EPA/600/R-08-087
                                  July 2008
ANALYSIS OF INDIRECT EMISSIONS
BENEFITS OF WIND, LANDFILL GAS,
   AND MUNICIPAL SOLID WASTE
             GENERATION
                    By

      Ezra D. Hausman, Jeremy Fisher, and Bruce Biewald
                Synapse Energy
                 22 Pearl Street
               Cambridge, MA 02139
              Order No.: EP07C000079
                 Project Officer

               Joseph F. DeCarolis
        Air Pollution Prevention and Control Division
       National Risk Management Research Laboratory
           Research Triangle Park, NC 27711

-------
                                   Notice

The U.S. Environmental Protection Agency, through its Office of Research and Development,
funded and managed the research described here under Order Number EP07C000079 to
Synapse Energy Economics. It has been subjected to the Agency's peer and administrative
review and has been approved for publication as an EPA document.
                                Foreword

The U. S. Environmental Protection Agency (EPA) is charged by Congress with protecting the
Nation's land, air, and water resources. Under a mandate of national environmental laws, the
Agency strives to formulate and implement actions leading to a compatible balance between
human activities and the ability of natural systems to support and nurture life. To meet this
mandate, EPA's research program is providing data and technical support for solving
environmental problems today and building a science knowledge base necessary to manage our
ecological resources wisely, understand how pollutants affect our health, and prevent or reduce
environmental risks in the future.

The National Risk Management Research Laboratory (NRMRL) is the Agency's center for
investigation of technological and management approaches for preventing and reducing risks
from pollution that threaten  human health and the environment. The focus of the Laboratory's
research program is on methods and their cost-effectiveness for prevention and control of
pollution to air, land, water,  and subsurface resources; protection of water quality in public water
systems; remediation of contaminated sites, sediments and ground water; prevention and control
of indoor air pollution; and restoration of ecosystems. NRMRL collaborates with both public and
private sector partners to foster technologies that reduce the cost of compliance and to anticipate
emerging problems. NRMRL's research provides solutions to environmental  problems by:
developing and promoting technologies that protect and improve the environment;  advancing
scientific and engineering information to support regulatory and policy decisions; and providing
the technical support and information transfer to ensure implementation of environmental
regulations and strategies at the national, state, and community levels.

This publication has been produced as part of the Laboratory's strategic long-term  research plan.
It is published and made available by EPA's Office of Research and Development to assist the
user community and to link  researchers with their clients.
                                                              Sally Gutierrez, Director
                                       National Risk Management Research Laboratory
                          Indirect Emissions Analysis • i

-------
                              Abstract

A number of techniques are introduced to calculate the hourly indirect emissions benefits
of three types of renewable resources: wind energy, Municipal Solid Waste (MSW)
combustion, and Landfill Gas (LFG) combustion. These techniques are applied to each of
the EPA's eGRID subregions in the continental United States in order to derive hourly,
seasonal, or annual (as appropriate) coefficients for use in evaluating the indirect
emissions benefits of such projects in each region.

For wind  power impacts, simulated wind project power profiles are derived using publicly
available wind data scaled to a typical turbine height for new wind projects and
transformed through a  power curve for a  proxy turbine. The results for one region are
compared to the output of an existing project, although the limitations of this type of
comparison, as well as the limitations of representing large areas of the country with
single proxy curves, are discussed. LFG and MSW are found to have flat, "always on"
profiles based on the limited data available.

The regional, hourly power profiles for each type of resource are combined with the
hourly indirect emissions coefficients to yield annual indirect emissions benefits for each
type of resource for each eGRID subregion.

For each GWh of renewable energy produced each year, the indirect CO2 emissions
benefit is found to be between 600 and 1100 tons of CO2 depending on the region, with
coal-dependent regions having the highest indirect emissions benefit. The indirect NOx
and SO2 emissions benefit also depends  on the regional fuel mix, as well as the
stringency of environmental regulation in each region. These indirect emissions benefits
vary between 500 and  5,000 pounds of NOx per GWh, and between 200 and 1,300
pounds of SO2 per GWh. With some exceptions, these results are robust regardless of
whether the renewable resource has a base load profile (such as MSW and LFG) of
output which varies diurnally and  seasonally (such as wind). However, as the calculated
benefits do  vary in some cases depending on the analytical approach, care must be used
in selecting  an  appropriate calculation methodology for each application.

This report was submitted in fulfillment of Order Number EP07C000079 under the
sponsorship of the EPA.
                      Indirect Emissions Analysis • ii

-------
                            Table of Contents
Notice	i
Foreword	i
Abstract	ii
List of Tables	iv
List of Figures	v
Acronyms and Abbreviations	vii
List of Appendices	viii
1. Introduction and Summary of Conclusions	1
2. Operating Data on Renewable Energy Technologies	6
     A. Wind Generation: Existing and Synthetic Projects	6
         Data Sources for Synthetic Generation	7
         Generation of Synthetic Wind Power Data	8
         Comparing Empirical and Simulated Wind Generation	10
     B. Landfill Gas	14
     C. Municipal Solid Waste	14
3. Hourly generation and emissions data	15
         Characterization of emission rates	16
         Data filtering	16
4. Emissions Displacement Methods	17
     A. Limiting Cases	18
     B. Hourly Average Emissions Rate (HAER)	19
     C. Emissions Slope Factor	22
     D. Empirical Incremental Emissions Rate (EIER)	29
     E. Load-Following Incremental Emissions Rate (LFIR)	30
     F. Flexibility-Weighted  Hourly Average Emissions Rate (FW-HAER)	31
     G. Comparison of Indirect Emissions Rates	35
5. Indirect Emissions Results	37
6. Conclusions	44
Appendix A:	46
Appendix B:	48
Appendix C:	52
Appendix D:	56
                            Indirect Emissions Analysis • ill

-------
                           List of Tables

Table 1. Seasonal and annual emissions slope factors and R-squared
         values for C02	24

Table 2. Seasonal and annual emissions slope factors and R-squared
         values for S02	25

Table 3. Ozone season and non-ozone season emissions rate
         slopes for NOx	26

Table 4. Guide to the summary figures	37

Table 5. Annual indirect emissions reduction rates associated with an
         incremental GWh of wind energy in each of the eGRID
         subregions for 2005	39

Table 6. Annual indirect emissions reduction rates associated with an
         incremental GWh of landfill gas of MSW generation in each of
         the eGRID subregions for 2005	40

Table 7. "Shape impact" for indirect emissions calculations for each
         pollutant in each eGrid region	40
                        Indirect Emissions Analysis • iv

-------
                           List of Figures
Figure 1.    Indirect emissions results for C02, NOx, and 862 for wind,
           municipal solid waste and landfill gas generation projects	4
Figure 2.    Wind power curve for a GE 1.5se wind turbine	9
Figure 3.    Proxy wind turbine power curve plotted against histograms of
           raw and scaled wind speeds  at a sample site in the Midwest	10
Figure 4.    Comparison of power output  distribution at sample wind
           power site in the Midwest with predicted output based on
           scaled wind data from a nearby meteorological station	12
Figure 5.    Comparison of hourly synthetic and observed wind power
           time series at nearby locations in the Midwest	13
Figure 6.    Cumulative distribution of the hourly deviation between
           observed and synthetic generation time series	13
Figure 7.    Distribution of C02 emission  rates for fossil generating plants
           by reported fuel type	18
Figure 8.    Representations of C02 hourly average emissions rate
           (HAER) in New England	21
Figure 9.    Determination of the regional emissions slope factor in the
           RFCE (Reliability First/Central) region	23
Figure 10.  Bifurcation of the NOx emissions slope factor for the RFCE
           region	23
Figure 11.  Regional slope factors for C02	27
Figure 12.  Seasonal and regional slope  factors for S02	27
Figure 13.  Regional slope factors for NOx	28
Figure 14.  Flexibility coefficient vs. generator size in New England and
           California	33
Figure 15.  Hourly dispatch and flexibility coefficient for three plants of
           differing capacities in the FRCC subregion	34
Figure 16.  Comparison of the distribution of indirect emissions
           coefficients for C02, S02,  and NOx based on methods
           described in the text	36
                         Indirect Emissions Analysis • v

-------
Figure 17.   Annual indirect C02 emissions impact associated with one
           incremental GWh of wind energy in each of the eGRID
           subregions for 2005	41

Figure 18.   Annual indirect NOx emissions impact associated with one
           incremental GWh of wind energy in each of the eGRID
           subregions for 2005	41

Figure 19.   Annual indirect 862 emissions impact associated with one
           incremental GWh of wind energy in each of the eGRID
           subregions for 2005	42

Figure 20.   Annual indirect C02 emissions impact associated with one
           incremental GWh of landfill gas or municipal solid waste
           energy in each of the eGRID subregions for 2005	42

Figure 21.   Annual indirect NOx emissions impact associated with an
           incremental GWh of landfill gas or municipal solid waste
           energy in each of the eGRID subregions for 2005	43

Figure 22.   Annual indirect 862 emissions impact associated with an
           incremental GWh of landfill gas or municipal solid waste
           energy in each of the eGRID subregions for 2005	43
                        Indirect Emissions Analysis • vi

-------
           Acronyms and Abbreviations
AWEA         American Wind Energy Association
BTU          British Thermal Unit
CO2          Carbon Dioxide
DOE          U.S. Department of Energy
eGRID         Emissions & Generation Resource Integrated Database
EIER          empirical incremental emission rate (defined in this study)
EPA          United States Environmental Protection Agency
FW-HAER     flexibility-weighted hourly average emission rate (defined in this study)
GE           General Electric
GW          gigawatt
GWh          gigawatt-hour
HAER         hourly average emission rate (defined in this study)
kW           kilowatt
kWh          kilowatt-hour
LFG          landfill gas (generating resource)
LFIR          load-following incremental emissions rate (defined in this study)
mmBTU       million British thermal units
MSW         municipal solid waste (generating resource)
MW          megawatt
MWh          megawatt-hours
NCDC         National Climatic Data Center
NERC         North American Electric Reliability Corporation
NOAA         National Oceanic and Atmospheric Administration
NOx          mono-nitrogen oxides (NO and NO2)
NREL         National Renewable Energy Laboratory
SO2          sulfur dioxide
WBAN         Weather Bureau Army Navy
                   Indirect Emissions Analysis • vii

-------
                     List of Appendices
Appendix A

Identification and map of eGRID subregions

Appendix B

Synthetic wind power time series (hourly percent of capacity) for each eGRID subregion,
represented as color intensity plots

Appendix C

Summary tables of raw and scaled wind speed and synthetic wind power time series

Appendix D

Color intensity maps of hourly indirect CO2, NOx and SO2 emissions rates for each
eGRID subregion, using several alternative calculation methodologies
                     Indirect Emissions Analysis • viii

-------
1.  Introduction and Summary of Conclusions
        The purpose of this report is to develop, analyze, and report upon methods of quantifying
        the indirect emissions benefit associated with renewable energy generation resources. In
        particular, the goal is to conduct a national assessment of indirect emissions reductions
        from wind, LFG and MSW by NERC subregion.1 Because  conventional electricity
        generation resources can differ widely in their emissions characteristics, the
        determination of exactly which conventional resource or resources would be running but
        for the contribution of the renewable resource would be the ideal first step in calculating
        indirect emissions benefits. If we knew which resources  were "displaced" by the
        renewable resource, we could simply  multiply the emission rate of those resources by the
        total number of megawatt-hours (MWh)  displaced,  accounting for line losses, and an
        exact answer would be obtained. The answer thus obtained would vary from region to
        region, and from hour-to-hour within any given region, depending on current and
        expected market conditions.

        Unfortunately, in large, interconnected, security-constrained2 electricity markets,
        determining which units would be displaced  in any  given hour would require perfect
        information and a complex dispatch model. Even under such ideal circumstances, the
        analysis would be prohibitively time-consuming and complex for each hour. Thus our goal
        in the current study is to develop and  apply proxy methods that, while imperfect, are
        applicable to and useful for the general estimation  of the indirect emissions benefits for
        renewable energy projects anywhere  in  the continental United States. Over the course of
        this report, we develop and apply these  methods to estimate the indirect emissions
        benefits  of three kinds of renewable energy projects, for carbon dioxide (CO2), sulfur
        dioxide (SO2) and oxides of nitrogen (NOx),  in each of the 22 eGRID3 subregions of the
        continental United States.

        An "indirect emissions benefit" may be a real reduction in total emissions in a region or,
        under cap-and-trade regulation such as that in much of the United States for the
        pollutants SO2 and NOx, it may be an opportunity to release emissions allowances for
        some other use. The calculation methods explored in this report are appropriate for either
        application, but the impacts of these scenarios are quite different in terms of their ultimate
        effect on pollutant emissions. In  the case of cap-and-trade regulation, displacing air
        emissions provides an opportunity for the allowance holder to sell or bank valuable
        emissions allowances. There may also be a societal benefit in terms of reducing the
        market price  of emissions allowances, thereby reducing  this component of the cost of
1 EPA RFQ # RFQ-OH-07-00015, as posted at:
 https://www.fbo.gov/?s=opportunity&mode=form&id=556118954b3728ae31326t>103396bcf6&tab=core&_cview=0.
2
 "Security constrained" refers to the need for system operators to dispatch generating units to meet load while respecting
 the loading on transmission lines and interfaces within their operating limits. Because of these security constraints,
 generating units cannot simply be dispatched in merit order, from least expensive to most expensive running costs. In
 many cases more expensive units must be run out of merit order due to transmission transfer limits.
 The Emissions & Generation Resource Integrated Database (eGRID) is a comprehensive inventory of environmental
 attributes of electric power systems, (http://www.epa.gov/cleanenergv/energy-resources/egrid/index.html.) Definitions and a
 map of the eGRID subregions are shown in Appendix A. eGRID subregions are roughly coincident with NERC (North
 American Electric Reliability Corporation) subregions.
                               Indirect Emissions Analysis • 1

-------
        electricity. Finally, there may ultimately be a total emissions reduction if the lower cost
        and greater availability of allowances permits regulators to tighten the total emissions
        cap.

        This report is organized as follows:

        •   In Section 2, we develop and present operating characteristics of wind, landfill gas,
           and municipal solid waste energy resources in each of the eGRID subregions  of the
           United States. We develop hourly operating profiles for a "typical" such resource in
           each region, normalized to produce one GWh4 of electricity per year.
        •   In Section 3, we collect and present data on conventional power plants in each
           eGRID region based on the EPA's Acid Rain database.5 We characterize the data
           and discuss our approach to data filtering and error handling, in preparation for
           characterizing the indirect emissions benefits associated with changes in dispatch.
        •   In Section 4, we implement and evaluate a number of approaches to applying the
           filtered and cleaned EPA  data for the purposes of calculating hourly avoidable
           emissions factors. These  factors  can then be applied to renewable resource
           production profiles to calculate annual indirect emissions benefits. In addition,  we
           present hourly indirect emissions results for each pollutant, region, and calculation
           approach under consideration, and we discuss which  approaches make sense for
           which kinds of analyses.
        •   In Section 5 we apply the indirect emissions calculation methods to determine
           indirect emissions benefits associated with each resource type under consideration
           for each region and pollutant.
        •   Finally, in Section 6, we discuss  our results and present suggestions for further
           research in this area.
        Fundamental to the analysis in this report, all of which is based on retrospective data, is
        the primary focus on short-term, operational response of the electric system to the
        presence of new renewable energy resources. This analysis is thus most appropriate for
        analyzing the impact of a relatively small quantity of new resources during the first few
        years  of operation, during which the displaced resources are likely to be the most flexible,
        load-following units on the system. Over the long term, it is likely that the capacity  mix of
        an electric system will be altered as a result of the addition of new renewable energy
        resources. Specifically, new generating capacity investments may be avoided or
        deferred, or existing capacity  may be retired, until the system returns to equilibrium in
        terms  of the balance of base load  and load-following resources. Ultimately, it may  be
        expected that the proportion of flexible units will be restored to what it would have  been
        without the new resource. In this sense, as it matures and is incorporated in larger
        quantities, renewable energy  technology will ultimately displace more base load capacity
4
 A GWh (gigawatt-hour) is one thousand megawatt-hours, approximately the amount of power that could be produced by
 one of the largest power plants in the United States in one hour. For perspective, the largest new single wind turbines
 are about 3 MW in capacity but produce, on average, about one MWh of electricity per hour. Thus one of the largest
 wind turbines would be expected to produce about nine GWh of electricity each year.  In general, wind projects have
 more than one wind turbine, and can have up to 200 or more.
 Available from the EPA Clean Air Markets data website, at http://camddataandmaps.epa.gov/gdm/
                               Indirect Emissions Analysis • 2

-------
      on the system. While we touch on this important issue in some parts of this report, it is
      not a primary focus.6

      The indirect emissions benefit calculated for any kind of renewable energy resource
      depends on a number of questions:

      •   Where is the resource located?
      •   What is the pollutant of interest?
      •   Is the time period of interest historical, in the near future, or several years in the
          future?
      •   Is the resource base load, dispatchable, or intermittent and  nondispatchable?
      •   If intermittent and nondispatchable, what is the expected hourly and seasonal profile
          of the resource?
      Because of the wide range of applications reflected in the  possible answers to these
      questions, the best method to use in calculating indirect emissions benefits can vary with
      the application. In this report we present a number of methods, and apply them to three
      types of renewable resources and three pollutants in 22 regions of the United States.
      Even this level of detail is almost certainly insufficient for precise analysis of the
      emissions benefits of individual projects. However, it leads to illuminating  and useful
      results that will allow for a  far more detailed understanding of the issues and a better first
      approximation of the indirect emissions benefits than has previously been available.

      We find:

      •   The indirect emissions benefits of renewable energy for all pollutants vary
          significantly by region, and that these differences can  be quantified and applied in
          calculating the indirect emissions reductions.
      •   For each GWh of renewable energy produced each year, the indirect CO2 emissions
          benefit is found to be between 600 and 1000 tons of CO2, with coal-dependent
          regions having the highest indirect emissions benefit.
      •   For each GWh of renewable energy produced each year, the indirect NOx emissions
          benefit varies between 500 and 5,000 pounds of NOx.
      •   For each GWh of renewable energy produced each year, the indirect SO2 emissions
          benefit varies between 200 and 1,300 pounds of SO2.
      •   These results are robust regardless of whether the renewable resource has a base
          load profile (such as MSWand LFG) of output which varies diurnally and seasonally
          (such as wind).
      •   Different calculation approaches are appropriate for different types of resources and
          different applications.
For a thorough treatment of this issue, see 2005 Synapse report Using Electric System Operating Margins and Build
Margins: Quantification of Carbon Emission Reductions Attributable to Grid Connected COM Projects, available at
http://www.synapse-energy.com/
                             Indirect Emissions Analysis • 3

-------
(a)
     •   More research is needed to establish which methods of calculating indirect emissions
        changes most accurately reflect real-world dispatch over a range of timescales and
        system conditions.
     Of the methods introduced here, we find the flexibility-weighted hourly average emission
     rate (FW-HAER) to be the most appropriate for near-term (<3 year) projections of indirect
     emissions impact from wind  resources, because it best represents the response of the
     electric system to rapid fluctuations in system load. We find the seasonal slope factor
     approach to be most appropriate for calculating the indirect emissions impact from base
     load resources such as municipal solid waste and landfill gas electricity generation.

     The annual indirect emissions results for these resources based on 2005 data, assuming
     one GWh of energy was produced by each over the course of the year, are shown in
    ^Figure 1T With some notable exceptions, the indirect CO2 and NOx emissions benefits of
     the different renewable energy resources is generally similar; this benefit diverges most
     dramatically by resource type for SO2. This divergence probably reflects the fact that in
     many regions, base  load resources like MSW and LFG can displace coal, which emits
     SO2, while wind initially primarily replaces gas, which does not emit SO2. However, with
     increasing penetration and geographic diversity of wind generation it is likely that this
     resource will begin to displace more base load coal, and the indirect emissions impact of
     wind generation will  be closer to that of MSW and LFG.
           1,200 -,
           1 ,000
         CD
O
O
'o
(A
c
o
T3
8
jn
Q.
            800
            600
            400
            200
                    X
                    <
                    0
                       0
                  O
                  0
                  CC
                                      DWind
                                                      BMSW/LFG
       Figure 1. Indirect emissions results for (a) COz (b) NOx; and (c) SC>2 for wind, municipal solid
       waste (MSW) and landfill gas (LFG) generation projects. (Continued on next page)
                           Indirect Emissions Analysis • 4

-------
(b)
b,UUU -
_c
O
9. A nnn
X
o
'o
m ^ nnn
c
o
CL
"m 9 nnn
ro
Q.
(A
b
1 nnn





1-1

5
z
N
<
-
flffl
-


i-.


-i

x H o iu <
s ฐ. o O i

r
p.




_


p





















r



-i
r

-i

-i





r
'


= LiJCLaziQ-LiJ5ง5O>O
)5|o^^^oo|zco||co^>
= z-zz zcccccccecococo^coWco
DWind • MSW/LFG

(C)
I4,UUU -
19 nnn
i
O -in nnn -
0
Q_
6"
co R nnn
c
D c nnn
CL
T3
8
ro ^ nnn -
(A
b
9 nnn -
n -






fln-rll"



•I


-,

[



imr



r

-


r

-i

r

-

[




\
-



r
Iffl
-1

r





-

r

                                          DWind
                                                             MSW/LFG
                               Indirect Emissions Analysis • 5

-------
2.  Operating Data on  Renewable Energy Technologies
       In order to determine the indirect emissions reduction impacts of electricity production
       from renewable energy sources, it is necessary to first establish what kinds of
       conventional resources they are displacing. As a first step, in this section of the report we
       explore the operational characteristics of three types of renewable energy resources:
       wind power generation, municipal solid waste combustion, and landfill gas. We develop
       profiles for these resources  representative of each eGRID subregion of the continental
       United States7 based on historic data and operating characteristics of these resources.
       (The eGRID subregions are identified and mapped in Appendix A.) Once we have
       determined their hourly output profiles over the course of a historic year, we can begin an
       assessment of which resources they would have been likely to displace.

       In the section that follows, we will explore corresponding operational and dispatch
       characteristics of conventional resources. Following that we combine the results of these
       research tasks, characterizing renewable energy output and assessing displaceable
       conventional resources, to make a final assessment of which fossil resources would have
       been likely to be displaced by renewables in each eGRID subregion. Finally, we use the
       results of this combined analysis to determine factors representing the  indirect emissions
       benefit for each subregion.

A. Wind Generation: Existing and Synthetic Projects
       The American Wind Energy Association  (AWEA) estimates that there are approximately
       8.3 gigawatts (GW) of installed wind capacity in the United States as of the 4th quarter of
       2007, with another 4.1 GW under construction.8 Texas and California dominate the
       current market, but there are underutilized regions throughout the Midwest. Some states
       with the greatest wind potential, such as the corridor stretching from  North Dakota south
       through Kansas, have yet to significantly tap this resource. In general, there are
       extensive opportunities to expand the rapid growth in wind generation,  resulting in the
       further displacement of fossil fuel generation and decreased emissions of greenhouse
       gases and other pollutants. This large resource potential motivates the need to
       understand in more detail the emissions benefits associated with renewable energy
       projects in each region.

       Ideally, to characterize the output profiles for new wind generation projects, we would
       draw empirical wind power output datasets from a broad sampling of existing projects
       around the country. Unfortunately, wind generation projects are not required to publicly
       report hourly generation, and thus there  are limited opportunities to calculate expected
       emissions reductions from existing resources. In addition, "real" wind projects present
       numerous hurdles that make them problematic as a source of research data for a broad-
       based study of operational characteristics. For example, there is generally no uniformity,
       and often no available information, on the technical details of existing resources.  With
 eGRID subregions are almost coincident with NERC subregions. We chose to use eGRID subregions for this analysis
 as this is the system used by EPA for emissions reporting.
 American Wind Energy Association, 2007. Updated information can be found at http://www.awea.org/projects/.
                              Indirect Emissions Analysis • 6

-------
       operational data available on few projects, what may appear to be regional differences in
       operating characteristics would be heavily influenced by unknown differences in
       technology, turbine size, turbine height, vintage, and specific local conditions associated
       with each individual project. In addition, the geographical coverage of available wind
       power output data is extremely limited. In fact, we were only able to find one source for
       such data, from the U.S. Department of Energy's National Renewable Energy Laboratory
       (NREL). This dataset covered  only a small number of projects with limited geographic
       distribution. Even with this source we were asked not to present specific, identifiable
       details or operational data to avoid disclosure of proprietary information, limiting its
       usefulness for our study.

       Our solution to these limitations is to simulate wind power projects based on
       meteorological data gathered at locations where wind projects are likely to be built in
       each region, determined based on the quality of the wind resource. While wind power
       data are scarce, wind speed data are abundant and readily available at reasonably high
       resolution throughout the United States.  Near-surface hourly wind speed time series data
       can be scaled to reflect a typical turbine  height, and then converted with the help of a
       "power curve"9 into corresponding simulated wind power output time series. We have
       attempted to perform a comparison  of one of our simulated wind power time series with
       the output of a  nearby physical wind-power project; however, the results are difficult to
       interpret. The wind power and  the wind data are not from exactly the same location or
       height, and even such small differences  can be extremely significant in determining the
       output of a resource. The technology and vintage represented by our power curve is also
       unlikely to be well  matched to the actual technology and vintage of the  specific wind
       turbines for which  we have data. Even so, this approach provides a useful means of
       evaluating the potential impacts of wind energy production on emissions. The data
       sources and comparison results are described  in more detail below.

    Data Sources for Synthetic Generation
       Over 1700 meteorological stations collect hourly data throughout the United States and
       territories for the National Climate Data Center (NCDC) of the  National Oceanic and
       Atmospheric Administration (NOAA). Our primary source of wind data for generating
       synthetic wind power time series was NCDC's Weather Bureau Army Navy (WBAN)
       stations, distributed throughout the United States. To estimate the temporal
       characteristics  of potential wind power generation, and thus expected indirect emissions
       reductions for each eGRID subregion, we needed to determine likely locations for new
       wind generation projects. We did this by reviewing the highest resolution available 50-
       meter wind speed  estimates in each eGRID subregion, as cataloged by the NREL,10 to
       determine where favorable wind resources were  likely to be located based on average
       wind speeds at 50-meter hub heights. We assumed that the minimum economic wind
       speed threshold for a wind  power project would be at Class 4 wind speeds, or winds of
9
 A "power curve" is a graph which relates wind speed to expected power production for a specific wind turbine technology
 and configuration. Actual power output data from wind turbines generally exhibit considerable scatter around the
 expected power curve, but this remains the best means of characterizing this relationship.
10 USDOE NREL Wind and Hydropower Electricity Program, 2007. Available online at:
 http://www.eere.energy.gov/windandhydro/windpoweringamerica/wind_maps.asp
                              Indirect Emissions Analysis • 7

-------
    15.7-16.8 miles per hour (mph). Once the area was selected in this manner, the closest
    high-quality NCDC station wind data location was selected for each subregion. Twenty-
    two NCDC meteorological stations near these locations were selected as proxies to
    represent the temporal pattern of wind generation expected at these sites. The locations
    of the sites and summary wind characteristics from the selected NCDC WBAN stations
    may be found in Appendix C.

    It must be noted at the outset that the geographic granularity of the wind data is far from
    satisfactory. Wind conditions can change significantly over spatial scales as small as a
    few kilometers or less, so representing each eGRID subregion with a single site entails a
    considerable simplification. We have tried to choose sites which are reasonably
    representative of promising wind power locations for each region, understanding that
    considerably more research and analysis would be needed to pick a truly representative
    site, if indeed a single site can stand in for a large region. The same can be said for solid
    waste and landfill gas projects considered in our analysis: there is simply not enough
    information available to confidently predict the behavior of all potential future sites. This is
    an inevitable shortcoming of attempting to represent site-specific characteristics on a
    national scale, and it should be kept in mind by the reader in interpreting our results.

Generation of Synthetic Wind Power Data
    WBAN meteorological stations are typically located at local and regional airports, or near
    rural towns, at 10 meters (33 ft) above ground level. Due to surface friction, wind speeds
    near the ground are significantly lower than speeds at expected 50-100 m windmill hub
    heights. In addition, the WBAN stations collect meteorological data in irregular intervals,
    with increments of approximately one hour. To create hourly synthetic time series, we
    used the simple average of the two measurements surrounding each hour. Next we
    normalized wind speeds to Class 4 speeds by scaling the time series for each station to
    an average speed of 16.25 mph, as follows:

                                  we     woo   16.25
                                  wst. = wst. * -=
                                      '•>      *•>    TJ7"O
    Where:

    WSjj = Adjusted wind speed during hour; and station j

    WSฐij= Unadjusted wind speed during hour; and station j
    and WSj = Average unadjusted wind speed at station j over all hours.

    Wind scales non-linearly with height, increasing in speed as the influence of ground-level
    friction decreases. However, we assume that a proximal Class 4 wind site would have
    similar temporal characteristics to the meteorological station, and thus our approximation
    is to scale the WBAN wind data linearly to reflect Class 4 characteristics. This approach
    encompasses a host of decisions, design parameters, and operational characteristics into
    a single scalar that represents, roughly, the decisions made by the builder of a wind
    power facility to optimally exploit a local resource.
                          Indirect Emissions Analysis • 8

-------
       To convert the scaled wind data into synthetic wind power data, it was necessary to
       select a proxy technology with an appropriate wind power conversion function. We used
       a wind power curve based on the General Electric (GE) 1.5se (1.5 MW nameplate
       capacity) wind turbine to estimate potential generation associated with the wind time
       series at each selected meteorological station. The power curve, shown in/igure 2,, is
       determined by a number of factors including the area swept by the blades, the "pitch" or
       angle of the blades (which can be variable in different wind conditions), and
       characteristics of the turbine gearbox. Wind turbines require a minimum amount of wind
       to begin generating power, known as the cut-in speed, and are designed to stall, or even
       brake, at higher wind velocities (the cut-out speed) to avoid damage.11 Once again, we
       do not know a priori what technology or set of parameters would be most appropriate for
       a wind turbine in each area, but by scaling the winds to match a proxy turbine we
       generate a series that preserves temporal variability while representing a project which is
       optimized to local conditions.

              Power Curve
                 1,800
                           246

                         -GE 1.5 xle
 8  10  12  IH

-GE1.5sl/sle
 16   18   20   22   24

— GE 1.5s/se  m/s
Figure 2. Wind power curve fora GE 1.5se (1.5 MW) wind turbine.
Source: Http://www.gepower.com/prod_serv/products/wind_turbines/en/downloads/ge_15_brochure.pdf

       The GE 1.5se proxy turbine is a common mid-sized turbine used on-shore; larger
       turbines are generally used offshore due to high wind conditions. The design tower height
       for these turbines is 54.7 meters (m), with 35.25 m blades sweeping an area of 3,904 m2.
       The cut-in speed is 4 meters per second, and the cut-out speed is 25 meters per second.
       To generate the wind power time series, we relied upon a GE-supplied power curve
       which was imported as a look-up table for our analysis; the look-up table indexes the
       power curve and returns an estimated generation value for each hour given the
       calculated wind speed at turbine height.
  Tony Burton, David Sharpe, Nick Jenkins, Ervin Bossanyi, Wind Energy Handbook, John Wiley & Sons, 2001.
                              Indirect Emissions Analysis • 9

-------
       For purposes of our analysis, we approximated the GE wind power curve with a sigmoid
       function:
                                      Q =
                                              a
                                          l + e
                                               b-cWS
(2)
       Where Q is output (kW) and l/l/S is wind speed (m-s~). The best fit to the GE curve in
       these units was obtained with parameter values a=1500, b=8.1, and c=0.9. The curve
       was truncated to zero at the cut-out speed of 25 m-s"1 The resulting curve is shown in
       Figure 3. along with reported and scaled wind data.
20% n
18%
16%
14%

0 12%
.n
o 10%
ง 8%
1
6%
4%
2%
0%

n


1
ll . ^
1 234567

r
r n

f
[4t J.
/

[




1


/


fl



X



•




i 	 1 Raw Wind Speed
l=l Scaled Wind Speed
— Power Curve

J J Lfl 0 H n n „ „ 	
8 9 10111213141516171819202122232425262728
- 1800
- 1600
o
- 1400 2
01
c
Ol
- 1200 o>
- 1000 m _
-800 | ~
II—
- 600 i.
-400 %
- 200 ฃ
- 0
Wind speed (mph)
Figure 3. Proxy wind turbine power curve plotted against histograms of raw and scaled wind speeds at a
sample site in the Midwest. "Scaled" wind speeds are adjusted to reflect class 4 winds at a turbine height of
50-100 meters using Equation 1.

    Comparing Empirical and Simulated Wind Generation
       As noted earlier, we were able to find one source of wind power data for comparison with
       our synthetic power profiles. The wind power time series that we were able to obtain
       represent twelve operational projects associated with the NREL wind farm monitoring
       program. The NREL data cover twelve sites including two in Minnesota, one in Iowa, one
       in Oklahoma, four in Texas, and four in Oregon. Total capacities ranged from 25 MWto
       230  MW, and the turbines ranged from 0.6 to1.6 MW each. While  NREL was able to
       make 2005 production records from these  projects available for research purposes, site-
       identifying information remains confidential. The turbines represent primarily plain and
       low-ridge wind farms; the farms in Iowa and Minnesota are distributed throughout active
       agricultural land, while the Oregon, Oklahoma, and Texas sites partially overlap grazing
       lands.
                             Indirect Emissions Analysis • 10

-------
To test whether our simulated wind generation time series were realistic and
representative of the output profile for a physical wind project at the same location, we
selected an NREL wind project from a midwestern U.S. location and compared the
recorded output with the simulated production time series generated using wind data
from a nearby location. The test project is in a region with Class 4 and Class 5 wind
speeds, running 145 turbines of 750 KWeach, fora  total capacity of 108 MW. A nearby
WBAN meteorological station was located, and wind records for this location extracted
from the NCDC data. The wind data were transformed through the power curve (Equation
2) to generate hourly power output.

Such a comparison has a number of complications, including:

•   Gaps in wind data. In the case of the selected series, there are 84 gaps larger than
    2 hours in duration, and 9 gaps longer than 24 hours in duration. The longest gap is
    nearly a week in duration. We have no way of knowing whether these were neglected
    because they were zero, or they were very high, or they simply represent random
    data errors, Thus it is impossible to  determine what effect their omission may have  on
    our analysis.
•   Gaps in power output data. The output time series for the installed project has 552
    missing hours in the sample year (6% of all hours) which were removed from the
    analysis. As with the missing wind data, it is impossible to determine what effect their
    omission may have on our analysis.
•   Granularity of wind data. Wind speed data are taken approximately once per hour,
    meaning that variations in wind speed within each  hour are unknown. In addition, the
    wind data show recording artifacts, such as certain values that rarely appear in the
    record including  any measurements between approximately 0.5 and 2.75 mph. We
    have not attempted to perform any smoothing to compensate for this.
•   Representativeness of wind site.  While the wind data  are taken  from a
    meteorological station in proximity to the turbine, the station is clearly not measuring
    winds identical to those that drive the turbine. This lack of correspondence to actual
    values is exemplified by the average wind speed in the meteorological record of only
    9.71 mph; to scale to the Class 4 average of 16.25 mph  winds, we multiply all values
    by 1.67, consistent with Equation 1. In addition, a wind power installation can contain
    hundreds of turbines, each responding to highly  localized conditions and each
    subject to maintenance schedules and other operational constraints which are  not
    captured in the power curve.
•   Idealization of power curve transformation. Each hour's scaled wind speed  is
    transformed into an hour of power output assuming a perfect generator response to
    steady winds,  neglecting any effects such as generator inertia which may dampen
    the response.  Wind speeds are implicitly assumed to hold steady for exactly one hour
    before instantly changing to the next hour's value,  again with an instantaneous
    response from the turbine.
The comparison of simulated and observed power output is  shown in ^Figure \ In this
analysis we take the maximum recorded output for the year  as the maximum output for
the station, although this value is only about 90% of the nameplate capacity of the
                      Indirect Emissions Analysis • 11

-------
       installation. This assumption presumably corrects for scheduled maintenance and forced
       outages which would affect some subset of turbines at all times.

       The simulated output curve shows an exaggerated peak at 5%-10% of capacity relative
       to the recorded power output series, and a peak at the 100% output level which is absent
       in the observed data. These peaks likely reflect the granularity issues identified above for
       the wind records, which can reflect just one or two observations for each hour while the
       measured power output reflects inevitable variations in wind  speed and output over the
       course of each hour. For the simulated project the overall capacity factor over the course
       of the study period is 32.8%, while for the observed project the overall capacity factor is
       30.3%.
    30% -,
    25%
    20%
   ฐ  15% -

   I
   fe  10% -
  Q.
     5% -

     0%
                                                              D Simulated Generation
                                                              • Obserwd Generation

                                                     O\O ,O\O
                                      Percent of maximum output
                                                                    O\O ,O\O
Figure A. Comparison of power output distribution at sample wind power site in the Midwest with predicted
output based on scaled wind data from a nearby meteorological station. Output levels for observed
generation are relative to the maximum recorded output for the year.

       The simulated and observed generation, as well as the difference between the two, are
       shown on color charts in/igure 5^ These charts represent each hour of the year as a cell
       on a 24-hour by 365-day matrix, with each cell color-coded according the scheme shown.
       ^Figure e^shows a cumulative distribution of the hourly deviation between the two, in
       percentage points. The simulated output exhibits a pronounced diurnal cycle which is not
       present in the project data. This discrepancy could be due to any number of factors,
       including light winds which blow through the evening at hub-heights which do not appear
       in the NCDC dataset (NCDC does not record wind speeds less than 3 mph at ground
       level) or more sensitive wind turbines with lower stalling speeds than the proxy turbine,
       which would produce a generally flatter distribution of power generation throughout the
       day. This discrepancy may also simply reflect the difference in the location of the two
       sites. The two time series  are within 20 percentage points of each other during
       approximately 65% of the  6,434 hours recorded in both datasets.
                              Indirect Emissions Analysis • 12

-------
      A) Project generation (% of capacity)
      B) Expected generation from scaled wind speed (% of maximum)
 I  15
    20
Jai      Feb     Mar      Apr     May      Jun      JiJ      Aug

                                          Percent scale (A-B)
                                                                    Sep
                                                                            Oct
                                                                                    Nov
                                                                                            Dec
              10
                        20
                                 30
                                           40
                                                    50
                                                              60
                                                                       70
                                                                                 80
                                                                                          90
                                                                                                   100
     C) Difference: expected generation (% of maximum) minus observed project power (% of capacity)
    5
   10
   15
   20
    Jai
            Feb    Mar     Apr     May
                                       Jun      Jul     Aug
                                          Percent scale (C)
Sep
                                                                           Oct
                                                                                   Nov
                                                                                           Dec
   -100       -80       -60       -40       -20        0        20        40        60        80        100
Figure 5. Comparison of hourly synthetic and observed wind power time series at nearby locations in the
Midwest. Each plot shows 8,760 hours in 2005, with hour of the day on the vertical axis and day of the year
on the horizontal. Missing data shown in black. Top: Observed wind power output (% of maximum); Middle:
Synthetic wind power output (% of capacity) based on scaled wind speed data; Bottom: Hourly deviation of
synthetic from observed time series.
                             10%
                                   20%
                                         30%   40%   50%   60%   70%
                                            Non-Exceedence Hours (%)
                                                                            90%   100%
Figure 6. Cumulative distribution of the hourly deviation (in absolute percentage points relative to maximum
output) between observed and synthetic generation time series.
                                  Indirect Emissions Analysis • 13

-------
        While there are clearly significant discrepancies between these two datasets, we feel that
        the NCDC meteorological dataset is a reasonable, if by no means ideal, resource for
        estimating temporal patterns and potential offsets from projects which might be built in
        each eGRID region.

        The hourly synthetic output generated for each of the eGRID subregions is shown in
        Appendix B as 24-hour by 365-day color intensity plots. Monthly summary tables of raw
        and scaled wind speeds, as well as synthetic wind power, may be found in  Appendix C.
        The hourly data are available in electronic format upon request.

B.  Landfill Gas
        To determine the operational and dispatch characteristics of landfill-gas-fired generating
        facilities, we contacted several landfill gas (LFG) project developers.12 We received
        hourly generation data for several landfill gas projects from Granger Electric Company
        and Granger Energy, LLC, ("Granger") which together own and operate six landfill gas
        projects in Michigan, and  10  LFG projects in Indiana, Ohio, and Pennsylvania.

        According to Granger, the generation profile of LFG projects generally does not differ
        significantly by time or season.13 The primary factors that impact the energy output of
        LFG projects have to do with ambient conditions. The warmer and the wetter the climate,
        the higher the methane content of the landfill gas, such that more power can be produced
        from this resource.

        High temperatures  also affect the plant's output by generally reducing the temperature
        gradient that drives the turbine. This temperature effect is minimal for the reciprocating
        engines which are used by Granger. Gas turbines appear to be the predominant choice
        for LFG projects by other companies. Gas turbine outputs are more susceptible to
        variation in temperature. This characteristic largely mirrors the behavior of base load
        generating plants, and system  operators can reasonably expect a constant and
        predictable stream  of energy output from these facilities. Thus we conclude that the
        fossil-fuel-fired generation displaced by landfill gas in any region of the country will be
        similar in operational and emissions characteristics to typical base load resources in that
        same region.

C.  Municipal  Solid Waste
        Municipal Solid Waste (MSW) generators, also known as waste-to-energy facilities, are
        currently operating  in 27 states and burning roughly 95,000 tons of garbage each day,
        generating roughly 2.5 GW of electricity.14'15 Even with this large volume of output, the
  Landfill gas project operators were identified by Rachel Goldstein of the EPA Landfill Methane Outreach Program:
  http://www.epa.gov/lmop/.
  Granger does have one project which operates as peaking generation, but this appears to be an anomaly and does not
  affect our conclusions about general LFG generation characteristics.
14
  New York State Department of Environmental Conservation, http://www.dec.nv.aov/chemical/8979.html.
  Because some portion of the fuel supply for MSW generators is generally fossil-based, there is some dispute over
  whether this resource should be considered "renewable" or not. However, as a number of state programs include MSW
  as a renewable energy source (see dsireusa.org for updated details), we treat it as a renewable resource for the
  purposes of this report.
                              Indirect Emissions Analysis • 14

-------
       facilities are not fuel-limited; our research and discussions with facility operators reveals
       that waste-to-energy plants run at high capacity factors of 85% and above. As with landfill
       gas projects, these characteristics suggest that they serve as base load capacity, and
       that the fossil generation they are likely to displace is also base load capacity. Thus we
       conclude that the fossil generation displaced by MSWin any region of the country will be
       similar in operational and emissions characteristics to typical base load resources in that
       same region.
3.  Hourly generation and emissions data
       In this section we characterize the operational, dispatch, and emissions characteristics of
       conventional power plants in order to predict the change in dispatch and emissions
       resulting from the presence of new renewable energy resources.

       Electricity grids are often characterized as having a range of resource types including
       base load, intermediate, and peaking resources. These resources respond differently to
       variations in load over different timescales: put simply, base load resources run most
       hours regardless of load levels, intermediate resources ramp up and down frequently in
       response to changes  in system load, while peaking resources run only during the highest
       load periods. However, real-world electric grids are more complicated than this, and real-
       world resources have complex operating constraints and practices depending on factors
       including heat rate, ramping capability, demand and price in the energy and ancillary
       service markets, availability of competing resources, local transmission constraints,
       environmental constraints, operator or dispatcher discretion, and even the warrantee
       and/or service contracts on specific pieces of generating equipment. It is thus impossible
       to predict with specificity which resources would respond to perturbations in load or
       available low-cost  energy. However, historical data on unit operations can lend insight
       into which units are more or less likely to be displaceable at any point in time.16 To the
       extent that these operational data are coupled with emissions characteristics, they can be
       used to estimate the likely displaceable emissions at any point during a historical year.
       Such an approach would be more refined than using a simple average emission rate at
       any point in time to represent displaceable emissions,  because it would be more sensitive
       to the actual dispatch characteristics of individual generating plants.

       To implement this  approach, we obtained generation and emissions data for United
       States fossil-fuel powered generating plants for 2005 from the Hourly Emissions
       database of the EPA Clean Air Markets Programs.17 The database consists of hourly
       reported generation, heat rate, and total emissions of CO2, SO2, NOx, and mercury for all
       fossil fuel-fired generators with nameplate capacities above 15 MW. These data form the
       basis of all of the displaceable emissions analysis reported here. We use the hourly
  It is common to think of generating resources as responding only to changes in electricity demand; however, certain
 intermittent renewable resources, such as wind generation, change the load on fossil resources in a very similar manner
 to variations in load. Thus we treat the availability of energy from such resources in the same manner that we would a
 reduction in energy demand from consumers.
  EPA, 2007. Data Sets and Published Reports: Emissions Data Prepackaged Data Sets. Online at
 http://camddataandmaps.epa. gov/gdm/index.cfm?fuseaction=prepackaged. select
                              Indirect Emissions Analysis • 15

-------
    generation data to characterize the operational characteristics of generating units in each
    NERC subregion, and we use the associated emissions data to characterize the indirect
    emissions benefit that would be associated with displacing the output of each generating
    unit.

Characterization of emission rates
    Hourly emissions are reported in  the Clean Air Markets database for CO2, NOx, and SO2.
    Within the database, the data records include fields to indicate whether the reported
    emissions are measured or calculated. Calculated emissions are based on the
    composition of the fuel, the unit's heat rate, and the output of the unit. In general, it
    appears that most of the emissions data in the database for 2005 are measured.

    Pollutant emission rates are reported in the hourly emissions database as tons per million
    British thermal units (mmBtu) for  CO2, and pounds per mmBtu for NOx and SO2. For the
    purposes of this analysis it is necessary to consider emissions in units of mass per MWh
    output. However,  while emission  rates per unit of heat input may be fairly constant,
    emission rates per unit of energy output can vary considerably—this variation is due to
    changes in the efficiency (or "heat rate") of a generating unit with different output levels.
    For purposes of this analysis, we will use the overall average emissions rate for each unit
    as derived from the EPA hourly emissions database,  unless otherwise noted.

Data filtering
    The EPA emissions data contain  anomalous features which are indicative of either
    reporting or calculation errors. While the hourly generation data are used to characterize
    generator behavior and identify times at which the generator is operational, emission
    rates could be misleading if anomalous data were included.  Therefore, data are removed
    from the emission rate analysis according to the following rules:

    •  Generation, CO2,  or heat rate is reported as zero.
    •  The reported  heat rate is in the top 99.5th or bottom 0.5th percentile  (for a record with
       8,760 non-zero hourly values, this would remove the 44 highest and 44 lowest
       values).
    •  CO2 emission rate is above 2.5 tons of CO2 per megawatt-hour (tCO2 MWh"1) or less
       than zero.
    We did not filter on the basis of NOx or SO2 measurements  as we had no basis for
    judging reasonable emission rate limits.
                         Indirect Emissions Analysis • 16

-------
4.  Emissions Displacement  Methods
        In order to calculate the indirect emissions benefits associated with power production
        from renewable resources, it is necessary to determine which power plants will have their
        output curtailed as a result of the availability of the new resource. This need arises from
        the recognition that the emission rates of power plants which will be running in either
        case (with and without the new resource) do not affect the calculation; nor do those of
        power plants that would not be running in either case. In fact, it is only a small number of
        plants, as few as one marginal resource in any given dispatch  interval, that will determine
        the relevant emission rate. The difficulty comes in identifying which plant or plants to
        consider, and thus in identifying the appropriate emissions rate or rates for each interval.

        There are numerous factors that affect which power plants operate, are committed to
        operate, or are held in reserve at any point in time. These factors include:

        •   Offer price or short-run marginal cost;
        •   Constraints such as ramp rates, minimum up and down times, and other operating
           restrictions;
        •   Transmission or other dispatch constraints;
        •   Maintenance schedules;
        •   Unplanned outages and/or deratings; and
        •   Environmental constraints.
        Unfortunately, most of this information is not available from public sources for most
        resources. Using only the information which we have available from the EPA database
        (i.e., hourly generation, heat rate, and emissions),  it is impossible to reconstruct all of the
        factors that guide dispatch and to determine the precise marginal emissions rate in effect
        for any particular hour.

        Our solution is to infer operational characteristics from the suite of units in each region
        based on their patterns of operation and emissions over the course of the study year.
        There is a broad range of possible approaches to interpreting the data in this manner.
        Some of these approaches use emissions characteristics aggregated over a season or
        year, and some are more directly focused on hourly analysis.18 Each approach carries its
        own implicit model about system operation and response to new resources, and each
        may be more appropriate than others to address the impact of certain  kinds of resources
        or over specific time scales. In this analysis, we present five methods,  ranging in
        simplicity, design, and conceptual model, and explore the implications of each. These
        methods are:
  Earlier investigations based on extracting hourly performance information from the same EPA database used here:
           •   Connors, S., K. Martin, M. Adams, E. Kern, and B Asiamah-Adjei. 2005. "Emissions Reductions from
               Solar Photovoltaic (PV) Systems" Publication MIT LFEE 2004-003 Report.
           •   Berlinski, M. and S. Connors, "Economic and Environmental Performance of Potential Northeast
               Offshore Wind Energy Resources: Final Report", Report to the Offshore Wind Collaborative Pilot
               Research  Projects, January 2006.
                              Indirect Emissions Analysis • 17

-------
        •   Hourly Average Emissions Rate (HAER);
        •   Annual (and seasonal) emissions slope factor (Slope Factor);
        •   Empirical Incremental Emissions Rate (EIER);
        •   Load-Following Rate Incremental Emissions Rate (LFIR); and
        •   Flexibility-Weighted Hourly Average Emissions Rate (FW-HAER).

A.  Limiting Cases
        Before delving into each method in detail, it is useful to characterize the most extreme
        emission scenarios to bound the results of our analysis, particularly with respect to CO2.
        In producing a single MWh of electricity, power plants fueled by coal (the most carbon-
        intensive fossil fuel) typically produce somewhat over one ton of CO2; modern plants
        fueled by natural gas produce about 0.6 tons of CO2 per MWh.19 The exact emission rate
        for each plant depends on technology, age, maintenance, and specific fuel (i.e., some
        coals have a higher carbon content than others for the same energy output.) The ranges
        of reported hourly emission rates for gas, coal, and oil-fired plants in the EPA database
        are shown in/igure 7r While this figure highlights some of the anomalous data in the
        database, such as coal plants which report unrealistically low CO2 emissions, it does
        show that the reported data are generally consistent with  the expected range of emission
        rates by fuel type.
             35%
             30%
              0%
                            0.4
                                 0.6
                                       0.8    1     1.2
                                        Tons per MWh
                                                       1.4
                                                            1.6
                                                                  1.8
        Figure 7. Distribution of CO2 emission rates for fossil generating plants by reported fuel type.
        Note that reported primary fuel is not a precise categorization, as some units are capable of
        switching fuels and some plants may actually have units which burn different primary fuels.

        The difference between emission rates from coal and gas essentially bounds the range of
        outcomes in accounting for the displaced emissions associated with renewable
19
  http://www.eia.doe.gov/cneaf/electricitv/page/co2 report/co2emiss.pdf:. Values in the EPA database confirm these numbers:
  the median emission rate for a coal plant is 1.024 tCO2 MW1, while for a gas plant it is 0.697 tCO2 MW1
                              Indirect Emissions Analysis • 18

-------
       generation. Renewable resources cannot displace more carbon emissions than the
       emission rate of an inefficient coal plant on a per-MWh basis; nor is it likely that they will
       displace less CO2 for each MWh generated than the amount emitted by an efficient
       combined cycle gas plant. This is because other low-carbon sources of power (such as
       nuclear and hydro) have low running costs and are unlikely to be displaced.2021 In
       general, the result will lie somewhere between these two extremes.

       For NOx, the story is somewhat more complicated. NOx emissions tend to be bifurcated
       into ozone-season emissions (May through September), when emissions are regulated
       and many power plants are operating emissions control technology; and non-ozone
       season emissions, when they are not. As will be shown below, in some areas this
       bifurcation is more pronounced than in others. However, NOX emission rates generally
       range from zero to perhaps five or six pounds per MWh, with ozone season rates
       generally one or two pounds below  non-ozone season rates.

       SO2 emission rates depend upon the fuel used for generation, with natural gas having  no
       sulfur (and therefore producing no SO2) and certain types of coal having the most. Thus if
       only gas generation is displaced, there is no displacement of SO2 emissions. For this
       reason, we would expect SO2to have perhaps the largest regional displaceable
       emissions differences of all of the pollutants considered, with ranges from zero up to
       about eight pounds per MWh for regions that rely on coal  resources for load following
       capacity.

       To narrow these ranges, it is necessary to find a means to identify which types of plants
       are more likely to be displaced  in  any given region. All of the methods described in this
       section are designed to refine the estimate of displaced emissions based on the
       observed operational characteristics of the system as deduced from the EPA database.

       For clarity, we will develop the displacement methodologies with primary reference to
       carbon dioxide emissions since it is the primary anthropogenic cause of global  warming.
       However,  results will be presented for all three of the pollutants considered here.

B.  Hourly Average Emissions  Rate (HAER)
       One of the most conceptually straightforward methods for estimating displaced emissions
       is to use a simple hourly average. This method calls for adding the aggregate emissions
       from an eGRID subregion (for example, in tCO2) for each  hour of the study period and
       dividing this by aggregate fossil generation (MWh) for the  same hour:
                               HAER, =   - —                            (3)
  As noted earlier, displaced emissions are not necessarily equivalent to avoided emissions, as emissions policies and
  other factors can come into play in determining total emissions in any given area.
21
  There are exceptions to this statement - in some cases storage hydropower may be the most flexible resource
  available to follow load. However, storage hydro is energy-limited resource, so total generation from such a facility will
  not be affected by the presence of a new renewable resource.
                             Indirect Emissions Analysis • 19

-------
Where t denotes a particular hour, and the summations are over all of the generating
units in the region of interest.

Equation 3 produces the average emissions rate across all fossil units for a specific
region and hour, with the implicit assumption that every fossil plant running during that
hour has an equal probability of being displaced. This is clearly not the case on an hour-
to-hour basis: certain units have limited ramping capability and other operational and
economic constraints which dictate that they only be used as base load resources.
However, this method is fairly straightforward to understand  and apply, and as noted
earlier we cannot know for sure which units would be displaced in any hour. In addition,
as the market matures with higher levels of renewables penetration, it is reasonable to
imagine that units throughout the supply curve may ultimately be displaced.  Finally, this
method may be the most appropriate for base load renewables such as landfill gas or
municipal solid waste combustion, for which output is stable  and predictable. Thus it may
be reasonable to quantify displaced emissions using this approach if more of a long-term
or base load impact is under consideration for any of these reasons.

^Figure 8,.shows selected representations of the variations in  HAER for CO2 in the New
England (NEWE) eGRID subregion. The first representation  is a timeline of hourly
variations, demonstrating that the CO2 emissions rate fluctuates between  about 0.65 and
0.95 tCO2 per MWh, with an average of 0.73 tCO2/MWh and a standard deviation of 0.05
tCO2/MWh. the  second panel from the top shows a three-week period in the summer in
more detail. The second representation (third from top) is a histogram of hourly HAER
values. The final representation (bottom panel) recasts the dataset as a color intensity
image of HAER, with day of year (DOY) in 2005 along the horizontal and the hour of the
day on the vertical. The green / yellow band across the top and bottom of the image
indicates that the evening hours have higher average emission rates, probably reflecting
the lower contribution of natural gas units during these hours.

We can see from these graphs that there is significant diurnal variability to the data over
a fairly narrow range, with higher hourly average emission rates at night. This reflects the
low position in merit order for coal plants with higher CO2 rates, which dominate the
lower-usage hours at night, but also the fact that these same large plants  strongly
influence the overall average, in combination with lower-emitting gas plants, even during
peak times.
                      Indirect Emissions Analysis • 20

-------
  >
                                        Hourly Average Emissions Rate (HAER) in New England (NEWE)
       Jan       Feb      Mar      Apr      May       Jun       Jul       Aug       Sep      Oct       Nov      Dec       Jan
     0.9
  8  0.8
  QL  0.7
  LU
  I  0.6
      07/18
                 07/20
                            07/22
                                      07/24
                                                 07/26
                                                            07/28
                                                                       07/30
                                                                                  08/01
                                                                                             08/03
                                                                                                        08/05
                                                                                                                   08/07
IOUU
1ROO -
i4nn
o
o -lonn
C
(A
"- 1000 -
o
"s snn
o
^2
ERnn
z
400 -
9nn
n -










































1
ll..
                                         Emissions Rate Bin (tCO2/ MWh)


                                       Hourly Average Emissions Rate (HAER) in New England (NEWE)
                                                 150            200
                                                     Day of Year (2005)
Figure 8. Representations of CC>2 Hourly Average Emissions Rate (HAER) in New England. Top: full record,
with detail shown for three weeks in the summer of 2005; Middle: Histogram of hourly emissions rates
showing distribution throughout the year; Bottom: 24x365 color representation of emission rates, showing
diurnal cycle and seasonal variations.
                                   Indirect Emissions Analysis • 21

-------
C.  Emissions Slope Factor
       The emissions slope factor is an empirical estimate of the marginal emissions rate based
       on the assumption that the observed linear relationship between emissions and electricity
       output from fossil fuel units is the best indicator of the system's response to a change in
       load. This factor can be calculated as a simple regression relationship over an entire
       year, or on a seasonal basis. Because of seasonal differences in dispatch and operating
       constraints, there may be additional value in performing this analysis on a seasonal
       basis.  For the time period under consideration, total emissions and total fossil generation
       for each subregion are summed for each hour, based on the EPA database. When the
       hourly pairs  of emissions/output data points for a subregion are plotted on a scatter plot,
       the slope  of the regression line represents an empirical measure of the dependence of
       emissions on total electricity output in units of mass of pollutant emitted per MWh
       produced. If the correlation is high, this slope provides a reasonable estimate of the
       regional, seasonal avoidable emissions factor for a change in load or for the addition of
       emissions-free energy. This  metric makes no specific assumptions about which generator
       will be displaced by a change in system load, but instead relies on a robust diagnostic of
       the overall system response to variations in load.

       One interesting feature of this approach is that the line of best fit rarely has a y-intercept
       of zero. This may seem counterintuitive, as zero generation would surely be associated
       with zero emissions. However, zero generation is clearly a domain which is not of interest
       for this analysis, and the very lowest generating units on the supply curve are likely to
       have emissions characteristics that diverge from those of units that are more likely to be
       load-following. Were we to force the regression lines to cross through the origin, we
       would  obtain a much poorer estimate of the emission factor in each region.

       ^Figure ^illustrates the application of this method for emissions in the Reliability
       First/Central (RFCE) subregion. All three pollutants of interest are shown. CO2 and SO2
       emissions can be characterized as having a linear dependence on total MW output, with
       R2 values of 0.98 and 0.81, respectively.

       The NOx emissions data (Figure 10) do not lend themselves to this simple relationship.
       Readily apparent from the  Figure are two different trends. These trends represent the
       emissions in the ozone season, when most generating units are required to operate
       ozone control technology,  and the non-ozone season when they are not.22 Finally, there
       appears to be an up-tick in NOx emissions during very high load hours during the ozone
       season. This up-tick may represent units that disengage their NOx controls to increase
       output during very high load  hours.
22
  Complicating this analysis is the fact that many units were not required to report NOx output during the non-ozone
 season in 2005.
                             Indirect Emissions Analysis • 22

-------
            450,000
            400,000
          — 350,000
                   0    5,000 10,000  15,000 20,000  25,000 30,000 35,000 40,000 45,000
                                           MW Output
Figure 9. Determination of the regional Emissions Slope Factor in the RFCE (Reliability First/Central) region.
Total emissions of CC>2, NOx and SC>2 vs. MW output for each hour of 2005 are shown. A linear line of best
fit is calculated for each pollutant, and the slope of this fit determines the annual slope factor for the region.
The bifurcation of the NOx data reflects the differential operations of pollution control equipment during the
ozone (summer) vs. non-ozone seasons.
            120,000
                   0     5,000  10,000  15,000  20,000  25,000  30,000  35,000   40,000  45,000
                                               MW Output
Figure 10. Bifurcation of the NOx emissions slope factor (pounds of NOX per MW output) for the RFCE
region. Slopes for the ozone season (May through September) and non-ozone season hours are shown.
The difference reflects the required operation of ozone control technology during the ozone season. Also
apparent are excess NOx emissions during high-load hours during the ozone season.
                                 Indirect Emissions Analysis • 23

-------
       These figures and statistics suggest that the empirical slope factor approach is a well-
       defined and reasonable way to characterize the variation of emissions with load in this
       region, especially if some care is taken in subdividing the load to isolate meaningful
       trends. Unfortunately, the regression relationships are not as constrained for all pollutants
       and all regions. Figure  1 shows the seasonal and annual slope factors for CO2, and Table
       2,for SO2, for each eGRID region. Table ^shows analogous results for NOx separated
       into the ozone and non-ozone seasons. Also  shown are the corresponding R-squared
       metrics indicating the strength of the regression relationship underlying each slope factor.
       The R-squared values for CO2 are uniformly quite high, indicating strong confidence in
       these calculated slopes. For NOx the relationships are also generally quite strong, but for
       SO2 the relationships range from very strong  to extremely weak. This diagnostic should
       be taken into account in applying this approach.

       In most cases, the differences between the seasons are small, and some of these
       relationships require  significant error bars. As a rule we apply seasonal slopes in
       performing the emissions analysis,  but it is important to keep in mind that we do so with
       more confidence in some areas and for some pollutants than for others.

       Summary results for  the Slope Factor method for all eGRID subregions, including factors
       calculated forCO2 (tons/MWh), SO2 (Ibs/MWh), and NOx (Ibs/MWh) are displayed in
       ^Figure 11rfigure 12, andfiqure 13, respectively.

Table 1. Seasonal and annual emissions slope factors (left) and R-squared values (right) forCO2. Spring =
March-May; Summer = July-August; Fall = September-November; Winter = January, February, and
December.

Region
AZNM
CAMX
ERCT
FRCC
MROE
MROW
NEWE
NWPP
NYCW
NYLI
NYUP
RFCE
RFCM
RFCW
RMPA
SPNO
SPSO
SRMV
SRMW
SRSO
SRTV
SRVC
CO2 Slope Factor
Spring
0.59
0.56
0.69
0.73
0.80
0.97
0.63
0.82
0.66
0.55
0.69
0.78
0.81
0.95
0.74
0.93
0.70
0.66
0.93
0.83
0.81
0.94
Summer
0.53
0.50
0.59
0.69
0.77
0.88
0.65
0.71
0.55
0.48
0.66
0.69
0.67
0.85
0.70
0.93
0.60
0.60
0.86
0.70
0.76
0.79
Fall
0.56
0.53
0.64
0.73
0.83
0.93
0.61
0.78
0.63
0.61
0.60
0.75
0.75
0.89
0.86
0.96
0.66
0.64
0.95
0.79
0.87
0.88
Winter
0.55
0.48
0.67
0.72
0.96
0.91
0.68
0.70
0.64
0.51
0.76
0.81
0.93
0.92
0.89
1.08
0.65
0.64
0.99
0.80
0.88
0.90
Annual
0.58
0.52
0.66
0.74
0.83
0.95
0.65
0.82
0.56
0.53
0.66
0.76
0.71
0.89
0.79
0.97
0.66
0.61
0.94
0.80
0.84
0.85
R-Squared
Spring
0.95
0.91
0.94
0.98
0.94
0.98
0.94
0.97
0.83
0.84
0.95
0.99
0.81
0.98
0.85
0.94
0.87
0.94
0.99
0.99
0.97
0.97
Summer
0.97
0.99
0.99
0.99
0.97
0.99
0.98
0.95
0.97
0.90
0.97
0.99
0.97
0.99
0.92
0.97
0.98
0.97
0.99
0.98
0.97
0.99
Fall Winter
0.94
0.95
0.99
0.99
0.95
0.98
0.97
0.90
0.98
0.97
0.92
0.98
0.91
0.99
0.90
0.99
0.97
0.94
0.99
0.97
0.98
0.98
0.91
0.92
0.97
0.99
0.94
0.94
0.88
0.87
0.91
0.82
0.95
0.98
0.94
0.99
0.93
0.97
0.89
0.89
0.98
0.96
0.97
0.99
Annual
0.96
0.97
0.97
0.99
0.95
0.98
0.95
0.95
0.93
0.92
0.95
0.98
0.93
0.99
0.90
0.97
0.94
0.96
0.99
0.98
0.97
0.98
                             Indirect Emissions Analysis • 24

-------
Table 2. Seasonal and annual emissions slope factors (left) and R-squared values (right) for SO2.

Region
AZNM
CAMX
ERCT
FRCC
MROE
MROW
NEWE
NWPP
NYCW
NYLI
NYUP
RFCE
RFCM
RFCW
RMPA
SPNO
SPSO
SRMV
SRMW
SRSO
SRTV
SRVC
SO2 Slope Factor
Spring
0.29
0.31
1.49
5.68
4.65
6.59
3.04
2.15
1.51
2.94
6.11
12.99
7.40
13.66
0.06
5.91
1.52
1.41
6.69
8.75
6.56
10.84
Summer
0.19
0.03
0.51
5.08
3.59
4.36
3.26
1.07
1.01
1.59
5.98
6.39
4.77
10.38
1.00
7.02
1.03
1.53
6.58
5.26
6.36
7.29
Fall
0.34
0.33
0.80
5.30
4.18
5.06
1.87
1.25
1.35
3.01
4.93
8.58
6.85
12.02
2.03
7.84
1.52
1.53
6.64
7.28
7.19
9.76
Winter
0.07
0.16
1.52
5.70
6.20
5.22
4.33
0.97
2.07
3.05
7.00
9.73
9.68
14.73
0.82
7.26
2.13
1.91
9.02
9.14
9.23
10.24
Annual
0.06
0.18
0.86
5.19
4.15
4.95
2.98
1.85
0.99
2.07
6.14
8.58
5.87
11.00
1.11
6.59
1.43
1.52
7.08
7.57
7.07
8.60
R-Squared
Spring
0.06
0.05
0.32
0.94
0.70
0.82
0.38
0.74
0.47
0.33
0.84
0.87
0.72
0.76
0.00
0.45
0.21
0.35
0.72
0.81
0.45
0.90
Summer
0.10
0.01
0.16
0.92
0.68
0.82
0.78
0.28
0.70
0.54
0.88
0.79
0.78
0.90
0.19
0.85
0.43
0.62
0.83
0.64
0.70
0.89
Fall Winter
0.11
0.10
0.26
0.93
0.34
0.75
0.27
0.09
0.77
0.75
0.76
0.78
0.79
0.92
0.49
0.90
0.53
0.32
0.89
0.67
0.72
0.90
0.00
0.03
0.27
0.79
0.79
0.74
0.63
0.05
0.83
0.55
0.79
0.85
0.72
0.93
0.15
0.79
0.41
0.36
0.70
0.75
0.72
0.92
Annual
0.00
0.07
0.25
0.91
0.47
0.80
0.40
0.47
0.51
0.50
0.85
0.81
0.72
0.84
0.19
0.79
0.38
0.47
0.79
0.77
0.66
0.89
                                Indirect Emissions Analysis • 25

-------
Table 3. Ozone season and Non-Ozone season emissions rate slopes for NOx. Ozone season = May
through September.


Region
AZNM
CAMX
ERCT
FRCC
MORE
MROW
NEWE
NWPP
NYCW
NYLI
NYUP
RFCE
RFCM
RFCW
RMPA
SPNO
SPSO
SRMV
SRMW
SRSO
SRTV
SRVC
NOx Slope
Factor
Ozone Non-Ozone
season
0.79
0.20
0.83
2.56
1.37
3.26
1.10
2.17
2.08
1.88
1.61
1.86
2.06
2.10
1.73
4.06
2.21
1.68
1.17
1.40
1.80
1.61
Season
0.92
0.62
0.76
2.63
1.93
4.43
1.34
2.37
1.70
2.24
2.15
3.50
3.29
5.18
2.61
5.56
2.58
1.51
3.55
2.26
4.38
3.60
R-Squared
Ozone
season
0.54
0.21
0.88
0.95
0.77
0.90
0.83
0.88
0.89
0.82
0.93
0.96
0.87
0.95
0.55
0.88
0.94
0.71
0.50
0.80
0.77
0.96
Non-Ozone
Season
0.28
0.42
0.59
0.90
0.64
0.85
0.52
0.61
0.52
0.74
0.82
0.93
0.71
0.87
0.67
0.85
0.74
0.66
0.67
0.67
0.73
0.90
                              Indirect Emissions Analysis • 26

-------
CN
ai ^
Q.
o
s ฐ
ra *~
o:
W
C O5
.2 CD
W
ง0>0
poo2iZw2งwi->
^LLiiSQ-Q-cejSceceK
"-cccecewwc/j^www
Figure 11. Regional slope factors forCC>2 (tons/MWh)
CD

•* _
Q)
Q.
0
55 ?!
0)
ra
ซ 0
-
c
o
'in
.i2 ฐฐ
u
8 w
ro
c
o •*
in
ro
5
CM
Q

X Spring (MAM)
n Summer (JJA) |
+ Fall (SON) x
• Wnter (DJF) ^
o Annual
0 V
D i
• • _ _ ™
B • •

B X ฑ A * * n
_-x HQt x ft ซ
B^g. + ฐ
pt.
H *
^•> H~ ^^ ^5 rn 1 ^P IK
K i p$T IQ rn w
Ur fm 1 — 1 vv
i'i'l'g'g'i'si'g^'s's'g'g'i'i'si'i's1^1^1
|c1-ฃl|zl|^z^^-l8585glggg
Figure 12. Seasonal and regional slope factors for SO2 (Ibs/MWh)
                                Indirect Emissions Analysis • 27

-------
SL
ฃ
OT

Si
13
o:


o
'in



m
x
O


"ra

o
in
ro
01
V)
                             X Ozone Season
n Not Ozone Season

                                     D
                                                                  X
      D      D
                                                                                 D
            9 n  a
                          n

                          x
                           X
               X
Figure 13. Regional slope factors for NOx (Ibs/MWh)
                                Indirect Emissions Analysis • 28

-------
D. Empirical Incremental Emissions Rate (EIER)
       The previously discussed methods for estimating emissions factors rely upon the
       empirical dependence of total emissions on total output over a predefined period, which
       could be a year, a season, or any other time period. As a result, they produce emissions
       factors characteristic of the selected time period, but not indirect emissions factors at an
       hourly level. However, it is possible to take this analysis to its limit and assess the hourly
       slope factor by dividing the change in total emissions from one hour to the next by the
       change in total output for the same hour. This approach results in the empirical
       incremental emissions rate (EIER), which may be expressed mathematically as follows:

                           EIER  _C02(t)-C02(t-l}
                                '   MW(t)-MW(t-V)

       The implicit model underlying this approach is that units which are not marginal or
       displaceable in any hour will have constant output and emissions, so that the metric will
       be sensitive only to units which change  their output from hour to hour. This theoretically
       weights the metric by the change in output for these "marginal" plants. In contrast, the
       hourly average emission rate  approach  was weighted by the total output of each unit,
       even for units whose output is insensitive to changes in load.

       While the EIER has the attractive feature that it is a direct estimator of the dependence of
       change in emissions on change in total  generation for each  hour, there are certain
       considerations which  must be made in its application. First, the estimator does not
       distinguish  between units that change their output because they are in a normal ramping
       cycle, for example in anticipation of daily changes in load from those that change their
       output in response to shorter term load  fluctuations. This may be problematic as the
       former may be large units and exhibit large changes in load, so they are likely to
       dominate the metric during the hours when their output varies. Next is a numerical
       stability issue: if the change in total generation is small, the calculated EIER could be
       dominated  by reporting and round-off errors in the emissions data. Thus we choose
       (somewhat arbitrarily) to eliminate hours from consideration if the change in total
       generation  in the region from the previous hour is less than  100 MW. In this case we
       carry forward the EIER from the previous hour.

       Finally, the EIER can be less than zero  in some hours. This phenomenon occurs when
       total reported generation output in some region rises or falls during an hour by an amount
       greater than the 100 MW threshold, but total reported emissions changes in the other
       direction. This is  not necessarily reporting error; it may well be that changes in dispatch,
       due to transmission constraints or some other consideration, have enough  of an impact
       on emission rates to offset the effect of  changing total output. This instability, and the
       wide scatter in the hourly results for this approach, cast some doubt on the value of this
       approach for predicting hour-to-hour indirect emissions benefits associated with
       renewable energy resources.
                             Indirect Emissions Analysis • 29

-------
E. Load-Following  Incremental Emissions Rate (LFIR)
       The load-following incremental emissions rate is our name for a hybrid of the hourly
       average emissions rate and the empirical incremental emissions rate introduced by
       Stephen Connors in a number of studies of avoided emissions in the United States.2324
       The implicit model underlying the LFIR is that units which change their output in the same
       direction that load is changing during any dispatch period can be treated as load
       following. Thus the LFIR approach is designed to yield the average emission rate of the
       units that change  output in the direction of changing load during each hour under
       consideration. However, as Connors notes,25 units which are held  for spinning reserves
       or automatic generation control may actually be load following units, and would be
       missed by the first criterion. Thus the list of "load following" units must include any unit
       whose last change in output has been in the direction  of the change in total load, even if
       subsequent hours have produced only small  changes in output (less than 2.5% of their
       maximum.) Finally, Connors examines the range of unit behavior and finds that any unit
       whose output is between 55% and 90% of its capacity is likely to be providing spinning
       reserve service, and designates these units as load following as well.26

       Following this logic, we find that a large proportion of generating units on the system
       most of the time would be designated as "load following" in the 2005 EPA database; in
       many regions, one-third to one-half of all resources earn this designation more than 50%
       of the time. This brings the effect of this approach close to that of the HAER approach
       based on all units running, despite the significantly more complicated analysis required.
       At the same time, because the units (and hence emission rates) may change from hour-
       to-hour even when total load is changing only slightly, this  method suffers from even
       greater numerical instability than the EIER approach.

       Connors describes the logic for selecting load following units as "excessively inclusive"27
       in designating too large a proportion of available units in each hour. To remedy this, he
       weights each unit's contribution for each hour by its MW change in output from the
       previous hour. This method may serve to mitigate some of the distortion from large base
       load resources that happen to make it into the list of "load-following" units by
       coincidentally ramping along with load, but only for the hours in which they do not happen
       to be ramping. For the hours when they are, they will clearly dominate the calculated
       average, as their changes in load will be much greater than that for smaller, truly load-
       following units.

       In summary, we find that the LFIR  approach suffers from many of the same distortions as
       both the EIER approach, and the HAER approach. The LFIR is numerically unstable,
       producing emissions rates which are highly and unrealistically variable throughout the
       year, while at the same time it gives undue weight to large units which are likely to have
23 Connors et al. (2005).
24
  Berlinski, and Connors (2006).
25 Ibid., p. 1-4.
26 Connors et al., 2005, Table 1.1, p. 1 -6.
27 Ibid, p. 1-10.
                             Indirect Emissions Analysis • 30

-------
       larger changes in output from hour to hour but are not necessarily likely to be load-
       following. Although we have calculated and tabulated the estimated indirect emissions
       benefits using the LFIR for comparison, we conclude that it does not produce a useful
       metric for estimating displaced emission factors on electricity grids.

F. Flexibility-Weighted Hourly Average Emissions Rate (FW-HAER)
       The final method considered, FW-HAER is also based on the premise that only certain
       units are likely to be  displaced by variable output renewable energy resources. This
       approach attempts to identify load following units based on their operational behavior
       over the course of the entire study period, rather than during individual hours as with the
       LFIR method. The FW-HAER approach assigns to each  unit a score based on how
       readily it appears to shift output; this score is then used as a weighting coefficient in
       calculating the indirect emissions coefficient for each hour.

       While it is impossible to determine exactly which units would be load following at any
       given time given the  complexities of dispatch in real electric systems, it is reasonably
       straightforward to determine which tend to behave in a more or less flexible  manner.
       Large base load units are generally scored lower than smaller units. Units that appear to
       spend a lot of time ramping in response to load during the year, or that are often found
       partially dispatched,  are ranked as highly displaceable whenever they are running.

       To capture these dynamics, the FW-HAER approach tracks how often a plant is in the
       process of ramping relative to the number of hours it operates during the year, as shown
       in Equation 5:

                                      77 _    ramping,i
                                       ' ~ N
                                            operating,!

       Where F is the flexibility coefficient and  i refers to an individual unit. A  "ramping" hour is
       defined as any hour  in which the change in the unit's output is greater than or equal to
       2.5% of its maximum capacity; by dividing the number of ramping hours  by the total
       number of hours in which the unit operates, we obtain a unit flexibility coefficient which
       represents the proportion of its operating hours which appear to be in "ramping" mode.

       Once the flexibility coefficient of each unit on the system is  calculated, the indirect
       emissions rate for each hour of operations is calculated as shown in Equation 6:
                                                                              (6)

       Where ERit is the emission rate for unit i during hourt and the summations are overall
       units operating during that hour, and F, is as in Equation 5. Again, this method does not
       weight by either unit size or by the change in output in any given hour. Only units which
       appear to be more flexible in their output over the year have increased influence on the
       displaceable emissions rate. This approach avoids giving  undue weight to large, inflexible
                             Indirect Emissions Analysis • 31

-------
units, and also exhibits substantially less volatility than methods that are based on hourly
changes in output.

The FW-HAER approach has several advantages for estimating the emissions impacts of
variable output resources. It does a much better job than any other approach in
narrowing the list of candidate units that dominate the indirect emissions calculation.
Unlike the other approaches considered here, the FW-HAER avoids giving undue weight
to large, inflexible units. FW-HAER is much more numerically stable than either the  EIER
approach or the LFIR approach, yielding a much smaller range of displaceable emissions
rates through the year. Finally, because of this stability and more refined identification of
the most flexible units in any region, it does the best job of differentiating regions by their
true indirect emissions rates. These features are demonstrated in the figures below.

^Figure l^shows the relationships among flexibility coefficient, generator size, and
number of operating hours for (a) New England and (b) California. The size of each  circle
represents the observed number of operating hours in 2005. In both regions, we see that
the very highest flexibility coefficients are  attributed to small units which run infrequently;
when running, these may also be the first units displaced. In California,  it is apparent that
the largest units are not displaceable; in New England, there does not appear to be  much
of a relationship between unit size and flexibility.

^Figure l^shows hourly generation profiles for three sample plants in  Florida, exhibiting
typical profiles for low, medium, and high flexibility coefficients.
                       Indirect Emissions Analysis • 32

-------
(a)
NEWE
1

0.9

O.B
0.7
C
| 06
:& 0.5
'x
il 04

0,3

02
0.1
n



o
8
D
o O
-. o o

V^/ r~\ ( }
o /T^.
O o oO ~O
o ^ O r
CD ' s*$i t-C^~^ C^)
o W Q~/O ฐ
o o

. o o ^ QO OQ Q o 1 ฉ o o o 0 100 200 300 400 500 600 700 Hours in operation • <100 O 8^60 (b) CAMX 1 0.9 n A U.o 0.7 c 1 ฐ6 3? 0 5 j= 'x ~ 04 03 0,1 n Hours in operation • <100 O 8760 % „ " 1 ฐ o - ftO o O 8 ฐ ฐ cP " |ฐ ฐ$> o ' 0ฐ 3 Qฐ^ p\ ^o yi) ฐ - c9 o o (t[v^ yfj o o ฐ rt3\ ฐ O ^ o O -0ฐ o^^^ 0 r>9 (C' Oฐ ^ 0 0 0 800 900 1000 0 100 200 300 400 500 600 700 800 900 1000 Generator maximum capacity (MW) Generator maximum capacity (MW) Figure 14. Flexibility coefficient vs. generator size in New England (NEWE) and California (CAMX). Each circle represents a single generating plant in the indicated region, and the size reflects total operating hours in 2005. Maximum size is 8760 hours. Indirect Emissions Analysis • 33


-------
(a)
           200
        .1  100
        ฃ
        0)

        I    „
             Jan
           200 r
        o  100
            07/28
                                                                     ^h
                      Feb
                               Mar
                                        Apr
May
                                                           Jun
                                                                    Jul
                                     Aug
Sep
                                                                                               Oct
                                                                                                         Nov
                                                                                                                  Dec
                                                                                                                           Jan
                                  07/31
                                                        08/03
                                                                              08/06
                                                                                                    08/09
                                                                                                                           08/12
(b)
          1000
       .2   500
             Jan
          1000
       ?
       1,
       I  500
       0)
       I    „
            07/28
                      Feb
                               Mar
Apr      May
                                                           Jun
                                                                    Jul
                                                                             Aug
                                     Sep
                                                                                               Oct
                                                                                                         Nov
                                                                                                                  Dec
                                                                                                                           Jan
                                                        T       T
                                                        I   \   I
                                  07/31
                                                        08/03
                                                                              08/06
                                                                                                    08/09
                                                                                                                           08/12
(C)
           100
        ฐ   50
           100
        o   50 -
             Jan       Feb      Mar      Apr      May       Jun       Jul       Aug      Sep      Oct       Nov      Dec
                                                                                                                           Jan
- ft f I
! 1 J
/I A
,
f
ft
i
\

i


!l fi
1
\
            07/28
                                  07/31
                                                        08/03
                                                                              08/06
                                                                                                    08/09
                                                                                                                           08/12
            Figure 15.  Hourly dispatch and flexibility coefficient (F) for three plants of differing capacities in the FRCC
            subregion. Each lower graph is a summer detail of the 2005 profile shown above, (a) Polk power plant, F =
            0.08;  (b) Manatee gas plant, F = 0.55; (c) Debray peaker, F = 0.86.
                                               Indirect Emissions Analysis • 34

-------
G. Comparison of Indirect Emissions Rates
       It is instructive to compare the implications for indirect emission rates from each
       estimation approach. ^Figure 1 ^displays these results in three panels, for CO2, SO2, and
       NOx, respectively. Each panel is organized vertically by region, with indirect emission
       rates on the horizontal axis. For each region, the range (single standard deviation around
       the mean) of the hourly indirect emission rates may be read and compared for each of
       the calculation approaches considered. Methods depicted are the empirical incremental
       emission rate (EIER, solid line), the hourly average emission rate (HAER, angle
       brackets), and the flexibility-weighted average emission rate (FW-HAER, shaded
                 po
       rectangle).  Also shown is the annual (CO2, SO2) or seasonal (NOx) slope factor for
       each region, which is shown as a point as it is not an hourly value but a metric calculated
       from the annual or seasonal data.

       The graphs infigure 16^ illustrate the variations among regions, and within each region
       between the different calculation approaches. For example, the leftmost panel illustrates
       that for CO2, there is general agreement among the different calculation approaches in  all
       regions, although the slope factor tends to suggest a lower indirect emissions rate than
       other methods. The HAER and the FW-HAER approaches both yield values in a fairly
       tight range and are generally in agreement with each other, while the EIER approach
       yields a much broader range of hourly values reflecting its numerical instability. It is also
       clear from/igure 1^,that certain regions, such as the Midwest areas that rely heavily on
       coal generation,  have indirect emissions rates of over 1.0 to 1.2 tCO2 per MWh, while
       regions such as California and New England, which rely more heavily on gas, have
       indirect emissions rates of perhaps 0.6 to 0.8 tCO2 per MWh.

       Color intensity plots for each hourly method for each pollutant, and for each region are
       provided in Appendix D. These plots display the diurnal and seasonal variations in each
       displacement factor derived using each approach.  The data are also available in
       electronic format.
  We have not plotted the single-standard-deviation range based on the LFIR approach in Figures 14-16 because it is far
 too broad for this scale, reflecting the numerical instability of this approach.
                             Indirect Emissions Analysis • 35

-------











RFCE

RFCM







SRVC
[
(a)
• 1 — a *
*rr — i -
Bs .,
S~~cl •-.

]_< — >j

*S 	 J
\ 	 1* < >

* 'LT^J.



1 	 A f i
fr ^ s i
ฃ 	 1 •?
*i —  	
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Indirect emissions coefficients (t C02 / MWh)
AZNM










RFCE

RFCM







SRVC

(b)
• n f •-.

Bs




\ 	 • 	 I
*• >

f~< *


•i — ^j >
A.-- --.i

r*i < i
a,- -.


• I < > I
	 * i< >
0 2 4 6 3 10 12 14
Indirect emissions coefficients (Ibs S02 / MWh)
AZNM










RFCE

RFCM







SRVC
-1
(c)

r— i
Brf-VH



1 <*->





• t 	 Ql>
r~* — i rf **

r~s^ .-"-. -.
FTrSi

K P
2 and 862) or seasonal (for NOx; dark = ozone season) emissions slope
factors for each region.
                                                       Indirect Emissions Analysis • 36

-------
5.  Indirect Emissions Results
       In this section we present our assessment of what the annual indirect emissions
       reductions of NOx, SO2, and CO2 would have been in each of the 22 eGRID subregions
       in 2005, had an incremental renewable energy project (wind, landfill gas, or municipal
       solid waste) produced one gigawatt-hour (GWh) of energy over the course of the year.
       As discussed earlier, we find that landfill gas and municipal solid waste have identical
       base load energy profiles. Thus we have combined the analyses for these two resources
       into a single set of results.

       All of the results presented  here are based on a one-to-one displacement of fossil
       generation, assuming emissions-free generation from the renewable resource. While
       beyond the current scope, it would improve precision to incorporate line losses, as well
       as consideration of internal energy use at power plants. In general, renewable generation
       sites are remote from load centers,  meaning that more than one MW of generation is
       required to serve one MW of load. Fossil generation sites may be near to or remote from
       load. However, all fossil units have some parasitic load—again, more than one MWmust
       be produced in order to deliver one MW onto the system. The extent to which these
       cancel out in any given region, or the adjustments that should be made to accommodate
       them, is beyond the scope of this study.

       The calculated indirect emissions results are presented graphically as outlined injable 4,
       The results are also tabulated in/able 5, (wind projects) and/able 6,.(landfill gas and
       municipal solid waste.)

       Table 4. Guide to the summary figures.
Figure
figure 17.
figure 161
figure 19.
figure 20.
Figure 21.
Figure 22^
Resource Type
Wind
Wind
Wind
MSW, LFG
MSW, LFG
MSW, LFG
Pollutant
CO2
NOx
SO2
C02
NOx
S02
Methods
HAER, EIER, FW-
HAER
HAER, EIER,
FW-HAER
HAER, EIER,
FW-HAER
HAER, seasonal and
annual slope factors
HAER, Seasonal Slope
Factor
HAER, Seasonal and
Annual Slope Factors
Regions
All eGRID regions
All eGRID regions
All eGRID regions
All eGRID regions
All eGRID regions
All eGRID regions
       The results show significant regional differences in the indirect emissions of all pollutants,
       no matter what calculation method is used. They also show that the calculation method
       applied makes a large difference in some regions, while in other regions these
       differences are less significant. This leads to the question of which method  is most
       appropriate to use under what conditions.

       In our judgment, the FW-HAER best captures the short-term operational behavior of
       power plants because it reflects observations of short-term changes in output for
       individual plants which  respond flexibly to system perturbations. We recommend this
       approach in particular for shorter-term indirect emissions benefits forCO2, in response to
                             Indirect Emissions Analysis • 37

-------
variable output, non-dispatchable resources such as wind power. The regional FW-HAER
values presented in this paper can be applied for this purpose, where the definition of
short-term depends on the application but may be the first two or three years of a new
wind resource lifetime.

Over the longer term, and with greater penetration of wind resources, it is reasonable to
assume that the system will adjust by displacing  resources lower down in the dispatch
order, ultimately restoring the utilization rate of load-following capacity to accommodate
fluctuations in both load and wind resource output. In addition, NOxand SO2 are both
regulated under regional or national cap-and-trade regimes which make hourly indirect
emissions analysis less relevant for these pollutants.  Thus for longer-term indirect CO2
analysis, or for NOx and SO2, we recommend using either HAER or the emissions slope
factor. These approaches are based on the longer-term response to  perturbations over
the full resource portfolio, as opposed to reflecting primarily the marginal unit or units.
These two approaches (HAER and slope factor)  give generally similar results. The
primary difference between them is that the HAER size-weights base load fossil
resources running at any given hour, giving the most  weight to large  units regardless of
their flexibility in response to changes in output. Conversely, the slope factor method
tends to deemphasize continually operational resources regardless of size. As we find it
unlikely that the largest and lowest running cost resources would ever be displaced by
renewable energy projects, we prefer the slope factor approach for those regions and
pollutants for which the slope factor is well defined.

For landfill gas and municipal solid waste resources, the most likely type of generators to
be displaced are intermediate units, which are more costly to run than base load but
which do not have the flexibility of load-following  resources.  In this case, the best method
would again be either HAER or the slope factor approach, and we would again give slight
preference to the slope factor approach if the slope is well-defined.

Finally we note that most of our results, especially those for CO2, are relatively insensitive
to hourly or seasonal variations in wind power generation. Replacing the calculated wind
profiles with a flat power output profile (such as that characteristic of landfill gas or
municipal solid waste) results in small changes in the calculated indirect emissions
impact over the course of a year. Table 8 shows  this  result as the percent change in the
total annual indirect emissions impact for wind resources if the hourly profile is replaced
by a flat profile, typical of an LFG or MSW facility. The small changes suggest that
indirect emission factors and wind output profiles are  generally poorly correlated, and in
many cases little is gained from the effort to apply realistic profiles.

However, some of these differences are larger than others and  probably important—for
example, in some regions the impact of using a realistic wind-based shape on indirect
NOx emissions is as high as 10%. This different  probably reflects seasonal differences in
wind strength which are correlated with seasonal differences in the operation of NOx
control equipment.  Whether the effort required to establish realistic wind profiles pays
sufficient dividends in this type of analysis to be justified probably depends on the specific
application.
                      Indirect Emissions Analysis • 38

-------
Table 5. Annual indirect emissions reduction rates associated with an incremental GWh of wind energy in each of the eGRID subregions for 2005. Calculation
methods shown are: HAER: Hourly average emission rate; FW-HAER: Flexibility-weighted hourly average emission rate; EIER: Empirical incremental
emissions rate; LFIR: Load-following incremental rate; Slope factor Indirect emission rate based on seasonal relationship between emissions and output
across all hours in each season. For CO2 and SO2 the seasonal slope factors are based on conventional 3-month seasons; for NOx the slopes are defined for
the ozone and non-ozone regulation  seasons.
eGrid
Region
AZNM
CAMX
ERCT
FRCC
MROE
MROW
NEWE
NWPP
NYCW
NYLI
NYUP
RFCE
RFCM
RFCW
RMPA
SPNO
SPSO
SRMV
SRMW
SRSO
SRTV
SRVC
2005 Avoided CO2 (tons/GWh)
Slope
HAER EIER LFIR FW-HAER Factor
844 573 591 738 943
670 497 496 619 892
845 606 596 684 1,040
845 694 689 757 994
1,071 803 807 1,054 1,114
1,094 885 884 1,060 1,079
728 588 637 711 780
981 716 657 741 1,076
800 616 657 717 742
662 530 621 687 744
731 664 698 875 816
854 724 731 911 896
1,059 835 773 1,072 959
954 885 854 948 906
980 661 612 734 1,056
1,066 912 906 1,040 1,074
905 618 629 735 1,062
765 659 653 686 1,027
996 905 900 1,001 966
960 818 826 914 997
1,005 856 882 987 974
1,035 955 837 974 933
2005 Avoided NOx (Ibs/GWh)
Slope
HAER EIER LFIR FW-HAER Factor
2,768 1,335 1,260 1,641 927
1,238 259 287 411 508
1,058 863 877 785 775
2,645 2,811 2,757 1,675 2,584
2,981 1,902 2,346 4,380 1,810
4,456 3,407 3,388 4,461 4,101
1,177 1,004 944 913 1,253
3,267 2,400 2,101 2,310 2,547
1,821 2,073 957 1,206 1,846
1,623 2,277 1,160 1,317 2,158
2,016 1,991 1,864 2,172 1,967
3,062 2,691 2,485 3,039 3,009
3,194 3,277 3,057 2,988 2,912
3,579 3,852 3,496 3,435 4,219
2,963 1,675 1,011 1,805 2,377
4,060 4,019 3,203 3,955 4,966
2,642 2,662 2,610 2,151 2,466
1,638 1,912 1,900 1,407 1,616
2,664 2,415 2,489 2,717 2,839
2,769 2,300 2,504 2,919 2,089
3,755 3,270 3,252 3,708 3,499
3,349 3,177 2,685 3,405 3,029
2005 Avoided SO2 (Ibs/GWh)
Slope
HAER EIER LFIR FW-HAER Factor
1,782 1,335 1,049 801 218
1,138 259 202 139 216
3,920 863 1,300 1,074 1,116
4,857 2,811 5,254 3,519 5,454
7,678 1,902 5,147 9,310 4,738
6,887 3,407 5,075 7,850 5,451
4,177 1,004 2,691 2,737 3,126
2,546 2,400 (1,085) 1,493 1,350
1,452 2,073 1,353 1,251 1,545
4,204 2,277 2,804 3,513 2,773
6,277 1,991 6,367 9,737 6,015
13,596 2,691 9,117 12,842 9,881
9,363 3,277 6,448 7,511 7,420
12,250 3,852 11,037 13,443 12,994
1,848 1,675 708 1,699 895
6,949 4,019 6,910 8,953 6,909
3,912 2,662 1,454 1,331 1,531
2,693 1,912 2,148 1,404 1,611
6,739 2,415 6,348 7,941 7,332
11,308 2,300 9,293 10,796 7,909
9,004 3,270 9,463 12,374 7,469
11,289 3,177 9,302 12,135 9,902
                                                  Indirect Emissions Analysis • 39

-------
Table 6. Annual indirect emissions reduction rates associated with an
incremental GWh of landfill gas of MSW generation in each of the eGRID
subregions for 2005. Calculation  methods shown are: HAER: Hourly average
emission rate; Slope factor: Indirect emission rate based on seasonal
relationship between emissions and output across all hours in each  season.
For CC>2 and SC>2 the seasonal slope factors are based on conventional 3-
month seasons; for NOx the slopes are defined for the ozone and non-ozone
regulation seasons.
eGrid
Region
AZNM
CAMX
ERCT
FRCC
MROE
MROW
NEWE
NWPP
NYCW
NYLI
NYUP
RFCE
RFCM
RFCW
RMPA
SPNO
SPSO
SRMV
SRMW
SRSO
SRTV
SRVC
2005 Avoided
C02 (tons/GWh)
Slope
HAER Factor
838 943
667 893
852 1 ,042
859 995
1,086 1,117
1,098 1,080
732 777
982 1 ,078
795 740
660 740
736 812
862 898
1,058 958
956 905
980 1 ,055
1,069 1,075
909 1 ,062
764 1 ,025
999 970
961 997
1,004 975
1,032 932
2005 Avoided
NOx (I bs/ GWh)
Slope
HAER Factor
2,718 901
1,226 476
1 ,069 775
2,632 2,582
2,996 1,764
4,459 4,022
1,138 1,232
3,274 2,472
1 ,739 1 ,887
1,549 2,110
2,004 1,917
2,904 2,829
3,103 2,787
3,303 3,900
2,952 2,278
4,088 5,012
2,653 2,464
1 ,660 1 ,628
2,458 2,482
2,654 1,976
3,366 3,175
3,096 2,790
2005 Avoided
S02 (Ibs/GWh)
Slope
HAER Factor
1,710 223
1,129 206
3,981 1,080
4,806 5,442
7,834 4,649
6,891 5,307
4,140 3,119
2,553 1,362
1 ,339 1 ,482
4,103 2,645
6,306 6,003
13,720 9,424
9,380 7,164
12,187 12,688
1 ,830 974
6,949 7,005
3,958 1 ,549
2,665 1,593
6,773 7,226
11,269 7,601
8,927 7,327
11,168 9,529
Table 7. "Shape impact" for indirect emissions calculations for each pollutant
in each eGrid region. The shape impact is the change in calculated indirect
emissions if an hourly wind profile is replaced with a constant generation
profile with the same total output.
eGrid
Region
AZNM
CAMX
ERCT
FRCC
MROE
MROW
NEWE
NWPP
NYCW
NYLI
NYUP
RFCE
RFCM
RFCW
RMPA
SPNO
SPSO
SRMV
SRMW
SRSO
SRTV
SRVC
CO2 Shape
Impact
Slope
HAER Factor
-0.7% -0.1%
-0.4% 0.0%
0.9% 0.2%
1.7% 0.1%
1.5% 0.3%
0.3% 0.1%
0.5% -0.4%
0.2% 0.2%
-0.7% -0.2%
-0.3% -0.5%
0.6% -0.4%
0.9% 0.2%
-0.1% -0.1%
0.1% -0.2%
0.0% -0.1%
0.3% 0.1%
0.4% 0.1%
-0.1% -0.2%
0.3% 0.5%
0.1% 0.0%
-0.1% 0.1%
-0.3% 0.0%
NOx Shape
Impact
Slope
HAER Factor
-1.8% -2.8%
-0.9% -6.3%
1.0% -0.1%
-0.5% -0.1%
0.5% -2.5%
0.1% -1.9%
-3.3% -1.7%
0.2% -2.9%
-4.5% 2.2%
-4.6% -2.2%
-0.6% -2.5%
-5.1% -6.0%
-2.8% -4.3%
-7.7% -7.6%
-0.4% -4.2%
0.7% 0.9%
0.4% -0.1%
1.3% 0.7%
-7.7% -12.6%
-4.2% -5.4%
-10.4% -9.3%
-7.6% -7.9%
SO2 Shape
Impact
Slope
HAER Factor
-4.0% 2.2%
-0.8% -4.6%
1.5% -3.2%
-1.0% -0.2%
2.0% -1.9%
0.1% -2.6%
-0.9% -0.2%
0.3% 0.8%
-7.8% -4.1%
-2.4% -4.6%
0.5% -0.2%
0.9% -4.6%
0.2% -3.4%
-0.5% -2.4%
-1.0% 8.8%
0.0% 1.4%
1.2% 1.1%
-1.0% -1.1%
0.5% -1.5%
-0.3% -3.9%
-0.9% -1.9%
-1.1% -3.8%
                                                            Indirect Emissions Analysis • 40

-------
      1,200 -,
      1,000
        800
    O
    O
        600
        400
        200
                       HAER
                                               DEIER
                                                                       DFW-HAER
Figure 17. Annual indirect CC>2 emissions impact associated with one incremental GWh of wind energy in each of the
eGRID subregions for 2005.
       5,000 -,
4 000

O 3 QQQ
0
Q_
X
z
{/)
c 2 000
o
Q_
1 nnn
500
n



-

...
•






5


1
"i



n









~|



















ri

1-

|-|




n
•
-l 1


1




•























-1




r

_






,
_







n r











xi-OLuaiucLaziQ-^J555o?
|I^MIil^lfe^lilซMis

• HAER DEIER OFW-HAER
n







-






\ 1
Figure 18. Annual indirect NOx emissions impact associated with one incremental GWh of wind energy in each of the
eGRID subregions for 2005.
                                Indirect Emissions Analysis • 41

-------
16,000 -i
14 000
12 000
5 1 n nnn
0
0
Q_
j^ Q nnn
CO
V)
•Q
C
^ c nnn
Q_
4 000
2 000
n _







Ilk

Tl

1
1
-





-



n



[

1


1


n


r

n

-I ]






-i -


1


n


-l


•





-r

               X   H  O  LU
                                     Q-gpQ,LiJ5ggO
                     IHAER
                                              OEIER
                                                                      DFW-HAER
Figure 19. Annual indirect SC>2 emissions impact associated with one incremental GWh of wind energy in each of the
eGRID subregions for 2005.
       1,200 -,
       1,000
        800
        600
        400
        200
    o
    o
                    HAER
                                       Seasonal Slope Factor
D Annual Slope Factor
Figure 20. Annual indirect CC>2 emissions impact associated with one incremental GWh of landfill gas or municipal
solid waste energy in each of the eGRID subregions for 2005.
                                Indirect Emissions Analysis • 42

-------
            6,000
            5,000
            4,000
         O 3,000
         ฐ- 2,000
            1,000
I
                      X  H  O   LU
                                            8-g=]9;W5g<00>gO>0
                                    IHAER
                                                               • Seasonal Slope Factor
     Figure 21. Annual indirect NOx emissions impact associated with an incremental GWh of landfill gas or municipal
     solid waste energy in each of the eGRID subregions for 2005.
16,000 -i
1 4 000
1 2 000
5 10 000
0
Q_
j^ Q nnn
CO
V)
•Q
C
^ c nnn
Q_
4 nnn
2 000







-

1 L
5 X
y <
< o
1

k.
-,


•
-,







•-i


"llfti
^ ง ง o ง | ง ^
S uL i K ^ 1 ^ '
"-2 5 z z z
IHAER

1



n
-I




-I





^ >~ LL- L_ n
- z a: it ^
• Seasonal Slope Factor
b
r



D- Z
5 CL
o: co

11
O
73

|
1 n
r




-,




> g O > O
/> ^ CO "5 CO
D Annual Slope Factor
Figure 22. Annual indirect SO2 emissions impact associated with an incremental GWh of landfill gas or municipal solid
waste energy in each of the eGRID subregions for 2005.
                                     Indirect Emissions Analysis • 43

-------
6.  Conclusions
       In this analysis we have produced and applied a number of approaches for calculating
       indirect emissions impacts and provided a general discussion of which might be more
       appropriate under different circumstances. We have calculated indirect emissions
       benefits associated with CO2, NOx, and SO2 for wind, landfill gas, and municipal solid
       waste electricity generating resources, in each of the 22 eGRID subregions of the
       continental United States.

       The indirect emissions benefit calculated  for any kind of renewable energy resource
       depends on a number of questions:

       •   Where is the resource located?
       •   What is the pollutant of interest? Is there a cap and trade system in place for this
           pollutant?
       •   Is the time period of interest historical, in the near future, or several years in the
           future?
       •   Is the resource base load, dispatchable, or intermittent and  nondispatchable?
       •   If intermittent and nondispatchable, what is the expected hourly and seasonal profile
           of the resource?
       •   We have found that there are important regional differences in indirect emissions
           which can be quantified in ways that will be useful for estimating the emissions
           benefits of renewable energy projects in each region.
       •   Because of the wide range of applications reflected in the possible answers to the
           questions posed above, the best method to use in calculating indirect emissions is
           almost certainly application-specific. We believe that these methodologies provide
           useful guidance for many applications, and a better understanding of the issues for
           approximation of indirect emissions benefits than has previously been available.
       •   The results presented here could be validated by a modeling analysis, using a full
           electric system dispatch model to simulate actual indirect emissions under realistic
           unit commitment and dispatch conditions. The results could then  be compared to
           those that would be derived using each of the methods applied here. This
           comparison would provide an empirical basis for establishing which approach is most
           consistent and accurate in predicting  indirect emissions  in any particular  region or
           under specific circumstances, at least in the context of the model.
       •   In summary, we have found that:
       •   The indirect emissions benefits of renewable energy for all pollutants vary
           significantly by region, and these differences can be quantified and applied in
           calculating indirect emissions impacts;
       •   Different calculation approaches are appropriate for different types of resources and
           different applications;
                             Indirect Emissions Analysis • 44

-------
In general, hourly profiles of renewable energy resources make only a modest
difference in quantifying annual indirect emissions benefits;
More research is needed to establish which methods of calculating indirect emissions
benefits most accurately reflect real-world dispatch over a range of timescales and
system conditions.
Finally, we note again that the geographic coverage of this analysis was by necessity
coarse, in the interest of presenting results for the entire continental United States,
particularly with respect to wind resources. For the analysis to be applied to calculate
or predict indirect emissions impacts for specific projects, the tables of hourly
avoidable emissions data in Appendix D could be applied directly to site-specific
operational and meteorological data. However, given the modest impact of
differences in hourly profiles noted above, the annual results presented here are
likely to be sufficient for most  purposes.
                   Indirect Emissions Analysis • 45

-------
      Appendix A:
Identification and map of eGRID subregions
      Indirect Emissions Analysis • 46
           Appendix A

-------
Figure A-1. Map of eGRID subregions. Subregions in Alaska and Hawaii are not considered in this analysis.
Acronyms are defined below. Source of figure: https://www.energvstar.gov/istar/pmpam/help/eGRID_Subregion_Map.htm.
              AZNM   Arizona and New Mexico
              CAMX   California
              ERCT   ERGOT (Texas)
              FRCC   FRCC (Florida)
              MROE   Midwest Reliability - East
              MROW  Midwest Reliability - West
              NEWE   New England
              NWPP   Northwest
              NYCW  New York City/ Westchester
              NYLI    New York - Long Island
              NYUP   New York - Upstate
RFCE   Reliability First - East
RFCM   Reliability First - Michigan
RFCW  Reliability First - West
RMPA   Rocky Mountain region
SPNO   Southwest Power Pool - North
SPSO   Southwest Power Pool - South
SRMV   SERC - Mississippi Valley
SRMW  SERC-Midwest
SRSO   SERC - South
SRTV   SERC - Tennessee Valley
SRVC   SERC - Virginia/Carolina
                                      Indirect Emissions Analysis'
                                             Appendix A
     47

-------
                     Appendix B:
Color intensity plots of synthetic wind power time series (hourly percent of
capacity) for each eGRID subregion. Each grid shows hour of the day on
the vertical axis and days of the year on the horizontal axis. Color scale
for all charts is shown below and at the bottom of the last page.

These data were obtained by scaling publicly available ground-level
meteorological data to a typical turbine height and  class 4 wind strength
(Equation 1 in the text) and then transforming into wind power output
using Equation 2.

The hourly data are available in electronic format upon request.
                         Percent Capacity (%)
      10
20
30
40
50
60
70
80
90
100
                     Indirect Emissions Analysis • 48
                           Appendix B

-------
                                      AZNM
 Jan    Feb   Mar   Apr   May   Jun    Jul    Aug   Sep    Oct    Nov   Dec
                                      CAMX
 Jan    Feb   Mar
Apr   May    Jun    Jul
                  ERCT
                                             Aug   Sep
Nov   Dec
 Jan    Feb   Mar
Apr   May    Jun    Jul
                  FRCC
                                             Aug   Sep   Oct    Nov   Dec
        Feb   Mar   Apr   May   Jun    Jul    Aug   Sep    Oct    Nov   Dec
                                      MROE
 5
10
15
20
 Jan    Feb   Mar   Apr   May
             Jun    Jul    Aug    Sep
                 MROW
                                                         Oct    Nov   Dec
 Jan    Feb   Mar   Apr   May
             Jun    Jul    Aug    Sep
                  NEWE
                                                                Nov   Dec
18
Jan    Feb    Mar    Apr   May
                                 Jun    Jul    Aug   Sep
                                      NWPP
                                      Oct    Nov   Dec
 Jan    Feb   Mar   Apr   May   Jun    Jul    Aug   Sep    Oct    Nov   Dec
                         Indirect Emissions Analysis • 49
                                 Appendix B

-------
                                     NYCW
       Feb    Mar    Apr   May    Jun    Jul    Aug   Sep    Oct    Nov    Dec
Jan    Feb    Mar   Apr   May    Jun    Jul    Aug   Sep    Oct    Nov    Dec
                                     NYUP
       Feb    Mar    Apr   May    Jun    Jul    Aug   Sep    Oct    Nov    Dec
       Feb    Mar
Apr   May    Jun    Jul    Aug   Sep
                  RFCM
Oct   Nov    Dec
       Feb    Mar
Apr   May    Jun    Jul    Aug   Sep
                  RFCW
Oct   Nov    Dec
Jan    Feb    Mar    Apr   May
             Jun    Jul    Aug   Sep
                  RMPA
      Nov    Dec
Jan    Feb    Mar    Apr   May
             Jun    Jul    Aug   Sep
                  SPNO
Oct   Nov    Dec
Jan    Feb    Mar    Apr   May    Jun    Jul    Aug   Sep    Oct    Nov    Dec
                        Indirect Emissions Analysis • 50
                                Appendix B

-------
                                       SPSO
 Jan    Feb   Mar
Apr   May    Jun    Jul    Aug    Sep    Oct    Nov   Dec
                  SRMV
 Jan    Feb   Mar    Apr   May
             Jun    Jul    Aug    Sep
                  SRMW
       Nov   Dec
 5
10
15
20
 Jan    Feb   Mar    Apr   May
             Jun    Jul    Aug    Sep
                  SRSO
Oct    Nov   Dec
        Feb   Mar
Apr   May    Jun    Jul    Aug    Sep    Oct    Nov   Dec
                  SRTV
        Feb   Mar
Apr   May    Jun    Jul    Aug    Sep    Oct    Nov   Dec
                  SRVC
 Jan    Feb   Mar
Apr   May    Jun    Jul    Aug    Sep    Oct    Nov   Dec


            Percent Capacity (%)
         10      20      30      40      50      60      70      80
                                                   90     100
                         Indirect Emissions Analysis • 51
                                 Appendix B

-------
               Appendix C:
Summary tables of raw and scaled wind speed and synthetic wind power
time series. Monthly summaries are shown for each eGRID subregion.

The hourly data are available in electronic format upon request

The proxy wind turbine locations are shown below, overlaid on map of all
U.S. WBAN stations.
         • ...T . ^/^^"^Htef^a^
              % .fit.VTJe:    ซ3I*?**
         .   •   •ป.%... I*."**.**!*;* • • V_
           •{•  ••!••• ••<••ซ •••• •!ป• *•• *t*ปH*
         ^5Jฎ^^Sง5
                    o -j^t^ •••• ••fOwt *
                     •ซ•!•!••••••• ซNซ.%*
• iป  • ••
• ••••ซ./•
   *•
                       i
                                    •9
Index
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
NERC
Subregion
AZNM
CAMX
ERCT
FRCC
MROE
MROW
NEWE
NWPP
NYCW
NYLI
NYUP
RFCE
RFCM
RFCW
RMPA
SPNO
SPSO
SRMV
SRMW
SRSO
SRTV
SRVC
WBAN
Number
23055
23187
23042
12844
14898
24012
14739
24150
94789
94789
4725
93730
94860
14848
24018
23065
23047
12916
93822
13874
13893
93729
Location
GUADALUPE PASS, TX
SANDBERG, CA
LUBBOCK, TX
WEST PALM BEACH, FL
GREEN BAY, Wl
DICKINSON, ND
BOSTON, MA
LIVINGSTON, MT
NEW YORK, NY
NEW YORK, NY
BINGHAMTON, NY
ATLANTIC CITY, NJ
GRAND RAPIDS, Ml
SOUTH BEND, IN
CHEYENNE, WY
GOODLAND, KS
AMARILLO, TX
NEW ORLEANS, LA
SPRINGFIELD, IL
ATLANTA, GA
MEMPHIS, TN
CAPE HATTERAS, NC
Latitude
31.5
34.44
33.4
26.41
44.31
46.48
42.22
45.42
40.4
40.4
42.13
39.28
42.53
41.43
41.1
39.22
35.13
29.59
39.51
33.38
35.04
35.14
Longitude
-104.49
-118.43
-101.49
-80.06
-88.07
-102.48
-71.01
-110.27
-73.48
-73.48
-75.59
-74.28
-85.31
-86.2
-104.49
-101.41
-101.43
-90.15
-89.41
-84.26
-89.59
-75.37
               Indirect Emissions Analysis'
                    Appendix B
                  52

-------
from
 Monthly average interpolated wind speed (mph)
representative WBAN station in each NERC subregion

AZNM
CAMX
ERCT
FRCC
MROE
MROW
NEWE
NWPP
NYCW
NYLI
NYUP
RFCE
RFCM
RFCW
RMPA
SPNO
SPSO
SRMV
SRMW
SRSO
SRTV
SRVC
Jan
19.63
14.08
10.61
10.49
9.23
11.78
12.32
15.95
12.33
12.33
8.11
8.15
9.49
9.23
11.03
10.20
11.87
8.94
9.86
8.92
9.75
10.60
Feb
21.52
13.99
10.36
8.74
8.34
10.66
10.86
14.12
11.71
11.71
7.23
8.77
8.23
8.40
11.36
9.73
12.04
8.79
9.37
8.72
8.60
9.51
Mar
22.01
13.37
13.24
8.99
9.12
13.27
12.21
15.85
13.01
13.01
8.37
10.40
8.90
9.26
13.99
13.26
13.01
8.39
11.32
10.08
10.39
11.27
Apr
20.21
14.91
14.15
10.33
9.59
14.23
11.29
12.57
12.40
12.40
8.35
10.72
10.23
10.53
13.03
14.24
14.29
9.93
10.93
8.96
9.35
12.02
May
16.42
14.43
12.49
8.44
9.27
12.74
11.53
12.15
10.97
10.97
8.00
8.30
8.68
8.58
11.14
12.76
11.71
7.47
9.26
7.71
7.38
9.59
Jun
15.69
14.16
13.87
8.85
8.04
10.95
9.75
11.63
10.28
10.28
6.72
8.17
7.90
8.08
11.01
12.91
14.10
6.12
7.52
7.66
7.03
8.65
July
15.94
12.23
10.74
8.97
6.98
10.83
10.45
9.35
9.42
9.42
7.00
6.85
7.36
7.05
9.86
12.76
12.50
7.08
6.34
5.82
7.10
7.44
Aug
13.56
11.38
9.23
7.67
6.44
10.56
9.18
10.30
9.75
9.75
7.22
6.26
7.32
6.11
9.80
9.74
10.55
5.45
5.74
7.03
6.64
7.25
Sept
14.98
11.39
10.37
9.07
6.41
10.52
9.69
12.52
10.31
10.31
7.36
6.94
8.32
7.19
9.70
12.28
13.01
8.33
6.35
7.61
7.04
9.04
Oct
16.82
13.43
9.69
10.69
7.36
11.14
12.67
12.35
12.10
12.10
7.58
9.58
7.98
6.94
10.07
11.23
11.86
8.57
7.60
8.48
6.81
10.18
Nov
19.53
15.19
11.24
10.40
10.58
13.14
11.02
18.28
13.11
13.11
10.18
9.36
12.82
12.31
14.79
12.64
13.02
8.30
12.36
8.80
9.43
9.69
Dec
19.26
13.98
10.86
8.48
7.67
13.25
11.49
19.86
12.68
12.68
8.87
8.85
9.93
9.88
13.47
11.57
10.90
9.02
9.58
9.33
8.46
9.52
                Indirect Emissions Analysis • 53
                      Appendix C

-------

AZNM
CAMX
ERCT
FRCC
MROE
MROW
NEWE
NWPP
NYCW
NYLI
NYUP
RFCE
RFCM
RFCW
RMPA
SPNO
SPSO
SRMV
SRMW
SRSO
SRTV
SRVC
Monthly average interpolated repr
scaled to Class 4 (16.25 mph average
Jan
38.64
28.04
21.49
13.75
8.36
16.50
14.79
18.86
18.14
23.51
14.24
13.85
18.69
16.96
20.79
13.88
21.61
12.18
20.22
12.60
13.77
15.10
Feb
42.36
27.86
20.98
11.45
7.55
14.93
13.03
16.70
17.23
22.32
12.68
14.91
16.23
15.44
21.40
13.23
21.92
11.97
19.22
12.31
12.14
13.55
Mar
43.33
26.63
26.82
11.79
8.26
18.60
14.66
18.74
19.14
24.81
14.69
17.68
17.54
17.02
26.36
18.03
23.68
11.43
23.23
14.24
14.68
16.06
Apr
39.78
29.68
28.65
13.55
8.68
19.94
13.55
14.87
18.24
23.64
14.65
18.22
20.16
19.35
24.56
19.36
26.02
13.53
22.42
12.65
13.21
17.13
May
32.32
28.74
25.29
11.06
8.40
17.85
13.83
14.37
16.13
20.91
14.04
14.10
17.11
15.76
20.99
17.36
21.31
10.18
19.00
10.89
10.42
13.67
Jun
30.89
28.19
28.09
11.61
7.28
15.34
11.70
13.75
15.12
19.60
11.80
13.88
15.58
14.85
20.76
17.55
25.66
8.34
15.42
10.83
9.93
12.33
esentative wind speed
for each NERC subregion
July
31.38
24.35
21.76
11.76
6.33
15.18
12.54
11.06
13.86
17.95
12.28
11.64
14.51
12.96
18.58
17.36
22.76
9.64
13.01
8.21
10.03
10.61
Aug
26.70
22.67
18.69
10.06
5.83
14.80
11.02
12.18
14.35
18.59
12.67
10.63
14.43
11.23
18.47
13.24
19.20
7.43
11.78
9.92
9.38
10.33
Sept
29.48
22.68
21.00
11.90
5.80
14.74
11.63
14.81
15.16
19.64
12.91
11.79
16.40
13.21
18.28
16.70
23.67
11.34
13.03
10.75
9.95
12.89
Oct
33.12
26.74
19.62
14.02
6.66
15.61
15.20
14.60
17.80
23.07
13.30
16.28
15.73
12.75
18.97
15.26
21.59
11.67
15.59
11.98
9.62
14.51
Nov
38.44
30.24
22.76
13.64
9.58
18.40
13.23
21.61
19.29
24.99
17.87
15.91
25.27
22.63
27.87
17.19
23.70
11.30
25.36
12.43
13.32
13.81
Dec
37.92
27.84
21.99
11.12
6.95
18.57
13.80
23.48
18.66
24.18
15.56
15.05
19.57
18.16
25.39
15.73
19.83
12.29
19.66
13.18
11.94
13.57
Indirect Emissions Analysis • 54
         Appendix C

-------
Monthly average calculated wind power output from 1
             in each NERC subregion (dropout at 56
.5 MW proxy turbine
mph)

AZNM
CAMX
ERCT
FRCC
MROE
MROW
NEWE
NWPP
NYCW
NYLI
NYUP
RFCE
RFCM
RFCW
RMPA
SPNO
SPSO
SRMV
SRMW
SRSO
SRTV
SRVC
Jan
883.49
971.55
680.73
286.16
58.62
439.92
338.89
572.73
546.35
822.24
335.36
356.54
638.43
506.60
644.99
290.58
752.59
202.22
677.87
253.66
258.66
353.12
Feb
919.73
941.26
670.23
166.17
23.10
343.82
213.95
479.22
504.41
767.69
324.15
394.50
477.22
413.12
659.34
217.16
766.10
177.89
624.12
227.19
200.36
226.16
Mar
910.01
865.85
929.82
190.93
51.41
573.12
272.13
574.52
589.69
848.34
368.73
540.16
605.68
549.15
895.00
515.08
815.82
160.16
797.53
310.51
335.47
407.72
Apr
823.61
1048.52
902.22
268.04
64.85
655.90
241.16
402.75
577.93
849.13
377.29
584.68
721.99
606.09
835.90
587.51
878.05
283.73
796.78
235.63
242.64
500.17
May
879.14
1101.11
899.96
137.00
47.93
518.31
286.37
358.02
406.64
677.99
294.03
341.90
549.50
457.05
664.56
515.45
735.24
114.13
637.27
127.92
97.73
228.59
Jun
1056.20
1005.35
1121.26
198.17
28.74
386.22
119.42
323.25
330.33
606.39
158.21
271.12
423.01
350.10
643.43
497.59
1004.26
67.34
390.96
139.60
87.17
172.78
July
1102.16
924.58
769.51
193.28
14.36
329.69
156.53
185.49
276.92
531.39
171.78
169.22
363.23
281.30
566.71
490.45
859.04
106.24
274.41
91.58
94.79
90.20
Aug
988.58
835.30
577.06
149.79
11.58
343.77
107.23
229.19
288.96
556.41
197.17
164.24
354.43
208.01
526.45
233.00
625.42
62.86
204.13
92.51
87.37
88.40
Sept
990.65
803.95
692.23
199.06
17.25
336.63
138.02
368.41
336.61
649.84
225.59
237.97
452.75
304.92
522.83
451.34
912.29
207.91
305.80
97.90
108.31
237.99
Oct
901.21
906.92
604.10
267.63
16.27
391.38
363.12
354.81
572.02
812.71
261.58
498.45
446.94
264.16
602.56
366.46
697.96
207.15
480.58
210.36
100.74
312.39
Nov
849.34
1037.77
675.85
256.74
113.47
559.79
251.55
786.22
633.20
851.24
491.64
501.62
890.43
780.63
813.64
414.62
848.84
209.46
898.76
218.47
289.18
281.43
Dec
987.33
971.48
672.41
164.98
21.29
532.01
254.57
880.60
574.40
858.86
368.78
415.26
673.93
548.80
838.43
362.94
601.09
216.39
648.20
209.75
168.19
262.02
                       Indirect Emissions Analysis • 55
                             Appendix C

-------
                      Appendix  D:
Color intensity maps of hourly indirect C02, NOx and S02 emissions rates
for each eGRID subregion, using several alternative calculation
methodologies.
Each grid shows hour of the day on the vertical axis and days of the year
on the horizontal axis. Color scale is shown at the bottom of each page.
The  indirect emissions metrics shown are:
 Figure                      Title                          Page
   D-1     Hourly Average CO2 Emissions Rate                       56
   D-2    Empirical Incremental CO2 Emissions Rate                  59
   n .    Flexibility-Weighted Hourly Average CO2 Emissions            R0
   D'3    Rate                                                b2
   D-4    Load Following Incremental CO2 Emissions Rate              65
   D-5    Hourly Average NOx Emissions Rate                       68
   D-6    Empirical Incremental NOx Emissions Rate                  71
   n _    Flexibility-Weighted Hourly Average NOx Emissions            7.
   u~'     Rate                                                /4
   D-8    Load Following Incremental NOx Emissions Rate              77
   D-9    Hourly Average SO2 Emissions Rate                       80
  D-10    Empirical Incremental SO2 Emissions Rate                  83
  _.  ..    Flexibility-Weighted Hourly Average SO2 Emissions            _„
  D'11    Rate                                                86
  D-12    Load Following Incremental SO2 Emissions Rate              89
The data are available in electronic format upon request.
                      Indirect Emissions Analysis • 56
                             Appendix D

-------
        D-l: Hourly Average COi Emissions Rate(tons per MWh)


                                 AZNM
 5
10
15
20
•>,

                                 CAMX
                                 ERCT
 5
10
15
20
   'I HiMf ซ•>'*
-------
          D-l (continued): Hourly Average COi Emissions Rate


                                   NYCW
 5
10
15
20
 5
10
15
20
'sr-.-i  -.in
                                   NYLI
                                   NYUP
                                   RFCE
 5
10
15
20

                                   RFCM
                                   RFCW
 5
10
15
20
                    100
                                   RMPA
5
10
15
20
I I I V I I
: :
r ' -
' ' I 1 1 1 1 1
                                   SPNO
                             150
                                       200
                                                 250
                              Emissions Rate (tons/MWh)
                                     0.9
                                                          300
                                                                    350
                     Indirect Emissions Analysis • 58
                            Appendix D

-------
     D-l (continued): Hourly Average COi Emissions Rate (continued)
                                         SPSO
 5
10
15
20


                                         SRMV
10
15
20
                                         SRMW
 5
10
15
20
                                         SRSO
5
10
15
20
i i i i i i
-
-
i i i i i i
 5
10
15
20
r
                                         SRTV

                                         SRVC
            50
                       100
                                   150
                                              200
                                                         250
                                                                    300
                                                                               350
                     0.7
                                   Emissions Rate (tons/MWh)
                                0.8
                                            0.9
                         Indirect Emissions Analysis • 59
                                 Appendix D

-------
      D-2: Empirical Incremental COi Emissions Rate (tons per MWh)
                                         AZNM
                                         CAMX
 5
10
15
20
                                         ERCT
                                         FRCC
                                         MROE
                                         MROW
                                         NEWE
                                         NWPP
                       100
                                   150
                                              200
                                                         250
                                                                    300
                                                                                350
                                   Emissions Rate (tons/MWh)
                         Indirect Emissions Analysis • 60
                                 Appendix D

-------
       D-2 (continued): Empirical Incremental COi Emissions Rate
                                  NYCW
                                  NYLI
 5
10
15
20
 5
10
15
20
                                  NYUP
                                  RFCE
 Pijฃll$^
 4 jJrlป*s|iJjB&AJf>ซ'a"*\j!i{'>ฃ Ovf " ' !'?rirT-i =1 ?Cr •  ••' -. 'r''~J- ""''<••!'* IJ^v^-j"iJ.l'S^}i 1,'-If•!• T^l
 fe isfcSH^SS^s3rafi^'-^^ jf^^^^M^IMEi^
                                  RFCM
          50
                                  RFCW
                                  RMPA
                                  SPNO
                                                            '

                   100
                             150

                                      200
                                                      -. •• -
                                               250
                                                        300
                                                                  350
                             Emissions Rate (tons/MWh)
                                    0.9
                     Indirect Emissions Analysis • 61
                           Appendix D

-------
        D-2 (continued): Empirical Incremental COi Emissions Rate


                                       AZNM
                                       CAMX
                                       ERCT
 5
10
15
20
                                       FRCC
                                       MROE
                                      MROW
 aa&gi&Xi&ti^^^
                                              - .*. Vt •  •. I • . If1  ..  J- J •. .  *f
                                       NEWE
                                       NWPP
                      100
                                 150
                                           200
                                                      250
                                                                300
                                                                           350
                                 Emissions Rate (tons/MWh)
                                         0.9
                       Indirect Emissions Analysis • 62
                               Appendix D

-------
         D-2 (continued): Empirical Incremental COi Emissions Rate


                                         SPSO
 5
10
15
20
 5
10
15
20
                                         SRMV
                                         SRMW
                                         SRSO
 5
10
15
20
 5
10
15
20
                                          SRTV
                                          SRVC

             50
                        100
                                   150
                                              200
                                                          250
                                   Emissions Rate (tons/MWh)
                                            0.9
                                                                     300
                                                                                350
                                                                   1.1
                                                                              1.2
                         Indirect Emissions Analysis • 63
                                 Appendix D

-------
       D-3: Flexibility-Weighted Hourly Average COi Emissions Rate
                               (tons per MWh)
 5
10
15
20
               III
                                         AZNM
 5
10
15
20
                                         CAMX
 5
10
15
20
                                         ERCT
                                                          .
 5
10
15
20
~-  T
                                         FRCC
                                         MROE
 5
10
15
20

                                         MROW
 5
10
15
20
                                         NEWE

                       100
                                         NWPP
                                   150
                                              200
                                                         250
                                                                    300
                                                                                350
                                   Emissions Rate (tons/MWh)
                                            0.9
                         Indirect Emissions Analysis • 64
                                 Appendix D

-------
 D-3 (continued): Flexibility-Weighted Hourly Average COi Emissions Rate
                                        NYCW
 5
10
15
20
              f
     ff *\ a $
 5
10
15
20
                                        NYLI
                                        NYUP
                                        RFCE
 5
10
15
20
,
                                                              -  -    -
rr1  f  "•*
                                I
 5
10
15
20
                                        RFCM
           I  hi <  ^'iHMr
                                        RFCW
5
10
15
20
i i i. _ ' i i
l -
-
ii [ [ i i [
 5
10
15
20
                                        RMPA
                                        SPNO
                       100
                                  150
                                            200
                                                       250
                                                                  300
                                                                             350
                                  Emissions Rate (tons/MWh)
                                          0.9
                        Indirect Emissions Analysis • 65
                                Appendix D

-------
 D-3 (continued): Flexibility Weighted Hourly Average COi Emissions Rate
                                         SPSO
                                         SRMV
                                         SRMW
                                         SRSO
10
15
20
 5
10
15
20
                                         SRTV
                                         SRVC
 5
10
15
20
',
             50
                       100
                                   150
                                              200
                                                         250
                                                                    300
                                                                                350
                                   Emissions Rate (tons/MWh)
                                            0.9
                         Indirect Emissions Analysis • 66
                                 Appendix D

-------
   D-4: Load-Following Incremental COi Emissions Rate (tons per MWh)
                                        AZNM
 5
10
15
20
                                        CAMX
                                         ERCT
 5
10
15
20
                                        FRCC
                                        MROE
                                        MROW
 5
10
15
20
                                        NEWE
                                        NWPP
            50
                       100
                                  150
                                             200
                                                        250
                                                                   300
                                                                              350
                                  Emissions Rate (tons/MWh)
                                           0.9
                         Indirect Emissions Analysis • 67
                                Appendix D

-------
     D-4 (continued): Load-Following Incremental COi Emissions Rate
                                        NYCW
                                         NYLI
                                        NYUP
                                        RFCE
 5
10
15
20
                                        RFCM

                                        RFCW
 5
10
15
20

                                        RMPA
            50
                                        SPNO
                       100
                                  150
                                             200
                                                        250
                                                                   300
                                                                              350
                                  Emissions Rate (tons/MWh)
                                           0.9
                        Indirect Emissions Analysis • 68
                                Appendix D

-------
     D-4 (continued): Load-Following Incremental COi Emissions Rate
                                     SPSO
                                     SRMV

                                     SRMW
 5
10
15
20

                                     SRSO
  >v-T.,:;iv. •', tfjl i-:.-.'•'• -t-?~ . -•'•"I;
  M^mjMfit
 5
10
15
20
                                                                         4?ฃ
                                     SRTV
                                     SRVC
 5
10
15
20
  iffซM5l
  ^> ?i-f&••••> --r^s. T.V-II*--
                         l •'"^•'^•'Iv.'--1.- - •  -, '•?•
           50
                     100
                               150
                                         200
                                                   250
                                                             ..
                                                              300
                                                                        350
                                Emissions Rate (tons/MWh)
                                       0.9
                                                           1.1
                                                                      1.2
                      Indirect Emissions Analysis • 69
                              Appendix D

-------
       D-5: Hourly Average NOx Emissions Rate (pounds per MWh)


                                         AZNM
 5
10
15
20
                                         CAMX
                           I     j
                                         ERCT
 5
10
15
20
                                         FRCC
 5
10
15
20
                                         MROE
                                         MROW
                                         NEWE
10
15
20
\-
                                         NWPP


 5
10
15
20
             50
                       100
                                   150
                                              200
                                                         250
                                                                    300
                                                                                350
                                   Emissions Rate (Ibs/MWh)
           0     0.5     1     1.5
                                            2.5
                                                         3.5
                         Indirect Emissions Analysis • 70
                                 Appendix D

-------
            D-5 (continued): Hourly Average NOx Emissions Rate


                                         NYCW
 5
10
15
20
'II
                                          NYLI
ioF
15h
20 \-
                                          NYUP
 5
10
15
20
                                                           _
 5
10
15
20
 5
10
15
20
 5
10
15
20
 5
10
15
20
                                          RFCE
5
10
15
20
i i i i
:
I I I I I I
I
—
I
                                          RFCM
                                         RFCW
                                          RMPA

             50
                        100
                                         SPNO

                                   150
                                              200
                                                          250
                                                                     300
                                    Emissions Rate (Ibs/MWh)
                                                                                 350
           0     0.5      1      1.5
                                            2.5
                                                          3.5
                         Indirect Emissions Analysis • 71
                                 Appendix D

-------
           D-5 (continued): Hourly Average NOx Emissions Rate
                                         SPSO
 5
10
15
20
                                         SRMV
 5
10
15
20

20
                                         SRMW
fill
I 1 1 I
;
i i * i
                                         SRSO
 5
10
15
20
                                         SRTV
                                         SRVC
                                   150
                                              200
                                                         250
                                                                     300
                                                                                350
                 0.5
                              1.5
                                   Emissions Rate (Ibs/MWh)
                                            2.5
                                                         3.5
                         Indirect Emissions Analysis • 72
                                 Appendix D

-------
    D-6: Empirical Incremental NOx Emissions Rate (pounds per MWh)


                                         AZNM
 5
10
15
20
 5
10
15
20
 5
10
15
20
                                         CAMX
                                         ERCT
                                         FRCC
                                         MROE
                                         MROW


                                         NEWE
                                         NWPP
                       100
                                   150
                                              200
                                                         250
                                                                    300
                                                                                350
                                   Emissions Rate (Ibs/MWh)
           0     0.5     1     1.5
                                            2.5
                                                         3.5
                         Indirect Emissions Analysis • 73
                                 Appendix D

-------
        D-6 (continued): Empirical Incremental NOx Emissions Rate


                                         NYCW
                                          NYLI
                                          NYUP
                                          RFCE
                                                             ^
 5
10
15
20
 5
10
15
20
                                         RFCM
                                         RFCW
                                         RMPA
                                         SPNO
             50
                        100
                                   150
                                              200
                                                          250
                                                                     300
                                                                                350
                                    Emissions Rate (Ibs/MWh)
           0     0.5      1
                               1.5
                                            2.5
                                                          3.5
                         Indirect Emissions Analysis • 74
                                 Appendix D

-------
D-6 (continued): Empirical Incremental NOx Emissions Rate
                                 SPSO
                                 SRMV
                                 SRMW
                                 SRSO
                                 SRTV
                100
                                 SRVC
                           150
                                      200
                                                  250
                                                             300
                                                                        350
                            Emissions Rate (Ibs/MWh)
   0     0.5     1     1.5
                                    2.5
                                                 3.5
                                                               4.5
                 Indirect Emissions Analysis • 75
                         Appendix D

-------
       D-7: Flexibility-Weighted Hourly Average NOx Emissions Rate

                             (pounds per MWh)
 5
10
15
20
 5
10
15
20
                                        AZNM
                                        CAMX
                                        ERCT
                                        FRCC
                                        MROE
 5
10
15
20
                                       MROW
 5
10
15
20
                                        NEWE
                                             IT I
                                        NWPP
 5
10
15
20
              ''
                	

                                                                     *< <' ซ.iH
            50
                       100
                                  150
                                            200
                                                       250
                                                                  300
                                                                             350
                0.5      1     1.5
                                  Emissions Rate (Ibs/MWh)
                                          2.5
                                                       3.5
                        Indirect Emissions Analysis • 76
                                Appendix D

-------
 D-7 (continued): Flexibility-Weighted Hourly Average NOx Emissions Rate


                                  NYCW
 5
10
15
20
 5
10
15
20
 5
10
15
20


                                  NYLI

                                  NYUP


                     n

                                  RFCE
                                  RFCM
 5
10
15
20
                                  RFCW
 5
10
15
20
       |.
                                  RMPA
10
15
20
 5
10
15
20
.ill
-.-j     ..   -^
                                  SPNO
          50
                   100
                            150
                                      200
                                               250
                                                        300
                                                                 350
                             Emissions Rate (Ibs/MWh)
         0    0.5    1     1.5
                                    2.5
                                               3.5
                    Indirect Emissions Analysis • 77
                           Appendix D

-------
 D-7 (continued): Flexibility-Weighted Hourly Average NOx Emissions Rate
                                         SPSO
 5
10
15
20
  E
                                         SRMV
10
15
20
                                        SRMW
                                         SRSO
 5
10
15
20
  'Iff.
       W;
                                         SRTV

 5
10
15
20
            50
                       100
                                         SRVC
                                  150
                                             200
                                                         250
                                                                    300
                                   Emissions Rate (Ibs/MWh)
           0     0.5     1      1.5
                                           2.5
                                                        3.5
                                                                      4.5
                         Indirect Emissions Analysis • 78
                                Appendix D

-------
 D-8: Load-Following Incremental NOx Emissions Rate (pounds per MWh)
                                        AZNM
                                        CAMX
                                        ERCT
 5
10
15
20
                                        FRCC
 5
10
15
20
                                        MROE
                                        MROW
                                        NEWE
                                        NWPP
            50
                       100
                                  150
                                             200
                                                        250
                                                                   300
                                                                              350
                                   Emissions Rate (Ibs/MWh)
           0     0.5     1      1.5     2     2.5     3     3.5     4     4.5     5
                        Indirect Emissions Analysis • 79
                                Appendix D

-------
D-8 (continued): Load-Following Incremental NOx Emissions Rate

                                   NYCW
                                    NYLI
                                   NYUP

                                   RFCE

                                   RFCM
                                   RFCW
                                   RMPA
                                   SPNO
                  100
                             150
                                        200
                                                   250
                                                              300
                                                                         350
                              Emissions Rate (Ibs/MWh)
      0     0.5     1      1.5
                                      2.5
                                                   3.5
                    Indirect Emissions Analysis • 80
                           Appendix D

-------
     D-8 (continued): Load-Following Incremental NOx Emissions Rate
                                        SPSO
                                        SRMV
                                        SRMW
                                        SRSO

 5
10
15
20
                                        SRTV
                                        SRVC
                       100
                                  150
                                             200
                                                        250
                                                                   300
                                   Emissions Rate (Ibs/MWh)
           0     0.5     1
                                           2.5
                                                        3.5
                                                                     4.5
                                                                              350
                        Indirect Emissions Analysis • 81
                                Appendix D

-------
        D-9: Hourly Average SOi Emissions Rate (pounds per MWh)
                                        AZNM
                                        CAMX
 5
10
15
20
                                        ERCT
10
15
20
r
                                        FRCC
                                        MROE
5
10
15
20
I I I I I H <
-
-
I I I I I Mi .il
I
—
_
[
                                        MROW
 5
10
15
20
 5
10
15
20
                                        NEWE

                       100
                                        NWPP
                                  150
                                             200
                                                        250
                                                                   300
                                                                              350
                                   Emissions Rate (Ibs/MWh)
                        Indirect Emissions Analysis • 82
                                Appendix D

-------
JgW*"
I O
20
           D-9 (continued): Hourly Average SOi Emissions Rate


                                     NYCW
                                     NYLI
                                     NYUP
 5
10
15
20
                               _,
                                                                _
 5
10
15
20
 5
10
15
20
                                     RFCE
II  •   Mlljf Hi  ii
                                     RFCM
                                     RFCW
                                                                           !
                                     RMPA
                                     SPNO
 5
10
15
20
           50
                     100
                               150
                                         200
                                                   250
                                                              300
                                Emissions Rate (Ibs/MWh)
                                                                        350
                      Indirect Emissions Analysis • 83
                              Appendix D

-------
            D-9 (continued): Hourly Average SOi Emissions Rate


                                         SPSO
 5
10
15
20
 5
10
15
20
 5
10
15
20

[  .I
                                         SRMV
                                        SRMW
5
10
15
20
i i i i i i
-
-
i i i i i i
                                         SRSO
                                         SRTV
                                         SRVC
                                          ^^^^^
            50
                       100
                                  150
                                             200
                                                        250
                                   Emissions Rate (Ibs/MWh)
                                                      10
                                                                    300
                                                                               350
                         Indirect Emissions Analysis • 84
                                Appendix D

-------
    D-10: Empirical Incremental SOi Emissions Rate (pounds per MWh)
                                        AZNM
                                        CAMX
15
20
                                        ERCT
 5
10
15
20
                                        FRCC
                                        MROE
                                        MROW
                                        NEWE
                                        NWPP
                       100
                                  150
                                             200
                                                        250
                                                                   300
                                                                              350
                                   Emissions Rate (Ibs/MWh)
                        Indirect Emissions Analysis • 85
                                Appendix D

-------
        D-10 (continued): Empirical Incremental SOi Emissions Rate


                                         NYCW
 5
10
15
20
 5
10
15
20
10
15
20
 5
10
15
20
                       100
                                         NYLI
                                         NYUP
                                         RFCE
                                         RFCM
                                         RFCW
                                         RMPA
                                         SPNO
                                  150
                                             200
                                                         250
                                   Emissions Rate (Ibs/MWh)
                                                                    300
                                                                               350
                         Indirect Emissions Analysis • 86
                                 Appendix D

-------
        D-10 (continued): Empirical Incremental SOi Emissions Rate


                                        SPSO
 5
10
15
20
 5
10
15
20
                                        SRMV
                       100
                                        SRMW
                                        SRSO
                                         SRTV
                                        SRVC
                                  150
                                             200
                                                        250
                                                                   300
                                                                              350
                                   Emissions Rate (Ibs/MWh)
                                                      10
                         Indirect Emissions Analysis • 87
                                Appendix D

-------
      D-ll: Flexibility-Weighted Hourly Average SOi Emissions Rate

                             (pounds per MWh)
10
15
20

                                        AZNM
                                        CAMX
                                        ERCT
                                        FRCC


                                        MROE
                                        MROW
 5
10
15
20

 5
10
15
20
                                        NEWE
                                        NWPP
                       100
                                  150
                                             200
                                                        250
                                                                   300
                                                                              350
                                  Emissions Rate (Ibs/MWh)
                                                    10
                                                                       14
                        Indirect Emissions Analysis • 88
                                Appendix D

-------
D-ll (continued): Flexibility-Weighted Hourly Average SOi Emissions Rate
                                        NYCW
 5
10
15
20
                                         NYLI
\
i
        WWP
                                     .
                                         NYUP
                                         RFCE
                                        RFCM
 5
10
15
20
 5
10
15
20
                                        RFCW
                                        RMPA
                                        SPNO
 5
10
15
20

   1 .

            50
                       100
                                  150
                                             200
                                                        250
                                                                   300
                                                                              350
                                   Emissions Rate (Ibs/MWh)
                                                     10
                                                                        14
                         Indirect Emissions Analysis • 89
                                Appendix D

-------
D-ll (continued): Flexibility-Weighted Hourly Average SOi Emissions Rate
                                      SPSO
                                      SRMV
                                     SRMW
                                  If ซ<*i JIM
 5
10
15
20
 5
10
15
20
!	1
                                      SRSO
            50
                      100
                                      SRTV
                                      SRVC
                                150
                                          200
                                                     250
                                                               300
                                                                         350
                                 Emissions Rate (Ibs/MWh)
                                                 10
                                                          12
                                                                   14
                       Indirect Emissions Analysis • 90
                              Appendix D

-------
 D-12: Load-Following Incremental SOi Emissions Rate (pounds per MWh)
                                       AZNM
 5
10
15
20
                                       CAMX
 5
10
15
20
                                        ERCT
                                        FRCC
 5
10
15
20
                                       MROW
                                       NEWE
                                       NWPP
                       100
                                 150
                                            200
                                                       250
                                                                  300
                                                                            350
                                  Emissions Rate (Ibs/MWh)
                        Indirect Emissions Analysis • 91
                                Appendix D

-------
     D-12 (continued): Load-Following Incremental SOi Emissions Rate
                                        NYCW
                                        NYLI
 5
10
15
20
                                        NYUP
                                        RFCE
                                        RFCM
 5
10
15
20
                                        RFCW
                                        RMPA
                                        SPNO
                       100
                                  150
                                            200
                                                       250
                                                                  300
                                                                             350
                                  Emissions Rate (Ibs/MWh)
                        Indirect Emissions Analysis • 92
                                Appendix D

-------
     D-12 (continued): Load-Following Incremental SOi Emissions Rate


                                       SPSO
                                       SRMV
                                       SRMW
                                       SRSO
 5
10
15
20
                                        SRTV
                                        SRVC
                       100
                                 150
                                            200
                                                       250
                                                                  300
                                                                             350
                                  Emissions Rate (Ibs/MWh)
                                                     10
                        Indirect Emissions Analysis • 93
                                Appendix D

-------