0
$76.0
$28.0
$78.0
$24.0
$160.0
$38.0
$61.0
$29.0
$22.0
$36.0
$39.0
$38.0
$16.0
$110.0
$66.0
$61.0
$30.0
$40.0
$22.0
75ppb
$26.0
$280.0
$17.0
$14.0
$21.0
$47.0
$19.0
$51.0
$28.0
$29.0
$2.9
$38.0
$47.0
$28.0
$22.0
100 ppb
$280.0
$3.0
$25.0
$29.0
$38.0
$22.0
All estimates rounded to two significant figures. As such, totals will not sum down columns.
b All estimates provided reflect the engineering cost of the identified control strategy analysis, incremental to a
2020 baseline.
c Total annualized costs were calculated using a 7% discount rate.
dThese values represent partial attainment costs for the identified control strategy analysis. There were locations
not able to attain the alternative standard being analyzed with identified controls only.
6-6
-------
6.1.3 Extrapolated Costs
Prior to presenting the methodology for estimating costs for unspecified
emission reductions, it is important to provide information from EPA's Science Advisory
Board (SAB) Council Advisory on the issue of estimating costs of unidentified control
measures.4
812 Council Advisory, Direct Cost Report, Unidentified Measures
(charge question 2.a):
"The Project Team has been unable to identify measures that yield
sufficient emission reductions to comply with the National Ambient
Air Quality Standards (NAAQS) and relies on unidentified pollution
control measures to make up the difference. Emission reductions
attributed to unidentified measures appear to account for a large
share of emission reductions required for a few large metropolitan
areas but a relatively small share of emission reductions in other
locations and nationwide.
"The Council agrees with the Project Team that there is little
credibility and hence limited value to assigning costs to these
unidentified measures. It suggests taking great care in reporting
cost estimates in cases where unidentified measures account for a
significant share of emission reductions. At a minimum, the
components of the total cost associated with identified and
unidentified measures should be clearly distinguished. In some
cases, it may be preferable to not quantify the costs of
unidentified measures and to simply report the quantity and share
of emissions reductions attributed to these measures.
"When assigning costs to unidentified measures, the Council
suggests that a simple, transparent method that is sensitive to
the degree of uncertainty about these costs is best. Of the three
approaches outlined, assuming a fixed cost/ton appears to be the
4 U.S. Environmental Protection Agency, Advisory Council on Clean Air Compliance Analysis (COUNCIL),
Council Advisory on OAR's Direct Cost Report and Uncertainty Analysis Plan, Washington, DC. June 8,
2007.
6-7
-------
simplest and most straightforward. Uncertainty might be
represented using alternative fixed costs per ton of emissions
avoided."
EPA has considered this advice and the requirements of E.G. 12866 and OMB
circular A-4, which provides guidance on the estimation of benefits and costs of
regulations.
As indicated above the identified control costs do not result in attainment of the
selected or alternative standards in four areas. In these areas, unspecified emission
reductions needed beyond identified controls will likely be necessary to reach
attainment.
Taking into consideration the above SAB advice, we estimated the costs of
unspecified future emission reductions using a fixed (annualized) cost per ton approach.
In previous analyses we have estimated the extrapolated costs using other marginal cost
based approaches in addition to the fixed cost per ton approach. We examine the data
available for each analysis and determine on a case by case basis the appropriate
extrapolation technique. Due to the limited number of control measures applied in this
analysis across all sectors, we concluded that it would not be credible to establish a
marginal cost-based approach or a representative value for the costs of further S02
emission reductions. We also recognize that the emissions from EGUs are the largest
for these areas. In addition, there is also limited information on S02 controls applied to
non-EGUs beyond the scope of this analysis, especially for small sources. For these
reasons, we have relied upon a simple fixed cost approach utilized for that analysis to
represent the fixed cost of unspecified emission reductions for this analysis. The
primary estimate presented is $15,000 (2006$), with sensitivities of $10,000/ton and
$20,000/ton. Use of $15,000/ton as a fixed cost estimate is commensurate with the
cost of nonEGU S02 control measures as applied in the PM2.5 RIA three years ago. This
fixed costs is also much higher than reported costs for S02 controls such as wet FGD
scrubbers for industrial boilers are reported to be up to at least $5,200/ton (2006$).5
Also, this estimate is considerably greater than the current and futures prices for S02
emissions allowances traded for compliance with the CAIR program.6 Finally, as
5 Applicability and Feasibility of NOx, SO2, and PM Emissions Control Technologies for Industrial,
Commercial, and Institutional (ICI) Boilers. NESCAUM, November 2008. Available on the Internet at
http://www.nescaum.org/documents/ici-boilers-20081118-final.pdf/.
6 The Evolving SO2 Allowance Market: Title IV, CAIR, and Beyond. Palmer, Karen, Resources for the
Future and Evans, David, US EPA/OPEI, July 13, 2009. Available on the Internet at
http://www.rff.org/Publications/WPC/Pages/090713-Evolving-SO2-Allowance-Market.aspx.
6-8
-------
mentioned above, the use of a fixed cost per ton of $15,000/ton is consistent with what
an advisory committee to the Section 812 second prospective analysis on the Clean Air
Act Amendments suggested in June 2007 for estimating the costs of reductions from
unidentified controls.
The estimation of costs for emission reductions needed to reach attainment
many years in the future is inherently difficult. We expect that additional control
measures that we were not able to identify may be developed by 2020. As described
later in this chapter, our experience with Clean Air Act implementation shows that
technological advances and development of innovative strategies can make possible
cost effective emissions reductions that are unforeseen today, and can reduce costs of
some emerging technologies over time. But we cannot precisely predict the amount of
technology advance in the future. The relationship of the cost of additional future
controls to the cost of control options available today is not at all clear. Available,
currently known control measures increase in costs per ton beyond the range of what
has ever been implemented and because they are not currently required can not serve
as an accurate representation of expected costs of implementation. Such measures
would still not provide the needed additional control for full attainment in the analysis
year 2020. History has shown that when faced with potentially costly controls
requirements, firms could adapt by changing their production process or innovate to
develop more cost effective ways of meeting control requirements. We recognize that a
single fixed cost of control of $15,000 per ton of emissions reductions does not account
for the significant emissions cuts that are necessary in some areas and so its use
provides an estimate that is likely to differ from actual future costs. Yet, the limited
emission controls dataset applied for the identified control strategy analysis significantly
limits our ability to estimate full attainment costs using more sophisticated methods.
In the economics literature there are a variety of theoretical ways to estimate
the cost of more stringent emissions reductions than can be achieved by known
technologies. One method would be to estimate the cost of reducing all remaining tons
by simply extrapolating the cost curve using data on cost and effectiveness of all known
controls. This method can imply the last ton of reductions costs an amount which is
thousands of times higher than the fixed cost presumed above (i.e., $15,000 per ton).
This result is highly unlikely given the uncertainty surrounding the assumptions implicit
in this estimate (e.g. projecting 10 years into the future, not including factors for
technological innovation and improvements, not including societal and economy wide
changes from dealing with climate change). Such a result does not necessarily mean
that such costs will be incurred, because of uncertainties about future control
6-9
-------
technology, economic activity and the possibility of deferment of full attainment dates.
Another variant on this approach is to develop a method which simulates technological
change by causing shifts in the cost curve over time to reflect that innovation can
reduce costs of control.
In addition, it is theoretically possible to consider the cost of a geographic area
changing to a different type of economic structure over time (e.g. moving from a one
type of manufacturing to another or from manufacturing to a more service oriented
economy) as another way to predict the cost of meeting a tighter standard. This would
be a challenging, data intensive exercise that would be very area specific. Nationwide
estimates would have to be built from an area by area basis. In some areas, mobile
sources may be a significant source of emissions; some areas are experimenting with
congestion pricing as a means of restructuring how people and goods travel to reduce
emissions.
In the absence of more robust methods for estimating these costs, EPA is
following the SAB advice to keep the approach simple and transparent. If commentors
have different assumptions about the cost of attainment, it is easy for them to calculate
the cost of attaining a tighter standard using the fixed cost formula. EPA is going to
continue to work on most robust methods of developing these estimates. EPA will
continue to improve methods of estimating the costs of full attainment when health-
based standards require emissions cuts greater than can be achieved by all known
engineering controls. Over the course of the next several months EPA, in partnership
with OMB and interested federal agencies will be investigating different ways of
estimating these extrapolated full attainment costs, including consideration of ways of
incorporating technological change and other factors. In addition, EPA is looking into
developing approaches to characterize different future states of the world. These
scenarios (similar to the goal of the IPCC scenarios for the outcome of climate change,
for example) would allow us to consider a range of possibilities. Many criteria pollutant
emissions result from combustion processes used to make energy, transport goods and
people and other industrial operations. Our alternative futures could represent
different types of power generation that could become more prevalent under different
circumstances. For example, in one scenario solar or wind power would prevail leading
to reductions in the burning of coal for power generation. In contrast, in another
scenario coal use remains consistent with current usage but is subject to more
emissions reductions. Another could presume significant inroads for electric vehicles.
EPA will be considering this approach as another method for projecting a range of
possibilities for the cost of attaining a tighter standard. This research will include a
6-10
-------
review of how best to characterize the likely adoption by 2020 (or similar target years)
of new technologies (e.g., solar, wind and others unrelated to fossil fuel combustion, as
well as more fuel-efficient vehicles), that are expected to have the ancillary benefit of
facilitating compliance with new standards for criteria air pollutants. It will also include
consideration of control measures that depend on behavioral change (such as
congestion pricing) rather than simply the adoption of engineering controls.
The approach outlined above represents a significant amount of theoretical and
applied analysis and the development of new methodologies for doing this analysis.
Data supporting our cost approach is in the S02 NAAQS RIA docket and we welcome
ideas from the public on suggestions for analytical methods to estimate these future
costs and plans to hopefully utilize portions of it in the proposed PM2.5 NAAQS RIA to
be released with the rest of the material accompanying the standard.
Table 6.3 presents the extrapolated costs for each alternative standard analyzed.
See Chapter 4 for a complete discussion of the air quality projections for these counties.
Table 6.3: Extrapolated Costs Estimated for the Alternative Standards
(Millions of 2006$) a'b
50 ppb 75 ppb 100 ppb
Total Extrapolated Costs
($10,000/ton): $ 1'2°° ^330 $180
Total Extrapolated Costs
($15,000/ton): ***» ^ $ 26°
Total Extrapolated Costs
($20,000/ton): $ 2'4°° ^ $ 35°
aAII estimates rounded to two significant figures. As such, totals will not sum down columns.
b Estimates of extrapolated costs are assumed using a 7% discount rate. Given the fixed cost per ton
approach used here, 3% discount rate estimates could not be calculated.
6-11
-------
6.1.4 Monitoring Costs
The final amendments would revise the technical requirements for S02 monitoring sites;
require the siting and operation of additional S02 ambient air monitors, and the reporting of
the collected ambient monitoring data to EPA's Air Quality System (AQS). We have estimated
the burden based on the monitoring requirements of this rule. Details of the burden estimate
are contained in the information collection request (ICR) accompanying the final rule.7 The ICR
estimates annualized costs of a new monitoring network at approximately $15 million per year
(2006 dollars).
6.1.5 Summary of Cost Estimates
Table 6.4 provides a summary of total costs to achieve the alternative standards in the
year 2020, and this summary includes the sensitivity estimates. As mentioned previously, we
use $15,000/ton as our primary estimate of the extrapolated costs on a per ton reduction basis,
and $10,000/ton and $20,000/ton are used as sensitivities. Using that estimate, we find that
the total annualized costs for the 75 ppb final standard in 2020 are $1.0 billion (2006$) using
seven percent as the discount rate and applying the primary estimate of the extrapolated costs,
and the costs for the other alternative standards range from $0.5 billion to $2.6 billion (2006$).
The portion of these costs accounted for by identified controls ranges from 59 percent for the
50 ppb standard to 64 percent for the 100 ppb standard. Hence, the portion of these costs
accounted for by extrapolated controls ranges from 41 percent for the 50 ppb standard to 36
percent for the 100 ppb standard.
Finally, Table 6.5 present the annual cost/ton for the identified controls by sector as
applied for the alternative standards in 2020. For each alternative standard, the annual
cost/ton for reductions from the non-EGU sector is the most expensive. For the 75 ppb final
standard, reductions from non-EGUs occur at $2,400/ton while the annual cost/ton for ECU
sector is $2,700/ton. All of these estimates are for reductions in 2020 in 2006 dollars and using
a seven percent discount rate.
The significant difference between the costs of identified controls alone and the cost of
achieving attainment (i.e. including both identified controls and emission reductions beyond
identified controls) in this and other areas reflects the limited information available to EPA on
7 ICR 2358.01, May 2009.
6-12
-------
the control measures that sources may implement. Although AirControlNET contains
information on a large number of different point source controls, we would expect that State
and local air quality managers would have access to additional information on the controls
available to the most significant sources.
Table 6.4: Total Annual Costs for Alternative Standards (Millions of 2006$)a'b
Identified Control Costs
Monitoring
Extrapolated Costs
Total Costs
; Costs
Fixed Cost
($10,000/ton)
dFixed Cost
($15,000/ton)
Fixed Cost
($20,000/ton)
Fixed Cost
($10,000/ton)
dFixed Cost
($15,000/ton)
Fixed Cost
($20,000/ton)
50 ppb
$ 2,600
$2.1
$ 1,200
$ 1,800
$ 2,400
$ 3,800
$ 4,400
$ 5,000
75 ppb
$ 960
$2.1
$330
$500
$670
$ 1,300
$ 1,500
$ 1,600
100 ppb
$ 470
$2.1
$180
$260
$350
$650
$730
$820
All estimates rounded to two significant figures. As such, totals will not sum down columns.
b All estimates provided reflect the engineering cost of the identified control strategy analysis, incremental to a
2020.
c Values reflect a 7% discount rate.
d Our primary estimate of extrapolated costs is, as mentioned earlier in this RIA, based on a fixed annual cost of
$15,000/ton. This estimate of extrapolated costs is incorporated into our estimate of total costs for the alternative
standards.
6-13
-------
Table 6.5: Annual Cost per Ton of Identified Controls applied for the Alternative Standards by
Emissions Sector (2006$)a'b
Emissions Sector 50 ppb 75 ppb 100 ppb
NonEGU $ 2,400 $ 2,700 $ 2,800
Area $ 2,500 $ 2,200 $2,100
ECU $ 2,700 $ 2,700 $ 2,800
aAII estimates rounded to two significant figures. As such, totals will not sum down columns.
b All estimates provided reflect the engineering cost of the identified control strategy analysis, incremental to a
2020 baseline.
6.1.6 Technology Innovation and Regulatory Cost Estimates
There are many examples in which technological innovation and "learning by doing"
have made it possible to achieve greater emissions reductions than had been feasible earlier, or
have reduced the costs of emission control in relation to original estimates. Studies8 have
suggested that costs of some EPA programs have been less than originally estimated due in part
to inadequate inability to predict and account for future technological innovation in regulatory
impact analyses.
Constantly increasing marginal costs are likely to induce the type of innovation that
would result in lower costs than estimated early in this chapter. Breakthrough technologies in
control equipment could by 2020 result in a rightward shift in the marginal cost curve for such
equipment (Figure 6.1)9 as well as perhaps a decrease in its slope, reducing marginal costs per
unit of abatement, and thus deviate from the assumption of a static marginal cost curve. In
addition, elevated abatement costs may result in significant increases in the cost of production
and would likely induce production efficiencies, in particular those related to energy inputs,
which would lower emissions from the production side.
Harrington et al. (2000) and previous studies cited by Harrington.
Harrington, W., R.D. Morgenstern, and P. Nelson. 2000. "On the Accuracy of Regulatory Cost Estimates." Journal of
Policy Analysis and Management 19(2):297-322.
9 Figure 6.1 shows a linear marginal abatement cost curve. It is possible that the shape of the marginal abatement
cost curve is non-linear.
6-14
-------
Figure 6.1: Technological Innovation Reflected by Marginal Cost Shift
Cost/
Ton
MC0
?o //
Slope = / Slope =
MC,
Induced Technology Shift
Cumulative SO2 Reductions
6.1.6.1 Examples of Technological Advances in Pollution Control
There are numerous examples of low-emission technologies developed and/or
commercialized over the past 15 or 20 years, such as:
Selective catalytic reduction (SCR) and ultra-low NOx burners for NOx emissions
Scrubbers which achieve 95% and even greater S02 control on boilers
Sophisticated new valve seals and leak detection equipment for refineries and
chemical plans
Low or zero VOC paints, consumer products and cleaning processes
Chlorofluorocarbon (CFC) free air conditioners, refrigerators, and solvents
Water and powder-based coatings to replace petroleum-based formulations
Vehicles far cleaner than believed possible in the late 1980s due to
improvements in evaporative controls, catalyst design and fuel control systems
for light-duty vehicles; and treatment devices and retrofit technologies for
heavy-duty engines
6-15
-------
• Idle-reduction technologies for engines, including truck stop electrification
efforts
• Market penetration of gas-electric hybrid vehicles, and clean fuels
• The development of retrofit technology to reduce emissions from in-use vehicles
and non-road equipment
These technologies were not commercially available two decades ago, and some were
not even in existence. Yet today, all of these technologies are on the market, and many are
widely employed. Several are key components of major pollution control programs and most of
the examples are discussed further below.
What is known as "learning by doing" or "learning curve impacts", which is a concept
distinct from technological innovation, has also made it possible to achieve greater emissions
reductions than had been feasible earlier, or have reduced the costs of emission control in
relation to original estimates. Learning curve impacts can be defined generally as the extent to
which variable costs (of production and/or pollution control) decline as firms gain experience
with a specific technology. Such impacts have been identified to occur in a number of studies
conducted for various production processes. Impacts such as these would manifest themselves
as a lowering of expected costs for operation of technologies in the future below what they
may have been.
The magnitude of learning curve impacts on pollution control costs has been estimated
for a variety of sectors as part of the cost analyses done for the Draft Direct Cost Report for the
second EPA Section 812 Prospective Analysis of the Clean Air Act Amendments of 1990.10 In
that report, learning curve adjustments were included for those sectors and technologies for
which learning curve data was available. A typical learning curve adjustment example is to
reduce either capital or O&M costs by a certain percentage given a doubling of output from
that sector or for that technology. In other words, capital or O&M costs will be reduced by
some percentage for every doubling of output for the given sector or technology.
T.P. Wright, in 1936, was the first to characterize the relationship between increased
productivity and cumulative production. He analyzed man-hours required to assemble
successive airplane bodies. He suggested the relationship is a log linear function, since he
observed a constant linear reduction in man-hours every time the total number of airplanes
assembled was doubled. The relationship he devised between number assembled and assembly
10 E.H. Pechan and Associates and Industrial Economics, Direct Cost Estimates for the Clean Air Act Second Section
812 Prospective Analysis: Draft Report, prepared for U.S. EPA, Office of Air and Radiation, February 2007.
Available at http://www.epa.gov/oar/sect812/mar07/direct_cost_draft.pdf.
6-16
-------
time is called Wright's Equation (Gumerman and Marnay, 2004)11. This equation, shown below,
has been shown to be widely applicable in manufacturing:
Wright's Equation: CN= C0 * Nb,
Where:
N = cumulative production
CN = cost to produce Nth unit of capacity
C0 = cost to produce the first unit
B = learning parameter = In (l-LR)/ln(2), where
LR = learning by doing rate, or cost reduction per doubling of capacity or
output.
The percentage adjustments to costs can range from 5 to 20 percent, depending on the
sector and technology. Learning curve adjustments were prepared in a memo by lEc supplied to
US EPA and applied for the mobile source sector (both onroad and nonroad) and for application
of various ECU control technologies within the Draft Direct Cost Report.12 Advice received from
the SAB Advisory Council on Clean Air Compliance Analysis in June 2007 indicated an interest in
expanding the treatment of learning curves to those portions of the cost analysis for which no
learning curve impact data are currently available. Examples of these sectors are non-EGU point
sources and area sources. The memo by lEc outlined various approaches by which learning
curve impacts can be addressed for those sectors. The recommended learning curve impact
adjustment for virtually every sector considered in the Draft Direct Cost Report is a 10%
reduction in O&M costs for two doubling of cumulative output, with proxies such as cumulative
fuel sales or cumulative emission reductions being used when output data was unavailable.
For this RIA, we do not have the necessary data for cumulative output, fuel sales, or
emission reductions for all sectors included in our analysis in order to properly generate control
costs that reflect learning curve impacts. Clearly, the effect of including these impacts would be
to lower our estimates of costs for our control strategies in 2020, but we are not able to include
such an analysis in this RIA.
11 Gumerman, Etan and Marnay, Chris. Learning and Cost Reductions for Generating Technologies in the National
Energy Modeling System (NEMS), Ernest Orlando Lawrence Berkeley National Laboratory, University of California
at Berkeley, Berkeley, CA. January 2004, LBNL-52559.
12 Industrial Economics, Inc. Proposed Approach for Expanding the Treatment of Learning Curve Impacts for the
Second Section 812 Prospective Analysis: Memorandum, prepared for U.S. EPA, Office of Air and Radiation, August
13, 2007.
6-17
-------
6.1.6.2 Influence on Regulatory Cost Estimates
Studies indicate that it is not uncommon for pre-regulatory cost estimates to be higher
than later estimates, in part because of inability to predict technological advances. Over longer
time horizons the opportunity for technical advances is greater.
• Multi-rule study: Harrington et al. of Resources for the Future13 conducted an
analysis of the predicted and actual costs of 28 federal and state rules, including 21 issued by
EPA and the Occupational Safety and Health Administration (OSHA), and found a tendency for
predicted costs to overstate actual implementation costs. Costs were considered accurate if
they fell within the analysis error bounds or if they fall within 25 percent (greater or less than)
the predicted amount. They found that predicted total costs were overestimated for 14 of the
28 rules, while total costs were underestimated for only three rules. Differences can result
because of quantity differences (e.g., overestimate of pollution reductions) or differences in
per-unit costs (e.g., cost per unit of pollution reduction). Per-unit costs of regulations were
overestimated in 14 cases, while they were underestimated in six cases. In the case of EPA
rules, the agency overestimated per-unit costs for five regulations, underestimated them for
four regulations (three of these were relatively small pesticide rules), and accurately estimated
them for four. Based on examination of eight economic incentive rules, "for those rules that
employed economic incentive mechanisms, overestimation of per-unit costs seems to be the
norm," the study said. It is worth noting here, that the controls applied for this NAAQS do not
use an economic incentive mechanism. In addition, Harrington also states that overestimation
of total costs can be due to error in the quantity of emission reductions achieved, which would
also cause the benefits to be overestimated.
Based on the case study results and existing literature, the authors identified
technological innovation as one of five explanations of why predicted and actual regulatory cost
estimates differ: "Most regulatory cost estimates ignore the possibility of technological
innovation ... Technical change is, after all, notoriously difficult to forecast... In numerous case
studies actual compliance costs are lower than predicted because of unanticipated use of new
technology."
It should be noted that many (though not all) of the EPA rules examined by Harrington
had compliance dates of several years, which allowed a limited period for technical innovation.
13 Harrington, W., R.D. Morgenstern, and P. Nelson. 2000. "On the Accuracy of Regulatory Cost Estimates." Journal
of Policy Analysis and Management 19(2):297-322.
6-18
-------
• Acid Rain 502 Trading Program: Recent cost estimates of the Acid Rain S02
trading program by Resources for the Future (RFF) and MIT have been as much as 83 percent
lower than originally projected by EPA.14 As noted in the RIA for the Clean Air Interstate Rule,
the ex ante numbers in 1989 were an overestimate in part because of the limitation of
economic modeling to predict technological improvement of pollution controls and other
compliance options such as fuel switching. The fuel switching from high-sulfur to low-sulfur coal
was spurred by a reduction in rail transportation costs due to deregulation of rail rates during
the 1990's Harrington et al. report that scrubbing turned out to be more efficient (95% removal
vs. 80-85% removal) and more reliable (95% vs. 85% reliability) than expected, and that
unanticipated opportunities arose to blend low and high sulfur coal in older boilers up to a
40/60 mixture, compared with the 5/95 mixture originally estimated.
Phase 2 Cost Estimates
Ex ante estimates $2.7 to $6.2 billion3
Ex post estimates $1.0 to $1.4 billion
3 2010 Phase II cost estimate in 1995$.
• EPA Fuel Control Rules: A 2002 study by EPA's Office of Transportation and Air
Quality15 examined EPA vehicle and fuels rules and found a general pattern that "all ex ante
estimates tended to exceed actual price impacts, with the EPA estimates exceeding actual
prices by the smallest amount." The paper notes that cost is not the same as price, but suggests
that a comparison nonetheless can be instructive.16 An example focusing on fuel rules is
provided in Table 6.6:
14 Carlson, Curtis, Dallas R. Burtraw, Maureen, Cropper, and Karen L Palmer. 2000. "Sulfur Dioxide Control by
Electric Utilities: What Are the Gains from Trade?" Journal of Political Economy 108(#6):1292-1326.
Ellerman, Denny. January 2003. Ex Post Evaluation of Tradable Permits: The U.S. SO2 Cap-and-Trade Program.
Massachusetts Institute of Technology Center for Energy and Environmental Policy Research.
15 Anderson, J.F., and Sherwood, T., 2002. "Comparison of EPA and Other Estimates of Mobile Source Rule Costs to
Actual Price Changes," Office of Transportation and Air Quality, U.S. Environmental Protection Agency. Technical
Paper published by the Society of Automotive Engineers. SAE 2002-01-1980.
16 The paper notes: "Cost is not the same as price. This simple statement reflects the fact that a lot happens
between a producer's determination of manufacturing cost and its decisions about what the market will bear in
terms of price change."
6-19
-------
Table 6.6: Comparison of Inflation-Adjusted Estimated Costs and Actual Price Changes for EPA
Fuel Control Rules3
Inflation-adjusted Cost Estimates (c/gal)
Gasoline
Phase 2 RVP Control (7.8 RVP-
Summer) (1995$)
Reformulated Gasoline Phase 1
(1997$)
Reformulated Gasoline Phase 2
(Summer) (2000$)
30 ppm sulfur gasoline (Tier 2)
Diesel
500 ppm sulfur highway diesel fuel
(1997$)
15 ppm sulfur highway diesel fuel
EPA DOE
1.1 1.8
3.1-5.1 3.4-4.1 8.
4.6-6.8 7.6-10.2 10
1.7-1.9 2.9-3.4
1.9-2.4 3.3
4.5 4.2-6.0
API
2-14.0
.8-19.4
2.6
(NPRA)
6.2
Other
0.5
7.4 (CRA)
12
5.7 (NPRA),
3.1 (AIAM)
2.2
4.2-6.1
(NPRA)
Actual Price
Changes (c/gal)
2.2
7.2(5.1, when
corrected to Syr
MTBE price)
N/A
N/A
Anderson, J.F., and Sherwood, T., 2002. "Comparison of EPA and Other Estimates of Mobile Source Rule Costs to
Actual Price Changes," Office of Transportation and Air Quality, U.S. Environmental Protection Agency. Technical
Paper published by the Society of Automotive Engineers. SAE 2002-01-1980.
• Chlorofluorocarbon (CFC) Phase-Out: EPA used a combination of regulatory,
market based (i.e., a cap-and-trade system among manufacturers), and voluntary approaches
to phase out the most harmful ozone depleting substances. This was done more efficiently than
either EPA or industry originally anticipated. The phaseout for Class I substances was
implemented 4-6 years faster, included 13 more chemicals, and cost 30 percent less than was
predicted at the time the 1990 Clean Air Act Amendments were enacted.17
The Harrington study states, "When the original cost analysis was performed for the CFC
phase-out it was not anticipated that the hydrofluorocarbon HFC-134a could be substituted for
CFC-12 in refrigeration. However, as Hammit18 notes, 'since 1991 most new U.S. automobile air
conditioners have contained HFC-134a (a compound for which no commercial production
technology was available in 1986) instead of CFC-12" (p.13). He cites a similar story for HCFRC-
141b and 142b, which are currently substituting for CFC-11 in important foam-blowing
applications."
Holmstead, Jeffrey, 2002. "Testimony of Jeffrey Holmstead, Assistant Administrator, Office of Air and Radiation,
U.S. Environmental Protection Agency, Before the Subcommittee on Energy and air Quality of the committee on
Energy and Commerce, U.S. House of Representatives, May 1, 2002, p. 10.
18 Hammit, J.K. (2000). "Are the costs of proposed environmental regulations overestimated? Evidence from the
CFC phaseout." Environmental and Resource Economics, 16(#3): 281-302.
6-20
-------
Additional examples of decreasing costs of emissions controls include: SCR catalyst costs
decreasing from $llk-$14k/m3 in 1998 to $3.5k-$5k/m3 in 2004, and improved low NOx
burners reduced emissions by 50% from 1993-2003 while the associated capital cost dropped
from $25-$38/kW to $15/kW19. Also, FGD scrubber capital costs have been estimated to have
decreased by more than 50 percent from 1976 to 2005, and the operating and maintenance
(O&M) costs decreased by more than 50% from 1982 to 2005. Many process improvements
contributed to lowering the capital costs, especially improved understanding and control of
process chemistry, improved materials of construction, simplified absorber designs, and other
factors that improved reliability.20
We cannot estimate the precise interplay between EPA regulation and technology
improvement, but it is clear that a priori cost estimation often results in overestimation of costs
because changes in technology (whatever the cause) make less costly control possible.
6.2 Economic Impacts
The assessment of economic impacts in Table 6.7 was conducted based on those source
categories which are assumed in this analysis to become controlled. The impacts presented
here are a comparison of the control costs to the revenues for industries affected by control
strategies applied for the 75 ppb final standard. Control costs are allocated to specific source
categories by North American Industry Classification System (NAICS) code.
19ICF Consulting. October 2005. The Clean Air Act Amendment: Spurring Innovation and Growth While Cleaning
the Air. Washington, DC. Available at http://www.icfi.com/Markets/Environment/doc_files/caaa-success.pdf.
20
Yeh, Sonia and Rubin, Edward. February 2007. "Incorporating Technological Learning in the Coal Utility
Environmental Cost (CUECost) Model: Estimating the Future Cost Trends of SO2, NOx, and Mercury Control
Technologies." Prepared for ARCADIS Geraghty and Miller, Research Triangle Park, NC 27711. Available at
http://steps.ucdavis.edu/People/slyeh/syeh-resources/Drft%20Fnl%20Rpt%20Lrng%20for%20CUECost_v3.pdf.
6-21
-------
Table 6.7: Identified Cost/Revenue Ratios by Affected Industry for Illustrative Control
Strategy for the Final SO2 Standard (75 ppb) in 2020 (Millions of 2006$)a'b'c
NAICS
Code
2211
311
312
322
324
325
326
327
331
332
333
336
611
Industry Description
Electric Power Generation,
Transmission and Distribution
Food Manufacturing
Beverage and Tobacco Product
Manufacturing
Paper Manufacturing
Petroleum and Coal Products
Manufacturing
Chemical Manufacturing
Plastics and Rubber Products
Manufacturing
Nonmetallic Mineral Product
Manufacturing
Primary Metal Manufacturing
Fabricated metal product
manufacturing
Machinery manufacturing
Transportation equipment
manufacturing
Educational services
3% Discount
Rated
699
55
1.3
$143
$245
$12.8
6.2
266
$
0.4
3.0
2.9
137
7% Discount
Rate
699
19.9
7.0
$31.2
$39.5
$12.8
6.2
43.5
$43.6
0.4
3.0
0.8
51.9
Industry
Revenue in
2007e
440,000
589,000
128,000
$170,000
$590,000
$720,000
211,000
128,000
$250,000
344,000
19,700
737,000
47,000
Cost/Revenue
Ratio
0.16%
<0.01%
<0.01%
< 0.01%
< 0.01%
< 0.01%
<0.01%
<0.01%
< 0.01%
< 0.01%
< 0.01%
< 0.01%
0.13%
All estimates rounded to two significant figures. As such, totals will not sum down columns.
b All estimates provided reflect the engineering cost of the identified control strategy analysis, incremental to a
2020 baseline.
c NAICS codes were unavailable for area source controls. These controls account for less than 2% of the total
identified control strategy costs.
d Total annualized costs were calculated using a 3% discount rate for controls which had a capital component and
where equipment life values were available. For the identified control strategy, data for calculating annualized
costs at a 3% discount was available for point sources. Therefore, the total annualized identified control cost value
presented in this referenced cell is an aggregation of engineering costs at 3% and 7% discount rate.
Source: U.S. Census Bureau 2007 Economic Census. Industry-level data on revenues can be found at
http://factfinder.census.gov/servlet/IBQTable? bm=y&-fds name=EC0700Al&- skip=0&-ds name=EC0700Al&-
lang=en.
f No data on budget or revenues for this NAICS code is included in the 2007 Economic Census.
6.3 Energy Impacts
This section summarizes the energy consumption impacts associated with control strategies
applied for the final S02 NAAQS of 75 ppb. The S02 NAAQS revisions do not constitute a
"significant energy action" as defined in Executive Order 13211; this information merely
represents impacts of the illustrative control strategy applied in the RIA. The rule does not
prescribe specific control strategies by which these ambient standards will be met. Such
6-22
-------
strategies will be developed by States on a case-by-case basis, and EPA cannot predict whether
the control options selected by States will include regulations on energy suppliers, distributors,
or users. Thus, EPA concludes that this rule is not likely to have any adverse energy effects as
defined in Executive Order 13211.
For this RIA, implementation of the control measures needed for attainment with the
alternative standards will likely lead to increased energy consumption among S02 emitting
facilities. In addition, because the energy consumption and impacts on various energy markets
associated with emission reductions beyond identified controls is uncertain, we only consider
the energy impacts associated with identified controls.
With respect to energy supply and prices, the analysis in Table 6.7 suggests that at the
electric power industry level, the annualized costs associated with the illustrative control
strategy for the final standard (75 ppb) represent only about 0.16 percent of its revenues in
2020. In addition, for the other industries affected under the 75 ppb standard, no other
industry has annualized costs of more than 0.13 percent of its revenues. As a result we can
conclude that impacts to supply and electricity price are small
6.4 Limitations and Uncertainties Associated with Engineering Cost Estimates
• EPA bases its estimates of emissions control costs on the best available information
from engineering studies of air pollution controls and has developed a reliable
modeling framework for analyzing the cost, emissions changes, and other impacts of
regulatory controls. The annualized cost estimates of the private compliance costs
are meant to show the increase in production (engineering) costs to the various
affected sectors in our control strategy analyses. To estimate these annualized costs,
EPA uses conventional and widely-accepted approaches that are commonplace for
estimating engineering costs in annual terms. However, our engineering cost analysis
is subject to uncertainties and limitations.
• One of these limitations is that we do not have sufficient information for all of our
known control measures to calculate cost estimates that vary with an interest rate.
We are able to calculate annualized costs at an interest rate other than 7% (e.g., 3%
interest rate) where there is sufficient information—available capital cost data, and
equipment life—to annualize the costs for individual control measures. For the vast
majority of nonEGU point source control measures, we do have sufficient capital cost
and equipment life data for individual control measures to prepare annualized capital
costs using the standard capital recovery factor. Hence, we are able to provide
6-23
-------
annualized cost estimates at different interest rates for the point source control
measures.
• For area source control measures, the engineering cost information is available only
in annualized cost/ton terms. We have extremely limited capital cost and equipment
life data for area source control measures. We know that these annualized cost/ton
estimates reflect an interest rate of 7% because these estimates are typically
products of technical memos and reports prepared as part of rules issued by EPA over
the last 10 years or so, and the costs estimated in these reports have followed the
policy provided in OMB Circular A-4 that recommends the use of 7% as the interest
rate for annualizing regulatory costs. Capital cost information for these area source
controls, however, is often limited since these measures are often not the traditional
add-on controls where the capital cost is well known and convenient to estimate. The
limited availability of useful capital cost data for such control measures has led to our
use of annualized cost/ton estimates to represent the engineering costs of these
controls in our cost tools and hence in this RIA.
• There are some unquantified costs that are not adequately captured in this
illustrative analysis. These costs include the costs of federal and State administration
of control programs, which we believe are less than the alternative of States
developing approvable SIPs, securing EPA approval of those SIPs, and Federal/State
enforcement. The analysis also did not consider transactional costs and/or effects on
labor supply in the illustrative analysis.
6-24
-------
Chapter 7: Estimates of Costs and Benefits
Synopsis
As discussed above, this RIA analyzes alternative primary standards of 50 parts per
billion (ppb), 75 ppb, and 100 ppb. Our assessment of the lower bound S02 target NAAQS
includes several key elements, including specification of baseline S02 emissions and
concentrations; development of illustrative control strategies to attain the standard in 2020;
and analyses of the control costs and health benefits of reaching the various alternative
standards. We also note that because it was not possible, in this analysis, to bring all areas into
attainment with the selected standard of 75 ppb in all areas using only identified controls, EPA
conducted a second step in the analysis, and estimated the cost of unspecified emission
reductions needed to attain the alternative primary NAAQS.
This analysis does not estimate the projected attainment status of areas of the country
other than those counties currently served by one of the approximately 488 monitors in the
current network. It is important to note that the rule would require a monitoring network
wholly comprised of monitors sited at locations of expected maximum hourly concentrations.
Only about one third of the existing S02 network may be source-oriented and/or in the
locations of maximum concentration required by the proposed rule because the current
network is focused on population areas and community-wide ambient levels of S02. Actual
monitored levels using the new monitoring network may be higher than levels measured using
the existing network. We recognize that once a network of monitors located at maximum-
concentration is put in place, more areas could find themselves exceeding the new S02 NAAQS.
However for this RIA analysis, we lack sufficient data to predict which counties might exceed
the new NAAQS after implementation of the new monitoring network. Therefore we lack a
credible analytic path to estimating costs and benefits for such a future scenario.
7.1 Benefits and Costs
We estimated the benefits and costs for four alternative S02 NAAQS levels: 50 ppb, 75
ppb, and 100 ppb (99th percentile). These costs and benefits are associated with an
incremental difference in ambient concentrations between a baseline scenario and a pollution
control strategy. As indicated above and in Chapter 4, several areas of the country may not be
able to attain some alternative standard using known pollution control methods. Because
some areas require substantial emission reductions from unknown sources to attain the various
standards, the results are very sensitive to assuming full attainment. For this reason, we
provide the full attainment and the partial attainment results for both benefits and costs.
7-1
-------
Costs
Our analysis of the costs associated with the range of alternative NAAQS focuses on S02
emission controls for electric generating units (ECU) and nonEGU stationary and area sources.
ECU, nonEGU and area source controls largely include measures from the Control Strategy Tool
(CoST), and the AirControlNET control technology database. For these sources, we estimated
costs based on the cost equations included in AirControlNET.
As indicated in the above discussion on illustrative control strategies, implementation of
the S02 control measures identified from AirControlNET and other sources does not result in
attainment with the selected NAAQS in several areas. In these areas, additional unspecified
emission reductions might be necessary to reach some alternative standard levels. In order to
bring these monitor areas into attainment, we calculated controls costs using a fixed cost per
ton approach similar to that used in the ozone RIA analysis. We recognize that a single fixed
cost of control of $15,000 per ton of emissions reductions does not account for the significant
emissions cuts that are necessary in some areas, and so its use provides an estimate that is
likely to differ from actual future costs.
Benefits
EPA estimated the monetized human health benefits of reducing cases of morbidity
among populations exposed to S02 and cases of morbidity and premature mortality among
populations exposed to PM2.5 in 2020 for the selected standard and alternative standard levels
in 2006$. Because S02 is also a precursor to PM2.5; reducing S02 emissions in the projected
non-attainment areas will also reduce PM2.5 formation, human exposure and the incidence of
PM2.5-related health effects. For the selected S02 standard at 75 ppb (99th percentile, daily 1-
hour maximum), the total monetized benefits would be $15 to $37 billion at a 3% discount rate
and $14 to $33 billion at a 7% discount rate. For an S02 standard at 50 ppb, the total
monetized benefits would be $34 to $83 billion at a 3% discount rate and $31 to $75 billion at a
7% discount rate. For an S02 standard at 100 ppb, the total monetized benefits would be $7.4
to $18 billion at a 3% discount rate and $6.7 to $16 billion at a 7% discount rate.
These estimates reflect EPA's most current interpretation of the scientific literature and
are consistent with the methodology used for the proposal RIA. These benefits are incremental
to an air quality baseline that reflects attainment with the 2008 ozone and 2006 PM2.5 National
Ambient Air Quality Standards (NAAQS). More than 99% of the total dollar benefits are
attributable to reductions in PM2.5 exposure resulting from S02 emission reductions. Higher or
lower estimates of benefits are possible using other assumptions; examples of this are provided
7-2
-------
in Figure 5.1 for the selected standard of 75 ppb. Methodological limitations prevented EPA
from quantifying the impacts to, or monetizing the benefits from several important benefit
categories, including ecosystem effects from sulfur deposition, improvements in visibility, and
materials damage. Other direct benefits from reduced S02 exposure have not been quantified,
including reductions in premature mortality.
When estimating the S02- and PM2.5-related human health benefits and compliance
costs in Table 7.1 below, EPA applied methods and assumptions consistent with the state-of-
the-science for human health impact assessment, economics and air quality analysis. EPA
applied its best professional judgment in performing this analysis and believes that these
estimates provide a reasonable indication of the expected benefits and costs to the nation of
the selected S02 standard and alternatives considered by the Agency. The Regulatory Impacts
Analysis (RIA) available in the docket describes in detail the empirical basis for EPA's
assumptions and characterizes the various sources of uncertainties affecting the estimates
below.
EPA's 2009 Integrated Science Assessment for Particulate Matter concluded, based on
the scientific literature, that a no-threshold log-linear model most adequately portrays the PM-
mortality concentration-response relationship. Nonetheless, consistent with historical practice
and our commitment to characterizing the uncertainty in our benefits estimates, EPA has
included a sensitivity analysis with an assumed threshold in the PM-mortality health impact
function in the RIA. EPA has included a sensitivity analysis in the RIA to help inform our
understanding of the health benefits which can be achieved at lower air quality concentration
levels. While the primary estimate and the sensitivity analysis are not directly comparable, due
to differences in population data and use of different analysis years, as well as the difference in
the assumption of a threshold in the sensitivity analysis, comparison of the two results provide
a rough sense of the proportion of the health benefits that occur at lower PM2.5 air quality
levels. Using a threshold of 10 u.g/m3 is an arbitrary choice (EPA could have assumed 6, 8, or 12
u.g/m3 for the sensitivity analysis). Assuming a threshold of 10 u.g/m3, the sensitivity analysis
shows that roughly one-third of the benefits occur at air quality levels below that threshold.
Because the primary estimates reflect EPA's current methods and data, EPA notes that caution
should be exercised when comparing the results of the primary and sensitivity analyses. EPA
appreciates the value of sensitivity analyses in highlighting the uncertainty in the benefits
estimates and will continue to work to refine these analyses, particularly in those instances in
which air quality modeling data are available.
Table 7.1 presents total national primary estimates of costs and benefits for a 3%
discount rate and a 7% discount rate. The net benefits were calculated by subtracting the total
7-3
-------
cost estimate from the two estimates of total benefits. As indicated above, implementation of
the S02 control measures identified from AirControlNET and other sources does not result in
attainment with the all target NAAQS levels in several areas. In these areas, additional
unspecified emission reductions might be necessary to reach some alternative standard levels.
The first part of the table, labeled Partial attainment (known controls), shows only those
benefits and costs from control measures we were able to identify. The second part of the
table, labeled Unidentified Controls, shows only additional benefits and costs resulting from
unidentified controls. The third part of the table, labeled Full attainment, shows total benefits
and costs resulting from both identified and unidentified controls. It is important to emphasize
that we were able to identify control measures for a significant portion of attainment for many
of those counties that would not fully attain the target NAAQS level with identified controls.
Note also that in addition to separating full and partial attainment, the table also separates the
portion of benefits associated with reduced S02 exposure (i.e., S02 benefits) from the additional
benefits associated with reducing S02 emissions, which are precursors to PM2.5 formation -
(i.e., the PM2.5 co-benefits). For instance, for the selected standard of 75 ppb, $2.2 million in
benefits are associated with reduced S02 exposure while $15 billion to $37 billion are
associated with reduced PM2.5 exposure.
7-4
-------
Table 7.1: Monetized Benefits and Costs to Attain Alternate Standard Levels in 2020 (millions
of 2006$)a
# Counties
Fully
Controlled
Partial
Attainment
(identified controls)
•D
« w
Unidentif
Contra
Full Attainment
50 ppb
75ppb
100 ppb
50 ppb
75 ppb
100 ppb
50 ppb
75 ppb
100 ppb
40
20
6
16
4
3
56
24
9
Discount
Rate
3%
7%
3%
7%
3%
7%
3%
7%
3%
7%
3%
7%
3%
7%
3%
7%
3%
7%
Monetized „„ .
Monetized PM2 5
S°* Co-Benefits''"
Benefits
b $30,000 to $74,000
$28,000 to $67,000
b $14,000 to $35,000
$13,000 to $3 1,000
b $6,900 to $17,000
$6,200 to $15,000
b $4,000 to $9,000
$3,000 to $8,000
b $1,000 to $3,000
$1,000 to $3,000
b $500 to $1,000
$500 to $1,000
$34,000 to $83,000
$31,000 to $75,000
$15,000 to $37,000
$14,000 to $34,000
$7,400 to $18,000
$6,700 to $16,000
Costs
$2,600
$960
$470
$1,800
$500
$260
$4,400
$1,500
$730
Net Benefits
$27,000 to $7 1,000
$25,000 to $64,000
$13,000 to $34,000
$12,000 to $30,000
$6,400 to $17,000
$5,700 to $15,000
$2,200 to $7,200
$1,200 to $6,200
$500 to $1,500
$500 to $2,500
$240 to $740
$240 to $740
$30,000 to $79,000
$27,000 to $71,000
$14,000 to $36,000
$13,000 to $33,000
$6,700 to $17,000
$6,000 to $15,000
a Estimates have been rounded to two significant figures and therefore summation may not match table estimates.
bThe approach used to simulate air quality changes for SO2 did not provide the data needed to distinguish partial
attainment benefits from full attainment benefits from reduced SO2 exposure. Therefore, a portion of theSO2
benefits is attributable to the known controls and a portion of the SO2 benefits are attributable to the unidentified
controls. Because all SO2-related benefits are short-term effects, the results are identical for all discount rates.
c Benefits are shown as a range from Pope et al (2002) to Laden et al. (2006). Monetized benefits do not include
unquantified benefits, such as other health effects, reduced sulfur deposition, or improvements in visibility.
dThese models assume that all fine particles, regardless of their chemical composition, are equally potent in
causing premature mortality because there is no clear scientific evidence that would support the development of
differential effects estimates by particle type. Reductions in SO2 emissions from multiple sectors to meet the SO2
NAAQS would primarily reduce the sulfate fraction of PM2.5. Because this rule targets a specific particle precursor
(i.e., SO2), this introduces some uncertainty into the results of the analysis.
7.2 Discussion of Uncertainties and Limitations
Air Quality, Emissions, and Control Strategies
The estimates of emission reductions associated with the control strategies described
above are subject to important limitations and uncertainties. We summarize these limitations
as follows:
• Actual State Implementation Plans May Differ from our Simulation: In order to reach
attainment with the proposed NAAQS, each state will develop its own
7-5
-------
implementation plan implementing a combination of emissions controls that may
differ from those simulated in this analysis. This analysis therefore represents an
approximation of the emissions reductions that would be required to reach
attainment and should not be treated as a precise estimate.
• Use of Existing CMAQ Model Runs: This analysis represents a screening level
analysis. We did not conduct new regional scale modeling specifically targets to S02;
instead we relied upon impact ratios developed from model runs used in the
analysis underlying the PM2.5 NAAQS.
• Unidentified controls: We have limited information on available controls for some of
the monitor areas included in this analysis. For a number of small non-EGU and
area sources, there is little or no information available on S02 controls.
Costs
• We do not have sufficient information for all of our known control measures to calculate
cost estimates that vary with an interest rate. We are able to calculate annualized costs
at an interest rate other than 7% (e.g., 3% interest rate) where there is sufficient
information—available capital cost data, and equipment life—to annualize the costs for
individual control measures. For the vast majority of nonEGU point source control
measures, we do have sufficient capital cost and equipment life data for individual
control measures to prepare annualized capital costs using the standard capital recovery
factor. Hence, we are able to provide annualized cost estimates at different interest
rates for the point source control measures.
• There are some unquantified costs that are not adequately captured in this illustrative
analysis. These costs include the costs of federal and State administration of control
programs, which we believe are less than the alternative of States developing
approvable SIPs, securing EPA approval of those SIPs, and Federal/State enforcement.
Additionally, control measure costs referred to as "no cost" may require limited
government agency resources for administration and oversight of the program not
included in this analysis; those costs are generally outweighed by the saving to the
industrial, commercial, or private sector. The Agency also did not consider transactional
costs and/or effects on labor supply in the illustrative analysis.
7-6
-------
Benefits
Although we strive to incorporate as many quantitative assessments of uncertainty,
there are several aspects for which we are only able to address qualitatively. These aspects are
important factors to consider when evaluating the relative benefits of the attainment strategies
for each of the alternative standards:
1. The 12 km CMAQ grid, which is the air quality modeling resolution, may be too coarse
to accurately estimate the potential near-field health benefits of reducing S02 emissions.
These uncertainties may under- or over-estimate benefits.
2. The interpolation techniques used to estimate the full attainment benefits of the
alternative standards contributed some uncertainty to the analysis. The great majority
of benefits estimated for the various standard alternatives were derived through
interpolation. As noted previously in this chapter, these benefits are likely to be more
uncertain than if we had modeled the air quality scenario for both S02 and PM2.5. In
general, the VNA interpolation approach will under-estimate benefits because it does
not account for the broader spatial distribution of air quality changes that may occur
due to the implementation of a regional emission control program.
3. There are many uncertainties associated with the health impact functions used in this
modeling effort. These include: within study variability (the precision with which a given
study estimates the relationship between air quality changes and health effects); across
study variation (different published studies of the same pollutant/health effect
relationship typically do not report identical findings and in some instances the
differences are substantial); the application of C-R functions nationwide (does not
account for any relationship between region and health effect, to the extent that such a
relationship exists); extrapolation of impact functions across population (we assumed
that certain health impact functions applied to age ranges broader than that considered
in the original epidemiological study); and various uncertainties in the C-R function,
including causality and thresholds. These uncertainties may under- or over-estimate
benefits.
4. Co-pollutants present in the ambient air may have contributed to the health effects
attributed to S02 in single pollutant models. Risks attributed to S02 might be
overestimated where concentration-response functions are based on single pollutant
models. If co-pollutants are highly correlated with S02, their inclusion in an S02 health
effects model can lead to misleading conclusions in identifying a specific causal
pollutant. Because this collinearity exists, many of the studies reported statistically
insignificant effect estimates for both S02 and the co-pollutants; this is due in part to the
loss of statistical power as these models control for co-pollutants. Where available, we
7-7
-------
have selected multipollutant effect estimates to control for the potential confounding
effects of co-pollutants; these include NYDOH (2006), Schwartz et al. (1994) and
O'Connor et al. (2008). The remaining studies include single pollutant models.
5. This analysis is for the year 2020, and projecting key variables introduces uncertainty.
Inherent in any analysis of future regulatory programs are uncertainties in projecting
atmospheric conditions and source level emissions, as well as population, health
baselines, incomes, technology, and other factors.
6. This analysis omits certain unquantified effects due to lack of data, time and resources.
These unquantified endpoints include other health effects, ecosystem effects, and
visibility. EPA will continue to evaluate new methods and models and select those most
appropriate for estimating the benefits of reductions in air pollution. Enhanced
collaboration between air quality modelers, epidemiologists, toxicologists, ecologists,
and economists should result in a more tightly integrated analytical framework for
measuring benefits of air pollution policies.
7. PM2.5 co-benefits represent a substantial proportion of total monetized benefits (over
99% of total monetized benefits), and these estimates are subject to a number of
assumptions and uncertainties.
a. PM2.5 co-benefits were derived through benefit per-ton estimates, which do not
reflect local variability in population density, meteorology, exposure, baseline
health incidence rates, or other local factors that might lead to an over-estimate
or under-estimate of the actual benefits of controlling directly emitted fine
particulates.
b. We assume that all fine particles, regardless of their chemical composition, are
equally potent in causing premature mortality. This is an important assumption,
because PM2.5 produced via transported precursors emitted from EGUs may
differ significantly from direct PM2.5 released from diesel engines and other
industrial sources, but no clear scientific grounds exist for supporting differential
effects estimates by particle type.
c. We assume that the health impact function for fine particles is linear within the
range of ambient concentrations under consideration. Thus, the estimates
include health benefits from reducing fine particles in areas with varied
concentrations of PM2.5; including both regions that are in attainment with fine
particle standard and those that do not meet the standard down to the lowest
modeled concentrations.
d. To characterize the uncertainty in the relationship between PM2.5and premature
mortality (which typically accounts for 85% to 95% of total monetized benefits),
we include a set of twelve estimates based on results of the expert elicitation
study in addition to our core estimates. Even these multiple characterizations
7-8
-------
omit the uncertainty in air quality estimates, baseline incidence rates,
populations exposed and transferability of the effect estimate to diverse
locations. As a result, the reported confidence intervals and range of estimates
give an incomplete picture about the overall uncertainty in the PM2.5 estimates.
This information should be interpreted within the context of the larger
uncertainty surrounding the entire analysis. For more information on the
uncertainties associated with PM2.5 co-benefits, please consult the PM2.5 NAAQS
RIA (Table 5.5).
While the monetized benefits of reduced S02 exposure appear small when compared to the
monetized benefits of reduced PM2.5 exposure, readers should not necessarily infer that the
total monetized benefits of attaining a new S02 standard are minimal. For this rule, the
monetized PM2.5 co-benefits represent over 99% of the total monetized benefits. This result is
consistent with other recent RIAs, where the PM2.5 co-benefits represent a large proportion of
total monetized benefits. This result is amplified in this RIA by the decision not to quantify S02-
related premature mortality and other morbidity endpoints due to the uncertainties associated
with estimating those endpoints. Studies have shown that there is a relationship between S02
exposure and premature mortality, but that relationship is limited by potential confounding.
Because premature mortality generally comprises over 90% of the total monetized benefits,
this decision may substantially underestimate the monetized health benefits of reduced S02
exposure.
In addition, we were unable to quantify the benefits from several welfare benefit
categories. We lacked the necessary air quality data to quantify the benefits from
improvements in visibility from reducing light-scattering particles. Previous RIAs for ozone (U.S.
EPA, 2008a) and PM2.5 (U.S. EPA, 2006a) indicate that visibility is an important benefit category,
and previous efforts to monetize those benefits have only included a subset of visibility
benefits, excluding benefits in urban areas and many national and state parks. Even this subset
accounted for up to 5% of total monetized benefits in the Ozone NAAQS RIA (U.S. EPA, 2008a).
We were also unable to quantify the ecosystem benefits of reduced sulfur deposition
because we lacked the necessary air quality data, and the methodology to estimate ecosystem
benefits is still being developed. Previous assessments (U.S. EPA, 1999; U.S. EPA, 2005; U.S.
EPA, 2009e) indicate that ecosystem benefits are also an important benefits category, but those
efforts were only able to monetize a tiny subset of ecosystem benefits in specific geographic
locations, such as recreational fishing effects from lake acidification in the Adirondacks. We
were also unable to quantify the benefits of decreased mercury methylation from sulfate
deposition. Quantifying the relationship between sulfate and mercury methylation in natural
7-9
-------
settings is difficult, but some studies have shown that decreasing sulfate deposition can also
decrease methylmercury.
7-10
-------
Chapter 8: Statutory and Executive Order Reviews
1.0 Executive Order 12866: Regulatory Planning and Review
Under section 3(f)(l) of Executive Order 12866 (58 FR 51735, October 4, 1993), this
action is an "economically significant regulatory action" because it is likely to have an
annual effect on the economy of $100 million or more. Accordingly, EPA submitted this
action to the Office of Management and Budget (OMB) for review under EO 12866 and
any changes made in response to OMB recommendations have been documented in the
docket for this action. In addition, EPA prepared a Regulatory Impact Analysis (RIA) of
the potential costs and benefits associated with this action. However, the CAA and
judicial decisions make clear that the economic and technical feasibility of attaining the
national ambient standards cannot be considered in setting or revising NAAQS, although
such factors may be considered in the development of State implementation plans to
implement the standards. Accordingly, although an RIA has been prepared, the results
of the RIA have not been considered by EPA in developing this final rule.
2.0 Paperwork Reduction Act
The information collection requirements in this final rule have been submitted
for approval to the Office of Management and Budget (OMB) under the Paperwork
Reduction Act, 44 U.S.C. 3501 et seq. The Information Collection Request (ICR)
document prepared by EPA for these proposed revisions to part 58 has been assigned
EPA ICR number 2370.01.
The information collected under 40 CFR part 53 (e.g., test results, monitoring
records, instruction manual, and other associated information) is needed to determine
whether a candidate method intended for use in determining attainment of the NAAQS
in 40 CFR part 50 will meet the design, performance, and/or comparability requirements
for designation as a Federal reference method (FRM) or Federal equivalent method
(FEM). We do not expect the number of FRM or FEM determinations to increase over
the number that is currently used to estimate burden associated with S02 FRM/FEM
determinations provided in the current ICR for 40 CFR part 53 (EPA ICR numbers
2370.01). As such, no change in the burden estimate for 40 CFR part 53 has been made
as part of this rulemaking.
The information collected and reported under 40 CFR part 58 is needed to
determine compliance with the NAAQS, to characterize air quality and associated health
8-1
-------
impacts, to develop emissions control strategies, and to measure progress for the air
pollution program. The amendments would revise the technical requirements for S02
monitoring sites, require the siting and operation of additional S02 ambient air
monitors, and the reporting of the collected ambient S02 monitoring data to EPA's Air
Quality System (AQS). This Information Collection is estimated to involve 102
respondents for a total approximate cost of $15,203,762 (total capital, and labor and
non-labor operation and maintenance) and a total burden of 207,662 hours. The labor
costs associated with these hours is $11,130,409. Included in the $15,203,762 total are
other costs of non-labor operations and maintenance of $1,104,377 and equipment and
contract costs of $2,968,975. In addition to the costs at the State and local air quality
management agencies, there is a burden to EPA of total of 14,749 hours and
$1,060,621. Burden is defined at 5 CFR 1320.3(b). State, local, and tribal entities are
eligible for State assistance grants provided by the Federal government under the CAA
which can be used for monitors and related activities.
An agency may not conduct or sponsor, and a person is not required to respond
to, a collection of information unless it displays a currently valid OMB control number.
The OMB control numbers for EPA's regulations in 40 CFR are listed in 40 CFR part 9.
3.0 Regulatory Flexibility Act
The Regulatory Flexibility Act (RFA) generally requires an agency to prepare a
regulatory flexibility analysis of any rule subject to notice and comment rulemaking
requirements under the Administrative Procedure Act or any other statute unless the
agency certifies that the rule will not have a significant economic impact on a substantial
number of small entities. Small entities include small businesses, small organizations,
and small governmental jurisdictions.
For purposes of assessing the impacts of this rule on small entities, small entity is
defined as: (1) a small business that is a small industrial entity as defined by the Small
Business Administration's (SBA) regulations at 13 CFR 121.201; (2) a small governmental
jurisdiction that is a government of a city, county, town, school district or special district
with a population of less than 50,000; and (3) a small organization that is any not-for-
profit enterprise which is independently owned and operated and is not dominant in its
field.
After considering the economic impacts of this proposed rule on small entities, I
certify that this action will not have a significant economic impact on a substantial
8-2
-------
number of small entities. This final rule will not impose any requirements on small
entities. Rather, this rule establishes national standards for allowable concentrations of
S02 in ambient air as required by section 109 of the CAA. American Trucking Ass''ns v.
EPA, 175 F. 3d 1027, 1044-45 (D.C. Cir. 1999) (NAAQS do not have significant impacts
upon small entities because NAAQS themselves impose no regulations upon small
entities). Similarly, the amendments to 40 CFR Part 58 address the requirements for
States to collect information and report compliance with the NAAQS and will not impose
any requirements on small entities.
4.0 Unfunded Mandates Reform Act
This action is not subject to the requirements of sections 202 and 205 of the
UMRA. EPA has determined that this proposed rule does not contain a Federal mandate
that may result in expenditures of $100 million or more for State, local, and tribal
governments, in the aggregate, or the private sector in any one year. The revisions to
the S02 NAAQS impose no enforceable duty on any State, local or Tribal governments or
the private sector. The expected costs associated with the monitoring requirements are
described in EPA's ICR document, but those costs are not expected to exceed $100
million in the aggregate for any year. Furthermore, as indicated previously, in setting a
NAAQS, EPA cannot consider the economic or technological feasibility of attaining
ambient air quality standards. Because the CAA prohibits EPA from considering the
types of estimates and assessments described in section 202 when setting the NAAQS,
the UMRA does not require EPA to prepare a written statement under section 202 for
the revisions to the S02 NAAQS.
With regard to implementation guidance, the CAA imposes the obligation for
States to submit SIPs to implement the S02 NAAQS. In this final rule, EPA is merely
providing an interpretation of those requirements. However, even if this rule did
establish an independent obligation for States to submit SIPs, it is questionable whether
an obligation to submit a SIP revision would constitute a Federal mandate in any case.
The obligation for a State to submit a SIP that arises out of section 110 and section 191
of the CAA is not legally enforceable by a court of law, and at most is a condition for
continued receipt of highway funds. Therefore, it is possible to view an action requiring
such a submittal as not creating any enforceable duty within the meaning of U.S.C. 658
for purposes of the UMRA. Even if it did, the duty could be viewed as falling within the
exception for a condition of Federal assistance under U.S.C. 658.
8-3
-------
EPA has determined that this final rule contains no regulatory requirements that
might significantly or uniquely affect small governments because it imposes no
enforceable duty on any small governments. Therefore, the rule is not subject to the
requirements of section 203 of the UMRA.
5.0 Executive Order 13132: Federalism
This final rule does not have federalism implications. It will not have substantial
direct effects on the States, on the relationship between the national government and
the States, or on the distribution of power and responsibilities among the various levels
of government, as specified in Executive Order 13132. The rule does not alter the
relationship between the Federal government and the States regarding the
establishment and implementation of air quality improvement programs as codified in
the CAA. Under section 109 of the CAA, EPA is mandated to establish NAAQS; however,
CAA section 116 preserves the rights of States to establish more stringent requirements
if deemed necessary by a State. Furthermore, this rule does not impact CAA section 107
which establishes that the States have primary responsibility for implementation of the
NAAQS. Finally, as noted in section E (above) on UMRA, this rule does not impose
significant costs on State, local, or tribal governments or the private sector. Thus,
Executive Order 13132 does not apply to this rule.
6.0 Executive Order 13175: Consultation and Coordination with Indian Tribal
Governments
Executive Order 13175, entitled "Consultation and Coordination with Indian
Tribal Governments" (65 FR 67249, November 9, 2000), requires EPA to develop an
accountable process to ensure "meaningful and timely input by tribal officials in the
development of regulatory policies that have tribal implications." This final rule does
not have tribal implications, as specified in Executive Order 13175. It does not have a
substantial direct effect on one or more Indian tribes, on the relationship between the
Federal government and Indian tribes, or on the distribution of power and
responsibilities between the Federal government and tribes. The rule does not alter the
relationship between the Federal government and tribes as established in the CAA and
the TAR. Under section 109 of the CAA, EPA is mandated to establish NAAQS; however,
this rule does not infringe existing tribal authorities to regulate air quality under their
own programs or under programs submitted to EPA for approval. Furthermore, this rule
does not affect the flexibility afforded to tribes in seeking to implement CAA programs
consistent with the TAR, nor does it impose any new obligation on tribes to adopt or
8-4
-------
implement any NAAQS. Finally, as noted in section E (above) on UMRA, this rule does
not impose significant costs on tribal governments. Thus, Executive Order 13175 does
not apply to this rule.
7.0 Executive Order 13045: Protection of Children from Environmental Health &
Safety Risks
This action is subject to Executive Order (62 FR 19885, April 23, 1997) because it
is an economically significant regulatory action as defined by Executive Order 12866,
and we believe that the environmental health risk addressed by this action has a
disproportionate effect on children. This final rule will establish uniform national
ambient air quality standards for S02; these standards are designed to protect public
health with an adequate margin of safety, as required by CAA section 109. The
protection offered by these standards may be especially important for asthmatics,
including asthmatic children, because respiratory effects in asthmatics are among the
most sensitive health endpointsfor S02exposure. Because asthmatic children are
considered a sensitive population, we have evaluated the potential health effects of
exposure to S02 pollution among asthmatic children. These effects and the size of the
population affected are discussed in chapters 3 and 4 of the ISA; chapters 3, 4, 7, 8, 9 of
the REA, and sections II.A through II.E of the preamble.
8.0 Executive Order 13211: Actions that Significantly Affect Energy Supply,
Distribution or Use
This rule is not a "significant energy action" as defined in Executive Order 13211,
"Actions Concerning Regulations That Significantly Affect Energy Supply, Distribution, or
Use" (66 FR 28355; May 22, 2001) because it is not likely to have a significant adverse
effect on the supply, distribution, or use of energy. The purpose of this rule is to
establish revised NAAQS for S02. The rule does not prescribe specific control strategies
by which these ambient standards will be met. Such strategies will be developed by
States on a case-by-case basis, and EPA cannot predict whether the control options
selected by States will include regulations on energy suppliers, distributors, or users.
Thus, EPA concludes that this rule is not likely to have any adverse energy effects.
9.0 National Technology Transfer and Advancement Act
Section 12(d) of the National Technology Transfer and Advancement Act of 1995
(NTTAA), Public Law 104-113, section 12(d) (15 U.S.C. 27) directs EPA to use voluntary
8-5
-------
consensus standards in its regulatory activities unless to do so would be inconsistent
with applicable law or otherwise impractical. Voluntary consensus standards are
technical standards (e.g., materials specifications, test methods, sampling procedures,
and business practices) that are developed or adopted by voluntary consensus
standards bodies. The NTTAA directs EPA to provide Congress, through OMB,
explanations when the Agency decides not to use available and applicable voluntary
consensus standards.
This rulemaking involves technical standards with regard to ambient monitoring
of S02. The use of this voluntary consensus standard would be impractical because the
analysis method does not provide for the method detection limits necessary to
adequately characterize ambient S02 concentrations for the purpose of determining
compliance with the revisions to the S02 NAAQS.
10.0 Executive Order 12898: Federal Actions to Address Environmental Justice in
Minority Populations and Low-Income Populations
Executive Order 12898 (59 FR 7629; Feb. 16, 1994) establishes federal executive
policy on environmental justice. Its main provision directs federal agencies, to the
greatest extent practicable and permitted by law, to make environmental justice part of
their mission by identifying and addressing, as appropriate, disproportionately high and
adverse human health or environmental effects of their programs, policies, and
activities on minority populations and low-income populations in the United States.
EPA has determined that this final rule will not have disproportionately high and
adverse human health or environmental effects on minority or low-income populations
because it increases the level of environmental protection for all affected populations
without having any disproportionately high and adverse human health effects on any
population, including any minority or low-income population. The rule will establish
uniform national standards for S02 in ambient air.
8-6
------- |