MEMORANDUM
TO:			Air and Radiation Docket EPA-HQ-OAR-2012-0401
FROM:		EPA Office of Transportation and Air Quality
DATE:			July 1, 2014
SUBJECT:	Support for Classification of Biofuel Produced from Waste Derived Biogas as Cellulosic Biofuel and Summary of Lifecycle Analysis Assumptions and Calculations for Biofuel Produced from Waste Derived Biogas
This memorandum provides additional support for EPA's determination that renewable fuel produced using biogas from landfills, separated MSW digesters, wastewater treatment facility digesters, and agricultural digesters should be classified as cellulosic biofuel.  It also provides information on the key input assumptions used for the lifecycle analysis of renewable fuel produced from waste derived biogas.  Data sources are presented to promote transparency for our lifecycle analysis.  This memorandum also describes the calculations made for the lifecycle analysis of renewable fuel produced from waste derived biogas and discusses, in depth, the choice of baseline used in the lifecycle analysis.

I. Summary
   
      In the June 14, 2013 Notice of Proposed Rulemaking (NPRM), EPA proposed to allow compressed natural gas (CNG), liquefied natural gas (LNG) and renewable electricity from landfill biogas to qualify for cellulosic renewable identification numbers (RINs), and requested comment on whether fuel derived from other sources of biogas should also be allowed to qualify for cellulosic RINs.  The proposal to treat landfill biogas as a cellulosic biofuel feedstock was based on available data and was supported by a memo to the docket (EPA-HQ-2012-0401). This document expands upon the original memo to the docket and provides a summary of additional information provided through the public comment process or developed in response to comments received.  Through the comment process, and additional EPA research, EPA has identified additional data that supports a finding that fuel derived from biogas from MSW digesters, municipal wastewater treatment facility digesters, and agricultural digesters is also predominantly derived from cellulosic components.  

II. Determination of the Cellulosic Content of Waste Derived Biogas

A. Biogas from Landfills
      
      For the purposes of this memorandum, the terms "cellulosic" or "cellulosic content" are intended to refer to the sum of the cellulose, hemicellulose, and lignin components of a feedstock.  However, when calculating the "adjusted cellulosic content" of a feedstock, which is a percentage value, the cellulosic content is divided by the volatile organic fraction because only this portion is reactive and can contribute to the production of biogas.  For municipal solid waste (MSW) in landfills, the non-organic fraction is composed of metal, cement and other inorganic materials that will not yield methane in the landfill. Review and analysis of available data characterizing MSW landfill material and its associated biogas indicate that the organic fraction of MSW is predominately cellulosic and the biogas generated from MSW is predominantly derived from the cellulosic components.  
      
      The Barlaz research cited in the June 2013 NPRM most directly attempts to answer the question of what percent of the MSW landfill biogas is derived from the cellulosic components of the MSW.  Results from this study are outlined in Table 1 and Figures 1 and 2, and show that methane does not appear until day 41, indicating that the initial decomposition was aerobic, involving loss of sugars to carbon dioxide and water.  During this time period, about 20% of the MSW was lost, including fractions of cellulose and hemicellulose (Figure 1).  Once the originally present oxygen was consumed, the reactors became anaerobic and methane production began (Figures 1 and 2).  Large amounts of both cellulose and hemicellulose were converted during this period (Figure 1), and lignin accumulated in the residue (Figure 2).  Overall, 71% of the cellulose and 77% of the hemicellulose were converted during the reaction, roughly half aerobically and half anaerobically (Table 1).  These data show that a substantial fraction of the cellulosic content of MSW is converted.  In fact, based on the abundance of different biochemicals in the MSW and theoretical methane potentials for the different components (e.g., cellulose, proteins), the authors determined that 91% of the total methane in this sample was derived from cellulosic components.  This research indicates that a predominant portion of the biogas from MSW landfills comes from anaerobic digestion of the cellulosic content of MSW, namely cellulose and hemicellulose. 

Table 1  -  Chemical composition of municipal solid waste (MSW) from different studies, the percent of each type of compound lost in a degradation experiment and the percent of total methane potential from the different components.


Barlaz et al. (1989)
 

                     % Composition (average of 14 studies)
                                % Composition 
                          % Lost during decomposition
                         % of total methane potential
Cellulose
                                   38 +- 14
                                      51
                                      71
                                      74
Hemicellulose
                                    8 +- 2
                                      12
                                      77
                                      15
Lignin
                                    16 +- 6
                                      15
                                       8
                                     0[*]
Protein
                                    3 +- 1
                                       4
                                     n.d.
                                       8
Volatile Solids
                                   69 +- 16
                                      79
                                     n.d.
                                      100
Average % "Cellulosic"[**]
                                 90 +- 30[***]
                                   100[****]
                                       
                                      91
[*]The authors assumed that lignin had a methane potential of zero because lignin is known to react very slowly under anaerobic conditions.
[**]Calculated relative to the measured concentration of volatile solids.
[***]Standard error calculated by propagation of error.
[****]Note that this value is artificially high due to measurement uncertainties


	
Figure 1  -  Chemical Composition of degrading MSW versus time: 2 % of original mass remaining for total mass, cellulose, hemicelluloses and lignin, with cumulative methane yield plotted on the right axis.  Methane first appeared (at very low concentration) on day 41.


Figure 2  -  Chemical Composition of degrading MSW versus time: percent composition of MSW over time, with the area between curves corresponding to the percent of total mass in that component.  The "other" material (white) includes metals, plastics and cement, as well as organic materials such as proteins and starches.  The circles show the cumulative production of methane in the reaction containers.

      Recent studies have also confirmed that the organic fraction of MSW is still predominantly comprised of cellulose, hemicellulose, and lignin. Specifically, the organic fraction of MSW has been largely unaffected by changes in waste management trends such as increased recycling.  Moreover, there is a strong and direct correlation between the amount of cellulosic materials present in the MSW and the amount of biogas that can be produced  -  demonstrating that anaerobic digestion is converting cellulosic biomass to biogas.   Based on the average of 14 studies, MSW contains 38% cellulose, 8% hemicellulose and 16% lignin, with smaller amounts of proteins, sugars and other organic materials (Table 1).  Since the average organic fraction of the MSW is 69%, the average adjusted cellulosic content of the landfill MSW is 90%. 
      
      MSW composition varies considerably depending on, for example, the amount of yard waste versus food waste versus metal waste deposited.  It is worth noting that the studies examined were published over 27 years (1982 to 2009), during a time period in which recycling expanded considerably, leading to substantial changes in the types of materials deposited in landfills.  Despite these changes, the cellulosic content of the organic fraction of MSW, the portion capable of contributing to biogas production, has not changed substantially during this time.[1] The data reviewed by EPA indicate that the vast majority of volatile solids in landfills are composed of cellulosic components[1].
      
	Eleazer et al. (1997) analyzed the decomposition of a variety of components of MSW, including leaves, paper, food and MSW itself.  They found a correlation between the percent of cellulose and hemicellulose in the MSW and the yield of methane from the different samples, with methane yield increasing as the cellulose and hemicellulose increased.  This study demonstrated that biogas yield is proportional to the cellulosic content of MSW, and supports the conclusion that landfill biogas is predominately derived from cellulosic components. 

	While the fraction of the biogas being generated from cellulosic components may vary slightly from location to location depending on the composition of the MSW, the volatile solids in MSW appear to always contain a large proportion of cellulosic materials, and a large proportion of the biogas generated will be derived from cellulose and hemicellulose.  Therefore, we conclude that fuel derived from landfill biogas is predominantly cellulosic, and that all volumes of fuel made from landfill biogas qualify for cellulosic biofuel RINs.  

B. Biogas from Separated MSW Digesters, Agricultural Digesters, and Municipal Wastewater Treatment Facility Digesters
      For the reasons described below, EPA is extending its assessment that biogas derived from MSW landfills is predominantly cellulosic to include biogas from separated MSW digesters, municipal wastewater treatment facility digesters, and agricultural digesters.    
      Organic wastes in the United States that may be available for, and susceptible to, anaerobic decomposition to produce biogas can be generally characterized as falling into one of four categories.  The first includes the organic fraction of municipal solid waste (OFMSW), the second is biosolids from municipal wastewater treatment facilities, the third is agriculture waste (including animal manure), and the fourth is wastes that do not fall into any of the first three categories. Almost all of the organic wastes anaerobically digested in the U.S. fall into one of the first three categories, and their adjusted cellulosic content is on average above 75%.  (Table 4).  Given our understanding of anaerobic digestion in MSW landfills, described above, we are confident that anaerobic digesters processing predominantly cellulosic materials, including materials from any of these first three general waste categories described above, would produce biogas that is predominantly cellulosic in origin. For our final rule we have identified three digester types that we believe will process wastes that are predominantly cellulosic: separated MSW digesters, municipal wastewater treatment facility digesters and agricultural digesters. Digesters processing the fourth category of organic wastes described above may accept large quantities of non-cellulosic materials, such as waste fats, oils and greases or industrial food wastes (e.g., food and beverage production wastes that are primarily composed of sugar or starch). While the materials comprising the fourth category may meet EPA's definition of renewable biomass, we do not have enough information at this time to determine that the biogas from these materials would be derived from predominantly cellulosic components. Therefore, biogas digesters primarily processing non-cellulosic materials such as those listed in the fourth category are considered non-cellulosic, and transportation fuels derived from biogas produced at these digesters may only qualify for cellulosic RINs for the cellulosic portions of the biogas and may produce advanced biofuel RINs for the non-cellulosic portions. 
      Finally, determinations regarding the cellulosic content of various types of organic wastes discussed in this memo apply only in the context of production of biogas through anaerobic digestion. Our determination that fuel derived from certain feedstocks that are anaerobically digested to produce biogas should be considered of cellulosic origin does not apply to fuels produced from the same organic wastes that are converted to fuel using a biomass conversion process other than anaerobic digestion. One of the reasons for this limitation is that the anaerobic digestion process is known to convert cellulosic components of organic wastes to biogas, and this cannot be extended to all biomass treatments.       
1. Separated MSW Digesters
      Current regulations allow generation of cellulosic RINs for the entire biogenic fraction of separated MSW, and require testing to determine what portion of the finished fuel is made from the biogenic portion of separated MSW. The test prescribed is a carbon-14 radio dating test (ASTM Method D-6866). EPA determined in the March 2010 RFS rule that biogas is not formed from non-biogenic compounds in landfills, and thus it was unnecessary to require the ASTM method in the context of landfill biogas.
      The biogenic portion of MSW may be processed in a separated MSW digester to produce biogas and biogas-derived transportation fuels.  As discussed in the March 2010 rule, the organic fraction of MSW generally consists of yard waste and heterogeneous post-consumer food waste.  The biochemical conversion processes by which organic material is converted to biogas, known as anaerobic digestion, are thermodynamically identical in both landfills and anaerobic digesters.  The only differences between the anaerobic conversion processes in these systems are kinetic (reaction rate).  While the environmental conditions in landfill systems favor the anaerobic digestion of organic matter thermodynamically, the conversion process is not optimized kinetically. In fact, landfills can be viewed as anaerobic digesters that have not been optimized for efficient reaction kinetics capable only of passively generating biogas over long periods of time. Waste digesters, by contrast, actively perform the anaerobic biochemical processes that landfills allow passively because they are designed to optimize system kinetics. Therefore, waste digesters more rapidly and efficiently convert biomass to biogas. For example, while processing the same amount of identical biomass, a waste digester system would generate more useable biogas on a substantially shorter time-scale than a landfill system.  However, despite kinetic differences, the bioconversion processes occurring in landfills and waste digesters are essentially identical. We considered  -  as we did for landfill biogas  -  whether to require biogas producers to use ASTM Method D - 6866 to identify the biogenic versus non-biogenic fractions of the fuel. However, since landfill gas and digester biogas are produced in an identical fashion, and it is not formed from non-biogenic compounds in landfills or anaerobic digesters, no purpose would be served in using the ASTM method in the context of MSW digester biogas. Therefore, EPA finds that fuels made from biogas derived from digesters processing separated MSW are biogenic, and qualify for cellulosic biofuel RINs for the entire volume of fuel produced. .   
      EPA notes that the CAA requires that renewable fuel be produced from "renewable biomass" as that term is defined in CAA 211(o)(1)(I).  MSW is not identified in the statute as renewable biomass.   However, EPA determined that "separated MSW" as defined in 40 CFR 80.1426(f)(5)(i)(C), qualifies as renewable biomass for purposes of renewable fuel production provided that it is collected according to a plan submitted and approved by EPA pursuant to 40 CFR 80.1426(5)(ii).  Material that could have been disposed of in a MSW landfill, but was diverted from it, does not automatically qualify on that basis as separated MSW.  Rather, such waste must qualify as a different type of renewable biomass, such as animal waste, separated food waste, or separated yard waste to be an eligible feedstock for renewable fuel production.   A "separated MSW digester" for purposes of today's rule is a digester processing separated MSW as defined in EPA regulations pursuant to an anaerobic digestion process.  
2. Agricultural Digesters
      There are approximately 250 agricultural digester projects in the U.S.  Of these agricultural digesters about 40% are co-digesters processing a mixture of animal manure, on-farm crop residues, and small amounts of other biogenic wastes.  There are no consistent data to portray the exact mixture of material being digested in individual digesters on a continuous basis because both the material composition and relative proportions of digested material varies over time. However, in aggregate, over 90% of the material processed in agricultural digesters is animal manure, and the remaining portion consists mainly of plant residues and other on-farm wastes that are predominately cellulosic in origin such as crop residues and separated yard waste. This section provides information on both animal manure and on-farm crop residues  -  the main components processed in agricultural digesters.  
      Data used to estimate the aggregate composition of animal manure were obtained from a comprehensive review published by the Pacific Northwest National Laboratory covering four typically digested categories of animal manure (dairy, cattle, swine, and poultry).  Although these studies reported the cellulose, hemicelluloses, and lignin on a total solids basis, they did not calculate an adjusted cellulosic content based on the fraction of the material that could be converted to biogas. To obtain an adjusted cellulosic content, we normalized the reported values based on the volatile organic fraction.  When possible (e.g., if the studies analyzed the same categories of manure) and if comparable (e.g., provided the breakdown of cellulosic components in the same way), differing data were averaged. A weighted average based on the type and volume of animal manure digested, based on data from AgSTAR was used to estimate the cellulosic content for animal manure in aggregate.  The adjusted cellulosic content was calculated by summing the cellulosic fractions (cellulose, hemicellulose, and lignin) and dividing by the volatile fraction capable of contributing to biogas production. The volatile fraction capable of contributing to biogas was assumed to be the volatile solids minus the organic nitrogen which would be converted to nitrogen gas or other inorganic nitrogen species rather than methane or carbon dioxide.  
Table 2  -  Data used to assess the cellulosic content of animal manure 


Animal Manure
 


                                    Cattle
                                      (%)
                                    Poultry
                                     (%) 
                                     Swine
                                      (%)
                                 Animal Manure
                                 (Aggregate %)
                                    Source
Cellulose
                                      25
                                      22
                                      22
                                      25
                                      (1)
Hemicellulose
                                      23
                                      12
                                      14
                                      21
                                       
Lignin
                                      19
                                       7
                                       7
                                      17
                                       
Volatile Solids
Organic Nitrogen
                                      84
                                       3
                                      62
                                       3
                                      76
                                       3
                                      82
                                       3
                                      (2)
Relative Volume
                                      79
                                       1
                                      20
                                      --
                                      (3)
Adjusted Cellulosic %
                                       
                                       
                                      80
                                       
      
      The aggregate cellulosic content of animal manure (highlighted in green in Table 2) was found to be 25% cellulose, 21% hemicellulose, and 17% lignin, which means the cellulosic components equal 63%. The volatile solids and organic nitrogen were found to be 82% and 3% respectively, which results in a volatile fraction of 79% capable of contributing to biogas production.  Based on these values, the average adjusted cellulosic content of the biomass treated in animal manure digesters is estimated to be 80%. While this estimate assumes that the digester material would be exclusively animal manure, it is common practice to co-digest mixtures of animal wastes, farm/crop residues and other wastes.  As described in a separate memo to the docket, materials that we have determined meet the RFS regulatory definition of "crop residue" have on average an adjusted cellulosic content of 84%.  The addition of such crop residues would therefore typically increase the overall cellulosic content of the biomass being treated in manure digesters.  Similarly, EPA has determined that "separated yard waste", as that term is defined in 40 CFR 80.1426(5)(i), should be deemed to be composed entirely of cellulosic materials. See 40 CFR 80.1426(5)(i)(A).  Such materials may also be added to waste digesters in an agricultural setting, perhaps in combination with manure and/or crop residues.  We have defined the term "agricultural waste digester" as those digesters processing animal manure, crop residue and separated yard waste, and we have specified in Table 1 to 80.1426 that fuel produced from biogas collected at such digesters qualifies as cellulosic biofuel for which cellulosic RINs may be generated for the entire volume.  EPA understands that incidental de minimis quantities of materials that are not predominantly cellulosic, or materials that are not renewable biomass, may be introduced into agricultural digesters.   This practice would not disqualify the resulting fuel from qualifying as renewable fuel under the RFS program providing that such incidental de minimis feedstock contaminants are either impractical to remove or are related to customary feedstock production and transport.  See 40 CFR 18.1426(f)(1).     
3. Municipal Wastewater Treatment Facility Digesters
      Municipal wastewater treatment facility (MWTF) digesters are anaerobic digesters that process the sludge, undissolved solids, and biosolids derived from municipal wastewater whether or not the facility is owned by a municipality.  The composition of wastes processed in MWTF tend to be heterogeneous, but are generally composed of human wastes and highly fibrous tissue products.   Wastes that are processed in municipal wastewater treatment facilities are often treated aerobically, but more and more are being configured to employ anaerobic digestion as part of the facility's primary treatment system, allowing biogas to be produced.  
      The data used to assess the cellulosic content of wastewater solids is presented in Table 3. 
Table 3  -  Data used to assess the cellulosic content of wastewater solids


Characteristics of Wastewater Solids 
 



Wastewater Type
                                      % C
                                      % H
                                      % L
                             %  Convertible Solids
                                     % AC
                                     % AH
                                     % AL
                                    Source
Primary
                                     29.3
                                       
                                       
                                     72.9
                                     40.2
                                       
                                       
                                Champagne, 2009
Primary
                                     11.5
                                       
                                     26.5
                                     74.1
                                     15.5
                                       
                                     35.8
                                 Cheung, 1997
Primary
                                      17
                                       
                                       9
                                     62.6
                                     27.2
                                       
                                     14.4
                                    Wang, 
                                     2008
AS
                                     13.8
                                       
                                       
                                     78.7
                                     17.5
                                       
                                       
                              Champagne, 2009[4]
AS
                                       9
                                      16
                                      13
                                     49.3
                                     18.3
                                     32.5
                                     26.4
                                     Wang,
                                    2008[6]
Biosolids
                                      14
                                      19
                                       8
                                     49.0
                                     28.6
                                     38.8
                                     16.3
                                     Wang,
                                    2008[6]

                                       
                                   Averages
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                     % AC
                                     % AH
                                     % AL
                                     % ACC
Average of AS and Biosolids Data
                                       
                                      22
                                      36
                                      21
                                      79
C =; cellulose; H = hemicellulose; L = Lignin; AC = adjusted cellulose; AH = adjusted hemicellulose; AL = adjusted lignin; AS =activated sludge; ACC = adjusted cellulosic content 
      While there are substantial data characterizing the solids composition of municipal wastewater, there are somewhat less data characterizing the composition of materials entering the digesters specifically.  We chose to use the subset of peer-reviewed data (highlighted in green in Table 3) that analyzes the activated sludge and biosolids for the purposes of calculating the average adjusted cellulosic content of materials that would be expected to enter wastewater treatment facility digesters.  This is because the broader data set did not include data for the fraction of hemicellulose in primary sludge and it does not adequately represent the material entering the digester.  As will be discussed forthwith, the material entering a digester is better represented by activated sludge and biosolids.
      The wastewater treatment process is designed to oxidize dissolved organics, remove nutrients such as aqueous nitrogen and phosphorus species, and separate suspended solids. A typical process will involve several steps as part of three broader systems: the primary treatment system (pretreatment, primary clarification and anaerobic digestion), the secondary treatment system (aeration, clarification, internal sludge recycle, and disinfection) and the tertiary treatment system (nutrient removal and solids handling). Anaerobic digestion at wastewater treatment facilities typically occurs as part of the process' primary treatment; however data characterizing primary sludge does not adequately represent the material entering the digesters because the primary sludge is separated from the wastewater and partially treated prior to entering the digester.  
      The material that enters the digester includes the undissolved solids that are recovered from the primary clarification tank and the solids that are allowed to settle out in a secondary clarification tank.  Therefore, the data for activated sludge and biosolids are more likely to represent the material entering the wastewater facility digesters.  In addition, the data related to activated sludge and biosolids is more consistent and comparable, and therefore provide a more robust estimate of the cellulosic content. 
      The data considered (highlighted in green in Table 3) offered consistent and comparable data, including values for cellulose, hemicellulose, and lignin individually, and the volatile solids representing the organic fraction of the wastewater solids that can be converted via anaerobic digestion was reported in each of these studies. The average adjusted cellulosic content was obtained by dividing the reported cellulosic fraction by the volatile organic fraction that can be converted to methane.   Based on the data for activated sludge and biosolids, the material entering the digesters is determined to be on average composed of 13% cellulose, 22% hemicellulose, and 13% lignin, for a total of 48% cellulosic components.  The volatile solids and organic nitrogen were found to be 64% and 3% respectively, which results in an organic fraction of 61%. When the cellulosic component is divided by the organic fraction that can be converted to methane, the average adjusted cellulosic content of the material used to generate the biogas through anaerobic digestion from wastewater treatment facilities is, on average, 79% (Table 4).  Therefore, we have determined that fuels made from biogas collected at MWTF digesters is predominantly cellulosic, such that the entire volume of fuel produced qualifies for cellulosic biofuel RINs.  
Table 4  -  Estimated average composition of anaerobically digested cellulosic biomass in the United States


Aggregate Composition of Digested Biomass 
 

                                    % OFMSW
                            % wastewater biosolids
                                % animal manure
Cellulose
                                      38
                                      13
                                      25
Hemicellulose
                                       8
                                      22
                                      21
Lignin
                                      16
                                      13
                                      17
Volatile Solids
Organic Nitrogen
                                      69
                                      --
                                      64
                                       3
                                      82
                                       3

                                       
                                       
                                       
                                       
Adjusted Cellulosic  Content
                                      90
                                      79
                                      80
Source: This table represents the aggregate data from Table 1, Table 2, and Table 3.

III. Lifecycle Analysis of Biogas-Derived Renewable Fuel Pathways
      Under anaerobic conditions, organic wastes naturally decompose to produce biogas, and controlled management of organic waste digestion can reduce GHG emissions that would have occurred alternatively.  The following sections describe the assumptions, analytical methods and conclusions regarding the lifecycle GHG emissions associated with renewable fuel produced from waste derived biogas. 
A. Assumptions used in the Lifecycle Analysis
Table 5 outlines key assumptions used in determining the lifecycle GHG emissions associated with renewable fuel produced from waste derived biogas.  EPA used these values in our lifecycle analysis as described in the subsequent sections.




Table 5  - Key Assumptions for the Lifecycle Analysis of Renewable Fuel Produced from Waste Derived Biogas
Category
Assumption
Source
Notes




On-Site Emissions from Combustion of Biogas (considers CH4 and N2O only)
Emissions from Flaring
1004 g CO2-eq/ mmBtu biogas
GREET1_2011
Value given is for Renewable Natural Gas
Emissions from Stationary Reciprocating Engines
9634 g CO2-eq/ mmBtu biogas

Value given is for Biogas
Emissions from Turbines
554 g CO2-eq/ mmBtu biogas

Value given is for Natural Gas; Emissions are assumed to be the same for Biogas
Emissions from Combined Cycle Gas Turbines
554 g CO2-eq/ mmBtu biogas

Value given is for Natural Gas; Emissions are assumed to be the same for Biogas
% of Gas-to-Energy Projects using Reciprocating Engines
70.6% of total capacity (in MW)
EPA LMOP Database of Operational Projects
Percentages exclude microturbines, which are used in 0.3% of projects
% of Gas-to-Energy Projects using Turbines
23.8% of total capacity (in MW)

Includes both gas and steam turbines; excludes microturbines
% of Gas-to-Energy Projects using Combined Cycle Gas Turbines
5.6% of total capacity (in MW)

 
Weighted Average  Emissions for Gas-to-Energy Projects
6965 g CO2-eq/ mmBtu biogas
 Calculated from data referenced in this table
 
 
 
 
 
Upstream Emissions
 
 
 
U.S. Average Electricity Production
219,823 g CO2-eq/ mmBtu electricity
RFS Final Rule
 







Table 5, continued
Renewable Electricity Generation
 
 
Efficiency of Electricity Generation from Biogas
11,700 Btu biogas/kWh
EPA LMOP LFGE Benefits Calculator
Weighted average of gas-to-energy projects
Electricity-Ethanol Conversion
22.6 kW-h/gal ethanol
RFS Final Rule[6]
 
Parasitic Loss Efficiencies for Reciprocating Engine-Generator Sets
93%
EPA LMOP Data
 
Parasitic Loss Efficiencies for Turbine-Generator Sets
88%

Use this value for Combined Cycle Gas Turbines as well
Electricity Demand for Biogas Collection Blowers
0.002 kWh/ft[3]

For flaring and gas-to-energy projects
U.S. Average Transmission and Distribution Losses
6.6%
U.S. EIA State Electricity Profiles
 
CNG and LNG
Energy Use for Gas Clean Up
0.030 MJe/MJ biomethane
GREET1_2011

Energy Use for Compression
0.016 MJe/MJ biomethane


Energy Use for Liquefaction
0.043 MJe/MJ biomethane



B. Calculations for the Lifecycle Analysis of Renewable Electricity Produced from Waste Derived Biogas.
	The lifecycle analysis of renewable electricity produced from waste derived biogas focused on emissions associated with production of the fuel.  We did not consider any emissions associated with transportation of the renewable electricity (although losses are accounted for), and no tailpipe emissions, so the only significant GHG emissions are derived from fuel production.
      The first step in determining the lifecycle GHG emissions associated with production of renewable electricity from waste derived biogas was to determine how much electricity (in mmBtu) could be produced from a given amount of biogas (in mmBtu) based on values for the efficiency of typical electricity generation at landfills or waste digesters provided by EPA's Landfill Methane Outreach Program (LMOP; Table 5) Table V.B.-1 in the Preamble outlines this calculation.  We then apportioned the electricity generation proportionately according to the generation technology used (engines, turbines, combined cycle turbines) using the % of total capacity values given above.  We used the parasitic loss efficiency values for engines and turbines (above) to determine the size of the parasitic losses for each category of electricity generation and subtracted these values from the generation values above.  We then summed the amounts of electricity remaining after these losses to determine the net amount of electricity produced from the biogas after parasitic losses.  We subtracted the amount of this energy that would be used to power the biogas collection system.  We also calculated how much of the electricity leaving the facility would be lost during transmission and distribution using data from EIA (see Table 5) and subtracted this amount from the total electricity.  In the end, we determined that 0.236 mmBtu electricity would be delivered to the consumer for each mmBtu biogas combusted.
	There were two components of electricity production that figured in the lifecycle analysis: on-site emissions and upstream emissions (Table V.B.-2 in the Preamble).  Based on the relationship between electricity production and biogas combustion, the emissions factor for flaring, and the weighted average emissions factor for gas-to-energy projects listed above (Table 5), we determined the amount by which on-site emissions at the biogas source would change.  We used the value for the U.S. average GHG emissions from electricity production to calculate the amount of emissions in the flaring baseline scenario that are due to the grid electricity used to power the biogas collection system.  This amount was assigned as a credit, since these emissions would be eliminated upon installation of a gas-to-energy project because renewable electricity would power the biogas collection system.  Adding together the values for on-site and upstream emissions yields lifecycle GHG emissions of 12 kg CO2-eq/mmBtu electricity, which is an 87% reduction in emissions from the petroleum gasoline baseline (Table V.B.-2 in the Preamble).  Because the drivetrains of electric vehicles are roughly 3 times as efficient as those of traditional internal combustion cars, we also calculated GHG emissions taking this improved efficiency into account.  For these calculations, we determined the GHG emissions for each lifecycle stage per mmBtu of a "fuel equivalent", assuming that it would take three times the energy of a liquid fuel to drive a car as far as a given amount of electricity.  The lifecycle GHG emissions from electricity produced from waste derived biogas considering this factor are 4 kg CO2-eq/mmBtu fuel equivalent, or a 96% reduction in GHG emissions from the petroleum baseline.  According to both of these calculations, renewable electricity produced from waste derived biogas could qualify as either cellulosic or advanced biofuel.
C. Calculations for the Lifecycle Analysis of Renewable CNG and LNG Produced from Waste Derived Biogas.
	As part of the March 2010 RFS final rule EPA determined that biogas produced from landfills, sewage waste treatment plants and manure digesters met the 50% lifecycle GHG requirements to be eligible to generate advanced RINs as part of the RFS program.  In this rulemaking we indicate that CNG and LNG produced from biogas from landfills, MSW digesters, wastewater treatment facility digesters, and agricultural digesters should be classified as cellulosic biofuel if it meets the 60% lifecycle GHG requirements.  Similar to renewable electricity, the only significant GHG emissions associated with the lifecycle analysis of CNG and LNG produced from waste derived biogas are derived from fuel production.  
	We considered the energy associated with cleaning the biogas and compressing it to either CNG or LNG based on the factors outlined in Table 5.  We compared this to the flaring baseline as discussed above and determined that CNG and LNG produced from waste derived biogas would result in over an 80% reduction in GHG emissions compared to the petroleum baseline.  According to these calculations, CNG and LNG produced from waste derived biogas could qualify as either cellulosic or advanced biofuel.
D. Choice of Baseline for Waste Derived Biogas Treatment
	When conducting a lifecycle analysis of greenhouse gas emissions for biofuels produced from waste derived biogas, it is important to choose an appropriate baseline for comparison.  The baseline assumption is important because, as per the approach outlined in the March 2010 RFS rule,[6] the lifecycle calculations are based on a scenario approach.  The results are determined based on a comparison of the biogas-based biofuel scenario to a baseline scenario that would have happened if the biogas was not used to produce transportation fuels.  There are two main components of the baseline to consider, how use of waste-derived biogas for transportation fuels would impact waste disposal, and how it would impact the current or alternative treatment of the biogas.  
	If waste management methods were impacted by use of biogas for transportation fuel, there could be indirect GHG emissions impacts.  However, waste management policies are typically controlled by state and local governments, and there are many unique factors that influence these decisions.  We have not seen any evidence or data to suggest that the RFS in general has had or will have a substantial impact on existing waste disposal practices across the U.S., and therefore we believe that there will not be significant GHG impacts associated with the biogas-based pathways adopted in this rule.  Therefore no GHG impacts from waste disposal changes are included in the GHG analysis.  This is consistent with the March 2010 RFS final rule which concluded that municipal solid waste has no agricultural or land use change GHG emissions associated with its production.  
	Biogas can be handled and/or processed in one of three primary ways: venting to the atmosphere, collection and flaring, and collection with use in a gas-to-energy or gas-to-fuel project.  Theoretically, any of the above scenarios could serve as a baseline scenario for our lifecycle analysis since all three occur at biogas source sites in the U.S.  We first examine how the prevalence of flaring and gas-to-energy projects at landfills has changed over time and what impacts this has had on total methane emissions from landfills.  We then outline the implications of using other possible baselines and discuss why we propose to use flaring as the baseline.
Historical uses for landfill biogas
	Figure 3, below, shows how the uses of landfill biogas in the U.S. changed over a 20-year period from 1990 to 2010 based on data from the Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2010.  During this time period, the amount of methane produced in landfills increased by 50%, however, the amount of methane released from these same landfills decreased by roughly 25% between 1990 and 2000 and has remained mostly constant since then.  This large reduction in methane emissions is due primarily to the institution of EPA regulations, finalized in 1996, requiring large landfills to capture and treat their biogas.  In parallel, the Landfill Methane Outreach Program (LMOP) has facilitated the development of an array of projects to collect and beneficially use landfill biogas, particularly from smaller, unregulated landfills.  These emissions reductions were achieved via a 12-fold increase in the amount of methane that was flared from 1990 to 2010 and a 6-fold increase in the amount of methane used in gas-to-energy projects over the same time period (Figure 3).  From 1990 to 2000, flaring and gas-to-energy projects increased more or less in tandem, but from about 2000 to 2005, gas-to-energy projects stabilized whereas flaring projects continued to increase.  Leading up to 2010, the importance of flaring projects declined slightly whereas gas-to-energy projects increased substantially, such that in 2010, the amounts of methane destroyed by each type of project were almost equivalent.  Today, most of the methane that is still emitted derives from small landfills that are not required to capture and treat their biogas.  Smaller proportions derive from leakage at landfills that do capture their biogas; EPA's Compilation of Air Pollution Emission Factors (AP-42) estimates collection efficiencies at such landfills typically range from 60% to 85%, with an average of 75% capture efficiency most commonly assumed.
Figure 3  -  Amounts of methane generated at, emitted from, flared at, and used in gas-to-energy projects at landfills in the U.S. from 1990-2010.[8]
Possible alternative baseline scenarios 
	The results of our landfill biogas lifecycle analysis would have been very different if, instead of landfills that flare their biogas, we had chosen as a baseline either landfills that do not capture and treat (e.g., that vent) their biogas or those with existing gas-to-electricity projects.  Here, we examine in detail the possible use of these three baselines for the renewable electricity pathway.
Renewable Electricity Pathway.  Table 5 below shows the results of lifecycle analyses for the renewable electricity pathway calculated using each of these three baseline scenarios.  For the scenario involving conversion of a landfill that is venting its methane to a gas-to-energy project, the primary change is in the on-site emissions at the landfill.  Under the venting scenario, the landfill would be releasing large amounts of methane, a potent greenhouse gas, to the atmosphere, so that collecting the methane to use in a gas-to-energy project would be accompanied by reductions in GHG emissions at the landfill of 1959 kgCO2-eq/mmBtu electricity (Table 6).  Another, smaller change relative to the flaring baseline is that because venting landfills do not collect their biogas, they would not have been using any electricity to power blowers for gas collection, so this is not a factor in this scenario.  The large reduction in methane emissions from the landfill results in a 2099% reduction in GHG emissions relative to the gasoline baseline upon conversion from a venting landfill to a landfill with a gas-to-electricity project.  Consideration of the improved efficiency of electric vehicles results in a 765% reduction in GHG emissions relative to the baseline (Table 6).  This demonstrates that renewable electricity would easily qualify as a cellulosic biofuel if we used landfills that vented (did not treat) their biogas as the baseline for comparison.
Table 6  -  Comparison of total lifecycle greenhouse gas emissions (kg CO2-eq/mmBtu electricity or gasoline) for renewable electricity produced from landfill biogas using three different baselines.  Numbers in parentheses show values calculated per mmBtu fuel equivalent, which considers the greater efficiency of electric vehicles compared to gasoline vehicles.
 Lifecycle Stage
                       From Landfills that Vented Biogas
                       From Landfills that Flared Biogas
              From Landfills that Converted Biogas to Electricity
                            2005 Gasoline Baseline
On-site emissions
                               -1959     (-653)
                                  25     (8)
                                       
 
Upstream (electricity production for blowers)
                                       
                                 -13     (-4)
                                       
 
New electricity generation
                                       
                                       
                                 220     (73)
 
                                                               Total Emissions:
                                -1959   (-653)
                                  12     (4)
                                 220     (73)
                                      98
                                                % Change from Gasoline Baseline
                               -2099%   (-765%)
                                 -87%   (-96%)
                                 124%   (-25%)


	The use of landfills that flare their biogas was discussed in the Preamble.  Under this scenario, installation of electricity generators would result in increased on-site emissions due to less efficient combustion of generators versus flares.  The other factor in this lifecycle analysis is reduced upstream emissions from grid electricity that was used to collect the biogas under a flaring scenario but would be replaced by renewable electricity upon installation of gas-to-electricity capability.  Overall emissions using a flaring baseline were 12 kg CO2/mmBtu electricity, corresponding to an 87% reduction in GHG emissions versus the gasoline baseline.  These reductions increase to 96% upon consideration of the increased efficiency of electric vehicles.
	In contrast, if our baseline is landfills that already capture their biogas and use it to generate electricity, there are no physical changes at the landfill, and the only change is that this renewable electricity is tracked to transportation uses.  In this case, the lifecycle GHG emissions increase relative to gasoline, as shown in Table 6.  Because there is no physical change at the landfill, neither on-site emissions nor upstream electrical production to power gas collection blowers changes.  However, this scenario could require additional electricity to be produced to replace that diverted for use in the transportation sector, and this new electricity, which we assume is derived from the grid, is accompanied by additional GHG emissions of 220 kg CO2-eq/mmBtu electricity.  The possible need for this additional electricity can be best understood by stepping through Figure 4 below.  The left side of the figure shows the baseline situation in which combustion of biogas at a landfill generates electricity that is added to the shared electrical grid.  Electricity from the grid is then withdrawn for a variety of uses.  The transportation sector is completely separate from these sectors.  However, if the electricity is directed to the transportation sector, as shown on the right side of the figure, all of these sectors are connected.  EPA is required by the Energy Independence and Security Act of 2007 (EISA) to compare the GHG emissions of renewable fuels to those of the petroleum (gasoline or diesel) fuel baseline.  One way of performing this analysis would be to compare electricity used in electrical vehicles (EVs) to baseline gasoline in gasoline vehicles.  As shown in Figure 4, the net result is that additional electricity is withdrawn from the shared grid to power EVs (that replace gasoline vehicles).  However, the demand for electricity for other uses would not change in this scenario (i.e., the balance of inputs and outputs from the grid box in Figure 4 must remain constant), which requires that the electricity diverted to transportation uses must be replaced with electricity from other sources.  This electricity comes with additional GHG emissions of 220 kg CO2-eq/mmBtu electricity (Table 6) and results in a 124% increase in GHG emissions compared with the gasoline baseline, such that renewable electricity using a gas-to-energy baseline would not qualify as a renewable fuel under the RFS program.  Consideration of the increased efficiency of electric vehicles reduces these emissions to a 25% decrease versus the baseline, which would allow the electricity to qualify as a renewable fuel but not as an advanced or cellulosic biofuel.	
Figure 4  -  Comparison of scenarios used for evaluating the renewable electricity pathway when using an existing gas-to-energy project as a baseline.  Black arrows represent fluxes of greenhouse gases or electricity that are relevant under both scenarios, red arrows represent fluxes that are important only when electricity is directed to transportation use, and blue arrows are fluxes that decrease when electricity is directed to transportation uses.
	However, in reality, the availability of additional renewable electricity would not have a direct impact on the numbers of EVs on the road and thus would not directly replace gasoline as a fuel.  Instead, the renewable electricity would replace other sources of electricity (e.g. from coal- or natural gas-fired power plants) as power for EVs.  This scenario would represent a diversion of renewable electricity from general uses to EVs, and a diversion of grid electricity from general and EV uses to more general uses.  This is effectively an exchange in the uses for the two types of electricity and would not result in any net change in GHG emissions from the baseline scenario.  Thus, if we compare the GHG emissions of a gas-to-energy project used for transportation to one directed into the shared grid, there would be no change in GHG emissions.  The electricity would not replace gasoline, there would be a 0% change from the petroleum baseline, and renewable electricity would still not qualify as a renewable fuel under the RFS program.  As stated above, EPA is required by statute to compare to a petroleum fuel baseline, resulting in a very different lifecycle analysis for the resulting renewable electricity.  However, renewable electricity does not qualify as a renewable fuel under either scenario using a gas-to-energy baseline, so these different ways of considering this question lead to the same end result.
Justification for the choice of a flaring baseline scenario
	Of these three possible baseline scenarios, we believe a flaring baseline is the most logical for the renewable electricity pathway.  First, flaring is a possible scenario at all waste digesters.  Second, assuming that the biogas is flared generally provides a worst case baseline as compared to a venting baseline.  If sources that are using flaring will achieve a 60% GHG reduction when converting to electricity production, sources that are venting their methane or portions of their methane will certainly do so as well.   As discussed in the Preamble, we do not consider landfills that vent their biogas to be a realistic baseline for any of these pathways, and the flaring baseline is the more conservative baseline compared to venting.  Moreover, venting landfills must be small in size and generate a relatively small amount of biogas, otherwise they would be required to capture and treat their biogas.  Accordingly, we expect that these landfills would not typically generate enough electricity from their biogas to justify the capital costs to install generators, so we expect that few of these landfills would convert to gas-to-energy projects Therefore, these facilities are extremely unlikely to draw biogas from the small, unregulated landfills that currently vent their biogas.  LMOP is currently working with some of the larger of these landfills to facilitate the installation of gas-to-energy projects, so such conversions are economically feasible for at least a subset of these landfills.  However, use of venting landfills as a baseline would result in the same classification as would use of flaring landfills as a baseline.  In both cases, the fuel produced from landfill biogas qualifies for the RFS program.
	In contrast with the use of a venting baseline, if we used landfills with existing gas-to-electricity projects as a baseline, these landfill biogas-based biofuels would all fail to qualify for the RFS program.  It is therefore especially important to consider which of these baselines is most appropriate.  If we use a flaring baseline and qualify renewable electricity under the RFS program, it is possible that landfills with existing gas-to-energy projects may divert their renewable electricity for use as a transportation fuel.  In these cases, the renewable electricity would already be produced and added to the grid so there would be no actual change at the landfill.  Additionally, Figure 3 shows that gas-to-energy projects have been expanding since 2003, which suggests that other incentives are already promoting installation of gas-to-energy projects at landfills.  In contrast, if we used facilities with existing gas-to-energy projects as the baseline for comparison, no renewable electricity from landfill biogas would qualify for the RFS program.  The use of this baseline would, therefore, exclude renewable electricity from facilities that converted from flaring (or even venting) from qualifying under RFS, even though these projects would be accompanied by real, large reductions in GHG emissions.  Because of this result, we have determined that use of existing gas-to-energy projects as a baseline would not be appropriate.  EPA believes that the Act should be interpreted and implemented to promote the growth in use of renewable fuels for transportation purposes, and to achieve GHG emissions reductions as a consequence.  EPA believes that use of a flaring baseline best accomplishes these objectives.  
      One option to deal with the discrepancy in lifecycle results between facilities that used to flare their biogas and those with existing gas-to-energy projects would be to qualify electricity produced from these different types of facilities differently.  In this case, renewable electricity produced from landfills that installed new gas-to-electricity projects would qualify as a cellulosic biofuel whereas renewable electricity produced from landfills with existing gas-to-electricity projects would not qualify as renewable fuel under the RFS program.  One problem with such a tiered approach is that landfills with existing gas-to-energy projects previously made the decision to install the gas-to-energy equipment either to replace flaring or instead of installing flares and thus are already best-performers.  Likewise, under such a system, electricity from all new gas-to-energy facilities would qualify as cellulosic biofuels, but electricity from existing facilities would not.  However, many of the new facilities may have installed gas-to-energy projects regardless of the RFS program, driven by the same incentives that motivated the existing facilities.  Given the existence of other incentives to install gas-to-energy capabilities, discriminating between existing and new gas-to-energy projects seems arbitrary in this light.  Additionally, the RFS program does not discriminate against facilities that are already producing renewable fuels, and in fact, the program grandfathers many existing facilities into the program.  It would therefore appear inconsistent with the RFS program to discriminate between facilities that are already creating renewable electricity and those that convert from flaring to gas-to-energy projects.
      We considered all of these factors in deciding on a baseline and believe that a flaring baseline is most appropriate.  The choice of a single baseline for all renewable fuels produced from waste-derived biogas does not discriminate based on prior use of the biogas and is thus consistent with the qualified pathways that have been approved to date under the RFS program.  As discussed above, most venting landfills are unlikely to install gas-to-energy projects because of their small size, thus venting is an inappropriate baseline.  Likewise, using existing gas-to-energy projects as a baseline is also inappropriate because this would exclude projects with legitimate GHG emissions reductions (e.g., those that are currently venting or flaring) from qualifying.  Use of landfills that flare their biogas as a baseline is appropriate because flaring is the main alternative to electricity generation or liquid fuel production at large landfills, and all existing and future gas-to-energy projects chose or will choose to generate electricity instead of simply flaring.  Using this baseline, all renewable electricity generated from landfill biogas would qualify as cellulosic biofuel.  Allowing renewable fuels produced from landfill biogas to generate RINs will provide an additional financial incentive for landfills to convert from flaring or venting to gas-to-energy projects and may thus help accelerate the adoption of this technology and lead to additional GHG emissions reductions.  

References 
