&EPA
United States
Environmental Protection
Agency
Solid Waste and
Emergency Response
(5305W)
EPA530-R-97-010
March 1997
Greenhouse Gas
Emissions from
Municipal Waste
Management
Draft Working Paper
Printed on paper that contains at least 20 percent postconsumer fiber
-------
-------
GREENHOUSE GAS EMISSIONS FROM
MUNICIPAL WASTE MANAGEMENT
DRAFT WORKING PAPER
Prepared for:
Office of Solid Waste
and
Office of Policy, Planning and Evaluation
U.S. Environmental Protection Agency
Prepared by:
ICF Incorporated
EPA Contract No. 68-W6-0029
Work Assignment 0-06
March 1997
-------
-------
TABLE OF CONTENTS
EXECUTIVE SUMMARY: BACKGROUND AND FINDINGS 1
ES.l GREENHOUSE GASES AND CLIMATE CHANGE 1
ES.2 WHAT IS THE UNITED STATES DOING ABOUT CLIMATE CHANGE? 2
ES.3 WHAT IS THE RELATIONSHIP OF MUNICIPAL SOLID WASTE TO
GREENHOUSE GAS EMISSIONS? 3
ES.4 WHY EPA PREPARED THIS REPORT AND HOW IT WILL BE USED 4
ES.5 HOW WE ANALYZED THE IMPACT OF MUNICIPAL SOLID WASTE
ON GREENHOUSE GAS EMISSIONS 4
ES.6 RESULTS OFTHE INITIAL ANALYSIS 8
ES.7 LIMITATIONS OF THE ANALYSIS H
1. METHODOLOGY 13
1.1 INTRODUCTION 13
1.2 THE OVERALL FRAMEWORK:
A STREAMLINED LIFE CYCLE INVENTORY 13
1.3 MSW MATERIALS CONSIDERED IN THE STREAMLINED
LIFE CYCLE INVENTORY ^3
1.4 KEY INPUTS AND BASELINES FOR THE STREAMLINED
LIFE CYCLE INVENTORY 14
1.5 HOWTHESE INPUTS ARE TALLIED AND COMPARED 18
1.6 SUMMARY ANALYSIS OF THE LIFE CYCLE STAGES 20
2. RAW MATERIALS ACQUISITION AND MANUFACTURING 25
2.1 GHG EMISSIONS FROM ENERGY USE IN RAW MATERIALS
ACQUISITION AND MANUFACTURING 25
2.2 NON-ENERGY GHG EMISSIONS FROM MANUFACTURING
AND RAW MATERIALS ACQUISITION :. 30
2.3 RESULTS 30
2.4 LIMITATIONS 31
3. FOREST CARBON SEQUESTRATION 43
3.1 MODELING FRAMEWORK 45
3.2 THE NORTH AMERICAN PULP AND PAPER MODEL (NAPAP) 47
3.3 THE TIMBER ASSESSMENT MARKET MODEL (TAMM)
AND THE AGGREGATE TIMBERLAND ASSESSMENT SYSTEM (ATLAS) 51
3.4 THE FOREST CARBON MODEL (FORCARB) 55
3.5 THE HARVESTED CARBON MODEL (HARVCARB) 57
3.6 RESULTS 62
3.7 LIMITATIONS ... 64
4. SOURCE REDUCTION AND RECYCLING 67
4.1 GHG IMPLICATIONS OF SOURCE REDUCTION 67
4.2 GHG IMPLICATIONS OF RECYCLING 69
4.3 LIMITATIONS 71
-------
5. COMPOSTING 75
5.1 POTENTIAL GREENHOUSE GAS EMISSIONS .........................." ' 75
5.2 POTENTIAL CARBON SEQUESTRATION 75
5.3 NET GHG EMISSIONS FROM COMPOSTING ....." 80
5.4 LIMITATIONS OF THE ANALYSIS 80
6. COMBUSTION 83
6.1 METHODOLOGY 84
6.2 RESULTS ................... 90
6.3 LIMITATIONS OF THE ANALYSIS '.'.'...................... 91
7. LANDFILLING 93
7.1 EXPERIMENTAL VALUES FOR METHANE GENERATION
AND CARBON SEQUESTRATION 94
7.2 FATES OF LANDFILL METHANE: CONVERSION TO CO2, EMISSIONS,
AND FLARING OR COMBUSTION WITH ENERGY RECOVERY 100
7.3 NET GHG EMISSIONS FROM LANDFILLING . 104
7.4 LIMITATIONS '.'.'.'.'.'.'.'.'.'.'.'.'.'.'. 106
8. COMPARISON OF OPTIONS 109
8.1 FULL LIFE CYCLE GREENHOUSE GAS EMISSIONS
FOR EACH WASTE MANAGEMENT OPTION HO
8.2 COMPARISONS OF THE WASTE MANAGEMENT OPTIONS .......... HO
-------
PREFACE
This document was prepared for the United States Environmental Protection Agency under
Contract No. 68-W2-0018. The work assignment was managed by Eugene Lee of the Office of Solid
Waste. Clare Lindsay of the Office of Solid Waste and Michael Podoisky and Brett Van Akkeren of the
Office of Policy, Planning and Evaluation provided substantive input and assistance.
Mr. Lee would like to thank the primary authors of this report, William Driscoll, Randy Freed, and
Sarah Stafford of ICF Incorporated. EPA also acknowledges the assistance of the following persons in the
development of this report:
Contributors
Frank Ackerman, Tufts University
Terry Boguski, Franklin Associates
Bill Franklin, Franklin Associates
Marge Franklin, Franklin Associates
Melissa Huff, Franklin Associates
John Stutz, Tellus Institute
Brian Zuckerman, Tellus Institute
Reviewers
Dan Abbasi, EPA, OPPE
Dana Arnold, EPA, OSW
Morton Barlaz, North Carolina State
University
Michael Cole, University of Illinois at Urbana-
Champaign
Truett DeGeare, EPA OSW
Bob Dellinger, EPA, OSW
Sy Friedrich, DOE
Linda Gaines, Argonne National Laboratory
George Garland, EPA, OSW
Rosalie Greene, EPA, OSW
Terry Grist, EPA, OSW
Terry Grogan, EPA, OSW
Richard Haynes, USDA, Forest Service
Linda Heath, USDA, Forest Service
Kathleen Hogan, EPA, OAR
Bill Hohenstein, EPA, OPPE
Peter Ince, USDA, Forest Service
Cindy Jacobs, EPA, OAR
Rich Kashmanian, EPA, OPPE
Steve Levy, EPA, OSW
Reid Lifset, Yale University
Richard McClimans, State University of New
York at Syracuse
Thea McManus, EPA, OSW
Sam Napolitano, EPA, OAR
William Seitz, H. John Heinz Center for
Science, Economics and Environment
Doreen Sterling, EPA, OPPE
Frank Stodolsky, Argonne National Laboratory
Andy Teplitzky, EPA, OSW
Susan Thorneloe, EPA, ORD
Carleton Wiles, NREL
Steven Winnett, EPA, OPPE
Lynda Wynn, EPA, OSW
This document is a preliminary draft and does not necessarily reflect the findings, views, or
policies of the Environmental Protection Agency. It will be revised and finalized after review of comments
received from the public.
-------
-------
EXECUTIVE SUMMARY? BACKGROUND AND FINDINGS
One of the largest environmental challenges facing the US is management of municipal solid waste
(MSW). In 1994, the US generated 209 million tons of MSW,1 and per-capita MSW generation rates have
continued to rise throughout the last decade. At the same time, the US recognizes climate change as a
serious issue, and is embarking on a number of actions to reduce the emissions of greenhouse gases
(GHGs) that cause climate change. This report examines how the two issues - MSW management and
climate change - are related, by examining how different management options for MSW may reduce or
increase GHG emissions.
Efforts to slow climate change focus on reducing emissions of carbon dioxide from energy use,
reducing methane emissions, and changing forestry practices to promote long-term storage of carbon in
trees. Different management options for MSW provide many opportunities to affect these processes,
directly or indirectly. This report integrates, for the first time, a wealth of information on GHG implications
of various MSW management options. The findings indicate which management practices are most
favorable, from a GHG perspective, for the most common materials in MSW, and for mixed MSW.
ES.l GREENHOUSE GASES AND CLIMATE CHANGE
Climate change is a serious international environmental concern and the subject of much research
and debate. Many, if not most, of the readers of this report will have a general understanding of the
greenhouse effect and climate change. However, for those who are not familiar with the topic, a brief
explanation follows.2
A naturally occurring shield of "greenhouse gases" (primarily water vapor, carbon dioxide,
methane, and nitrous oxide), comprising 1 to 2 percent of the Earth's atmosphere, traps radiant heat from
the Earth and helps warm the planet to a comfortable, livable temperature range. Without this natural
"greenhouse effect", the average temperature on Earth would be approximately 5 degrees Fahrenheit,
rather than the current 60 degrees Fahrenheit.3
Many scientists, however, are alarmed by a significant increase in the concentration of carbon
dioxide and other GHGs in the atmosphere. Since the pre-industrial era, atmospheric concentrations of
1 U.S. EPA Office of Solid Waste, Characterization of Municipal Solid Waste in the United States: 1995
Update, EPA 530-R-96-001, p. 26.
2 For more detailed information on climate change, please see The Climate Change Action Plan (October
1993); Inventory of US Greenhouse Gas Emissions and Sinks: 1990-1994, EPA 230-R-96-006 (November 1995);
and Climate Change 1995: The Science of Climate Change (J.T. Houghton, etal, eds.; Intergovernmental Panel on
Climate Change [IPCC]; published by Cambridge University Press, 1996). To request a list of additional documents
addressing climate change, call EPA's Climate Change "FAX on Demand" at (202) 260-2860 or access EPA's
global warming web site at www.epa.gov/globalwarming.
3 Houghton, J.T., et al, eds., Intergovernmental Panel on Climate Change, Climate Change 1995: The
Science of Climate Change (Cambridge, England: Cambridge University Press) 1996, pp. 57-58, and U.S. EPA,
Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-1994 (EPA-230-R-96-006) November 1995, p. 1.
DRAFT - March 1997 *
-------
carbon dioxide have increased by nearly 30 percent and methane concentrations have more than doubled.
There is a growing international scientific consensus that this increase has been caused, at least in part, by
human activity, primarily the burning of fossil fuels (coal, oil, and natural gas) for such activities as
generating electricity and driving cars.4
Moreover, there is a growing consensus in international scientific circles that the buildup of carbon
dioxide and other GHGs in the atmosphere will lead to major environmental changes such as: (1) rising sea
levels (that may flood coastal and river delta communities); (2) shrinking mountain glaciers and reduced
snow cover (that may diminish fresh water resources), and (3) the spread of infectious diseases and
increased heat-related mortality. The best current predictions suggest that the rate of climate change
attributable to GHGs will far exceed any natural climate changes that have occurred during the last 10,000
years.5
Many of these changes appear to be occurring already. Global mean surface temperatures have
already increased by about 1 degree Fahrenheit over the past century. A reduction in the Northern
Hemisphere's snow cover, a decrease in Arctic sea ice, a rise in sea level, and an increase in the frequency
of extreme rainfall events have all been documented.6
Such important environmental changes pose potentially significant risks to humans, social systems,
and the natural world. Of course, many uncertainties remain regarding the precise timing, magnitude, and
regional patterns of climate change and the extent to which mankind and nature can adapt to any changes.
It is clear, however, that no changes will be easily reversed for many decades or even centuries because of
the long atmospheric lifetimes of the GHGs and the inertia of the climate system.
ES.2 WHAT IS THE UNITED STATES DOING ABOUT CLIMATE CHANGE?
In 1992, world leaders and citizens from some 200 countries met in Rio de Janeiro, Brazil to
confront global ecological concerns. At this "Earth Summit," 154 nations, including the United States,
signed the Framework Convention on Climate Change, an international agreement to address the danger of
global climate change. The objective of the Convention is to stabilize GHG concentrations in the
atmosphere at a level, and over a time frame, that will minimize man-made climate disruptions.
By signing the Convention, countries make a voluntary commitment to reduce GHGs or take other
actions to stabilize emissions of GHGs at 1990 levels. All parties to the Convention are also required to
develop, and periodically update, national inventories of their GHG emissions. The US ratified the
Convention in October 1992.
One year later, President Clinton issued the US Climate Change Action Plan (CCAP). The plan
calls for returning US GHG emissions to their 1990 levels by the year 2000 through cost-effective
domestic actions and voluntary cooperation with states, local governments, industry, and citizens. The
Administration projected that without the Action Plan, net emissions of GHGs in the US could grow by
4 Climate Change 1995: The Science of Climate Change, op. cit. pp. 3-5.
5 Ibid., pp. 6,29-30, 156, and 371-372.
6 Ibid., pp. 26, 29-30,156, and 171.
DRAFT -- March 1997
-------
about 7 percent between 1990 and 2000. The Action Plan calls for GHG reductions of 108 million metric
tons of carbon equivalent (MTCE)7 to close the gap between the 1990 and projected 2000 levels.
To put these numbers in perspective, consider the following: the average American automobile
each year emits about 10,600 pounds of CO2 or 1.3 metric tons of carbon equivalent (MTCE). Therefore,
reducing GHG emissions by 108 million MTCE, the initial goal of the CCAP, is equivalent to taking 83
million cars off the road.
The CCAP outlines over 50 initiatives to reduce GHG emissions in the US. One of the initiatives
calls for accelerated source reduction and recycling of municipal solid waste through combined efforts by
EPA, the Department of Energy, and the Department of Agriculture.
ES.3 WHAT IS THE RELATIONSHIP OF MUNICIPAL SOLID WASTE TO GREENHOUSE
GAS EMISSIONS?
What does municipal solid waste have to do with rising sea levels, higher temperatures and GHG
emissions? Actually, a lot. For many wastes, the materials that we dispose are left over after a long series
of steps including: (1) extraction and processing of raw materials; (2) manufacture of products; (3)
transportation of materials and products to markets; (4) use by consumers; and (5) waste management.
At virtually every step along this "life cycle", the potential exists for GHG impacts. In its simplest
terms, waste affects GHGs by affecting one or more of the following:
(1) Energy consumption (specifically, combustion of fossil fuels) associated with making,
transporting, and using the product or material that becomes a waste.
(2) Non-energy-related manufacturing emissions, such as the carbon dioxide released when
limestone is converted to lime (which is needed for aluminum and steel manufacturing).
(3) Methane emissions from landfills where the waste is disposed.
(4) Carbon sequestration. Carbon sequestration refers to natural or man-made processes that
remove carbon from the atmosphere and store it for long time periods or permanently. A
store of sequestered carbon (e.g., a forest or coal deposit) is known as a carbon sink.
The first three mechanisms add GHGs to the atmosphere and contribute to global warming. The
fourth mechanism - carbon sequestration - reduces GHG concentrations by removing carbon dioxide from
the atmosphere. Forests are one mechanism for sequestering carbon; if more wood is grown than is
harvested in a given year, the amount of carbon from atmospheric carbon dioxide that is stored in trees
increases, and thus carbon is sequestered.
7 The greenhouse gases considered in this report - carbon dioxide, methane, nitrous oxide, and
perfluorocarbons - have very different heat trapping potentials. Throughout this report, we use "metric tons of
carbon equivalent" to express the heat trapping effect of emissions of these gases. International protocols establish
carbon dioxide as the reference gas for measurement of global warming, so by definition, the global warming
potential of carbon dioxide (the measure of its heat trapping potential) is 1.0. Because carbon dioxide is 27 percent
carbon, by weight, a metric ton of carbon dioxide emissions is 0.27 MTCE. As described in Chapter 1, methane and
nitrous oxide have much higher heat-trapping capacities than carbon dioxide.
DRAFT -- March 1997 3
-------
Different wastes and waste management options have different implications for energy
consumption, methane emissions, and carbon sequestration. Source reduction and recycling of paper
products, for example, reduce energy consumption, decrease landfill methane emissions, and increase
forest carbon sequestration.
ES.4 WHY EPA PREPARED THIS REPORT AND HOW IT WILL BE USED
Recognizing the potential for source reduction and recycling of municipal solid waste to reduce
GHG emissions, EPA included a source reduction and recycling initiative in the original Climate Change
Action Plan. At that time, EPA estimated that its portion of the source reduction and recycling initiative
could reduce annual GHG emissions by roughly 5.6 million MTCE by the year 2000, or about 5 percent of
the overall goal of the Action Plan. To make these projections, EPA used limited data on energy
consumption and forest carbon sequestration to estimate how a 5 percent increase in both source reduction
and recycling would affect GHG emissions in 2000.
It was clear then that a rigorous analysis would be needed, both to more accurately gauge the total
GHG emission reductions achievable through source reduction and recycling, and to identify which
materials found in MSW were most likely to bring about the greatest climate benefits if they were reduced
or recycled. Moreover, it was clear that all of the options for managing MSW should be considered. This
way, one could determine the relative GHG impacts of various waste management approaches for a given
material or mix of materials. To this end, the Office of Policy, Planning and Evaluation and the Office of
Solid Waste launched a major research effort. This report represents the initial results. We look forward to
input from interested reviewers to correct, expand, and clarify its contents.
We anticipate four potential applications for the GHG emission estimates provided here. First,
organizations that are interested in quantifying GHG emission reductions due to source reduction or
recycling may use these estimates for that purpose; in particular, EPA may use these estimates as the basis
for developing guidance for voluntary reporting of GHG reductions, as authorized by Congress in Section
1605(b) of the Energy Policy Act of 1992. Second, the estimates may be useful for evaluation of MSW
management options on a national, regional, state, or local basis. Third, EPA plans to use the estimates to
evaluate its progress in reducing US GHG emissions, by promoting source reduction and recycling through
programs such as WasteWi$e and Unit-Based Pricing, as part of the US Climate Change Action Plan.
Finally, this report may also assist other countries involved in developing GHG emissions estimates for
their solid waste streams.
ES.5 HOW WE ANALYZED THE IMPACT OF MUNICIPAL SOLID WASTE ON
GREENHOUSE GAS EMISSIONS
To measure the GHG impacts of municipal solid waste (MSW), one must first decide which
components of the waste to analyze. We surveyed the universe of materials and products found in MSW and
determined which were most likely to have the greatest impact on GHGs. These determinations were based on
the quantity generated, and on the different amounts of energy used to manufacture the product from virgin
DRAFT -- March 1997
-------
rather than recycled inputs. By this process, we limited the analysis to the following ten items:8'9
newspaper,
office paper,
corrugated cardboard,
aluminum cans,
steel cans,
HDPE (high density polyethylene) plastic,
LDPE (low density polyethylene) plastic,
PET (polyethylene terephthalate) plastic,
food scraps, and
yard trimmings.
The foregoing materials constitute 52 percent, by
weight, of municipal solid waste, as shown in Exhibit ES-1.10
We also examined the GHG implications of managing mixed
MSW in various ways.
Next, we developed a streamlined life cycle inventory
for each of the chosen materials. Our analysis is streamlined in
the sense that it examines GHG emissions only, and is not a
more comprehensive environmental analysis of all emissions
from municipal solid waste management options.11
Exhibit ES-1
Percentage of 1994 US Generation of
MSW for Materials in This Report
Material
Newspaper
Office paper
Corrugated cardboard
Aluminum cans
Steel cans
HDPE plastic
LDPE plastic
PET plastic
Food scraps
Yard trimmings
TOTAL
Percentage of
MSW Generation
(by Weight)
6.5%
3.2%
13.6%
0.8%
1.4%
1.9%
... 2.7%
0.5%
6.7%
14.6%
52%
Source: Franklin Associates, Ltd.,
Characterization of Municipal Solid Waste in the
United States: 1995 Update, EPA530-R-96-001.
8 Glass was not included in the analysis, in part because of the relatively small difference between the
amount of energy used in manufacturing glass from virgin rather than recycled inputs.
9 There is substantial opportunity for greenhouse gas reductions in other segments of the municipal solid
waste stream. EPA's Office of Solid Waste plans to investigate the potential for GHG reductions for selected
additional materials in MSW.
10 See Characterization of Municipal Solid Waste in the United States: 1995 Update. EPA 530-R-96-001
(March 1996).
11 EPA?s Office of Research and Development (ORD) is in the early stages of performing a more extensive
application of life cycle assessment for various waste management options for MSW. ORD's analysis will inventory
all emissions (air, water, and waste) associated with these options.
DRAFT -- March 1997
-------
We focused on those aspects of the life cycle that have the potential to emit GHGs as materials
change from their raw states, to products, to waste. Exhibit ES-2 shows the steps in the life cycle in which
GHGs are emitted, carbon sequestration is affected, and utility energy is displaced. As shown, we
examined the potential for these effects at the following points in a product's life cycle:
raw material acquisition (fossil fuel energy and other emissions, and change in forest
carbon sequestration);
manufacturing (fossil fuel energy emissions); and
waste management (carbon dioxide emissions associated with combustion and methane
emissions from landfills, both offset by some avoided utility fossil fuel use; and carbon
sequestration in landfills).
At each of these points, we also considered transportation-related energy emissions.
We did not analyze the GHG emissions associated with consumer use of products because energy
use for the selected materials is small (or zero) at this point in the life cycle, and in any case, the energy
consumed during use would be about the same whether the product was made from virgin or recycled
inputs.
We use a "zero-impact" scenario as the baseline for this analysis. In other words, for each material,
we assumed in the baseline that the material is not made in the first place - thus, there are no GHG
emissions attributable to the material. The current stock of carbon sinks (e.g., standing trees, existing
landfills) was the baseline against which changes in carbon sinks were measured. Using this consistent
baseline enabled direct comparison of the net GHG emissions of all waste management options: source
reduction, recycling, composting, combustion, and landfilling.
Exhibit ES-3 shows how the GHG sources and sinks are affected by each waste management
strategy. For example, the top row of the exhibit shows that source reduction12 results in no GHG
emissions from raw materials acquisition, manufacturing, or waste management; but, source reduction of
paper products increases forest carbon sequestration.
12 In this analysis, source reduction means using less of a given product without using more of some other
product-e.g., making aluminum cans with less aluminum ("lightweighting"), or double-sided rather than single-
sided photocopying. We did not consider source reduction of one product that would be associated with substitution
by another product - e.g., substituting plastic boxes for corrugated paper boxes. Nor did we estimate the potential
for source reduction of chemical fertilizers and pesticides with increased production and use of compost.
<> DRAFT -- March 1997
-------
Exhibit ES-2
GHG Sources and Sinks Associated with Materials in the MSW Stream
Inputs
Ore, trees,
petroleum,
energy, etc.
Energy
Life Cycle Stage1 GHG Emissions/Carbon Sinks2
Raw Materials
Acquisition
Manufacturing
1
Use
Waste
Management
1 Note that source reduction affects aU stages in the life cycle.
1 All life cycle stages include transportation energy-related emissions
(except that emissions from transporting products from manufacturers
to consumers were not counted in this analysis).
Energy-related emissions
Non-energy related emissions
Change in carbon storage in forests
Energy-related emissions
Energy-related emissions
Change in carbon storage in soils
CO2 emissions from plastics
^ N2O emissions
Credit for avoided fossil fuel use
CH4 emissions
- Uncontrolled
- Flared or recovered for energy (converted to CO2)
- Credit for avoided fossil fuel use
Credit for Carbon in long-term storage _,,,,
CO0023-1-1
-------
Exhibit ES-3
Components of Net Emissions for Various Municipal Solid Waste Management Strategies
Municipal
Solid Waste
Management
Strategy
Source
Reduction
Recycling
Composting
Combustion
Landfllling
Greenhouse Gas Sources and Sinks
Process and Transportation
GHGs from Raw Materials
Acquisition and
Manufacturing
No emissions/sinks
Decrease in GHG emissions
due to lower energy
requirements (compared to
manufacture from virgin
inputs) and avoided process
non-energy GHGs
No emissions/sinks3
Process and transportation
emissions
Process and transportation
emissions
Change in
Forest Carbon
Storage
Increase in
forest carbon
storage
Increase in
forest carbon
storage
No change
No change
No change
Change in
Soil Carbon
Storage
No change
No change
Increase in
soil carbon
storage
No change
No change
Waste Management GHGs
No emissions/sinks
No emissions/sinks
Compost machinery emissions
and transportation emissions
Nonbiogenic CO2, N2O
emissions, avoided utility
emissions, and transportation
emissions
Methane emissions, long-term
carbon storage, avoided utility
emissions, and transportation
emissions
* No manufacturing transportation GHG emissions are considered for composting of food scraps and yard
trimmings because these materials are not considered to be manufactured.
ES.6 RESULTS OF THE INITIAL ANALYSIS
The results of this initial research suggest that the solid waste management hierarchy,13 which
ranks management options for environmental benefits, is also generally valid from a GHG perspective.
Exhibit ES-4 compares the GHG impacts of source reduction, recycling, composting, combustion, and
landfllling for each of the ten materials we studied. Most strikingly, the findings indicate the following:
* Source reduction generally has much lower GHG emissions than the other waste
management strategies.14
13 The solid waste management hierarchy ranks management options, from those with the most
environmental benefits to those with the least, as follows: source reduction, including backyard composting;
recycling, including centralized composting; and combustion and landfllling.
14 As noted above, the only source reduction strategy analyzed in this study is lightweighting.
Consequently, the results shown here do not apply to material substitution.
8
DRAFT - March 1997
-------
Exhibit ES-4
Greenhouse Gas Emissions from Source Reduction and MSW Management Options
(Assuming Initial Production Using the Current Mix of Virgin and Recycled Inputs)
(MTCE/Ton) *
Material
Newspaper
Office Paper
Corrugated Cardboard
Aluminum Cans
Steel Cans
HOPE
LDPE
PET
Food Scraps
Yard Trimmings
Mixed MSW
Net Source
Reduction Emissions
-0.48
-0.53
-0.44
0.00
0.00
0.00
0.00
0.00
NA
NA
NA
Net Recycling
Emissions**
-0.37
-0.29
-0.30
-1.011
0.30
0.34
0.36
0.35
NA
NA
NA
Net Composting
Emissions***
NA
NA
NA
NA
NA
NA
NA
NA
0.00
0.00
NA
Net Combustion
Emissions**
0.40
0.46
0.32
2.97
0.47
1.22
1.38
1.38
-0.01
-0.02
0.04
Net Landfilling
Emissions**
0.28
1.09
0.44
2.97
0.88
0.73
0.88
0.99
0.09
-0.07
0.00
NA: Not applicable, or in the case of composting of paper, not analyzed.
*MTCE/ton: Metric tons of carbon equivalent per short ton of material. Material tonnages are on an as-managed (wet weight) basis.
"Includes emissions from the initial production of the material being managed, except for food waste, yard waste, and mixed MSW.
"There is considerable uncertainty in our estimate of net GHG emissions from composting; the values of zero are plausible values
based on assumptions and a bounding analysis.
1For a discussion of why recycling of aluminum cans reduces GHG emissions more than source reduction (based on the current mix of
virgin and recycled inputs), please see chapter four.
-------
Recycling generally has the second lowest GHG emissions.
Of the remaining management strategies, the strategy with the next lowest GHG emissions
differs for different materials.
Composting is a management option for food scraps and yard trimmings; the net
GHG emissions from composting, combusting, or landfllling these materials are
similar, given the uncertainty in the analysis.
Combustion has lower GHG emissions than landfllling for office paper,
corrugated cardboard, and steel cans, because office paper and corrugated
cardboard generate a substantial amount of methane when landfilled, and steel is
recovered for recycling at most MSW combustors.
Landfllling has lower GHG emissions than combustion for plastics and
newspaper, because combustion of plastic results in substantial nonbiogenic CO2
emissions, and landfllling of newspaper results in substantial carbon
sequestration.15
The net GHG emissions from combustion and landfllling are similar for
aluminum cans.
The ordering of combustion, landfllling, and composting is affected by (1) the GHG
inventory accounting methods, which do not count CO2 emissions from sustainable
biogenic sources,16 but do count emissions from sources such as plastics, and (2) a series
of assumptions on sequestration, future use of methane recovery systems, landfill gas
recovery system efficiency, ferrous metals recovery, and avoided utility fossil fuels. On a
site-specific basis, the ordering of results between a combustor and a landfill could be
different from the ordering provided here, which is based on national average results.
In sum, material-by-material and for MSW as a whole, source reduction and recycling have the
potential to substantially decrease emissions of GHGs.
The full life cycle GHG emissions for each of the first four waste management strategies - source
reduction, recycling, composting, and combustion - are compared to the GHG emissions from landfllling
in Exhibit ES-5. This exhibit shows the GHG values for each of the first four management strategies,
minus the GHG values for landfllling. This exhibit is provided because landfllling is often viewed as the
baseline waste management strategy. With this exhibit, one may compare the GHG emissions from other
waste management options to the GHG emissions from landfllling. Exhibit ES-5 shows that source
reduction and recycling have lower GHG emissions than landfllling for all of the materials considered.
15 Note, however, that from the standpoint of combustor operation, combustors typically rely on paper and
plastic as fuel.
16 Sustainable biogenic sources include paper and wood products from sustainably managed forests; when
these materials are burned or aerobically decomposed to CO2,the CO2 emissions are not counted. Our approach to
measuring GHG emissions from biogenic sources is described in detail in Chapter 1.
10
DRAFT - March 1997
-------
There are only small differences (less than 0.1 MTCE/ton) between composting food scraps and yard
trimmings, and landfilling these materials. Combustion has higher GHG emissions than landfilling for
newspaper and plastics, but has lower GHG emissions than landfilling for office paper and steel cans.
Differences between combustion and landfilling are small for corrugated cardboard, aluminum cans, food
scraps, yard trimmings, and mixed MSW.
We also conducted a sensitivity analysis to examine the GHG emissions from landfilling under
varying assumptions about the percentage of landfilled waste sent to landfills with methane recovery by the
year 2000. Our results showed that net GHG emissions for landfilling of mixed MSW are positive at lower
rates of methane recovery, and only turn negative (i.e., carbon sequestration in the landfill and avoided
electric utility emissions more than offset the methane emissions) when the LFG recovery rate exceeds 50
percent. For more details on this sensitivity analysis, see section 7.4 and Exhibit 7-6.
ES.7 LIMITATIONS OF THE ANALYSIS
When conducting this analysis, we used a number of analytical approaches and numerous data
sources, each with its own limitations. In addition, we postulated assumptions throughout the analysis.
Some of the major limitations follow:
The landfill analysis (1) is based on laboratory data from a single researcher, and (2) uses
as a baseline landfill methane recovery levels projected for the year 2000.
The forest carbon sequestration analysis (1) uses a point estimate for forest carbon
sequestration, whereas the system of models predicts changing net sequestration over time
and (2) assumes that no forested lands will be converted to non-forest uses as a result of
increased paper recycling.
The combustion analysis uses national average values for a number of parameters that may
not be representative of a given combustor facility.
The manufacturing GHG analysis is based on estimated industry averages for energy
usage, and in some cases the estimates are based on limited data. In addition, we used
values for the average GHG emissions per ton of material produced, not the marginal
emission rates per incremental ton produced. In some cases, the marginal emission rates
may be significantly different.
The composting analysis was limited by the lack of data on methane generation and
carbon sequestration resulting from composting; we relied on a theoretical approach to
estimate the values.
DRAFT --March 1997
11
-------
Exhibit ES-5
Greenhouse Gas Emissions of MSW Management Options Compared to Landfilling
(MTCE/Ton)
Material
Newspaper
Office Paper
Corrugated Cardboard
Aluminum Cans
Steel Cans
HOPE
LDPE
PET
Food Scraps
Yard Trimmings
Mixed MSW
Source Reduction
Net Carbon
Minus Landfilling Net Carbon
Current Mix of Inputs
-0.76
-1.62
-0.89
-2.97*
-0.88
-0.73
-0.88
-0.99
NA
NA
NA
100% Virgin Inputs
-1.07
-1.85
-1.15
-5.52*
-1.13
-0.73
-0.92
-1.19
NA
NA
NA
Recycling Net Carbon
Minus Landfilling
Net Carbon
-0.65
-1.38
-0.74
-3.98*
-0.58
-0.39
-0.52
-0.64
NA
NA
NA
Composting Net C
Minus Landfilling
Net Carbon
NA
NA
NA
NA
NA
NA
NA
NA
-0.09
0.07
NA
Combustion Net Carbon
Minus Landfilling
Net Carbon
0.12
-0.63
-0.12
0.00
-0.42
0.50
0.50
0.38
-0.10
0.05
0.04
*For a discussion of why recycling of aluminum cans reduces GHG emissions more than source reduction, please see chapter four.
-------
1. METHODOLOGY
1.1 INTRODUCTION
This chapter provides an overview of the methodology we used to calculate the greenhouse gas
(GHG) emissions associated with various management strategies for municipal solid waste (MSW). The
chapter begins with a brief discussion of the life cycle framework used for the analysis. Next, it explains
how we selected the ten materials that were analyzed. We then describe the specific GHG emissions and
emission offsets considered in calculating the net emissions associated with particular waste management
options. Finally, the chapter discusses the life cycle stages that we studied to identify the GHG impacts of
MSW management options. Succeeding chapters will describe how we analyzed each step in the life cycle.
1.2 THE OVERALL FRAMEWORK: A STREAMLINED LIFE CYCLE INVENTORY
Early in our analysis of the GHG benefits of source reduction and recycling, it became clear that
comparing source reduction and recycling to other waste management options would clarify where the
greatest GHG benefits could be obtained for particular materials in MSW, and help policymakers identify
the best options for GHG reductions. We determined that a streamlined application of life cycle assessment
would be the best way to make such comparisons.
A full life-cycle assessment (LCA) is an analytical framework for understanding the material inputs,
energy inputs, and environmental releases associated with manufacturing, using, and disposing a given
material. A full LCA generally consists of four parts: (1) goal definition and scoping, (2) an inventory of the
materials and energy used in all stages in the life of a product or process, and an inventory of environmental
releases throughout the product lifecycle; (3) an impact assessment that examines potential and actual human
health effects related to the use of resources and environmental releases; and (4) an assessment of the change
that is needed to bring about environmental improvements in the product or processes.
A full life-cycle assessment is beyond the scope of this analysis. Rather, this report is a streamlined
application of a life cycle assessment that is limited to an inventory of the emissions and other
environmental impacts related to global warming; we did not assess air, water, or environmental impacts
that did not have a direct bearing on climate change. Moreover, we did not attempt, as part of this analysis,
to assess human health impacts or environmental improvements needed.17
1.3 MSW MATERIALS CONSIDERED IN THE STREAMLINED LIFE CYCLE INVENTORY
We made initial rough estimates of the potential for source reduction and recycling of MSW to reduce
GHG emissions for the President's Climate Change Action Plan in 1993. However, it was clear that a more
rigorous analysis would be needed to determine the GHG emissions associated with source reduction and
recycling and to identify which materials in MSW were most likely to reduce GHG emissions if source
reduced or recycled.
17 Note that EPA's Office of Research and Development (ORD) is in the early stages of performing a more
extensive life cycle inventory for various waste management options for MSW. ORD's analysis is inventorying all
emissions (air, water, and waste) associated with these options.
DRAFT - March 1997 13
-------
Each material in MSW has different GHG impacts depending on how it is made and disposed. To
determine which materials in MSW had the greatest potential to reduce GHG emissions if source reduced
or recycled, we performed a screening analysis of 37 of the most common materials and products found in
MSW. The screening analysis compared: (1) the GHG emissions from manufacturing each of the 37
materials from virgin or recycled inputs (based on the process and transportation energy requirements, and
fuel mix for each material); and (2) the projected source reduction and recycling rates for each material.
The information on energy requirements, fuel mix, and recycling rates was estimated independently by two
groups with experience in MSW and life cycle assessment: Franklin Associates, Ltd. and the Tellus
Institute. Then, ICF Incorporated ranked the materials by their potential for GHG reductions: for each
material, ICF (1) averaged the two estimates for energy requirements and fuel mix, then (2) used those
averages, together with estimates of the GHG emissions per unit of fuel used, to estimate GHG reductions
per ton of product source reduced or recycled, and finally (3) used the estimated GHG reductions per ton,
together with the averaged estimates of the potential tonnage of source reduction and recycling, to estimate
the total GHG reduction potential for each material.
While the screening analysis was general in nature and employed many assumptions, the underlying
data provided by Franklin Associates and the Tellus Institute overlapped a great deal. The energy and
recycling data provided by both groups indicated that the same eight manufactured materials had the
greatest potential to reduce GHG emissions if they were source reduced or recycled. We chose to limit the
life cycle assessment to these eight materials:
newspaper,
office paper,18
corrugated cardboard,
aluminum cans,
steel cans,
HDPE (high density polyethylene) plastic,
LDPE (low density polyethylene) plastic, and
PET (polyethylene terephthalate) plastic.19
To round out the analysis, we also examined the GHG implications of various management strategies for
food waste, yard trimmings, and mixed MSW.
1.4 KEY INPUTS AND BASELINES FOR THE STREAMLINED LIFE CYCLE INVENTORY
Evaluating the GHG emissions of waste management requires analysis of three factors: 1) GHG
emissions throughout the life cycle of the material (including the chosen disposal option); 2) the extent to
which carbon sinks are affected by manufacturing and disposing the material; and 3) the extent to which
the management option recovers energy that can be used to replace electric utility energy, thus reducing
utility GHG emissions. In addition, a baseline year must be selected so that changes may be measured in
comparison to conditions in that baseline year. Each of these factors warrants further discussion.
18 Office paper refers to the type of paper used in computer printers and photocopiers.
19 Glass was not included in the analysis, partly because of the relatively small difference between the
amount of energy used in manufacturing glass from virgin versus recycled inputs.
14 DRAFT -- March 1997
-------
GHGs Emissions Relevant to Waste:
The most important GHGs for purposes of
analyzing MSW management options are
carbon dioxide, methane, nitrous oxide, and
perfluorocarbons. Of these, carbon dioxide
(CO2) is by far the most common GHG emitted
in the US. Most carbon dioxide emissions result
from energy use, particularly fossil fuel
combustion. A great deal of energy is
consumed when a product is made and then
thrown away. This energy is used in the
following stages: 1) extracting and processing
raw materials; 2) manufacturing products; 3)
managing products at the end of their useful
lives; and 4) transporting materials and
products between each stage of their life cycles.
We estimated energy-related GHG emissions at
all of these stages, except for transportation of
products to consumers (because GHG
emissions from transportation to consumers will
vary little among the options considered). Much
of this report is devoted to explaining how we
quantified the energy used - and the resulting
carbon dioxide emissions - at each stage in the
life cycle of any given material in MSW.
Energy consumed in connection with consumer
use of products is not evaluated, because energy
use for the selected materials is small (or zero)
at this point in the life cycle, and in any case,
the energy consumed during use would be
about the same whether the product was made
from virgin or recycled inputs.
Comparing GHGs
Carbon dioxide, methane, and nitrous oxide
are very different gases when it comes to their heat-
trapping potential. An international protocol has
established carbon dioxide as the reference gas for
measurement of heat-trapping potential (also known
as global warming potential). By definition, the
global warming potential of one kilogram (kg) of
carbon dioxide is 1.
Methane, which has a much higher heat-
trapping capacity, has a global warming potential of
24.5. This means that one kg of methane has the
same heat-trapping potential as 24.5 kg of CO2.
Nitrous oxide, a more potent GHG, has a
global warming potential of 270.
Perfluorocarbons have extremely high
global wanning potentials: 6,300 for CF4 and
12,500 for C2F6.
In this report, emissions of carbon dioxide,
methane, nitrous oxide, and perfluorocarbons have
been converted to their "carbon equivalents."
Because CO2 is 12/44 carbon by weight, one metric
ton of CO2 is defined as 12/44 or 0.27 metric tons
of carbon equivalent (MTCE). The MTCE value
for one metric ton of each of the other gases is
determined by multiplying its global warming
potential by a factor of 12/44. (All data provided
here are from US EPA, Inventory of US
Greenhouse Gas Emissions and Sinks: 1990-1994,
November 1995, p. 3.)
Methane (CH4), a more potent GHG, is
produced when organic waste decomposes in an
oxygen-free (anaerobic) environment, such as a
landfill. Methane from landfills is the largest
source of methane in the US;20 these emissions are addressed in Chapter 7.
Nitrous oxide (N2O), another GHG, results from the use of commercial and organic fertilizers and
fossil fuel combustion, as well as other sources. For this analysis, we estimated nitrous oxide emissions
from waste combustion.
20 U.S. EPA, Inventory of US Greenhouse Gas Emissions and Sinks: 1990-1994, EPA 230-R-96-006,
November 1995, p. ES-10.
DRAFT -- March 1997
15
-------
Perfluorocarbons (CF4 and C2F6) are emitted during the reduction of alumina to aluminum in the
primary smelting process. The source of fluorine for CF4 and C2F6 is the molten cryolite (Na3AlF6) in
which the reduction of alumina occurs. Perfluorocarbons are formed when the fluorine in cryolite reacts
with the carbon in the anode (a carbon mass of paste, coke briquettes, or prebaked carbon blocks), and in
the carbon lining that serves as the cathode. Although the quantities of perfluorocarbons emitted are small,
these gases are significant because of their high global warming potential.
The baseline against which total GHG emissions are calculated is the zero-emissions scenario, in
which no product is made. Thus, in the baseline, there are no GHG emissions in any of the following life
cycle stages: raw materials acquisition, manufacturing, or MSW management.
Carbon Stocks and Carbon Sequestration Relevant to Waste: Carbon, like many other elements,
cycles throughout earth's air, water, land, and biota. A carbon stock (or sink) is a point in the carbon cycle
where carbon is stored. While the carbon is stored, it is not in the atmosphere contributing to the
"greenhouse effect" (i.e., the trapping of heat close to the earth's surface). Examples of carbon stocks are
forests, oceans, oil fields, and landfills.
"Carbon sequestration" is the opposite of GHG emissions. With carbon sequestration, carbon is
removed from the carbon cycle and added to a carbon stock. For example, when a forest removes carbon
from the atmosphere and converts it to wood at a faster pace than the trees are harvested (or decompose),
this is known as forest carbon sequestration. Likewise, if organic matter added to a landfill does not
decompose into methane or carbon dioxide, and enters into long-term storage, it is said to be
"sequestered."
The baseline against which future carbon stocks are measured is the current set of carbon stocks.
For the forest carbon stock, using the current stock of forest carbon as the baseline is based on an
assumption that the forest will be harvested on a sustainable basis (i.e., trees will be grown at a rate at least
equal to the rate at which they are cut).21 Thus, we assume in the baseline that harvesting trees results in no
diminution of the forest carbon stock and no additional carbon dioxide in the atmosphere. On the other
hand, forest carbon sequestration increases as a result of source reduction or recycling of paper products
because both source reduction and recycling cause annual tree harvests to drop below the annual growth of
forests. Consequently, source reduction and recycling "get credit" for increasing the forest carbon stock,
whereas other waste management options (composting, combustion, and landfilling) do not.
Landfills are another means by which carbon is removed from the atmosphere. Landfill carbon
stocks increase over time because much of the organic matter placed in landfills does not decompose,
especially if the landfill is located in an arid area. However, not all carbon in landfills is counted in
determining the extent to which landfills are carbon stocks. For example, the analysis does not count
plastic in landfills toward carbon sequestration. Plastic in a landfill represents merely a transfer from one
carbon stock (the oil field containing the petroleum or natural gas from which the plastic was made) to
21 Assuming a sustainable harvest is reasonable because the US is currently experiencing net reforestation:
that is, more trees are being planted and grown than the US is consuming. This may come as a surprise to some who
live in areas of the country that are being rapidly developed. However, changes in agricultural management
practices, increased productivity per agricultural acre, and other factors are causing large areas of land that were
once cultivated or otherwise disturbed by man to revert to forest. In addition, the average mass of Wood per forested
acre is increasing because, on average, forests are getting older, and the average tree is getting bigger and storing
more carbon.
16
DRAFT -- March 1997
-------
another carbon stock (the landfill); thus, there has been no change in the overall amount of carbon stored.
On the other hand, that portion of organic matter (such as yard trimmings) that does not decompose in a
landfill represents an addition to a carbon stock, because it would have largely decomposed into CO2 if left
to deteriorate on the ground.
While changes in fossil fuel carbon stocks (i.e., reductions in oil field stores that result from the
extraction and burning of oil resources) are not measured directly in this analysis, the reduction in fossil
fuel carbon stocks is indirectly captured by counting the CO2 emissions from fossil fuel combustion in
calculating GHG emissions.
Avoided Electric Utility GHG Emissions Relevant to Waste: When a waste is used to generate
electricity (either through combustion or recovery of methane from landfills), it displaces utility fossil fuels
that would otherwise be consumed. Fossil fuel combustion is the single largest source of GHGs in the US.
When waste is substituted for fossil fuel to generate electricity, the GHG emissions from burning the waste
are offset by the avoided electric utility GHG emissions.
Baseline Year: For most parts of the analysis, we selected as the baseline year the most recent year
for which data were available. For the baseline landfill methane recovery rate, however, we used values
projected for the year 2000. For paper recovery, we made annual projections through 2010 that enabled us
to develop an average value for the period from 1996 through 2010.22 In both cases, we developed future
scenarios because some of the underlying factors that affect GHG emissions are changing rapidly, and we
are seeking to define relationships (e.g., between tonnage of waste landfilled and methane emissions) that
represent an average over the next several years.
In the case of landfill methane, there are three EPA programs that reduce methane emissions:
one that requires landfill gas recovery at large landfills; one that promotes recovery of landfill
methane on a voluntary basis at smaller landfills; and another that promotes source reduction
and recycling (which results in less methane-producing waste being landfilled). In estimating
the landfill methane emission reductions due to source reduction and recycling, we needed to
account for the planned increase in landfill methane capture. Otherwise, EPA would be
counting landfill methane emissions reductions twice: once for landfill methane capture, and
once for source reduction and recycling. Because the programs to regulate landfill gas and
promote voluntary methane recovery will be fully effective by 2000 (dramatically increasing
methane recovery), by using a baseline year of 2000 we avoided double counting.
For paper recovery, earlier analyses had indicated that the marginal impact of increased paper
recovery on forest carbon sequestration changes over time; the impact also differs depending
on the initial paper recovery rate and how that rate changes over time. To estimate the impact
of increased paper recovery on forest carbon sequestration, we needed to account for these
influences. First, we developed a baseline projection for paper recovery rates. We began with
a projection, from the American Forest and Paper Association, that paper recovery rates will
continue to increase from about 35 percent in 1994 to 50 percent by 2000. Then we
developed a baseline scenario for paper recovery whose trajectory passes through 50 percent
in 2000, with continued modest increases in the following years, Because we needed to
22Actually, the models we used simulated carbon sequestration through 2040, but we selected a value based
oh average conditions through 2010.
DRAFT - March 1997 17
-------
estimate the effect of efforts (e.g., by EPA) to enhance recovery beyond the baseline
projected rates, we developed a plausible scenario for enhanced paper recovery rates, and
then compared the predicted forest carbon sequestration under the two scenarios. (Our
approach is fully described in chapter 3).
1.5 HOW THESE INPUTS ARE TALLIED AND COMPARED
Exhibit 1-1 shows the GHG sources and carbon sinks associated with the manufacture of various
materials, and the post-consumer management of these materials as wastes. As shown in the exhibit, GHGs
are emitted from: (1) the pre-consumer stages of raw materials acquisition and manufacturing; and (2) the
post-consumer stage of waste management. No GHG emissions are attributed to the consumer's use of any
product.
To calculate the net GHG implications of a waste management strategy for a given material, one
must determine the difference between: (1) the GHG emissions associated with that material; and (2) any
increases in carbon stocks and/or displaced fossil fuel combustion that offset these emissions. The formula
for net GHG emissions is as follows:
Net GHG emissions =
Gross GHG emissions - (increase in carbon stocks + avoided utility
GHG emissions)
Comparing GHG emissions and carbon sinks for each manufacturing and waste management option
with a consistent baseline allows the net GHG emissions for each option to be compared. From these
comparisons, one may identify which options have the lowest net GHG emissions. For example, when a
material is source reduced (i.e., some or all of it is not produced), GHG emissions throughout the life cycle
are avoided. In addition, when paper products are source reduced, additional carbon is sequestered in
forests.
Similarly, when a material is recycled, the GHG emissions from making an equivalent amount of
material from virgin inputs are avoided. However, there are GHG emissions from making the material from
recycled inputs. Generally, recycling reduces GHG emissions because, in most cases, manufacturing a
product from recycled inputs requires less energy than making the product from virgin inputs, and thus
reduces energy-related GHG emissions. In the case of paper, recycling also results in additional carbon
sequestration in forests.
If a waste is not source reduced or recycled, it may be either composted (if it is organic matter),
combusted, or landfilled. In any of these cases, the full GHG emissions associated with making the
material/product are counted. These GHG emissions may be augmented by methane emissions from
landfills (which themselves may be offset to some degree by energy recovery at landfills or landfill carbon
sequestration). If the wastes are combusted, there may be an offset for avoided utility emissions.
18
DRAFT -- March 1997
-------
Exhibit 1-1
GHG Sources and Sinks Associated with Materials in the MSW Stream
Inputs.
Ore, trees,
petroleum,
energy, etc.
Energy
Energy
Life Cycle Stage1 GHG Emissions/Carbon Sinks2
Raw Materials
Acquisition
i
Manufacturing
Use
Waste
Management
1 Note that source reduction affects all stages in the life cycle.
1 All fife cycle stages include transportation energy-related emissions
(except that emissions from transporting products from manufacturers
to consumers were not counted in this analysis).
Energy-related emissions
Non-energy related emissions
Change in carbon storage in forests
Energy-related emissions
Energy-related emissions
Change in carbon storage in soils
CO2 emissions from plastics
^ N2O emissions
Credit for avoided fossil fuel use
CH4 emissions
- Uncontrolled
- Flared or recovered for energy (converted to C02)
- Credit for avoided fossil fuel use
Credit for Carbon in long-term storage
c6oQ23-1-1
-------
Exhibit 1-2 indicates how the GHG sources and sinks have been counted for each MSW
management strategy to estimate net GHG emissions. For example, the top row of the exhibit shows that
source reduction results in no GHG emissions or sinks (long-term carbon storage) from raw materials
acquisition and manufacturing, soil carbon storage, or waste management; however, there is an increase in
forest carbon sequestration.
Exhibit 1-2
Components of Net Emissions for Various Municipal Solid Waste Management Strategies
Municipal
Solid Waste
Management
Strategy
Source
Reduction
Recycling
Composting
Combustion
Landfilling
Greenhouse Gas Sources and Sinks
Process and Transportation
GHGs from Raw Materials
Acquisition and
Manufacturing
No emissions or sinks
Decrease in GHG emissions
due to lower energy
requirements (compared to
manufacture from virgin
inputs) and avoided process
non-energy GHGs
No emissions or sinks
Process and transportation
emissions
Process and transportation
emissions
Change in
Forest Carbon
Storage
Increase in
forest carbon
storage
Increase in
forest carbon
storage
No change
No change
No change
Change in
Soil Carbon
Storage
No change
No change
Increase in
soil carbon
storage
No change
No change
Waste Management GHGs
No emissions or sinks
No emissions or sinks
Compost machinery emissions,
and transportation emissions
Nonbiogenic CO2, N2O
emissions, avoided utility
emissions, and transportation
emissions.
Methane emissions, long-term
carbon storage in landfill,
avoided utility emissions,
transportation emissions, and
landfill machinery emissions.
1.6 SUMMARY ANALYSIS OF THE LIFE CYCLE STAGES
The following sections of this chapter explain the life cycle diagram presented in Exhibit 1-1, and
outline the GHG emissions and carbon sinks at each stage of the product life cycle. These GHG emissions
and carbon sinks are described in detail, and quantified for each material, in chapters 2 through 7.
20
DRAFT - March 1997
-------
GHG Emissions and Carbon Sinks Associated With Raw Materials Acquisition and
Manufacturing
The top left of Exhibit 1-1 shows inputs for raw materials acquisition. These are virgin inputs used
to make various materials including ore used to make metal products, trees used to make paper products,
and petroleum or natural gas used to make plastic products. Fuel energy used to obtain or extract these
material inputs is also shown.
The inputs used in manufacturing are: (1) energy, and (2) either virgin raw materials or recycled
materials. In the exhibit these inputs are identified with arrows that point to the box labeled
"Manufacturing."
The GHG emissions associated with raw materials acquisition and manufacturing are: (1) process
energy GHG emissions, (2) transportation energy GHG emissions, and (3) process non-energy GHG
emissions (for aluminum, steel, plastics, and office paper.) Each type of emission is described below.
Changes in carbon storage in forests are also associated with raw materials acquisition for paper products.
This analysis assumes no GHG impacts at the raw materials acquisition and manufacturing stages
for source reduction. Source reduction is assumed to entail more efficient use of a given material - for
example, "lightweighting," double-sided photocopying, or extension of a product's useful life. No other
material substitutions are assumed for source reduction; therefore, no corresponding increases in
production and disposal of other materials are analyzed that could result in GHG emissions.23
Process Energy GHG Emissions: Process energy GHG emissions consist of CO2 emissions from the
combustion of fuels used in raw materials acquisition and manufacturing. CO2 emissions from combustion
of biomass are not counted as GHG emissions (see box on Biogenic Sources of CO2 below).
The majority of process energy CO2 emissions are from combustion of fuels used directly, e.g., to
operate ore mining equipment or to fuel a blast furnace. Fuel is also needed to extract the oil or mine the
coal that is ultimately used to produce energy; thus CO2 emissions from this "pre-combustion energy" are
counted in this category as well. When electricity generated by combustion of fossil fuels is used in
manufacturing, the CO2 emissions from the fossil fuels are also counted.
To estimate process energy GHG emissions, we first obtained estimates of both the total amount of
process energy used per ton of product (measured in British thermal units or BTUs), and the fuel mix (e.g.,
diesel oil, natural gas, fuel oil). Next, we used emissions factors for each type of fuel to convert the amount
of each type of fuel used to the GHG emissions that are produced.
In the case of recycling, we found that making a material from recycled inputs generally requires
less process energy (and uses a different fuel mix) than making the material from virgin inputs.
Details of our methodology for estimating process energy GHG emissions is provided in Chapter 2.
23 Although material substitution is not considered here, it remains a high priority issue for future EPA
research.
DRAFT - March 1997 21
-------
CO2 Emissions from Biogenic Sources
The US and all other parties to the Framework Convention on Climate Change agreed to develop
inventories of GHGs for purposes of (1) developing mitigation strategies and (2) monitoring the progress of
those strategies. The Intergovernmental Panel on Climate Change (IPCC) developed a set of inventory
methods to be used as the international standard. (IPCC, IPCC Guidelines for National Greenhouse Gas
Inventories (three volumes), no date.) In selecting the methodologies used in this report to evaluate emissions
and sinks of GHGs, we attempted to be consistent with IPCC's guidance.
One of the elements of the IPCC guidance that deserves special mention is the approach used to
address CO2 emissions from biogenic sources. For many countries, the treatment of CO2 releases from
biogenic sources is most important when addressing releases from energy derived from biomass (e.g., burning
wood), but this element is also important when evaluating waste management emissions (for example, the
decomposition or combustion of grass clippings or paper). The carbon in paper and grass trimmings was
originally removed from the atmosphere by photosynthesis, and under natural conditions, it would eventually
cycle back to the atmosphere as CO2 due to degradation processes. The quantity of carbon that these natural
processes cycle through the earth's atmosphere, waters, soils, and biota is much greater than the quantity
added by anthropogenic GHG sources. But the focus of the Framework Convention on Climate Change is on
anthropogenic emissions - emissions resulting from human activities and subject to human control - because it
is these emissions that have the potential to alter the climate by disrupting the natural balances in carbon's
biogeochemical cycle, and altering the atmosphere's heat-trapping ability.
Thus, for processes with CO2 emissions, if (a) the emissions are from biogenic materials and (b) the
materials are grown on a sustainable basis, then those emissions are considered to simply close the loop in the
natural carbon cycle, that is they return to the atmosphere CO2 that was originally removed by photosynthesis.
In this case, the CO2 emissions are not counted. (For purposes of this analysis, biogenic materials are paper,
yard trimmings, and food scraps.) On the other hand, CO2 emissions from burning fossil fuels are counted
because these emissions would not enter the cycle were it not for human activity. Likewise, CH4 emissions
from landfills are counted - even though the source of carbon is primarily biogenic, CH4 would not be emitted
were it not for the human activity of landfilling the waste, which creates anaerobic conditions conducive to
CH4 formation.
Note that this approach does not distinguish between the timing of CO2 emissions, provided that they
occur in a reasonably short time scale relative to the speed of the processes that affect global climate change -
as long as the biogenic carbon would eventually be released as CO2, it does not matter whether it is released
virtually instantaneously (e.g., from combustion) or over a period of a few decades (e.g., decomposition on the
forest floor).
Transportation Energy GHG Emissions: Transportation energy GHG emissions consist of CO2
emissions from the combustion of fuels used to transport raw materials and intermediate products to the
final manufacturing or fabrication facility. We based our estimates of transportation energy GHG
emissions on: 1) the amounts of raw material inputs and intermediate products used in manufacturing one
ton of each material; 2) the average distance that each raw material input or intermediate product is
transported; and 3) the transportation modes and fuels used. For the amounts of fuel used, we used data on
the average fuel consumption per ton-mile for each mode of transportation. Then we used an emission
factor for each type of fuel to convert the amount of each type of fuel consumed, to the GHG emissions
produced.
22
DRAFT -- March 1997
-------
More detail on our methodology to estimate transportation energy GHG emissions is provided in
Chapter 2.
Process Non-Energy GHG Emissions: Some GHG emissions occur directly in the manufacture of
certain materials and are not associated with energy consumption. In this analysis, we refer to these
emissions as process non-energy emissions. For example, the production of steel or aluminum requires
lime (calcium oxide, or CaO), which is produced from limestone (calcium carbonate, or CaCO3); the
manufacture of lime results in CO2 emissions. Other process non-energy GHG emissions are associated
with production of plastics, office paper, and tissue paper. In some cases, process non-energy GHG
emissions are only associated with production using virgin inputs; in other cases, these emissions result
when either virgin or recycled inputs are used. These emissions are described in Chapter 2.
Carbon Sinks: The only carbon sink in the stages of raw materials acquisition and manufacturing is
the additional carbon sequestration in trees associated with source reduction or recycling of paper products.
Our methodology for estimating forest carbon sequestration is described in Chapter 3.
GHG Emissions and Carbon Sinks Associated With Waste Management
As shown at the bottom of Exhibit 1-1, there are, depending on the material, up to four waste
management options once a material is manufactured: recycling, composting, combustion, and landfilling.
This section describes the GHG emissions and carbon sinks associated with these four waste management
options.
In this analysis, source reduction is measured by the amount of material that would otherwise be
produced but is not being produced because of a program promoting source reduction. Thus, with source
reduction there are no emissions from MSW management.
Recycling: When a material is recycled, it is used in place of virgin inputs in the manufacturing
process. Thus, the only GHG emission consequences are those from manufacturing a material from
recycled rather than virgin inputs (including transportation GHGs and avoided GHGs from raw materials
acquisition); there are no GHG emissions at the MSW management stage. (If the product made from the
recycled material is later composted, combusted, or landfilled, the GHG emissions at that point would be
attributed to the product that was made from the recycled material.) Chapter 4 details GHG emissions from
recycling.
Most of the materials considered in this analysis are modeled as being recycled in a "closed loop"
(e.g., newspapers are recycled into new newspapers). However, office paper and corrugated boxes are
modeled as being recycled in an "open loop" (i.e., they are recycled into more than one product):
Office paper is modeled as being recycled into either office paper or tissue paper; and
Corrugated boxes are modeled as being recycled into either corrugated boxes or folding
boxes.
By developing GHG estimates for the manufacture of all four of these products, we were able to estimate
the GHG implications of "open loop" recycling of office paper and corrugated boxes. We recognize that
other materials are recycled in open loop processes, but due to limited resources, we could not analyze all
open loop processes.
DRAFT - March 1997
23
-------
Composting: When organic materials are composted, most of their organic mass quickly
decomposes to CO2. The materials that may be composted (e.g., leaves, brush, grass, food waste,
newspapers) are all originally produced by trees or other plants. As described in the text box above, the
CO2 emitted from these materials during composting is biogenic CO2, and thus is not counted in GHG
emissions.
There is some potential for the composting of yard trimmings to result in production of more humic
material (natural organic polymers, which degrade at a slow rate) than is produced when yard trimmings
are left to decompose in the yard. This process may act to enhance long-term carbon storage in soils to
which compost is applied.
Although composting may result in some production of methane (due to anaerobic decomposition in
the center of the compost pile) compost researchers concluded that the methane is almost always oxidized
to CO2 before it escapes from the compost pile.
Because the CO2 emissions from composting are biogenic, and there are generally no methane
emissions, the only GHG emissions from composting result from transportation of compostable materials
to composting facilities, and mechanical turning of the compost piles. Carbon cycling in compost
operations is discussed in Chapter 5.
Combustion: When waste is combusted, two GHGs are emitted: CO2 and N2O. Non-biogenic CO2
emitted during combustion (i.e., CO2 from plastics) is counted toward the GHG emissions associated with
combustion, but biogenic CO2 is not. Because most waste combustors produce electricity that substitutes
for utility-generated electricity, the net GHG emissions are calculated by subtracting the utility GHG
emissions avoided from the gross GHG emissions. GHG emissions from combustion are described in
Chapter 6.
Landfilling: When organic matter is landfilled, some of this matter decomposes anaerobically and
releases methane, a potent GHG. Some of the organic matter never decomposes at all; instead it becomes
sequestered carbon. (Landfilling of metals and plastics does not result in either methane emissions or
carbon sequestration).
At some landfills, virtually all of the methane produced is released to the atmosphere. The gross
GHG emissions from these landfills consist of the methane emissions. At other landfills, methane is
captured for flaring or combustion with energy recovery (i.e., electricity production). Most of the captured
methane is converted to CO2, which is not counted as a GHG because it is biogenic. With combustion of
methane for energy recovery, credit is given for the electric utility GHG emissions avoided. Regardless of
the fate of methane, credit is given for the landfill carbon sequestration associated with landfilling of some
organic materials. GHG emissions and carbon sinks from landfilling are described in Chapter 7.
24
DRAFT -- March 1997
-------
2. RAW MATERIALS ACQUISITION AND MANUFACTURING
To estimate the GHG emissions and sinks for the full life cycle of MSW materials, we needed to
estimate the GHG emissions associated with raw materials acquisition and manufacturing. This chapter
describes how we estimated these emissions for eight materials: newspaper, office paper, corrugated boxes,
aluminum cans, steel cans, and three types of plastic (LDPE, HDPE, and PET).
In manufacturing, substantial amounts of energy are used in the acquisition of raw materials and in
the manufacturing process itself. In most processes, the majority of this energy comes from fossil fuels.
Combustion of fossil fuels results in emissions of CO2, a greenhouse gas, and trace amounts of other GHGs
that are not included in the analysis. In addition, manufacturing of some materials also results in GHG
emissions that are not associated with energy consumption. Section 2.1 addresses energy-related CO2
emissions, and section 2.2 covers non-energy GHGs.
2.1 GHG EMISSIONS FROM ENERGY USE IN RAW MATERIALS ACQUISITION AND
MANUFACTURING
To begin our analysis, we estimated the GHG emissions from fossil fuel combustion for both (1)
raw materials acquisition and manufacturing (referred to here as "process energy"), and (2) transportation
(referred to as "transportation energy").
In this analysis, process energy GHG emissions consist primarily of COj.24 The majority of CO2
emissions are from combustion of fuels used directly, e.g., to operate mining equipment or to fuel a blast
furnace. Because fuel is also needed for "pre-combustion" activities (such as oil exploration and extraction,
coal mining and beneficiation, and natural gas production), CO2 emissions from "pre-combustion"
activities are also counted in this category. When electricity is used in manufacturing, the CO2 emissions
from the fuels burned to produce the electricity are also counted. In general, making a material from
recycled inputs requires less process energy than making the material from virgin inputs.
Transportation energy GHG emissions consist of CO2 emissions from combustion of fuels used to
transport raw materials and intermediate products to the final manufacturing or fabrication facility. For
transportation of recycled inputs, this analysis considers transportation (1) from the curbside to the
materials recovery facility (MRF),25 (2) from the MRF to a broker, and (3) from a broker to the plant or
mill where the recycled inputs are used. The transportation values for recycled inputs also generally
include the energy used to process the inputs at a MRF.26 Transportation of finished manufactured goods to
24 1
* Note, however, that CO2 emissions from combustion of biomass are not counted as GHG emissions (as
described in Chapter 1). For example, paper manufacturing uses biomass as a fuel.
25 A MRF processes recovered materials from the municipal solid waste (MSW) stream. Some MRFs take
mixed MSW and separate recyclable materials. Other MRFs accept only source-separated recyclable materials.
MRFs may crush, shred, or bale recyclable materials to make them ready for the scrap materials market,
26 The one exception is that data provided by Franklin Associates, Ltd. do not include the energy used in a
MRF to sort paper products.
DRAFT - March 1997 25
-------
consumers is not included in the analysis. We did not consider the global warming impacts of
transportation emissions of nitrogen oxides (NOX); such emissions contribute indirectly to climate change.27
This omission would tend to slightly understate the GHG impacts from transportation.
We also considered the methane emissions associated with producing, processing, and transporting
coal, oil, and natural gas. Methane is emitted during the various stages of fossil fuel production because
methane is trapped within coal and oil deposits, and because natural gas consists largely of methane.
We developed separate estimates for GHG emissions from process and transportation energy for
virgin inputs and recycled inputs, generating a total of four separate GHG emissions estimates for each
material: (1) process energy with virgin inputs, (2) process energy with recycled inputs, (3) transportation
energy with virgin inputs, and (4) transportation energy with recycled inputs.
Methodology
We developed GHG emission estimates for each material based on two sets of data: (1) the amount
of each type of fuel used to make one ton of the material, and (2) the "carbon coefficient" for each fuel (a
factor that translates the energy value of fuel combusted into the mass of GHGs emitted).
Our methodology in using these two sets of data to estimate process and transportation energy GHG
emissions is best illustrated by an example. To estimate process energy GHG emissions from the
production of one ton of newspapers from virgin inputs, we multiplied the amount of each type of fuel
used (as measured in million British thermal units, or BTUs) times the carbon coefficient for that type of
fuel (as measured in metric tons of carbon equivalent, or MTCE, per million BTUs). Each of these
multiplications yielded an estimate, for one of the fuels used to make newspaper, of the amount of GHGs
emitted (in MTCE) from the combustion of that fuel when one ton of newspaper is made. The total process
energy GHGs from making one ton of newspaper is simply the sum of the GHG estimates across the
different fuels used. To estimate the GHG emissions when electricity is used, we used the national average
mix of fuels used to make electricity.
We estimated GHGs from the energy used to transport raw materials for making one ton of a given
product (e.g., newspapers) in the same way: the amount of each fuel used was multiplied by its carbon
coefficient, and the resulting values for each of the fuels were summed.
To count "pre-combustion" energy, we scaled up the amount of each fuel combusted during
manufacture by the amount of energy needed to produce that fuel. In this approach, we used the
simplifying assumption that when oil is produced, oil is used as the energy source in oil production, while
natural gas is used for natural gas production, etc.
We developed GHG estimates for raw materials acquisition and manufacturing for each of the eight
manufactured materials of the ten materials considered in this report. We also developed GHG estimates
27 Because the Intergovernmental Panel on Climate Change (IPCC) has not established a method for
estimating the global warming implications of emissions of nitrogen oxides, we have not attempted such an
estimation.
26
DRAFT -- March 1997
-------
for tissue paper and folding boxes to enable us to estimate the GHG implications of increased recycling of
office paper and corrugated boxes, respectively, in an "open loop." Thus, the exhibits in this chapter show
data not only for the eight materials of interest, but also for tissue paper and folding boxes. For steel cans,
we developed GHG estimates for virgin production using the basic oxygen furnace process, and for
recycled production using the electric arc furnace process.28
For the first set of data that we needed (the amounts of each type of fuel used for process and
transportation energy), we obtained two independent sets of estimates from two consulting firms that have
expertise in lifecycle analysis, including process and transportation energy analysis: Franklin Associates
Ltd. (FAL), and the Tellus Institute (Tellus). For the second set of data (carbon coefficients), we used data
from the Energy Information Administration of the US Department of Energy29 for all fuels except diesel
fuel and electricity; for the latter fuels we used data from the American Council for an Energy-Efficient
Economy.30 The carbon coefficient for electricity was based on the weighted average carbon coefficients
for all fuels used to generate electricity in the US.31
Because the carbon coefficients from these sources accounted for only the CO2 emissions from
combustion of each type of fuel, we added to these carbon coefficients (1) the average amount of methane
emitted during the production, processing, and transportation of fossil fuels, and (2) the average CO2
emissions from oil production, due to the flaring of natural gas. To estimate these GHG emissions
associated with fossil fuel production, we used data from EPA, the US Department of Energy, and the
Intergovernmental Panel on Climate Change. We calculated the average GHG emissions associated with
US production of coal, oil, and natural gas. The resulting estimates for GHG emissions from fossil fuel
28 Note that when recovered steel cans are used as inputs to an electric arc furnace, the resulting steel is not
suited for milling to the thinness of steel sheet needed for use in making new steel cans. Thus, a more precise
approach would have been to model recovery of steel cans as an open loop process, in which recovered steel cans
are made into some other steel product. By modeling recovery of steel cans as a closed loop process, we implicitly
assumed that each ton of steel produced from recovered steel cans in an electric arc furnace displaces a ton of steel
produced from virgin inputs in a basic oxygen furnace; we believe this is a reasonable assumption. (For the
fabrication energy required to make steel cans from steel, we used the values for fabrication of steel cans from steel
produced in a basic oxygen furnace.)
29 Energy Information Administration, U.S. Department of Energy, Draft Emissions of Greenhouse Gases
in the United States 1989-1994, DOE/EIA-0573-annual (Washington, D.C.: U.S. Department of Energy), in press
1995, cited in U.S. Environmental Protection Agency, Inventory of U.S. Greenhouse Gas Emissions and Sinks:
1990-1994 (Washington, D.C.: U.S. EPA), November 1995, pp. A-8 to A-13.
30 R. Neal Elliott, Carbon Reduction Potential from Recycling in Primary Materials Manufacturing"
(Berkeley, CA: American Council for an Energy-Efficient Economy), February 8,1994, p. 14.
31 FAL and Tellus reported the BTU value for electricity in terms of the BTUs of fuel combusted to
generate the electricity used at the factory, rather than the (much lower) BTU value of the electricity that is
delivered to the factory. Thus, FAL and Tellus had already accounted for the efficiency of converting fuels to
electricity, and the losses in transmission and distribution of electricity; and we did not need to account for these
factors in the carbon coefficient for electricity.
DRAFT - March 1997
27
-------
production were 1.07 kilograms of carbon equivalent per million BTUs (kg C/million BTU) for coal, 0.23
kg C/million BTU for oil, and 0.82 kg C/million BTU for natural gas.32
The carbon coefficients that reflect both CO2 and methane emissions are provided in Exhibit 2-1 (all
exhibits are provided at the end of this chapter).
The process and transportation GHG values are shown in summary form in Exhibit 2-2. For each
product and each type of input (virgin or recycled), we summed the estimates for process and
transportation GHG emissions based on the FAL data, and then repeated the summation using the Tellus
data. Both sets of summed estimates are listed in Exhibit 2-2 in columns "b" (for virgin inputs) and "c"
(for recycled inputs). Although these estimates do not represent minimum or maximum values, we believe
that they do portray the variability in actual industry values for each material.
We also estimated the energy-related GHG emissions from manufacturing each material from the
current mix of virgin and recycled inputs. To do so, we averaged the two estimates for each material based
on FAL and Tellus data; the results are shown in column "e." (The remaining two columns of Exhibit 2-2
are discussed later in this chapter.)
The FAL and Tellus values for energy use are provided in Exhibits 2-3 through 2-10. Exhibits 2-3
through 2-6 present the FAL data - providing, in turn, the data used to estimate energy-related GHG
emissions for products manufactured from virgin inputs, and then the data for energy-related GHG
emissions for products manufactured from recycled inputs.33 Exhibits 2-7 through 2-10 present the Tellus
data, which are organized in the same way.34
For most materials, both FAL and Tellus provided data for fuels used in manufacturing processes
that use (1) 100 percent virgin inputs and (2) 100 percent recycled inputs.35 To estimate the types and
32 Memorandum from William Driscoll (ICF) to Michael Podolsky and Clare Lindsay (U.S. EPA),
"Fugitive Methane Emissions from Production of Coal, Natural Gas, and Oil," August 8, 1995, updated to use
global warming potential for methane of 24.5.
33 Note that when newspaper is made from virgin inputs, a substantial amount of biomass fuel (e.g., from
tree bark) is used; when newspaper is made from recycled inputs, no biomass fuel is used.
34 Note that in Exhibits 2-7 and 2-9, Tellus included values for the energy content of steam used in
manufacturing. We translated these steam energy values into fuel inputs as follows: (1) we assumed that the energy
content of the fuels combusted was converted into steam energy at a conversion efficiency of 85 percent; (2) for
paper products, made from virgin or recycled inputs, we used a fuel mix for steam of 40 percent oil, 33 percent
biomass, 17 percent natural gas, and 10 percent coal; and (3) for non-paper products made from virgin or recycled
inputs, we used a fuel mix for steam of 50 percent natural gas, 25 percent coal, and 25 percent oil (based on ICF
professional judgment).
35 The three exceptions were (1) the FAL data for corrugated boxes made from virgin inputs, for which
FAL provided data for manufacture from 90.2 percent virgin inputs and 9.8 percent recycled inputs, (2) the FAL
data for steel cans made from virgin inputs, for which FAL provided data for manufacture from 80 percent virgin
inputs and 20 percent recycled inputs, and (3) the Tellus data for steel cans made from virgin inputs, for which
Tellus provided data for manufacture from 90 percent virgin inputs and 10 percent recycled inputs. We extrapolated
from these data (and the corresponding values for production using 100 percent recycled inputs) to obtain estimates
of the energy inputs for manufacturing these materials from 100 percent virgin inputs.
28
DRAFT - March 1997
-------
amounts of fuels used for process and transportation energy, FAL and Tellus relied on published data
(such as engineering handbooks and published production data), and on personal contacts with industry
experts. FAL and Tellus counted all energy, no matter where it was used. For example, much aluminum
produced in the US is made from bauxite that is mined and processed into alumina in other countries. The
energy required for overseas bauxite mining and processing is counted in the analysis. In addition, it does
not matter where recycled inputs are made into remanufactured products. For example, if office paper that
is recovered in the US is remanufactured into paper products in Asia, the energy savings from
remanufacture using recycled rather than virgin inputs are counted.
Neither the FAL nor the Tellus transportation data reflect transportation of the finished
manufactured product to the retailer and consumer. This omission is only important in estimating the GHG
reductions associated with source reduction. It is not relevant in analyzing GHG implications of recycling
compared to other post-consumer management options, because the amount of transportation energy from
the factory to the consumer is about the same whether the product is manufactured from virgin inputs or
recycled inputs. Even for the source reduction analysis, we expect that the transportation energy from
factory to consumer would represent a very small fraction of the total process and transportation energy.
After FAL and Tellus had developed their initial estimates of process energy intensity and fuel mix,
we reviewed and verified the data by analyzing significant discrepancies between the estimates provided
by the two firms. Where discrepancies were found, we reviewed the most critical assumptions and data
elements that each firm used, and identified circumstances where it would be appropriate for one firm to
revise its assumptions or update its data sources.12 The effect of this process was to arrive at estimates by
the two firms that were closer to each other and, we expect, that more accurately reflect the energy used in
raw materials acquisition and manufacturing of the materials considered. Nevertheless, we recognize that
different manufacturers making the same product use somewhat different processes with different energy
requirements and fuel mixes, and that there are limited data on the extent to which various processes are
used. Thus, our goal was to estimate as accurately as possible the national average GHG emissions for the
manufacture of each material from virgin and recycled inputs.
In order to make the best use of all available data, for each material we averaged the FAL and Tellus
final estimates of GHG emissions for manufacturing the material from virgin inputs, and then did the same
for recycled inputs. These averaged values are used in all of the computations displayed in the executive
summary and in Chapter 8, which present overall results of the analysis.
Complete documentation of the FAL and Tellus data on the types and amounts of fuels used for
process and transportation energy, including data sources, is provided in the Appendix to this report.
12 For example, some of the data issues that we reviewed and decided on were (1) the fuel mix to assume
for electricity used to manufacture aluminum (the national average fuel mix for generating electricity was used,
because electricity generated from all types of fuel is sold as a single commodity through interconnected regional
grids), (2) whether to include the "pre-combustion" energy for fossil fuels, i.e., the energy required to extract,
refine, and deliver the fuels (pre-combustion energy was counted), (3) whether to use data for use of recovered
materials in "closed loop" or "open loop" processes (we used "closed loop" data except for office paper and
corrugated boxes), and (4) what loss rates should be used (we averaged the FAL and Tellus loss rates).
DRAFT - March 1997 29
-------
2.2 NON-ENERGY GHG EMISSIONS FROM MANUFACTURING AND RAW MATERIALS
ACQUISITION
We also accounted for three additional sources of GHGs in manufacturing processes that are not
related to energy use:
When limestone (calcium carbonate, or CaCO3) is converted to lime (calcium oxide, or
CaO), CO2 is emitted. Significant quantities of lime are used in the production of steel,
aluminum, and, to a much lesser extent, office paper.
Methane emissions from natural gas pipelines and processing of natural gas are associated
with the manufacture of plastic products.
Perfluorocarbons (CF4 and C2F6) are emitted during aluminum smelting.
In most cases, process non-energy GHG emissions are only associated with production using virgin
inputs. In the case of steel, however, these emissions result when either virgin or recycled inputs are used
(because lime is used in the production of steel from recycled as well as virgin inputs).
The process non-energy GHGs for each material are shown in the last column of Exhibits 2-3 and 2-
5 (for manufacture from virgin inputs and recycled inputs, respectively), and are repeated in column "f' of
Exhibit 2-2. Our source for all these data, except the perfluorocarbon emissions, is an appendix to a report
prepared for the EPA Office of Policy, Planning, and Evaluation.13 Our source for the perfluorocarbon
emissions is a memorandum prepared by ICF.14
23 RESULTS
Our estimates of the total GHG emissions from raw material acquisition and manufacturing for each
material are shown in Exhibit 2-2, column "g." To obtain these estimates, we summed the energy-related
GHG emissions (column "e") and the non-energy GHG emissions (column "f').
The process and transportation GHG values that were developed as described earlier in this chapter
are shown in the second to last columns of Exhibits 2-3 and 2-5, and in the last columns of Exhibits 2-4
and 2-6 through 2-10 (the last columns of Exhibits 2-3 and 2-5 show the process non-energy GHG
emissions, as noted above).
Because we had two independent sets of data on the amounts of each type of fuel used in making
each product, we were able to develop both range estimates and point estimates of the energy-related GHG
values for manufacturing each material from virgin or recycled inputs, and from the current mix of virgin
and recycled inputs. Li this report, for purposes of analyzing the GHG emissions associated with the
13 Memorandum from William Driscoll, Randy Freed, and Sarah Stafford (ICF) to Brett Van Akkeren (U.S.
EPA), "Detailed Analysis of Greenhouse Gas Emissions Reductions from Increased Recycling and Source
Reduction of Municipal Solid Waste," July 29, 1994, p. 48 of the Appendix prepared by Franklin Associates, Ltd.,
dated July 14,1994.
14 Memorandum from William Driscoll, Doug Keinath, and Randy Freed (ICF) to Eugene Lee and Clare
Lindsay (U.S. EPA), "Perfluorocarbon Emissions from Aluminum Smelting," March 27, 1996.
30
DRAFT - March 1997
-------
manufacturing stage of the product lifecycle, we are using the values in column "g" for total manufacturing
GHG emissions (i.e., averages of point estimates). Depending on the inputs being considered, the
appropriate value for total GHG emissions is used (i.e., the value for manufacture from virgin inputs,
recycled inputs, or the current mix of virgin and recycled inputs).
2.4 LIMITATIONS
There are numerous limitations to the analysis of the GHG emissions associated with raw materials
acquisition and manufacturing, as described below.
The approach used in this analysis provides values for the average GHG emission rates per ton of
material produced, not the marginal emission rates per incremental ton produced. In some cases, the
marginal emission rates may be significantly different. For example, reducing production of plastic
products from virgin inputs may not result in a proportional decrease in methane emissions from natural
gas pipelines and natural gas processing. Natural gas pipeline methane emissions are determined by the
operating pressure in natural gas pipelines, and the number and size of leaks in the pipeline. Consequently,
the amount of natural gas consumed at one end of the pipeline (e.g., to make plastic) does not affect the
level of pipeline methane emissions in a direct, linear way.15 As another example, long-term reductions in
electricity demand could selectively reduce demand for specific fuels, rather than reducing demand for all
fuels in proportion to their representation in the current average fuel mix. This analysis estimates average
carbon conversion rates largely because the marginal rates are much more difficult to estimate.
Nevertheless, we believe the average values provide a reasonable approximation of the GHG emissions.
In addition, the analysis assumes that the GHG emissions from manufacturing a given product
change in a linear fashion as the percentage of recycled inputs moves from 0 percent to 100 percent. In
other words, the analysis assumes that both the energy intensity and the fuel mix change in linear paths
over this range. However, it could be that GHG emissions from manufacturing move in a non-linear path,
(e.g., some form of step function) when the percentage of recycled inputs changes, due to capacity limits in
manufacturing or due to the economics of manufacturing processes.
The transportation energy required for the final stage of transportation (to the consumer) was not
considered. Consequently, some carbon emissions reductions for "lightweighted" products for these
transportation stages were not considered; these savings are likely to be negligible.
Finally, this static analysis does not consider potential future changes in energy usage per unit of
output. Reductions in energy inputs, due to efficiency improvements, could occur in either virgin input
processes or recycled input processes. Efficiency improvements will directly result in carbon emissions
reductions, and may change the amount of carbon reductions possible through increased recycling or
source reduction.
15 Bob Lott, Gas Research Institute, personal communication with William Driscoll, ICF Incorporated, June
30,1995.
DRAFT -March 1997 31
-------
Exhibit 2-1
Carbon Coefficients
For Selected Fuels
Fuel Type
Gasoline
LPG
Distillate Fuel
Residual Fuel
Diesel
Oil/Lubricants
Steam (non-paper products)
Steam (paper products)
National Average Fuel Mix for Electricity
Coal Used for Electricity
Coal Used by Industry (Non-Coking Coal)
Natural Gas
Other (Petroleum Coke)
kgCOz-Cfrom
Combustion Per
Million BTUs
19.43
17.02
19.95
21.49
20.80
20.24
18.70
13.12
16.24
25.71
25.61
14.47
27.85
kg COrCfrom Fugitive
Methane Emissions
Per Million BTUs
0.23
0.23
0.23
0.23
0.23
0.23
0.73
0.34
0.68
1.07
1.07
0.82
0.23
kg COrC
Emitted Per Million
BTUs Consumed
19.66
17.25
20.18
21.72
21.03
20.47
19.43
13.46
16.92
26.78
26.68
15.29
28.08
-------
Exhibit 2-2
Greenhouse Gas Emissions from the Manufacture of Selected Materials
(Metric Tons of Carbon Equivalent (MTCE) per Ton of Product)
(a)
Tvpe of Product
Newspaper
Office Paper
Tissue Paper
Corrugated Boxes
Folding Boxes
HOPE
LDPE
PET
(b)
Virgin Input Combined
Process and Transportation
Energy GHG Emissions
(MTCE Per Ton of Product
Made With Viraln Incuts)
FAL
Est
0.54
0.57
0.67
0.28
0.42
4.30
0.79
0.52
0.63
0.98
Tellus
Est
0.56
0.53
0.51
0.48
0.51
3.73
0.98
0.79
1.06
1.30
(e)
Recycled Input Combined
Process and Transportation
Energy GHG Emissions
(MTCE Per Ton of Product
Made With Virgin Inputs)
FAU
Est
0.39
0.50
0.50
0.34
0.38
0.69
0,28
0.25
0.23
0.41
Tellus
Est.
0.38
0.42
0.37
0.54
0.56
0.76
0.31
0.32
0.43
0.50
(d)
Percent Recycled
Inputs in the Current
Mix of Virgin and
Recycled Inputs
FAL
Est
37%
27%
36%
54%
39%
1%
8%
28%
Tellus
Est
33%
29%
43%
53%
46%
1%
8%
26%
(e)
Current Mix Combined
Process and Transportation
Energy GHG Emissions
(MTCE Per Ton of Product
Made with the Current Mix of
Virgin and Recycled Inputs
FAL
Est
0.49
0.55
0.67
0.30
0.42
2.35
0.60
0.51
0.60
0.82
Tellus
Est
0.50
0.50
0.51
0.51
0.51
2.17
0.67
0.79
1.01
1.0E
(0
Process
Non-Energy GHG
Emissions (MTCE Per
Ton of Product
Virgin
Inputs
0.00
0.01
0.01
0.00
0.00
1.49
0.24
0.07
0.07
0.04
Recycled
Inputs
0.00
0.00
0.00
0.00
0.00
0.00
0.24
0.00
0.00
0.00
Current
Mix
0.00
0.01
0.01
0.00
0.00
0.69
0.24
0.07
0.06
0.03
Average Combined
Process and Transportation
Energy and Process
Non-Energy GHG Emissions
(MTCE Per Ton of Product)
Virgin
Inputs
0.56
0.60
0.38
0.47
5.51
1.12
0.72
0.91
1.18
Recycled
Inputs
0.46
0.43
0.44
0.47
0.72
0.53
0.29
0.33
0.46
current
Mix
0.53
0.60
0.40
0.47
2.95
0.87
0.72
0.87
0.98
Explanatory notes: To estimate the GHG emissions from manufacturing, we first estimated the process and transportation GHG emissions
when 100 percent virgin inputs, or 100 percent recycled inputs, are used. For each product and each type of input (virgin or recycled), we first
summed the estimates for process and transportation GHG emissions based on the FAL data, and then repeated the summation using the Tellus data.
These summed estimates are shown above in columns "b" (for virgin inputs) and "c" (for recycled inputs). Two summed estimates are shown for
each material in each column: the "FAL estimate" and the "Tellus estimate."
Next we estimated the GHG emissions from manufacturing each material from the current mix of virgin and recycled inputs. We began with
estimates of the percentage of recycled inputs currently used in the manufacture of each material, as shown in column "d." We used these
percentages to develop a weighted average value for the GHG emissions associated with the manufacture of each material from the current mix of
virgin and recycled inputs. Specifically, we used the FAL estimate of the percentage of recycled inputs in the current mix, together with the FAL
estimates for GHG emissions from manufacture using virgin or recycled inputs, to develop FAL estimates of GHG emissions from manufacture
using the current mix of virgin and recycled inputs (labeled "FAL estimate" in column "e"). We repeated the process using the Tellus data (labeled
"Tellus estimate" in column "e").
-------
Explanatory notes for Exhibit 2-2 (continued): Column "f' shows estimates of the process non-energy GHG emissions from
manufacturing. First this column shows the process non-energy GHG emissions when virgin inputs are used. Then it shows the emissions when
recycled inputs are used (these values are simply copied from the final columns of Exhibits 2-3 and 2-5). Finally, column "f' shows the process non-
energy GHG emissions from manufacturing each product from the current mix of virgin and recycled inputs. The values for the current mix are the
weighted averages of the values for virgin and recycled inputs, based on the percentage of recycled inputs used in the current mix (as shown in
column "d").
The total GHG emissions from manufacturing are shown in column "g." This column shows total GHG emissions when a product is
manufactured from virgin or recycled inputs, or from the current mix of virgin and recycled inputs. To obtain these values, we first developed two
estimates of the GHG emissions for each material and each set of inputs. One estimate is based on FAL data, and the other is based on Tellus data
(these estimates included both energy-related GHG emissions and process non-energy GHG emissions). The values in column "g" are the averages
of the estimates based on FAL and Tellus data.
-------
Exhibit 2-3 (Franklin Data)
Amount of Carbon Produced Per Ton of Product Manufactured from Virgin Inputs
Process GHGs Only
Type of Product
Newspaper
Office Paper
Tissue Paper
Corrugated Boxes
Folding Boxes
Aluminum Cans
Steel Cans
HOPE
LDPE
Process Energy
(Million BTUs Per
Ton of Product)
33.96
54.80
52.09
30.01
40.12
243.53
31.58
30.71
37.68
50.51
Average Fuel Mix (in Percent)
Gasoline
0.00
1.99
2.29
0.00
2.79
0.00
0.21
0.10
0.08
0.05
LPG
0.06
0.00
0.00
0.00
0.00
0.01
0.00
0.03
0.03
0.05
Distillate Fuel
0.08
0.01
3.35
0.01
1.44
0.21
5.06
0.23
0.19
5.88
Residual Fuel
0.49
4.34
13.19
1.62
5.88
1.17
0.35
0.72
0.58
15.56
Biomass
6.53
50.07
40.88
56.06
47.87
0.00
0.00
0.00
0.00
0.00
Diesel
0.82
0.00
0.00
1.21
0.00
5.81
0.00
0.00
0.00
0.00
Electricity
57.54
24.75
18.90
19.67
18.22
78.41
21.02
42.46
51.11
51.66
Coal
1.07
9.78
11.95
8.75
10.29
1.47
53.90
0.00
0.00
6.14
Natural Gas
33.41
9.06
9.44
12.68
13.52
12.91
19.45
56.46
48.01
20.67
Total
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
Process Energy
Carbon Emissions
(MTCETon of Product)
0.52
0.62
0.25
0.40
4.18
0.70
0.49
0.61
0.92
Process Non-Energy
Carbon Emissions
(MTCEHon of Product)
0.01
0.01
0.00
0.00
1.49
0.24
0.07
0.07
0.04
-------
Exhibit 2-4 (FranWin Data)
Amount of Carbon Produced Per Ton of Product Manufactured from Virgin Inputs
Transportation GHGs Only
Type of Product
Newspaper
Office Paper
Tissue Paper
Corrugated Boxes
Folding Boxes
Aluminum Cans
Steel Cans
HDFE
LDFE
PET
Transportation Energy
(Million BTU* Per
Ton of Product)
0.77
2.46
2.46
1.43
1.01
5.73
4.60
1.15
1.15
3.27
Average Fuel Mix (In Percent)
Basel
98.59
99.43
99.43
99.79
99.19
37.53
98.24
54.50
54.50
79.65
Residual Oil
1.14
0.43
0.43
0.18
0.59
62.07
1.76
19.32
19.32
16.63
Natural Gas
0.17
0.11
0.11
0.02
0.20
0.00
0.00
24.66
24.66
2.42
Bectrlclty
0.10
0.03
0.03
0.01
0.02
0.40
0.00
1.52
1.52
1.31
Total
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
Transportation Biorgy Carbon Emissions
(Metric Tons of Carbon
Equivalent Per Ton of Product)
0.02
0.05
0.05
0.03
0.02
0.12
0.10
0.02
0.02
0.07
-------
Exhibit 2-5 (Franklin Data)
Greenhouse Gas Emissions Per Ton of Product Manufactured from Recycled Inputs
Process GHGs Only
Newspaper
Office Paper
Tissue Paper
Corrugated Boxes
Folding Boxes
Aluminum Cans
Steel Cans
HOPE
LDPE
Process Energy
(Million BTUs Per
23.01
26.46
26.46
15.95
18.90
40.34
11.78
12.68
11.43
0.00
0.00
0.00
0.00
0.00
0.00
0.01
0.00
0.00
Average Fuel Mix (in Percent
0.22
0.00
0.00
0.13
0.00
0.00
0.17
0.21
0.23
OlF
14.29
14.29
0.01
0.00
0.00
0.07
0.00
0.00
0.05
13.26
13.26
1.29
3.22
3.10
0.03
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.66
0.00
0.00
0.00
0.00
0.00
0.00
59.65
48.64
48.64
44.81
36.23
39.96
77.28
99.79
99.77
99.88
0.95
0.00
0.00
30.08
22.45
0.00
0.65
0.00
0.00
0.00
23.81
23.81
23.00
38.10
56.94
21.80
0.00
0.00
Total
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
Process Energy
Carbon Emissions
(MICE Per Ton of Product)
0.47
0.47
0.31
0.35
0.65
0.20
0.21
0.19
Carbon Emissions
(MTCE Per Ton of Product)
0.00
0.00
0.00
0.00
0.00
0.24
0.00
0.00
-------
Exhibit 2-6 (Franklin Data)
Anount of Carbon Produced Per Ton of Product Manufactured from Recycled Inputs
Transportation GHGs Only
Type of Product
Nawspaper
Office Paper
Tissue Paper
Corrugated Boxes
Folding Boxes
Aluminum Cans
Steel Cans
HOPE
LDFE
PET
Transportation Energy
(Million BTUS Per
Ton of Product)
0.75
1.61
1.61
1.23
1.29
1.65
4.03
1.74
1.74
1.74
Diesel
98.67
100.00
100.00
99.90
99.92
100.00
99.99
100.00
100.00
100.00
Average Fuel Mix (In Percent)
Residual Oi
1.08
0.00
0.00
0.10
0.08
0.00
0.01
0.00
0.00
0.00
Natural Gas
0.15
0.00
0.00
0.00
0,00
0.00
0.00
0.00
0.00
0.00
Electricity
0.10
0.00
0.00
0.00
0.00
0,00
0.00
0.00
0.00
0.00
Total
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
Transportation Energy Carbon Emissions
(Metric Tons of Carbon
Equivalent Per Ton of Product)
0.02
003
v.vo
n AO
u.uo
003
U.Wi.>
ona
W.UO
0.03
0.08
0.04
0.04
0.04
-------
Exhibit 2-7 (Tellus Data)
Amount of Carbon Produced Per Ton of Product Manufactured from Virgin Inputs
Process GHGs Only
Type of Product
Newspaper
Office Paper
Tissue Paper
Corrugated Boxes
Folding Boxes
Aluminum Cans
Steel Cans
HOPE
LDPE
PET
Process Biergy
(Million BTUs Per
Ton of Product)
34.11
35.18
33.22
32.07
34.05
216.24
42.10
37.29
51.78
62.51
Average Mix of Biergy Sources
Gasoline
0.46
0.89
0.94
0.86
0.81
0.00
0.03
0.00
0.00
0.00
Diesel
0.35
0.71
0.75
0.70
0.66
0.00
0.36
8.10
6.91
5.61
Oil
0.27
5.00
5.29
4.90
4.61
1.93
2.35
0.00
0.00
0.00
Steam
28.45
77.00
74.17
82.58
80.82
1.08
6.15
1.69
5.03
27.37
Bectricity
70.47
16.41
18.84
10.95
13.09
72.01
34.66
23.09
31.21
34.99
Coal
0.00
0.00
0.00
0.00
0.00
1.25
0.33
0.00
0.00
0.00
Natural Gas
0.00
0.00
0.00
0.00
0.00
23.68
5.71
42.27
35.81
10.89
Mher Fuels
0.00
0.00
0.00
0.00
0.00
0.05
50.41
24.85
21.03
21.14
Total
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
Process Biergy Carbon Bnlssions
(Metric Tons of Carbon
Equivalent Per Ton of Product)
0.54
0.51
0.48
0.46
0.49
3.62
0.96
0.72
0.99
1.25
-------
Exhibit 2-8 (Tellus Data)
Amount of Carbon Produced Per Ton of Product Manufactured from Virgin Inputs
Transportation GHGs Only
Type of Product
Newspaper
Office Paper
Tissue Paper
Corrugated Boxes
Folding Boxes
AlurrinumCans
Steel Cans
HOPE
LOPE
PET
Transportation Energy
(Million BTUs Per
ton of Product)
0.58
1.21
1.21
1.08
1.08
5.29
0.91
3.72
3.83
2.48
Average Fust Mix (in Percent)
Diesel
100.00
100.00
100.00
100.00
100.00
100.00
100.00
53.25
53.19
57.44
Natural Gas
0.00
0.00
0.00
0.00
0.00
0.00
0.00
46.75
46.81
42.56
Total
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
Transportation Energy Carbon Emissions
(Metric Tons of Carbon
Equivalent Per Ton of Product)
0.01
0.03
0.03
0.02
0.02
0.11
0.02
0.07
0.07
0.05
-------
Exhibit 2-9 (Tellus Data)
Amount of Carbon Produced Per Ton of Product Manufactured from Recycled Inputs
Process GHGs Only
Type of Product
Newspaper
Office Paper
Tissue Paper
Corrugated Boxes
Folding Boxes
AlurrinumCans
Steel Cans
HOPE
LDPE
PET
Process Biergy
(Million BTUs Per
Ton of Product)
18.52
20.80
0.94
27.31
29.23
46.04
17.01
17.85
23.29
27.84
Average Mix of Biergy Sources
Gasoline
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Diesel
0.26
0.23
0.24
0.17
0.16
0.10
0.35
0.27
0.20
0.17
Oil
0.00
0.00
0.00
0.00
0.00
2.57
0.00
0.00
0.00
0.00
Steam
41.88
62.85
58.66
69.47
67.11
0.00
0.00
0.00
0.00
0.00
Bectricity
57.86
36.92
41.10
30.36
32.72
35.51
97.92
99.73
99.80
99.83
Coal
0.00
0.00
0.00
0.00
0.00
0.00
0.48
0.00
0.00
0.00
Natural Gas
0.00
0.00
0.00
0.00
0.00
61.82
1.25
0.00
0.00
0.00
Other Fuels
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Total
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
Process Biergy Carbon Bnissions
(Metric Tons of Carbon
Equivalent Per Ton of Product)
0.33
0.38
0.37
0.51
0.54
0.74
0.29
0.30
0.39
0.47
-------
Exhibit 2-10 (Tellus Data)
Amount of Carbon Produced Per Ton of Product Manufactured from Recycled Inputs
Transportation GHGs Only
Typo of Product
Newspaper
Office Paper
Tissue Paper
Corrugated Boxes
Folding Boxes
Alum'numCans
Steel Cans
HOPE
LDPE
PET
Transportation Energy
(Million BTUs Per
Ton of Product)
2.13
1.87
0.00
1.33
0.83
0.90
0.82
0.83
1.56
1.56
Average Rial Mix (In Percent)
Diesel
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
Natural Gas
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0,00
0.00
Total
100.00
100.00
100.00
100.00
100,00
100.00
100.00
100.00
100.00
100.00
Transportation Enorgy Carbon Bnisslons
(Metric Tons of Carbon
Equivalent Per Ton of Product)
0.04
0.04
0.00
0.03
0.02
0.02
0.02
0.02
0.03
0.03
-------
3. FOREST CARBON SEQUESTRATION
This chapter presents estimates of the forest carbon sequestration associated with recovering or
source reducing newspaper, office paper, and corrugated cardboard.
One of the large-scale processes that influences the cycling of carbon is the uptake or release of
carbon from forests. When trees are cleared for agriculture or other activities, carbon is released (generally
in the form of CO2). On the other hand, when forests are planted and allowed to continue growing, they
absorb atmospheric CO2 and store it in the form of cellulose and other materials. When the rate of uptake
exceeds the rate of release, carbon is said to be sequestered. In the US, uptake by forests has exceeded
release since about 1977, primarily due to forest management activities and the reforestation of previously
cleared areas. This net storage of carbon in forests represents a large and important process - EPA
estimates that the annual net CO2 flux (i.e., the excess of uptake minus release) in US forests was about
125 million metric tons of carbon equivalent in 1990-92,16 offsetting about 9 percent of US energy-related
CO2 emissions.
When paper products are source reduced or recycled, trees that would otherwise be harvested are
left standing. Li the short term, this results in a larger amount of carbon remaining sequestered, because the
standing trees continue to store carbon, whereas paper production and use tends to release carbon.17 In the
long term, some of the short-term benefits disappear as market forces result in less planting of new
managed forests than there would otherwise be, so that there is comparatively less forest acreage in trees
that are growing rapidly (and thus sequestering carbon rapidly).18-19'20
16 US EPA, Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-1994, EPA 230-R-96-006, p. 75.
Estimates are not available for more recent years because the last national forest inventory was completed in 1992.
17 Viewed from another perspective, when trees are harvested, the trees that are planted to replace them
store very little carbon in the early years of their growth.
18 Note also that when a ton of paper is source reduced or recycled, the trees that would otherwise have
been harvested to make new paper do not all remain unharvested. Instead, as the demand for trees falls with
increased source reduction and recycling, the price of trees also falls, and consequently some additional trees are
harvested for non-paper purposes (e.g., construction); the use of wood for these other purposes may or may not
result in carbon sequestration.
19 Note that some analysts project that in the long run, global demand for forest products will increase more
rapidly than global paper recovery rates, so that even with increased paper recovery, forested acreage is likely to
increase in the long run.
20 Some analysts have suggested that more efficient municipal waste combustors and increased paper
combustion rates, combined with more intensive tree planting, could result in reduced GHG emissions [Electric
Power Research Institute, "Paper Recycling: Impact on Electricity-Use, Electro-Technology Opportunities," Report
RP-3228-06 (1993), cited in Gaines, Linda L. and Frank Stodolsky, "Energy Implications of Recycling Packaging
Materials" (Argonne, IL: Argonne National Laboratory) 1994], but we did not analyze this issue.
DRAFT -- March 1997 43
-------
Considering the importance of forest carbon sequestration as a process affecting net US GHG
emissions, and given the fact that paper products are a large and growing portion of the forest products
market, we recognized that a thorough examination of forest carbon sequestration was warranted for this
study. Moreover, the complexity and long time frame of carbon storage in forests, coupled with the
importance of market dynamics that determine land use behavior, dictated the use of state-of-the-art
models to evaluate the effect of source reduction and recycling of paper products on forest carbon
sequestration. This chapter describes our method for applying models to estimate the effect of forest carbon
sequestration associated with paper recovery for recycling. We used the results from our analysis of paper
recovery to estimate the effect of source reduction on forest carbon sequestration.
We worked with the US Forest Service (USFS) to use a system of models of the US forest sector to
estimate the amount of forest carbon sequestration per incremental ton of newspaper, office paper, and
corrugated cardboard recovered for recycling. Because the models did not allow us to separately estimate
the forest carbon sequestration associated with recovery of each of the three types of paper, we obtained a
single estimate for the sequestration from recovering any type of paper.
Working with USFS, we used five linked
models of the forest sector to estimate the impacts
of increased recovery of paper products on forest
carbon sequestration. The first model projects the
decline in US pulpwood harvests when paper
recovery increases. The second and third models
use the outputs of the first model, together with
other inputs and assumptions, to estimate the
extent to which reduced pulpwood harvests due to
paper recovery result in lower US timber harvests
and increase timber inventories. The fourth and
fifth models use the outputs of the second and
third models, and estimate how the increased
timber inventories and decreased timber harvests
due to paper recovery translate into (1) increased
forest carbon sequestration and (2) changes in
carbon sequestration in wood-in-use carbon sinks
(e.g., wood used in home construction). We used
the USFS system of models because (1) they are
the most sophisticated models available in
modeling the species composition, inventory, and
growth of forests, and (2) these models had been
used previously to analyze climate change
mitigation options for the Climate Change Action
Plan.
In brief, we found that recovering one ton
of paper results in incremental forest carbon
sequestration of 0.73 metric tons of carbon
equivalent (MTCE). We converted this single estimate for recovering any type of paper into three separate
estimates for source reducing each of three different types of paper. We developed separate estimates for
Performance of the USFS Forest Models
Researchers have never formally
assessed the accuracy of the USFS models of
the forest sector. However, informal
assessments of the models' accuracy indicate
that the models are fairly reliable. For example,
Peter Ince of the Forest Service reports that
projections of pulpwood harvests made by the
NAPAP model have been within five percent
of actual harvests. In analyses that compare the
forest impacts of a policy scenario to those of a
baseline scenario (such as the analysis
described in this chapter), the USFS model
results are probably reasonably accurate. This
is because most of the uncertainty in the model
results is due to assumptions that apply to both
the baseline and policy scenarios - assumptions
about population growth, economic growth,
tree growth, and land use changes. Any error in
these assumptions would tend to bias the
results in the baseline and policy scenarios in
the same direction. Thus, when the outcomes
of the baseline and policy scenarios are
compared, errors in the assumptions tend to
cancel each other out.
44
DRAFT-March 1997
-------
source reduction based on the inputs displaced by source reduction - either 100 percent virgin inputs, or a
mix of virgin and recycled inputs (this mix is different for each type of paper).
If one assumes that source reduction displaces 100 percent virgin inputs, we estimated that source
reduction of any of the three types of paper results in forest carbon sequestration of 0.73 MTCE per ton -
the same as for paper recovery. On the other hand, if one assumes that source reduction displaces the mix
of virgin and recycled inputs currently used in manufacturing, source reduction of one ton of newspaper,
office paper, or corrugated cardboard results in forest carbon sequestration of, respectively, 0.48, 0.56, and
0.44 MTCE.
The remainder of this chapter is divided into seven parts. Section 3.1 provides an overview of the
linkages between the five models used in the analysis. Sections 3.2 through 3.5 describe the five models in
greater detail, and briefly discuss the inputs, assumptions, and outputs for each model. Section 3.6 presents
the results of the analysis, and Section 3.7 discusses the limitations of individual models, and of the
analysis as a whole.
3.1 MODELING FRAMEWORK
Working with USFS, we used five computer models, each representing a different aspect of the
forest sector, to track the effects of increased paper recovery on the forest sector, and to estimate forest
carbon sequestration due to paper recovery. The five models are linked in the sense that outputs from one
model were used as inputs to the next model. Exhibit 3-1 shows how the models are linked.
Our overall analysis proceeded as follows:
1) We developed two future recovery scenarios - a projected baseline paper recovery rate for the
year 2000 of 50 percent, and a hypothetical year 2000 paper recovery rate of 55 percent - as inputs to the
North American Pulp and Paper (NAPAP) model (the model is described in Section 3.2). A 50 percent
recovery rate was used for the baseline scenario because the paper industry projects a 50 percent recovery
rate by the year 2000 in the absence of further government policies to promote recycling. We used a 55
percent recovery rate for the high recovery scenario because (1) we considered this to be a plausible
recovery rate with additional government programs to promote recycling, and (2) this recovery rate
corresponds to EPA's goal of increasing recovery of MSW in the Climate Change Action Plan. We
assumed that EPA policies to promote source reduction and recycling would end in 2000, and that over the
next 15 years, the recovery rates under both scenarios would continue to rise and would converge in the
year 2016 at 57 percent. (We assumed convergence so that we could isolate the long-term carbon
sequestration effects that might result from increasing paper recovery in the near term.) The paper recovery
rates for both scenarios were then projected to rise slowly from 57 percent in 2016 to 61 percent in 2040.
This adjustment to the model incorporated our assumption that the current trend of increasing paper
recovery rates would continue into the future.
The NAPAP model was then run to model the pulpwood harvests from 1985 to 2040 that would be
associated with (1) the baseline paper recovery rate and (2) the high paper recovery rate.
2) The outputs from NAPAP for projected pulpwood harvests in the two scenarios were used as
inputs to the Timber Assessment Market Model (TAMM), which projects US timber harvests, and the
Aggregate Timberland Assessment System (ATLAS) model, which projects timber growth and changes in
the US forest inventory (where inventory is a function of both growth and harvests). The TAMM and
DRAFT -- March 1997
45
-------
Exhibit 3-1
USFS Models of the Forest Sector
Macro
Economic
Data
Forest
Inventory
Data
NAPAP
(pulp and paper)
AREA
MODELS
(forest acreage)
I
Timberland
area
Prices
TAMM
(timber harvests)
Inventory
Growth
Removals
Carbon
accounting
forest inventory,
forest growth
parameters
ATLAS
(forest growth)
Acres, inventory
growth, removals
FORCARB
(forest carbon)
All harvest
HARVCARB
(harvested carbon)
-Denotes 'hand' linkages between models/modelers requiring manipulation of data.
46
C63056
DRAFT - March 1997
-------
ATLAS models are described more fully in Section 3.3. The TAMM and ATLAS models were run, using
the NAPAP inputs, to generate estimates of US harvest levels and forest inventories for each year through
2040, for both the baseline and high recovery scenarios.
3) The outputs from TAMM and ATLAS for forest harvest levels and forest inventories in the two
scenarios were used as inputs to the Forest Carbon (FORCARB) model, described in Section 3.4, which
projects forest carbon sequestration. The FORCARB model produced, as outputs, estimates of US forest
carbon sequestration for each year through 2040, for both the baseline and high recovery scenarios.
4) FORCARB outputs were also used as inputs to the HARVCARB (Harvested Carbon) model,
which tracks the flow of carbon in wood products (see Section 3.5).
3.2 THE NORTH AMERICAN PULP AND PAPER MODEL (NAPAP)
The NAPAP model is a linear optimization model,21 which uses forecasts of the US economy (e.g.,
growth in population and in the economy) to estimate the quantity of hardwood and softwood trees
harvested for pulpwood in North America each year.22'23 The model predicts the quantity of pulpwood
harvested each year based on estimated demand and supply curves; the quantity harvested is the quantity at
which these curves intersect.
Inputs to the NAPAP Model
Major inputs to the NAPAP model are:
macroeconomic forecast data (e.g., estimates of US population growth, and growth in per-
capita gross domestic product),
21 A linear optimization model begins with a set of constraints (e.g., profits = revenues - costs; costs = labor
costs + equipment costs + administrative costs + overhead costs) and an objective function (e.g., maximize profits).
The model uses principles of matrix algebra to find the solution (e.g., the total level of output) at which the
objective function is optimized (e.g., profits are maximized).
22 The description of the NAPAP model in this section is drawn from a memorandum from Peter Ince, US
Department of Agriculture Forest Service, Forest Products Laboratory, to Michael Podolsky, US EPA, entitled
"Alternate Recycling Scenarios," dated September 15,1995, and a telephone conversation between Peter Ince and
William Driscoll of ICF Incorporated, November 22,1995.
23 A number of analyses have been conducted using results from the NAPAP models. These include: (1)
USDA Forest Service, RPA Assessment of the Forest and Rangeland Situation in the United States -1993 Update,
USDA Forest Service Forest Resource Report No. 27 (Washington, DC: USDA Forest Service) 1994,75 pp.; (2)
Haynes, Richard W., Darius M. Adams, and John R. Mills, 1995, The 1993 RPA Timber Assessment Update, USDA
Forest Service General Technical Report RM-GTR-259 (Fort Collins, CO: Rocky Mountain Forest and Range
Experiment Station) 1995, 66 pp.; (3) Ince, Peter J., 1995, What Won't Get Harvested Where and When: The Effects
of Increased Paper Recycling on Timber Harvest, Yale School of Forestry and Environmental Studies Program on
Solid Waste Policy, Working Paper #3 (New Haven, CT: Yale University) 75 pp.; and (4) Environmental Defense
Fund, Paper Task Force Recommendations for Purchasing and Using Environmentally Preferable Paper: Final
Report of the Paper Task Force (New York, NY: Environmental Defense Fund) 1995,245 pp.
DRAFT - March 1997
47
-------
24
paper manufacturing capacity as of a baseline year,
manufacturing costs (apart from wood, fiber, labor, and energy) for each different paper
manufacturing process, and
assumed levels of future harvests from public forests.
Equations and Assumptions Used in the NAPAP Model
The NAPAP model incorporates equations for the following mathematical functions:
estimated pulpwood supply functions (reflecting an increasing supply of pulpwood at
increasing market prices) for three regions in the US (west, south, and north) and two regions
in Canada,
estimated supply functions for four principal categories of recovered paper - newspapers,
corrugated boxes, mixed papers, and the aggregate of pulp substitutes and high-grade
deinking categories - in each supply region (the supply functions reflect an increasing supply
of recovered paper at increasing market prices),
an unlimited supply of labor and energy at the market price in each supply region,
a fixed-quantity supply function for pulpwood residues,
demand functions25 for all thirteen principal categories of paper and paperboard products
produced in North America26 (the demand functions reflect increasing demand due to
population growth and growth in the gross domestic product, and decreasing demand at
increasing market prices),
functions for changes in paper manufacturing capacity (including capacity for both virgin and
recycled inputs), based on the ratio of profitability of new capacity to the capital cost of new
capacity,27 and
The baseline year for paper manufacturing capacity is 1986. The model predicts how capacity for each
paper manufacturing process changes each year from 1986 onward. The model's predictions for paper
manufacturing capacity in 1995, based on the 1986 baseline as updated, were within five percent of actual 1995
paper manufacturing capacity. (Peter Ince, USFS, telecon with William Driscoll, ICF, October 18, 1996).
25 Separate demand functions are incorporated for US domestic demand, Canadian domestic demand, and
demand from various trading regions for exported paper products from the US and Canada.
26 These paper grades include newsprint, coated and uncoated free sheet, coated and uncoated groundwood,
linerboard, and corrugating medium.
27 The model assumes that when demand for paper increases, the investment in paper manufacturing
capacity that is needed to meet demand will be made in those types of capacity where the ratio of profitability to
capital cost is the highest.
48
DRAFT - March 1997
-------
the ratio of the amount of paper recovered to the amount actually used in manufacturing new
paper, after accounting for discards during processing and losses during manufacturing.
The major assumptions of the NAPAP model include basic assumptions of economic analysis - i.e.,
that markets are perfectly competitive and that paper manufacturers seek to maximize their profits. Because
owners of private forests may not always act to maximize their profits, NAPAP assumes that they will
continue historical patterns of economic behavior (which USFS has modeled through econometric
methods). In addition, the model assumes (1) particular levels of harvests from public forests, and (2)
specific future technology options.28 Finally, the NAPAP pulpwood supply functions are the same for both
the baseline and the high recycling scenario. In other words, the supply functions do not incorporate
market feedbacks to account for changes in the age structure of forests or the acreage of forested land (The
age structure of forests could change as increased paper recovery reduces tree harvests, so that on average
trees grow longer; forested acreage could change if higher paper recovery led to decreased demand for
pulpwood and lower pulpwood prices, leading some landowners to convert forested land to farmland or
ranchland).29
Modifications of NAPAP for the High Recovery Scenario
Three adjustments to the NAPAP model were needed to generate the baseline and high recovery
scenarios. First, the USFS replaced paper recovery rates that would ordinarily be estimated by the model
itself with the recovery rates specified for the two scenarios (e.g., for the year 2000, 50 percent in the
baseline scenario, and 55 percent in the high recovery scenario). The cumulative amounts of paper
recovered under the baseline and high recovery scenarios are shown in Exhibit 3-2.
A second adjustment of the model was needed to maintain consistency with the estimates of paper
production and consumption based on projections prepared by the American Forest and Paper Association
(AFPA). Franklin Associates, Ltd. (FAL), a consulting firm with expertise in analyzing the paper
manufacturing and paper recycling industries, developed estimates for the year 2000 that corresponded to
AFPA projections of a 50 percent paper recovery rate. To match the FAL estimates, the USFS revised the
NAPAP model by making an upward adjustment to the demand functions for paper products, which
resulted in increased projections of paper demand and increased estimates of the equilibrium quantity of
paper produced.30
28 The model assumes that certain technologies that existed in 1995 but were not yet commercialized (e.g.,
two newsprint processes with higher yields) would enter the commercial marketplace in the period from 1995-2000.
29 The NAPAP pulpwood supply functions incorporate projections of timber inventories over time from a
prior run of the linked TAMM and ATLAS models. Ideally, the NAPAP portion of this analysis would have used
two separate projections of timber inventories over time: one projection based on the baseline paper recovery
scenario, and another based on the high paper recovery scenario. NAPAP has recently been revised so that it may
now be run iteratively with TAMM and ATLAS; however, NAPAP did not have that capability at the time this
analysis was conducted.
30 Specifically, the USFS adjusted the NAPAP model by increasing the elasticity of demand for paper
products so that it reflected the historical relationship between (1) paper demand and (2) population and per capita
gross domestic product. "Elasticity of demand" is the extent to which a change in the price of a good will affect the
quantity of the good demanded, and is defined as the percentage change in quantity divided by the percentage
change in price that induced the change in quantity. For example, if the quantity demanded goes down by 2 percent
DRAFT -- March 1997 49
-------
Exhibit 3-2
Cumulative Paper Recovery
Under the Baseline and High Recovery Scenarios
(Million Short Tons)
Year
A. Baseline Scenario
B. High Recovery Scenario
C. Incremental Paper Recovery
Under the High Recovery
Scenario (B-A)
2000
536
556
20
2010
1143
1189
46
2020
1893
1975
81
2030
2795
2876
81
2040
3808
3890
81
4.000
3,500
3.000
Cumulative Paper Recovery
Under the Baseline and High Recovery Scenarios
[1.500
51.000
I BaseBno Scenario
I High Recovery Scenario
fiH^^H l^^^ns 9Sss
I I
2000
2010
2030
when the price goes up by one percent, the elasticity of demand is -2. (Specifically, this is the "own-price elasticity"
of demand - because it is measured with respect to the price of the good in question, as distinct from "cross-
elasticity" of demand - which would be measured with respect to the price of a different good.)
50
DRAFT - March 1997
-------
In the third adjustment, trade in forest products between the US and Canada was fixed in the model
at levels projected in recent USFS studies. The result was that any change in North American pulpwood
harvests due to increased US paper recovery would be shown in the NAPAP outputs as a change in US
pulpwood harvests. This adjustment allowed us to model the full forest carbon effects of increased paper
recovery in the US as if those effects occur entirely in the US.
The NAPAP model, thus modified, was run to obtain estimates of US pulpwood harvests, from
2000 to 2040, for the baseline and high recovery scenarios.
Outputs of the NAPAP Model
The principal outputs of the NAPAP model, for each of the two1 scenarios modeled, are annual US
pulpwood harvests from the present to the year 2040. These harvests are broken down into four categories
of pulpwood: (1) softwood roundwood, (2) softwood residues, (3) hardwood roundwood, and (4)
hardwood residues. The NAPAP estimates of pulpwood harvests for each scenario - for selected years
from 1995 to 2040 - are shown in Exhibit 3-3. As the exhibit shows, the NAPAP model projected that
higher paper recovery rates until the year 2016 would result in pulpwood harvests that would be
substantially below the baseline from 1995 to 2000 (because the recovered paper substitutes for pulp that
would otherwise be made from trees). From 2005 to 2010, the higher recovery scenario would result in
slightly higher pulpwood harvests than under the baseline,31 and from 2020 onward, pulpwood harvests
would be the same under the baseline and high recovery scenarios (because after 2016 the paper recovery
rates would be the same in both scenarios).32
3.3 THE TIMBER ASSESSMENT MARKET MODEL (TAMM) AND THE AGGREGATE
TEMBERLAND ASSESSMENT SYSTEM (ATLAS)
TAMM and ATLAS are spatial equilibrium models.33 TAMM models US timber harvests through
2040, and ATLAS models changes in US forest growth, and inventory of growing stock volume, through
2040.34 The two models are interrelated, because timber harvests depend in part on timber inventory, and
31 Pulpwood harvests are projected to be higher between 2005 and 2010 under the high recycling scenario
due to the modeled consequences of reduced pulpwood harvests before 2005. Because pulpwood harvests before
2005 are projected to be lower under the high recycling scenario, more pulpwood remains to be harvested in later
years The increasing supply of pulpwood ready for harvest reduces pulpwood prices, leading to modeled increases
in industry demand for non-paper uses. The increased industry demand results in slightly higher pulpwood harvests
after 2005.
32 Note that under the baseline scenario, pulpwood harvests are projected to decline between 2000 and
2005. This is because the increase in paper recycling during this period is projected to be greater than the increase in
paper consumption.
33 A spatial equilibrium model is an optimization model (see footnote 21) that accounts for costs of
transportation of products from producing regions to consuming regions.
34 The descriptions of the TAMM and ATLAS models are drawn from Richard W. Haynes et al, Alternative
Simulations of Forestry Scenarios Involving Carbon Sequestration Options: Investigation of Impacts on Regional
and National Timber Markets, US Department of Agriculture Forest Service, Pacific Northwest Station, August 5,
1993. Two articles which give a more detailed description of the TAMM model are (1) Adams, D.M. and R.W.
DRAFT -- March 1997
51
-------
Exhibit 3-3
U.S. Pulpwood Harvests as Predicted by the NAPAP Model
for Selected Years (Million Cubic Feet)
Year
Baseline Scenario
Hgh Paper Recovarv Scenario
1995
7,152
6,982
2000
7,230
6,858
2010
7,328
7,362
2020
7,808
7,808
2030
7,989
7,989
2040
8,173
8,173
U.S. Pulpwood Harvests as Predicted by the NAPAP Model
2030 2035 2040
Haynes, The 1980 Softwood Timber Assessment Market Model: Structure, Projections, and Policy Simulations
Forest Science Monograph No. 22 (Washington, DC: USDA Forest Service) 1980,62 pp., and (2) Adams, D.M. and
R.W. Haynes, A Spatial Equilibrium Model of US Forest Products Markets for Long-Range Projection and Policy
Analysis. In Andersson etal., eds., Systems Analysis in Forestry and Forest Industries, TIMS Studies in the
Management Sciences 21(1986)73-87. Two journal articles which describe analyses based on the TAMM model are
(1) Adams, D.M. and R.W. Haynes, Softwood Timber Supply and the Future of the Southern Forest Economy,
Southern Journal of Applied Forestry 15(1991):31-37, and (2) Adams, D.M and R.W. Haynes, 1991, Estimating the
Economic Impacts of Preserving Old-Growth on Public Lands in the Pacific Northwest, The Northwest
Environmental Journal 6(2):439-441.
52
DRAFT - March 1997
-------
timber inventory depends in part on prior harvest levels. (This interrelationship is shown graphically in
Exhibit 3-1 with arrows going in both directions between the two models.) To obtain consistency in the
projections of the two models, an iterative process is used. TAMM outputs for timber removals are used as
inputs to ATLAS, and the resulting ATLAS outputs for forest growth and inventory are used as inputs to
TAMM. This cycle is continued until the difference in projections between one cycle and the next has been
reduced to an acceptably small amount. Note that, to reduce the costs of modeling in this analysis, no hand
linkages were made to transfer price estimates from TAMM back to the Area Models (see Exhibit 3-1), nor
to transfer timberland area estimates from the Area Models back to ATLAS. Implicitly, the forested area
was modeled as being unaffected by increased paper recovery rates.
TAMM's estimates of timber harvests are based on four factors: (1) estimated demand for timber
products, such as softwood lumber, based on projected macroeconomic data (e.g., growth in population
and in the economy), (2) estimates of pulpwood harvests from the NAPAP model, (3) estimates of
fuelwood harvests (held constant at recent levels), and (4) estimates of annual forest growth from ATLAS.
The ATLAS estimates of forest growth and inventory are based on (1) the previous year's
inventory, (2) timber harvests from TAMM, and (3) estimated forest growth parameters.
Inputs to the TAMM Model
Major inputs to the TAMM model are:35
US pulpwood harvests, from the NAPAP model, .
US fuelwood harvests, from a fuelwood model,
assumed levels of future timber harvests from public forests, from USFS harvest plans,
US net imports of forest products, from a trade model,
changes in US forested acreage over time, from a prior run of forest area models,36
growth in forest inventory, from the ATLAS model,
macroeconomic forecast data, e.g., on US housing starts, housing repairs, and remodeling,
and
installed capacity as of 1990 for producing timber products, such as lumber or plywood, from
harvested trees.
35 Inputs to the TAMM model are documented in Haynes, R.W., An Analysis of the Timber Situation in the
United States: 1989-2040, Gen. Tech. Rep. RM-199 (Ft. Collins, Colorado: USDA Forest Service, Rocky Mountain
Forest and Range Experiment Station) 1990,286 pp.
36 Although in the NAPAP portion of this analysis, timber inventories over time were not affected by the
different paper recovery rates in the two different scenarios analyzed, in the TAMM and ATLAS models, timber
inventories were estimated independently for the two different scenarios.
DRAFT - March 1997
53
-------
Equations and Assumptions Used in the TAMM Model
The TAMM model incorporates equations for the following:
estimated timber product supply functions (reflecting an increasing supply of timber products
at increasing market prices) for eight regions in the US, and
estimated demand functions for US demand for all major uses of lumber and plywood
(reflecting decreasing demand for such products at increasing market prices).
Also, changes in supply capacity for timber products are predicted by the model, based on anticipated
changes in relative regional profitability or rate of return from capital investment.37
The major assumptions of the TAMM model include38
general assumptions of competitive markets, increasing demand for wood products with
increasing economic activity, profit maximization by owners of lumber and plywood mills,
and continued historical patterns of economic behavior by owners of forest land (these
behavior patterns may not be strictly profit maximizing), and
specific assumptions regarding particular levels of public harvests, and projected changes in
technology.
In addition, TAMM and ATLAS assume (1) specified levels for net imports of softwood products, and (2)
no net imports of hardwood lumber.
. Inputs to the ATLAS Model
Major inputs to the ATLAS model, for each simulation year, are:
forest inventory at the beginning of the previous period, from a prior ATLAS model ran,
forest removals during the previous period, from the TAMM model,
changes in forest acreage, from a prior run of a modified version of the Southern Area
Model, and
Specifically, TAMM uses an assumption that changes in capital investment are a function of past
changes in output (i.e., that manufacturers' expectations about the profitability of capital investment are based on
past changes in output).
38 Assumptions of the TAMM model are documented in the following two reports: (1) Haynes, R.W., An
Analysis of the Timber Situation in the United States: 1989-2040, Gen. Tech. Rep. RM-199. (Fort Collins,
Colorado: USDA Forest Service, Rocky Mountain Forest and Range Experiment Station) 1990, 286 pp.; and (2)
Haynes, R.W., D.M. Adams, and J.R. Mills, The 1993 RPA Timber Assessment Update, Gen. Tech. Rep. RM-GTR-
259 (Fort Collins, Colorado: USDA Forest Service, Rocky Mountain Forest and Range Experiment Station) 1995,
66pp.
54
DRAFT -- March 1997
-------
state-by-state data on the number of forested acres, and the volume of timber per forested
acre (shown as "Forest Inventory Data" in Exhibit 3-1).
Equations and Assumptions of the ATLAS Model
The ATLAS model incorporates equations that allow the model to simulate shifts in forest
management intensities and consequent changes in yields. Projected shifts in forest management intensities
are based upon (1) the modeled prices of forest products, (2) the costs of various management practices,
and (3) the timber yields associated with each management practice.
The only major assumption in the ATLAS model is that owners of private forests manage their
forests at the level of intensity indicated by recent average forest planting rates. Otherwise, the model is
very simple, relying on a basic mathematical proposition that forest inventory in any period equals forest
inventory in the previous period, plus growth, minus harvests.
Outputs of the TAMM/ATLAS Models
The outputs of the linked TAMM and ATLAS models are projections, through 2040, of US
inventories of forest growing stock volumes (i.e., the volume of trees growing in forests), annual US
sawtimber harvests, and forest growth.
We used the TAMM/ATLAS data on forest growing stock inventories as inputs to FORCARB.
Exhibit 3-4 shows the growing stock inventories of privately owned forest lands in the US as projected by the
TAMM/ATLAS models. As the exhibit shows, forest growing stock inventories range from one to two billion
cubic feet higher under the high recovery scenario than under the baseline scenario for the entire simulation
period.
3.4 THE FOREST CARBON MODEL (FORCARB)
The Forest Carbon Model (FORCARB) projects US forest carbon storage (including soil, forest floor,
and understory carbon) each year through 2040, based on outputs from the TAMM/ATLAS linked models.39
Inputs to the FORCARB Model
The major inputs to the FORCARB model are the following:
forest growing stock inventories - by tree species, age, and region - from the linked
TAMM/ATLAS models, and
39 The description of the FORCARB model here is drawn from Birdsey, Richard A., and Linda S. Heath,
Carbon Sequestration Impacts of Alternative Forestry Scenarios -Draft (Radnor, PA: US Department of
Agriculture Forest Service, Global Change Research Program), April 1993, pp. 47-51. A number of studies
analyzing forest issues using the FORCARB and HARVCARB models have been published in journal articles.
Among these are three which also explain the FORCARB and HARVCARB models. These three articles are: (1)
Plantinga, A.J. and R.A. Birdsey, 1993, "Carbon fluxes resulting from US private timberland management,"
Climatic Change 23:37-53; (2) Heath, L.S. and R.A. Birdsey, 1993, "Carbon trends of productive temperate forests
of the coterminous United States," Water, Air, and Soil Pollution 70:279-293; and (3) Heath, L.S. and R.A. Birdsey,
1993, "Impacts of alternative forest management policies on carbon sequestration on US timberlands," World
Resource Review 5:171-179.
DRAFT - March 1997
55
-------
Exhibit 3-4
Growing Stock Inventories of
Privately Owned Forest Lands in the US
As Projected by the TAMM/ATLAS Models
(Billion Cubic Feet)
Year
Baseline Scenario
Hgh Paper Recovery Scenario
1995
478
478
2000
488
489
2010
515
517
2020
532
534
2030
541
544
2040
545
548
Growing Stock Inventories of Privately Ovmed Forest Lands in the US as
Projected by the TAMM/ATLAS Models
2010 2015
56
DRAFT - March 1997
-------
the percentage carbon composition for different species of trees, as grown in different forest
regions.
Assumptions of the FORCARB Model
The Forest Service tracks information in TAMM/ATLAS in terms of growing stock volume, i.e.,
the merchantable portion of trees. Tree volume is larger than growing stock volume, due to additional
volume in non-merchantable portions of the tree such as roots and branches. The FORCARB model uses
the simplifying assumption that tree volume is a constant multiple of growing stock volume. Carbon in the
tree volume in the US forest industry is then estimated based on the percentage carbon content of different
species of trees.
When a tree is harvested, FORCARB no longer counts the carbon remaining in the non-
merchantable portion of the tree (e.g., tree roots) following harvest. In other words, FORCARB uses the
simplifying modeling assumption that the carbon in the non-merchantable portion of the tree is
immediately lost from storage, i.e., converted to CO2 emissions.
Outputs of the FORCARB Model
The FORCARB model produces as outputs estimates of total US forest carbon inventories, and
estimates of sawtimber and pulpwood harvests, for each year through 2040. The amount of forest carbon
sequestration in a given year equals the increase in forest carbon inventories during that year. (If forest
carbon inventories decrease, that implies negative forest carbon sequestration.)
Exhibit 3-5 shows the projected carbon inventories of US forests, as predicted by the FORCARB
model, for the baseline and high paper recovery scenarios. The forest carbon inventories on which these
annual changes were based counted carbon in trees and understory (e.g., small trees), but not carbon in the
soil and forest floor. These carbon stocks were not included because of the high level of uncertainty in
estimating and modeling their carbon content.
Exhibit 3-6 shows the change in US forest carbon inventories, expressed as an annual average for
decades from 2000 to 2040. Inventories increase more quickly under the high recycling scenario than
under the baseline recycling scenario, through the decade ending 2010. After 2010, the rate of increase in
forest carbon inventories is essentially the same for both scenarios. This is because the paper recovery rate
is modeled as converging in 2016 to the same rate in both scenarios.
3.5 THE HARVESTED CARBON MODEL (HARVCARB)
The Harvested Carbon Model (HARVCARB) can be thought of as a spreadsheet model that
projects the disposition of harvested wood across four different potential fates, for 50 years into the
future.40 The spreadsheet would consist of estimates of the percentage of each of four types of wood that
will be found in each of four potential fates at ten-year intervals. The four potential fates are (1) products (a
"wood-in-use" sink), (2) landfills, (3) combustion for energy, and (4) aerobic decomposition. There is
some change in the fate of a wood product over time: wood products that are in use in the early years are
40 This description of HARVCARB is based on a description in Birdsey, Richard A. and Linda S. Heath, op
cit, pp. 50-51.
DRAFT - March 1997 57
-------
Exhibit 3-5
US Forest Carbon Inventory, Trees and Understory
As Predicted by the FORCARB Model
(Million Metric Tons of Carbon)
Year
A. Baseline Scenario
B. High Paper Recovery Scenario
C. Incremental Carbon Stored
Under the High Paper Recovery
Scenario (B-A)
2000
8,641
8,665
24
2010
9,076
9,118
42
2020
9 322
9,364
42
2030
9,442
9,480
38
2040
9,497
9,537
40
9600
9400
9200
9000
8800
8600
8400
8200
8000
US Forest Carbon Inventory
As Predicted by the FORCARB Model
(Trees and Understory)
2000
2040
58
DRAFT - March 1997
-------
Exhibit 3-6
Average Annual Change
In US Forest Carbon Inventories
As Predicted by the FORCARB Model
(Million Metric Tons of Carbon)
Time Period
A. Baseline Scenario
B. Ugh Paper Recovery
Scenario
C. Incremental Annual Forest
Carbon Sequestration in the
Hgh Paper Recovery
Scenario [B-A1
Decade
Ending
2000
45.5
47.9
2.4
Decade
Ending
2010
43.5
45.3
1.8
Decade
Ending
2020
24.6
24.6
0
Decade
Ending
2030
12
11.7
-0.3
Decade
Ending
2040
5.5
5.7
0.2
Average Annual Change in US Forest Carbon
Inventories
d Baseline Scenario
High Paper Recovery Scenario
2000 2010 2020 2030
Docad* Ending
2040
DRAFT - March 1997
59
-------
likely to be landfilled or combusted in later years. The four different types of wood considered in the
model are softwood and hardwood pulpwood, and softwood and hardwood sawtimber. The model has
separate fate estimates for three regions of the US - west, south, and north.
The HARVCARB model was developed in three steps. First, output from the TAMM model was
used to estimate the mix of wood products made from harvested wood. Second, data on the fate of various
wood products over a 50-year period were collected from research studies. Third, the wood products were
allocated to their various fates over time to allow simulation of the fates of all wood products.41
We combined the average annual sawtimber and pulpwood harvest estimates from FORCARB, with
the fate estimates in the HARVCARB spreadsheet, to obtain estimates of the amount of carbon from
harvested wood that would be found in each of the four potential fates for 50 years into the future.
Inputs to the HARVCARB Model
For this analysis the USFS used, as the only input to the HARVCARB model, the annual sawtimber
and pulpwood harvests (from the FORCARB model).
Assumptions of the HARVCARB Model
The HARVCARB model assumes that the disposition patterns for the four types of wood over a 50-
year period do not change (e.g., it does not assume any change in the proportion of wood combusted for
energy).
Outputs of the HARVCARB Model
In this analysis, HARVCARB provided outputs for the amount of carbon (1) retained in wood-in-
use sinks, (2) landfilled, (3) combusted for energy, and (4) aerobically decomposed, for each year from
1995 to 2040. Because other parts of our analysis address landfills and combustion, and aerobic
decomposition has no GHG effects, we used only the estimates of the amount of carbon retained in wood-
in-use sinks (a form of carbon sequestration). We included this amount in our estimate of total "forest
carbon," even though this carbon is stored in locations outside of forests.
Exhibit 3-7 shows the wood-in-use sinks for the baseline and high recovery scenarios from 1990 to
2040, as predicted by the HARVCARB model. As shown in the exhibit, the wood-in-use sinks are slightly
less under the high recovery scenario than under the baseline scenario. The HARVCARB model predicts
this result because under the high recovery scenario, tree harvests are reduced; thus, under the fixed
proportions of the fates of wood assumed in HARVCARB, less wood is available for each of the fates for
wood products, including wood-in-use sinks. As noted above, HARVCARB uses fixed proportions for the
disposition of harvested wood (e.g., paper, housing, and furniture). With increased paper recovery, wood
prices would be expected to decline (due to reduced demand), and more wood would probably be used for
housing and furniture. Since HARVCARB does not account for any change in the price of wood, and its
impacts on wood-in-use sinks, the values in Exhibit 3-7 probably underestimate the amount of carbon in
wood-in-use sinks under the high recovery scenario.
41 Linda Heath, USDA Forest Service, Pacific Northwest Research Station, telephone conversation with
William Driscoll, ICF Incorporated, November 29, 1995.
60 DRAFT -- March 1997
-------
Exhibit 3-7
US Cumulative (Since 1990) Wood-in-Use Sinks
as Predicted by the HARVCARB Model
(Million Metric Tons of Carbon)
Year
A. Baseline Scenario
B. High Paper Recovery Scenario
C. Change in Carbon Storage in
Wood-in-Use Sinks [B-A]
2000
733
726
-7
2010
1,216
1,208
-8
2020
1,634
1,630
-4
2030
2,028
2,026
-2
2040
2,381
2,379
-2
Change in Carbon Storage in Wood-in-Use Sinks
Between the Baseline and High Paper Recovery Scenarios,
as Predicted by the HARVCARB Model
(Million Metric Tons of Carbon)
2000
2010
2020
2030
2040
DRAFT -March 1997
61
-------
3.6 RESULTS
As noted in the introduction to this chapter, we first obtained estimates of the forest carbon
sequestration42 from recovery of paper, and then used those estimates to develop estimates of the forest
carbon sequestration from source reduction of paper.
We estimated the forest carbon sequestration per ton of paper recovered at various points in the
future by dividing the cumulative difference in forest carbon between the high recovery and baseline
scenarios by the cumulative difference in the amount of paper recovered between the two scenarios. To
estimate the forest carbon sequestration in each scenario, we summed the forest carbon sequestration
estimates generated by the FORCARB model and the wood-in-use sink estimates generated by the
HARVCARB model.
The USFS projected forest carbon inventories under the baseline and high recovery scenarios at
several points in time (i.e., 2000, 2010,2020, 2030, and 2040). The estimates of incremental forest carbon
sequestration per ton of paper recovered vary across time, as shown in Exhibit 3-8. Note that the estimates
of incremental forest carbon sequestration decline from 2000 to 2020, and then stabilize.
An important goal of this analysis is to develop "conversion factors" or point estimates that enable
policymakers and the public to quantify and compare the GHG impacts from managing specific waste
materials in specific ways. In the near term, these conversion factors will be used to estimate progress made
by the US in meeting its commitment to reduce GHG emissions by the year 2000. Therefore, the year 2000 is
a key endpoint for this analysis. We chose the forest carbon sequestration factor for the period ending in 2010
as the best approximation of the forest carbon benefits from increasing source reduction and recycling by the
year 2000. This value - 0.73 MTCE per short ton of paper recovered - falls between the higher value for 2000
and the lower values for later years in the simulation period. We selected this value to approximate the short-
term carbon sequestration benefits of source reduction and recycling because it balances the following: (1)
relatively high carbon sequestration benefits will be achievable by the year 2000; (2) actions taken to bring
about increases in source reduction and recycling by the year 2000 will have lingering effects beyond the year
2000, (3) forest carbon sequestration benefits drop somewhat over time; and (4) there is more uncertainty
associated with the long-term carbon sequestration effects and market response (because model predictions far
into the future are more uncertain than near-term predictions). In sum, we believe that the value for the year
2010 strikes the best balance in capturing the relatively higher short-term benefits of forest carbon
sequestration, and recognizing that these benefits decline over time.
As noted above, we did not consider in this analysis carbon sequestration in forest soils or the forest
floor. Had we also counted the incremental carbon sequestration in these carbon sinks, paper recovery
would have had a forest carbon sequestration value of 0.96 MTCE per ton of paper recovered.43
Next, we used the forest carbon sequestration estimate for paper recovery to develop estimates for
source reduction, as shown in Exhibit 3-9. We estimated source reduction values under two assumptions:
that source reduction displaces virgin inputs, and that it displaces the current mix of virgin and recycled
41 As noted earlier, the term forest carbon sequestration is intended to include both the carbon stored in
forests and the carbon stored in wood-in-use sinks.
43 Based on the cumulative rates through 2010.
62
DRAFT - March 1997
-------
Exhibit 3-8
Increased Forest Carbon Storage Per Ton of Paper Recovered
Cumulative Change
Between the Baseline
and High Paper Recovery
Scenarios for:
A. Forest Carbon Stocks* (million MICE)
B. Wood-in-Use Stocks (million MICE)
C. Incremental Carbon Stored (million MTCE) [A+B]
D. Incremental Paper Recovery (million short tons)
E. Incremental Carbon Sequestration (MTCE/ton) [C/D]
2000
24.0
-7.0
17.0
19.7
0.86
2010
41.9
-8.0
33.9
46.2
0.73
2020
42.2
-4.0
38.2
81.4
0.47
2030
39.7
-2.0
37.7
81.4
0.46
2040
41.9
-2.0
39.9
81.4
0.49
'Includes trees and understory; excludes soils and forest floor
Increased Forest Carbon Sequestration Per Ton of Paper Recovered
2000
2040
DRAFT -March 1997
63
-------
inputs.*14 We estimated that the forest carbon sequestration for source reduction, assuming displacement of
virgin inputs, is the same as the forest carbon sequestration for paper recovery. Although this approach for
estimating the effects of source reduction does not consider the loss rates associated with paper recovery,
we believe it is a reasonable first approximation. To estimate the forest carbon sequestration for source
reduction assuming displacement of the current mix of inputs, we used an additional factor, i.e., the
percentage of virgin inputs in the current mix of inputs. For this calculation (column "a" in Exhibit 3-9),
we account for the fact that displacement of recycled inputs does not have any impact on forest carbon
sequestration.
EXHIBIT 3-9
Forest Carbon Sequestration
Per Ton of Paper Recovered or Source Reduced
(a)
Material
Newspaper
Office Paper
Corrugated
Cardboard
(b)
Recovery
-Recovering
One
Incremental
Ton of Paper
(MTCE)
0.73
0.73
0.73
(c)
Source Reduction -
Assuming
Displacement of
One Ton of of
Paper Made from
Virgin Inputs (= b)
(MTCE)
0.73
0.73
0.73
(d)
Percent
Virgin Inputs
in the
Current Mix
of Inputs
65%
76%
60%
(e)
Source Reduction
Assuming Displacement
of One Ton of Paper
Made from the Current
Mix of Virgin and
Recycled Inputs (=b*d)
(MTCE)
0.48
0.56
0.44
3.7 LIMITATIONS
Any analysis based on a complex system of models is subject to the limitations introduced by each
model in the system. The limitations of each component model derive from (1) the assumptions made in
44 Source reduction may displace 100 percent virgin inputs if the quantity of paper recovered does not
change with source reduction, and all recovered paper is used to make new paper. In that case, if the quantity of
paper manufactured is reduced through source reduction, all of the reduction in inputs would come from virgin
inputs. It is more likely, however, that source reduction reduces both virgin and recycled inputs. Economic theory
states that when a good (e.g., paper) is manufactured, alternative inputs (e.g., virgin and recycled materials) are used
in amounts such that the marginal cost of using either type of input is the same. Thus, when less paper is
manufactured due to source reduction, the use of both virgin and recycled inputs will decline in amounts such that
the marginal cost of using either input is again the same. Because we did not have data to estimate the marginal
reductions in virgin and recycled inputs when paper is source reduced, we used the current mix of virgin and
recycled inputs in paper manufacture as a proxy.
64
DRAFT - March 1997
-------
developing the model, (2) the input equations used in the model, and (3) the potential impact of factors not
included in the model. Because of these limitations, the actual behavior of markets for paper and other
forest products (and the actual choices made by owners of private forest lands) could differ from those
predicted by the system of forest models. We believe that most of these limitations would tend to bias
estimates under the baseline and high recycling scenarios in the same direction - so that the estimated
differences between the two scenarios should be relatively accurate. However, some limitations could
result in unequal bias in the estimates, leading to biased estimates of the differences.
This section first discusses limitations that could bias the estimates, because they are of greater
concern. Limitations that could bias both scenarios in the same direction are listed next. This section
concludes with a brief discussion of the uncertainties introduced by the choice of a time period over which
incremental forest carbon sequestration is estimated.
Limitations Expected to Bias the Results
Three limitations in the system of forest sector models could result in biased estimates of the
incremental forest carbon sequestration from increased paper recycling. They are as follows:
The modeling system does not account for any conversion of US forest land to farmland or
rangeland that might occur in response to lower prices for pulpwood due to higher paper
recycling rates. The NAPAP model did not account for potential changes in timber inventory
in the near term (due to lower harvests associated with higher paper recovery), nor potential
changes in forest acreage in the longer term (if higher paper recovery depresses pulpwood
prices enough to induce landowners to convert forested acreage to other uses). The TAMM
and ATLAS models likewise did not allow for long-term changes in forested acreage due to
increased paper recovery.
NAPAP does not account for any effects of lower pulpwood prices (due to higher paper
recycling rates) on net exports of US pulpwood to non-Canadian markets. Lower pulpwood
prices would be expected to result in increased exports, and possibly changes in foreign
timber inventories. This effect is expected to be insignificant, though, because US pulpwood
exports are currently less than one percent of US pulpwood production.
HARVCARB does not account for the reduced timber prices expected under a high paper
recycling scenario, and the resulting increase expected in the amount of wood used for
housing, furniture, and other wood-in-use sinks. If this increase were accounted for, the
system of models would show a somewhat higher carbon sequestration benefit associated
with recycling.
This analysis did not consider carbon sequestration in forest soils and forest floors, because
of the high level of uncertainty in projecting changes in carbon storage. Nonetheless,
projections of carbon storage in forest soils and floors under the baseline and high recycling
scenarios, as generated by the FORCARB model, suggest that incremental carbon storage
45 Telephone conversation between Richard Haynes, US Department of Agriculture Forest Service, Pacific
Northwest Research Station, and William Driscoll, ICF Incorporated, December 4, 1995.
DRAFT - March 1997
65
-------
under the high recycling scenario could be slightly higher than shown here, if storage in soils
and the forest floor were included.
Limitations Not Expected to Bias the Results
We expect that several limitations in the system of forest models would bias - to about the same
extent- the estimates of forest carbon sequestration in the baseline and high recycling scenarios - and thus
would not result in significant bias in the estimate of the difference in forest carbon sequestration between
the two scenarios. These limitations are as follows:
The macroeconomic forecasts used in the models (e.g., for population growth and growth in
per-capita gross domestic product) are simply forecasts, and may turn out to be inaccurate.
The historical supply and demand functions used in the models may change in the future. For
example, (1) demand for newspapers may drop sharply due to competition from electronic
news media, or (2) improved technologies or tree diseases not anticipated in the models may
significantly change the cost of producing forest products.
Future harvests from public forest lands may be different from those projected.
The Use of a Point Estimate for Forest Carbon Sequestration
As shown in Exhibit 3-8, estimates of forest carbon sequestration due to increased paper recycling
vary over time. As noted above, in choosing a single point estimate, we selected the time period that best
balances the competing criteria of (1) capturing the long-term forest carbon sequestration effects, and (2)
limiting the uncertainty inherent in projections made well into the future. The range of forest carbon
sequestration estimates over time, and the limitations of the analysis discussed above, indicate that there is
considerable uncertainty in the point estimate selected. Li comparison to the estimates of other types of
GHG emissions and sinks developed in other parts of this analysis, the magnitude of forest carbon
sequestration is relatively high; based on these forest carbon sequestration estimates, source reduction and
recycling of paper are found to have substantial net GHG reductions. Because paper products comprise the
largest share of municipal waste generation (and the largest volumes of waste managed through recycling,
landfilling, and combustion), it is important to bear in mind the uncertainty in the forest carbon
sequestration values when evaluating the results of this project.
66
DRAFT -- March 1997
-------
4. SOURCE REDUCTION AND RECYCLING
This chapter presents estimates of GHG emission reductions and carbon sequestration resulting
from source reduction and recycling of eight materials: newspaper, office paper, corrugated boxes,
aluminum cans, steel cans, and three types of plastic (LDPE, HDPE, and PET).
To estimate GHG emissions associated with source reduction and recycling (and other MSW
management options), we used a baseline scenario in which the material is not manufactured (as shown in
Exhibit 1-2). Based on this measurement convention, we estimated that source reduction results in no
GHG emissions for all materials. Moreover, source reduction of paper results forest carbon sequestration
(as discussed in Chapter 3), which is treated as a GHG reduction. (In this analysis, source reduction is
assumed to entail either material "lightweighting" or extension of a product's useful life. We assume no
substitution by another material or product, and thus we assume no offsetting GHG emissions from another
material or product. Thus the data should not be used directly for estimating GHG impacts of source
reduction that involves material substitution.46)
Manufacturing from recycled inputs generally requires less energy than manufacturing from virgin
inputs. Thus, manufacturing from recycled inputs generally results in lower GHG emissions than
manufacturing from virgin inputs. Our estimates of the GHG implications of recycling, which are
developed in this chapter, show that recycling reduces GHG emissions for each of the eight materials
studied.
4.1 GHG IMPLICATIONS OF SOURCE REDUCTION
When a material is source reduced (i.e., less of the material is made), the greenhouse gas emissions
associated with making the material and managing the post-consumer waste are avoided. In addition, when
paper products are source reduced, trees that would otherwise be harvested are left standing and continue
to grow, so that carbon remains sequestered in forests (as described in Chapter 3). The additional carbon
sequestered due to source reduction is counted in the same way as a reduction in GHG emissions.
As discussed above, under the measurement convention used in this analysis, source reduction has
(1) zero manufacturing GHG emissions, (2) positive forest carbon sequestration benefits for paper products
(as estimated in Chapter 3), and (3) zero waste management GHG emissions. Exhibit 4-1 presents the
GHG implications of source reduction. The values for forest carbon sequestration were copied from
Exhibit 3-8.
46 The GHG impacts of source reduction involving material substitution could be estimated based on (1) the
data provided in this report for the material that is source reduced, (2) the mass substitution rate for the material that
is substituted, and (3) data in this report for the material substituted. If source reduction involves substitution of a
product not analyzed in this report, one would also need to assume that the final fabrication energy per ton of
substitute product is similar to the final fabrication energy per ton of product analyzed in this report.
DRAFT -- March 1997 67
-------
Exhibit 4-1
Greenhouse Gas Emissions for Source Reduction
(MTCE/Ton of Material Source Reduced)
Material
Newspaper
Office Paper
Corrugated Cardboard
Aluminum Cans
Steel Cans
HOPE
LDPE
PET
GHG Emissions
from Raw Materials
Acquisition and
Manufacturing
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Change In Forest Carbon Storage
(Minus sign Indicates incremental carbon storage)
Source Reduction
Displaces Current
Mix of Virgin and
Recycled Inputs
0.48
-0.53
0,44
0.00
0.00
0.00
0.00
0.00
Source
Reduction
Displaces Virgin
Inputs
0.73
-0.73
-0.73
0.00
0.00
0.00
0.00
0.00
NetGHGs
Waste
Management
GHGs
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Source reduction
Displaces Current
Mix of Virgin and
Recycled Inputs
0.48
-0.53
0.44
0.00
0.00
0.00
0.00
0.00
Source
Reduction
Displaces Virgin
Inputs
-0.73
-0.73
0.73
0.00
0.00
0.00
0.00
0.00
-------
In order to compare source reduction to another solid waste management option, we added the GHG
savings from source reduction to the GHG emissions avoided by not using another solid waste
management option (e.g., landfilling). With this approach, we determined the overall difference in GHG
emissions between (1) source reducing one ton of material and (2) manufacturing and then managing
(post-consumer) one ton of the same material. Such comparisons are made in the executive summary
chapter and in Chapter 8 of this report. Overall, source reduction has lower GHG emissions than the other
waste management options.
4.2 GHG IMPLICATIONS OF RECYCLING
When a material is recycled, it is used in place of virgin inputs in the manufacturing process, rather
than being disposed of and managed as waste. As with source reduction of paper products, recycling of
paper products also results in forest carbon sequestration (as described in Chapter 3), which is counted in
the same way as a reduction in GHG emissions.
Although most of the materials considered are modeled as being recycled in a "closed loop" (e.g.,
newspapers are recycled into new newspapers), two of the products considered - office paper and
corrugated boxes - are modeled as being recycled in an "open loop" (i.e., they are recycled into more than
one product). Office paper is modeled as being recycled into either office paper or tissue paper, in
proportions of 45 percent and 55 percent, respectively. Corrugated boxes are modeled as being recycled
into either corrugated boxes (70 percent) or folding boxes (30 percent). By developing GHG estimates for
all four of these products we were able to estimate the GHG implications of "open loop" recycling of office
paper and corrugated boxes.47
To compare recycling to another solid waste management option such as landfilling, we compared
the total GHG emissions from manufacturing and then recycling, to the GHG emissions from
manufacturing and then landfilling. Specifically, we subtracted (1) the GHG emissions from
manufacturing, minus the avoided GHG emissions from remanufacture using recycled (rather than virgin)
inputs, from (2) the GHG emissions from manufacturing and then landfilling.48 Overall, recycling has
lower GHG emissions than all other waste management options except for source reduction.
When any material is recovered for recycling, some portion of the recovered materials are unsuitable
for use as recycled inputs (these materials are discarded either in the recovery stage or in the
remanufacturing stage). Consequently, less than one ton of material is generally made from one ton of
recovered inputs. These losses may be quantified as "loss rates." We obtained estimates of loss rates from
Franklin Associates, Ltd. and the Tellus Institute. For each material, we then averaged the estimated rates
from the two firms. The loss rates for each material are shown in Exhibit 4-2.
47
7 Note that this modeling approach does not fully reflect the prevalence and diversity of open loop
recycling. For example, (1) office paper and corrugated cardboard are recycled into a variety of manufactured
products, not just the two products we selected for each, and (2) additional materials are also recycled in an open
loop.
48 We assumed that recycling does not change demand for the products made from recycled materials.
Thus, we assumed that each incremental ton of recycled inputs would displace an equivalent amount of virgin
inputs. To estimate the avoided GHG emissions from remanufacture from recycled inputs, we compared the GHG
emissions from manufacturing a material from 100 percent virgin inputs, to the GHG emissions from manufacturing
the material from 100 percent recycled inputs.
DRAFT--March 1997 69
-------
Exhibit 4-2
Loss Rates For Recovered Materials
(a)
Material
Newspaper
Office Paper
Corrugated Cardboard
Aluminum Cans
Steel Cans
HOPE
LDPE
PET
(b)
Percent of
Recovered
Materials
Retained in the
Recovery Staqe
90
88
92
95
98
87
87
87
(0)
Tons of
Product Made
Per Ton of
Recycled Inputs
in the
Manufacturing Staqe
035
0,75
0.84
037
1.00
1.00
1.00
1.00
(d)
Tons of
Product Made
Per Ton of
Recovered
Materials
0.77
0.66
0.77
0.83
0.97
0.87
037
0.87
Explanatory notes: The value in column "b" accounts for losses such as recovered newspapers that were unsuitable for recycling because they
were too wet. Column "c" reflects process waste losses at the manufacturing plant or mill. Column "d" is the product of the values in columns "b"
and "c."
-------
Exhibit 4-3 shows the greenhouse gas implications of recycling each material. The estimates in this
exhibit account for the loss rates for each material. Thus, the exhibit shows the GHG emissions, in metric
tons of carbon equivalent (MTCE), per short ton of material recovered (rather than emissions per ton of
material made with recycled inputs).
In addition, Exhibit 4-3 sums, for each material, the differences between manufacture from virgin
and recycled inputs for (1) energy-related greenhouse gas emissions (both in manufacturing processes and
transportation), (2) process non-energy-related greenhouse gas emissions, and (3) forest carbon
sequestration.
4.3 LIMITATIONS
Because the data presented in this chapter were developed earlier in Chapters 2 and 3, the
limitations discussed in those chapters also apply to the values presented here. Three other limitations are
as follows:
There may be GHG impacts from disposal of industrial wastes, particularly paper sludge at
paper mills. Because of the complexity of analyzing these second-order effects, and the lack
of data, we did not include them in our estimates. We did perform a screening analysis for
paper sludge, however, based on (1) data on sludge generation rates and sludge composition
(i.e., percentage of cellulose, hemicellulose, lignin, etc. in sludge), and (2) professional
judgment on the methane generation rates for cellulose, etc. The screening'analysis indicated
that net GHG emissions (methane emissions minus carbon sequestration) from paper sludge
are probably on the order of 0.00 MTCE per ton of paper made from virgin inputs to 0.01
MTCE per ton for recycled inputs. Our worst case bounding assumptions indicated
maximum possible net GHG emissions ranging from 0.03 to 0.11 MTCE per ton of paper
(depending on the type of paper and whether virgin or recycled inputs are used).
There is uncertainty in the loss rates - some materials recovery facilities and manufacturing
processes may recover or use recycled materials more or less efficiently than estimated here.
We used a simple representation of recycling as mostly closed loop. We considered open
loop processes for only two products, and even there our open loop model was simplified -
we considered only two products that might be made from each original product.
DRAFT -- March 1997
71
-------
Exhibit 4-3
Greenhouse Gas Emissions for Recycling
(MTCE/Ton of Material Recovered)
(a)
Material
Newspaper
Office Paper
Corrugated Cardboard
Aluminum Cans
Steel Cans
HOPE
LDPE
PET
(b)
Recycled Input
Credit*:
Process Energy
GHG
-0.14
-0.08
0.03
-2.66
-0.57
-0.31
-0.44
-0.58
(c)
Recycled Input
Credit*:
Transportation
Energy GHG
0.01
-0.01
0.00
-0.07
-0.01
-0.02
-0.01
-0.02
(d)
Recycled Input
Credit*:
Process Non-
Energy GHG
0.00
0.00
0.00
-1.24
0.00
-0.06
-0.06
-0.03
(e)
Forest Carbon
Sequestration
-0.73
-0.73
-0.73
0.00
0.00
0.00
0.00
0.00
(f)
GHG Reductions
From Using
Recycled Inputs
Instead of
Virgin Inputs
-0.86
-0.82
-0.70
-3.97
-0.57
-0.38
-0.51
-0.63
Material that is recycled post-consumer is then substituted for virgin inputs in Hie production of new products. This credit
represents the difference in emissions that results from using recycled inputs rather than virgin inputs. It accounts for
toss rates in collection, processing, and remanufacturing. Recycling credit is based on weighted average of closed and open
loop recycling for office paper and corrugated cardboard. However, all other estimates are only for the products themselves.
Explanatory notes: Columns "b" and "c" show the reduction in process energy GHGs and transportation energy GHGs from making each material
from recycled inputs, rather than virgin inputs. The values in columns "b" and "c" are based on (1) the difference in energy-related GHG emissions
between making one ton of the material from 100% virgin inputs and from 100% recycled inputs, multiplied by (2) the estimated tons of material
manufactured from one ton of material recovered, after accounting for loss rates in the recovery and remanufacturing stages. We first estimated the
values in columns "b" and "c" based on data provided by Franklin Associates, Ltd. (FAL), as shown in Exhibits 2-2 through 2-5. Then we estimated
the same values based on data provided by the Tellus Institute, as shown in Exhibits 2-6 through 2-9. Finally, we averaged the two sets of estimates
to obtain the values shown in columns "b" and "c." Note that for corrugated cardboard, the process energy GHG emissions are higher when using
recycled inputs than when using virgin inputs (as shown by the positive value in column "b" for corrugated cardboard). This is because the
manufacture of corrugated cardboard from virgin inputs uses a high proportion of biomass fuels - whose biogenic CO2 emissions are not counted as
GHG emissions (see the discussion of biogenic CO2 emissions in Chapter 1). Still, because of forest carbon sequestration, the net
-------
Explanatory notes for Exhibit 4-3 (continued):
GHG emissions from recycling corrugated cardboard are lower than the net GHG emissions from the re-manufacture of corrugated cardboard from
virgin inputs. Also note that the transportation GHGs for newsprint from recycled inputs are higher than for newsprint from virgin inputs. This is
because Tellus estimated much higher transportation energy for recycled inputs than for virgin inputs (FAL estimated nearly equal transportation
energy).
For column "d," which presents the process non-energy GHG emissions from recycling, we used (1) data provided by FAL showing the
difference in process non-energy GHG emissions between making one ton of the material from 100% virgin inputs, and from 100% recycled inputs
(as shown in the second to last column of Exhibits 2-2 and 2-4) multiplied by (2) the estimated amount of material manufactured (in tons) from one
ton of material recovered, after accounting for loss rates in the recovery and remanufacturing steps.
Next, in column "e," the exhibit shows the estimated forest carbon sequestration from recycling of paper products, as estimated in Chapter 3.
The last column (column "f') sums columns "b" through "e" to show the GHG implications of recycling each material.
-------
THIS PAGE INTENTIONALLY LEFT BLANK.
74
DRAFT - March 1997
-------
5. COMPOSTING
This chapter presents estimates of greenhouse gas (GHG) emissions and carbon sequestration from
composting of yard trimmings and food scraps.49
Composting consists of the aerobic decomposition of organic materials. In controlled composting
operations, organic materials are typically placed in piles that have sufficient moisture and aeration for
aerobic microorganisms (e.g., bacteria) to decompose the materials. Aeration may be provided by turning
the piles; this prevents the development of low-oxygen conditions in the piles which could lead to
anaerobic decomposition, with its associated noxious odors and methane generation. Nitrogen may be
added to a compost pile to achieve a carbon/nitrogen ratio that is optimal for rapid composting.
As organic materials are composted, they are converted into a form of organic matter known as
humus. When the compost is added to soil, the humus decomposes further. At both stages of
decomposition, much of the carbon in the original material is released in the form of carbon dioxide.
Because this carbon dioxide is biogenic in origin, it is not counted as a greenhouse gas (as explained in
Section 1.6). However, it is conceivable that composting could result in (1) methane emissions from
anaerobic decomposition, or (2) long-term carbon sequestration in the form of undecomposed carbon
compounds. In addition, with centralized composting there are non-biogenic CO2 emissions from
collection and transportation of the organic materials to the central composting site, and from mechanical
turning of the compost pile.50 Therefore, we investigated the extent to which composting might result in (1)
methane emissions, (2) carbon sequestration in soils to which compost is applied (for yard trimmings, we
considered the incremental carbon sequestration from composting, beyond the carbon sequestration
expected when yard trimmings are left in place on the ground) and (3) CO2 emissions from transportation
of compostable materials, and turning of the compost piles.
Our analysis suggests that composting, when properly done, does not result in methane generation,
and results in minimal carbon sequestration for yard trimmings. For centralized composting, slight GHG
emissions result from transportation of material to be composted and mechanical turning of the compost.
Overall, centralized composting of yard trimmings probably has no net GHG emissions (measured as GHG
emissions minus carbon sequestration). Similarly, backyard composting of food scraps is estimated to have
no net GHG emissions.
5.1 POTENTIAL GREENHOUSE GAS EMISSIONS
Two potential types of GHG emissions are associated with composting - (1) methane from
anaerobic decomposition, and (2) non-biogenic CO2 from transportation of compostable materials, and
turning of the compost piles.
49 Although paper and mixed MSW can be composted, we did not analyze the GHG implications of
composting them because of time and resource constraints.
50 CO2 emissions from delivery of compost to its final destination were not counted, because (1) compost is
a marketable product and (2) CO2 emissions from transportation of other marketable, finished goods to consumers
have not been counted in other parts of this analysis.
DRAFT--March 1997 75
-------
Methane. To research the methane issue, we first conducted a literature search for articles on
methane generation from composting. We identified no relevant articles published between 1991 and early
1995, and thus decided not to continue searching for earlier articles. Because the literature search was
unproductive, we contacted several researchers from universities and the US Department of Agriculture to
discuss the potential for methane generation, based on the nature of carbon flows during composting. Our
methane analysis was based on their expert opinions.
The researchers we contacted stated that well-managed compost operations usually do not generate
methane because they typically maintain an aerobic environment with proper moisture content to
encourage aerobic decomposition of the materials. They also said that even if methane is generated in
anaerobic pockets in the center of the compost pile, the methane is most likely oxidized when it reaches the
oxygen-rich surface of the pile. Several of the researchers commented that anaerobic pockets are most
likely to develop when too much water is added to the compost pile; however, they noted that this problem
rarely occurs because compost piles are much more likely to be watered too little, rather than too much.
For backyard composting, the compost pile is rarely large enough to permit anaerobic conditions to
develop, even in the center of the pile (i.e., all parts of the pile are close enough to the surface to remain
oxygenated).
We concluded from the available information that methane generation from backyard and
centralized compost piles is negligible.
Carbon Dioxide from Transportation of Materials and Turning of Compost. Next, we estimated the
indirect carbon dioxide emissions associated with collecting and transporting yard trimmings to centralized
compost facilities, and taming the compost piles. We began with estimates developed by Franklin
Associates, Ltd. for the amount of diesel fuel required, for one ton of yard trimmings,51 to (1) collect and
transport the yard trimmings to a central composting facility (363,000 BTUs), and (2) turn the compost
piles (221,000 BTUs).52 We converted these estimates to units of metric tons of carbon equivalent (MTCE)
per ton of yard trimmings, based on a carbon coefficient of 0.0208 MTCE per million BTUs of diesel fuel.
This resulted in an estimate of 0.01 MTCE of indirect CO2 emissions per ton of material composted in a
centralized facility. There are no indirect CO2 emissions from backyard composting, because there is no
significant use of machinery to transport materials or to turn the compost pile.
5.2 POTENTIAL CARBON SEQUESTRATION
We also evaluated the effect on carbon storage of composting yard trimmings and food scraps.
Yard Trimmings. For yard trimmings, our analysis compared the amount of long-term carbon
storage when yard trimmings are composted (and subsequently applied to soil) to the amount of carbon
storage when the trimmings are left directly on the ground to decompose. Because we Were unable to find
data allowing us to quantify incremental carbon storage, we used a bounding analysis to estimate the upper
and lower limits of the magnitude of this phenomenon.
51 Measured on a wet weight basis, as MSW is typically measured.
S2 Franklin Associates, Ltd., The Role of Recycling in Integrated Solid Waste Management to the Year 2000
(Stamford, CT: Keep America Beautiful), pp. 1-27,1-30, and 1-31.
76 DRAFT-March 1997
-------
During the process of decomposition, organic materials typically go through a series of steps before
finally being converted to CO2 (as well as water and other reaction products). The intermediate compounds
that are formed, and the lifetime of these compounds, can vary widely depending on the chemical
composition of the parent compound; the availability of oxygen and nutrients; the population of
microorganisms capable of degrading the compounds; temperature and moisture conditions; and many
other factors. To evaluate the potential of composting to enhance carbon storage, a useful simplification is
to view decomposition as a process consisting of two phases:
a rapid degradation phase, lasting for a few months to a few years, where the readily
degradable materials are converted mostly to CO2, and to a much lesser extent to humic
materials, and
a slow degradation phase, lasting much longer, where the humic materials are slowly
degraded to CO2.
Composting is designed to accelerate the pace of the first phase. It promotes rapid decomposition of
organics, thus reducing their volume. Some evidence suggests that composting produces a greater
proportion of humus than that typically formed when yard trimmings are left directly on the ground. The
conditions in the two settings are different - the heat generated within compost piles favors "thermophilic"
(heat-loving) bacteria, which tend to produce a greater proportion of stable, long-chain carbon compounds
than do bacteria that predominate at ambient surface temperatures. These long-chain carbon compounds
include lignin and humic materials (humic acids, fulvic acids, and humin).
For our analysis, we assumed that in soils where trimmings (i.e., grass clippings, leaves, and
branches) are left in place, there is no net accumulation of carbon in the soil. This assumption is consistent
with the observation that the quantity of carbon emitted from soils as carbon dioxide each year is typically
in equilibrium with the quantity of additional carbon introduced into the soil each year by roots, leaf litter,
and branches.53 We used this scenario as our baseline against which to measure incremental carbon storage
attributable to composting.
The incremental storage is a function of three principal factors:
(1) The amount of carbon in each material (grass, leaves, branches),
(2) The additional proportion of carbon converted into humus when trimmings are composted,
rather than left in place, and
(3) The rate at which humus is degraded to CO2.
We obtained point estimates for the first factor from a series of experiments by Dr. Morton Barlaz,
which are described later in Chapter 7. As in other parts of the analysis, we assumed that yard trimmings
53 Alexander, Martin, Introduction to Soil Microbiology, Second Edition (Malabar, Florida: Krieger
Publishing Company) 1991, p. 133.
DRAFT - March 1997 77
-------
comprise 50 percent grass clippings, 25 percent leaves, and 25 percent branches, by weight.54 We used
professional judgment to develop lower and upper bound estimates for the second and third factors, and
then combined the estimates in a bounding analysis.
As an upper bound on the incremental humus formation, we assumed that composting can result in
conversion of up to 25 percent more of the carbon to humus than the "baseline" conversion rate (i.e., if
residues were left on the ground).55 (This upper bound implies, for example, that if 10 percent of the
carbon is in a relatively stable form following decomposition at ambient temperatures, then 35 percent of
the carbon would be relatively stable after composting.) For a lower bound, we used a value of 5 percent as
the incremental portion of carbon that is converted to stable carbon compounds.
We also developed a range for the half-life of stable carbon compounds in soil. Radiocarbon dating
of soils has shown that the long-chain carbon compounds in some soil samples can be hundreds or
thousands of years old.56 As noted above, the decay rate of individual compounds is highly site- and
compound-specific; to account for this heterogeneity, we used wide bounds - from a half-life of 20 years to
a half-life of 2,000 years. We assumed that humus decomposition is a first-order decay process (i.e., the
proportional decrease in concentration is constant over time).
Combining the two bounds for incremental humus formation (5 percent and 25 percent) and the two
bounds for half-lives (20 years and 2,000 years) resulted in four scenarios for the bounding analysis. We
estimated the incremental carbon storage implied by each scenario over a period of 100 years.
The results of our bounding analysis are shown graphically in Exhibit 5-1. The upper bound on the
incremental carbon storage from composting is about 0.05 MTCE per ton of yard trimmings (shown in the
top left of the graph); the lower bound is about 0.001 (shown in the bottom right of the graph). With the
rapid decomposition (20 year half-life) assumption, incremental storage is quite sensitive to the time period
over which carbon storage is considered - values at 20 years are sixteen times as high as values at 100
years. Under the slow decomposition assumption, there is little difference in incremental storage for all
periods up to 100 years.
Food Scraps. We also estimated the carbon storage from backyard composting of food scraps. Data
were not available on the amount of carbon sequestered in humus when food scraps are composted. We
assumed that backyard composting of food scraps converts all of the carbon in food scraps to CO2, and that
none of the carbon is sequestered in humus. To the extent that backyard composting of food scraps may
sequester carbon, our results would understate the net carbon sequestration resulting from composting this
material.
54 This professional judgment estimate for the percentage composition of yard trimmings (as a national
average) was provided by Nick Artz of Franklin Associates, Ltd. (FAL) in a telephone conversation with William
Driscoll of IGF Incorporated, November 14,1995. Subsequently, FAL obtained and provided data showing a wide
range of percentage breakdowns for yard waste composition in different states; the percentage composition used
here is within that range.
55 Memorandum from Michael Cole, University of Illinois at Urbana-Champaign to George Garland, U.S.
EPA Office of Solid Waste, February 1,1996.
16 Allison, F.E., Soil Organic Matter and Its Role in Crop Production (Elsevier Scientific Publishing Co.)
1973, pp. 157-8.
78 DRAFT - March 1997
-------
Incremental C Storage
(MTCESton)
I
-------
5.3 NET GHG EMISSIONS FROM COMPOSTING
Exhibit 5-2 presents the estimated net greenhouse gas emissions from composting. Our analysis
indicated that composting is a process that produces virtually no greenhouse gas emissions, and is not
likely to represent a significant carbon sink. For centralized yard trimmings composting, the transportation
emissions are probably balanced (and could well be exceeded) by additional carbon storage. Given the
large tonnage of yard trimmings composted annually, and the remaining uncertainties in this analysis, this
is an area that would benefit from further study. For backyard food waste composting, we estimated no net
GHG emissions.
EXHIBIT 5-2
Net Greenhouse Gas Emissions from Composting
(In Metric Tons of Carbon Equivalent Per Short Ton of Material Composted)
Material
Yard
Trimmings
Food
Scraps
CH*
0
NA
Centralized Composting
Transport
CO,
0.01
NA
C seq
0.001 to
0.05
NA
NetC
0.009 to
-0.49
NA
Backyard Composting
CH4
NA
0
Transport
CO,
NA
0
Cseq
NA
0
NetC
NA
0
5.4 LIMITATIONS OF THE ANALYSIS
The analyses in this chapter are limited by the lack of data on methane generation and carbon
sequestration that result from composting. Because of inadequate data, we relied on a theoretical approach
to estimate the values (and in the case of carbon sequestration from composting of food scraps, we
assumed zero carbon sequestration).
Our analysis did not consider the GHG emissions that might be avoided if compost displaces some
chemical fertilizers (or peat moss, fungicides, pesticides, and other products applied to soil and plants).
The manufacture of chemical fertilizers requires energy, and thus is associated with some level of energy-
related GHG emissions. We also did not analyze the extent to which compost may reduce the need for
pesticides.57-58 For the most part, compost is applied for its soil amendment properties, rather than for
purposes of fertilization or pest control.
57 For example, the use of compost may eliminate the need for soil fumigation with methyl bromide (an
ozone-depleting substance) to kill plant pests and pathogens.
80
" EPA plans to investigate in 1997 the GHG impacts of substituting compost for fertilizer.
DRAFT - March 1997
-------
Moreover, we did not consider other environmental benefits of composting, and of using compost as
a soil amendment. For example, adding compost to soil increases the soil's capability to retain moisture
and nutrients. This helps to reduce storm runoff, thus preserving topsoil and reducing siltation of streams
and rivers. In the future, this may allow continued farming in areas that might have more frequent droughts
due to climate change, Adding compost to soil also improves soil tilth and reduces soil density, i.e., it
makes the soil easier to till, allows plant roots to go deeper, increases the likelihood that new plantings
become established, and helps plants to grow larger. Finally, we did not consider the value of composting
in reducing the amount of waste landfiUed, and extending the useful lifetime of landfills.
DRAFT -- March 1997
81
-------
THIS PAGE INTENTIONALLY LEFT BLANK.
82
DRAFT - March 1997
-------
6. COMBUSTION
This chapter presents estimates of the net GHG emissions from combustion of (1) each of the ten
materials considered in this analysis, and (2) mixed MSW.
Combustion of municipal solid waste (MSW) results in emissions of CO2 (because nearly all of the
carbon in MSW is converted to CO2), and N2O, a potent greenhouse gas. Note, however, that CO2 from
burning biomass sources (such as paper products and yard trimmings) is not counted as a GHG, because it is
biogenic (as explained in Section 1.6).
Combustion of MSW with energy recovery in a waste-to-energy (WTE) plant also results in avoided
CO2 emissions in two other industrial sectors. First, the electricity produced by a WTE plant displaces
electricity that would otherwise be provided by an electric utility power plant. Because most utility power
plants burn fossil fuels, and thus emit CO2, the electricity produced by a WTE plant reduces utility CO2
emissions. These avoided GHG emissions must be subtracted from the GHG emissions associated with
combustion of MSW. Second, most MSW combusted with energy recovery in the US is combusted in WTEs
that recover ferrous metals (e.g., steel).59 The ferrous metals that are recovered are then recycled. Steel from
recycled steel requires less energy than steel produced from iron ore, resulting in lower CO2 emissions. Thus,
the additional recycling of steel associated with MSW combustion reduces CO2 emissions in steel
manufacturing.
We analyzed the net GHG emissions from combustion of mixed MSW, and the following individual
materials:
newspaper,
office paper,
corrugated cardboard,
aluminum cans,
steel cans,
HDPE plastic,
LDPE plastic,
PET plastic,
yard trimmings, and
food scraps.
Net emissions consist of (1) emissions of non-biogenic CO2 and N2O minus (2) avoided GHG emissions in
the electric utility and steel sectors. There is some evidence that as combustor ash ages, it absorbs CO2 from
the atmosphere; however, we did not count CO2 absorbed because we estimated the quantity absorbed to be
59 We did not consider any recovery of materials from the MSW stream that may occur before MSW is
delivered to the combustor. We considered such prior recovery to be unrelated to the combustion operation - unlike
recovery of steel from combustor ash, an activity that is an integral part of the operation of many combustors.
DRAFT - March 1997
83
-------
less than 0.01 MTCE per ton of MSW combusted.60 Combustion also results in emissions of nitrogen oxide
(NO), nitrogen dioxide (NO^, and sulfur dioxide (SO2) - all of which contribute indirectly to climate
change.61 However, we did not consider these gases in this analysis because there is no widely accepted
method to estimate their contribution to climate change.62
Our results showed that combustion of mixed MSW has small positive net GHG emissions (in
absolute terms). Combustion of paper products, food scraps, and yard trimmings results in negative net GHG
emissions. Running steel cans through a combustor likewise results in negative net GHG emissions.
Combustion of plastic produces large positive net GHG emissions, and combustion of aluminum cans results
in small positive net GHG emissions. The reasons for each of these results are presented throughout this
chapter.
6.1
METHODOLOGY
Our general approach was to estimate the (1) gross emissions of CO2 and N2O from MSW
combustion (including emissions from transportation of waste to the combustor, and ash from the combustor
to a landfill) and (2) CO2 emissions avoided due to displaced electric utility generation and increased
production of steel from recycled inputs.63 To obtain an estimate of the net GHG emissions from MSW
combustion, we subtracted the GHG emissions avoided from the direct GHG emissions. We estimated the
net GHG emissions from waste combustion per ton of mixed MSW, and per ton of each selected material in
MSW. The remainder of this section describes how we developed these estimates.
Estimating Direct CO2 Emissions from MSW Combustion
The carbon in MSW has two distinct origins. Some of the carbon in MSW is biomass carbon (i.e.,
carbon in plant and animal matter that was converted from CO2 in the atmosphere through photosynthesis).
The remaining carbon in MSW is from non-biomass sources, e.g., plastic and synthetic rubber derived from
petroleum.
We did not count the biogenic CO2 emissions from combustion of biomass, as described in Section
1.6. On the other hand, we did count CO2 emissions from combustion of non-biomass components of MSW
*° Based on data provided by Dr. Jurgen Vehlow [of Karlsruhe, Germany's Institut fur Technische
Chemie], we estimated that the ash from one ton of MSW would absorb roughly 0.004 MTCE of CO2.
61 These gases contribute indirectly to climate change either because they are transformed in the
atmosphere into a greenhouse gas or gases, or because they influence the atmospheric lifetimes of greenhouse gases.
62 Because the Intergovernmental Panel on Climate Change (IPCC) has not established a method for
estimating the global warming implications of emissions of these gases, we have not attempted such an estimation.
Note, however, that NO and NO2 emissions are believed to increase global warming, whereas SO2 is believed to
counteract global wanning (by forming sulfate aerosols that reflect sunlight, and that aid in the formation of clouds,
which also reflect sunlight).
63 A comprehensive evaluation would also consider the fate of carbon remaining in combustor ash.
Depending on its chemical form, carbon may be aerobically degraded to CO2, anaerobically degraded to CH4, or
remain in a relatively inert form and be sequestered. Unless the ash carbon is converted to CH4 (which we
considered to be unlikely), the effect on the net GHG emissions would be very small.
84
DRAFT -- March 1997
-------
- plastic, textiles, and rubber. Overall, only a small portion of the total CO2 emissions from combustion are
counted as GHG emissions.
For mixed MSW, we used the simplifying assumptions that (1) all carbon in textiles was non-
biomass carbon, i.e., petrochemical-based plastic fibers such as polyester (this is a worst-case assumption),
and (2) the category of "rubber and leather" in EPA's MSW characterization report64 was composed almost
entirely of rubber. Based on these assumptions, we estimated that there are 0.11 pounds of non-biogenic
carbon in the plastic, textiles, rubber, and leather contained in one pound of mixed MSW.65 We assumed that
98 percent of this carbon would be converted to CO2 when the waste was combusted, with the balance going
to the ash. Then, we converted the 0.11 pounds of non-biomass carbon per pound of mixed MSW to units of
metric tons of carbon equivalent (MTCE) per ton of mixed MSW combusted. The resulting value for mixed
MSW is 0.10 MTCE per ton of mixed MSW combusted,66 as shown in Exhibit 6-1.
We estimated that HDPE and LDPE are 84 percent carbon, while PET is 57 percent carbon.67
(accounting for a moisture content of 2 percent). We again assumed that 98 percent of the carbon in the
plastic is converted to CO2 during combustion. The values for CO2 emissions, converted to units of MTCE
per ton of plastic combusted, are shown in column "b" of Exhibit 6-1.
Estimating N2O Emissions from Combustion of Waste
Recent studies compiled by the Intergovernmental Panel on Climate Change (IPCC) show that
MSW combustion results in measurable emissions of N2O (nitrous oxide), a greenhouse gas with a high
global warming potential (GWP).68 The IPCC compiled reported ranges of N2O emissions, per metric ton of
waste combusted, from six classifications of MSW combustors. We averaged the midpoints of each range
and converted the units to MTCE of N2O per short ton of MSW; the resulting estimate is 0.01 MTCE of
N2O emissions per ton of mixed MSW combusted. Because the IPCC did not report N2O values for
combustion of individual components of MSW, we used the 0.01 value not only for mixed MSW, but also as
a proxy for all components of MSW, except for aluminum and steel cans.69
64 U.S. EPA, Office of Solid Waste and Emergency Response, Characterization of Municipal Solid Waste
in the United States: 1994 Update, November 1994.
65ICF Incorporated, "Work Assignment 239, Task 2: Carbon Sequestration in Landfills," memorandum to
Michael Podolsky, Clare Lindsay, and Brett Van Akkeren of EPA, April 28, 1995, Exhibit 2-A, column "o."
66 Note that if we had used a best-case assumption for textiles, i.e., assuming they had no petrochemical-
based fibers, the resulting value for mixed MSW would have been 0.09 MTCE per ton of mixed MSW combusted.
67 ICF Incorporated, op cit., Exhibit 1-A, column "n."
68 Intergovernmental Panel on Climate Change, Greenhouse Gas Inventory Reference Manual, Volume 3,
(undated) p. 6-33. The GWP of N2O is 270 times that of CO2.
69 This exception was made because at the relatively low combustion temperatures found in MSW
combustors, most of the nitrogen in N2O emissions is derived from the waste, not from the combustion air. Because
aluminum and steel cans do not contain nitrogen, we concluded that running these metals through an MSW
combustor would not result in N2O emissions.
DRAFT - March 1997 85
-------
r
Exhibit 6-1
Gross Emissions of Greenhouse Gases From MSW Combustion
(MTCEfl-on)
(a)
Material
Combusted
Newspaper
Office Paper
Corrugated Cardboard
Aluminum Cans
Steel Cans
HOPE
LDPE
PET
Food Scraps
Yard Trimmings
Mixed MSW
(b)
Combustion CO2
Emissions From
Non-Biomass Per
Ton Combusted
0.00
0.00
0.00
0.00
0.00
0.75
0.75
0.51
0.00
0.00
0.10
(c)
Combustion N2O
Emissions
Per Ton
Combusted
0.01
0.01
0.01
0.00
0.00
0.01
0.01
0.01
0.01
0.01
0.01
(d)
Transportation CCs
Emissions
Per Ton
Combusted
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
(e)
Gross GHG
Emissions Per
Ton Combusted
0.02
0.02
0.02
0.01
0.01
0,77
0.77
0.53
0.02
0.02
0.12
Estimating Indirect CO2 Emissions from Transportation of Waste to the WTE Plant
Next, we estimated the indirect CO2 emissions from the transportation of waste. For the indirect CO2
emissions from transporting waste to the WTE plant, and ash from the WTE plant to a landfill, we used an
estimate for mixed MSW developed by Franklin Associates, Ltd. (FAL).70 We then converted the FAL
estimate from pounds of CO2 per ton of mixed MSW to MTCE per ton of mixed MSW. This resulted in an
estimate of 0.01 MTCE of CO2 emissions from transporting one ton of mixed MSW, and the resulting ash.
We assumed that transportation of any individual material in MSW would use the same amount of energy as
transportation of mixed MSW.
Estimating Gross Greenhouse Gas Emissions from Combustion
To estimate the gross GHG emissions per ton of waste combusted, we summed the values for
emissions from combustion CO2, combustion N2O, and transportation CO2. The gross GHG emissions
estimates, for mixed MSW and for each individual material, are shown in column "e" of Exhibit 6-1.
Estimating Utility CO2 Emissions Avoided
Most WTE plants in the US produce electricity. Only a few produce steam, and few if any
cogenerate electricity and steam. Thus, in our analysis, we assumed that the energy recovered with MSW
combustion would be in the form of electricity. Our analysis is shown in Exhibit 6-2. We used three data
elements to estimate the avoided electric utility CO2 emissions associated with combustion of waste in a
70 Franklin Associates, Ltd., The Role of Recycling in Integrated Solid Waste Management to the Year 2000
(Stamford, CT: Keep America Beautiful, Inc.) September 1994, p. 1-24.
86
DRAFT - March 1997
-------
Exhibit 6-2
Avoided Utility GHG Emissions from MSW Combustion
(a)
Material Combusted
Newspaper
Office paper
Corrugated cardboard
Aluminum cans
Steel cans
HOPE
LDPE
PET
Yard trimmings
Food scraps
Mixed MSW
(b)
Energy Content
(BTUs per
pound)
7,950 a
6,800 a,b
7,043 a
-335 c
-210 c
18,687 a
18,687 a
9,702 d,e
2,800 f
2,370 a
5,358 a
(c)
Energy
Content
(1E6
BTUs per
ton)
15.9
13.6
14.1
-0.7
-0.4
37.4
37.4
19.4
5.6
4.7
10.7
(d)
Combustion
System
Efficiency
(Percent)
13.6%
13.6%
13.6%
13.6%
13.6%
13.6%
13.6%
13.6%
13.6%
13.6%
13.6%
(e)
Emission
Factor for
Utility-
Generated
Electricity
(MTCE/1E6
BTUs of
electricity
delivered)
0.05112
0.05112
0.05112
0.05112
0.05112
0.05112
0.05112
0.05112
0.05112
0.05112
0.05112
(f)
Avoided Utility
CC-2 Per Ton
Combusted
(MTCE)
0.11
0.09
0.10
0.00 *
0.00 *
0.26
0.26
0.13
0.04
0.03
0.07
'The amount of energy absorbed by one ton of steel or aluminum cans in an MSW combustor would, if not absorbed, result in
less than 0.01 MTCE of avoided utility CO2.
a MSW Fact Book.
b We used the MSW Fact Book's value for mixed paper as a proxy for the value for office paper.
c We developed these estimates based on data on the specific heat of aluminum and steel,
and calculated the energy required to raise the temperature of aluminum or steel from
ambient temperature to the temperature found in a combustor (about 750° Celsius). We
obtained the specific heat data from Incropera, Frank P. and David P. DeWitt,
Introduction to Heat Transfer, Second Edition (New York: John Wiley & Sons) 1990,
pp. A3-A4.
d Gaines and Stodolsky.
e For PET plastic, we converted the value of 9,900 BTUs/pound dry weight, to 9,702
BTUs/pound wet weight, to account for a moisture content of 2 percent.
f Procter and Redfem, Ltd. and ORTECH International.
g Berenyi and Gould.
WTE plant: (1) the energy content of mixed MSW and of each separate waste material considered, (2) the
combustion system efficiency in converting energy in MSW to delivered electricity, and (3) the electric
utility CO2 emissions avoided per kilowatt-hour of electricity delivered by WTE plants.
Energy content. For the energy content of mixed MSW, we used a value of 5,358 BTUs per pound
of mixed MSW combusted, which was the mean value based on data from 133 (of the total of 171)
municipal waste combustors in the US that provided data in a survey by Governmental Advisory Associates
DRAFT -March 1997
87
-------
(GAA).71 This estimate is about the midpoint of the range of values (4,500 to 6,500 BTUs per pound)
reported by FAL,72 and is slightly higher than the 4,800 BTUs per pound value reported in EPA's MSW
Fact Book.73 For the energy content of specific materials in MSW, we evaluated three sources: (1) EPA's
MSW Fact Book (a compilation of data from primary sources), (2) a report by the Canadian government,74
and (3) a report by Argonne National Laboratories.75 We assume that the energy contents reported in the first
two of these sources were for materials with moisture contents typically found for the materials in MSW (the
sources implied this, but did not explicitly state it). The Argonne study reported energy contents on a dry
weight basis.
Combustion system efficiency. To estimate the combustion system efficiency of WTE plants, we
began with GAA's reported net value of 471 kWh generated by WTE plants per ton of mixed MSW
combusted.76 This value is consistent with the net 480 kWh generated per ton of mixed MSW reported by
FAL.77 Next, we considered losses in transmission and distribution of electricity, and used the US average
transmission and distribution loss of 9 percent,78 to estimate that 429 kWh are delivered per ton of waste
combusted.
We then used the value for the delivered kWhs per ton of waste combusted to derive the implicit
combustion system efficiency (i.e., the percentage of energy in the waste that is ultimately delivered in the
form of electricity). To determine this efficiency, we first estimated the BTUs of MSW needed to deliver one
kWh of electricity. We divided the BTUs per ton of waste by the delivered kWh per ton of waste to obtain
the BTUs of waste per delivered kWh. The result is 25,000 BTUs per kWh. Next we divided the physical
constant for the energy in one kWh (3,412 BTUS) by the BTUs of MSW needed to deliver one kWh, to
estimate the total system efficiency at 13.6 percent.79 This relatively low efficiency of combustion is due in
71 Berenyi, Eileen B. and Robert N. Gould, Resource Recovery Yearbook 1993-94 (New York, NY:
Governmental Advisory Associates, Inc.) 1993, p. 47.
72 Franklin Associates, Ltd., p. 1-16.
73 U.S. Environmental Protection Agency, Office of Solid Waste, MSW Fact Book, Version 2.0
(Washington, D.C.: U.S. Environmental Protection Agency) April 1995.
74 Procter and Redfern, Ltd. and ORTECH International, Estimation of the Effects of Various Municipal
Waste Management Strategies on Greenhouse Gas Emissions, Part II (Ottawa, Canada: Environment Canada, Solid
Waste Management Division, and Natural Resources Canada, Alternative Energy Division), September 1993.
75 Gaines, Linda, and Frank Stodolsky, "Mandated Recycling Rates: Impacts on Energy Consumption and
Municipal Solid Waste Volume" (Argonne, IL: Argonne National Laboratory) December 1993, pp. 11 and 85.
76 Berenyi and Gould, op cit, p. 46.
77 Franklin Associates, Ltd., op cit, p. 1-21.
78 U.S. Department of Energy, Energy Information Administration, Annual Energy Review 1993
(Washington, D.C.: Energy Information Administration) July 1994, p. 252.
79 Note that the total system efficiency is the efficiency of translating the energy content of the fuel into the
energy content of delivered electricity. The relatively low system efficiency of 13.6 percent reflects losses in (1)
converting energy in the fuel into steam, (2) converting energy in steam into electricity, and (3) delivering
88
DRAFT -- March 1997
-------
part to the absorption of heat energy by moisture (as well as metals and glass) in MSW. We recognize that
combustors dedicated to a single type of high-BTU waste, such as paper, may be able to achieve a higher
combustion system efficiency, but we did not estimate the net GHG emissions from such combustors.
Moreover, it is likely that new combustion capacity would have a higher efficiency than the average existing
combustion facility.
Electric utility carbon emissions avoided. To estimate the avoided utility CO2 from waste
combustion, we used the results in columns "c" and "d," together with a "carbon coefficient" of 0.051
MTCE emitted per million BTUs of utility-generated electricity (delivered), based on the national average
fuel mix used by utilities.80 This approach implicitly uses the average fuel mix as a proxy for the fuels
displaced at the margin when utility-generated electricity is displaced by electricity from WTE plants. The
actual carbon reductions could vary depending on which type of fuel used to generate electricity is displaced
at the margin. The resulting estimates for utility carbon emissions avoided for each material are shown in
column "f' of Exhibit 6-2.
Approach to Estimating CO2 Emissions Avoided Due to Increased Steel Recycling
We next estimated the avoided CO2 emissions from increased steel recycling made possible by steel
recovery from WTE plants for (1) mixed MSW and (2) steel cans.
For mixed MSW, we estimated the amount of steel recovered per ton of mixed MSW combusted,
based on (1) the amount of MSW combusted in the US, and (2) the amount of steel recovered, post-
combustion. Ferrous metals are recovered at sixty-five MSW combustion facilities in the US. These facilities
account for approximately 75 percent of the 32 million tons of MSW combusted per year, and recover a total
of about 532,000 tons per year of ferrous metals.81 By dividing 532,000 tons by 32 million tons, we
estimated that 0.02 tons of steel are recovered per ton of MSW combusted (as a national average).
For steel cans, we first estimated the national average proportion of steel cans entering WTE plants
that would be recovered. As noted above, approximately 75 percent of MSW destined for combustion goes
to facilities with a ferrous recovery system; at these plants, approximately 98 percent of the steel cans would
be recovered.82 We multiplied these percentages to estimate the weight of steel cans recovered per ton of
steel cans combusted - about 0.74 tons per ton.
Finally, to estimate the avoided CO2 emissions due to increased recycling of steel, we multiplied (1)
the weight of steel recovered by (2) the avoided CO2 emissions per ton of steel recovered. Thus, we
estimated avoided CO2 emissions of 0.42 MTCE per ton for steel cans, and 0.01 MTCE per ton for MSW, as
shown in column "d" of Exhibit 6-3.
electricity. The losses in delivering electricity are the transmission and distribution losses, estimated at 9 percent.
80 R. Neal Elliott, "Carbon Reduction Potential from Recycling in Primary Materials Manufacturing"
(Washington, D.C.: American Council for an Energy-Efficient Economy) February 8, 1994, p. 14.
81 Telephone conversation, David Sussman, Senior Vice President for Environmental Affairs, Ogden
Corporation, with William Driscoll, ICF, April 4,1995.
82 Telephone conversation, David Sussman with William Driscoll, May 12, 1995.
DRAFT -- March 1997 89
-------
Exhibit 6-3
Avoided GHG Emissions Due to
Increased Steel Recovery from MSWCombustors
(a)
Material Combusted
Newspaper
Office paper
Corrugated cardboard
Aluminum cans
Steel cans
HOPE
LDPE
PET
Yard trimmings
Food scraps
Mixed MSW
(b)
Tons of Steel
Recovered Per Ton
of Waste Combusted
(tons)
NA
NA
NA
NA
0.735
NA
NA
NA
NA
NA
0.017
(c)
Avoided CO2
Emissions Per Ton of
Steel Recovered
(MTCE/ton)
NA
NA
NA
NA
0.57
NA
NA
NA
NA
NA
0.57
(d)*
Avoided CO2
Emissions Per Ton of
Waste Combusted
(MTCE/ton)
NA
NA
NA
NA
0.42
NA
NA
NA
NA
NA
0.01
'The value in column "d" is a national average, and is weighted to reflect 98 percent recovery at the 75 percent
of facilities that recover ferrous metals.
6.2 RESULTS
The results of our analysis are shown in Exhibit 6-4. The results from the last columns of Exhibits 6-
1, 6-2, and 6-3 are shown in columns "b," "c," and "d," respectively, of Exhibit 6-4. The net GHG emissions
from combustion of each material are shown in column "e." These net values represent the gross GHG
emissions (column "b"), minus the avoided GHG emissions (columns "c" and "d"). As stated earlier, these
net GHG emissions estimates are absolute values, not values relative to some other waste management
option.
We estimate that combustion of mixed MSW has slightly positive net GHG emissions of 0.04
MTCE per ton. Combustion of paper products has negative net GHG emissions of approximately -0.1
MTCE per ton, because CO2 emissions from burning paper are not counted and fossil fuel burning by
utilities is avoided. Combustion of food scraps and yard trimmings (two other forms of biomass) also has
negative net GHG emissions, but of a smaller magnitude (-0.01 to -0.02 MTCE per ton of material).
Combustion of plastics results in substantial net GHG emissions estimated at 0.39 to 0.51 MTCE
per ton. This is primarily because of the high content of non-biomass carbon in plastics. Also, when
combustion of plastic results in electricity generation, the utility carbon emissions avoided (due to displaced
utility fossil fuel combustion) are much less than the carbon emissions from the combustion of plastic. This
is largely due to the lower system efficiency of WTE plants, compared to electric utility plants. Recovery of
ferrous metals at combustors results in negative net GHG emissions, estimated at -0.41 MTCE per ton of
steel cans, due to the increased steel recycling made possible by ferrous metal recovery at WTE plants.
Combustion of aluminum cans, on the other hand, results in slight positive net GHG emissions of 0.01
MTCE per ton, due to the energy used in transporting the cans to the WTE plant.
90
DRAFT -- March 1997
-------
Exhibit 6-4
Net GHG Emissions from MSW Combustion
(a)
Material Combusted
Newspaper
Office paper
Corrugated cardboard
Aluminum cans
Steel cans
HOPE
LDPE
PET
Yard trimmings
Food scraps
Mixed MSW
(b)
Gross GHG
Emissions Per Ton
Combusted
(MTCEAon)
0.02
0.02
0.02
0.01
0.01
0.77
0.77
0.53
0.02
0.02
0.12
(c)
Avoided Utility COz
Per Ton Combusted
(MTCEAon)
0.11
0.09
0.10
0.00
0.00
0.26
0.26
0.13
0.04
0.03
0.07
(d)
Avoided COz
Emissions Per Ton
Combusted Due to
Steel Recovery
(MTCEAon)
NA
NA
NA
NA
0.42
NA
NA
NA
NA
NA
0.01
(e)
Net GHG Emissions
from Combustion
(MTCEAon)
-0.09
-0.07
-0.08
0.01
-0.41
0.51
0.51
0.39
-0.02
-0.01
0.04
6.3 LIMITATIONS OF THE ANALYSIS
The reliability of the analysis presented in this chapter is limited by the reliability of the various data
elements used. The most significant limitations are as follows:
Combustion system efficiency of WTE plants may be improving. A survey of planned WTE
plants shows an expected efficiency improvement of 14 percent over current plants.83 If
efficiency improves, more utility CO2 will be displaced per ton of waste combusted
(assuming no change in utility emissions per kWh), and the net GHG emissions from
combustion of MSW will decrease.
The reported ranges for N2O emissions were broad; in some cases the high end of the range
was 10 times the low end of the range. Research has indicated that N2O emissions vary with
the type of waste burned. Thus, the average value used for mixed MSW and for all MSW
components should be interpreted as an approximate value.
For mixed MSW, we assumed that all carbon in textiles is from synthetic fibers derived
from petrochemicals (whereas, in fact, some textiles are made from cotton, wool, and other
natural fibers). Because we assumed that all carbon in textiles is non-biogenic, we counted
all of the CO2 emissions from combustion of textiles as GHG emissions. This assumption
will slightly overstate the net GHG emissions from combustion of mixed MSW, but the
magnitude of the error is small because textiles represent only a small fraction of the MSW
stream. Similarly, the MSW category of "rubber and leather" contains some biogenic carbon
83 Berenyi and Gould, op cit, p. 46.
DRAFT - March 1997
91
-------
from the leather. By not considering this small amount of biogenic carbon, the analysis
slightly overstates the GHG emissions from MSW combustion.
Because the makeup of a given community's mixed MSW may vary from the national
average, the energy content may also vary from the national average energy content that we
used in this analysis. For example, MSW from communities with a higher or lower than
average recycling rate may have a different energy content, and MSW with more than the
average proportion of dry leaves and branches will have a higher energy content.
In our analysis, we used the national average recovery rate for steel. Where waste is sent to a
WTE plant vf ith steel recovery, the net GHG emissions for steel cans will be slightly lower
(i.e., more negative). Where waste is sent to a WTE plant without steel recovery, the net
GHG emissions for steel cans will be the same as for aluminum cans (i.e., close to zero).
We used in this analysis the national average fuel mix for electricity as the proxy for fuel
displaced at the margin when WTE plants displace utility electricity. If some other fuel or
mix of fuels is displaced at the margin (e.g. coal), the avoided utility CO2 would be different
(e.g., for coal, the avoided utility CO2 would be about 0.025 MTCE per ton higher for
mixed MSW, and the net GHG emissions would be 0.01 MTCE instead of 0.04 MTCE per
ton).
Combustors dedicated to a single type of high-BTU waste, such as paper, may be able to
achieve a higher combustion system efficiency, but we did not estimate the net GHG
emissions from such combustors.
92
DRAFT--March 1997
-------
7. LANDFILLING
This chapter presents estimates of GHG emissions and carbon sequestration from landfilling of (1)
each of the ten materials considered in this analysis, and (2) mixed MSW.
When food scraps, yard trimmings, and paper are landfilled, anaerobic bacteria degrade the
materials, producing methane (a potent GHG) and carbon dioxide. The carbon dioxide is not counted as a
GHG because it is biogenic, as explained in Section 1.6. Because metals do not contain carbon, they do not
generate methane when landfilled. Plastics are essentially not biodegradable, and therefore do not generate
any methane.
Because food scraps, yard trimmings, and paper are not completely decomposed by anaerobic
bacteria, some of the carbon in these materials is sequestered in the landfill. However, carbon in plastic that
remains in the landfill is not counted as sequestered carbon (as described in Section 1.4).
Transportation of waste materials to a landfill results in anthropogenic carbon dioxide emissions,
due to the combustion of fossil fuels in the vehicles used to haul the wastes.
For this study, we estimated the methane emissions, anthropogenic carbon dioxide emissions, and
carbon sequestration that will result from landfilling each type of organic waste, and from landfilling mixed
MSW. We accounted for the extent to which methane will be flared at some landfills, and will be combusted
for energy recovery at others. In both cases, we projected future landfill gas (LFG) recovery rates based on a
significant increase in the use of LFG recovery systems due to a new EPA rule that requires gas recovery at
large MSW landfills.84
Our results showed that landfilling of office paper results in substantial positive net GHG emissions,
and that landfilling of food scraps and grass have small positive net GHG emissions (in absolute terms). For
these three materials, the net GHG emissions from methane generation exceed the carbon sequestration (for
the fraction of these materials that does not degrade in landfills). For all of the other materials that we
examined, landfilling results in negative net GHG emissions in absolute terms - ranging from slight negative
net emissions for corrugated boxes, mixed MSW, and yard trimmings,85 to moderate negative net emissions
for branches, leaves, and newspaper. For these materials, carbon sequestration exceeds the net GHG
emissions from methane generation (after accounting for projected LFG recovery).
The results would differ if a different assumption were used for the percentage of landfill methane
recovered in the year 2000. At lower (e.g., current) rates of LFG recovery, the net GHG emissions of office
84 The rule requires a well-designed and well-operated landfill gas collection system at landfills that (1)
have a design capacity of at least 2.5 million metric tons, or 2.5 million cubic meters, and (2) emit more than 50
metric tons of nonmethane organic compounds per year (Federal Register, Vol. 61, No. 49, p. 9905, March 12,
1996).
85 Yard trimmings were estimated to consist of 50 percent grass clippings, 25 percent leaves, and 25
percent branches from trees and shrubs.
DRAFT - March 1997 93
-------
paper, food scraps, and grass increase further, and the net GHG emissions of corrugated boxes, yard
trimmings, and mixed MSW turn from negative to positive.
7.1 EXPERIMENTAL VALUES FOR METHANE GENERATION AND CARBON
SEQUESTRATION
To estimate methane emissions and carbon sequestration from landfilling of specific materials, we
used data from laboratory experiments conducted by Dr. Morton Barlaz.86 The experiments provided data on
(1) the amount of methane generated by each type of organic material, when digested by bacteria in
anaerobic conditions simulating those in a landfill, and (2) the amount of carbon remaining, undecomposed
(i.e., sequestered) at the end of the experiment.
Experimental Design
Barlaz placed each type of organic waste and mixed MSW in separate reactor vessels, in which he
maintained anaerobic conditions similar to those in a landfill, but controlled to favor maximum methane
generation. Barlaz measured the amount of methane generated in each reactor, and the amount of
undecomposed carbon remaining in each reactor at the end of the experiment. Each material was tested in
four reactors, and the results from each were averaged.87
At the start of the experiment, Barlaz dried a sample of each material, and analyzed the amount of
cellulose, hemicellulose, and lignin (and, for food scraps, protein) in each material. Cellulose, hemicellulose,
and protein partly decompose in a landfill, resulting in methane generation; lignin is relatively stable and
non-decomposable under anaerobic conditions.
Portions of each material were weighed, placed in two-liter plastic containers (i.e., reactors), and
allowed to decompose anaerobically under warm, moist conditions designed to accelerate decomposition.
The reactors were seeded with a small amount of well-decomposed refuse containing an active population of
methane-producing microorganisms (the "seed"), to ensure that methane generation was not limited due to
an insufficient population of microorganisms. To promote degradation, water was cycled through each
reactor. Nitrogen and phosphorus were then added so that methane generation would not be limited by a lack
of these nutrients.
The reactors were allowed to run for periods varying from three months to two years. The
experiment ended for each reactor when one of two conditions were met: (1) no measurable methane was
being emitted (i.e., any methane that was being emitted was below the detection limits of the analytical
equipment), or (2) a curve generated mathematically from an analysis of the reactor's prior methane
generation indicated that the reactor had produced at least 95 percent of the methane that it would produce if
allowed to run forever.
84 Barlaz's work was funded by EPA's Air and Energy Engineering Research Laboratory under the
supervision of Susan Thorneloe.
87 Barlaz, Morton, "Measurement of the Methane Potential of the Paper, Yard Waste, and Food Waste
Components of Municipal Solid Waste," unpublished paper, Department of Civil Engineering, North Carolina State
University, Raleigh, NC, 1994.
94 DRAFT - March 1997
-------
Barlaz measured the amount of methane generated during the experimental period, and subtracted
the amount of methane attributable to the seed in order to obtain the amount of methane generated by the
material being tested. At the end of the experiment, he opened the reactors, drained the leachate, dried and
weighed the contents, and analyzed the percentage composition of cellulose, hemicellulose, and lignin (and,
for food scraps, protein) in the remaining contents. He then measured the percentage of total volatile solids
in the remaining contents. This amount included the cellulose, hemicellulose, lignin, and protein, and any
other carbon-containing components such as waxes and tannins.
We used the experimental results that Barlaz obtained to first estimate the amount of each carbon-
containing component remaining that was attributable to the seed,88 and then to estimate the amount of
carbon for each material that remained. We assumed that the experiment reflected landfill conditions, and
that organic carbon remaining undegraded in the reactors would also remain undegraded over the long term
in landfills, i.e., it would be sequestered.
Methane Generation: Experimental Data and Adjusted Values
The amount of methane generated by each type of organic material (after deducting the methane
attributable to the seed), is shown in column "b" of Exhibit 7-1.89
As a check on his experimental results, Barlaz estimated the amount of methane that would have
been produced if all of the cellulose, hemicellulose, and protein from the waste material that was
decomposed during the experiment had been converted to equal parts of methane and carbon dioxide
(methane-producing microorganisms generate equal amounts, by volume, of methane and carbon dioxide
gas).90 Barlaz referred to this amount as the material's "methane potential." He then calculated the
percentage of the methane potential for each material accounted for by the sum of (1) the measured methane
generation, and (2) the amount of methane that could be formed from the carbon in the leachate that was
removed from the reactor, and from the carbon in the refuse that remained in the reactor at the end of the
experiment.91 The resulting percentages of the methane potential accounted for are shown in column "c" of
Exhibit 7-1. Methane potential not accounted for could be due to either (1) leaks of methane, (2)
measurement error, or (3) carbon in the cell mass of microorganisms (which was not measured).
Methane recovery was below 85 percent of the "methane potential" for four materials: office paper,
food scraps, leaves, and branches. In using Barlaz's data, we needed to make a choice regarding how to
allocate this missing carbon. We chose to assume that some of it had been converted to microorganism cell
mass, and the remainder had been degraded. Barlaz postulated a higher methane yield based on assumptions
that (1) five percent of the carbon in cellulose and hemicellulose (and protein in the case of food scraps) that
was degraded was converted into the cell mass of the microbial population, and (2) 90 percent of the carbon-
88 Dr. Barlaz tested seed alone to be able to control for the amount of methane generation and carbon
sequestration that was attributable to the seed.
89 Personal communication from Dr. Morton Barlaz to Clare Lindsay, U.S. EPA, July 20,1995.
90 Ibid. Lignin was not considered in this check because cellulose, hemicellulose, and protein account for
nearly all of the methane generated.
91 Note that any carbon that was converted to cell mass in microorganisms was not considered in this
calculation.
DRAFT -- March 1997 95
-------
containing compounds that were degraded but not converted to cell mass were converted to equal parts of
methane and carbon dioxide. The "corrected yields," based on these assumptions, are shown in column "d"
of Exhibit 7-1.
We decided, in consultation with Barlaz, to use the "corrected yields" for leaves, branches, and
office paper because we believed that these values were more realistic than the measured yields.92
The methane values that we used for each material (either the measured yield, or the "corrected"
yield) are shown again in column "f' of Exhibit 7-1. In order to maintain consistent units with the other parts
of our analysis, we converted the units for methane generation from milliliters per dry gram of waste, to
metric tons of carbon equivalent (MTCE) per wet ton of waste.93 The resulting values are shown in column
"g" of Exhibit 7-1. The value for yard trimmings is a weighted average of the values for grass, leaves, and
branches, based on an assumption that yard trimmings are composed of 50 percent grass, 25 percent leaves,
and 25 percent branches (on a wet weight basis).94
Carbon Sequestration: Experimental Data and Calculations
To estimate the amount of carbon sequestered when each material is landfilled, we used data from
Barlaz1 s experiments on the amount of carbon-containing components in (1) the material placed in the
reactors at the start of the experiment, (2) the "seed" added to each reactor, and (3) the material remaining in
each reactor at the end of the experiment. We also used data on the total dry weight of both the sample of
waste material and the seed placed in each reactor.
In essence, we used these data to estimate carbon sequestration by calculating the amount of carbon
remaining in each reactor at the end of the experiment, and then subtracting the amount of carbon remaining
that was attributable to the seed. The difference between the two values is the amount of carbon from the
waste material that remained in the reactor, undecomposed, at the end of the experiment. Because the
conditions in the reactor simulated landfill conditions, approximately this amount of carbon would be
sequestered if the material were landfilled. For each material, we averaged the carbon sequestration values
for the four reactors. Our results are shown in Exhibit 7-2; our approach to estimating carbon sequestration
is described in more detail in the explanatory notes accompanying that exhibit.
92 For food scraps, however, even though the methane potential recovery percentage was lower than 85
percent, we used the measured yield, as shown in column "b." We made this choice for food scraps because the
"corrected yield" for food scraps was greater than the maximum possible yield (shown in column "e" of the
exhibit). Barlaz had calculated the maximum possible yield for each material based on the methane yield if all of the
cellulose, hemicellulose, and protein in the material: (1) decomposed and (2) were converted to equal parts of
methane and carbon dioxide.
93 To make the conversion, we used the ratio of dry weight to wet weight for each material and a global
wanning potential of 24.5 for methane.
94 As noted in chapter 5, this professional judgment estimate for the percentage composition of yard
trimmings (as a national average) was provided by Nick Artz of Franklin Associates, Ltd. in a telephone
conversation with William Driscoll of ICF Incorporated, November 14,1995.
96 DRAFT -- March 1997
-------
Exhibit 7-1
Methane Yield and Methane Generation for Solid Waste Components
(a)
Material
Newspaper
Office Paper
Corrugated Boxes
Food Scraps
Grass
Leaves
Branches
Yard Trimmings
Mixed MSW
(b)
Average Measured
Methane Yield
(ml per dry gm)
74.2
217.3
152.3
300.7
144.3
30.5
62.6
92.0
(C)
Percentage of
"Methane Potential"
Accounted For
98.0
55.5
87.7
77.4
89.3
75.2
82.8
97.6
(d)
"Corrected"
Methane Yield
(ml per dry gram)
NA
346.0
NA
386.2
NA
56.0
76.3
NA
(e)
Maximum Possible
Methane Yield
(ml per dry gram)
239.4
398.2
279.7
357.6
153.2
108.0
224.9
157.6
(f)
Selected
Methane Yield
(ml per dry gm)
742
346.0
152.3
300.7
144.3
56.0
76.3
92.0
(9)
Selected
Methane Yield
(MICE /wet ton)
0.302
1.408
0.626
0.391
0.250
0.194
0.198
0.223
0.319
Explanatory note: Because Dr. Barlaz based his measurements on the dry weight of each material, the units throughout most of this exhibit
are also provided in terms of methane generation per unit of dry weight. But because MSW is measured in its wet form (i.e., in tons of MSW "as is;"
not dried), the units are converted to methane generation per wet ton in the last column of the exhibit. Note that no values are shown for yard
trimmings until the last column of the exhibit, because the yard trimmings value is based on the estimated proportions of grass, leaves, and branches
contained in yard trimmings, weighed on a wet basis.
-------
Exhibit 7-2
Carbon Sequestration for Solid Waste Components
(a)
Material
Newspaper
Office Paper
Corrugated boxes
Food Scraps
Grass
Leaves
Branches
Yard Trimmings
Mixed MSW
(b)
Ratio of Carbon
Sequestration
to Dry Weight
(gmC/drygm)
0.40
0.04
0.26
0.31
0.22
0.42
0.38
0.20
(o)
Ratio of Dry
Weight to
Wet Weight
0.94
0.94
0.95
0.30
0.40
0.80
0.60
0.80
(d)
Ratio of Carbon
Sequestration
to Wet Weight
(gm C/wet gm)
0.38
0.04
0.25
0.09
0.09
0.34
0.23
0.19
0.16
(e)
Amount of
Carbon
Sequestered
(MTCEperWetTon)
0.34
0.04
0.23
0.08
0.08
0.30
0.21
0.17
0.14
Explanatory notes:
(1) To determine the amount of carbon remaining in each reactor at the end of the experiment, we used (1) the
measured amount of each carbon-containing component cellulose, hemicellulose, lignin, protein (for food scraps),
and total volatile solids remaining in each reactor,95 and (2) estimates of the amount of carbon in each carbon-
containing component. To estimate the second data element, we used the following data sources:
Cellulose: we used the chemical formula to determine that cellulose is 44.4 percent carbon (on a mass basis).
Hemicellulose: There are various types of hemicellulose; we used a composite chemical formula to estimate
that the carbon content is 45.5 percent.96
Lignin: The Encyclopedia of Chemical Technology reports the "average elementary analysis for wood lignin"
for coniferous species to be 63.8 percent carbon.98
Protein: There are many types of protein; we used a carbon content of 53.8 percent from a composite
composition for protein.99
93 Personal communication from Dr. Morton Barlaz to Randy Freed, ICF Incorporated, July 19, 1995.
96 U.S. Environmental Protection Agency, Estimate of Methane Emissions from U.S. Landfills
(Washington, D.C.: U.S. EPA) September 1994, p. 6.
98 Kirk-Othmer, Encyclopedia of Chemical Technology, Third Edition (New York: John Wiley & Sons)
1981, Vol. 14, p. 298.
99 Barlaz, Morton A. and Robert K. Ham, "The Use of Mass Balances for Calculation of the Methane
Potential of Fresh and Anaerobically Decomposed Refuse," in Proceedings from the GRCDA 13th Annual
International Landfill Gas Symposium March 27-29,1990 (Silver Spring, MD: GRCDA - The Association of Solid
Waste Management Professionals) 1990, p. 232.
98
DRAFT-March 1997
-------
Explanatory notes for Exhibit 7-2 (continued):
Other carbon-containing components (i.e., total volatile solids minus cellulose, hemicellulose, lignin, and
protein): These components consist largely of waxes and tannins, so we used the average carbon content from
two relevant compounds - the wax in Douglas fir bark (58.7 percent carbon) and tannic acid (53.65 percent
carbon).100 The average of these values was 56 percent carbon.
(2) We next estimated the amount of carbon remaining in each reactor that was attributable to the seed. Barlaz used the
same type of seed for all materials except food scraps. For the predominant type of seed, we first determined the average
amount of carbon remaining in the four seed reactors (i.e., those reactors containing only seed) at the end of the
experiment: 14.6 percent of the total dry weight of the seed entering the reactor. For all materials but food scraps, we used
this percentage, together with the dry weight of the seed entering each reactor, to estimate the amount of carbon
sequestration in each reactor that was attributable to seed.101
(3) We used the estimates for the amount of carbon remaining in each reactor, and the amount attributable to seed, to
develop an average ratio, for each material, of (1) the amount of carbon sequestered (after deducting the amount
attributable to seed) to (2) the dry weight of waste material placed in the reactor. These ratios are shown in column "b" of
the exhibit.
(4) Because MSW is typically measured in terms of its wet weight, we needed to convert the ratios for carbon sequestered
as a fraction of dry weight to carbon sequestered as a fraction of wet weight. To do this, we used the estimated ratio of dry
weight to wet weight for each material. These ratios are shown in column "c" of the exhibit. For most of the materials, we
used data from an engineering handbook.102 For grass, leaves, and branches, we used data provided by Barlaz.103 To
determine the ratio of carbon sequestration to wet weight of each material, we multiplied the values in columns "b" and
"c." The results are shown in column "d."
(5) For consistency with the overall analysis, we converted the carbon sequestration values for each material to units of
metric tons of carbon equivalent (MTCE) sequestered per short ton of waste material landfilled. The resulting values are
shown in column "e" of the exhibit.
(6) We also used Barlaz's data for mixed MSW to estimate the percentage of carbon in mixed MSW that is sequestered in
a landfill. Specifically, we used data on (1) the amount of carbon in each sample of mixed MSW initially placed in a
reactor (based on the amounts of cellulose, hemicellulose, lignin, and other volatile solids in each sample, and the
percentage carbon content of these components), and (2) the amount of carbon remaining in each mixed MSW reactor at
the end of the experiment. The average percentage of carbon sequestered in the four samples of mixed MSW was 50
percent.
100 The molecular formula for Douglas fir bark wax is from Kirk-Othmer, Encyclopedia of Chemical
Technology, Third Edition, (New York: John Wiley & Sons) 1984, vol. 24, p. 470. The molecular formula for tannic
acid is from Merck & Co., The Merck Index, Eleventh Edition (Rahway, NJ: Merck & Co., Inc.) 1989, p. 9027.
101 For the seed used for food scraps, Dr. Barlaz collected data on the amount of cellulose, hemicellulose,
and lignin remaining in the seed reactors at the end of the experiment, but did not collect this data on the protein
remaining in the seed reactors. Thus, for food scraps, we assumed that the amounts of protein remaining in each
reactor that had originated from the waste and from the seed, respectively, were in the same proportion as the
amounts of protein in waste and protein in seed placed in the reactor initially.
102 Tchobanoglous, George, Hilary Theisen, and Rolf Eliassen, Solid Wastes: Engineering Principles and
Management Issues (New York: McGraw-Hill Book Co.) 1977, pp. 58 and 60.
103 Dr. Morton Barlaz, personal communication with Joanne Colt, ICF Incorporated, April 25,1995.
DRAFT - March 1997 99
-------
7.2 FATES OF LANDFILL METHANE: CONVERSION TO CO2, EMISSIONS, AND
FLARING OR COMBUSTION WITH ENERGY RECOVERY
In this analysis, we accounted for (1) the conversion in the landfill of some portion of landfill
methane to CO2, and (2) the capture of methane, either for flaring or for combustion with energy recovery
(in either case, the captured methane is converted to CO2).104 Exhibit 7-3 presents this analysis.
The exhibit begins with the methane generation per wet ton of each material, which is shown in
column "b" (the values were simply copied from the last column of Exhibit 7-1). The next three sections of
the spreadsheet calculate net GHG emissions from methane generation for each of three categories of
landfills: (1) landfills without LFG recovery, (2) landfills with LFG recovery that flare LFG, and (3) landfills
with LFG recovery that generate electricity from the LFG. The second to last section of the spreadsheet
shows the expected percentage of landfills in each category in 2000. The final column shows the weighted
average GHG emissions from methane generation across all types of landfills in 2000.
To estimate MSW methane emissions from each category of landfill, we first needed to estimate the
percentage of landfill methane that is oxidized near the surface of the landfill. We estimated that 10 percent
of the landfill methane that is generated is either chemically oxidized or converted by bacteria to CO2, and
that the remaining 90 percent is available for atmospheric methane emissions.105
To estimate MSW methane emissions from landfills with LFG recovery, we used an estimate that
these landfills will have an average LFG recovery efficiency of 85 percent by 2000.106 In Exhibit 7-3 we
show the percentage of methane that will not be captured by these landfills (i.e., 15 percent) in two columns
(once for each of the two categories of landfills with LFG recovery).
To estimate net GHG emissions from methane generation for landfills that combust LFG to generate
electricity, we estimated the utility GHG emissions avoided per unit of methane combusted for energy
recovery (our calculations to develop this estimate are shown in Exhibit 7-4.
We also projected the percentage of MSW disposed in each category of landfill in 2000. We
estimated that by the year 2000, when large landfills with substantial LFG emissions will have been required
to recover LFG, 58 percent of all landfill methane will be generated at landfills with recovery systems, and
42 percent will be generated at landfills without LFG recovery.107 Of the 58 percent of all methane generated
at landfills with LFG recovery, 91 percent (or 53 percent of all methane) is expected to be generated at
landfills that use LFG to generate electricity, and 9 percent (or 5 percent of all methane) at
104 The CO2 that is emitted is not counted as a GHG because it is biogenic in origin (as described in Section
1.6).
105 U.S. EPA, Office of Air and Radiation, Anthropogenic Methane Emissions in the United States:
Estimates for 1990 (Washington, D.C.: U.S. EPA) April 1993, page 4-20.
106 Memorandum from Cindy Jacobs of the U.S. EPA Atmospheric Pollution Prevention Division to
Michael Podolsky of the U.S. EPA Office of Policy, Planning and Evaluation, July 25,1995.
107 Based on data on (1) year 2000 MSW landfill methane generation of 64.5 million MTCE, (2) year 2000
landfill methane recovery of 40.0 million MTCE, and (3) projected year 2000 landfill methane recovery efficiency
of 85 percent (all from the memorandum from Cindy Jacobs, op cit.).
100 DRAFT - March 1997
-------
Exhibit 7-3
Net GHG Emissions from Methane Generation
(a)
Material
Newspaper .
Office Paper
Core. Boxes
Food Scraps
Grass
Leaves
Branches
Yard Trimmings
MxedMSW
(b)
CH4
Generation
(MTCB
Wet Ton)
0.302
1.408
0.626
0.391
0.250
0.194
0.198
0.223
0.319
Methane from Landfills Without
Methane Recovery
(c)
Percentage
Of Methane
Not ,
Oxidized
toC02
90%
90%
90%
90%
90%
90%
90%
90%
90%
(
-------
Explanatory notes for Exhibit 7-3 (continued):
(3) To estimate the net GHG emissions from landfills with LFG recovery that generate electricity, we estimated the methane emissions and subtracted
the avoided utility C02 emissions when methane is used to generate electricity. The calculations (and values) for column "i" are identical to those for
column "g." Columns "j" and "k" account for avoided utility CO2 emissions; the value in column "j" comes from Exhibit 7-4. Column "1" equals
column "i" minus column "k."
(4) The expected percentage of methane from each type of landfill in 2000 is shown in columns "m," "n," and "o." The value for landfills with LFG
recovery are based on the values of 58 percent of methane being generated at landfills with LFG recovery, and 9 percent of these landfills flaring the
LFG rather than generating electricity.
(5) Finally, to estimate the total net GHG emissions from landfilling of each type of material, we used the GHG emissions for each category of landfill
and the percentage of methane generated at each type of landfill to develop a weighted, average across all landfills. The results are shown in column
"p."
-------
Exhibit 7-4
Calculation to Estimate Utility GHGs Avoided
Through Combustion of Landfill Methane
Step
Metric tons CH^MTCE CH4
Grams Chymetric ton CH4
Cubic ft. CHVgram CH4
BTUs/cubicft.CH4
kWh electricity generated/BTU
kWh electricity delivered/kWh electricity generated
BTUs/kWh electricity delivered
Kg. utility C avoided/BTU delivered electricity
Metric Tons avoided utility C/kg utility C
Ratio of MTCE avoided utility C per MTCE CH4
Value
0.15
1.00E+06
0.05
1,000
0.00008
0.91
3,412
5.112E-05
0.001
0.09
Source
1/((1 2/44)*24.5): Global warming potential of 24.5 for CH4
Physical constant
1/20: 20 grams per cubic foot of methane at standard temperature and pressure
"Opportunity for LF Gas Energy Recovery in FL [Working Draft]," USEPA/OAR May 95, p. 2-1 1
1/13,000: from "Opportunity" report p. 2-11, assumes use of internal combustion engines
U.S. DOE, EIA, "Annual Energy Review 1993 (Washington, DC: DOE/EIA) July 1994, p. 252
Physical constant
51 .12 kg C/mmBTU del'd electricity, from carbon coefficients table
1000 kg per metric ton
Product from multiplying all factors
-------
landfills that flare LFG.108 By basing our analysis on projected LFG recovery by the year 2000 (and the
projected LFG recovery efficiency in 2000), we avoided double-counting of GHG reductions between
programs that reduce landfilling and programs that increase recovery of landfill methane.
Our results are shown in the final column of Exhibit 7-3. The materials with the highest rates of net
GHG emissions from methane generation - office paper, corrugated boxes, food scraps, and newspaper -
also have the highest gross methane generation, as shown in column "b" of Exhibit 7-3. The recovery of
methane at landfills reduces the methane emissions for each material in proportionate amounts, but does not
change the ranking of materials by methane emissions. Yard trimmings and mixed MSW have the lowest
rates of net GHG emissions from methane generation.
The three sections of the exhibit providing GHG emissions estimates for each category of landfill (in
columns "d," "g," and "1") may be used by local MSW planners to estimate GHG emissions from MSW
from a given community. For this purpose, one should add to the values in the appropriate column the
estimated transportation GHG emissions (the national average used in this study is 0.01 MTCE per ton), and
subtract estimated carbon sequestration (as shown for each material in Exhibit 7-2).
In a separate analysis, EPA has estimated that in 2000, when most landfill methane will be captured,
landfills will emit 24.5 million MTCE of methane,109 as compared to 68.2 million MTCE in 1994."°
Anthropogenic Carbon Dioxide Emissions from Transportation of Wastes to a Landfill
We next estimated the anthropogenic carbon dioxide emissions from transporting waste materials to
a landfill. We began with estimates provided by Franklin Associates, Ltd. for the amount of diesel fuel
required per ton of waste material for (1) collecting and transporting the material to a landfill (297,000
BTUs), and (2) operating the landfill equipment (231,000 BTUs).111 We converted these estimates to units of
metric tons of carbon equivalent (MTCE) per short ton of yard trimmings, based on a carbon coefficient of
0.0208 MTCE per million BTUs of diesel fuel. This resulted in an estimate of 0.01 MTCE of anthropogenic
CO2 emissions per short ton of material landfilled.
73 NET GHG EMISSIONS FROM LANDFILLING
To determine the net GHG emissions from landfilling each material (in absolute terms), we began
with the net GHG emissions from methane generation, subtracted carbon sequestration, and added
transportation CO2 emissions. The results are shown in Exhibit 7-5.
Only one material, office paper, has net GHG emissions when landfilled of at least 0.1 MTCE per
ton (because its methane emissions far exceed its landfill carbon sequestration). Food scraps, corrugated
108 Memorandum from Cindy Jacobs, op cit.
109 Ibid.
110 U.S. EPA, Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-1994, EPA 230-R-96-006
(November 1995) p. 80.
111 Franklin Associates, Ltd., The Role of Recycling in Integrated Solid Waste Management to the Year
2000 (Stamford, CT: Keep America Beautiful), p. 1-5.
104 DRAFT - March 1997
-------
Exhibit 7-5
Net GHG Emissions from Landfilling
(a)
Material
Newspaper
Office Paper
Corrugated Cardboard
Aluminum Cans
Steel Cans
HOPE
LDPE
PET
Food Scraps
Grass
Leaves
Branches
Yard Trimmings
Mixed MSW
(b)
Net GHG
Emissions from
CH« Generation
(MTCE/WetTon)
0.13
0.58
0.26
0.00
0.00
0.00
0.00
0.00
0.16
0.10
0.08
0.08
0.09
0.13
(c)
Net Carbon
Sequestration
(MTCE/WetTon)
0.34
0.04
0.23
0.00
0.00
0.00
0.00
0.00
0.08
0.08
0.30
0.21
0.17
0.14
(d)
GHG Emissions
from
Transportation
(MTCE/WetTon)
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
(e)
Net GHG
Emissions from
Landfilling
(MTCE/WetTon)
-0.21
0.56
0.04
0.01
0.01
0.01
0.01
0.01
0.09
0.04
-0.21
-0.12
-0.07
0.00
Explanatory notes: The net GHG emissions from methane generation, shown in column "b" of the exhibit, were simply copied from column "p" of
Exhibit 7-3. The carbon sequestration values in column "c" of this exhibit were copied from column "e" of Exhibit 7-2. The GHG emissions from
transportation in column "d" of this exhibit were developed as described in the text. The net GHG emissions for each material were determined by
summing the values in columns "b" and "d," and then subtracting the value in column "c." The results are shown in column "e."
-------
boxes, and grass have lower net GHG emissions; the metals and plastics have small transportation-related
emissions (0.01 MTCE per ton). All the other materials appear to have carbon sequestration that exceeds
their methane emissions, and so result in negative net GHG emissions when landfilled.112
7.4 LIMITATIONS
Perhaps the most important caveat is that the analysis is based on a single set of laboratory
experiments conducted by Dr. Morton Barlaz. While researchers other than Barlaz have conducted
laboratory studies that track the degradation of mixed MSW, Barlaz's experiments were the only ones we
identified that rigorously tested different materials individually. Barlaz is recognized as an expert on the
degradation of different fractions of MSW under anaerobic conditions, and his findings with respect to the
methane potential of mixed MSW are within the range used by landfill gas developers. Nevertheless, given
the sensitivity of the landfill results to estimated methane generation and carbon sequestration, we recognize
that more research is needed in this area.
Another important caveat is that we estimated that 58 percent of MSW landfill methane generated in
the year 2000 would be generated at landfills with LFG recovery systems. This would be an increase from
the estimated 17 percent of landfill methane generated at landfills with LFG recovery in 1995. The net GHG
emissions from landfilling each material are quite sensitive to the LFG recovery rate. Because of the high
global warming potential for methane, small changes in the LFG recovery rate by the year 2000 could have a
large effect on the net GHG impacts of landfilling each material, and on the ranking of landfilling relative to
other MSW management options. The effects of different rates of LFG recovery by the year 2000 are shown
in Exhibit 7-6. Column "b" of the exhibit shows net GHG emissions at the 1995 recovery rate of 17 percent.
The remaining columns show net GHG emissions at increasing LFG recovery rates, up to a 60 percent
recovery rate (rounded up from 58 percent, the rate projected for 2000). As the exhibit shows, the net GHG
emissions for landfilling mixed MSW are positive at lower rates of recovery, and turn negative only when
the LFG recovery rate exceeds 50 percent. At the local level, the GHG emissions from landfilling MSW are
quite different depending on whether the local landfill has LFG recovery, as shown in Exhibit 7-3.
On the national level, this analysis was based on LFG recovery levels expected by the year 2000.
Because some landfill methane emissions prior to 2000 will not be recovered at the year 2000 levels,
keeping organic materials out of landfills prior to the year 2000 will have GHG benefits in excess of those
estimated here. A related point is that the analysis does not account for the timing of methane generation,
which can occur for years after waste is landfilled. The values listed in this chapter represent total methane
generated, over time, per ton of waste landfilled. To the extent that LFG recovery rates shift dramatically
over time, these shifts are not reflected in the analysis. In addition, landfills with LFG recovery will be
permitted, under EPA regulations, to remove the LFG recovery equipment when three conditions are met:
(1) the landfill is permanently closed, (2) LFG has been collected continuously for at least 15 years, and (3)
the landfill emits less than 50 metric tons of nonmethane organic compounds per year.113 Although the
removal of LFG recovery equipment will permit methane from closed landfills to escape into the
112 Note that the components of yard trimmings - grass, leaves, and branches - have substantially different
net GHG emissions when landfilled. Grass has small positive net GHG emissions, while leaves and branches have
substantial negative GHG emissions (due to landfill carbon sequestration).
113 Federal Register, Vol. 61, No. 49, p. 9907.
106
DRAFT - March 1997
-------
Exhibit 7-6
Net GHG Emissions from Landfilling
Sensitivity Analysis: Varying the Percentage of Waste Disposed
at Landfills with Methane Recovery
(a)
Material
Newspaper
Office Paper
Corrugated Cardboard
Food Scraps
Grass
Leaves
Branches
Yard Trimmings
Mixed MSW
(b)
17%
of waste disposed
at landfills with
LFG recovery
-0.10
1.04
0.26
0.22
0.12
-0.15
-0.05
0.01
0.11
(c)
30%
of waste disposed
at landfills with
LFG recovery
-0.14
0.89
0.19
0.18
0.09
-0.17
-0.07
-0.01
0.07
(d)
40%
of waste disposed
at landfills with
LFG recovery
-0.16
0.77
0.14
0.15
0.07
-0.19
-0.09
-0.03
0.05
(e)
50%
of waste disposed
at landfills with
LFG recovery
-0.19
0.65
0.08
0.11
0.05
-020
-0.10
-0.05
0.02
(f)
60%
of waste disposed
at landfills with
LFG recovery
-0.21
0.54
0.03
0.08
0.03
-0.22
-0.12
-0.07
-0.01
Mote: Of the methane that is captured, we assumed that 9% is flared and 91% is recovered for energy.
Explanatory note: In every case we assumed that for methane that is captured, the proportions flared versus combusted for energy recovery are the
same as shown in columns "n" and'
-------
atmosphere, the amounts of methane emitted should be relatively small, because of the relatively long time
period required for LFG collection before LFG recovery equipment is removed.
It is also likely that ongoing shifts in the use of landfill cover and liner systems could influence the
rate of methane generation and collection. As more landfills install effective covers and implement controls
to keep water and other liquids out, conditions will be less favorable for degradation of organic wastes. Over
the long term, it is possible that this will result in a decrease in methane generation and an increase in carbon
sequestration. Moreover, Dr. Barlaz believes that the methane yields from his laboratory experiments are
likely to be higher than methane yields in a landfill, because the laboratory experiments were designed to
generate the maximum amount of methane possible. If the methane yields used in this analysis are higher
than yields in a landfill, the net GHG emissions from landfilling organic materials would be lower than
estimated here.
We assumed that once wastes are disposed in a landfill, they are never removed. In other words, we
assumed that landfills are never "mined." (A number of communities have mined their landfills - removing
and combusting the waste - in order to create more space for continued disposal of waste in the landfill.) To
the extent that landfills are mined in the future, it is incorrect to assume that carbon sequestered in a landfill
will remain sequestered. For example, if landfilled wastes are later combusted, the carbon that was
sequestered in the landfill will be oxidized to CO2 in the combustor.
For landfilling of yard trimmings (and other organic materials), we assumed that all carbon storage
in a landfill environment is incremental to the storage that occurs in a non-landfill environment. In other
words, we assumed that in a baseline where yard trimmings are returned to the soil, all of the carbon is
decomposed relatively rapidly (i.e., within several years) to CO2, and there is no long-term carbon storage.
This approach differs somewhat from the one used in the chapter on composting, where we estimated the
incremental carbon storage without regard to the absolute value of carbon storage in the baseline. To the
extent that long-term carbon storage occurs in the baseline, the estimates of net GHG emissions reported
here are understated.
Finally, our spreadsheet analysis is limited by the assumptions that were made at various steps in the
analysis, as described throughout this chapter. The key assumptions that have not already been discussed as
limitations are the assumptions used in developing "corrected" methane yields for organic materials in
MSW. Because of the high global warming potential of methane, a small difference between estimated and
actual methane generation values would have a large effect on the GHG impacts of landfilling, and on the
ranking of landfilling relative to other MSW management options.
108
DRAFT -- March 1997
-------
8. COMPARISON OF OPTIONS
Earlier chapters of this report examined the GHG emissions from (1) raw materials acquisition and
manufacturing (and forest carbon sequestration for paper products) and (2) waste management. In other
words, the earlier chapters analyzed the two separate components of a complete life cycle analysis; the first
part of the cycle being manufacturing, the second, waste management.114 This chapter combines information
from the earlier chapters to present a picture of the full life cycle GHG emissions associated with
manufacturing, and then managing as waste, each of the materials considered in this analysis.
This chapter compares the life cycle GHG emissions for the five municipal solid waste (MSW)
management options analyzed in this report (source reduction, recycling, composting, combustion, and
landfilling), for each of the eight manufactured materials considered. (These materials are newspaper, office
paper, corrugated boxes, aluminum cans, steel cans, and three types of plastic - LDPE, HOPE, and PET.) In
addition, this chapter presents the GHG emissions from waste management for composting, combusting, or
landfilling food scraps, yard trimmings, and mixed MSW (a full life cycle analysis for food scraps and yard
trimmings is not appropriate because neither is manufactured; for mixed MSW - a composite of dozens of
manufactured materials - developing a weighted average GHG emission rate is beyond the scope of this
project.)
Using the estimates contained in this chapter of the full life cycle GHG emissions from (1) raw
materials acquisition and manufacturing, and (2) waste management, one can compare any waste
management option, for a given material, to any other waste management option. In this chapter we have
provided an exhibit that compares each of four waste management options - source reduction, recycling,
composting, and combustion - to landfilling (which is currently the most commonly used waste management
option).
Our results show that source reduction has lower GHG emissions than all other options for all eight
of the manufactured materials considered, if one assumes that source reduction will displace the use of
virgin inputs. After source reduction, recycling has the next lowest GHG emissions. If one assumes that
source reduction displaces the current mix of virgin and recycled inputs to manufacturing, recycling results
in greater GHG reductions than source reduction for aluminum cans. Composting of food scraps and yard
trimmings has GHG emissions in the same range as combustion and landfilling. Finally, comparing
landfilling and combustion, landfilling has lower GHG emissions than combustion for newspaper and
plastics; combustion has lower GHG emissions than landfilling for office paper and corrugated cardboard, as
well as steel cans (since steel is recovered for recycling at most combustors); and emissions are similar for
aluminum cans, food scraps, yard trimmings, and mixed MSW.
114 The one exception is Chapter 4, on source reduction, which for the sake of clarity showed both (1) the
GHG emissions from the manufacturing stage (i.e., zero, except for credits for forest carbon sequestration for paper
products) and (2) the waste management GHGs (i.e., zero, because the material was not produced).
DRAFT - March 1997 109
-------
8.1 FULL LEFJE CYCLE GREENHOUSE GAS EMISSIONS FOR EACH WASTE
MANAGEMENT OPTION
This section presents the full life cycle GHG emissions for each waste management option, for each
material considered. These emissions are shown in five exhibits that recapitulate the GHG emissions and
sinks analyzed in detail in earlier chapters.
Exhibit 8-1 shows the life cycle GHG reductions associated with source reduction. In brief, the
exhibit shows that source reduction of paper products results in GHG reductions (due to forest carbon
sequestration), whereas source reduction of other materials results in no GHG emissions. This same exhibit
was presented earlier in this report as Exhibit 4-1.
Exhibit 8-2 shows the life cycle GHG emissions associated with manufacturing and then recycling
each of the materials considered. The values in the first column of the exhibit show the GHG emissions
associated with the initial manufacture of each material, using the current mix of virgin and recycled inputs.
The next four columns show the GHG reductions associated with using recycled inputs in place of virgin
inputs when the material is remanufactured. The final column, which simply sums the others, shows the
overall life cycle GHG implications of manufacturing and then recycling each material. The net carbon
values for paper products are negative, primarily due to the forest carbon sequestration benefits of recycling.
The net carbon value for aluminum is also negative, because the GHG emissions avoided by displacing 100
percent virgin inputs in the remanufacturing stage are larger than the GHG emissions from manufacturing
using the current mix of virgin and recycled inputs. The net carbon values for steel cans and plastics are
positive, indicating that manufacturing and then recycling the products results in net GHG emissions.
Exhibit 8-3 presents the life cycle GHG emissions from manufacturing and then combusting each of
the materials considered. As the exhibit shows, manufacturing and then combusting each material results in
net GHG emissions for nearly all of the manufactured materials. For food scraps and yard trimmings, the
GHG emissions are slightly negative.
Exhibit 8-4 shows the GHG emissions from manufacturing and then landfilling each material. The
final column shows the net GHG emissions from landfilling.
We have not provided an exhibit for composting. As described in Chapter 5, we performed a
bounding analysis and concluded that GHG emissions from composting are zero or close to zero.
8.2 COMPARISONS OF THE WASTE MANAGEMENT OPTIONS
The full life cycle GHG emissions for each waste management option and each material are
compared in Exhibit 8-5. As the exhibit shows, when the full life cycle, including manufacturing, is
considered, source reduction dominates all other options (i.e., it has lower GHG emissions than any other
option on a ton-per-ton basis) for all materials except aluminum cans. For aluminum cans, recycling has
lower GHG emissions than source reduction only because the recycling value in this exhibit assumes that
increased recycling results in displacement of virgin inputs, whereas the source reduction value assumes that
source reduction results in displacement of the current mix of virgin and recycled inputs. If source reduction
were assumed to displace virgin inputs, it would have lower GHG emissions than recycling in both cases.
110
DRAFT -- March 1997
-------
Exhibit 8-1
Greenhouse Gas Emissions for Source Reduction
(MTCE/Ton of Material Source Reduced)
Material
Newspaper
Office Paper
Corrugated Cardboard
Aluminum Cans
Steel Cans
HOPE
LDPE
PET
GHG Emissions
from Raw Materials
Acquisition and
Manufacturing
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Change in Forest Carbon Storage
(Minus sign indicates incremental carbon storage)
Source Reduction
Displaces Current
Mix of Virgin and
Recycled Inputs
-0.48
-0.53
-0.44
0.00
0.00
0.00
0.00
0.00
Source
Reduction
Displaces Virgin
Inputs .
-0.73
-0.73
-0.73
0.00
0.00
0.00
0.00
0.00
Net GHGs
Waste
Management
GHGs
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Source reduction
Displaces Current
Mix of Virgin and
Recycled Inputs
-0.48
-0.53
-0.44
0.00
0.00
0.00
0.00
0.00
Source
Reduction
Displaces Virgin
Inputs
-0.73
-0.73
-0.73
0.00
0.00
0.00
0.00
0.00
-------
Exhibit 8-2
Recycling of Post-Consumer Material
(GHG Emissions in MTCE/Ton)
Material
Newspaper
Office Paper
Corrugated Cardboard
Aluminum Cans
Steel Cans
HOPE
LDPE
PET
Food Scraps
Yard Trimmings
Mixed MSW
Manufacturing
GHG
(Current Mix of Inputs)
0.49
0.53
0.40
2.96
0.87
0.72
0.87
0.98
NA
NA
NA
Recycled Input
Credit*
Process Energy
GHG
-0.14
-0.08
0.03
-2.66
-0.57
-0.31
-0.44
-0.58
NA
NA
NA
Recycled Input
Credit*
Trans. Energy
GHG
0.01
-0.01
0.00
-0.07
-0.01
-0.02
-0.01
-0.02
NA
NA
NA
Recycled Input
Credit*
Process Non-
Energy GHG
0.00
0.00
0.00
-1.24
0.00
-0.06
-0.06
-0.03
NA
NA
NA
Forest Carbon
Sequestration
-0.73
-0.73
-0.73
0.00
0.00
0.00
0.00
0.00
NA
NA
NA
Net Carbon
(Post-
Consumer)
-0.37
-0.29
-0.30
-1.01
0.30
0.34
0.36
0.35
NA
NA
NA
Material that is recycled post-consumer is then substituted for virgin inputs in the production of new products. TOs credit
represents the difference in emissions that results from using recycled inputs rather than virgin inputs. It accounts for
loss rates in collection, processing, and remanufacturing. Recycling credit is based on weighted average of closed and open
loop recycling for office paper and corrugated cardboard. However, all other estimates are only for the products themselves.
-------
Exhibit 8-3
Combustion of Post-Consumer Material
(GHG Emissions in MTCE/Ton)
Material
Newspaper
Office Paper
Corrugated Cardboard
Aluminum Cans
Steel Cans
HOPE
LDPE
PET
Food Waste
Yard Waste
Mixed MSW
Manufacturing
GHG
(Current Mix)
0.49
0.53
0.40
2.96
0.87
0.72
0.87
0.98
NA
MA
NA
Transportation
to Combustion
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
CC-2 from
Combustion
0.00
0.00
0.00
0.00
0.00
0.75
0.75
0.51
0.00
0.00
0.10
N20 from
Combustion
0.01
0.01
0.01
0.00
0.00
0.01
0.01
0.01
0.01
0.01
0.01
Avoided
Utility GHG
-0.11
-0.09
-0.10
0.00
0.00
-0.26
-0.26
-0.13
-0.03
-0.04
-0.07
Avoided
GHG: Steel
Recovery
0.00
0.00
0.00
0.00
-0.42
0.00
0.00
0.00
0.00
0.00
-0.01
Net Carbon
(Post-
Consumer)
0.40
0.46
0.32
2.97
0.47
1.22
1.38
1.38
-0.01 *'
-0,02 *
0.04 *
* Excludes manufacturing GHG emissions.
-------
Exhibit 8-4
Landfilling of Post-Consumer Material
(GHG Emissions in MTCE/Ton)
Material
Newspaper
Office Paper
Corrugated Cardboard
Aluminum Cans
Steel Cans
HOPE
LDPE
PET
Food Scrap
Yard Trimmings
Mixed MSW
Manufacturing
GHG
(Current Mix of Inputs)
0.49
0.53
0.40
2.96
0.87
0.72
0.87
0.98
NA
MA
NA
Transportation
to Landfill
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
Net
Landfill
CH4
0.13
0.58
0.26
0.00
0.00
0.00
0.00
0.00
0.16
0.09
0.13
Landfill
Carbon
Sequestration
0.34
0.04
0.23
0.00
0.00
0.00
0.00
0.00
0.08
0.17
0.14
Net Carbon
(Post-
Consumer)
0.28
1.09
0.44
2.97
0.88
0.73
0.88
0.99
0.09*
-0.07 *
0.00*
* Excludes manufacturing GHG emissions.
-------
Exhibit 8-5
Greenhouse Gas Emissions from Source Reduction and MSW Management Options
(Assuming Initial Production Using the Current Mix of Virgin and Recycled Inputs)
(MTCEATon)
Material
Newspaper
Office Paper
Corrugated Cardboard
Aluminum Cans
Steel Cans
HOPE
LDPE
PET
Food Scraps
Yard Trimmings
Mixed MSW
Net Source
Reduction Emissions
-0.48
-0.53
-0.44
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Net Recycling
Emissions*
-0.37
-0.29
-0.30
-1.01
0.30
0.34
0.36
0.35
NA
NA
NA
Net Composting
Emissions**
NA
NA
NA
NA
NA
NA
NA
NA
0.00
0.00
NA
Net Combustion
Emissions*
0.40
0.46
0.32
2.97
0.47
1.22
1.38
1.38
-0.01
-0.02
0.04
Net Landfilling
Emissions*
0.28
1.09
0.44
2.97
0.88
0.73
0.88
0.99
0.09
-0.07
0.00
'Includes emissions from the initial production of the material being managed, except for food waste, yard waste, and mixed MSW.
"There is considerable uncertainty in our estimate of net GHG emissions from composting; the values of zero are plausible values
based on assumptions and a bounding analysis.
-------
After source reduction, Exhibit 8-5 shows that recycling has the next lowest GHG emissions.
Recycling has lower GHG emissions than combustion or landfilling for all eight of the manufactured
materials analyzed in this study.
Between combustion and landfilling, the strategy with the next lowest GHG emissions differs for
different materials. Combustion has lower GHG emissions than landfilling for office paper, corrugated
cardboard, and steel cans, because office paper and corrugated cardboard generate a substantial amount of
methane when landfilled, and steel is recovered for recycling at most MSW combustors. Landfilling has
lower GHG emissions than combustion for plastics and newspaper, because combustion of plastic results in
substantial nonbiogenic CO2 emissions, and landfilling of newspaper results in substantial carbon
sequestration. The net GHG emissions from combustion and landfilling are similar (given the range of
uncertainty in the values developed in this analysis) for aluminum cans. Composting is a management option
for food scraps and yard trimmings; the net GHG emissions from composting, combusting, or landfilling
these materials are similar, given the uncertainty in the analysis.
The ordering of these options is affected by (1) the GHG inventory accounting methods, which do
not count CO2 emissions from sustainable biogenic sources, but do count emissions from sources such as
plastics, and (2) a series of assumptions on sequestration, future use of methane recovery systems, recovery
system efficiency, ferrous metals recovery, and avoided utility fossil fuels. On a site-specific basis, the
ordering of results between a combustor and a landfill could be different from the ordering provided here,
which is based on national average results.
The full life cycle GHG emissions for each of the first four waste management strategies source
reduction, recycling, composting, and combustion - are compared to the GHG emissions from landfilling in
Exhibit 8-6. This exhibit shows the GHG values for each of the first four management strategies, minus the
GHG values for landfilling. This exhibit is provided because landfilling is often viewed as the baseline
waste management strategy. With this exhibit, one may easily compare the GHG emissions from other waste
management options to the GHG emissions from landfilling.
We close with a final note about the limitations of these GHG emission estimates, and their potential
uses. We based our analysis on what we believed to be the best available data; where necessary, we made
assumptions that we believe are reasonable. However, the accuracy of the estimates is limited by the
assumptions made, and by limitations in the data sources. We have discussed these limitations throughout
this report.
We anticipate four potential applications for the GHG emission estimates provided here. First,
organizations that are interested in quantifying GHG emission reductions due to source reduction or
recycling may use these estimates for that purpose; EPA may soon use these estimates as the basis for
developing guidance for voluntary reporting of GHG reductions, as authorized by Congress in Section
1605(b) of the Energy Policy Act of 1992. Second, the estimates may be useful for evaluation of MSW
management options on a national, regional, state, or local basis. Third, EPA plans to use the estimates to
evaluate its progress in reducing US GHG emissions by promoting source reduction and recycling through
programs such as WasteWi$e and Unit-Based Pricing, as part of the US Climate Change Action Plan.
Finally, this report may also assist other countries involved in developing GHG emissions estimates for their
solid waste streams.
116
DRAFT - March 1997
-------
Exhibit 8-6
Greenhouse Gas Emissions of MSW Management Options Compared to Landfilling
(MTCEm>n)
Material
Newspaper
Office Paper
Corrugated Cardboard
Aluminum Cans
Steel Cans
HOPE
LDPE
PET
Food Scrap
Yard Trimmings
Mixed MSW
Source Reduction
Net Carbon
Minus Landfilling Net Carbon
Current Mix of Inputs
-0.76
-1.62
-0.89
-2.97
-0.88
-0.73
-0.88
-0.99
NA
NA
NA
100% Virgin Inputs
-1.07
-1.85
-1.15
-5.52
-1.13
-0.73
-0.92
-1.19
NA
NA
NA
Recycling Net Carbon
Minus Landfilling
Net Carbon
-0.65
-1.38
-0.74
-3.98
-0.58
-0.39
-0.52
-0.64
NA
NA
NA
Composting Net Carbon
Minus Landfilling
Net Carbon
NA
NA
NA
NA
NA
NA
NA
NA
-0.09
0.07
NA
Combustion Net Carbon
Minus Landfilling
Net Carbon
0.12
-0.63
-0.12
0.00
-0.42
0.50
0.50
0.38
-0.10
0.05
0.04
-------
------- |