HEALTH 0000
^BENEFITS
> per kilowatt hour
Public Health Benefits per kWh
of Energy Efficiency and Renewable
Energy in the United States:
A Technical Report
vvEPA
United States
Environmental Protection
Agency
State and Local
Energy and Environment Program
July 2019
-------
Contents
Acknowledgments 1
Executive Summary 2
When to use benefits-per-kWh screening values? 2
When not to use benefits-per-kWh values? 2
Benefits-per-kWh screening values 2
Understanding the Values 4
How to use BPK values? 4
Introduction 5
Background 6
Methods 8
Overview of Approach 8
Modeling Scenarios Development 10
Project, Program, and Policy Size Assumptions 13
Electricity and Emissions Modeling 14
Air Quality and Health Impact Modeling 15
Developing the Health Benefits-per-kWh Estimates 16
Uncertainty 17
Limitations 19
Results 21
Discussion 27
Conclusions 29
References 30
Appendix A: AVoided Emissions and geneRation Tool (AVERT) 35
Appendix B: Co-Benefits Risk Assessment (COBRA) Health Impacts Screening and Mapping
Tool 37
Appendix C: Sensitivity Analyses on Project, Program, or Policy Size and Peak Energy-
Efficiency Definition 40
Appendix D: Top 200 Hours of Demand Benefit-per-kWh Results 49
Appendix E: Health Impact Functions 50
Appendix F: Health Benefits Valuation 52
Appendix G: Detailed Benefits-per-kWh Results 53
Appendix H: Conversions 56
-------
Acki :s
This report was developed by EPA's State and Local Energy and Environment Program within
the Climate Protection Partnerships Division of EPA's Office of Atmospheric Programs. Denise
Mulholland led a technical team of EPA experts to develop this report, including Emma
Zinsmeister, Robyn DeYoung, Nikolaas Dietsch, Neal Fann, and Elizabeth Chan, with
significant analytic support from David Cooley and Kait Siegel of Abt Associates.
The EPA technical team would like to thank Carolyn Snyder, Director of EPA's Climate
Protection Partnerships Division, and Julie Rosenberg, Chief of EPA's State and Local Branch,
for their leadership on and support of this project.
The team would also like to thank the following technical experts for comments they provided
on earlier drafts of this report: Susan Annenberg (George Washington University), James
Critchfield (EPA), Tom Eckman (Lawrence Berkeley National Laboratory), Mimi Goldberg
(DNV GL), Etan Gumerman (Nicholas Institute, Duke University), Sara Hayes (American
Council for an Energy-Efficient Economy), Ed Holt (Ed Holt and Associates), David Hoppock
(Maryland Public Service Commission), Jeff Loiter (Optimal Energy), Dev Millstein (Lawrence
Berkeley National Laboratory), Jason West (University of North Carolina), and Ryan Wiser
(Lawrence Berkeley National Laboratory).
In addition, EPA would like to thank Pat Knight and Nina Peluso from Synapse Energy
Economics Inc. and Ariel Horowitz from the Massachusetts Clean Energy Center (formerly
from Synapse Energy Economics, Inc) for their assistance with AVERT for this analysis.
1
-------
Executive Summary
EPA has developed a set of values that help state and local government policymakers and other
stakeholders estimate the monetized public health benefits of investments in energy efficiency
and renewable energy (EE/RE) using methods consistent with those EPA uses for health benefits
analyses at the federal level. It's important to note that EPA is continually reviewing methods
and assumptions for quantifying public health benefits. The values presented here and the
associated documentation will be updated as appropriate to reflect any future changes in methods
or assumptions.
When to use benefits-per-kWh screening values?
Benefits per kilowatt-hour (BPK) values are reasonable
approximations of the health benefits of state EE/RE
investments that can be used for preliminary analysis
when comparing across state and local policy scenarios to
indicate direction and relative magnitude.
Examples of analyses where it would be appropriate to use
them include:
• Estimating the public health benefits of regional,
state, or local-level investments in EE/RE projects,
programs, and policies
• Understanding the cost-effectiveness of regional,
state, or local-level EE projects, programs, and
measures
• Incorporating health benefits in short-term
regional, state, or local policy analyses and
decision-making
When not to use benefits-per-kWh values?
BPK values are not a substitute for sophisticated analysis and should not be used to justify or inform
federal regulatory decisions. They are based on data inputs, assumptions, and methods that approximate
the dynamics of energy, environment, and health interactions and include uncertainties and limitations, as
documented in the technical report.
Benefits-per-kWh screening values
EPA used a peer reviewed methodology and tools to develop a set of screening-level regional
estimates of the dollar benefits per kilowatt-hour from four different types of EE/RE initiatives.
• Uniform Energy Efficiency - Energy efficiency programs, projects, and measures that
achieve a constant level of savings over time,
• Peak Energy Efficiency - Energy efficiency programs, projects, and measures that
achieve savings during 12pm-6pm when energy demand is high (i.e. peak),
• Solar Energy - Programs, projects, and measures that increase the supply of solar energy
available (e.g. utility-scale and rooftop solar generation), and
• Wind Energy - Programs, projects, and measures that increase the supply of wind
available (e.g. wind turbines).
Audience for BPK screening values
Stakeholders interested in
approximating the outdoor air quality-
related public health benefits from
EE/RE, including:
• State and local energy, air quality,
or public health agencies
• Public utility commissions
• Energy efficiency and renewable
energy project developers
• Industry organization
• Nongovernmental organizations
• Other researchers
2
-------
Table ES.l. 2017 Benefits-per-kWh Values (cents per kWh, 2017 USD)1
3% Discount Rate
7% Discount Rate
Region
Project Type
2017 C/kWh
2017 C/kWh
2017 C/kWh
2017 C/kWh
(low estimate)
(high estimate)
(low estimate)
(high estimate)
Uniform EE
0.48
1.08
0.42
0.96
California
EE at Peak
0.52
1.17
0.46
1.04
Solar
0.51
1.15
0.45
1.03
Wind
0.48
1.09
0.43
0.97
Uniform EE
3.51
7.95
3.14
7.09
Great Lakes/ Mid-
EE at Peak
3.57
8.08
3.19
7.21
Atlantic
Solar
3.67
8.29
3.27
7.39
Wind
3.35
7.59
2.99
6.77
Uniform EE
2.31
5.23
2.06
4.66
Lower Midwest
EE at Peak
2.11
4.77
1.88
4.25
Solar
2.19
4.96
1.96
4.42
Wind
2.35
5.32
2.10
4.74
Uniform EE
1.65
3.73
1.47
3.33
Northeast
EE at Peak
2.24
5.07
2.00
4.52
Solar
1.94
4.38
1.73
3.91
Wind
1.58
3.56
1.41
3.18
Uniform EE
1.13
2.55
1.01
2.28
Pacific Northwest
EE at Peak
1.12
2.54
1.00
2.27
Solar
1.17
2.64
1.04
2.35
Wind
1.13
2.55
1.01
2.27
Uniform EE
1.03
2.32
0.92
2.07
Rocky Mountains
EE at Peak
0.98
2.21
0.87
1.98
Solar
0.99
2.25
0.89
2.01
Wind
1.07
2.41
0.95
2.15
Uniform EE
1.78
4.02
1.58
3.58
Southeast
EE at Peak
1.87
4.24
1.67
3.78
Solar
1.83
4.15
1.64
3.70
Wind
1.76
3.98
1.57
3.55
Uniform EE
0.71
1.62
0.64
1.44
Southwest
EE at Peak
0.70
1.59
0.63
1.42
Solar
0.73
1.64
0.65
1.46
Wind
0.77
1.73
0.68
1.54
Uniform EE
1.58
3.58
1.41
3.19
Texas
EE at Peak
1.39
3.13
1.24
2.80
Solar
1.42
3.22
1.27
2.87
Wind
1.63
3.69
1.45
3.29
Uniform EE
3.12
7.06
2.78
6.30
Upper Midwest
EE at Peak
2.75
6.22
2.45
5.55
Solar
2.89
6.53
2.58
5.83
Wind
3.20
7.23
2.85
6.45
1 In addition to using these regional values, users can also use EPA's AVERT and COBRA tools to develop more specific analyses,
such as state- or county-level health benefits estimates. For more information on other more sophisticated options for
modeling health benefits for or how to quantify the electricity impacts of energy efficiency and renewable energy, see the EPA
report, Quantifying the Multiple Benefits of Energy Efficiency and Renewable Energy: A Guide for State and Local Governments.
3
-------
Understanding the Values
EPA created BPK values
Mapping Tool. BPK values
are:
using existing tools, including
EPA's AVoided Emissions
and geneRation Tool
(AVERT) and CO-Benefits
Risk Assessment (COBRA)
Health Impacts Screening and
• Available for each of
the four project types
for each of the ten
AVERT regions
shown in the map
below
Northwest
(NW)
Upper
Midwest
IWMW)
treat Lakw I Mid
Atlantic (EMW.i
i Rocky
Mountains
(RM)
Lower
Midwest
(SC)
California*
(CA)
Southeast
(SE)
Figure ES.l. AVERT Regions.
• Based on 2017
electricity generation data and emissions, population, baseline mortality incidence rate,
and income growth projections
• Presented in 2017 dollars and reflecting the use of either a 3% or a 7% discount rate as
recommended by EPA's Guidelines for Preparing Economic Analyses (2010)
• Calculated using the same health impact functions EPA uses for regulatory impact
analyses, including the calculation of low estimates of mortality using health impact
functions that assume people are not very sensitive to changes in PM2.5 levels and high
estimates of mortality using functions that assume people are more sensitive to changes
in PM2.5
How to use BPK values?
States and communities interested in having screening-level estimation of outdoor air quality-
related health impacts of energy efficiency or renewable energy can multiply the BPK values,
presented in Table ES. 1 in cents per kilowatt hour, by the number of kWh saved from EE or
generated from RE to estimate potential health benefits from projects in dollars saved. Users
should keep in mind there are uncertainties associated with any modeled estimates when
interpreting or reporting results.
4
-------
!! 11 I ' '1 "' ! il'Mi
State and local government policymakers have increasingly been asking for the
U.S. Environmental Protection Agency's (EPA's) help in understanding the opportunities for
using energy efficiency and renewable energy (EE/RE) to reduce air pollution and improve
public health. Many recognize that EE/RE projects, programs, and policies can reduce air
pollution emissions from the electric power sector either by decreasing demand for electricity
generation or by displacing fossil fuel-based generation with zero-emitting sources of generation.
They also recognize that these avoided emissions may lead to tangible public health benefits,
such as reducing the number of premature deaths, incidences of respiratory and cardiovascular
illnesses, and missed work and school days.2 However, in many cases, state and local decision-
makers are not quantifying or fully reflecting the health benefits of existing or planned EE/RE
projects, programs, and policies in their decision-making processes. EPA has found that state and
local decision-makers may not be fully aware of or confident in the available quantification tools
and methods; or they lack the time, resources, or expertise needed to quantify the health benefits.
EPA seeks to address this gap by providing state and local governments and their stakeholders
with tools and information to estimate the public health benefits of EE/RE. In particular, EPA
has developed screening-level regional estimates of the benefits per kilowatt-hour (kWh) of
EE/RE projects, programs, and policies.3 The goal of these estimates is to create credible and
comparable values (i.e., factors) that stakeholders, such as state and local governments, EE/RE
project developers, and nongovernmental organizations (NGOs), can use to estimate health
benefits of EE/RE projects, programs, and policies. EPA has also sought to ensure that these
values are easy to use, and do not require state and local governments or other users to download
specific modeling software packages.
This report describes EPA's approach for developing this set of screening-level estimates of the
monetized health benefits per kWh that represent the benefits from fossil fuel-based generation
reduced or avoided as a result of EE, solar, and wind projects, programs, and policies. The
estimates use a 2017 profile of the electricity system to represent the benefits in the near-term of
EE/RE projects, programs, and policies that have already been or are about to be implemented.
The resulting health benefits-per-kWh (BPK) values can be used by state and local governments,
EE/RE project developers, and other stakeholders to develop a more complete picture of the
public health benefits of existing or proposed EE/RE projects, programs, and policies. Note that
because BPK values provide a screening-level estimate of health benefits of EE/RE, they may
not be appropriate for certain analyses, such as federal air quality rulemaking. It's also important
to note that EPA is continually reviewing its methods and assumptions for quantifying public
health benefits. The health BPK values presented here and the associated documentation will be
updated as appropriate to reflect any future changes in EPA methods or assumptions.
2 The Health Effects Institute (2018) estimates that in 2016,105,669 premature deaths in the United States were
attributable to air pollution [93,376 due to fine particulate matter (PM2 5) and 12,293 due to ozone (03)].
3 These estimates include the contiguous United States, but do not include Alaska and Hawaii. These states are not
included in the AVoided Emissions and geneRation Tool (AVERT) used to estimate impacts of EE/RE on air
pollution emissions because they do not report emissions data for most of their electric generating units (EGUs) to
EPA. Alaska and Hawaii are also not included in the Co-Benefits Risk Assessment (COBRA) Health Impacts
Screening and Mapping Tool used to estimate the health impacts of EE/RE because they were not included in the air
quality modeling originally used to develop the tool.
5
-------
Background
Electricity generation in the United States is essential to our economy but it also results in
significant emissions of air pollution, depending upon how it is generated. In 2014, the electricity
generation sector emitted more than 1 million tons each of nitrogen oxides (NOx) and sulfur
dioxide (SO2); and more than 170,000 tons of PM2.5, which is more than the PM2.5 emissions of
highway vehicles in that year (EPA 2018). Emissions of these pollutants can result in serious
health impacts, including premature mortality, non-fatal heart attacks, asthma exacerbations, and
other respiratory diseases. EPA's retrospective analysis of the Clean Air Act (CAA) found that
approximately 85 percent of the public health benefits of air quality regulations are due to PM
reductions, with the remainder coming from other air pollutants, such as ozone (O3) (EPA
2011b).
While the U.S. electric power sector has historically been a significant source of air pollution, the
sector has undergone rapid change in recent years. Between 2007 and 2016, coal and oil
generation sources combined have decreased from just over 50 percent of the U.S. generation
resource mix to 31 percent; and renewables, including wind, solar, and geothermal, have
increased from just over 1 percent to nearly 7 percent of the resource mix (Table 1). Similarly,
electricity savings from energy-efficiency programs were over 180 terawatt hours (TWh) in
2016, an increase of more than 115 percent from 2008 (IEI 2017). All of these changes amount
to a cleaner U.S. electric power sector with reduced emissions and health impacts.
Table 1. U.S. Generation Resource Mix, 2007-2016
Generation Resource Mix (percent)
Year
Coal
Oil
Gas
Other
Fossil
Biomass
Hydro
Nuclear
Wind
Solar
Geo-
Thermal
Other/
Unknown
2007
48.5
1.6
21.7
0.5
1.3
5.8
19.4
0.8
0.0
0.4
0.1
2009
44.5
1.1
23.3
0.3
1.4
6.8
20.2
1.9
0.0
0.4
0.1
2012
37.4
0.7
30.3
0.4
1.4
6.7
18.0
3.4
0.1
0.4
0.1
2014
38.7
0.7
27.5
0.5
1.6
6.2
19.5
4.4
0.4
0.4
0.1
2016
30.4
0.6
33.8
0.3
1.7
6.4
19.8
5.6
0.9
0.4
0.1
Source: EPA eGRID.
In order to help state and local governments quantify the health benefits of EE/RE, EPA first
needed to understand the current state of the scientific literature to determine if there are best
practices or factors that states could apply. EPA commissioned a literature review that examined
more than 60 studies for BPK values in order to better understand current methods and health
benefits of EE/RE projects, programs, and policies (EPA 2017). Through the literature review,
EPA found that the results varied depending on the approach used, the benefits included, and the
geographic focus of the analysis. Therefore, the resulting sets of BPK values identified in the
literature review were not easily comparable to one another.
Lawrence Berkley National Laboratory (LBL), for example, published several studies examining
both the prospective and retrospective health benefits from wind, solar, and renewable portfolio
standard (RPS) programs across the United States (Table 2). The benefits reported by each study
are an average value of health benefits calculated using multiple different air quality and health
impact models, including the Air Pollution Emission Experiments and Policy Analysis Model
6
-------
(AP2), EPA's benefit-per-ton methodology, EPA's CO-Benefits Risk Assessment (COBRA)
Health Impacts Screening and Mapping Tool, and the Estimating Air Pollution Social Impact
Using Regression (EASIUR) model. Overall, these studies provide a range nationally between
2,60/kWh and 10.10/kWh for recent years, and between 0.40/kWh and 8,20/kWh when looking
prospectively. Other studies included in the literature review generated a different range of
results that were not directly comparable to the LBL estimates, typically because they used a
variety of models or included additional benefits. For example, some of the models used in
studies identified in the literature review include non-health, welfare benefits, such as avoiding
damages from decreased timber and agricultural yields, reduced visibility, accelerated
depreciation of materials, and reductions in recreation services; results from these studies may be
higher than the values calculated using models that focus solely on health benefits.
Table 2. Public Health Benefits from wind, solar, and RPS program across the US
Program Evaluated
Benefit-per-kWh (c/kWh)
Source
2013 RPS programs
2.60/kWh- 10.10/kWh
Barbose et al. 2016
2015 Wind energy
7.3/0kWh
Millstein et al. 2017
2015 Solar energy
40/kWh
Millstein et al. 2017
2015-2050 RPS Programs
2.70/kWh - 8.20/kWh
Mai et al. 2016
2050 Wind energy
0.40/kWh - 2.20/kWh
Wiser et al. 2016a
2050 Solar energy
0.70/kWh - 2.60/kWh
Wiser et al. 2016b
The literature review also identified two key gaps across all available estimates. While several
studies estimated the benefits per kWh in specific regions, particularly the Northeast and
California, there is no comprehensive set of monetized health benefits per kWh from EE/RE for
all U.S. regions. The national numbers provided by LBL do not appropriately represent regional
differences in the specific composition of electricity generation throughout the United States and
therefore do not account for regional differences in emissions. Additionally, the values from the
literature are not methodologically consistent, and can therefore not be compared with
confidence. These gaps limit practitioners' abilities to include health benefits in the assessments
of EE/RE projects or programs, or policy costs and benefits.
This study fills these gaps identified in the literature review by quantifying and presenting easy-
to-use health BPK values for a range of EE/RE types that are comparable with each other and
cover all regions in the United States. These BPK values are calculated in a similar fashion to
EPA's existing estimates of monetized public health benefits-per-ton of emissions reductions in
that both sets of estimates take health benefits and divide them by an amount of emissions or
generation reduction (Fann et al. 2009).4
4 EPA has used the benefits-per-ton estimates in multiple regulatory impact assessments for air quality regulations,
such as the Mercury and Air Toxics Standards; the New Source Performance Standards for Petroleum Refineries;
and the National Emission Standards for Hazardous Air Pollutants for Industrial, Commercial, and Institutional
Boilers and Process Heaters. For more information, see https://www.epa.gov/economic-and-cost-analvsis-air-
pollution-regulations/regulatorv-impact-analvses-air-pollution.
7
-------
In general, the literature review examined common approaches to estimating BPK values and
identified a series of best practices for estimating these values in the United States. The best
practices include:
1. Establish a set of public health BPK values for interventions in specific regions, rather
than a single national value, to account for regional differences in generation and air
pollution control technologies.
2. Establish separate BPK values for different types of EE/RE projects, programs, and
policies, such as wind, solar, uniform EE, and EE at peak, to account for how different
technologies impact the load (i.e. demand) curve.5
3. Establish BPK values for interventions of varying capacity to capture the benefits
stemming from EE/RE interventions that can displace power from baseload, intermediate
load, and peaking units.
4. Account for changes in primary and secondary PM2.5 emissions and, whenever feasible,
changes in O3 concentrations in health BPK values, to capture the majority of health
impacts from outdoor air pollution.6
5. Use emissions, population, and income datasets from the same year to maintain internal
consistency.
The BPK values included in this report are estimated using a method informed by these best
practices. EPA also sought input on the methods for this analysis from outside experts in energy
modeling, health benefits estimation, electricity system operations, and EE/RE policy and
deployment. The remainder of this report describes the methods used to estimate screening-level
BPK values and results of the analysis. The report also contains technical appendices with more
information on the tools and models used in the analysis, as well as the results of sensitivity
analyses performed to address uncertainty in the estimates.
Methods
In this section, EPA provides a general overview of the approach used to estimate the near-term
benefits per kWh of EE/RE,7 and then discusses in more detail the electricity, emissions, and
health impact modeling steps used to develop the BPK values.
Overview of Approach
EPA's approach for estimating the screening-level health benefits per kWh of EE/RE projects,
programs, and policies involves a six-step process:
1. Estimate changes in fossil-based electricity generation due to representative EE/RE
projects, programs, and policies.
2. Estimate changes in air pollution emissions (NOx, SO2, and PM2.5) due to changes in
fossil-based generation.
5 See the Energy-Efficiency Scenarios section on page 6 of this report for definitions of uniform EE and EE at peak.
6 EPA's retrospective analysis of the CAA found that approximately 85 percent of the public health benefits of air
quality regulations are due to PM reductions, rather than O3 (EPA 201 lb).
7 The "near term" is defined as approximately the next five years, which is discussed in more detail in the
Limitations section on page 15.
8
-------
3. Estimate changes in ambient concentrations of air pollution due to changes in emissions
of primary PM2.5 and precursors of secondary PM2.5.8
4. Estimate changes in public health impacts due to changes in ambient concentrations of
PM2.5.
5. Estimate the monetary value of changes in public health impacts.
6. Divide the monetized public health benefits by the change in generation to determine the
health benefits per kWh (c/kWh),
This approach follows well-established methodologies for estimating the magnitude and
economic value of public health benefits of air pollution emissions reductions, which have been
documented in the literature (e.g., Dockins et al. 2004, Fann et al. 2012) and used in recent EPA
Regulatory Impact Analyses (RIAs). Based on these established methodologies, EPA chose not
to include reductions of carbon dioxide (CO2) in this analysis because it is generally only
included in studies that assess climate and welfare impacts in addition to public health impacts.
In order to quantify public health benefits in the near-term, EPA developed a set of values for the
year 2017. To carry out the approach for these estimates, EPA used two peer-reviewed Agency
tools, the AVoided Emissions and geneRation Tool (AVERT)9 and the COBRA tool.10 Figure 1
depicts the approach outlined above as it relates to the tools used in this analysis. These tools are
described further in Appendix A: A Voided Emissions and geneRation Tool and Appendix B: Co-
Benefits Risk Assessment (COBRA) Health Impacts Screening and Mapping Tool.
»100 MW of wind
»100 MW of solar
~ 500 GWh of uniform EE
»200 GWh of EE at peak
hours (12 p.m. to 6 p.m.)
Estimate air quality
changes (primary and
secondary PM2 5)
Estimate monetized public
health benefits of changes
Estimate changes in
electricity generation
Estimate changes in
emissions of NOx S02, and
primary PM2 5
Aggregate health benefits
for each EE/RE scenario
Divide health benefits by
total electricity displaced
COBRA
AVERT
BPK Factors
^Sscenarios
Figure 1. BPK Approach.
8 Primary PM2 s refers to the direct emissions of PM from EGUs. Secondary PM2 s is created as emissions of SO: and
NOx [and other pollutants such as ammonia and volatile organic compounds (VOCs)] undergo chemical reactions in
the atmosphere.
9 EPA AVERT; see https://www.epa.gov/statelocalenergv/avoided-emissions-and-generation-tool-avert.
111 EPA COBRA Health Impacts Screening and Mapping Tool; see https://www.epa.gov/statelocalenergy/co-
benefits-risk-assessment-cobra-screening-model.
9
-------
Modeling Scenarios Development
EPA considered multiple scenarios to estimate changes in electricity generation and emissions
due to EE/RE projects, programs, and policies. During the scenario development process, EPA
sought input from technical experts in EE/RE modeling and analysis, and refined the scenarios
based on their comments. For a description of how these scenarios were used to estimate changes
in electricity generation and emissions, see the Electricity and Emissions Modeling section on
page 14, as well as Appendix A: A Voided Emissions and gene Ration Tool (AVERT).
Renewable Energy Scenarios
For RE, EPA chose to model separate scenarios for wind and solar projects. These projects have
different impacts on the timing of generation (i.e., solar only generates during the daytime while
wind can generate during more hours of the day) and may therefore have different impacts on
emissions reductions in each region. EPA modeled both the wind and solar in AVERT as 100-
megawatt (MW) projects. The assumptions EPA made in choosing this project size are discussed
in more detail in the Project, Program, and Policy Size Assumptions section (page 13).
EPA considered modeling several other RE scenarios before deciding to model 100-MW
projects. These scenarios included:
• A mix of wind and solar to estimate a portfolio of RE projects. However, EPA decided
that to estimate the benefits of a single mix of wind and solar would have limited value,
given than most states and regions have different mixes of wind and solar generation.
Therefore, EPA decided to provide separate estimates of BPK values for wind and solar
generation.
• Separate scenarios for utility-scale and rooftop solar generation, as AVERT allows the
user to model those technologies separately. However, the supply curves for these project
types are very similar and result in similar emissions reductions per kWh. Therefore, for
simplicity, EPA is reporting a single value for solar projects, which is based on a utility-
scale solar project modeled in AVERT.
Stakeholders can use the individual wind and solar values to evaluate the benefits of a mix of
wind and solar generation in a particular region. The impacts of EE/RE projects, programs, and
policies are additive as long as the additional capacity does not exceed 15 percent of fossil
generation in any given hour in a region. This cap on capacity is a limit set by EPA and is due to
the fact that AVERT is a historical dispatch model that is limited in its ability to estimate
emissions reductions for projects, programs, or policies that may significantly alter the
generation mix in a region. Capacity added beyond this 15 percent cap may have a different
impact on emissions that is not captured by the model. For more information on project size
limits when using AVERT, see the Project, Program, and Policy Size Assumptions (page 10),
Uncertainty (page 14), and Limitations (page 16) sections in this report.
En ergy-Efficien cy Seen alios
EPA developed two scenarios for EE projects, programs, and policies: uniform EE and EE at
peak. EPA modeled uniform EE in AVERT as a 500-gigawatt hour (GWh) reduction in
electricity demand, distributed evenly throughout all hours of the year. EPA modeled EE at peak
as a 200-GWh reduction distributed evenly (but exclusively) during the limited hours of 12 p.m.
10
-------
to 6 p.m. on weekdays. The assumptions EPA made in choosing this project, program, and policy
size are discussed in more detail in the Project, Program, and Policy Size Assumptions section
on page 13.
Uniform EE is based on a constant reduction in electricity demand applied evenly to all hours of
the year. This assumes that an EE intervention would reduce demand for electricity to the same
degree during all hours of the day and for all seasons. For example, installing energy-efficient
exit signs (which operate 24 hours a day, seven days a week) will result in constant or uniform
reductions, because the signs lower demand during all hours of the year. Alternatively, a mixed
portfolio of EE strategies that, taken together, saves electricity in a relatively uniform pattern
throughout the year can be viewed as a uniform EE intervention.11
The EE at peak scenario assumes that EE reductions occur only during certain times of the day
when demand is highest (often called "peak hours"). In states with warmer climates this is often
the afternoon hours in the summer, while colder states have peak hours during winter mornings;
some states have both morning and afternoon peak hours. Air conditioners are an example of a
technology that largely impacts the load curve during summer peak hours. Air-conditioning
(A/C) units often consume more electricity during peak times when people return home from
work or school. Installing an energy-efficient air conditioner is, therefore, an example of a
measure that largely affects generation during peak hours.
The types of EGUs that typically operate on the margin during peak hours often differ from those
that operate on the margin at other times of the day.12 Peaking units are generally natural gas
units that can ramp up and down quickly compared to baseload coal or nuclear units that
typically operate 24 hours a day. Because emissions from these types of power plants can vary
significantly, the reduction in emissions will likely also vary for different types of EE
interventions.13 Note that interventions that result in load reductions during the peak hours may
also result in load reductions during off-peak hours. For example, an energy-efficient A/C unit
will result in decreased demand in all hours in which it is in use, even though the largest
reductions will occur during peak hours. Nevertheless, because these types of EE interventions
result in significant load reductions during peak hours, it is useful to examine the difference in
benefits provided by load reductions during peak hours compared to those from a more uniform
load reduction.
In order to model the EE at peak scenario, it is necessary to select a window of time along the
load curve as representative of system peak. However, there is currently no universally agreed-
11 An example of how a portfolio of EE programs can save electricity relatively uniformly throughout the day is
demonstrated in the graphic on page 3 of the Southern California Edison Preferred Resources Pilot Annual Update
for 2018: https://www.sce.com/wps/wcm/con.nect/el.34c4a9-aff0-4ddf-a8a0-
cf9d5a0e3304/:» 1 ~ rHPAnnnalReport.pdf?MOD=AJPERES.
12 EPA defines EGUs on the margin as "the last units expected to be dispatched, which are most likely to be
displaced by energy efficiency or renewable energy." For more information, see chapter 3 of the EPA report,
Quantifying the Multiple Benefits of Energy Efficiency and Renewable Energy: A Guide for State and Local
Governments', https://www.epa.gov/statelocalenergv/anantifving-mnltiple-benefits-energv-efficiencv-and-
renewab le-e ne rev -guide-state
13 For example, natural gas single cycle turbines are well-suited to serve peak load because of their quick start-up
capability, but these units generally have higher NOx emissions than natural gas combined cycle plants, which are
more efficient and typically serve intermediate or even baseload demand.
11
-------
upon definition of peak hours. When electric utilities are managing the operations of existing
EGUs, they often define the peak period based on the hour of day. Utilities know that demand
tends to increase in the afternoons in the summer and early mornings/late evenings in the winter
and adjust their operations accordingly. EPA compared various definitions of the peak period to
determine which definition to use for estimating the EE at peak BPK values.
EPA reviewed definitions of peak hours from several utilities in different parts of the country
(Figure 2). The definitions of peak hours differed slightly among the utilities (e.g., some are from
2 p.m. to 6 p.m., some include morning hours, some differ by season). EPA conducted a
sensitivity analysis by modeling the same generation reduction for each utility's definition of
peak, including seasonal variations. For example, Duke Energy defines the peak period in the
winter from 6 a.m. to 9 a.m. and in the summer from 1 p.m. to 6 p.m.; while Pacific Gas and
Electric (PG&E) defines the peak period only during 1 p.m. to 7 p.m. in the summer but does not
include a peak period in the winter. The sensitivity analysis involved running scenarios for all 10
AVERT regions using the definitions of the peak period, discussed in more detail in Appendix C:
Sensitivity Analyses on Project, Program, or Policy Size and Peak Energy-Efficiency Definition.
This analysis found that the differences in the definition of peak hours do not result in large
differences in emissions reductions within each region when modeled in AVERT. Therefore,
EPA chose to use the general definition of 12 p.m. to 6 p.m. on weekdays for peak hours, as this
scenario also generated similar emissions reductions compared to the other definitions in all
regions. The results of the sensitivity analysis on the definition of peak hours are discussed in
more detail in Appendix C: Sensitivity Analyses on Project, Program, or Policy Size and Peak
Energy-Efficiency Definition.
In addition to defining the peak period based on the hour of day, it can also be defined as the top
hours of demand during the year. Utilities generally use this approach to determine whether and
when to build new capacity, because they must ensure they have enough capacity to meet even
the highest days of demand (e.g., the peak period could be based on the top 200 hours of
demand). In most regions, these high periods of demand are concentrated in the hottest summer
afternoons. By contrast, defining the peak period as 12 p.m. to 6 p.m. on weekdays includes
more than 1,500 hours during the year. EPA conducted a sensitivity analysis to compare these
definitions of the peak period by estimating emissions reductions in all 10 AVERT regions in
2017 using both a "top 200 hours approach" and an "hour of day approach" to define the peak
period. The results of this sensitivity analysis show large differences in the emissions rate in
some regions. The full results of this sensitivity analysis are discussed in Appendix C: Sensitivity
Analyses on Project, Program, or Policy Size and Peak Energy-Efficiency Definition.
After consultation with energy-sector experts, EPA ultimately determined that the hour-of-day
approach is more relevant for this analysis. Only very-specific technological interventions or EE
programs or policies would coincide with just the top 200 hours of demand, and the use of this
definition would, therefore, not accurately capture all the benefits from broader programs or
policies.
The two definitions of the peak period described above are used for different purposes by electric
utilities—the hour-of-day approach is used to manage existing capacity and the top-hours-of-
demand approach is used to plan for additional capacity. EPA asserts that most independent
developers, nonprofits, and state/local users of these BPK values will be more interested in
12
-------
capturing the impacts of an EE project, program, or policy on the existing or projected fleet of
EGUs, rather than planning for additional capacity, and therefore the Agency reports values
using the hour-of-day approach as the primary BPK values for EE at peak in this analysis.
However, if a utility is planning to use BPK values to estimate the health benefits of an EE
project, program, or policy in order to avoid investing in new generation, transmission, and
distribution, then the top-hours-of-demand approach may be more appropriate. BPK values
calculated using a top 200 hours approach are shown in Appendix D: Top 200 Honrs of Demand
Benefit-per-kWh Results.
Nevertheless, this definition of the peak period should inform how BPK values are used. If an EE
project, program, or policy results in generation reductions only during the top 200 hours of
demand, then it may have a different emissions profile and, therefore, different health benefits
than the type of EE at peak modeled here. Analysts have the option of developing their own
custom BPK estimates using AVERT and COBRA if the estimates EPA provides do not fit their
unique circumstances.
Project, Program, and Policy Size Assumptions
EPA modeled the RE projects assuming a project,
project, or policy size of 100 MW and modeled
the EE projects assuming generation reductions
of 500 GWh for uniform EE scenarios and 200
GWh for EE during peak hours. It is possible that
larger EE/RE projects could displace a different
set of EGUs, resulting in disproportionately
larger (or smaller) emissions reductions and
health benefits. To determine whether the project size would have a large effect on BPK
estimates, EPA conducted a sensitivity analysis by running AVERT with five different project
sizes, ranging from 100 MW to 2,000 MW for RE and 100 GWh to 2,000 GWh for EE. The
results from each AVERT run were entered into COBRA to estimate the health benefits. The
results from both AVERT and COBRA demonstrated strong linear relationships (R2 = 0.9996-
1.0). This means that the BPK values were nearly constant across all the project sizes tested in
the sensitivity analysis. As a result, the results presented here used a single assumption about
project size for each technology type. The full results for this sensitivity analysis are shown in
Appendix C: Sensitivity Analyses on Project, Program, or Policy Size and Peak Energy-
Efficiency Definition.
EPA chose the 100 MW and 200 and 500 GWh sizes for RE and EE projects, programs, and
policies respectively because they are large enough to generate significant emissions reductions
but small enough that they do not displace more than 15 percent of fossil generation in any given
hour. AVERT is a historical dispatch model that is limited in its ability to estimate emissions
reductions for projects, programs, or policies that may significantly alter the generation mix in a
region. EPA recommends that users avoid modeling scenarios in which the EE/RE project,
program, or policy would reduce more than 15 percent of fossil-fuel generation in any given
hour.14 The size an individual project, program, or policy can range widely before hitting that
14 In general, EE/RE impacts greater than 15 percent of regional fossil-load could influence the historical dispatch
patterns that AVERT's statistical module is based upon. AVERT should not be used to change dispatch based on
future economic or regulatory conditions, such as expected fuel prices, emissions prices, or specific emissions limits.
RE/EE Scenarios
• 100 MW of added wind capacity
• 100 MW of added solar capacity
• 500 GWh of uniform EE
• 200 GWh of EE during peak hours
(12 p.m. to 6 p.m., weekdays)
13
-------
limit, depending on the amount of fossil generation in each region. For example, in the California
region, a 400-MW solar project would exceed that limit. In the Southeast, however, a solar
project could be as large as 14,000 MW before hitting the 15-percent threshold. Table 3 lists the
15-percent thresholds in all regions for the scenarios included in this report. Furthermore, EPA
also recommends users avoid estimating emissions reductions for projects less than roughly 1
MW because the resulting emissions reductions estimated by the model are too small to be
distinguished from the underlying variation in the baseline data.
Table 3. AVERT 15-percent Threshold of Fossil-Fuel Generation in 2017
Region
Wind
(MW)
Utility Solar
(MW)
Uniform EE
(GWh)
EE at Peak
(GWh)
California
469
309
1,825
340
Great Lakes/Mid-Atlantic
6,223
9,568
36,582
9,733
Lower Midwest
1,388
1,999
8,341
2,318
Northeast
1,235
1,429
4,386
1,660
Northwest
1,191
827
4,275
843
Rocky Mountains
610
604
3,461
868
Southeast
19,465
14,496
55,084
15,111
Southwest
1,437
1,177
7,504
1,602
Texas
2,425
3,705
13,025
4,240
Upper Midwest
2,106
3,467
14,562
3,845
Electricity and Emissions Modeling
To estimate the changes in electricity generation and associated changes in emissions due to
EE/RE projects, programs, and policies (steps 1 and 2 in the overall approach), EPA used
AVERT. AVERT uses hourly emissions and generation data reported to EPA by EGUs to
determine the air pollution emissions per kWh from each generating unit, as well as the
probability that a given unit will be operating during a given hour.15 AVERT uses this
information to estimate which fossil-fired units will likely be affected by EE/RE projects,
programs, and policies; and the amount of emissions displaced or avoided. The results from
AVERT are the estimated emissions reductions of NOx, SO2, and PM2.5 from the modeled EE or
RE project, program, or policy. The results from AVERT are presented at the county, state, and
regional levels.
The 2017 estimates in this analysis were developed using actual emissions and generation of
fossil-fired EGUs in 2017, which are built into the latest version of AVERT. The assumptions
about how AVERT uses historical data to estimate emissions reductions are discussed in more
detail in Appendix A: AVoided Emissions and geneRation Tool (AVERT).
EPA developed separate estimates for each of the 10 AVERT regions (Figure 2) in order to take
into account regional differences in generation power plant fuel mixes and air pollution control
15 Facilities are required under 40 CFR Part 75 to report information on emissions, heat rate, and generation to
EPA's Clean Air Market Division (CAMD) for EGUs 25 MW or larger.
14
-------
CaHfornia
AAnf
Nunneasi
m)
technologies.16 These regions are based on the Emissions & Generation Resource Integrated
Database (eGRID) subregions and NERC regions. EPA modeled each scenario, outlined above,
in each region in 2017; 40 estimates of emissions reductions were developed.
Air Quality and Health Impact Modeling
Once EPA developed estimates of emissions reductions by applying AVERT for all scenarios,
EPA used the COBRA Health Impacts Screening and Mapping Tool to complete steps 3, 4, and
5 of the approach—estimating changes in ambient air quality, impacts on public health, and
monetized health benefits from emissions reductions, respectively.
COBRA uses a reduced-form air quality model called the Phase II Source-Receptor (S-R) Matrix
to develop screening-level estimates of how changes in emissions at source counties will affect
ambient PM2.5 concentrations in receptor counties. The S-R Matrix was developed using
multiple runs from the
Climatological Regional
Dispersion Model (CRDM), a
more sophisticated air quality
model, and it is intended as a
screening-level tool, which can
be run more quickly than the
full model. COBRA accounts
for both primary (i.e., directly
emitted) PM2.5 emissions and
the formation of secondary
PM2.5 in the atmosphere from
the reaction of SO2 and NOx
with ammonia (NH3).
COBRA also uses Fig"re 2-AVERT ReSions-
concentration-response (C-R)
functions from the epidemiological literature to determine how changes in ambient PM2.5
concentrations will impact health outcomes, such as premature mortality, non-fatal heart attacks,
asthma exacerbations, and other respiratory symptoms. Finally, COBRA uses established
valuation functions from the economic literature to estimate the monetary value of each health
outcome. C-R and valuation functions used in COBRA are consistent with those used in EPA's
Environmental Benefits Mapping and Analysis Program (BenMAP) and in RIAs conducted by
the Agency. COBRA assumes that National Ambient Air Quality Standards (NAAQS) are met in
all states and counties, and, therefore, estimates incremental health benefits from reduced
exposure below the standards.17 The result from COBRA is the estimated avoided public health
outcomes from emissions reductions and the monetary value of those avoided public health
outcomes. The results from COBRA are presented at the county level. For more information on
Northwest
(NW)
Rocky
Mountains
-------
the COBRA tool, see Appendix B: Co-Benefits Risk Assessment (COBRA) Health Impacts
Screening and Mapping Tool; for detailed information on the C-R functions used in COBRA, see
Appendix E: Health Impact Functions; and for detailed information on the valuation functions
used in COBRA, see Appendix F: Health Benefits Valuation.
For this analysis, EPA used the 2017 baseline emissions inventory housed in COBRA v3.0.
Given that AVERT uses 2017 data, EPA did not make any changes to the baseline data in either
AVERT or COBRA.
County-level emissions reductions from each AVERT run were entered into the COBRA tool.
This tool allows users to select from multiple emissions tiers, or categories of emissions sources,
in order to more accurately determine the health impacts due to reductions in emissions from that
category. COBRA takes into account the height of the smokestacks of the emissions sources in
each emissions tier, which impacts the modeled transport of pollution.18 EPA entered emissions
reductions using the tier for Fuel Combustion from Electric Utilities.
COBRA also gives users the ability to choose between a 3 percent or 7 percent discount rate that
will be used in the economic analyses completed by the model.19 Following the Agency's
Guidelines for Preparing Economic Analyses (EPA 2010), EPA ran scenarios using both the
3 percent and 7 percent discount rates. This allowed EPA to evaluate the effect of the discount
rate on monetized health benefits of EE/RE projects, programs, and policies.
For each discount rate, COBRA reports a low and high estimate of the monetary value of the
health benefits impacts, based on the use of different C-R functions (e.g., different mortality
functions). Specifically, the low and high estimates are derived using two sets of assumptions
from the literature about the sensitivity of adult mortality and non-fatal heart attacks to changes
in ambient PM2.5 levels. EPA used these low and high estimates for both the 3 percent and
7 percent discount rates to report the total health benefits of all scenarios as a range.
Developing the Health Benefits-per-kWh Estimates
AVERT presents results at the county and regional levels, whereas COBRA only presents results
at the county level. EPA aggregated the total county-level results from each COBRA scenario
and developed the monetized health BPK estimates (0/kWh) for each region and each scenario
by dividing the total monetized health benefits ($) from COBRA by the total regional-level
reduction in generation (kWh) from AVERT.
While the inputs to COBRA are based on emissions reductions occurring in each AVERT
region, the COBRA results also include health benefits that occur outside the region(s) where
modeled emissions reductions occur. This is because COBRA accounts for the transport of
pollution to air sheds located downwind of an emissions source. For example, emissions
reductions from EGUs in the Great Lakes/Mid-Atlantic region will likely result in health benefits
within that region and also in neighboring regions downwind of the power plant smokestacks,
18 For example, the highway vehicles tier assumes all emissions are at the ground level; while the electric utilities
tier assumes emissions are from taller smoke stacks, which result in the transport of pollution across farther
distances.
19 COBRA accounts for most health impacts during only the year of the analysis (i.e., 2017). However, the C-R
functions for premature mortality and nonfatal heart attacks are based on a 20-year increase in incidence. Therefore,
the benefits from avoiding these specific health impacts are discounted to determine their present value.
16
-------
such as the Northeast region, due to the interstate transport of air pollution. In the BPK
calculations, EPA aggregated the total health benefits calculated by COBRA for each scenario to
account for all of the health benefits that occur both within the AVERT region where the
emissions reductions occur, and in other regions that also experience health benefits from those
emissions reductions. This approach is consistent with other EPA estimates of monetized public
health benefits per ton of emissions reductions (Fann et al. 2009).
Screening-level health benefits per kWh of each scenario are estimated using the following
equation:
HealthBenefitstus
BPKtr = — ;—— '¦
GenerationChangetr
Where:
BPKt.r = Monetized public health benefits per kilowatt-hour (0/kWh) for
each EE/RE technology type (/) and AVERT region (r)
HealthBenefitst.us = Aggregated monetized public health benefits from emissions
reductions for each type of EE/RE technology type (I) for the
contiguous United States (us)
GenerationChangetr = Change in electricity generation for each EE/RE technology type
(t) and AVERT region (r).
Uncertainty
As described above, EPA calculated the BPK values using a suite of models that are each
affected by various sources of uncertainty. While data limitations prevent EPA from quantifying
these uncertainties, the Agency can qualitatively characterize the sources and magnitude of the
uncertainties from electricity and emissions modeling, and air quality and health impact
modeling. EPA discusses here these sources of uncertainty, as well as steps taken within the
models and by EPA to mitigate this uncertainty. This discussion also includes an assessment of
whether each source of uncertainty leads to an overestimate or underestimate of the BPK values,
where possible. In addition, this section also includes a discussion of the uncertainty over the
length of time into the future that these values can be used for analysis. EPA does not attempt to
quantify the uncertainty in the BPK values (e.g., by calculating a confidence interval around each
estimate). Readers interested in reviewing a comprehensive quantitative analysis of the
uncertainty of the impacts of PM on public health should consult the RIA for the PM NAAQS
(EPA 2013).
Uncertainty in Electricity and Emissions Modeling
EPA identified three main sources of uncertainty stemming from estimating EE/RE-related
emissions reductions using AVERT. Estimates in AVERT are calculated using a single
assumption about project size. These estimates could, therefore, be sensitive to project size, and
under- or overestimate reductions if applied to larger or smaller projects. As discussed in the
Project, Program, and Policy Size Assumptions section above on page 13, to address this
uncertainty, EPA conducted sensitivity analyses varying the project size from 100 MW to 2,000
MW of added capacity for wind and utility solar, and varying EE definitions. This analysis is
17
-------
discussed in detail in Appendix C: Sensitivity Analyses on Project, Program, or Policy Size and
Peak Energy-Efficiency Definition; and shows that changes in project size do not have a large
impact on the resulting BPK values.
Uncertainties also exist in the cohort of marginal units AVERT simulates when there are changes
in demand or RE generation within an AVERT region. The core emissions, heat rate, and
generation information AVERT uses is based on historical datasets that utilities report to EPA's
Clean Air Market Division (CAMD) for EGUs 25 MW or larger. AVERT's statistical module
uses probability distributions of how EGUs operated historically in every hour of a base year to
determine which cohort of EGUs are on the margin. Refer to Appendix A for more details on
AVERT's operations.20 Additionally, AVERT does not report results for cases that are not above
the level of reportable significance. This prevents AVERT from falsely reporting emissions
outcomes of very small EE/RE project, program, or policy impacts. For example, AVERT does
not report any emissions impacts less than 10 lbs. of a criteria air pollutant and does not report
any results less than 10 tons of CO2. Furthermore, there is some uncertainty in how the regions
are defined. Although AVERT regions are based on eGRID subregions and NERC regions, the
electricity grid is interconnected and there are transfers of electricity across regions. AVERT
does not currently account for these transfers since this could lead to isolating impacts within a
region that may affect power plants outside of the region. This could result in either an
overestimate or an underestimate of the emissions impacts, depending on which regions are
transferring electricity.
Additionally, AVERT only includes fossil fuel-generating units. However, some states, such as
California, experience a curtailment of generation from renewable sources when there is an
oversupply of electricity generation during certain hours of the year. Curtailment is defined as "a
reduction in the output of a generator from what it could otherwise produce given available
resources, typically on an involuntary basis" (Bird et al. 2014, p. 1). By assuming that only fossil
sources are displaced and not accounting for the fact that some renewable sources could be
displaced, the BPK results could overestimate the health benefits of EE/RE. For more
information on this issue, see the Limitations section on page 19.
Uncertainty in Air Quality and Health Impact Modeling
EPA identified sources of uncertainty from using COBRA to model changes in air quality, health
impacts, and the value of those impacts. The largest source of uncertainty in the COBRA tool is
the S-R Matrix, which consists of fixed transfer coefficients that reflect the relationship between
emissions at source counties and ambient air pollution concentrations at receptor locations. Even
though the S-R Matrix was developed as a screening-level tool using a more advanced model
(CDRM), it still represents a simplification of the transport of air pollution, and it is less
sophisticated than a photochemical grid model, such as the Community Multiscale Air Quality
Modeling System (CMAQ), which would quantify the non-linear chemistry governing the
formation of PM2.5 in the atmosphere. Due to the uncertainty surrounding the S-R Matrix,
COBRA is considered a screening-level tool; for more detailed estimates of air quality changes,
20 For more information on AVERT's statistical module, refer to Appendix D in the AVERT User Manual:
https://www.eDa.gov/statelocalenergY/avert-user-manual.
18
-------
more sophisticated models should be used.21 However, COBRA has been used extensively in the
peer-reviewed literature and has been compared favorably to the estimates from CALPUFF, a
more sophisticated air quality model (Levy et al. 2003). It is not clear whether the uncertainty
with the S-R Matrix leads to an overestimate or underestimate of the BPK values.
The C-R and valuation functions used in COBRA to estimate and monetize public health impacts
are another source of uncertainty. The functions used in COBRA do not represent the complete
body of epidemiological literature but are consistent with those used in recent EPA regulatory
analyses. Additionally, COBRA addresses uncertainty in some C-R functions by using
two separate approaches to estimate the incidence of mortality and nonfatal heart attacks and
reports high and low values. The valuation function that accounts for a majority of the benefits is
the value of a statistical life, which is a well-established value that has been used in many EPA
regulatory analyses.22
Uncertainty in Modeling into the Future
The baselines used in AVERT are constructed from emissions and generation data reported to
EPA for the year 2017. Estimating health benefits for future years using 2017 BPK values results
in some uncertainty. EPA suggests that AVERT should not be used to estimate emissions
reductions more than five years into the future; this limitation is discussed in the Limitations
section, below. In most cases, forecasting the electricity sector is based on assumptions about
future fuel prices, emissions constraints, electricity markets, and technological advancements, as
well as other aspects of the U.S. economic and regulatory systems. These factors can be used in
sophisticated analyses to forecast retirements and additions of EGUs and determine dispatch.
AVERT, however, does not take these factors into account, which limits its ability to forecast
changes in emissions in the future. The average emissions rates from electricity generation have
been declining over the past several years for most regions. If these trends continue, the 2017
BPK values would be an overestimate of the benefits of EE/RE in future years.
Limitations
The BPK values are subject to the same limitations as the results of the AVERT and COBRA
tools. Limitations discussed in this section include the timeframe for which the BPK values may
be used; types of projects, programs, or policies that can be evaluated; modeling limitations
regarding the curtailment of renewables; modeling limitations regarding energy storage;
pollutants that are included in the analysis; and benefits beyond the scope of the tools.
Timeframe of the BPK Values
Estimates of emissions reductions from AVERT are based on actual 2017 emissions data
reported to EPA by EGUs 25 MW or larger, while the emissions baseline in COBRA is based on
a projection for 2017. Therefore, there are limitations in using the estimates produced by these
tools to evaluate projects, programs, and policies into the future. For example, if the electricity
grid continues to get cleaner, resulting in fewer emissions per kWh of generation, the BPK
21 For more information on other more sophisticated options for modeling health benefits for energy efficiency and
renewable energy, see chapter 4 of the EPA report, Quantifying the Multiple Benefits of Energy Efficiency and
Renewable Energy: A Guide for State and Local Governments, https://www.epa. gov/statetoca le nergy /qua ntify i n
multiple-benefits-energy-efficiency-and-renewable-energy-guide-state.
22 For more information on the value of a statistical life, please see EPA's Mortality Risk Valuation web page at
https://www.epa.gov/environ.mental-econom.ics/mortality-risk-valuation.
19
-------
values would decrease. EPA recommends not using AVERT to evaluate scenarios more than
five years into the future; the BPK values have a similar limitation. The emission rates at EGUs
will likely continue to change in the coming years, in response to regulations, fuel prices, and
changes in electricity demand, such as from electric vehicles. These BPK values should therefore
not be used to estimate the benefits of EE/RE past 2022.
EPA has also explored the development of BPK values for future years. As EE/RE projects,
programs, and policies are often planned years in advance, it would be useful to have BPK
values that are based on electricity and emissions modeling projections for years after 2022 (the
limit of the 2017 values). However, EPA decided to focus on the development of the 2017 BPK
values before developing a set of future values. Future BPK values, if developed, will be based
on the most up-to-date electricity and emissions modeling that is available to EPA.
Project, Program, or Policy to Be Evaluated
EPA advises against using AVERT to estimate emissions reductions for projects that are too
small (~ 1 MW) or too big (no greater than 15 percent of regional fossil demand). The absolute
amount can differ by region but can be as low as 1,000 MW. For this reason, the BPK values will
have the same limitations in terms of the size of the project, program, or policy for which they
can be used.
In addition, as mentioned above, EPA modeled the EE at peak scenario by reducing generation
only during 12 p.m. to 6 p.m. on weekdays. If a particular EE measure reduces demand during a
very different time, such as only during the hottest days of the summer, then the benefits per
kWh may be different, as discussed in Appendix C.
Modeling Limitations Related to Curtailing Renewable Energy Generation
AVERT estimates emissions reductions resulting from the displacement of fossil fuel-generating
units by sources of EE/RE. However, the real-world dispatch of EGUs is not this simple, and as
renewables continue to be added to the electricity supply, some states are beginning to see the
curtailment of RE sources in periods of oversupply of generation. Generators are curtailed to
ensure the reliability of the grid, usually when there is more electricity generation than demand
or there is transmission congestion. Because fossil fuel units have higher marginal costs than
renewables (due to the cost of the fuel), they are typically curtailed more often than renewables.
However, in some states with a large proportion of generation from renewables, such as
California, there have been curtailments of renewables.23 Because AVERT does not model
existing RE sources, it cannot capture the potential curtailment of renewables. For this reason,
the emissions reductions and BPK values from EE/RE projects, programs, and policies may be
overestimated.
Modeling Limitations Regarding Energy Storage
AVERT currently does not include assumptions concerning energy storage. Advancements in
energy storage may make the storage of generation from renewables more viable, leading to
increased displacement of different fossil fuel-generating units at different times of the day. For
example, a solar panel generating during daylight hours could store its electricity for
23 See, for example, a factsheet on curtailments from the California Independent System Operator (ISO):
https://www.eaiso.com/DocxHnents/CurtailmentFastFacts.pdf.
20
-------
consumption during the evening hours. It is unclear whether this limitation leads to an
overestimate or underestimate of the BPK values.
Pollutants Beyond the Scope of the Tools
AVERT does not model reductions in emissions of NH3 or volatile organic compounds (VOCs)
associated with changes in electricity generation; therefore, EPA did not include changes in
emissions of these pollutants in their analysis. However, the electricity generation sector was
responsible for less than 1 percent of the NH3 and VOC emissions in the United States in 2014,
according to the National Emissions Inventory (EPA 2018). Similarly, COBRA does not
estimate the formation of O3; therefore, EPA did not examine the health impacts due to changes
in O3 concentrations. For these reasons, the BPK values may slightly underestimate the total
health benefits of emissions reductions from EE/RE projects, programs, and policies. It should be
noted that EPA's retrospective analysis of the CAA found that approximately 85 percent of the
public health benefits of air quality regulations are due to PM reductions, rather than O3
reductions (EPA 201 lb).
AVERT does model emissions of CO2; however, EPA chose not to include reductions of CO2 in
this analysis. Reductions in CO2 are generally only included in studies that assess climate and
welfare impacts in addition to public health impacts, which is beyond the scope of this study.
Although emissions of CO2 and climate change may be linked with some public health impacts,
such as increased heat stress or incidence of vector-borne diseases, COBRA does not estimate
those particular health impacts. The health impacts due to EE/RE projects, programs, and
policies and corresponding BPK values may therefore be underestimated.
Benefits Beyond the Scope of the Analysis
Finally, COBRA estimates and values health benefits due to emissions reductions, but it does not
include other types of benefits, such as avoiding damages from decreased timber and agricultural
yields, reduced visibility, accelerated depreciation of materials, and reductions in recreation
services. For this reason, the BPK values presented here may be an underestimate compared to
similar values calculated using other tools that include both health and welfare benefits, such as
the AP2 Model (Muller and Mendelsohn 2018).
U Its
In this section, EPA presents the results of the electricity and emissions modeling, as well as the
BPK values for 2017.
Emissions Reductions
EPA's AVERT was used to estimate changes in fossil-generated electricity and emissions
reductions from EE/RE projects, programs, and policies. AVERT outputs used in this analysis
include displaced generation (MWh) and emissions reductions of SO2, NOx, and PM2.5 (tons).
Complete regional-level outputs from AVERT can be found in Appendix G: DetailedBenefits-
per-kWh Results.
On average, the SO2 emissions reductions from EE/RE are approximately 0.85 lbs. per megawatt
hour (MWh), with large regional variation. In general, the regional variation in emissions
reductions is greater than the variation across EE/RE technology types. The California region has
the smallest reduction in SO2 emissions per MWh for all types of EE/RE projects, programs, and
21
-------
policies (Figure 3). In 2017, the largest reduction in SO2 emissions per MWh for all types of
EE/RE occurred in the Upper Midwest region. In six of the ten regions, wind projects delivered
the largest SO2 reductions per MWh. EE at peak projects show the largest SO2 reductions per
MWh in the Northeast and Southeast, and solar projects deliver the largest reductions in the
Great Lakes/Mid-Atlantic and Pacific Northwest regions. EE at peak projects resulted in the
lowest SO2 reductions per MWh in six of the ten regions.
2017 S02 Avoided Emissions Rate
2.0
1.8
1.6
^ 1.4
5 1.2
1.0
£ 0.8
_Q
-1 0.6
0.4
0.2
0.0
h
.O g. £-
IS?
^ ^ JF JP „V
^ $
& J" ^
gj" 4?
& V°
i
Uniform EE
I EE at Peak
I Solar
Wind
Average
Figure 3. Avoided S02 Emissions Rates for EE/RE Projects, Programs, and Policies in 10 AVERT Regions in
2017.
There is also substantial regional variation in the NOx avoided emissions rates, with an average
of 0.91 lbs./MWh in 2017. The California region again has the smallest reduction in NOx
emissions per MWh; and the Rocky Mountain region sees the largest reduction in emissions for
all types of EE/RE projects, programs, and policies, except EE at Peak (Figure 4). EE at peak
projects result in the largest NOx emissions reduction per MWh in the majority of the regions,
with wind and solar projects delivering the largest reductions in the others. It also appears SO2
and NOx avoided emissions rates may have an inverse relationship. In all regions where the SO2
avoided emissions rates are below average, the NOx avoided emissions rates are higher than the
SO2 rate; and in all regions where the SO2 avoided emissions rates are above average, the NOx
avoided emissions rates are lower than the SO2 rate.
22
-------
2017 NOv Avoided Emissions Rate
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
& <3- 4-
^
^ _v£"
y ^
* . J> /• /• J J>
-------
Ben efits-per-kWh Values
The county-level emissions reductions from AVERT were entered into the appropriate counties
of the COBRA tool to estimate the health benefits of each EE/RE scenario. These benefits reflect
the sum of the PM2.5 benefits from the changes in electric sector emissions of NOx, SO2, and
PM2.5 and reflect the range based on adult mortality functions (e.g., Krewski et al. 2009, Lepeule
et al. 2012). The total health benefits from COBRA for each scenario were divided by the
corresponding displaced generation values in each region as estimated by AVERT in order to
calculate benefits per kWh. Values were calculated for low and high estimates using both 3
percent and 7 percent discount rates; however, only the 3 percent results are discussed in the
main body of this report, as the 7 percent results have the same general trends. The low and high
estimates are derived in COBRA using two different C-R functions from the literature to
estimate the sensitivity of adult mortality and non-fatal heart attacks to changes in ambient PM2.5
levels.24 A detailed results table, including values calculated using a 7 percent discount rate, can
be found in Appendix G: DetailedBenefits-per-kWh Results. COBRA reports results in 2017
U.S. dollars (USD).
EE/RE projects, programs, and policies in California deliver the lowest public health benefits per
kWh in all scenarios (Figure 6).The largest benefits per kWh can be seen in the Great
Lakes/Mid-Atlantic region, followed by the Upper Midwest. Regions such as the Pacific
Northwest, Rocky Mountains, and Southwest, which had low SO2 but high NOx avoided
emissions rates, have lower-than-average benefits per kWh. This is due in part to the fact that
SO2 converts to secondary PM in the atmosphere more readily than NOx, and therefore results in
more health impacts per ton than NOx. The Northeast region values are of note, as EE at peak
and solar projects deliver above-average benefits per kWh, despite having below-average SO2,
NOx, and PM2.5 emissions rates. A full list of EPA's 2017 BPK values can be found in Table 4.
24 More information about the C-R functions used in COBRA can be found in Appendix E: Health Impact
Functions.
24
-------
2017 Benefits-per-kWh (3% Discount Rate, Low Estimate)
4.000
3.500
3.000
o 2.500
rvl
2.000
> 1.500
o 1.000
0.500
0.000
2017 Benefits-per-kWh (3% Discount Rate, High Estimate)
9.00
8.00
„ 7.00
£i 6.00
o
OL 5.00
€ 4.00
^ 3.00
° 2.00
1.00
0.00
¦
_<<>s ^ J? &
&
* f /
v°
& J? J? & J? J
# ^ ^ ^ ^
* o*°
-------
Table 4. 2017 Benefits-per-kWh Values (2017 USD)
Region
Project Type
3% Discount Rate
7% Discount Rate
2017 c/kWh
(low estimate)
2017 ^/kWh
(high estimate)
2017 c/kWh
(low estimate)
2017 ^/kWh
(high estimate)
Calitomia
Uniform EE
0.48
1.08
0.42
0.96
EE at Peak
0.52
1.17
0.46
1.04
Solar
0.51
1.15
0.45
1.03
Wind
0.48
1.09
0.43
0.97
Great Lakes/Mid-Atlantic
Uniform EE
3.51
7.95
3.14
7.09
EE at Peak
3.57
8.08
3.19
7.21
Solar
3.67
8.29
3.27
7.39
Wind
3.35
7.59
2.99
6.77
Lower Midwest
Uniform EE
2.31
5.23
2.06
4.66
EE at Peak
2.11
4.77
1.88
4.25
Solar
2.19
4.96
1.96
4.42
Wind
2.35
5.32
2.10
4.74
Northeast
Uniform EE
1.65
3.73
1.47
3.33
EE at Peak
2.24
5.07
2.00
4.52
Solar
1.94
4.38
1.73
3.91
Wind
1.58
3.56
1.41
3.18
Pacitic Northwest
Uniform EE
1.13
2.55
1.01
2.28
EE at Peak
1.12
2.54
1.00
2.27
Solar
1.17
2.64
1.04
2.35
Wind
1.13
2.55
1.01
2.27
Rocky Mountains
Uniform EE
1.03
2.32
0.92
2.07
EE at Peak
0.98
2.21
0.87
1.98
Solar
0.99
2.25
0.89
2.01
Wind
1.07
2.41
0.95
2.15
Southeast
Uniform EE
1.78
4.02
1.58
3.58
EE at Peak
1.87
4.24
1.67
3.78
Solar
1.83
4.15
1.64
3.70
Wind
1.76
3.98
1.57
3.55
Southwest
Uniform EE
0.71
1.62
0.64
1.44
EE at Peak
0.70
1.59
0.63
1.42
Solar
0.73
1.64
0.65
1.46
Wind
0.77
1.73
0.68
1.54
Texas
Uniform EE
1.58
3.58
1.41
3.19
EE at Peak
1.39
3.13
1.24
2.80
Solar
1.42
3.22
1.27
2.87
Wind
1.63
3.69
1.45
3.29
Upper Midwest
Uniform EE
3.12
7.06
2.78
6.30
EE at Peak
2.75
6.22
2.45
5.55
Solar
2.89
6.53
2.58
5.83
Wind
3.20
7.23
2.85
6.45
26
-------
Discussion
The BPK values represent estimates of the monetized annual public health benefits resulting
from emissions reductions associated with EE/RE projects, programs, and policies. There are
different values for each combination of region and EE/RE intervention type (i.e., wind, solar,
uniform EE, and EE at peak). It should be noted that the total benefits from EE/RE projects,
programs, and policies in any region will include health benefits both within and outside of that
region.
The results show that there are larger differences in benefits per kWh across regions than across
EE/RE technologies. This is likely due to differences in the fossil fuel mix used for generation
across regions. For example, California has low BPK values because its generation comes
largely from natural gas, which has low emissions rates. These emissions rates are similar
regardless of the EE/RE technology displacing the fossil generation. However, in other regions
such as the Northeast, there is more variation across technology types. In the case of the
Northeast, the fossil generation operating during the peak period has higher emissions rates than
the generation operating during other times of the day. Therefore, EE at peak and solar power,
which displace generation during the daytime peak hours, have higher benefits per kWh than
wind or uniform EE, which displace generation in more hours of the day. However, emissions
are only one factor in the estimation of BPK values. The estimated health benefits are also
affected by the population of the areas impacted by the emissions reductions. Areas with more
people affected by changes in air quality will have a greater cumulative health benefit. For
example, the Southwest has higher NOx and PM2.5 emissions rates than the Northeast, although
both regions have similar SO2 rates. However, the Northeast has larger benefits per kWh for all
technology types; this is due in part to the Northeast's higher population density relative to the
Southwest.
In most cases though, the regional variation in BPK values is driven by differences in both
population and emissions rates. For example, the Upper Midwest has higher avoided SO2 and
NOx emissions rates compared to the Great Lakes/Mid-Atlantic region in 2017. However, the
Great Lakes region has 5 to 30 percent higher BPK values compared to the Upper Midwest.
There are several possible reasons for this, including that the avoided PM2.5 emissions rates in
the Great Lakes/Mid-Atlantic region are approximately double those in the Upper Midwest, and
that the Great Lakes/Mid-Atlantic region is more densely populated than the Upper Midwest
region.
The BPK values presented here are similar in magnitude to values reported in the literature.
McCubbin and Sovacool (2013) found that wind generation in California between 1987 and
2006 delivered 0,40/kWh to 4,70/kWh in health benefits. EPA's low California results are
similar to these results, approximately 0.40/kWh, but the high estimate (1.10/kWh) is more than
double McCubbin and Sovacool's (2013) estimate. Buonocore et al. (2016) examined EE/RE
benefits in New Jersey and Maryland, an area similar to EPA's Great Lakes/Mid-Atlantic region.
Again, EPA's results are similar, but slightly lower for all technology types, except EE at peak,
compared to those seen in the literature review (Table 5).
27
-------
Table 5. Comparison of EPA and Buonocore et al. (2016) Benefits-per-kWh Values
Project Type
Buonocore et al. Results for
New Jersey and Maryland
(2012 c/kWh)'
EPA Results for
Great Lakes/Mid-Atlantic Region
(2017 ^/kWh)
Uniform EEa
9.4-15
3.1-7.9
EE at Peak3
1.4-10
3.2-8.1
Solar
6.3-15
3.3-8.3
Wind
8.1-17
3.0-7.6
" Referred to as baseload or peak demand side management (DSM) in the Buonocore et al. study.
In addition to being similar to other BPK values from the literature, EPA's results are similar to
the cost of EE/RE projects, programs, and policies. This suggests that the health benefits of
EE/RE projects, programs, and policies can help offset all or part of the cost of these
technologies. According to a study by the LBL (2015), the average cost of "saved electricity" or
EE is 0.0460/kWh (Figure 7; 2012 USD). EPA's estimates for the benefits of EE projects range
from 0.40/kWh (1,20/kWh) in California to 4.00/kWh (8.00/kWh) in the Great Lakes/Mid-
Atlantic region using the low and high estimates.
The BPK values are also largely similar to the cost of new RE capacity. According to Lazard's
annual Levelized Cost of Energy Analysis, the cost of wind energy is between 3,00/kWh and
6,00/kWh, and the cost of utility scale solar is between 4,30/kWh and 5,30/kWh (Lazard 2017).
According to EPA's results, the average benefits per kWh for both wind and solar projects are
approximately 1,60/kWh and 3,70/kWh for the low and high values, respectively. Therefore,
without considering any of the other non-health benefits of EE/RE technologies, up to half of the
costs of wind and solar projects could be covered by EPA's low health benefit estimates, and
nearly all of the costs could be covered by EPA's high health benefit estimates. For some
regions, the health benefits of EE/RE entirely outweigh their costs. For example, the high
estimates, using both the 3 and 7 percent discount rates, for wind projects in the Great
Lakes/Mid-Atlantic region are greater than Lazard's (2017) levelized costs; the high estimate
BPK values in the Upper Midwest are also larger than the levelized costs for wind energy.
Similarly, the BPK values (high estimates, 3 percent discount rate) for solar projects in the Lower
Midwest, Great Lakes/Mid-Atlantic, and Upper Midwest regions are greater than or equal to the
higher end of the levelized costs.
28
-------
MA : so ore
Hi
VT I so 055
WA so 05-
S0.Q51
R3 | 1^^^H|^H$oo5o
CA I
SO MT
OK ^^^^^¦!SC045
j».o«
so 027
PA ^ U.S. Average ¦ Participant Co£
AZ I SC roe SO 046*Wtl Of Saved Etectncity
M0 1 so 035 B Program Administrator Cosi
ME oi Saves Electnoly
NM 1
$0.00 $0.01 $002 $0.03 30.04 $0.05 $0.06 S0.07 £0.08 £0.09
Lewliied Tout Cost of Saved Electricity (2012SfltWh)
Figure 7. Levelized Cost of EE Programs by State. Source: LBL 2015.
By generating these health benefits per kWh values for EE/RE, EPA hopes to address the gap in
the literature and provide health BPK values that cover all regions in the United States and cover
key EE/RE project, program, and policy types. Such health benefits estimates may have several
uses. For example, state public utility commissions (PUCs) and state energy offices (SEOs) may
use estimates of the monetized public health benefits of EE as an input to portfolio-level, cost-
benefit analyses; or program-specific, cost-effectiveness tests. Policymakers or financial
institutions could also use these estimates to develop a fuller accounting of the benefits of
investments in EE/RE. Finally, EE/RE developers, state and local public health administrators,
NGOs, and the general public can use these estimates to quantify the public health benefits of
existing or proposed EE/RE projects, programs, and policies. Please note that this is not an
exhaustive list of uses for BPK values. Furthermore, because the BPK values provide a
screening-level estimate, they may not be appropriate for certain analyses, such as federal air
quality rulemaking.
In addition, as discussed in the Limitations section on page 16, one area of additional research
includes developing BPK values for future years. Such values would be based on modeling the
electricity sector to estimate emissions rates in future years and would allow for the projection of
benefits from EE/RE projects, programs, and policies in years beyond 2022 (the current limit of
the 2017 values).
Conclusions
State and local governments are increasingly interested in quantifying the public health value of
emissions reductions from EE/RE so that they can fully reflect these benefits in policy decision-
making processes. Some studies have quantified the benefits but have used different approaches
29
-------
and assumptions, making it difficult for others to adopt or credibly compare the health benefits
estimates on a per-kWh basis.
EPA has developed regional-level BPK screening values to further these analyses and fill the gap
for this type of analysis in the literature. By using the AVERT and COBRA tools, EPA
developed regional BPK values for uniform EE, EE at peak, wind, and solar projects, programs,
and policies, which incorporate the benefits of SO2, NOx, and PM2.5 emissions reductions.
Although results vary by region, on average, EPA found that EE/RE programs delivered benefits
of 1.70/kWh to 3,90/kWh in the United States in 2017 (using a 3 percent discount rate).
EPA believes that these health benefit screening values may be useful to a wide range of
stakeholders seeking to estimate the public health benefits of EE/RE projects, programs, and
policies, including state PUCs, SEOs, policymakers, financial institutions, EE/RE developers,
state and local public health administrators, NGOs, and the general public.
References
Babin, S. M., H. S. Burkom, R. S. Holtry, N. R. Tabernero, L. D. Stokes, J. O. Davies-Cole, K.
DeHaan, and D. H. Lee. 2007. Pediatric patient asthma-related emergency department
visits and admissions in Washington, DC, from 2001-2004, and associations with air
quality, socio-economic status and age group. Environmental Health, 6, 9.
Barbose, G., R. Wiser, J. Heeter, T. Mai, L. Bird, M. Bolinger, A. Carpenter, G. Heath,
D. Keyser, J. Macknick, A. Mills, and D. Millstein. 2016. A retrospective analysis of
benefits and impacts of US renewable portfolio standards. Energy Policy 96:645-660.
Bell, M. L., K. Ebisu, R. D. Peng, J. Walker, J. Samet, S. L. Zeger, and F. Dominici. 2008.
Seasonal and Regional Short-term Effects of Fine Particles on Hospital Admissions in
202 US Counties, 1999-2005. American Journal of Epidemiology, 168(11), 1301-1310.
Bird, L., J. Cochran, and X. Wand. 2014. Wind and Solar Energy Curtailament: Experience and
Practice in the United States. National Renewable Energy Laboratory, Golden, CO.
Buonocore, J., P. Lucklow, G. Norris, J.D. Spengler, B. Biewald, J. Fisher, and J. Levy. 2016.
Health and climate benefits of different energy-efficiency and renewable energy choices.
Nature Climate Change 6:100-105.
Dockery, D. W., J. Cunningham, A. I. Damokosh, L. M. Neas, J. D. Spengler, P. Koutrakis, and
F. E. Speizer. 1996. Health Effects of Acid Aerosols On North American Children -
Respiratory Symptoms. Environmental Health Perspectives, 104(5), 500-505.
Dockins, C.K., K. Maguire, N. Simon, and M. Sullivan. 2004. Value of Statistical Life Analysis
and Environmental Policy: A White Paper. U.S. Environmental Protection Agency,
National Center for Environmental Economics, Golden, CO.
EPA. 2010. Guidelines for Preparing Economic Analyses. U.S. Environmental Protection
Agency, Washington, DC. Available: https://www.epa.gov/environmental-
economics/guidelines-preparing-economic-analvses.
30
-------
EPA. 2011a. Regulatory Impact Analysis for the Final Mercury and Air Toxics Standards.
U.S. Environmental Protection Agency. Available:
https ://www3. epa. gov/ttnecas 1 /regdata/RI As/matsriafinal .pdf.
EPA. 201 lb. The Benefits and Costs of the Clean Air Act from 1990 to 2020. Final Report -
Rev. A. April. U.S. Environmental Protection Agency Office of Air and Radiation.
Available: https://www.epa.gov/sites/production/files/2Q15-
07/documents/fullreport rev a.pdf.
EPA. 2013. Regulatory Impact Analysis for the Final Revisions to the National Ambient Air
Quality Standards for Particulate Matter. U.S. Environmental Protection Agency.
Available: https://www3.epa.gOv/ttnecasl/regdata/R.IA.s/finalria.pdf.
EPA. 2015. Technical Support Document: Preparation of Emissions Inventories for the
Version 6.2, 2011 Emisisons Modeling Platform. U.S. Environmental Protection Agency.
Available: https://www.epa.gov/sites/production/files/2015-
10/documents/201 1 „ _v'1 _aV5 ''taismod tsd aug2015.pdf.
EPA. 2017. Estimating Public Health Benefits of Energy Efficiency and Clean or Renewable
Energy Interventions: Technical Support for Benefits- and Emissions-Related Software
Models. Prepared by Abt Associates for U.S. Environmental Protection Agency, Office
of Atmospheric Programs.
EPA. 2018. National Emissions Inventory, 2014, Version 2. U.S. Environmental Protection
Agency. Available: https://www.epa.gov/air-emissions-inventories/2014-national-
emissions-inventory-nei-data.
Fann, N., K.R. Baker, and C.M. Fulcher. 2012. Characterizing the PM2.5-related health benefits
of emission reductions for 17 industrial, area and mobile emission sectors across the U.S.
Environment International 49:141-151.
Fann, N., C.M. Fulcher, and 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
Quality, Atmosphere & Health 2:69-176.
Glad, J.A., L.L. Brink, E.O. Talbott, P.C. Lee, X. Xu, M. Saul, and J. Rager. 2012. The
Relationship of Ambient Ozone and PM2.5 Levels and Asthma Emergency Department
Visits: Possible Influence of Gender and Ethnicity. Archives of Environmental &
Occupational Health. Vol 62 (2): 103- 108.
Health Effects Institute. 2018. State of Global Air 2018. Special Report. Health Effects Institute,
Boston, MA. Available: http://www.stateofglobalair.org/sites/default/files/soga-2018-
report.pdf.
IEI. 2017. Energy Efficiency Trends in the Electric Power Industry. Institute for Electric
Innovation. Available:
http://www.edisonfoimdation.net/iei/piiblications/Documents/IEI Energy%20Efficiency
-------
Kloog, I., B.A. Coull, A. Zanobetti, P. Koutrakis, and J.D. Schwartz. 2012. Acute and Chronic
Effects of Particles on Hospital Admissions in New-England. PLoS ONE. Vol 7 (4): 1-8.
Krewski, D., M. Jerrett, R. T. Burnett, R. Ma, E. Hughes, Y. Shi . . ., and B. Tempalski. 2009.
Extended follow-up and spatial analysis of the American Cancer Society study linking
particulate air pollution and mortality. Research Reports Health Effects Institute (140), 5-
114; discussion 115-136.
Lazard. 2017. Lazard's Levelized Cost of Energy Analysis. Version 11.0. Unsubsidized
Levelized Cost of Energy Comparison. Available:
https://www.lazard.com/perspective/levelized-cost-of-energy-!
LBL. 2015. The Total Cost of Saving Electricity through Utility Customer-Funded Energy
Efficiency Programs: Estimates at the National, State, Sector and Program Level.
Lawrence Berkley National Lab. Available: https://emp.lbl.gov/sites/all/files/total-cost-
of-saved-energy.pdf.
Lepeule J., F. Laden, D. Dockery, and J. Schwartz. 2012. Chronic exposure to fine particles and
mortality: an extended follow-up of the Harvard Six Cities study from 1974 to 2009.
Environmental Health Perspectives, 120(7), 965-970.
Levy, J.I., A.M. Wilson, J.S. Evans, and J.D. Spengler. 2003. Estimation of primary and
secondary particulate matter intake fractions for power plants in Georgia. Environ Sci
Technol 37:5528-5536.
Mai, T., R. Wiser, G. Barbose, L. Bird, J. Heeter, D. Keysey, V. Krishnan, J. Macknick, and
D. Millstein. 2016. A Prospective Analysis of the Costs, Benefits, and Impacts of
US Renewable Portfolio Standards. National Renewable Energy Laboratory and
Lawrence Berkeley National Laboratory, Golden, CO and Berkeley, CA.
Mar, T. F., T. V. Larson, R. A. Stier, C. Claiborn, and J. Q. Koenig. 2004. An analysis of the
association between respiratory symptoms in subjects with asthma and daily air pollution
in Spokane, Washington. Inhalation Toxicology, 16(13), 809-815.
Mar, T. F., J. Q. Koenig, and J. Primomo. 2010. Associations between asthma emergency visits
and particulate matter sources, including diesel emissions from stationary generators in
Tacoma, Washington. Inhalation Toxicology, 22(6), 445-448.
McCubbin, D. and B. Sovacool. 2013. Quantifying the health and environmental benefits of
wind power to natural gas. Energy Policy 53:429-441.
Millstein, D., R. Wiser, M. Bolinger, and G. Barbose. 2017. The climate and air-quality benefits
of wind and solar power in the United States. Nature Energy 2:17134.
Moolgavkar, S. H. 2000a. Air Pollution and Hospital Admissions for Chronic Obstructive
Pulmonary Disease in Three Metropolitan Areas in the United States. Inhalation
Toxicology, 12(Supplement 4), 75-90.
Moolgavkar, S. H. 2000b. Air pollution and hospital admissions for diseases of the circulatory
system in three U.S. metropolitan areas. Journal of the Air & Waste Management
Association, 50(7), 1199-1206.
32
-------
Muller, N.Z. and R. Mendelsohn. 2018. The Air Pollution Emission Experiments and Policy
Analysis Model (APEEP) Technical Appendix. Yale University School of Forestry and
Environmental Studies, New Haven, CT. Available:
https://www.researchgate.net/profile/Mcholas Muller/publication/253359043 The Air
Pollution Emission Experiments and Polk lysis Model APEEP Technical App
en.dix/links/0Jc t726274000000/The-Air-Pollution-Emission-Experiments-and-
Policv-Analysis-Mode inical-Appendix.pdf.
Ostro, B. D. 1987. Air Pollution and Morbidity Revisited: A Specification Test. Journal of
Environmental Economics and Management, 14, 87-98.
Ostro, B. D., and S. Rothschild. 1989. Air Pollution and Acute Respiratory Morbidity - an
Observational Study of Multiple Pollutants. Environ Res, 50(2), 238-247.
Ostro, B., M. Lipsett, J. Mann, H. Braxton-Owens, and M. White. 2001. Air pollution and
exacerbation of asthma in African-American children in Los Angeles. Epidemiology,
12(2), 200-208.
Peng, R. D., H. H. Chang, M. L. Bell, A. McDermott, S. L. Zeger, J. M. Samet, and F. Dominici.
2008. Coarse particulate matter air pollution and hospital admissions for cardiovascular
and respiratory diseases among Medicare patients. Jama, 299(18), 2172-2179.
Peng, R. D., M. L. Bell, A. S. Geyh, A. McDermott, S. L. Zeger, J. M. Samet, and F. Dominici.
2009. Emergency admissions for cardiovascular and respiratory diseases and the
chemical composition of fine particle air pollution. Environ Health Perspectives, 117(6),
957-963.
Peters, A., D. W. Dockery, J. E. Muller, and M.A. Mittleman. 2001. Increased particulate air
pollution and the triggering of myocardial infarction. Circulation, 103(23), 2810-2815.
Pope, C. A., D. W. Dockery, J. D. Spengler, and M. E. Raizenne. 1991. Respiratory Health and
PmlO Pollution - a Daily Time Series Analysis. American Review of Respiratory Disease,
144(3), 668- 674.
Pope, C. A., J. B. Muhlestein 3rd, H. T. May, D. G. Renlund, J. L. Anderson, and B. D. Home.
2006. Ischemic heart disease events triggered by short-term exposure to fine particulate
air pollution. Circulation, 114(23), 2443-2448.
Russell, M. W., D. M. Huse, S. Drowns, E. C. 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.
Schwartz, J., and L. M. Neas. 2000. Fine particles are more strongly associated than coarse
particles with acute respiratory health effects in schoolchildren. Epidemiology, 11(1), 6-
10.
Sheppard, L. 2003. Ambient Air Pollution and Nonelderly Asthma Hospital Admissions in
Seattle, Washington, 1987-1994. In: Revised Analyses of Time-Series Studies of Air
Pollution and Health (pp. 227-230). Boston, MA: Health Effects Institute.
33
-------
Slaughter, J. C., E. Kim, L. Sheppard, J. H. Sullivan, T. V. Larson, and C. Claiborn. 2005.
Association between particulate matter and emergency room visits, hospital admissions
and mortality in Spokane, Washington. J Expo Anal Environ Epidemiology, 15(2), 153-
159.
Sullivan, J., L. Sheppard, A. Schreuder, N. Ishikawa, D. Siscovick, and J. Kaufman. 2005.
Relation between short-term fine-particulate matter exposure and onset of myocardial
infarction. Epidemiology, 16(1), 41-48.
Wiser, R., M. Bolinger, G. Heath, D. Keyser, E. Lantz, J. Macknick, T. Mai, and D. Millstein.
2016a. Long-term implications of sustained wind power growth in the United States:
Potential benefits and secondary impacts. Applied Energy 179:146-158.
Wiser, R., T. Mai, D. Millstein, J. Macknick, A. Carpenter, S. Cohen, W. Cole, B. Frew, and
G. Heath. 2016b. On the Path to SunShot: the Environmental and Public Health Benefits
of Achieving High Penetrations of Solar Energy in the United States. National Renewable
Energy Laboratory, Golden, CO.
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.
Woodruff, T. J., J. Grillo, and K.C. Schoendorf. 1997. The relationship between selected causes
of post neonatal infant mortality and particulate air pollution in the United States.
Environmental Health Perspectives, 105(6), 608-612.
Zanobetti, A., and J. Schwartz. 2006. Air pollution and emergency admissions in Boston, MA.
Journal of Epidemiology & Community Health, 60(10), 890-895. doi: 60/10/890 [piil
10.1136/jech.2005.039834 [doi]
Zanobetti, A., M. Franklin, and J. Schwartz. 2009. Fine particulate air pollution and its
components in association with cause-specific emergency admissions. Environmental
Health 8, 58-60.
34
-------
Appendix A: AVoided tmissions and geneRation Tool (AVERT)
AVERT analyzes changes in fossil-fired electricity generation from solar, wind, and EE
programs in 10 unique regions of the continental United States (Figure A - l).23 The AVERT
regions take into account the
fact that customers' electricity
demand is met jointly by
generation resources throughout
a region, rather than from a
single power plant.26 AVERT
provides estimates of changes in
NOx, SO2, PM2.5, and CO2
emissions at the regional, state,
and county levels.
In AVERT, the impacts on
emissions from wind and solar
electricity generation are
modeled using the annual
electricity generation capacity in
MWs of the renewable project.
AVERT uses these capacity inputs to estimate the amount of electricity generation (in megawatt
hours) the project(s) would produce. Capacities can be entered separately for wind and two types
of solar installations: utility-scale and rooftop.
AVERT uses hourly data reported to EPA's CAMD by EGU. Data are available from 2007 to
2017. These data include gross generation; steam output; heat input; and emissions of SO2, NOx,
and CO2. Hourly emissions of PM2.5 are calculated using data from the National Emissions
Inventory.
AVERT uses hourly data on NO v, SO2, and CO2 emissions to estimate the impact of EE/RE
projects, programs, and policies on emissions. AVERT uses the hourly generation data to
determine the probability of whether a particular unit will be operating in a given hour of the
year. The tool also uses hourly emissions data to estimate the emissions from electricity
generation from that unit. AVERT provides built-in assumptions about the capacity factors of RE
technologies to estimate the annual amount of generation an RE project will produce, and the
likely hours in which it will be operating.27 For example, AVERT uses data from the National
Renewable Energy Laboratory to estimate the likely hours of the year a solar project would
generate electricity in each region. Users are able to develop their own site- or region-specific
25 Although in some regions solar or hvdroelectricity may be on the margin, AVERT assumes they are must-take
resources and fossil-fired electricity generators are the only generators affected by increased EE/RE.
16 Note that while there are imports and exports of electricity across regions, AVERT does not explicitly model these
transfers.
® AVERT reflects regional capacity factors for renewable generation, based on actual wind projects from AWS
Truepower and solar projects modeled in the National Renewable Energy Laboratory's PV Watts tool, reflecting the
availability of sun and wind resources in each region. See Appendix C of AVERT's user manual for details.
Northwest
(NW)
California
(CA)
Rocky
Mountains
(RM)
Upper
Midwest
(WMW)
Lower
Midwest
Southeast
(SE)
Figure A -1. AVERT Regions
35
-------
renewable energy load profiles for use in AVERT; however, this study used the built-in capacity
factor assumptions. For EE projects, programs, and policies, the hours of the year they reduce
electricity demand can be input directly by the user or it can be based on the top hours of demand
in each region.
AVERT then determines which fossil units would likely be operating during the hours that the
EE/RE project, program, or policy is operating or reducing demand, to determine the units that
would be displaced by the EE/RE project, program, or policy. AVERT estimates the emissions
reductions that would occur as a result of that displacement based on the emissions rate at each
unit. The resulting estimated reductions in generation and emissions are reported at the county,
state, and regional levels.
36
-------
Li'-fi"^ '.o-Benefits Risk A;.-- -nvrii »l' -L-l*' ¦ ! I • i -
Screening and Mapping Too!
COBRA v3.0 includes preloaded projected emissions baselines for 2017, which is estimated
using data from EPA's 2011 Version 6.2 Air Emissions Modeling Platform (2011 v6.2 platform).
Emissions from the electric generating sector in the 2011 v6.2 platform are projections of
emissions in 2017 from the Integrated Planning Model (IPM) Power Sector Modeling Platform
(v5.14). The air emissions platform also contains emissions projections from other sources
besides EGUs, such as nonpoint sources, mobile sources, fires, and other point sources. EPA has
used the emissions modeling platform for several recent air pollution rules, including the Final
2015 NAAQS for O3, the 2011 National Air Toxics Assessment (NATA), and the proposed
update to the Cross-State Air Pollution Rule (CSAPR). The 2017 emissions baselines contain
projected emissions that reflect federal and state measures (promulgated or under
reconsideration) as of December 2014, including:
• The CSAPR,
• A Federal regulatory measure for achieving the 1997 NAAQS for ozone and fine
particles,
• The Mercury and Air Toxics Standards (MATS),
• Actions EPA had taken to implement the Regional Haze Rule,
• The Cooling Water Intakes Rule [316(b)],
• The disposal of Coal Combustion Residuals (CCR) from Electric Utilities Rule, and
• State regulations in place as of December 2014.
The assumptions underlying the emissions inventories are detailed in the Technical Support
Document: Preparation of Emissions Inventories for the Version 6.2, 2011 Emissions Modeling
Platform (EPA 2015).
COBRA also includes a reduced-form air quality model, the Phase II S-R Matrix, to estimate
how changes in air pollution emissions impact ambient air quality. The S-R Matrix is based on
the Climatological Regional Dispersion Model (CRDM) and consists of fixed-transfer
coefficients that reflect the relationship between emissions at source counties and ambient air
pollution concentrations at receptor locations. To calculate the pollutant concentration at a
destination county, transfer coefficients are used in the following equation:
D/ = II^^xFSxf™t
i c
Where:
Df =
Concentration of pollutant 5 at destination county j (ug/m3)
Esi,c =
Emission of pollutant 5 from emissions category c in source county i (tons/year)
Ts . . —
1 C,l,J —
Transfer coefficient for pollutant 5 from source county i to destination county j
from emissions category c (sec/m3)
p
Ionic conversion factor for pollutant 5
37
-------
Funit = Unit conversion factor (28,778 |ig-year/ton-sec)
Ionic conversion factors used in the equation above are molecular weight ratios. These are used
to adjust the transfer coefficients to reflect the concentration of precursors to secondarily formed
particulate species. Standard molecular weights and ionic conversion factors are listed in Table B
- 1 and Table B - 2.
Table B -1. Standard Molecular Weights
Species
Symbol
Standard Molecular Weight
Nitrate Ion
NO3-
62.0049
Sulfate Ion
S042"
96.0626
Bisulfate
HSO4
97.07054
Sulfur Dioxide
SO2
64.0638
Nitrogen Dioxide
NO 2
46.055
Ammonia
NH3
17.03052
Ammonium Ion
nh4+
18.03846
Ammonium Nitrate
NH4N03
80.04336
Ammonium Bisulfate
NH4HS04
115.109
Ammonium Sulfate
(NH4)2S04
132.13952
Table B - 2. Ionic Conversion Factors
Species
Ionic Conversion Factors
PM2.5, Secondary Organic Aerosols
1
SO2 -> SO42-
96.0626/64.0638
no2^no3-
62.0049/46.0055
nh3^nh4+
18.03846/17.03052
COBRA accounts for the formation of secondary PM2.5 from NOx and SO2 emissions through
atmospheric chemistry and air pollution transport.28- 29 COBRA focuses only on primary and
secondary PM2.5, and it does not currently estimate the formation of other pollutants such as O3.
Secondary PM2.5 is formed when sulfate (SO42") and nitrate (NO3") ions react with ammonium
(NH4+) to form ammonium bisulfate (NH4HSO4), ammonium sulfate [(NH4)2S04], and
ammonium nitrate (NH4NO3). In COBRA NH4+reacts first with SO42" to form NH4HSO4 and
(NH4)2S04. If any NH4+remains, it then reacts with NO3" to form NH4NO3. As this method is
simpler than the modeling completed using more sophisticated air quality models, COBRA
results are also calibrated to measured PM2.5 concentration data obtained from EPA for 2011.
Again due to the uncertainty surrounding the S-R Matrix, COBRA is treated as a screening-level
tool.
28 The ambient pollution in a given area is a result of local and upwind pollutant emissions. Winds can transport
pollutants across state and regional boundaries, so emissions reductions in one region often affect air quality and
human health in downwind regions.
29 For more information about the S-R Matrix used by COBRA, see the User's Manual for the COBRA Health
Impact Screening and Mapping Tool, Appendix A (https://www.epa.gov/statelocalenergy/users-manual-co-benefits-
risk-assessment-cobra-screening-model).
38
-------
Once COBRA estimates the changes in PM2.5 concentrations at the county level, it then uses C-R
functions to determine the change in public health impacts from a change in ambient air quality.
The C-R functions embedded in COBRA are taken from epidemiological studies; and are
consistent with the methods used by EPA to estimate the health impacts of air pollution rules,
including MATS.30 The output of these functions is the number of avoided premature deaths,
heart attacks, hospital admissions for respiratory and cardiovascular-related illnesses, incidences
of acute bronchitis, upper and lower respiratory symptoms, asthma exacerbations or emergency
room visits, minor restricted activity days, and illness-related work loss days. See Appendix E for
a list of the epidemiological studies and more information about the C-R function used in
COBRA.
Finally, COBRA applies estimates of the value of avoiding public health impacts to determine
the monetary benefits associated with reductions in air pollution. Values used in COBRA were
used in recent EPA RIAs, including analyses for the rule mentioned above. They were derived
using a variety of methods that estimate how much people are willing to pay to reduce the risk of
a health incident or the cost of illness (COI), which includes direct medical costs and opportunity
costs.31 The value of avoiding premature adult mortality, also known as the value of a statistical
life (VSL), is generally responsible for more than 95 percent of the monetized benefits of
emissions reductions. The VSL used in COBRA to estimate the value of avoided adult mortality
ranges from approximately $7.5 million to $8.4 million (in 2010 USD), assuming a discount rate
of seven percent and three percent, respectively. This VSL value, based on 26 published studies,
is identical to the values used by EPA in regulatory analyses of air pollution rules. The value of
other health impacts, such as non-fatal heart attacks, hospitalizations, and asthma exacerbations,
are smaller and based on the COI. For example, the value of non-fatal heart attacks ranges
between $31,446 and $263,795, and the value of hospital admissions ranges between $15,430
and $41,002 per incident. See Appendix F for a complete list of the values used in COBRA.
30 For a complete list of recent RIAs of EPA air pollution rules, see https://www.epa.gov/economic-and~cost-
analvsis-air-pollution-regulations/regulatorv-impact-analvses-air-pollution. Many of these analyses use a benefits-
per-ton approach, developed by EPA (Fann et al. 2012). COBRA uses most of the same C-R functions as those used
in the benefits-per-ton approach. For a list and description of the epidemiological studies used by COBRA to
estimate adverse health effects, see the User's Manual for the COBRA Health Impact Screening and Mapping Tool,
Appendix C (https://www.epa.gov/statelocalenergv/users~manual~co~benefits~risk~assessment~cobra~screening~
model').
31 For more information about the economic values used by COBRA to estimate the economic value of avoiding
adverse health effects and how they were derived, see the User's Manual for the COBRA Health Impact Screening
and Mapping Tool, Appendix F (https://www.epa.gov/statelocalenergv/nsers~mannal~co~benefits~risk~assessment~
cobra-screening-model').
39
-------
Li'-fi"'/ v-n-iii -ityAnaK •' >ri I I K"l! irVi. '! I-' 'h'" •• 'Ml'"!
jfinition
EPA conducted sensitivity analyses using AVERT and the COBRA tool to determine the extent
to which modeling scenario assumptions might impact the BPK results. EPA analyzed
two different types of potential sensitivity: the size of the EE/RE project, program, or policy
studied; and the definition of EE at peak.
Sensitivity Analysis on Project, Program, or Policy Size Assumptions
EPA examined the potential sensitivity of the BPK values to assumptions about project size by
modeling BPK values for five different project sizes: from 100 MW to 2,000 MW added
capacity for the wind and utility solar modeling options in AVERT, and from 100 GWh to
2,000 GWh of displaced generation for the EE modeling options.
The results of these model runs illustrate that there is a strong linear relationship between project
size and emissions reductions (R2 = 0.9996-1.0, Figure C - 1). The results from AVERT were
then input into COBRA to assess the sensitivity of emissions reductions on health impacts. These
results also show that there is a strong linear relationship between the amount of emissions
reductions and health impacts (Figure C - 2).
The results of this sensitivity analysis indicate that the project size does not have a large impact
on the marginal BPK results (i.e., a larger project does not generate disproportionately larger
marginal benefits or have a higher BPK result than a smaller project). The resulting BPK values
from these model runs with different project sizes are consistent with this; for each region and
project, program, or policy type modeled, the results are within 0.10 per kWh (Table C - 1). As a
result, this analysis presents BPK values modeled using only a single assumption about project
size.
Note, however, an extremely large EE/RE project, program, or policy could displace more than
the marginal EGUs and extend into the baseload units, which may have a different emissions
profile. See the Limitations section on page 16 of this report for more information about the
limitations on project size for which the BPK values should be used.
40
-------
California, Solar
1.600.000
R- = 0.999
1,400,000
1.200.000
1,000.000
800.000
600.000
400.000
200.000
1000 1500
Project size (MW)
» NOx A PM25
California, Wind
1,200,000
1,000,000
800,000
600,000
400,000
200,000
1000 1500
Project size (MW)
~ S02 ¦ NOx x PM25
California, Uniform EE
Northeast, Solar
800,000
JS 700,000
500.000
C 500,000
-o 4OQ00C
x:
¦g 200,000
E 100.000
2.000,000
R = 0.9998
L5GD DOC
¦o 1.000.000
a
500.000
R -0.9999
1000 1500
Project size (GWh)
~ S02 "NOx ±PM25
1000 1500
Project size (MW)
~ S02 ¦ NOx , PM25
Northeast, Wind
Northeast, Uniform EE
1.600,000
JS 1.400,000
1,200,000
r 1,000,000
see. ore
600.000
400.000
200,000
1,200.000
1.000.000
800.000
600,000
400,000
:cc xe
1000 1500
Project size (MW)
¦ NOx APM25
1000 1500
Project size (GWh)
• S02 ¦ NOx a PM25
Southeast, Solar
Southeast, Wind
6.000.000
5.CCC.0CC
4.000,000
3.000.000
2.000,000
1.CCC.0C0
4.000.000
JS 3,500,000
3,000,000
« 2,500,000
¦o 2.000.000
1.500.000
S 1,000,000
500.000
500 1000 1500
Project size (MW)
~ S02 ¦ NOx *PM25
1000 1500
Project size (MW)
i NOx PM25
Southeast, Uniform EE
3.500
3.000,
: see
2.000.
1 .SCO
1.C00.
1000 1500
Project size (GWh)
~ S02 ¦ NOx * PM25
Figure C -1. AVERT Sensitivity for Project, Program, or Policy Size.
-------
California, Solar
= -
1000 1500
Project size |MW)
~ Health Benefits (low) ¦ Health Benefits (high)
California, Wind
1000 1500
Project size |MW)
~ Health Benefits (low) ¦ Health Benefits (high)
25
California, Uniform EE
Northeast, Solar
R" =0.99
£ 10
R =09233-—
1000 1500
Project size (GWh)
~ Health Benefits (low) ¦ Health Benefits (high)
1000 1500
Project size (MW)
~ Health Benefits (low) ¦ Health Benefits (high)
Northeast, Wind
Northeast, Uniform EE
Ic 100
swm
R3 =0_9SS9-^
500 1000 1500 2000
Project size (MW)
~ Health Benefits (tew) ¦ Health Benefits (high)
1000 1500
Project size (GWh)
~ Health Benefits (tow) ¦ Health Benefits (high)
Southeast, Solar
Southeast, Wind
500 1000 1500 2000
Project size (MW)
~ Health Benefits (tow) ¦ Health Benefits (high)
1000 1500
Project size (MW)
~ Health Benefits (tow) ¦ Health Benefits (high)
Southeast, Uniform EE
e 5C
1000 1500
Project size (GWh)
~ Health Benefits (tow) ¦ Health Benefits (high)
Figure C - 2. COBRA Sensitivity Analysis for Project Size.
42
-------
Table C -1. Results from Sensitivity Analysis on Project, Program, or Policy Size.
Emissions Reductions (tons) from
AVERT
Health Benefits (million USD)
from COBRA
Benefits per kWh (c/kWh)
Region
Project Type
Capacity
(MW/GWh)
Displaced
Generation
(MWh)
SO2
NOx
PM2.5
Health Benefits
(low)
Health Benefits
(high)
Low Estimate
High Estimate
100
120,370
84
56
6
1.67
3.77
1.4
3.1
500
602,150
418
281
30
8.33
18.86
1.4
3.1
Wind
1,000
1,204,500
837
562
60
16.60
37.58
1.4
3.1
1,500
1,806,580
1,256
842
91
24.85
56.25
1.4
3.1
2,000
2,408,940
1,676
1,124
121
33.06
74.83
1.4
3.1
100
169,440
121
90
9
2.33
5.28
1.4
3.1
500
847,250
601
449
46
11.52
26.08
1.4
3.1
Southeast
Solar
1,000
1,694,380
1,205
897
92
22.96
51.98
1.4
3.1
1,500
2,541,750
1,807
1,342
137
34.27
77.57
1.3
3.1
2,000
3,388,780
2,408
1,788
183
45.52
103.04
1.3
3.0
100
104,950
72
51
5
1.40
3.17
1.3
3.0
500
524,940
359
257
27
6.99
15.81
1.3
3.0
Uniform EE
1,000
1,049,980
716
514
54
13.90
31.47
1.3
3.0
1,500
1,575,040
1,073
771
82
20.79
47.06
1.3
3.0
2,000
2,099,990
1,432
1,027
109
27.64
62.57
1.3
3.0
100
152,050
5
26
3
0.75
1.69
0.5
1.1
500
761,630
25
129
15
3.55
8.02
0.5
1.1
Wind
1,000
1,522,830
50
257
30
6.93
15.67
0.5
1.0
1,500
2,284,090
75
386
45
10.35
23.39
0.5
1.0
2,000
3,044,890
99
514
60
13.75
31.09
0.5
1.0
100
194,640
6
36
4
1.04
2.34
0.5
1.2
500
971,730
31
174
19
4.84
10.94
0.5
1.1
California
Solar
1,000
1,945,550
62
346
39
9.51
21.50
0.5
1.1
1,500
2,923,700
93
523
59
14.26
32.22
0.5
1.1
2,000
3,899,550
126
704
79
18.98
42.93
0.5
1.1
100
104,510
3
19
2
0.56
1.27
0.5
1.2
500
522,680
17
94
10
2.65
5.99
0.5
1.1
Uniform EE
1,000
1,045,830
34
187
21
5.13
11.59
0.5
1.1
1,500
1,568,940
51
279
31
7.59
17.16
0.5
1.1
2,000
2,091,230
68
369
41
10.02
22.66
0.5
1.1
43
-------
Emissions Reductions (tons) from
AVERT
Health Benefits (million USD)
from COBRA
Benefits per kWh (c/kWh)
Region
Project Type
Capacity
(MW/GWh)
Displaced
Generation
(MWh)
SO2
NOx
PM2.5
Health Benefits
(low)
Health Benefits
(high)
Low Estimate
High Estimate
100
174,470
29
37
3
2.72
6.14
1.6
3.5
500
873,200
141
187
17
13.37
30.20
1.5
3.5
Wind
1,000
1,748,100
275
369
35
26.24
59.26
1.5
3.4
1,500
2,620,800
407
549
52
38.81
87.64
1.5
3.3
2,000
3,495,010
537
727
69
51.37
116.02
1.5
3.3
100
157,170
32
46
4
3.01
6.72
1.9
4.3
500
787,140
157
227
19
14.83
33.50
1.9
4.3
Northeast
Solar
1,000
1,573,340
306
448
39
29.42
66.45
1.9
4.2
1,500
2,361,630
449
660
58
43.65
98.56
1.8
4.2
2,000
3,146,030
590
869
77
57.51
129.88
1.8
4.1
100
104,880
18
25
2
1.72
3.91
1.6
3.7
500
524,150
88
126
11
8.57
19.36
1.6
3.7
Uniform EE
1,000
1,048,680
175
252
23
16.99
38.36
1.6
3.7
1,500
1,573,550
262
377
34
25.31
57.15
1.6
3.6
2,000
2,098,790
347
501
45
33.58
75.85
1.6
3.6
44
-------
Sensitivity Analysis on Definition of Peak Energy Efficiency
As discussed in the main text of this report, EPA considered two different definitions of the peak
period to model EE at peak projects, programs, and policies. One approach is based on defining
the peak period as certain hours of the day. The other approach is based on defining peak as the
top hours of demand during the year (e.g., the top 200 hours with the highest demand).
EPA conducted two sensitivity analyses related to the definition of peak. The first examined the
difference in emissions reductions based on using different hours of the day as the peak period.
This sensitivity analysis modeled the same total generation reduction but spread through
different hours of the day, including seasonal variations in some cases. Different hours of the day
and seasonal variations were taken from the definitions of the peak period used by five electric
utilities in different parts of the country (Figure C - 3). After modeling the definitions with
AVERT, EPA plotted the resulting avoided emissions rates to determine whether there were
large differences in emissions reductions based on differences in the hours of the day defined as
the peak period. This sensitivity analysis was conducted for all AVERT regions. The results
show that over the course of a year, there are only slight differences in avoided emissions rates in
most regions due to differences in the hours of the day and seasons defined as the peak period
(Figure C - 4). In some of the PG&E scenarios larger differences in avoided emissions rates can
be seen, but this may be attributable to the fact that PG&E was the only utility to define peak
hours as only occurring during the summer months. Uniform EE rates are included as a point of
reference but were not used to determine the final EE at peak scenario. As a result, EPA used a
single composite definition of 12 p.m. to 6 p.m. on weekdays as the definition of the peak period
for modeling purposes in this analysis.
Hour of the Day
Entity Type State Season 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Duke Energy (Apr. 1 - Sept. 30)
Utility NC
Summer
Duke Energy (Oct. 1- Mar. 31)
Utility NC
Winter
PG&E (May 1-Oct. 31)
Utility CA
Summer
PG&E (Nov. 1- Apr. 30)*
Utility CA
Winter
Entergy Texas (May 1- Oct. 31)
Utility TX
Summer
Entergy Texas (Nov. 1- Apr. 30)
Utility TX
Winter
Northern States Power
Utility MN
Year Round
PublicService Co. of Colorado
Utility CO
Year Round
_
_
_
_
_
_
_
_
__
__
__
__
*PG&E currently only has summer peak hours
= Peak
= Off Peak
Figure C - 3. Definitions of Peak Hours from Different Entities in the Electric Sector.
EPA also conducted a sensitivity analysis to determine the difference in emissions reductions
using an hour-of-day approach to define the peak period compared to using a top-hours-of-
demand approach. In this case, EPA modeled the same generation reduction, but spread it
differently in different hours of the year. In the hour-of-day approach, EPA reduced generation
only during 12 p.m. to 6 p.m. on weekdays. In the top-hours-of-demand approach, EPA used the
same total generation reduction but spread the reductions only to the top 200 hours of demand.
The results show large differences in many regions in the emissions reductions resulting from the
same amount of generation reduction, depending on whether the hour-of-day approach or top-
hours-of-demand approach was used to define the peak period (Figure C - 5). For example, in the
Northeast, using the top-hours-of-demand approach results in much higher emissions reductions
compared to the hour-of-day approach. This is likely due to the use of distillate oil backup units
that are used in the Northeast during periods of high demand. When the generation reductions are
45
-------
confined only to this period, it affects only these high-emitting units. Nevertheless, as discussed
in the report, EPA chose to use the hour-of-day approach to define the peak period, as EPA
determined it to be the more relevant definition for most EE/RE projects, programs, and policies
based on expected uses for the BPK values.
46
-------
S02 Avoided Emission Rates
2.5
2.0
J 1.5
« 1.0
0.5
0.0
#• J1 J? ^ &
^ J? ^ ^
\L>
/" * s y * r /
e?X JF v°
I Duke Energy
I Entergy Texas
Northern States Power
I PG&E
I Public Svc. Co. of Colorado
I Uniform EE
I Time of Day (12 p.m. - 6 p.m.]
©•
NCX Avoided Emission Rates
J? J- & j? & J?
^ ^ ^ KS^
I Duke Energy
I Entergy Texas
Northern States Power
I PG&E
I Public Svc. Co. of Colorado
I Uniform EE
I Time of Day (12 p.m. - 6 p.m.]
0.25
0.20
0.15
0.10
0.05
0.00
PM2 5 Avoided Emission Rates
II
IIih II..II
1 ¦¦ IIIIIIII
I Duke Energy
I Entergy Texas
Northern States Power
I PG&E
¦ Public Svc. Co. of Colorado
^ J? & <& ^ .„+* J?
& XT
<5- ti> -O <5" <5" <
^ J? t0<* J? ^ / JP
y s s s s s / s / e
^ ^ ^ ¦ Time of Day (12 p.m. - 6 p.m.]
^ *
of
&
Figure C - 4. Results of Sensitivity Analysis of Definition of Peak Period Based on Different Hours of the Day.
47
-------
S02 Avoided Emissions Rates
2.5
2
1.5
1
0.5
0
California
I
Great
Lakes/Mid
Atlantic
III.I.!.
Lower
Midwest
Northeast
ll.l
Pacific
Northwest
Rocky
Mountains
Southeast
Southwest
Texas
Upper
Midwest
NCX Avoided Emissions Rates
California
Northeast
Pacific
Northwest
Rocky
Mountains
Southeast
Southwest
Great
Lakes/Mid
Atlantic
Lower
Midwest
Upper
Midwest
PM2 5 Avoided Emissions Rates
0.20
| 0.15
^ 0.10
5 0.05
0.00
E
n
I
i
l
mm
California
Great
Lakes/Mid
Atlantic
Lower
Midwest
Northeast
Pacific
Northwest
Rocky
Mountains
Southeast
Southwest
Texas
Upper
Midwest
Figure C - 5. Results of Sensitivity Analysis Comparing Emissions Reductions Using Hour-of-Day Approach and Top-Hours-
of-Demand Approach to Define the Peak Period.
48
-------
Appendix D: Top 200 Hours of Demand Benefit-per-kWh Results
Table D - 1 includes the complete modeling results from AVERT and COBRA used to calculate the BPK values for the top 200 hours of demand
analysis in each region.
Table D -1. Complete AVERT and COBRA Results for Top 200 Hours of Demand Analysis (3 percent and 7 percent discount rate; 2017 USD)
Region
Discount
Rate
Results from AVERT
SO2
Emissions
Rate
(lb./MWh)
NOx
Emissions
Rate
(lb./MWh)
PM2.5
Emissions
Rate
(lb./MWh)
Results from COBRA
c/kWh
(low)
c/kWh
(high)
Displaced
Generation
(MWh)
SO2
Reduced
(lbs.)
NOx
Reduced
(lbs.)
PM2.5
Reduced
(lbs.)
$ Total Health
Benefits (low)
$ Total Health
Benefits
(high)
California
3
200,230
3,680
33,130
9,530
0.01838
0.16546
0.04760
1,868,183.33
4,221,243.69
0.93
2.11
Great Lakes/Mid-Atlantic
3
205,510
217,960
233,420
35,760
1.06058
1.13581
0.17401
7,353,520.30
16,631,254.33
3.58
8.09
Lower Midwest
3
203,670
3,080
373,040
16,210
0.01512
1.83159
0.07959
1,679,175.59
3,798,562.75
0.82
1.87
Northeast
3
197,440
171,450
210,820
15,640
0.86837
1.06777
0.07921
9,242,207.78
20,874,650.58
4.68
10.57
Pacific Northwest
3
202,330
173,090
228,080
18,200
0.85548
1.12727
0.08995
2,198,711.54
4,972,898.14
1.09
2.46
Rocky Mountains
3
195,720
63,550
226,500
11,870
0.32470
1.15727
0.06065
1,602,727.29
3,625,354.74
0.82
1.85
Southeast
3
201,440
152,400
248,990
24,350
0.75655
1.23605
0.12088
4,045,381.73
9,155,691.94
2.01
4.55
Southwest
3
193,640
7,450
265,600
13,160
0.03847
1.37162
0.06796
1,398,221.15
3,163,872.51
0.72
1.63
Texas
3
197,530
59,330
261,410
13,400
0.30036
1.32339
0.06784
2,243,773.58
5,075,140.16
1.14
2.57
Upper Midwest
3
205,770
133,580
256,210
20,200
0.64917
1.24513
0.09817
3,150,193.28
7,124,723.56
1.53
3.46
California
7
200,230
3,680
33,130
9,530
0.01838
0.16546
0.04760
1,667,429.97
3,765,217.37
0.83
1.88
Great Lakes/Mid-Atlantic
7
205,510
217,960
233,420
35,760
1.06058
1.13581
0.17401
6,561,493.57
14,833,891.17
3.19
7.22
Lower Midwest
7
203,670
3,080
373,040
16,210
0.01512
1.83159
0.07959
1,498,471.30
3,388,096.40
0.74
1.66
Northeast
7
197,440
171,450
210,820
15,640
0.86837
1.06777
0.07921
8,248,584.90
18,620,340.06
4.18
9.43
Pacific Northwest
7
202,330
173,090
228,080
18,200
0.85548
1.12727
0.08995
1,962,089.91
4,435,443.96
0.97
2.19
Rocky Mountains
7
195,720
63,550
226,500
11,870
0.32470
1.15727
0.06065
1,430,318.98
3,233,616.56
0.73
1.65
Southeast
7
201,440
152,400
248,990
24,350
0.75655
1.23605
0.12088
3,609,761.12
8,166,235.14
1.79
4.05
Southwest
7
193,640
7,450
265,600
13,160
0.03847
1.37162
0.06796
1,247,815.35
2,821,961.89
0.64
1.46
Texas
7
197,530
59,330
261,410
13,400
0.30036
1.32339
0.06784
2,002,718.23
4,527,067.37
1.01
2.29
Upper Midwest
7
205,770
133,580
256,210
20,200
0.64917
1.24513
0.09817
2,811,049.78
6,354,838.97
1.37
3.09
49
-------
Appendix E: Health Impact Functions
The health impact functions in the COBRA model were prepared by Abt Associates in close
consultation with EPA and rely on an up-to-date assessment of the published scientific literature
to ascertain the relationship between ambient PM2 5 concentrations and adverse human health
effects. Table E - 1 summarizes the key values from the epidemiological studies in COBRA used
to estimate adverse health impacts of PM2.5. The output of each health impact function is the
estimated number of incidences of each health outcome given a change in air pollution
concentrations.
Total results in COBRA and in this report are reported for a low and high estimate of health
impacts, which is a result of multiple C-R functions being used to calculate mortality and
nonfatal heart attacks. The high estimate uses the Lepeule et al. (2012) mortality estimate and the
Peters et al. (2001) non-fatal heart attack estimates. The low estimate uses the Krewski et al.
(2009) mortality estimates and the remaining four acute myocardial infarction estimates. See
Appendix C of the COBRA User's Manual for more information.
Table E -1. Key Health Impact Values in COBRA
Endpoint
Author
Age
Location
Metric
Beta
Standard
Error
Functional
Form
Mortality, All Cause
Krewski et al. (2009)
30-99
116 U.S. cities
Annual
0.005827
0.000963
Log-linear
Mortality, All Cause
Lepeule et al. (2012)
25-99
6 eastern cities
Annual
0.013103
0.003347
Log-linear
Mortality, All Cause
Woodruff et al.
(1997)
Infant
86 cities
Annual
0.003922
0.001221
Logistic
Acute Myocardial
Infarction, Nonfatal
Peters et al. (2001)
18-99
Boston, MA
24-hour
average
0.024121
0.009285
Logistic
Acute Myocardial
Infarction, Nonfatal
Pope et al. (2006)
18-99
Greater Salt
Lake City, Utah
24-hour
average
0.00481
0.001992
Logistic
Acute Myocardial
Infarction, Nonfatal
Sullivan et al. (2005)
18-99
King Comity,
Washington
24-hour
average
0.001980
0.002241
Logistic
Acute Myocardial
Infarction, Nonfatal
Zanobetti and
Schwartz (2006)
18-99
Greater Boston
area
24-hour
average
0.005300
0.002213
Logistic
Acute Myocardial
Infarction, Nonfatal
Zanobetti et al.
(2009)
18-99
26 U.S.
communities
24-hour
average
0.00225
0.000592
Log-linear
Hospital Admissions, All
Cardiovascular (less
myocardial infarctions)
Bell et al. (2008)
65-99
202 U.S.
counties
24-hour
average
0.0008
0.00011
Log-linear
Hospital Admissions, All
Cardiovascular (less
myocardial infarctions)
Moolgavkar (2000b)
18-64
Los Angeles,
CA
24-hour
average
0.0014
0.00034
Log-linear
Hospital Admissions, All
Cardiovascular (less
myocardial infarctions)
Peng et al. (2008)
65-99
108 U.S.
counties
24-hour
average
0.00071
0.00013
Log-linear
Hospital Admissions, All
Cardiovascular (less
myocardial infarctions)
Peng et al. (2009)
65-99
119 U.S. urban
counties
24-hour
average
0.00068
0.00021
Log-linear
Hospital Admissions, All
Cardiovascular (less
myocardial infarctions)
Zanobetti et al.
(2009)
65-99
26 U.S.
communities
24-hour
average
0.00189
0.00028
Log-linear
Hospital Admissions, All
Respiratory
Zanobetti et al.
(2009)
65-99
26 U.S.
communities
24-hour
average
0.00207
0.00045
Log-linear
50
-------
Endpoint
Author
Age
Location
Metric
Beta
Standard
Error
Functional
Form
Hospital Admissions, All
Respiratory
Kloogetal. (2012)
65-99
New England
area (6 states)
24-hour
average
0.0007
0.00096
Log-linear
Hospital Admissions,
Asthma
Babin et al. (2007)
0-17
Washington,
DC
24-hour
average
0.002
0.00434
Log-linear
Hospital Admissions,
Asthma
Sheppard (2003)
0-17
Seattle, WA
24-hour
average
0.00332
0.00104
Log-linear
Hospital Admissions,
Chronic Lung Disease
Moolgavkar (2000a)
18-64
Los Angeles,
CA
24-hour
average
0.0022
0.00073
Log-linear
Emergency Room Visits,
Asthma
Mar et al. (2010)
0-99
Greater
Tacoma,
Washington
24-hour
average
0.0056
0.0021
Log-linear
Emergency Room Visits,
Asthma
Slaughter et al.
(2005)
0-99
Spokane,
Washington
24-hour
average
0.0029
0.0027
Log-linear
Emergency Room Visits,
Asthma
Glad et al. (2012)
0-99
Pittsburgh, PA
24-hour
average
0.0039
0.0028
Logistic
Acute Bronchitis
Dockery et al. (1996)
8-12
24 communities
Annual
0.027212
0.017096
Logistic
Asthma Exacerbation,
Cough
Mar et al. (2004)
6-18
Vancouver,
CAN
24-hour
average
0.01906
0.009828
Logistic
Asthma Exacerbation,
Cough
Ostro et al. (2001)
6-18
Los Angeles,
CA
24-hour
average
0.000985
0.000747
Logistic
Asthma Exacerbation,
Shortness of Breath
Mar et al. (2004)
6-18
Vancouver,
CAN
24-hour
average
0.01222
0.013849
Logistic
Asthma Exacerbation,
Shortness of Breath
Ostro et al. (2001)
6-18
Los Angeles,
CA
24-hour
average
0.002565
0.001335
Logistic
Asthma Exacerbation,
Wheeze
Ostro et al. (2001)
6-18
Los Angeles,
CA
24-hour
average
0.001942
0.000803
Logistic
Minor Restricted Activity
Days
Ostro and Rothschild
(1989)
18-64
Nationwide
24-hour
average
0.007410
0.000700
Log-linear
Lower Respiratory
Symptoms
Schwartz and Neas
(2000)
7-14
6 U.S. cities
24-hour
average
0.019012
0.006005
Logistic
Upper Respiratory
Symptoms
Pope et al. (1991)
9-11
Utah Valley
24-hour
average
0.0036
0.0015
Logistic
Work Loss Days
Ostro (1987)
18-64
Nationwide
24-hour
average
0.004600
0.000360
Log-linear
51
-------
Appendix F: Health Benefits Valuation
Table F - 1 presents the mean estimate of the unit values used in COBRA to estimate the
monetary value of the health effects. The unit values are based on published estimates of the
costs of treating the illness (which can include both direct medical costs and costs of lost
productivity), or the willingness-to-pay (WTP) to avoid the illness or to reduce the risk of
premature death (i.e., VSL). The unit values based on WTP estimates reflect the expected growth
in real income over time. This is consistent with economic theory, which argues that WTP for
most goods (such as health risk reductions) will increase if real incomes increase. See Appendix
F of the COBRA User's Manual for more information.
Table F -1. COBRA Value of Health Effects
Health Endpoint
Age
Range
Unit Value (2017 USD at the 2017 income level)
3% Discount Rate
7% Discount Rate
Mortality3
25-99
$9,447,115
$8,414,395
Infant Mortality15
0-0
$10,529,882
$10,529,882
Acute Myocardial Infarction, Nonfatal0
0-24
$37,250
$35,220
Acute Myocardial Infarction, Nonfatal0
25^14
$50,495
$47,077
Acute Myocardial Infarction, Nonfatal0
45-54
$56,772
$52,696
Acute Myocardial Infarction, Nonfatal0
55-64
$150,083
$136,238
Acute Myocardial Infarction, Nonfatal0
65-99
$37,250
$35,220
Acute Myocardial Infarction, Nonfatal"1
0-24
$182,617
$182,617
Acute Myocardial Infarction, Nonfatal"1
25^14
$195,861
$194,475
Acute Myocardial Infarction, Nonfatal"1
45-54
$202,138
$200,094
Acute Myocardial Infarction, Nonfatal"1
55-64
$295,450
$283,637
Acute Myocardial Infarction, Nonfatal"1
65-99
$182,617
$182,617
Hospital Admissions, All Cardiovascular (less-acute
myocardinal infraction)
18-64
$45,922
$45,922
Hospital Admissions, All Cardiovascular (less-acute
myocardinal infraction)
65-99
$43,252
$43,252
Hospital Admissions, All Respiratory
65-99
$36,621
$36,621
Hospital Admissions, Asthma
0-17
$17,282
$17,282
Hospital Admissions, Chronic Lung Disease
18-64
$22,791
$22,791
Asthma Emergency Room Visits (Smith et al. 1997)
0-99
$520
$520
Asthma Emergency Room Visits (Stanford et al. 1999)
0-99
$435
$435
Acute Bronchitis
8-12
$534
$534
Lower Respiratory Symptoms
7-14
$24
$24
Upper Respiratory Symptoms
9-11
$37
$37
Minor Restricted Activity Days
18-64
$76
$76
Work Loss Days
18-64
$179
$179
Asthma Exacerbation (cough, shortness of breath, or
wheeze)
6-18
$64
$64
a Mortality value after adjustment for 20-year lag.
b Infant mortality value is not adjusted for 20-year lag.
c Based on Russell (1998).
d Based on Wittels (1990).
52
-------
Appendix G: Detailed Benefits-per-kWh Results
Table G - 1 includes the complete modeling results from AVERT and COBRA used to calculate the BPK values for each region and technology
type.
Table G -1. Complete AVERT and COBRA Results for 2017 (3 percent and 7 percent discount rate; 2017 USD)
Region
Project
Type
Discount
Rate
Results from AVERT
so2
Emissions
Rate
Ob./MWh)
NOx
Emissions
Rate
Ob./MWh)
pm25
Emissions
Rate
Ob./MWh)
Results from COBRA
c/k\Vh
0ow)
c/k\Vh
(high)
Displaced
Generation
(MWh)
so2
Reduced
Obs.)
NOx
Reduced
Obs.)
pm25
Reduced
Obs.)
S Total Health
Benefits
0ow)
S Total Health
Benefits
(high)
California
Uniform EE
3
522.060
37.410
165.980
23.090
0.07166
0.31793
0.04423
2.484.934
5.617.248
0.48
1.08
EE at Peak
3
200.130
13.110
57.400
8.620
0.06551
0.28681
0.04307
1.036.707
2.343.226
0.52
1.17
Solar
3
194.380
13.780
62.970
8.940
0.07089
0.32395
0.04599
990.413
2.238.698
0.51
1.15
Wind
3
151.660
11.150
49.200
6.720
0.07352
0.32441
0.04431
727.998
1.645.588
0.48
1.09
Great Lakes/
Mid-Atlantic
Uniform EE
3
521.980
606.820
478.260
102.050
1.16253
0.91624
0.19551
18.347.102
41.496.455
3.51
7.95
EE at Peak
3
198.470
228.410
192.760
38.100
1.15085
0.97123
0.19197
7.094.665
16.046.296
3.57
8.08
Solar
3
153.580
184.740
144.290
30.320
1.20289
0.93951
0.19742
5.629.211
12.731.861
3.67
8.29
Wind
3
226.120
243.150
210.750
44.120
1.07531
0.93203
0.19512
7.585.426
17.156.015
3.35
7.59
Lower
Midwest
Uniform EE
3
526.240
919.370
588.070
48.430
1.74705
1.11749
0.09203
12.162.120
27.515.418
2.31
5.23
EE at Peak
3
199.500
305.300
239.080
18.030
1.53033
1.19840
0.09038
4.204.249
9.511.507
2.11
4.77
Solar
3
188.940
303.910
220.330
17.280
1.60850
1.16614
0.09146
4.140.826
9.368.042
2.19
4.96
Wind
3
352.100
632.180
386.340
32.440
1.79546
1.09725
0.09213
8.272.975
18.716.683
2.35
5.32
Northeast
Uniform EE
3
528.750
138.810
212.330
22.240
0.26252
0.40157
0.04206
8.736.861
19.732.529
1.65
3.73
EE at Peak
3
200.980
73.510
103.110
9.900
0.36576
0.51304
0.04926
4.510.929
10.187.690
2.24
5.07
Solar
3
157.560
47.320
73.000
7.260
0.30033
0.46332
0.04608
3.056.040
6.901.902
1.94
4.38
Wind
3
175.560
47.150
64.520
7.090
0.26857
0.36751
0.04039
2.769.336
6.254.873
1.58
3.56
Pacific
Northwest
Uniform EE
3
520.420
441.810
620.760
41.960
0.84895
1.19281
0.08063
5.874.146
13.285.524
1.13
2.55
EE at Peak
3
199.410
170.770
237.720
16.510
0.85638
1.19212
0.08279
2.242.063
5.070.811
1.12
2.54
Solar
3
173.790
150.810
212.850
14.230
0.86777
1.22475
0.08188
2.028.039
4.586.878
1.17
2.64
Wind
3
220.430
183.720
261.500
17.740
0.83346
1.18632
0.08048
2.483.003
5.615.855
1.13
2.55
53
-------
Region
Project
Type
Discount
Rate
Results from AVERT
so2
Emissions
Rate
Ob./MWh)
NOx
Emissions
Rate
Ob./MWh)
pm25
Emissions
Rate
Ob./MWh)
Results from COBRA
c/k\Vh
0ow)
c/k\Vh
(high)
Displaced
Generation
(MWh)
so2
Reduced
Obs.)
NOx
Reduced
Obs.)
pm25
Reduced
Obs.)
S Total Health
Benefits
0ow)
S Total Health
Benefits
(high)
Rocky
Mountains
Uniform EE
3
521.840
288.080
687.930
17.370
0.55205
1.31828
0.03329
5.359.681
12.124.457
1.03
2.32
EE at Peak
3
199.330
102.310
244.880
7.340
0.51327
1.22852
0.03682
1.951.438
4.414.401
0.98
2.21
Solar
3
197.300
101.990
251.560
6.950
0.51693
1.27501
0.03523
1.961.721
4.437.674
0.99
2.25
Wind
3
306.950
176.630
412.520
9.920
0.57544
1.34393
0.03232
3.274.588
7.407.658
1.07
2.41
Southeast
Uniform EE
3
524.860
386.360
455.140
50.980
0.73612
0.86716
0.09713
9.319.089
21.094.677
1.78
4.02
EE at Peak
3
199.380
151.520
201.430
20.870
0.75996
1.01028
0.10467
3.733.770
8.451.618
1.87
4.24
Solar
3
168.790
126.600
160.410
17.150
0.75004
0.95035
0.10161
3.097.209
7.010.795
1.83
4.15
Wind
3
120.130
88.370
99.640
11.330
0.73562
0.82943
0.09431
2.110.045
4.776.336
1.76
3.98
Southwest
Uniform EE
3
521.270
122.320
472.260
36.220
0.23466
0.90598
0.06948
3.726.818
8.432.593
0.71
1.62
EE at Peak
3
200.550
38.430
194.520
13.990
0.19162
0.96993
0.06976
1.408.299
3.186.512
0.70
1.59
Solar
3
226.200
51.830
211.860
16.040
0.22913
0.93660
0.07091
1.640.241
3.711.328
0.73
1.64
Wind
3
213.890
61.000
192.250
15.530
0.28519
0.89883
0.07261
1.636.309
3.702.391
0.77
1.73
Texas
Uniform EE
3
526.050
655.200
324.990
37.900
1.24551
0.61779
0.07205
8.317.315
18.817.958
1.58
3.58
EE at Peak
3
199.800
198.090
137.810
13.420
0.99144
0.68974
0.06717
2.767.366
6.260.993
1.39
3.13
Solar
3
182.460
191.100
120.880
12.360
1.04735
0.66250
0.06774
2.593.074
5.866.753
1.42
3.22
Wind
3
295.870
387.990
176.990
21.620
1.31135
0.59820
0.07307
4.823.340
10.912.883
1.63
3.69
Upper
Midwest
Uniform EE
3
524.750
991.470
675.020
48.130
1.88941
1.28636
0.09172
16.377.183
37.044.470
3.12
7.06
EE at Peak
3
199.200
316.380
248.570
17.600
1.58825
1.24784
0.08835
5.475.798
12.385.741
2.75
6.22
Solar
3
167.860
283.770
207.880
14.960
1.69052
1.23841
0.08912
4.848.718
10.967.473
2.89
6.53
Wind
3
360.360
699.550
471.900
33.180
1.94125
1.30952
0.09207
11.521.813
26.061.951
3.20
7.23
California
Uniform EE
7
522.060
37.410
165.980
23.090
0.07166
0.31793
0.04423
2.217.845
5.010.362
0.42
0.96
EE at Peak
7
200.130
13.110
57.400
8.620
0.06551
0.28681
0.04307
925.286
2.090.072
0.46
1.04
Solar
7
194.380
13.780
62.970
8.940
0.07089
0.32395
0.04599
883.962
1.996.832
0.45
1.03
Wind
7
151.660
11.150
49.200
6.720
0.07352
0.32441
0.04431
649.750
1.467.800
0.43
0.97
Great Lakes/
Mid-Atlantic
Uniform EE
7
521.980
606.820
478.260
102.050
1.16253
0.91624
0.19551
16.370.600
37.011.474
3.14
7.09
EE at Peak
7
198.470
228.410
192.760
38.100
1.15085
0.97123
0.19197
6.330.389
14.312.011
3.19
7.21
Solar
7
153.580
184.740
144.290
30.320
1.20289
0.93951
0.19742
5.022.792
11.355.790
3.27
7.39
Wind
7
226.120
243.150
210.750
44.120
1.07531
0.93203
0.19512
6.768.256
15.301.768
2.99
6.77
Lower
Midwest
Uniform EE
7
526.240
919.370
588.070
48.430
1.74705
1.11749
0.09203
10.852.982
24.542.104
2.06
4.66
EE at Peak
7
199.500
305.300
239.080
18.030
1.53033
1.19840
0.09038
3.751.710
8.483.698
1.88
4.25
Solar
7
188.940
303.910
220.330
17.280
1.60850
1.16614
0.09146
3.695.110
8.355.735
1.96
4.42
Wind
7
352.100
632.180
386.340
32.440
1.79546
1.09725
0.09213
7.382.465
16.694.159
2.10
4.74
Northeast
Uniform EE
7
528.750
138.810
212.330
22.240
0.26252
0.40157
0.04206
7.796.956
17.601.071
1.47
3.33
EE at Peak
7
200.980
73.510
103.110
9.900
0.36576
0.51304
0.04926
4.025.772
9.087.337
2.00
4.52
Solar
7
157.560
47.320
73.000
7.260
0.30033
0.46332
0.04608
2.727.332
6.156.418
1.73
3.91
Wind
7
175.560
47.150
64.520
7.090
0.26857
0.36751
0.04039
2.471.374
5.579.212
1.41
3.18
54
-------
Results from AVERT
so2
NOx
pm25
Results from COBRA
Region
Project
Type
Discount
Rate
Displaced
Generation
(MWh)
so2
Reduced
Obs.)
NOx
Reduced
Obs.)
pm25
Reduced
Obs.)
Emissions
Rate
Ob./MWh)
Emissions
Rate
Ob./MWh)
Emissions
Rate
Ob./MWh)
S Total Health
Benefits
0ow)
S Total Health
Benefits
(high)
c/k\Vh
0ow)
c/k\Vh
(high)
Uniform EE
7
520.420
441.810
620.760
41.960
0.84895
1.19281
0.08063
5.241.959
11.849.667
1.01
2.28
Pacific
EE at Peak
7
199.410
170.770
237.720
16.510
0.85638
1.19212
0.08279
2.000.768
4.522.771
1.00
2.27
Northwest
Solar
7
173.790
150.810
212.850
14.230
0.86777
1.22475
0.08188
1.809.784
4.091.150
1.04
2.35
Wind
7
220.430
183.720
261.500
17.740
0.83346
1.18632
0.08048
2.215.783
5.008.918
1.01
2.27
Uniform EE
7
521.840
288.080
687.930
17.370
0.55205
1.31828
0.03329
4.783.050
10.814.286
0.92
2.07
Rocky
EE at Peak
7
199.330
102.310
244.880
7.340
0.51327
1.22852
0.03682
1.741.497
3.937.384
0.87
1.98
Mountains
Solar
7
197.300
101.990
251.560
6.950
0.51693
1.27501
0.03523
1.750.671
3.958.141
0.89
2.01
Wind
7
306.950
176.630
412.520
9.920
0.57544
1.34393
0.03232
2.922.287
6.607.187
0.95
2.15
Uniform EE
7
524.860
386.360
455.140
50.980
0.73612
0.86716
0.09713
8.315.459
18.814.863
1.58
3.58
Southeast
EE at Peak
7
199.380
151.520
201.430
20.870
0.75996
1.01028
0.10467
3.331.679
7.538.222
1.67
3.78
Solar
7
168.790
126.600
160.410
17.150
0.75004
0.95035
0.10161
2.763.662
6.253.109
1.64
3.70
Wind
7
120.130
88.370
99.640
11.330
0.73562
0.82943
0.09431
1.882.797
4.260.129
1.57
3.55
Uniform EE
7
521.270
122.320
472.260
36.220
0.23466
0.90598
0.06948
3.325.818
7.521.287
0.64
1.44
Southwest
EE at Peak
7
200.550
38.430
194.520
13.990
0.19162
0.96993
0.06976
1.256.774
2.842.148
0.63
1.42
Solar
7
226.200
51.830
211.860
16.040
0.22913
0.93660
0.07091
1.463.756
3.310.247
0.65
1.46
Wind
7
213.890
61.000
192.250
15.530
0.28519
0.89883
0.07261
1.460.238
3.302.273
0.68
1.54
Uniform EE
7
526.050
655.200
324.990
37.900
1.24551
0.61779
0.07205
7.423.071
16.785.407
1.41
3.19
Texas
EE at Peak
7
199.800
198.090
137.810
13.420
0.99144
0.68974
0.06717
2.469.838
5.584.731
1.24
2.80
Solar
7
182.460
191.100
120.880
12.360
1.04735
0.66250
0.06774
2.314.285
5.233.080
1.27
2.87
Wind
7
295.870
387.990
176.990
21.620
1.31135
0.59820
0.07307
4.304.745
9.734.164
1.45
3.29
Uniform EE
7
524.750
991.470
675.020
48.130
1.88941
1.28636
0.09172
14.613.805
33.041.446
2.78
6.30
Upper
EE at Peak
7
199.200
316.380
248.570
17.600
1.58825
1.24784
0.08835
4.886.201
11.047.331
2.45
5.55
Midwest
Solar
7
167.860
283.770
207.880
14.960
1.69052
1.23841
0.08912
4.326.639
9.782.324
2.58
5.83
Wind
7
360.360
699.550
471.900
33.180
1.94125
1.30952
0.09207
10.281.224
23.245.697
2.85
6.45
55
-------
Appendix H: Conversions
Table H - 1 lists common conversions used throughout this report.
Table H -1. Common Conversions.
Original Units
Multiply by
To Obtain
0/kWh
1,000
0/MWh
0/kWh
1,000,000
0/GWh
56
-------
U.S. Environmental Protection Agency
State and Local Energy and Environment Program
1200 Pennsylvania Ave, NW (6202A)
Washington, DC 20460
epa.gov/statelocalenergy
452R18001
July 2019
V
H
------- |