Lifecycle Greenhouse Gas Analysis of an Anaerobic CodigestionFacility Processing Dairy Manure and Industrial Food WasteJacqueline H. Ebner,∥ Rodrigo A. Labatut,† Matthew J. Rankin,∥ Jennifer L. Pronto,† Curt A. Gooch,†

Anahita A. Williamson,∥,§ and Thomas A. Trabold*,∥,¶

∥Golisano Institute for Sustainability, Rochester Institute of Technology, Rochester, New York 14623, United States†PRO-DAIRY Program, Biological and Environmental Engineering, Cornell University, Ithaca, New York 14853, United States§New York State Pollution Prevention Institute, Rochester Institute of Technology, Rochester, New York 14623, United States¶Center for Sustainable Mobility, Rochester Institute of Technology, Rochester, New York 14623, United States

*S Supporting Information

ABSTRACT: Anaerobic codigestion (AcoD) can address food waste disposaland manure management issues while delivering clean, renewable energy.Quantifying greenhouse gas (GHG) emissions due to implementation ofAcoD is important to achieve this goal. A lifecycle analysis was performed onthe basis of data from an on-farm AcoD in New York, resulting in a 71%reduction in GHG, or net reduction of 37.5 kg CO2e/t influent relative toconventional treatment of manure and food waste. Displacement of gridelectricity provided the largest reduction, followed by avoidance of alternativefood waste disposal options and reduced impacts associated with storage ofdigestate vs undigested manure. These reductions offset digester emissionsand the net increase in emissions associated with land application in the AcoDcase relative to the reference case. Sensitivity analysis showed that usingfeedstock diverted from high impact disposal pathways, control of digesteremissions, and managing digestate storage emissions were opportunities to improve the AcoD GHG benefits. Regional andparametrized emissions factors for the storage emissions and land application phases would reduce uncertainty.

■ INTRODUCTION

According to the U.S. Environmental Protection Agency(EPA), methane (CH4) emissions from manure managementcontributed 53 Mt carbon dioxide equivalents (CO2e) to totalU.S. anthropogenic CH4 emissions in 2012.1 Moreover,between 1990 and 2010 they rose 68%, with dairy farmemissions increasing 115% during the same period.1 The EPAattributes this increase, despite a general decrease in nationaldairy populations, to the shift toward larger dairy facilitieswhich utilize liquid-based manure management systems.1

Landfilling of solid waste and treatment of wastewater havealso been large sources of anthropogenic CH4 emissions,contributing 103 and 12.8 Mt CO2e, respectively, to thenational inventory in 2012.1 Anaerobic digestion (AD) has thepotential to mitigate these impacts by effectively capturing andutilizing CH4 emissions and offsetting fossil fuel emissions.Manure management via AD can also reduce odors and

increase farm nutrient management flexibility, while AD of foodwaste can allow food waste generators to respond to increasingregulation of landfilling and land application of organics.Combining food waste with manure is particularly attractive asit often improves farm-based digester economics due toimproved biogas yield as well as additional revenue in theform of “tipping fees” generated from importing food waste.For these reasons, anaerobic codigestion (AcoD) has been

promoted, particularly in areas with strong dairy and foodprocessing industries, such as Upstate New York whichcurrently has 33 on-farm digesters (of the approximately 244in the US).2

Many studies have been conducted concerning the environ-mental performance of biogas production with varied resultsand objectives (a review is contained in Table S.1). Some of thevariation in results can be attributed to a lack ofcomprehensiveness where significant phases of the lifecyclewere neglected. In two separate, comprehensive, comparativestudies, Poeschl et al.3,4 and Borjesson and Berglund5,6 useddata from literature to model a variety of state-of-the-art biogasproduction systems for Germany and Sweden, respectively.Their results showed that lifecycle impacts varied greatly andwere significantly affected by the feedstock, reference system,size and operation of the AD facility, and end-use technology.Dressler et al.7 compared three biogas plants in Germany andconcluded further that regional parameters such as agriculturalpractices, soil, and climate also influenced results. Thus, asBorjesson and Berglund suggested, environmental studies of

Received: October 6, 2014Revised: July 31, 2015Accepted: August 4, 2015Published: August 4, 2015

Article

pubs.acs.org/est

© 2015 American Chemical Society 11199 DOI: 10.1021/acs.est.5b01331Environ. Sci. Technol. 2015, 49, 11199−11208

biogas systems should be based on data referring to the specificlocal conditions valid for the actual biogas system.5

This study analyzed climate change impacts for an anaerobicdigester that codigests manure and industrial food waste(IFW). Data on feedstock, digester operation, and effluentproperties were combined with regional parameters whenavailable (e.g., climate and soil characteristics) to provide anestimate of greenhouse gas (GHG) impacts for a state-of-the-art AcoD in the Northeastern U.S. Data collected throughinterviews was used to model a reference case, representing thebusiness-as-usual food waste disposal and manure managementpractices in lieu of AcoD. This allowed for an analysis of theconsequential impacts incurred. Results are reported on anannual basis and based upon the functional unit of one metricton (t) of influent processed.There are few peer-reviewed studies of the environmental

impact of AcoD in the United States. Several case studies havepresented calculations of impacts using GHG registryprotocols; however, portions of the lifecycle have beenneglected, such as the feedstock reference case emissions,digestate storage emissions, and fertilizer displacementimpacts.8−11 Furthermore, they have often been modeledusing theoretical assumptions such as number of cows ratherthan empirical data.While comprehensive European studies exist, there are

significant regional differences that affect environmental impactanalysis. For example, common European feedstock of pigslurry and energy crops are not prevalent in New York State

(NYS) where AD is primarily dairy manure based with a strongshift toward AcoD with IFW. Feedstock composition influencesemissions as it affects feedstock transportation and pretreat-ment, as well as biogas production. Comparative studies haveconsidered MSW and IFW feedstocks, but the reference caseshave either been excluded12 or modeled to reflect Europeandisposal practices (i.e., incineration or composting).3−6,13 Thedisposal pathways for IFW feedstock in this study werereported to be land application, diversion to animals,wastewater treatment, and wastewater treatment followed bylandfilling. (Landfilling of organic waste is not banned in NYSat this time.)Thus, one novel contribution of this study was inclusion of

the impacts of diverting IFW for use in AcoD. In addition, acomprehensive analysis of a US on-farm anaerobic codigesterwas conducted. In doing so, regional differences such as limitedregulation of CH4 releases, the use of open-air storage pits,regional electric grid mix, and climate and soil conditions wereconsidered. Emission factors and a detailed methodology wereprovided for use in analyzing similar implementations, plus gapsin national and regional factors were identified to guide futureresearch.

■ METHODS

A lifecycle assessment (LCA) methodology was appliedconsidering both direct and indirect GHG emissions. Directemissions consisted of CH4 and nitrous oxide (N2O) releasesdue to biochemical processes, as well as carbon dioxide (CO2)

Figure 1. System boundaries and process flow for the reference and AcoD cases. Boxes represent individual process steps. Dashed boxes indicate asystem expansion to include indirect emissions avoided due to displaced processes. Reference case emissions can also be considered an expansion toinclude avoided processes. Percentages shown in the reference case indicate mass composition of industrial food waste for each pathway. The symbol∗ indicates a pathway not included in the base scenario analyzed but used in the sensitivity analysis.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b01331Environ. Sci. Technol. 2015, 49, 11199−11208

11200

emissions due to the combustion of fossil fuels. Biogenic CO2emissions, such as CO2 released during biogas combustion,were considered part of the photosynthetic carbon cycle andnot included.1 Indirect emissions consisted of upstreamemissions derived from the provision of energy or materialsused in the process and products or services that were avoidedas a result of the AcoD process, such as grid electricity orinorganic (commercial) fertilizer production. Emissions asso-ciated with the construction, maintenance, and decommission-ing of the AcoD plant were not included as these werepreviously reported to be <1% of gross GHG emissions.14

GHG impacts were evaluated in terms of CO2e using the latestIntergovernmental Panel on Climate Change (IPCC) FifthAssessment Report (AR5) 100-year Global Warming Potential(GWP) factors of 28, 265, and 1 for CH4, N2O, and fossil CO2,respectively.15

The reference and AcoD scenarios are shown in Figure 1 anddescribed as follows:

• Reference Case: Liquid manure slurry was collected andstored in an uncovered earthen pit until land-applied (viasurface spreading or injection) as organic fertilizer whenweather, crop, and field conditions allowed, following acomprehensive nutrient management plan. IFW treat-ment was modeled on the basis of the alternativetreatment reported for each of the IFW feedstock. Theseincluded land application (84%), wastewater treatmentplant (WWTP) followed by landfill (14%), landfill (1%),WWTP (1%), and diversion to feed animals (modeled asa sensitivity analysis).

• AcoD case: Food waste was transported to the AcoDfacility, combined with manure produced on-site and fedinto the digester. Biogas produced by the anaerobicdigester was combusted to generate electricity which wasexported to the grid. The digestate was fed into long-term uncovered earthen storage and recycled to croplandas described above for manure.

IFW, manure, and digester effluent (digestate) characteristicsas well as digester operational data (Table S.2) were used tomodel the AcoD and reference cases. The data was based uponan on-farm AcoD in Western New York operating since January2012. Data for the calendar year of 2013 was selected from acomprehensive monitoring study following the EPA ASERTTIreporting protocol and supplemented as needed (Table S.2,with additional detail found in the full report16). During the 12-month period under study, the AcoD facility blended 27% IFWwith manure from approximately 1800 cows. The 8.3 ML,continuous stirred tank reactor operated at an averagetemperature of 41 °C, with a hydraulic retention time (HRT)of 28 days and organic loading rate (OLR) of 2.1 kg VS/m3d.The process was multistage with a secondary biomass/gasstorage tank with an approximate 4-day HRT. Electricity wasgenerated using a 1.426 MW engine generator set. The systemrecovered 13% of the thermal energy produced from the biogasto provide heat to the process.Because the objective of this study was to compare the

impact of AcoD relative to alternative treatment of the samefood waste and manure, unrelated factors were controlled tothe extent possible. For example, although the farm under studyswitched from a flush manure handling system to a scrapesystem concurrent with AcoD implementation, both scenarioswere modeled as scrape systems. Furthermore, while the farmutilized a screw-press separator, both reference and AcoD

systems were modeled without solid−liquid separation, in orderto utilize the data available and because solid−liquid separationcan be implemented independently of AcoD. GHG emissionswere estimated throughout the process for both scenarios bycombining the empirical data with emission factors gatheredfrom literature as described in the following paragraphs (andTable S.3).

Reference Case Emissions. Dairy Manure Storage.Manure storage emissions were calculated per the IPCCmethodology (Tier 3) for reporting of GHG emissions due tolivestock.17 CH4 generated from the anaerobic decompositionof manure was based upon the volatile solids content (VSM)and the biomethane potential (B0,M) of the manure, along witha methane conversion factor (MCFi,,j) dependent upon themanure management system and climate. The MCFls,ny for aliquid slurry management system in NYS was obtained from theU.S. GHG Inventory, which was modeled to include monthlytemperature variation and account for monthly VS content ofliquid slurry stored.18

Direct N2O emissions result from the processes ofnitrification and denitrification. These emissions were estimatedas a portion of the total Kjeldahl nitrogen (TKNM) stored usingthe IPCC default emission factor (EF3 = 0.005) for dairymanure liquid slurry storage with a natural crust cover.17,18 Twosources of indirect emissions were calculated, indirect N2Oresulting from atmospheric deposition of volatilized nitrogen(primarily in the form of ammonia, NH3) and indirect N2Oresulting from leaching and runoff. These emissions werecalculated using the IPCC default emission factors (EF4 = 0.01and EF5 = 0.0075, respectively) to estimate the portion ofvolatilized N (FracGASMS) or runoff/leached N (Fracrunoffleach)converted to N2O−N.

17,18 FracGASMS,ls = 0.26 for dairy liquid/slurry management was taken from the U.S. Inventory of GHGemissions for NH3 emissions.16 The inventory used aFracrunoff,ls,ma = 0.007 for liquid/slurry management in themid-Atlantic region derived from the EPA’s Office of Waterrunoff data (as losses from leaching were stated to be small).18

Land Application of Manure. Net GHG emissions fromland application include the provision and combustion of fossilfuel to transport and spread manure, direct and indirectemissions due to subsequent biodegradation, emissions relatedto fertilizer displacement, and long-term carbon sequestration.The emission factor reported by Møller et al.12 for trans-portation and field spreading based upon an average distance tothe field of 20 km was scaled to the 11 km transportationdistance reported in this study, resulting in an emission factorof 0.8 kgCO2e/t applied.Direct N2O emissions were determined by applying the

default IPCC emission factor (EF1 = 0.0125) to estimate theportion of N applied converted to N2O, where N applied is themeasured total N (TKNM) minus N2O losses15 and anadditional 2% of N due to N2

19 and NO20 losses duringstorage and land application. Indirect N2O emissions due tovolatilization and leaching/runoff were also calculated for landapplication. Volatilization of N applied to land (FracGASM) isknown to be affected by several variables including timing,application rate, and application technique.19,20 However, as USfactors particular to these variables were not available, the IPCCdefault (FracGASM = 0.20; EF4 = 0.01) was used. Precipitationand soil hydrological group data for Wyoming County, NewYork18 estimated a low probability of leaching thereforeFracrunoff as described above was used to calculate indirectN2O resulting from runoff/leaching.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b01331Environ. Sci. Technol. 2015, 49, 11199−11208

11201

Land applying dairy manure returns valuable nutrients to thesoil thereby displacing inorganic (commercial) fertilizer use.Mineral-N is readily available for uptake by crops grown in theseason of application; however, it is subject to losses throughNH3 volatilization, denitrification, and nitrate leaching. OrganicN is more stable but over time is mineralized and becomesplant available. Inorganic fertilizer displacement was calculatedusing a mass balance approach to sum the N that will beavailable for plant uptake. Mineral-N, as measured via totalammonaical-N (TANM), was adjusted to subtract losses duringstorage and land application. This was added to 52%(MinFactor,ny) of organic N that was estimated to be plantavailable within 3 years, on the basis of a mineralization profilefor liquid dairy manure in NYS.21 Phosphorus (P) availability isassumed to be 90% of P applied.22 Potassium (K) displacementwas not considered separately as it is often included in N and Pinorganic fertilizer blends. Lifecycle GHG emission factors forfertilizer production were taken from the mean values reviewedby Wood and Cowie.23 In addition to production emissions,displacing inorganic fertilizer displaces N2O emissionsassociated with inorganic fertilizer application, replacing themwith those of organic fertilizer. These emissions were calculatedaccording to the IPCC protocol for indirect and direct N2Oemissions due to land application of inorganic fertilizer.17

Finally, carbon in manure can be biochemically or biophysicallystabilized in soil, resulting in carbon sequestration (CS). Risseet al.22 reviewed several studies of manure application andestimated 8−38% of C applied remained sequestered fortemperate and frigid regions. A nominal value of 13% of VS isused in the present study. This value was chosen on the basis ofthe lignin content of manure24 as it was reasoned that whileapplication rates, tillage practices, climate, and crop rotation allaffect carbon sequestration rates, over a very long time, thecomposition of the substrate has the largest influence on carbonremaining. Furthermore, this was consistent with the approachto estimating CS for landfilling, which was based upon substratedegradability experiments.Food Waste Disposal. Log records maintained by digester

personnel tracked the quantity and source of imported IFW(Table S.4). Most of the IFW (84%) was dairy processingwaste, consisting of any combination of whey, wastewater, ormilk products. Grease trap waste (GTW) and effluent fromdissolved air floatation (DAF) wastewater treatment con-stituted 14%. The remaining 2% of the IFW influent wascomprised of tomato processing waste and wastewaters fromdistilleries and wineries (Table S.5).Interviews with the waste generators or haulers were

conducted to ascertain where the waste would have gone hadit not been diverted to the AcoD facility. The predominatealternative disposal scenarios consisted of WWTP/Landfilldisposal of GTW/DAF, land application of dairy processingwaste, and the remaining 2% split between landfill and WWTP.Land Application of Food Waste. Land application of dairy

processing wastewater has been practiced in the United Statesfor over 50 years.25 The emissions associated with transportingthe waste to a farm for land application were calculated using anaverage transport distance of 100 km.26 Emissions due tooperation of farm equipment for spreading and direct andindirect N2O emissions were calculated similarly to thosedescribed above in relation to manure. Farm spreadingequipment emissions were calculated as described above inrelation to manure. The portion of N volatilized (FracGASdairy)and leached (FracLEACHdairy)were estimated from studies of

whey land application.25 Nominal estimates of N and P werederived from a survey of the literature (Table S.6A).25,27,28

Fertilizer displacement was calculated on the basis of 20%mineralization of organic N25 and applying the emission factorsfor fertilizer production and fertilizer emissions discussedabove. CS data was not available specifically for dairy processingwaste; therefore, it was estimated by applying data on thebiodegradable fraction of dairy wastewater relative to that ofdairy manure24 to arrive at 10% of VS applied.

WWTP/Landfill Disposal of GTW/DAF. Although wastehaulers reported that GTW and DAF were disposed of at theWWTP, interviews with the WWTP operator revealed thesewastes were actually combined untreated with wastewatersludge to achieve the solids content required for landfilldisposal.29 Thus, the GHG emissions associated with thedisposal of GTW/DAF included the impacts of transportingthe waste to the WWTP, plus the treatment of the waste at thelandfill. Transport emissions were calculated using a transportdistance of 50 km to the WWTP.26 Landfill emissions werecalculated as described in the following paragraph based uponthe characteristics of GTW and DAF.

Landfill Disposal. Landfill emissions consist of thoseassociated with fossil fuel used to collect the waste and operatethe landfill, plus the net emissions due to the waste decay in thelandfill. The emission factor for transport and operation of thelandfill was taken from the EPA’s WARM model.30 CH4generation at the landfill was calculated using a multiphased,first-order decay model adapted (to sum emissions over a 30-year rather than a 10-year period) from that used by theClimate Action Reserve (which is based on the UNFCC CleanDevelopment Mechanism).31,32 A specific decay rate constantwas not available for individual IFW constituents butexperiments by de la Cruz and Barlaz33 provided one for thebroad category of food waste which was used. Medianbiomethane potentials from a survey of the literature (TableS.6B) were used for the various IFW.34−39 Landfill gas (LFG)captured was estimated using a gas capture factor (GC),representing the fraction of landfills in the State with LFGrecovery systems30,31 and a landfill capture efficiency scheduleto model the efficiency of LFG collection over time.40

Electricity generated from recovered LFG was calculated onthe basis of a conversion efficiency and plant capacity factorobtained from the EPA’s landfill outreach program.30 Avoidedgrid emissions were calculated on the basis of the offset ofnonbaseload electricity generation, assuming the regional gridmix.41 Finally, 0.08 kg C/kg dry food waste was estimated toremain sequestered in the landfill.42

Municipal WWTP Disposal of Wastewater. Data onwastewater treatment emissions are limited and highly variable.The value of 0.518 kgCO2e/m

3 wastewater from the EcoInventv.2.2 database was applied to the small percentage ofwastewater that was diverted from a WWTP.43

Diversion to Feed Animals. While the primary alternativetreatment of dairy processing waste was reported to be landapplication, a sensitivity analysis was performed to analyze theimpact of diverting dairy waste to feed cows. A transportationdistance of 100 km to the farm was assumed. On the basis ofthe nutritional content of the dairy waste to the cows, 0.05 kgof corn was calculated to be displaced by 1 kg of dairyprocessing waste.44 Displaced GHG impacts due to cultivationand production of corn/maize animal feed were obtained fromthe EcoInvent v.2.2. database.43

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b01331Environ. Sci. Technol. 2015, 49, 11199−11208

11202

AcoD Case Emissions. Food Waste Hauling. Delivery logswere used to calculate emissions associated with transportationof the food waste to the digester using the freight transportemission factor and the distance and the weight for eachdelivery (Table S.4).26 A total of 1537 trips from 15 wastegenerators were made, ranging from 22 to 194 km, with a 40km average one-way transport distance and average payload of22t.Digester Operation. Digester emissions consist of direct

emissions due to leaks or incomplete combustion as well asindirect emissions offset by electricity generated. Canadian andGerman studies reported fugitive emissions ranging from 2.1%to 3.1% of CH4 utilized.

45,46 The nominal value of 3% of gasutilized was used. However, Liebetrau et al.46 noted that whenleaks and malfunctions were eliminated, near zero fugitiveemissions were measured. Conversely, releases of biogas wereobserved through emergency vents due to overpressureconditions in the reactor or when flaring was not possible.Therefore, a sensitivity analysis was performed using the IPCCdefault uncertainty range of 0−10%.18 This range also allowsfor consideration of emissions due to flaring of biogas, whichwere minimal during the period of study due to issues related toflare operation but were reported to be on average 21% of gasproduced in a study of seven NYS AD plants.47 Site suppliedmeasurements of gen-set exhaust reported 1314 ppmv dry CH4,which equated to 2.5% of the CH4 utilized. This was consistentwith reported values for incomplete combustion, which rangedfrom 0.4% to 3.28%.45,46 N2O exhaust emissions were a smallercontribution at 0.03gN2O/m

3CH4 utilized, which is alsoconsistent with the range reported in the literature (0.02−1.75g N2O/m

3 CH4 utilized).46,48

Excess electricity beyond a parasitic load to operate pumpsand mixers of about 12% of electricity generated was exportedto displace grid electricity (Table S.2). Avoided emissions werecalculated on the basis of a nonbaseload emissions factor from

the USEPA eGRID database for the Northeast regional(NPCC) grid mix.41

Digestate Storage. Similar to the storage of manure,uncovered storage of liquid digestate can generate CH4 overtime. It has been shown that CH4 emissions due to storage ofdigested manure are lower than those of raw manure due to VSdestruction during the digestion process.49,50 However, it hasalso been shown that just as codigestion of substrates withmanure increases biogas production, codigested slurries showhigher residual CH4 emissions than manure-only slurries.49 Themain factors influencing digestate residual emissions are VScontent, degree of degradation, and storage temperature.51,52

While data specific to US conditions was not available, severalEuropean studies of digestate emissions were reviewed. Hansenet al.51 observed that temperatures in storage tanks directly fedfrom digesters were mainly affected by effluent temperature andranged from 20 to 40 °C. Batch studies of European AcoDsamples incubated in this range had a mean value of 0.054m3CH4/kgVS stored (Table S.7).51−54 This equates to 1.6m3CH4/t digestate which is consistent with the results of astudy of 61 AcoD plants in Germany which reported averageresidual CH4 potential of 1.5 m3CH4/t digestate for multistageprocesses.55

Modeling of digestate nitrous emissions is complex andinfluenced by many factors. Although reduction in organicmatter typically prevents formation of a surface crust which isassociated with lower N2O formation, several studies havereported increases in digestate N2O storage emissions relativeto untreated manure.49,50 Therefore, digestate direct N2Oemissions were calculated using the IPCC default emissionfactor (EF3) for manure storage.It has been argued that emission factors based upon mineral-

N rather than total N more closely model the volatilization andleaching/runoff processes.19 Furthermore, the digestion processincreases mineral-N content. However, in this case, although

Table 1. Summary of GHG Impacts for a Unit of Waste (kg CO2e/t Waste), Annually (t CO2e/yr) and Normalized by Mass ofInfluent Processed (kg CO2e/t Influent) for the Reference and AcoD Casesa

Reference case AcoD case

climate change impact perfunctional unit

climate change impact perfunctional unit

process phasekgCO2e/t wasteb tCO2e/yr

ckgCO2e/t influentc process phase

kgCO2e/t wasteb tCO2e/yr

ckgCO2e/t influentc

IFW disposal 60.1 1926 16.0 IFW transport 4.0 129 1.1dairy waste 0.2 7 0.1GTW 747.0 1493 12.4DAF 155.5 384 3.2food processing waste 127.1 42 0.4wastewater 0.5 0 0.0

digester/gen-set emissions 2243 18.6displaced grid emissions (4247) (35.3)

manure storage 73.2 6463 53.7 digestate storage 4691 39.0manure land application 14.6 1286 10.7 digestate land application 1543 12.8displaced inorganic fertilizer (10.7) (946) (7.9) displaced inorganic fertilizer (981) (8.2)carbon sequestration (27.0) (2382) (19.8) digestate carbon

sequestration(1543) (12.8)

net reference emissions 6348 52.8 net AcoD emissions 1836 15.3net impactd (4512) (37.5)reduction 71%

aPositive values indicate emissions, and negative values () indicate a reduction in emissions. bEmissions associated with the treatment of a singlewaste stream/manure. cEmissions based upon the combined codigestion influent processed. dNet impact considers the replacement of the referenceprocess by the AcoD process.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b01331Environ. Sci. Technol. 2015, 49, 11199−11208

11203

mineral content of the feedstock was increased duringdigestion, the mineral content of the digestate (TAND) wassimilar to that of raw manure (TANM) (Table S.8). Thus, whileemissions modeling based upon mineral-N content mayprovide a more accurate estimation, error due to using theIPCC (TKN based) methodology will be comparable for boththe AcoD and reference cases and thus have minimal impactson net results. It has also been suggested that elevated pH andlower dry matter content, as found in digestate, may beconducive to higher volatilization. However, because studies ofAcoD are lacking it is difficult to distinguish the magnitude ofthese effects vs the impact of higher TAN content in studies ofdigested manure vs undigested manure. Therefore, the IPCCdefault factors were also used to calculate indirect N2Oemissions and uncertainty analyzed.17,18

Land Application of Digestate. Importing food wasteincreased the volume of organic fertilizer being land applied(Table S.8). This resulted in increased transportation distanceto the fields for spreading from 11 km for raw manure to 19 kmfor AcoD digestate. The emission factor provided by Møller etal.12 was scaled and applied to calculate transport and spreadingemissions.Despite observing elevated pH and lower organic matter

content, field experiments by Amon et al.50 and Clemens etal.49 observed no significant difference in N2O as a percentageof mineral-N during land application of digested manure vsuntreated manure. Therefore, direct and indirect N2Oemissions, N losses, and fertilizer displacement were modeledusing the IPCC methodology as described for manure andanalyzed through sensitivity analysis.Little data exists concerning CS for AcoD digestate. Bruun et

al.56 used an agronomic model to analyze inorganic fertilizersupplemented with digestate from MSW vs composted MSWand observed that, as time increased, the difference between CSrates between the two treatments decreased resulting in nearlyidentical rates after 100 years. Therefore, 12% of carbon appliedwas used to model digestate CS which was the weightedaverage of the raw manure CS rate and the IFW CS rate.

■ RESULTS AND DISCUSSION

Comparison of Reference Case to AcoD Case. AnnualGHG impacts and emission factors per ton processed for thereference case and the AcoD case were compared (Table 1). Itis important to consider that the impacts of a given fooddisposal and manure management pathway can be displaced bythose of an alternative pathway, but the treatment of manureand food waste must be achieved. Thus, the reference case canbe considered a system expansion to account for the processesdisplaced by AcoD.Annual net GHG impacts were reduced by 4512 t CO2/yr or

37.5 kgCO2e/t influent treated. This is a 71% reduction for theAcoD case relative to the reference case. Displacement of gridelectricity emissions was the largest contribution (avoiding4247 t CO2e/yr or 35.3 kg CO2e/t influent). The benefit ofavoiding alternative IFW disposal (1926 t CO2e/yr or 16.0 kgCO2e/t influent) was much greater than the impact of haulingfood waste to the digester (129 t CO2e/yr or 1.1 kg CO2e/tinfluent). This was driven by GTW/DAF which avoidedWWTP/landfill emissions (747.0 kg CO2e/t GTW/DAF),although these only constituted 4% of the total influent.Impacts of digestate storage relative to manure storage resultedin (14.7) kgCO2e/t influent, where the net benefit of lower VSovercame the increase in digestate volume due to importedIFW. In both cases, land application resulted in a net benefitwith fertilizer displacement and carbon sequestration benefitsoffsetting direct and indirect fossil fuel and N2O emissions.Land application of digestate had lower net benefit than that ofmanure due to greater volume being land applied, increasedtransportation distance to the field, and lower carbonsequestration in the AcoD process.The largest source of direct emissions was CH4 (5778 t

CO2e/yr or 48 kg CO2e/t influent for the AcoD case and 7602t CO2e/yr or 63.2 kg CO2e/t influent for the reference case).N2O direct and indirect emissions contributed 2535 t CO2e/yr(or 21.1 kg CO2e/t influent) in the AcoD case and 2210 tCO2e/yr (18.4 kg CO2e/t influent) for the reference case.N2O emissions were larger in the land application phases than

Figure 2. Contribution of greenhouse gases to climate change impacts annually (t CO2e/yr) and based upon mass of influent processed (kg CO2e/tinfluent) for phases of the reference and AcoD cases. Top black bars represent net emissions for each case.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b01331Environ. Sci. Technol. 2015, 49, 11199−11208

11204

during storage, and direct emissions were larger than indirect(Figure 2). Direct fossil fuel emissions had a minor impact.Carbon sequestration offset emissions 1543 t CO2e/yr or (12.8kg CO2e/t influent) for the AcoD case and 2879 t CO2e/yr (or23.9 kg CO2e/t influent) for the reference case. Avoided fossilfuel use also contributed an offset (5228 t CO2e/yr or 43.3 kgCO2e/t influent) for the reference case and 1284 t CO2e/yr (or10.7 kg CO2e/t influent) for the reference case.Impact of Feedstock Composition. A sensitivity analysis

modeled three scenarios where IFW composition and

alternative disposal treatment varied. Reference case emissionsand biogas production were estimated on the basis of thecharacteristics of the feedstock and biogas utilization, andelectricity conversion efficiencies calculated in this study wereapplied (as explained in Ebner et al.57). Net AcoD benefitvaried significantly (Figure 3a−d). Highly degradable GTWcodigestion increased the net benefit by an order of magnitude(Figure 3b) to 29 969 t CO2e/yr or 249.2 kg CO2e/t influent.This was due to avoidance of significant landfill emissions aswell as an increase in displaced grid electricity with only a

Figure 3. Sensitivity analyses. Comparison of AcoD, reference case GHG emissions, and net benefit in response to variation and uncertainty inparameters.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b01331Environ. Sci. Technol. 2015, 49, 11199−11208

11205

minor increase in fugitive emissions. Diverting whey fromfeeding animals (Figure 3d) resulted in the lowest net benefit1030 t CO2e/yr or 8.6 kg CO2e/t influent. While benefits ofavoiding raw manure storage still enabled a net benefit,diverting whey from feeding animals incurred emissions due toproduction impacts of replacement feed, with whey feedstockproviding only moderate CH4 production.Impact of CH4 Losses. Uncertainty in estimating CH4

storage emissions can have a large impact on the results(Figures 3h and S.1a,b). However, nominal factors for GHGemissions resulted in a loss of 8.8% of CH4 utilized. CapturingCH4 generated during storage eliminates atmospheric emis-sions and displaces grid emissions. This would more thandouble the net AcoD benefit to 9526 t CO2e/yr or 79.2 kgCO2e/t influent (Figure 3e).Similarly, nominal fugitive emissions were modeled as 3% of

CH4 utilized. Two scenarios explored the impact ofuncontrolled CH4 releases and leaks. If CH4 leaks werereduced to zero, the net emissions would be reduced to 47.6 kgCO2e/t influent (Figure 3f). However, emissions of 10% arequite possible through poor system inspection and uncon-trolled releases, which would reduce the net emissions to 13.9kg CO2e/t influent (Figure 3g).Nitrous Emissions. N2O emission estimates are subject to

both uncertainty and variability. A simulation that varied all ofthe IPCC parameters related to land application N2O emissionswithin their uncertainty ranges was used to analyze uncertainty(Table S.9). GHG impacts from N2O emissions varied by anorder of magnitude (from 481 t CO2e/yr to 6476 t CO2e/yr).The indirect emission factor (EF4) for volatized N was found tohave the largest impact (based on coefficient of correlation).However, varying EF4 alone had little effect on net resultsbecause it was applied to both the reference and AcoD cases.The effect of a high indirect emission factor was greater whenthere is a difference between reference and AcoD casevolatilization rates such as when the feedstock compositionresults in a TAND that differs significantly from TANM. Inaddition, variability in NH3 emissions can arise from applicationtechnique and fertilization rates. Field experiments of manureapplication reported NH3 emission varying from 2% of theTAN applied for slurry injection on arable land to 74% forbroadcast surface spreading on grassland.58 Low emissiontechniques have the added benefit of preserving the amount ofN remaining to displace inorganic fertilizer. Again, net impactswill be greatest when there is a difference between the referenceand AcoD case application techniques. These uncertainty andvariability impacts were explored through a sensitivity analysiswhere the volatilization rate for digestate was modeled to behigher than that of raw manure and a high indirect emissionrate was assumed (Figure 3i). The result was a 35% reductionin net benefit. Thus, research or modeling to better understandindirect N2O emissions can be important, especially in caseswhere elevated TAND is anticipated or when care is not takento minimize NH3 volatilization.Fertilizer Displacement and Carbon Sequestration

(CS). Some studies have neglected fertilizer displacement orwhen included have not considered inorganic fertilizeremissions.2−7 Although this analysis includes the impact offertilizer displacement, it is important to point out that theimpact will only be realized if the nutrients are required by thesystem and do in fact replace inorganic fertilizer use. This iscomplicated by several factors, including the imbalance ofnutrients in manure, difficulties in estimating nutrient

availability, and low concentration of nutrients, makingtransport of organic fertilizer over long distances costly. It isunclear if a bias exists between the AcoD and reference cases. Asensitivity analysis assuming no change in fertilization practicesin the AcoD case despite the import of food waste nutrientsshowed minimal impact (Figure S.4). However, betterunderstanding of true fertilization displacement due to AcoDmay still be important, especially in cases where high nutrientcontent feedstock is imported or when large portions of thefeedstock are not land applied in the reference caseCS has also been inconsistently applied in waste treatment

LCAs, often neglected or narrowly analyzed. Long-term studiesof CS across different treatment pathways and for differentsubstrates are lacking. In this study, CS was consistentlyestimated across pathways as the long-term nondegradablefraction based upon substrate composition. A sensitivityanalysis of a lower carbon sequestration factor for digestate(CSD = 0.1) relative to manure (CSM = 0.2) more than halvedthe net benefit (Figure 3j). Thus, research to better understandCS based upon waste composition for key waste treatmentpathways may be important.

Other Factors and Study Limitations. Operatingparameters (i.e., HRT, OLR) and performance issues(mechanical or biological) can impact digester performance.The capacity factor (CF) is a measure of the performance ofthe digester system and is defined as the electrical energygenerated by an engine gen-set relative to the maximumelectrical energy that could have been generated in the sametime period. It was calculated as 0.73 for this study. CF oftenimproves over time; however, a study of 7 New York on-farmdigesters reported an average CF of 0.57.47 Linear regression ofa sensitivity analysis of CF (Figure S.3) resulted in a change inelectricity generation of 110 MWh/percent CF and associatedGHG change of 59 t CO2e/percent change in CF, for theNPCC regional grid mix.While the impacts of several other parameters were explored

(Figures S4 and S5), it is not possible to generalize this study toall AcoD applications. This study analyzed climate changeimpacts of a state-of-the-art AcoD in Western New York,identifying key impacts and uncertainty. Furthermore, effort hasbeen made to provide a clear methodology to be applied toother AcoD implementations.

■ ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.est.5b01331.

Tables S.1−S.11 and Figures S.1−S.4 (PDF)

■ AUTHOR INFORMATIONCorresponding Author*Phone: (585) 475-4696; fax: (585) 475-5455; e-mail: tatasp@rit.edu.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSFunding provided by the New York State Pollution PreventionInstitute (NYSP2I) through a grant from the NYS Departmentof Environmental Conservation, which also provided aGraduate Research Assistantship for J.H.E. Funding was alsoprovided by the Cornell PRO-DAIRY Program. Any opinions,

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b01331Environ. Sci. Technol. 2015, 49, 11199−11208

11206

findings, conclusions or recommendations expressed are thoseof the author(s) and do not necessarily reflect the views of theDepartment of Environmental Conservation.

■ REFERENCES(1) Environmental Protection Agency (EPA). Inventory of U.S.Greenhouse Gas Emissions and Sinks: 1990−2012; USEPA: Wash-ington, DC, 2014; http://www.epa.gov/climatechange/Downloads/ghgemissions/US-GHG-Inventory-2014-Main-Text.pdf.(2) Agstar. Operating digester projects; http://www.epa.gov/agstar/projects (accessed December 15, 2014).(3) Poeschl, M.; Ward, S.; Owende, P. Environmental impacts ofbiogas deployment−Part I: Life cycle inventory for evaluation ofproduction process emissions to air. J. Cleaner Prod. 2012, 24, 168−183.(4) Poeschl, M.; Ward, S.; Owende, P. Environmental impacts ofbiogas deployment−Part II: life cycle assessment of multipleproduction and utilization pathways. J. Cleaner Prod. 2012, 24, 184−201.(5) Borjesson, P.; Berglund, M. Environmental systems analysis ofbiogas systemsPart I: Fuel-cycle emissions. Biomass Bioenergy 2006,30 (5), 469−485.(6) Borjesson, P.; Berglund, M. Environmental systems analysis ofbiogas systemsPart II: The environmental impact of replacingvarious reference systems. Biomass Bioenergy 2007, 31 (5), 326−344.(7) Dressler, D.; Loewen, A.; Nelles, M. Life cycle assessment of thesupply and use of bioenergy: impact of regional factors on biogasproduction. Int. J. Life Cycle Assess. 2012, 17 (9), 1104−1115.(8) Pronto, J. L.; Gooch, C. A. Greenhouse gas emission reductionsfor seven on-farm dairy manure-based anaerobic digestion systems−final results. In ASABE−International Symposium on Air Quality WasteManagement and Agriculture Conference Proceedings, Dallas, Texas, 13−16 September 2010; Vol. 7, pp 35−42 (DOI: 10.13031/2013.32617).(9) Bentley, C.; Gooch, C. A.; Pronto, J.; Scott, N. R.; McGlynn, S.Greenhouse gas emissions from a community anaerobic digester withmixed organic wastes. In ASABE Annual International Meeting,Pittsburgh, Pennsylvania, June 20−23, 2010; Paper No. 1009892.(10) Bartram, D.; Barbour, W. Estimating greenhouse gas reductionsfor a regional digester treating dairy manure. In 13th InternationalEmission Inventory Conference, Clearwater, Florida, June 9, 2004.(11) Artrip, K.; Shrestha, D.; Coats, E.; Keiser, D. GHG emissionsreduction from an anaerobic digester in a eairy farm: Theory andpractice. Appl. Eng. Agric. 2013, 29 (5), 729−737.(12) Møller, J.; Boldrin, A.; Christensen, T. H. Anaerobic digestionand digestate use: accounting of greenhouse gases and global warmingcontribution. Waste Manage Res. 2009, 27, 813−824.(13) Rodriguez-Verde, I.; Regueiro, L.; Carballa, M.; Hospido, A.;Lema, J. M. Assessing anaerobic co-digestion of pig manure withagroindustrial wastes: The link between environmental impacts andoperational parameters. Sci. Total Environ. 2014, 497-498, 475−483.(14) Meyer-Aurich, A.; Schattauer, A.; Hellebrand, H. J.; Klauss, H.;Plochl, M.; Berg, W. Impact of uncertainties on greenhouse gasmitigation potential of biogas production from agricultural resources.Renewable Energy 2012, 37, 277−284.(15) Intergovernmental Panel on Climate Change (IPCC). ClimateChange 2013: The Physical Science Basis. Contribution of WorkingGroup I to the Fifth Assessment Report of the IntergovernmentalPanel on Climate Change; Stocker, T. F., Qin, D.; Plattner, G.-K.;Tignor, M.; Allen, S.K.; Boschung, J.; Nauels, A.; Xia, Y.; Bex, V.;Midgley, P.M., Eds.; Cambridge University Press: Cambridge, UnitedKingdom and New York, NY, USA; 1535 pp; DOI:10.1017/CBO9781107415324.(16) Gooch, C.; Labatut, R. Anaerobic Digestion Monitoring atSynergy Biogas LLC; Cornell University Agreement No. 65961.Evaluation of the Continuously-Mixed Anaerobic Digester System atSynergy Biogas following the ERG Protocol (Jun 2012 to May 2014).www.manuremanagement.cornell.edu.

(17) National Greenhouse Gas Inventories Programme. IPCCguidelines for national greenhouse gas inventories. Eggleston, H. S.,Buendia, L., Miwa, K., Ngara, T., Tanabe, K., Eds.; IntergovernmentalPanel on Climate Change (IPCC), IGES: Japan, 2006; http://www.ipcc-nggip.iges.or.jp/public/2006gl/.(18) Environmental Protection Agency (EPA). Inventory of U.S.greenhouse gas emissions and sinks: 1990−2012. In Annex 3:Methodological descriptions for additional source or sink categories;http://www.epa.gov/climatechange/Downloads/ghgemissions/US-GHG-Inventory-2014-Annex-3-Additional-Source-or-Sink-Categories.pdf.(19) Velthof, C.; van Bruggen, C. M.; Groenestein; de Haan, B. J.;Hoogeveen, M. W.; Huijsmans, J. F. M. A model for inventory ofammonia emissions from agriculture in the Netherlands. Atmos.Environ. 2012, 46, 248−255.(20) Stehfest, E.; Bouwman, L. N2O and NO emission fromagricultural fields and soils under natural vegetation: summarizingavailable measurement data and modeling of global annual emissions.Nutr. Cycling Agroecosyst. 2006, 74, 207−228.(21) Ketterings, Q.; Klausner, S.; Czymmek, K. Nitrogen guidelines forfield crops in New York, Second Release, June 22, 2003; CornellUniversity Department of Crop and Soil Sciences Extension SeriesE03-16; Cornell University: Ithaca, NY, 2003.(22) Risse, L. M.; Cabrera, M. L.; Franzluebbers, A. J.; Gaskin, J. W.;Gilley, J. E.; Killorn, R.; Radcliffe, D. E.; Tollner, E. W.; Zhang, H.Land application of manure for beneficial reuse. In White Papers onAnimal Agriculture and the Environment; National Center for Manureand Animal Waste Management: Raleigh, N.C., 2001; 7.1−7.46.(23) Wood, S.; Cowie, A. A review of greenhouse gas emission factors forfertilizer production; For IEA Bioenergy Task 38; Research andDevelopment Division, State Forests of New South Wales,Cooperative Research Centre for Greenhouse Accounting, 2004.(24) Labatut, R. A.; Angenent, L. T.; Scott, N. R. Biochemicalmethane potential and biodegradability of complex organic substrates.Bioresour. Technol. 2011, 102 (3), 2255−2264.(25) Ghaly, A. E.; Mahmoud, N. S.; Rushton, D. G.; Arab, F.Potential environmental and health impacts of high land application ofcheese whey. Am. J. Agric. Biol. Sci. 2007, 2, 106−117.(26) U.S. Life Cycle Inventory (USLCI) Database; National RenewableEnergy Laboratory: Golden, CO, 2012; https://www.lcacommons.gov/nrel/search.(27) Kushwaha, J.; Srivastava, V. C.; Mall, I. D. An overview ofvarious technologies for the treatment of dairy wastewaters. Crit. Rev.Food Sci. Nutr. 2011, 51, 442−452.(28) Watkins, M.; Nash, D. Factory wastewaters, their use on landand possible environmental impacts − A Mini Review. Open Agric. J.2010, 4, 1−9.(29) Peletz, S. Monroe County Department of EnvironmentalServices. Personal communication, March 3, 2014.(30) Environmental Protection Agency (EPA). Waste ReductionModel (WARM), v13; Chapter: Landfilling; USEPA: Washington, DC;March 2015 http://epa.gov/epawaste/conserve/tools/warm/pdfs/Landfilling.pdf.(31) Climate Action Reserve (CAR). Organic Waste Digestion, ProjectProtocol, v.2.1; Climate Action Reserve: Los Angeles, CA, 2011;http://www.climateactionreserve.org/how/protocols/organic-waste-digestion/.(32) UNFCCC/CCNUCC. CDM Methodological to determinemethane emissions avoided from disposal of waste at a solid wastedisposal site. Report Annex10 v. 4; August 2008 [Retrieved January 15,2014]; http://cdm.unfccc.int/EB/026/eb26_repan14.pdf.(33) De la Cruz, F. B.; Barlaz, M. A. Estimation of waste componentspecific landfill decay rates using laboratory-scale decomposition data.Environ. Sci. Technol. 2010, 44, 4722−4728.(34) Baily, R. Anaerobic digestion of restaurant grease wastewater toimprove methane gas production and electrical power generationpotential. In Proceedings of the 80th Annual Technical Exhibition andConference of the Water Environment Federation, San Diego, CA,October 13−17, 2007; pp. 6793−6805.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b01331Environ. Sci. Technol. 2015, 49, 11199−11208

11207

(35) Razaviarani, V.; Buchanan, I. D.; Malik, S.; Katalambula, H.Pilot-scale anaerobic co-digestion of municipal wastewater sludge withrestaurant grease trap waste. J. Environ. Manage. 2013, 123, 26−33.(36) Davidsson, A.; Lovstedt, C.; la Cour Jansen, J.; Gruvberger, C.;Aspegren, H. Co-digestion of grease trap sludge and sewage sludge.Waste Manage. 2008, 28, 986−992.(37) Luste, S.; Luostarinen, S. Anaerobic co-digestion of meat-processing by-products and sewage sludge − Effect of hygienizationand organic loading rate. Bioresour. Technol. 2010, 101, 2657−2664.(38) Woon, S.; Othman, M. Anaerobic Digestion of Meat Wastes. InInternational Conference on Clean and Green Energy; IPCBEE Vol. 27;IACSIT Press: Singapore, 2012.(39) Dinuccio, E.; Balsari, P.; Gioelli, F.; Menardo, S. Evaluation ofthe biogas productivity potential of some Italian agro-industrialbiomasses. Bioresour. Technol. 2010, 101, 3780−3783.(40) Levis, J. M.; Barlaz, M. A. Is biodegradability a desirable attributefor discarded solid waste? Perspectives from a national landfillgreenhouse gas inventory model. Environ. Sci. Technol. 2011, 45(13), 5470−76.(41) Environmental Protection Agency (EPA). Emissions &Generation Resource Integrated Database (eGRID), 9th ed., v1.0; year2010 emissions; http://www.epa.gov/cleanenergy/energy-resources/egrid/index.html.(42) Staley, B. F.; Barlaz, M. A. Composition of municipal solid wastein the United States and implications for carbon sequestration andmethane yield. J. Environ. Eng. 2009, 135 (10), 901−909.(43) EcoInvent Centre. EcoInvent data v.2.2; EcoInvent Reports No.1−250; Swiss Centre for Life Cycle Inventories: Dubendorf, 2007.Available from www.ecoinvent.org.(44) Chase, L. E.; Overton, T. R.; Balbian, D. R. Using liquid whey indairy rations; Dairy nutrition fact sheet September 2013 (Revised);Cornell Cooperative Extension: Ithaca, NY, 2013.(45) Flesch, T.; Desjardins, R.; Worth, D. Fugitive methaneemissions from an agricultural biodigester. Biomass Bioenergy 2011,35, 3927−3935.(46) Liebetrau, J.; Reinelt, T.; Clemens, J.; Hafermann, C.; Friehe, J.;Weiland, P. Analysis of greenhouse gas emissions from 10 biogasplants within the agricultural sector. Water Sci. Technol. 2013, 67 (6),1370−1379.(47) Gooch, C. A.; Pronto, J.; Labatut, R. New York State EnergyResearch and Development Authority (NYSERDA), Biogas Dis-tributed Generation Systems Evaluation and Technology Transfer,NYSERDA Project No. 6597; Evaluation of seven on-farm anaerobicdigestion systems based on the ASERTTI monitoring protocol:Consolidated report and findings, PRO-DAIRY Program; Departmentof Biological and Environmental Engineering, Cornell University:Ithaca, NY, 2011; http://nmsp.cals.cornell.edu/publications/extension/Ndoc2003.pdf.(48) Nielsen, O. K.; Lyck, E.; Mikkelsen, M. H.; Hoffmann, L.;Gyldenkærne, S.; Winther, M.; Nielsen, M.; Fauser, P.; Thomsen, M.;Plejdrup, M. S.; Illerup, J. B.; Sørensen, P. B.; Vesterdal, L. Denmark’snational inventory report 2008 − Emission inventories 1990−2006;Submitted under the United Nations Framework Convention onClimate Change; NERI Technical Report no. 667; NationalEnvironmental Research Institute, University of Aarhus: Denmark,2008; http://www.dmu.dk/Pub/FR667.pdf.(49) Clemens, J.; Trimborn, M.; Weiland, P.; Amon, B. Mitigation ofgreenhouse gas emissions by anaerobic digestion of cattle slurry. Agric.,Ecosyst. Environ. 2006, 112 (2), 171−177.(50) Amon, A.; Kryvoruchko, V.; Amon, T.; Zechmeister-Boltenstern, S. Methane, nitrous oxide and ammonia emissions duringstorage and after application of dairy cattle slurry and influence orslurry treatment. Agric., Ecosyst. Environ. 2006, 112, 153−162.(51) Hansen, T. L.; Sommer, S. G.; Gabriel, S.; Christensen, T. H.Methane production during storage of anaerobically digestedmunicipal organic waste. J. Environ. Qual. 2006, 35, 830−836.(52) Menardo, S.; Gioelli, F.; Balsari, P. The methane yield ofdigestate: effect of organic loading rate, hydraulic retention time, andplant feeding. Bioresour. Technol. 2011, 102, 2348−2351.

(53) Gioelli, F.; Dinuccio, E.; Balsari, P. Residual biogas potentialfrom the storage tanks of non-separated digestate and digested liquidfraction. Bioresour. Technol. 2011, 102, 10248−10251.(54) Lehtomaki, A.; Huttunen, S.; Lehtinen, T.; Rintala, J. Anaerobicdigestion of grass silage in batch leach bed processes for methaneproduction. Bioresour. Technol. 2008, 99, 3267−3278.(55) Rieger, C.; Weiland, P. Biogas − Messprogramm I 61Biogasanlagen im Verlich; Erstellt durch Johann Heinrich vonThu-nen-Institut (VTI): Gulzow, Germany, 2009.(56) Bruun, S.; Hansen, T. L.; Christensen, T. H.; Magid, J.; Jensen,L. S. Application of processed organic municipal solid waste onagricultural land - a scenario analysis. Environmental Modeling andAssessment 2006, 11, 251−265.(57) Ebner, J.; Williamson, A.; Trabold, T. Quantifying thegreenhouse gas impact of pathways for treatment of secondaryresources generated in the food supply chain. In Proceedings of theASME 2015 Power and Energy Conversion Conference, San Diego,California, June 28−July 2, 2015.(58) Huijsmans, J.; Hol, J.; Vermeulen, G. Effect of applicationmethod, manure characteristics, weather and field conditions onammonia volatilization from manure applied to arable land. Atmos.Environ. 2003, 37 (26), 3669−3680.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b01331Environ. Sci. Technol. 2015, 49, 11199−11208

11208