Research paper

Bioenergy Potential from Food Waste in California Hanna M. Breunig,*,† Ling Jin,† Alastair Robinson,† and Corinne D. Scown†,‡

†Energy Technologies Area, Lawrence Berkeley National Laboratory, Berkeley, California 94720, United States ‡Joint BioEnergy Institute, Emeryville, California 94608, United States

*S Supporting Information

ABSTRACT: Food waste makes up approximately 15% of municipal solid waste generated in the United States, and 95% of food waste is ultimately landfilled. Its bioavailable carbon and nutrient content makes it a major contributor to landfill methane emissions, but also presents an important oppor- tunity for energy recovery. This paper presents the first detailed analysis of monthly food waste generation in California at a county level, and its potential contribution to the state’s energy production. Scenarios that rely on excess capacity at existing anaerobic digester (AD) and solid biomass combustion facilities, and alternatives that allow for new facility construction, are developed and modeled. Potential monthly electricity generation from the conversion of gross food waste using a combination of AD and combustion varies from 420 to 700 MW, averaging 530 MW. At least 66% of gross high moisture solids and 23% of gross low moisture solids can be treated using existing county infrastructure, and this fraction increases to 99% of high moisture solids and 55% of low moisture solids if waste can be shipped anywhere within the state. Biogas flaring practices at AD facilities can reduce potential energy production by 10 to 40%.

■ INTRODUCTION Inefficiencies occur at all stages of the food supply chain, linked to complex factors ranging from market conditions and weather to consumer preferences, and these inefficiencies translate to an abundance of food waste. The U.S. generated approximately 38 million tonnes of municipal food waste in 2014, approximately 95% of which was landfilled.1 An enormous amount of energy, water, land, and other resources go into producing nutrition for humans.2 A recent analysis of food waste estimated that $218 billion is spent in the U.S. on growing, processing, transporting, and disposing of food and byproducts that go uneaten.3

Furthermore, because food waste biodegrades four times faster than typical paper products and 10 times faster than wood waste, it releases methane from landfills more quickly than most other organic waste, with 34−51% of generated methane escaping typical landfill gas capture systems.4−6 Landfilling uneaten solid organic material not only contributes to climate change and occupies land resources, but also eliminates the possibility of cycling the valuable nutrients and energy in food back into the economy. First and foremost, policy measures are necessary to ensure

source-reduction through changes in consumer behavior and improved harvesting, processing, and transportation methods.7

However, source-reduction alone will not be a sufficient strategy. Americans consume raw produce and livestock that have both edible and inedible parts, from local and nonlocal sources, and in quantities that require some level of centralized production and distribution. These biological materials can only be used for their original purposeto provide nutrition and

sustenance to humansfor a short window of time, and maintaining the value of food is not always possible. Food waste-to-energy strategies can help meet renewable energy targets, greenhouse gas (GHG) reduction targets, air quality standards, and divert waste from landfills.8

In this paper, the potential for converting California’s food waste to electrical and thermal energy is analyzed, including organic waste from the food supply chain: agricultural production, postharvest handling and storage, processing and packaging, distribution, consumption, and end-of-life. The objectives of this study are to determine the quantity, locations, and temporal variation in food waste generation, use these results to model regional and subannual electricity and heat generation potential, and gain insight into the roles of policy and technology in overcoming challenges associated with food waste utilization. California serves as a useful starting point for building an analysis framework that can be applied to the U.S. or globally because of its diversity and significance in national food production (40% of U.S. vegetables, 20% of dairy, and 70% of fruits, tree nuts, and berry production by revenue).9,10

Although previous assessments have estimated the total annual energy potential from food waste from retail and consumer waste streams and from food processors in California,11−14 this study is the first to assess food waste production at the

Received: September 9, 2016 Revised: December 23, 2016 Accepted: January 10, 2017 Published: January 10, 2017

Policy Analysis

pubs.acs.org/est

© 2017 American Chemical Society 1120 DOI: 10.1021/acs.est.6b04591 Environ. Sci. Technol. 2017, 51, 1120−1128

D ow

nl oa

de d

vi a

F L

O R

ID A

I N

T L

U N

IV o

n N

ov em

be r

21 , 2

01 9

at 1

5: 15

:2 5

(U T

C ).

S ee

h tt

ps :/

/p ub

s. ac

s. or

g/ sh

ar in

gg ui

de li

ne s

fo r

op ti

on s

on h

ow t

o le

gi ti

m at

el y

sh ar

e pu

bl is

he d

ar ti

cl es

.

subannual scale and to develop a spatially and temporally explicit model that integrates feedstock production and energy infrastructure capacity to estimate potential energy production. By accounting for infrastructure, logistics, and storage limitations, this study provides a more robust assessment of potential electricity and thermal energy generation and highlights key challenges that must be overcome to maximize this potential. Background and Motivation. Despite recent drought

conditions, California produced over 400 types of agricultural commodities in 2014.15 With a state population of nearly 40 million people, a large amount of produce is processed and consumed in-state. In 2014, 5.2 million wet tonnes of food waste were sent to disposal facilities, up 13% from the 4.6 million wet tonnes disposed in 2008.16 Recent federal and state regulatory action has created incentives to reduce the generation of organic waste, including food waste, and to divert remaining waste to composting and transformation. In 2015, the U.S. Department of Agriculture (USDA) and Environmental Protection Agency (EPA) announced the first national food waste reduction goal: 50% reduction in postharvest losses at the retail- and consumer-levels by 2030. New requirements in California for source-separation and recycling of commercial organic waste (Assembly Bill (AB) 1826) are intended to reduce GHG emissions and create opportunities for recycling manufacturing facilities; however, transformation to energy is not counted toward the statewide 75% solid waste diversion goal for 2020, mandated in AB 341. In 2006, Governor Schwarzenegger signed Executive Order S- 06-06 mandating that 20% of renewable electricity comes from biomass; subsequent Bioenergy Action Plans have been released to promote technology innovation and guide market development for biobased products and energy.17

Diverting food waste for energy releases only biogenic carbon and is therefore considered renewable. Although some carbon in food waste would otherwise remain sequestered if the waste is landfilled, the resulting methane emissions outweigh this sequestration on a 100-year global warming potential (GWP) basis.6 Multiple technologies exist for converting organic materials, including food waste, into electricity, heat, transportation fuels like hydrogen, and chemical products.18

Extensive reviews have been conducted on the anaerobic digestion (AD) of the organic fraction of municipal solid waste,18 and food waste from retailers and consumers.19−22

Anaerobic digestion generates a methane-rich biogas and a nutrient-rich solid (digestate), the latter of which can be used as a low-carbon fertilizer.23 The methane can be cleaned and used onsite to generate electricity and heat in combined heat and power systems (CHP), injected into pipelines as renewable natural gas, or compressed into a biological natural gas (bioCNG) transportation fuel. Life-cycle assessments of incineration, composting, AD, and landfill treatment technol- ogies for food waste find that AD leds to the greatest reduction in carbon dioxide as long as biogas is captured and used for energy.23,24 Fats, oils, and grease (FOG) can be used in AD facilities or converted to liquid fuels. For example, yellow grease, the used cooking oil from the food industry, is a suitable feedstock for biodiesel.25 Not all food waste types are well suited for anaerobic digestion. Waste with moisture content (MC) below ∼50% and waste with high lignin content are better suited to thermochemical processes like combustion and gasification. Combustion of food waste like nut hulls and shells generates heat, electricity, and a nutrient-rich ash that can be

applied to land. A number of solid biomass power plants currently accept low moisture food waste like rice hulls and olive pits.26,27

■ MATERIALS AND METHODS Food waste is defined here as organic materials wasted within the food supply chain, including food waste generated during harvest, food processing, retail, and in eating establishments and consumers’ homes. For example, this distinction includes olive pits, but excludes olive tree branches and other woody/ herbaceous crop residues. In terms of transformation technologies, our study focuses on electricity and thermal energy generation. Direct combustion and AD of food waste are assessed because they are the most mature conversion technologies for food waste, are capable of handling highly heterogeneous food waste streams, and generate products which have established markets in California. Hydrogen or liquid fuels could become attractive in the future with technology advancements.28−30

Food Waste Meta-Analysis and Inventory. To deter- mine key methodological differences and the data quality/ completeness associated with existing studies, a meta-analysis of food waste inventories and energy potential literature is performed (Supporting Information (SI) Sections 1 and 2). Assumptions used to estimate food waste yields (e.g., fraction of MSW that is food waste) and potential electricity and heat generation (e.g., efficiencies) vary across previous assessments of California, and are used to recalculate and compare results.11−14,31,32 Results from previous assessments are compared with data from state33 and national15 agricultural surveys, municipal solid waste databases,34,35 and personal communication with food bioenergy program managers and food processors36−38 to develop the assumptions and method- ologies used in this study (SI, Sections 4.4, 4.5, 6).39,40

Waste production data is collection and disaggregation to develop a food waste inventory by month and county of origin for California for 2014. For ease of comparison, totals for high moisture solid (HMS) and low moisture solid (LMS) wastes are reported in bone dry tonnes (BDT). A bottom-up approach is used to estimate agricultural culls, where county level production data from the 2014 NASS33 for each type of produce is multiplied by a “cull multiplier”, which assumes that total available harvest is equal to the sum of reported production and cull production (SI, Section 3.1). Planting, harvesting, and peak harvesting dates in agricultural regions of California are characterized through a critical review of NASS agricultural survey and census data, and plant science literature for each crop (SI, Section 3.6). Cull production is distributed evenly over the harvesting time period, unless a peak harvesting or cull collection period is identified, in which case 80% of waste mass is distributed over the peak time period. County level food processing waste production was reported in a 2007 survey;13 annual waste production and locations from the survey are adjusted assuming constant waste yields, and that 2013 county level employment data can be used to scale 2007 production over time and space.41 New approaches are developed in this study to estimate county and monthly waste inventories for meat processors, distilleries, breweries, commercial bakeries, tortilla manufacturers, and fruit and olive pitters (SI, Section 3.2), while the approach developed in Williams et al. 2015 is used to model nut and rice hullers’ waste.11

Environmental Science & Technology Policy Analysis

DOI: 10.1021/acs.est.6b04591 Environ. Sci. Technol. 2017, 51, 1120−1128

1121

The mass of MSW generated in each county in 2014 is collected by quarter from the CalRecycle disposal database (SI, Section 3.3).16 The food waste fraction of MSW generated by retailers and consumers at the regional level is taken from a 2014 characterization study (SI, Table S6).39 Fats, oils, and grease production at the retail and consumer levels are determined using per-capita annual consumption data and waste yields from the USDA Economic Research Service (SI, Section 3.4).42,43 U.S. Census population data from 2014 is used to calculate total FOG generation at the county-level. Food waste technical availability is determined using two

indicators: (1) extent of source-separation practices and hauling networks, (2) strength of established markets for wastes (animal feed, rendering, etc.) (SI, Section 3.5).11−13 Technical availability is high for wastes like winery pomace, which are collected and stored during a production process and treated as wastes. Technical availability is very low for wastes like almond hulls which are used as animal feed, and low for wastes like vegetable culls which are difficult to collect and transport. Existing Anaerobic Digester Capacity. The EPA

estimates an excess AD capacity of 15−30% at roughly 140 wastewater treatment facilities in California.29,44 Excess capacity allows facilities to handle fluctuations in wastewater due to weather, population growth, and changes that occur when sources of wastewater and waste biomass relocate. Thus, near- term diversion of food waste can be achieved with existing AD infrastructure if there is sufficient capacity to codigest food waste alongside other organic feedstocks like wastewater solids.45,46 Specifically, high moisture solids (MC ≥ 55%), bakery wastes, and wastewater with high-biochemical oxygen demand (BOD) content are all candidates for codigestion. A meta-analysis of California WWTF databases and excess capacity estimates is included in the SI, Section 2. High-moisture solids are expensive to transport long

distances and are challenging and costly to store, although technologies are emerging for extending storage periods of food materials including wastes.47 Therefore, it is critical that AD capacity is matched to food waste production at appropriate spatial and temporal scales. Potentially available capacity is

estimated in this study for utilizing food waste in existing organic waste-to-energy facilities with wet, dry, or high-solids AD systems (SI, Section 4). Excess capacity in wet AD systems at WWTF (both municipal and private) is determined by calculating the potential to increase flow rate if facilities reduced their mean cell residence time (MCRT) to the EPA 40 CRF Part 503 regulation minimum of 15 days (eq 1).48 Adequate digestion of solids is necessary for producing biosolids that are stable and have low pathogen content, minimizing total biosolids produced, and generating biogas with higher methane content. Food waste is more biodegradable than wastewater solids, and thus more easily broken down in 15 days, however operators should take care when reducing their MCRT as the impact of changes on performance will vary between facilities.

τ = − −Q

V V Q

(15days) dilution (1)

where Q (million liter per day MLD) is the excess volumetric flow rate of food waste that could enter the digester, τ is the MCRT used at the facility (days), V is the volume of the digester (million L), and Qdilution is the volume of processing water needed to dilute food waste to 8% total solids (TS). Qdilution is adjusted in scenario alterative “d” (Table 1). Not all facilities provide data on digester volume (V) and

operational MCRT (τ), so an approach is developed to estimate flow rate increases based on average daily wastewater flow rate (Qinfluent). Data from 16 facilities in California that provided V, τ, and Qinfluent revealed that excess volumetric flow rate Q could increase Qinfluent by 0.1 to 2%.

32 These percentages are multiplied by Qinfluent for each WWTF, to give a range in excess capacity that is within an order of magnitude of the values calculated using engineering principles (eq 1), making it possible to model individual facilities, despite limited data. Capacity at operating organic waste-to-energy facilities are determined or approximated using available facility data on loading rates and waste composition (SI, Section 6).39 Finally, the excess volumetric flow rate Q [MLD] at each facility is converted to an excess mass loading rate [BDT/d] (SI, Equations S4−S7) and aggregated to the county and state level

Table 1. Electricity and Heat Generation Scenarios and Constraints. HMS = High Moisture Solids; TS = Total Solids

sensitivity analysis

scenario number and description feedstock availability

transport. extent parameter base case variation

variation label

1 energy from all food waste Gross WWTF influent scaling factor used to estimate food waste loading rates into AD

2% Qinfluent

0.1% Qinfluent

b

2 energy from technically available food waste Tech. Total solids specification for food waste slurry loaded into AD

8% TS no spec. c

3a energy from all food waste that can be treated using existing California infrastructure

Gross in-state Total solids specification for food waste slurry loaded into AD

8% TS 4% TS d

4a energy from all food waste that can be treated using existing county infrastructure

Gross in-county HMS storage duration limit 1 month 1 year e

5a energy from all technically available food waste that can be treated using existing California infrastructure

Tech. in-state methane utilizationa 85% 70−95% f

6a energy from all technically available food waste that can be treated using existing county infrastructure

Tech. in-county

aFor facilities that report biogas utilization. Facilities without biogas utilization are set at 0% utilization.

Environmental Science & Technology Policy Analysis

DOI: 10.1021/acs.est.6b04591 Environ. Sci. Technol. 2017, 51, 1120−1128

1122

(SI, Table S17); these rates are treated as constants. County or state aggregated food waste production is converted to [BDT/ d] for each month or for the whole year depending on the storage duration assumed in scenarios (alternative “f” Table 1). Existing Combustion Capacity. It is assumed that solid

fuel biomass power plants operating at less than 90% capacity factor (CF) are feedstock constrained and are capable of cofiring any dry biomass to reach a CF of 90%. Capacity factors are gathered from 2012 eGRID for all operating solid biomass power plants in California to determine excess capacity (eq 2).27,49 Power plants are cross-referenced with facility websites when possible to confirm operational status and feedstock mass loading rates. Combustion capacity and available waste biomass are adjusted for the mass of rice hulls, fruit and olive pits, and nut shells currently being diverted to specific solid biomass power plants.

= − ×P CF P(0.9 )i i NPi (2)

where P is the excess capacity (MW) at facility i, CF is the capacity factor (%), and PNP is the nameplate capacity (MW). Biogas combustion capacity is assumed to be unconstrained.

The analysis is rerun with facility specific capacity constraints reflected in biogas flaring practices (SI, Section 4.3). Electricity and Thermal Energy Generation Potential.

The six scenarios used to assess food waste bioenergy potential are summarized in Table 1. Potential energy generation from the conversion of gross food waste generated in California is estimated in Scenario 1. Potential energy generation from the conversion of only technically available food waste is estimated in Scenario 2. Scenarios 1 and 2 parallel the type of scenarios used in previous assessments.1

̀ Scenario 3 assumes that food

waste and FOG separated out of MSW are directed to organic waste-to-energy facilities until the statewide excess mass loading rate is met. Food waste and FOG not sent to organic waste-to- energy facilities are then sent to wet AD systems at WWTFs within the state. If excess capacity is still available, treatment of municipal wastes is followed by loading of state specific blends of processor food waste and then culls. This hierarchy is based on the following rationale: MSW is generated in urban areas near WWTFs and have centralized collection and hauling networks; organic waste-to-energy facilities have large upfront capital costs and have likely established contracts with MSW haulers; there are collection and disposal systems in place for food processors and less so for in-field culls. Scenario 4 is the same as Scenario 3 except food waste transportation and treatment is constrained to the county of origin. Scenarios 5 and 6 are the same as Scenarios 3 and 4, respectively, but

evaluate energy generation from technically available food waste. Scenario 6 is the most conservative scenario, as it assumes energy generation competes poorly with other markets for food waste, and that waste haulers will not export food waste to facilities in nearby counties. Variations b-f on Scenarios 3−6 reflect different operation practices at WWTF, and different storage limitations on HMS, than the base case (variation a). Methane production in Scenarios 1 and 2 is modeled using

wet-AD methane yield factors for specific types of HMS (including commercial bakery and tortilla wastes) (SI, Table S8). In the other scenarios, methane yields specific to the type of AD technologies at waste-to-energy facilities are used to model digestion of MWS food waste and FOG (SI, Table S10). Without in situ data, proxies are needed to estimate methane yields from blends of food waste collected throughout the state (Scenarios 3 and 5) or county (Scenarios 4 and 6) and sent to wet AD systems at WWTF. To develop these proxies, waste- specific methane yields are scaled by the volatile solids fraction in blends of culls or blends of food processor wastes and summed to estimate the methane yields for the blend. Survey data is used to identify the type of CHP technology

that is or potentially would be used at different scale facilities. Small facilities mainly use rich-burn and lean-burn engines, whereas large scale facilities use engines as well as micro- turbines, and fuel cells.50 Electric efficiency and average power- to-heat ratios are acquired from program performance standards (SI, Section 2.2).51 A CHP capacity factor of 85% and an energy content of 38.3 MJ/m3 (1027 BTU/ft3) methane are assumed. Combustion turbines, steam turbines, and combined cycles prime movers for CHP are not included as they are uncommon at WWTF,51 and because heat demand at WWTF reduces waste heat availability. Annual loss of methane due to flaring and fugitive emissions is estimated for each WWTF based on biogas utilization and flaring survey data (SI, Section 4.3).52 Net electricity generation from LMS is calculated based on an efficiency of 0.2, and using waste- specific dry-basis higher heating values (HHV) (SI, Table S7).11−13 It is assumed that LMS can be stored for up to a year to achieve steady loading rates to solid biofuel power plants in Scenarios 3−6. Additional details of the sensitivity analysis are provided in Table S18.

■ RESULTS Meta-Analysis. Recent assessments have estimated gross

production of food waste from MSW and from food processors in California at the state and annual level, as well as the fraction

Table 2. Previous Assessments of Food Waste Resource and Energy Potential in California. BDT = Bone Dry Tonnes; FOG = Fats, Oils, Grease

reference (Matteson & Jenkins 2007)12 (Williams et al. 2015)11 (Amon et al. 2012)13 (Kester 2015)32 this study (Scenario 1)

Gross Production [106 BDT/y] culls 0.8 processors 1.2 4.0 3.2 4.4 MSW - food waste 2.0 1.3 0.9 1.1 MSW - FOG 0.01 0.004 Electricity Generation [MWe] culls 97 processors 134 549 534 527 MSW - food waste 105 184 98 MSW - FOG 2 sum [MWe] 239 733 534 724

Environmental Science & Technology Policy Analysis

DOI: 10.1021/acs.est.6b04591 Environ. Sci. Technol. 2017, 51, 1120−1128

1123

that is technically available for energy production (SI, Section 1).11−13 With the inclusion of culled produce and new categories of food processing wastes, this study estimates 20% higher gross production of food waste. Previous studies have used combustion and AD to model energy production; however, this study is the first to model facility-specific excess combustion and AD capacity, technology- and feedstock- specific methane yields, and storage and transportation impacts. Despite generation potential varying between studies due to differences in food waste production, AD methane yields, and methane energy content, this study estimates a very similar generation potential from gross food waste (Scenario 1) as William et al. (Table 2).11 The range in excess AD capacity at WWTFs estimated in this study bounds values in previous assessments. Kester estimates that that 75% of all food waste from MSW could be treated in-state at WWTF (∼4100 BDT/ d). Using this estimate, it would appear that WWTF could treat 100% of technically available food waste from MSW (∼2570 BDT/d11 to ∼3030 BDT/d in this study). This study finds that 99% of technically available food waste from MSW could be treated in-county using the most optimiztic excess capacity assumptions. However, capacity assumptions that reflect existing food waste treatment projects ongoing in the state (Scenario 4b) resulted in treatment of only 23% of technically available food waste from MSW.

■ FOOD WASTE INVENTORY The monthly variation in HMS food waste production is less for food waste and FOG in MSW, spent grains, and meat residues than it is for culls and fruit and vegetable processors (SI Figure S6, Figure S7). Production of LMS varies significantly between harvest and nonharvest dates (SI Figure S8) and between urban and agricultural counties (SI Figure S9). Gross waste from food processors totals 4.4 million BDT/

y; for comparison, Williams et al. calculated a gross food processing residue generation of 3.9 million BDT in 2013.11

Statewide, 1.1 million BDT/y of food waste and 3560 BDT/y FOG is generated in MSW. Food waste from MSW is generated in all 58 counties in 2014, with the lowest production rate occurred in Alpine County in the fall (Oct-Dec) at 0.1 BDT/d and the highest production rate occurred during the same period in Los Angeles County (813 BDT/d). Uncertainty in the location of production is lowest for MSW and culls, and highest for food processor wastes, while uncertainty in tonnage is moderate for MSW and high for food processor wastes and culls. Food waste tonnage by type, county, and month is included in SI Section 5, Tables S13−S16 whereas statewide annual tonnage is provided in SI Table S19.

Electricity and Thermal Energy Generation Potential. The energy in all food waste generated in California could supply monthly electricity generation varying from 210 MWe in March to 1,490 MWe in September, with an annual average of 1120 MWe that is equivalent to 55% of installed biomass electric generation capacity in California (Scenario 1, Figure 1).53 Over 22 GJ of waste heat can be generated annually in addition to electricity. Nearly all gross HMS (99%) and 55% of gross LMS (including 97% of all rice hulls) could be converted to electricity and heat using only existing excess capacity, or capacity already dedicated to treating food waste like rice hulls (Scenario 3a). Multimonth storage of LMS evens out a majority of the seasonality seen in treatment of gross food waste (Scenario 1), resulting in a higher baseline at the state level. Electricity generation from technically available food waste (Scenario 2) is substantially lower than it is for gross food waste (Scenario 1), largely due to complete diversion of almond hulls to animal feed (Figure 1). Seasonality was even less pronounced in Scenario 5a, where technically available food

Figure 1. Total food waste converted to energy and total electricity generation potential per month are shown for Scenarios 1−6. Scenarios and variations are described in the table below the figure. Months are abbreviated in the x axis labels (J = January, A = April, J = July, O = October). Values are disaggregated by region and food waste type in SI, Section 7.

Environmental Science & Technology Policy Analysis

DOI: 10.1021/acs.est.6b04591 Environ. Sci. Technol. 2017, 51, 1120−1128

1124

waste is treated using existing excess capacity, as only 10% of culls are assumed to be diverted to bioenergy. Surprisingly, the biogas yield from state blends of culls is

relatively stable over the year, varying from 0.29 to 0.36 m3/kg TS (SI, Table S9). There is a slight dip in the summer due to production of an enormous amount of tomato culls, which have a low biogas yield, however root and tuber culls, which have a 36% higher biogas yield, are produced at the same time and help to offset the impact. Biogas yield from the state blend of food processing wastes stays around 0.44 m3/kg TS (high due to meat processing residues and spent grains), except for a noticeable dip from July through October when it drops to 0.34 m3/kg TS. This dip represents the most active season for wineries and fruit and vegetable processors, which generate residues with low biogas yields. For county level treatment of gross and technically available

food waste (Scenario 4a and Scenario 6a), monthly biogas yields are unique to each county and vary from 0.26 to 0.43 m3/ kg TS for cull blends and from 0.24 to 0.51 m3/kg TS for food processing residue blends. Seasonal variation in biogas yields resulting from changing composition of culls and food processing residues is limited in some counties and pronounced in others. Mismatch between the location of food waste production

and the location of energy facilities resulted in ∼60% lower average electrical generation in Scenario 4a than in Scenario 1 (Figure 2). Generation potential from HMS is further reduced by 25% when methane losses due to facility specific flaring practices are included (ranging from 11 to 38%). Local combustion capacity is limited in the Central Valley, where over 95% of gross LMS is produced; only 23% of gross LMS can be converted to energy in Scenario 4a even with multimonth storage. Calculations of existing excess AD capacity is more sensitive

to changes in the food waste to wastewater volumetric loading ratio (Scenario 3b) than the total solids specification for the

food waste slurry dilution (Scenario 3b,c). Scenario 3b reflects the operational practices used at WWTF currently accepting HMS; under these conditions, only food waste and FOG from MSW are treated, generating 25 MW. Wastewater from food processors is not included in these totals as it is unclear what fraction is already being treated at WWTFs (SI, Section 3.2.6). Annual wastewater from food processors contains 158410 BDT of BOD5 and has the potential generate 56 million m

3 of methane if 100% is codigested (20 MWe and 640 MJ waste heat).

■ DISCUSSION This study assesses the use of AD and direct combustion to convert food waste into electricity and thermal energy in California. Between 10% and 99% of gross HMS and can be digested using state AD infrastructure and in the same month of production, and between 10% and 66% can be digested in- county using AD infrastructure and in the same month of production. These large ranges reflect the uncertainty regarding excess capacity for food waste codigestion at WWTFs and organic waste-to-energy facilities. Accounting for technical availability (removing losses and currently utilized fractions) for waste best suited for AD results in potential utilization ranging from 37 to 100% for in-state, and 37 to 99% for in-county. Only 55% of gross LMS (including 97% of total rice hulls) can be converted to energy using excess capacity at in-state solid biomass power plants, while only 27% can be converted to energy if LMS must be utilized within the county of origin. This is concerning as over 90 MW of solid biomass installed capacity is going offline in 2016 alone.55 Additional LMS waste from forests, resulting from recent droughts and bark beetle infestations, will result in even more competition at composting and organic transformation facilities and solid biomass power plants.56 Fuel-switching at natural gas power plants is a possible solution for decreasing the new capacity needed to handle

Figure 2. Facility-level electricity generation capacities for treating food waste are mapped over county-level annual food waste-to-energy potential (Scenario 1e). New solid biomass combustion capacity and new AD infrastructure (converted to biogas combustion capacity [MWe] for ease of comparison) needed to reach gross potentials are shown in the maps on the right. Values reflect the assumption that facilities with existing excess AD capacity have unconstrained combustion capacity. Facility addresses and counties mapped using 2016 TIGER shapefiles.54

Environmental Science & Technology Policy Analysis

DOI: 10.1021/acs.est.6b04591 Environ. Sci. Technol. 2017, 51, 1120−1128

1125

LMS, but will likely require policy incentives, given the cost of retrofitting equipment for fuel-switching.57

As shown in Figure 2, the construction of 122 MWe of new AD capacity is still necessary for in-county utilization, even with the possibility of multimonth storage of HMS, as a number of counties with high cull and processor waste production have low populations and thus low AD capacity at local WWTFs. Widespread storage of HMS at AD treatment facilities in urban areas is unlikely in the near future due to cost, odor, and health concerns. Uncertainty, Data Gaps, and Future Work. Increased

availability of data on waste generation, and AD system capacity and operation would help reduce the uncertainty associated with estimates of energy generation potential. The temporal and spatial specificity of input data used to develop the food waste inventory is not uniform. Input data is collected from annual agricultural surveys, quarterly MSW disposal reports, and annual employment and per capita FOG consumption data. Input data is reported at the county level, while harvesting periods and MSW composition are reported at the regional level. Data needed to estimate excess capacity in AD systems is limited, and the uncertainty is bracketed through the use of low and high estimates in scenarios. Even with optimiztic assumptions regarding available existing excess capacity, twenty-three counties are likely to be capacity-limited at WWTFs year round, and 13 counties are capacity-limited at WWTFs during part of the year, resulting in 0.8 million BDT/y of food waste going untreated. Capacity at AD systems at food processing facilities, proposed food transformation facilities, and at dairy operations may be available to treat this waste, but are not included in this study due to limited data on current waste practices and AD design. Personal communication with operations managers at each facility will be essential to filling this data gap. Similar to the progression of the Billion-Ton studies, future work will estimate food waste fuel price points and system costs to constrain generation potential.25

The fraction of biogas currently generated at WWTF that is flared or vented remains another key source of uncertainty. It is a federal regulation that AD systems have flaring equipment installed and meet a conversion efficiency of methane to carbon dioxide between 95 and 99% during flaring events. Some facilities (∼30%) use the biogas for onsite CHP, but most facilities do not have CHP technology and either combust biogas in boilers (∼40%) or flare the biogas (∼30%).50 In 2011, total CHP capacity at WWTF in California was 63 MW.51 The use of flaring is driven by a range of factors including biogas quality56 and gas storage capacity.36 Data on deliberate venting is of course limited, as is data on fugitive methane emissions (leakage). An approximation is developed for modeling flaring of biogas (SI, Section 4.3), which revealed that nearly 100% of biogas from WWTF is flared, vented, or lost in 11 counties. Such practices need to change to capture the energy potential of food waste digestion. Policy Recommendations. Policy incentives to encourage

(1) the separation of food waste from MSW streams, (2) the transformation of food waste to energy (by including energy production as a diversion option in recycling policies), and (3) higher market values for energy byproducts, will be key in reducing food waste disposal and increasing food waste energy production. Regarding the third point, organic waste-to-energy facilities facing seasonal or weak markets for compost and byproducts are at risk if local policies prohibit food waste byproducts from entering landfills (e.g.: Alameda County

ACWMA Ordinance 2012-01). Furthermore, standards pro- hibiting the mixing of food waste with biosolids from human waste force AD facilities to reserve whole digesters for food waste treatment. Some barriers, like underreporting biogas flaring at AD facilities and incentivizing waste haulers to deliver organics to recycling and energy facilities may be quicker to resolve through policy than barriers like limited facility space for preprocessing and combustion equipment. Educational outreach to food processors, farmers, and WWTF managers is needed, as these actors frequently do not have the resources or experience to determine the best course of action for their unique waste streams or for becoming energy generators.

Global Perspective. As this study demonstrates, the importance of localized food waste management and data collection cannot be understated, as potential energy generation is dependent on the availability of blends of food waste which are suitable for processing in nearby systems with existing excess handling and conversion capacity. Developing countries generally lack centralized waste, recycling, wastewater, and energy infrastructure and regulation that could manage food waste at economies of scale.58 Industrialization leads to urbanization, as well as increasing per capita food con- sumption,59,60 and resulting growth and diversification of food supply chains will generate increasing quantities of food waste. Solutions broader than farm and home scaled biogas units will be needed to manage the waste.57 Research on the food, energy, and water nexus at the local and regional level can help stakeholders identify breakthroughs in technology and policy that provide food security, while enabling economic and sustainable flows of nutrients and energy.

■ ASSOCIATED CONTENT *S Supporting Information The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.est.6b04591.

Details of the meta-analysis, methods, and results sections, and a discussion on wastewater from food processors (PDF)

■ AUTHOR INFORMATION Corresponding Author *Phone: (510) 486-4046; e-mail: [email protected]. ORCID Hanna M. Breunig: 0000-0002-4727-424X Notes The authors declare no competing financial interest.

■ ACKNOWLEDGMENTS The research for this paper was financially supported by the California Energy Commission under agreement number EPC- 14-030. We thank S. Sherman, G. Kester, E. Bariani, K. Piscopo, N. Carr, H. Youngs, T. Pray, and P. Sethi for their insight and assistance gathering data. This work was also part of the DOE Joint BioEnergy Institute (http://www.jbei.org) supported by the U.S. Department of Energy, Office of Science, Office of Biological and Environmental Research, through contract DE- AC02-05CH11231 between Lawrence Berkeley National Laboratory and the U.S. Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a nonexclusive, paid-up, irrevocable,

Environmental Science & Technology Policy Analysis

DOI: 10.1021/acs.est.6b04591 Environ. Sci. Technol. 2017, 51, 1120−1128

1126

worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes.

■ REFERENCES (1) Advancing Sustainable Materials Management. 2014 Fact Sheet: Assessing Trends in material Generation, Recycling, Composting, Combustion with Energy Recovery and Landfilling in the United States; United States Environmental Protection Agency: Washington, DC, 2016; https://www.epa.gov/smm/advancing-sustainable-materials- management-facts-and-figures. (2) Hic,̧ C.; Pradhan, P.; Rybski, D.; Kropp, J.r.P. Food Surplus and Its Climate Burdens. Environ. Sci. Technol. 2016, 50 (8), 4269−4277. (3) A Roadmap to Reduce U.S. Food Waste by 20%; Rethink Food Waste Through Economics and Data (ReFED); United States, 2016; https://www.refed.com/downloads/ReFED_Report_2016.pdf. (4) Barlaz, M. A.; Chanton, J. P.; Green, R. B. Controls on landfill gas collection efficiency: instantaneous and lifetime performance. J. Air Waste Manage. Assoc. 2009, 59 (12), 1399−1404. (5) Staley, B. F.; Barlaz, M. A. Composition of municipal solid waste in the United States and implications for carbon sequestration and methane yield. J. Environ. Eng. 2009, 135 (10), 901−909. (6) WARM Component-Specific Decay Rate Methods: US EPA WAste Reduction Model (WARM); United States Environmental Protection Agency: Washington, DC, 2009; https://www3.epa.gov/warm/pdfs/ WARM_decay_rate_structure_10_30_2009.pdf. (7) Thyberg, K. L.; Tonjes, D. J.; Gurevitch, J. Quantification of Food Waste disposal in the United States: A meta-analysis. Environ. Sci. Technol. 2015, 49 (24), 13946−13953. (8) 2012 Bioenergy Action Plan; Publication number: CEC-300−2012; California Energy Commission, Efficiency and Renewables Division, Sacramento, CA, 2012; www.energy.ca.gov/bioenergy_action_plan/. (9) 2012 Census Highlights; United Stated Department of Agriculture, Washington, DC, 2012; http://www.agcensus.usda.gov/Publications/ 2012/Online_Resources/Highlights/. (10) Tom, M. S.; Fischbeck, P. S.; Hendrickson, C. T. Energy use, blue water footprint, and greenhouse gas emissions for current food consumption patterns and dietary recommendations in the US. Environ. Syst. Decis. 2016, 36 (1), 92−103. (11) An Assessment of Biomass Resources in California, 2013-Draft, PIER Contract 500−11−020; California Biomass Collaborative, Davis, CA, 2015; biomass.ucdavis.edu/publications. (12) Matteson, G. C.; Jenkins, B. Food and processing residues in California: Resource assessment and potential for power generation. Bioresour. Technol. 2007, 98 (16), 3098−3105. (13) California Food Processing Industry Organic Residue Assessment, CEC-500−2013−100; California Biomass Collaborative: Davis, CA, 2012; www.energy.ca.gov/2013publications/CEC-500-2013-100/ CEC-500-2013-100.pdf. (14) Tittmann, P.; Parker, N.; Hart, Q.; Jenkins, B. A spatially explicit techno-economic model of bioenergy and biofuels production in California. J. Transp. Geogr. 2010, 18 (6), 715−728. (15) California Agricultural Statistics Review 2013−2014; Fruit and Nut Crop; United States Department of Agriculture National Agricultural Statistics Service Pacific Regional Field Office (and the California Field Office): Sacramento, CA, 2015; https://www.nass. usda.gov/Statistics_by_State/California/Publications/California_Ag_ Statistics/. (16) Disposal-Facility-Based Characterization of Solid Waste in California, DRRR-2015−01546; CalRecycle; Sacramento, CA, 2015; www.calrecycle.ca.gov/Publications/Documents/1546/20151546.pdf. (17) Bioenergy Action Plan for California, CEC-600−2006−010; California Energy Commission Bioenergy Interagency Working Group, Sacramento, CA, 2006; http://www.energy.ca.gov/ bioenergy_action_plan/. (18) Current Anaerobic Digestion Technologies Used for Treatment of Municipal Organic Solid Waste; California Integrated Waste Manage-

ment Board: Davis, CA, 2008; www.calrecycle.ca.gov/Publications/ Documents/1275/2008011.pdf. (19) Co-Management of Municipal Solid Waste and Wastewater Treatment Plant Sludges Using an Anaerobic Composting Process; California Integrated Waste Management Board: Davis, CA, 1994; www.calrecycle.ca.gov/publications/Documents/Organics/2012015. pdf. (20) Ward, A. J.; Hobbs, P. J.; Holliman, P. J.; Jones, D. L. Optimization of the anaerobic digestion of agricultural resources. Bioresour. Technol. 2008, 99 (17), 7928−7940. (21) Long, J. H.; Aziz, T. N.; Francis, L.; Ducoste, J. J. Anaerobic co- digestion of fat, oil, and grease (FOG): a review of gas production and process limitations. Process Saf. Environ. Prot. 2012, 90 (3), 231−245. (22) Curry, N.; Pillay, P. Biogas prediction and design of a food waste to energy system for the urban environment. Renewable Energy 2012, 41, 200−209. (23) Schott, A. B. S.; Wenzel, H.; la Cour Jansen, J. Identification of decisive factors for greenhouse gas emissions in comparative life cycle assessments of food waste management−an analytical review. J. Cleaner Prod. 2016, 119, 13−24. (24) Levis, J. W.; Barlaz, M. A. What is the most environmentally beneficial way to treat commercial food waste? Environ. Sci. Technol. 2011, 45 (17), 7438−7444. (25) 2016 Billion-Ton Report: Advancing Domestic Resources for a Thriving Bioeconomy, Vol. 1: Economic Availability of Feedstocks; ORNL/TM-2016/160; Oak Ridge National Laboratory, Oak Ridge, TN, 2016; https://energy.gov/eere/bioenergy/2016-billion-ton- report. (26) Asmus, B., Bell Carter, CA. Personal Communication, January 2016. (27) Emissions & Generation Resource Integrated Database (eGRID) 2012. United States Environmental Protection Agency, Washington, DC, 2015; https://www.epa.gov/energy/egrid. (28) Ravindran, R.; Jaiswal, A. K. Exploitation of Food Industry Waste for High-Value Products. Trends Biotechnol. 2016, 34 (1), 58− 69. (29) Pham, T. P. T.; Kaushik, R.; Parshetti, G. K.; Mahmood, R.; Balasubramanian, R. Food waste-to-energy conversion technologies: Current status and future directions. Waste Manage. 2015, 38, 399− 408. (30) Sen, B.; Aravind, J.; Kanmani, P.; Lay, C.-H. State of the art and future concept of food waste fermentation to bioenergy. Renewable Sustainable Energy Rev. 2016, 53, 547−557. (31) Waste to Biogas Mapping Tool: Total Technical Potential estimate download; United States Environmental Protection Agency; Pacific Southwest, Region 9, https://epamap21.epa.gov/biogas/ updates.html. (32) Kester, G. California Association of Sanitation Agency (CASA), Sacramento, CA. Personal Communication and access to “Estimate of excess existing CA municipal wastewater treatment plant anaerobic digestion capacity available for organic waste” spreadsheet, February 2016. (33) California County Agricultural Commissioners’ Reports Crop Year 2013−2014; United States Department of Agriculture National Agricultural Statistics Service, Pacific Regional Field Office (and the California Field Office): Sacramento, CA, 2015: https://www.nass. usda.gov/Statistics_by_State/California/Publications/AgComm/ Detail/. (34) Disposal Reporting System (DRS): Single-year Countywide Origin Detail: 2014. 2015. CalRecycle, Sacramento, CA, 2016; http:// www.calrecycle.ca.gov/LGCentral/Reports/DRS/Origin/WFOrgin. aspx. (35) Generator-Based Characterization of Commercial Sector Disposal and Diversion in California; DRRR-2015-01543; CalRecycle, Sacra- mento, CA, 2015; www.calrecycle.ca.gov/Publications/Documents/ 1543/20151543.pdf. (36) Sherman, S., East Bay Municipal Utility District (EBMUD), Oakland CA. Personal Communication, February 2016.

Environmental Science & Technology Policy Analysis

DOI: 10.1021/acs.est.6b04591 Environ. Sci. Technol. 2017, 51, 1120−1128

1127

(37) Bariani, E. Bariani OliveOil, Zamora, CA. Personal Communi- cation, February 2016. (38) Carr, N., CalRecycle, Sacramento, CA. Personal Communica- tion, March 2016. (39) California Anaerobic Digestion Projects (a partial list, October 2015), CalRecycle, Sacramento, CA, 2015; www.calrecycle.ca.gov/ organics/conversion/ADProjects.pdf. (40) Anaerobic digestion of food waste; EPA-R9-WST-06-004; East Bay Municipal Utility District Report, Oakland, CA, 2008; https://archive. epa.gov/region9/organics/web/pdf/ebmudfinalreport.pdf. (41) County Business Patterns (CBP). 2013. United States Census Bureau, Washington, DC, http://www.census.gov/programs-surveys/ cbp.html. (42) Food Availability Data Service; Loss-Adjusted Food Availability Documentation; United Stated Department of Agriculture Economic Research Service (ERS), Washington, DC, 2013; https://www.ers. usda.gov/data-products/food-availability-per-capita-data-system/loss- adjusted-food-availability-documentation/. (43) National Nutrient Database for Standard Reference. United Stated Department of Agriculture, Agricultural Research Service National Agricultural Library, Beltsville, MD, 2015; https://www.ars.usda.gov/ northeast-area/beltsville-md/beltsville-human-nutrition-research- center/nutrient-data-laboratory/docs/usda-national-nutrient-database- for-standard-reference/. (44) Shang, Y.; Soroushian, F.; Whitman, E. J.; Zhang, Z. Co- digestion- Potential Increase of Renewable Energy Production from Waste for California. Proc. Water Environ. Fed. 2005, 2005, 6513− 6530. (45) Koch, K.; Helmreich, B.; Drewes, J. E. Co-digestion of food waste in municipal wastewater treatment plants: Effect of different mixtures on methane yield and hydrolysis rate constant. Appl. Energy 2015, 137, 250−255. (46) Koch, K.; Plabst, M.; Schmidt, A.; Helmreich, B.; Drewes, J. E. Co-digestion of food waste in a municipal wastewater treatment plant: Comparison of batch tests and full-scale experiences. Waste Manage. 2016, 47, 28−33. (47) Evaluation and Definition of Potentially Hazardous Foods - Chapter 3. Factors that Influence Microbial Growth; United States Food and Drug Administration: Silver Spring, MD, 2015; http://www.fda. gov/Food/FoodScienceResearch/SafePracticesforFoodProcesses/ ucm094145.htm. (48) Standards for the Use or Disposal of Sewage Sludge. In Code of Federal Regulations Title 40-CFR Part 503, Protection of Environment. 7−1-01 ed.. United States Environmental Protection Agency, Washington, DC, 1993.;https://www.epa.gov/biosolids/biosolids- laws-and-regulations. (49) California Forest Products and Biomass Power Plant Table. University of California, Berkeley Woody Biomass Utilization Group, Berkeley, CA, 2016; http://ucanr.edu/sites/WoodyBiomass/ Technical_Assistance/California_Biomass_Power_Plants. (50) Shen, Y.; Linville, J. L.; Urgun-Demirtas, M.; Mintz, M. M.; Snyder, S. W. An overview of biogas production and utilization at full- scale wastewater treatment plants (WWTPs) in the United States: challenges and opportunities towards energy-neutral WWTPs. Renew- able Sustainable Energy Rev. 2015, 50, 346−362. (51) Opportunities for Combined Heat and Power at Wastewater Treatment Facilities: Market Analysis and Lessons from the Field; United States Environmental Protection Agency Combined Heat and Power Partnership, 2011; https://www.epa.gov/chp/opportunities- combined-heat-and-power-wastewater-treatment-facilities-market- analysis-and. (52) National Biosolids Partnership Biogas Database, June 7, 2013 ed. Water Environment Federation, Alexandria, VA, 2013; http://epi9- prod.wef.org/. (53) Electric Generation Capacity & Energy. California Energy Commission Energy Almanac. Sacramento, CA, 2016; http://www. energy.ca.gov/almanac/.

(54) MAF/TIGER geographic database; 2016 TIGER/Line Shape- files. United States Census Bureau, Washington, DC, https://www. census.gov/geo/maps-data/data/tiger-line.html. (55) Solar Is in, Biomass Energy Is out  and Farmers Are Struggling to Dispose of Woody Waste; Los Angeles Times, Los Angeles, CA, January 14, 2016; http://www.abqjournal.com/705816/solar-is-in-biomass- energy-is-out-and-farmers-are-struggling-to-dispose-of-woody-waste. html. (56) Bark Beetles and Dead Trees; California Department of Forestry and Fire Protection, Sacramento, CA, 2016; http://www. readyforwildfire.org/Bark-Beetles-Dead-Trees/. (57) Nair, N.; Zhang, X.; Gutierrez, J.; Chen, J.; Egolfopoulos, F.; Tsotsis, T. Impact of Siloxane Impurities on the Performance of an Engine Operating on Renewable Natural Gas. Ind. Eng. Chem. Res. 2012, 51 (48), 15786−15795. (58) Thi, N. B. D.; Kumar, G.; Lin, C.-Y. An overview of food waste management in developing countries: current status and future perspective. J. Environ. Manage. 2015, 157, 220−229. (59) Buzby, J. C.; Hyman, J. Total and per capita value of food loss in the United States. Food Policy 2012, 37 (5), 561−570. (60) Parfitt, J.; Barthel, M.; Macnaughton, S. Food waste within food supply chains: quantification and potential for change to 2050. Philos. Trans. R. Soc., B 2010, 365 (1554), 3065−3081.

Environmental Science & Technology Policy Analysis

DOI: 10.1021/acs.est.6b04591 Environ. Sci. Technol. 2017, 51, 1120−1128

1128