Comparative carbon accounting of multi-stream food waste management: insights from a megacity case study

Background: Food waste (FW) management can contribute to emission reduction when low-carbon pathways such as anaerobic digestion or composting are adopted instead of landfilling.

Methods: This study quantified emissions across four treatment modules for five household (HHFW), one kitchen (KW), and two fruit and vegetable (FVW) waste sites.

Results: Baseline operations emitted 66,086.2 t CO₂, with HHFW and KW contributing 69.9% and 22.9%, respectively. Pollutant treatment dominated (93.4%), mainly from residue and sewage management. FW intensities ranged from 54.34–162.13 kg CO₂/t-HHFW, 122.1 kg CO₂/t-KW, and 59.2–84.3 kg CO₂/t-FVW. Resource recovery presented both offsets and burdens: grease recycling avoided up to −10.87 kg CO₂/t-KW, while composting added 74.7 kg CO₂/t-FVW. Compared to landfill disposal, the project reduced carbon intensity 7–16 fold, exceeding 800 kg CO₂/t at top-performing sites (V1 and H4). Relative to incineration, reductions were smaller and site-dependent, though most treatment streams still achieved net savings. Scenario optimization highlighted the transformative potential of advanced interventions: a Moderate pathway integrating anaerobic acid production (AOP) cut emissions 1.6–1.7 fold, while optimistic pathways, anaerobic digestion (AD) and insect bioconversion (BSFL), achieved net-negative emissions of −308.4 kg CO₂/t for HHFW and −117.41 kg CO₂/t for FVW, respectively.

Conclusion: These findings demonstrate that source separation, targeted resource recovery, and coupling bioenergy with value-added products can convert FW management from a carbon source to a net sink, supporting deep decarbonization strategies.

1 Introduction

Food waste (FW), a major and environmentally burdensome component of municipal solid waste (MSW), is rapidly increasing worldwide. Global FW totals about 1.6 billion tons per year (Tian et al., 2021). Across high-income countries, FW shares were 12%, 15%, 20%, 21%, 22%, 23%, 23%, 27%, and 30% of MSW for the United States, Germany, Switzerland, the Netherlands, France, Japan, Korea, the United Kingdom, and Singapore, respectively (Xu et al., 2015). In China, FW comprised 50%–60% of the 242.06 million tons of MSW generated in 2018, with an annual growth rate of 5.6%. (Xiao et al., 2022). Due to its high water content (80%–90%), low calorific value, and rapid biodegradability, FW produces leachate, odors, and methane (CH₄), hindering incineration and causing secondary pollution during landfilling, thereby disproportionately contributing to GHG emissions (Chen et al., 2020; Zhang et al., 2022). Beyond the solid-waste perspective, recent global assessments of the food system indicate that food production, processing, trade, and consumption account for roughly 30% of total final energy use and about one-quarter to one-third of anthropogenic GHG emissions, with around 6% of these emissions arising from food that is never consumed, underscoring the energy and environmental stakes of FW management (Corigliano et al., 2025). For example, landfill disposal of FW-rich MSW in eastern China emitted an average of 1.03 t CO₂/ton between 2007 and 2022, whereas source-separated collection with 60% sorting efficiency reduced emissions by 33 kg CO₂/t (Zhao et al., 2022). In response, China has implemented source separation policies across 46 megacities, achieving household participation rates as high as 83.6% in Shanghai and 49.7% in Nanjing (Chen et al., 2018; Wang et al., 2021). In Shenzhen, restaurant FW accounts for 14% MSW and its separation reduces carbon emissions by 10.6% (Yu and Li, 2021).

To improve resource recovery and treatment efficiency, FW in China is increasingly categorized by source, with household (HHFW), restaurant/canteen (KW), and market (FVW) streams accounting for 42%, 22%, and 10%–15% of source-separated waste (Zhang et al., 2024; Song et al., 2024). To the best of our knowledge, most existing life cycle analysis, treat FW as a homogeneous fraction, focusing on broad treatment comparisons such as landfill, incineration, composting, and anaerobic digestion (AD), conducted in Europe (Mondello et al., 2017), North America (Lee et al., 2020), and Asia (Liu et al., 2023; Zhang L. et al., 2025). These studies often overlook the impact of source-based classification on overall emissions. Evidence increasingly suggests that tailoring treatment technologies to FW stream characteristics can yield significant environmental benefits. For instance, FVW exhibits high suitability for producing liquid organic fertilizer via hydrothermal-biological integration (Wu et al., 2023; Abidin et al., 2024); KW, with its elevated lipid content, performs well in biogas recovery through AD (Weber, 2023; Naroznova et al., 2016); and protein-rich fractions of HHFW may be repurposed for animal feed or composted for high-quality fertilizers depending on nutrient profiles (Ho and Chu, 2018; Cui et al., 2023). In parallel, international studies on waste-to-energy systems show that coupling biogas production from organic residues with high-efficiency energy-conversion technologies, can further enhance energy recovery and reduce dependence on fossil fuels, highlighting the broader valorization potential of FW-derived biogas within low-carbon energy systems (Corigliano et al., 2011). Therefore, addressing the variations in carbon emissions across different FW streams is necessary to pinpoint the most appropriate treatment options.

Recent studies report wide variations in carbon emissions in FW treatment pathways, stemming from differences in system boundaries, energy baselines, and residue handling. Transport emissions alone range from 0.02 to 5.97 kg CO₂/t FW depending on haul distance, vehicle type, and facility centralization (Zhang L. et al., 2025). AD systems for biogas utilization can yield −81.4 to −72.8 kg CO₂/t FW for electricity generation and up to −88.5 kg CO₂/t FW when integrated with waste oil recovery (OIL) (Tian et al., 2021; Zeng et al., 2022). Insect-based bioconversion using black soldier fly larvae (BSFL) shows additional promise, with reported emissions from −24.8 kg CO₂/t FW in Singapore to 0.38 kg CO₂/t kW in Sweden and up to 88.4% GHG reduction compared to composting (COMP) (Ermolaev et al., 2019; Ramzy et al., 2025; Pang et al., 2020). The latter, although they produce fertilizers, account for the highest emissions as 926 kg CO₂/t in the USA or 123 kg CO₂/t in Korea (Kim et al., 2013; Perez et al., 2023). Organic acid production (OAP), primarily via acidogenic fermentation, achieved −6.86 kg CO₂/t FW (Zhang L. et al., 2025). However, emissions from digestate or residue treatment ranging from 93 to 125 kg CO₂/t FW (Zeng et al., 2022), are often underreported, limiting the reliability of net emission estimate. These discrepancies highlight the urgent need for a harmonized, source-specific framework that captures emissions across the entire FW management lifecycle.

This study conducts a stream-specific, process-based carbon accounting of FW management, focusing on HHFW, KW, and FVW. First, Scope 1- 3 emissions for collection, transport, treatment, waste valorization (OIL, COMP) and residue management (odor, sewage and pollutants) modules under harmonized system boundaries were quantified using empirical activity data. Second, within a scenario analysis framework, the study further compare lower-emission pathways (AD, VFA production, insect bioconversion), to guide climate-resilient, circular FW strategies. Additionally, this study incorporates regionalized parameters, providing benchmarking that aligns treatment choices with material characteristics. Additionally, by incorporating regionalized parameters, this study not only advances the understanding of FW steams management in China but also offers insights that can be replicate in other megacities.

2 Materials and methods

2.1 Study area, facilities and data collection

This study focused on Bao’an District, located in the western part of Shenzhen, Guangdong Province, China. The district covers an area of approximately 397 square kilometers and is home to over 3.9 million residents, generating a significant amount of FW. The enforcement of the “Shenzhen Municipal Domestic Waste Classification Management Regulations” on 1 September 2020, significantly improved the district’s waste management practices. To support the classification and treatment of FW, Bao’an District developed eleven decentralized, medium-sized, and technologically advanced facilities, collectively capable of processing nearly 1,500 tons of FW daily. This study selected eight of these facilities that were fully operational throughout 2023 for detailed analysis. For simplicity, the FW processing facilities are referred to as H1 to H5 for household food waste (HHFW), K1 for kitchen waste (KW), and V1 to V2 for fruit and vegetable waste (FVW). Detailed descriptions of these facilities, including full name, processing capacity, technological routing, pollutant control methods, and residue handling approaches, are presented in Supplementary Table S1.

By 2023, the integrated system for the collection, transportation, and treatment of HHFW, KW, and FVW, covered 622 residential quarters, 455 urban villages, 167 government agencies, and additional locations. The total annual volume of collected and transported waste in the district reached 515,379.56 tons, with a total transportation distance of 4,437,489.76 km (Source: Survey). Empirical data on collection and transport data were provided by the facility operator, including monthly energy bills and operational ledgers, collection point, vehicle type, transportation route, and weighbridge records at transfer stations. Additionally, facility operation teams provided data on electricity consumption and the types and quantities of chemicals or additives used during processing.

2.2 Accounting scope and boundaries

The study quantified emissions and potential reductions from multi-stream FW systems at different facilities and compares treatment pathways using scenario analysis. The system boundary encompasses all stages of FW management lifecycle, including collection, transportation, and downstream treatment. Treatment processes were grouped into three main categories: (1) pretreatment operations, such as sorting, crushing, three-phase separation, and others involving washing, dewatering and chemical treatments; (2) Resource utilization processes in (i) baseline scenario involves aerobic composting (COMP) and crude oil extraction (OIL), while in (ii) scenario analysis they involved organic acid production (OAP) through acidogenic fermentation (AF) and insect-based bioconversion (BIOCONV) using black soldier fly larvae (BSFL, Hermetia illucens). And (3) pollution control measures involving odor management, sewage treatment, and residue handling. An overview of the carbon accounting methodology in the current study is provided in Figure 1.

Figure 1

Figure 1. System boundary of FW management. The blue arrow represented the material-energy inputs, the red arrow represented GHG emissions.

COMP was implemented at H4, H5, and V2, with each facility incorporating its own systems for waste sorting, separation, and odor control. OIL extraction took place primarily at H2, H3, H5 and V1 including pretreatment. Incineration (INC) was the primary treatment method at K1, and residual treatment of H1. To evaluate alternative system layouts, decentralized BIOCONV, AD, and AF processes were also considered.

2.3 Carbon emission accounting methods

Carbon footprint accounting were conducted based on distinct process modules, including (i) Collection and Transportation Module, Pretreatment Module, Resource Recovery Module, and Pollution Control Module. Emissions were assessed for each module, considering direct, indirect, and avoided emissions associated with the project activities. For clarity, direct emissions are those from fuel combustion in vehicles or treatment processes; indirect emissions stem from inputs such as electricity consumption, chemicals used during treatment, and water used in pretreatment; and avoided emissions refer to emissions reductions achieved by substituting conventional practices with waste-to-energy technologies or other resource recovery methods, such as biodiesel production from waste oils or reduced synthetic fertilizer use through composting. The total carbon emissions were calculated as shown in Equation 1

$E_{system}=E_{direct}+E_{indirect}-E_{recovery}(1)$

where $E_{system}$ is carbon footprint of the current FW management system (t CO₂) or the emission intensity (kg CO₂/t FW) of the process activity. $E_{direct}$ refers to direct emissions, primarily from fuel combustion, and $E_{indirect}$ indirect emissions include emissions from purchased electricity, chemicals used, etc., during treatment process and $E_{recovery}$ represents the emissions avoided through energy and material recovery.

The emission reduction achieved through the FW treatment system was calculated by comparing the system’s carbon footprint with a baseline scenario. The baseline was defined as the conventional treatment of FW via landfilling or incineration, both accounted based on Supplementary Equations S30 and S31, respectively along with Supplementary Tables S4 and S5. The emission reduction was calculated using Equation 2. Absolute total emissions are expressed in tCO₂, while emission intensities are standardized per ton of food waste, reported as kg CO₂/t. The workflow of the current study is show in Figure 2.

$E_{reduction}=E_{system}-E_{baseline}(2)$

where $E_{reduction}$ refers to total emission reduction (t CO₂), $E_{baseline}$ represents the carbon emissions from the baseline treatment scenario (e.g., incineration or landfill).

Figure 2

Figure 2. Workflow of carbon accounting and scenario analysis for multi-stream food waste management.

2.3.1 Collection and transport emissions

The carbon footprint from waste collection and transport included contributions from both diesel and electric vehicles (Equation 3). Emissions from diesel vehicles were estimated using Supplementary Equation S2, while those from electric vehicles were calculated based on an electricity consumption and a grid electricity emission factor of 0.4403 t CO₂/MWh (Ministry of Ecology and Environment MEE, 2022). Carbon emissions were quantified based on the 100-year global warming potential (GWP100) (Eggleston et al., 2006). The greenhouse gases included in the assessment were carbon dioxide (GWP CO₂ factor: 1 t CO₂/t), methane (GWP CH₄ factor: 27.9 t CO₂/t), and nitrous oxide (GWP N₂O factor: 273 t CO₂/t) (Intergovernmental Panel on Climate Change IPCC, 2021). Transportation distances, diesel consumption, and electricity usage were reported in Supplementary Table S1.

$E_{Co2,trans}=E_{CO2,d}+E_{\text{CO}2,\mathrm{e}}(3)$

where $E_{Co2,trans}$ is the total emission to waste collection and transport (t CO₂), $E_{CO2,d}$ emission from diesel vehicles (t CO₂) and $E_{\text{CO}2,\mathrm{e}}$ is emission from electric vehicles (t CO₂).

2.3.2 Direct emissions

2.3.2.1 Resource recovery

During aerobic composting, the decomposition process generated direct GHG emissions of 4.0 g CH₄/kg and 0.3 g N₂O/kg (Ermolaev et al., 2019; Eggleston et al., 2006). Additionally, the conversion of crude grease to other products such as usable biodiesel, often subject to transesterification, emitted 85.2 tCO₂/t of oil produced (Li et al., 2014). The emission associated with resource recovery was computed following Equations 4 and Supplementary Equations S7–S24.

$E_{\text{Res}\mathrm{U}}=E_{\text{oil}}+E_{\text{comp}}+E_{\text{OAP}}+E_{\text{bioC}}+E_{\text{AD}}(4)$

where $E_{\text{ResU}}$ is the total carbon emissions in the resource utilization process (tCO₂), $E_{\text{oil}}$ is the emissions to crude oil extraction (tCO₂), $E_{\text{comp}}$ is emissions to composting process (tCO₂), $E_{\text{OAP}}$ is the emissions from the acid production process (tCO₂), $E_{\text{bioC}}$ is the emissions from the bioconversion using BSFL (tCO₂) and $E_{\text{AD}}$ is the emissions from anaerobic digestion process (tCO₂).

2.3.2.2 Pollution treatment

GHG emissions from COD and TN removal were estimated using methane and nitrous oxide emission factors of 0.0055 kg CH₄/kg COD and 0.00852 kg N₂O/kg TN, as reported by Yang et al. (Yang et al., 2025). After pretreatment, solid residues were mainly treated via co-incineration. According to the field investigation, 1 t of residue co-incineration required the addition of 1.1286 Nm³ of natural gas, which produced 25.05 kg of fly ash and 175.92 kg of slag, in addition to 189.92 kg of leachate.

2.3.3 Indirect emissions

2.3.3.1 Pretreatment

In addition to electricity, other forms of consumption caused indirect carbon emissions of water, and chemical consumption. The carbon emission factor for tap water was 0.82 kg CO₂/m³ (City Greenhouse Gas Working Group, 2022). The commonly used agent for three-phase separation is polyaluminum chloride (PAC) flocculant, the commonly used concentration is about 5%, and the corresponding carbon dioxide emission factor is 1.62 CO₂/kg (Zexing et al., 2025). The emission was calculated following Equation 5 and Supplementary Equations S3–S6.

$E_{\text{pre}\mathrm{T}}=E_{\text{sor}}+E_{C{O}_2,\text{cru}}+E_{3p\_\mathrm{O}}(5)$

where $E_{C{O}_2,\text{sor}}$ and $E_{C{O}_2,\text{cru}}$ are the Emissions from the sorting process (tCO₂) and crushing process (tCO₂), $E_{C{O}_2,3\mathrm{p}-\mathrm{o}}$ are the Emissions from three-phase separation and other pretreatment processes (tCO₂).

2.3.3.2 Resource recovery

The indirect emission factor associated with microbial inoculants is 0.27 tCO₂/t (City Greenhouse Gas Working Group, 2022). In anaerobic fermentation for acid production, microbial agents are added to boost efficiency, while alkaline agents like sodium hydroxide (NaOH) are used to regulate pH with emission of 1.59 kg CO₂/kg (City Greenhouse Gas Working Group, 2022). In the anaerobic digestion process, numerous chemicals are used including NaOH, Fe₂O₃, NaOCl, PAM, FeCl₃, and CaO,the emission is 18.4 kg CO₂/t FW (Yu et al., 2020).

2.3.3.3 Pollution treatment

Chemical deodorization can be achieved through biological or chemical scrubbers and plant-based sprays. In this model, microbial flora (fungus, 0.27 kgCO₂/kg), vegetable oil products (oils and fats, 0.96 kgCO₂/kg, 90% water), and sodium hydroxide solution (1.59 kgCO₂/kg, 90% water) are considered, and carbon black (2.12 kg CO₂/t) in the activated carbon process (City Greenhouse Gas Working Group, 2022). Sewage treatment includes external and self-treatment processes. Chemicals such as polyacrylamide and polyaluminum chloride 3.062 kgCO₂/kg (Braun et al., 2021) and 1.61 kgCO₂/kg (Liu et al., 2024), respectively. The emission associated with pollution management is accounted following Equation 6 and Supplementary Equations S25–S34.

$E_{\text{pol}}=E_{\text{od}}+E_{\text{sew}}+E_{\text{rsd}}(6)$

where $E_{\text{pol}}$ is the total pollution emissions generated during FW treatment (tCO₂), $E_{\text{od}}$ represents gaseous pollutants and malodorous compounds released during handling, pretreatment, and biological degradation of FW (tCO₂), $E_{\text{sew}}$ refers to wastewater discharges generated in the process, including leachate and liquid effluents from pretreatment and downstream steps (tCO₂), and $E_{\text{rsd}}$ accounts for the solid by-products and treatment residues that require further management, such as incineration etc (tCO₂).

2.3.4 Avoided emissions

The study estimated that composting leads to avoided emissions of 0.1543 by mitigating soil erosion and 0.1247 tCO₂/t through the substitution of synthetic fertilizers. Crude grease and fats were converted into biodiesel, which offsets emissions from conventional diesel production and use. The corresponding avoided carbon emission factor was −2.55 kg CO₂/kg of biodiesel (Yang et al., 2020; Bian et al., 2022). Kitchen waste fermentation broth contains 10 kg COD per ton of volatile acids, equivalent to 6.67 kg of methanol, avoiding 10.9 kg CO₂/t of broth (Zhang L. et al., 2025; Tang et al., 2018). Dried black soldier fly from 1 t of organic residue (5% yield) replaced an equal-protein amount of soybean meal, with an emission factor of 0.6 kg CO₂/kg (City Greenhouse Gas Working Group, 2022).

2.4 Sensitivity analysis

The Baseline scenario (S1) was defined to reflect existing FW treatment practices in 2023, with waste oil recovery carried out at H2, H3, H5, and K1, and composting applied at H5, H4, and V2. Back-end operations involved chemical or negative-pressure odor treatment, a mix of on-site and municipal sewage treatment, and partial residue incineration. Building on this configuration, two alternative strategies were designed. The Moderate scenario (S2) replaced composting with three-phase separation and anaerobic acid fermentation at selected sites (H1, H5, H4, and V2), while increasing residue recycling and reducing incineration to ≤20%; odor and sewage treatment followed the S1 setup. The Optimistic scenario (S3) further enhanced recycling by implementing 100% biogas production after grease recovery at H1–H5 and complete saprobic bioconversion at V1–V2, while maintaining the same back-end approach as S2, with residue recycling above 80% and incineration limited to 20%. The assumptions made in each scenario are as follows and summarized in Table 1.

Table 1

Table 1. Scenario assumptions for FW operation.

3 Results and discussion

3.1 Food waste material flow analysis

The quantity of FW generated in Baoan district, Shenzhen, during the 5-year period, alongside the FW sorting rate (Supplementary Figure S1). HHFW increased dramatically, rising nearly 30-fold from 10,655 tons in 2019 to 321,346 tons in 2023. KW also grew significantly, almost doubling over the same period, from 66,081 tons to 124,085 tons (an 88% increase). FVW showed more moderate fluctuations: it increased from 38,214 tons in 2019 to 69,814 tons in 2023, representing an 83% rise, with a temporary peak of 63,620 tons in 2022. Importantly, the FW sorting rate demonstrated continuous progress, improving from 5.8% in 2019 to 25.5% in 2023. This steady upward trend highlights 2023 as a turning point for advancing waste management strategies and in-depth study.

The mass flow of FW processing across various sites is summarized in Figure 3a. Among the 515,379.56 t of FW processed, HHFW was the most dominant with 62.35%, followed by KW at 24.08%, and FVW with 13.57%. This distribution is consistent with previous findings that highlight HHFW as the predominant FW category in urban settings (Herzberg et al., 2020). Three-phase separation was the most widely applied pretreatment method in this study, treating 68.19% of FW. This widespread application reflects a national trend in China, where thermal pretreatment using steam or hot water for three-phase separation has been reported to dominate KW management, accounting for 76.1% of currently operational projects (Wu et al., 2021). The high adoption rate is largely due to its effectiveness in separating waste oil of FW in improving the biodegradability of the liquid fraction for subsequent anaerobic digestion (Liu et al., 2022). Preprocessing operations, including sorting and crushing, were applied to 12.35% and 21.93% of the total weight, respectively, serving as essential steps for particle size reduction and impurity removal. Additionally, 41.83% underwent other forms of pretreatment, such as pulping and dewatering, aimed at improving downstream recovery efficiency. Regarding resource utilization, grease recovery accounted for the largest share of valorization pathways, representing 37.09% of FW amount. This highlights the growing emphasis on lipid extraction for bioenergy production and industrial reuse. Composting contributed 9.19%, supporting nutrient recycling and soil enhancement efforts. its lower share in urban areas is due to limited land and market demand. However, decentralized composting systems have been shown to offer sustainable and cost-effective alternatives. A study in Beijing demonstrated that decentralized composting can alleviate FW management gaps, reduce transport costs, and lower carbon emissions (Shen et al., 2025). In total, 44.52% of the FW was directed toward material or energy recovery routes, indicating meaningful progress toward circular economic objectives.

Figure 3

Figure 3. Mass-energy flow of FW transport and processing. (a) The mass flow is expressed in tons, and the percent value is proportional to total amount of FW in each subcategory. (b) Electricity (KWh/t) and diesel (L/t) consumed across each site of FW categories.

The energy consumption and intensity of FW collection and transportation is shown in Figure 3b and Supplementary Table S2. Electricity is the main energy source. HHFW consumes between 1.23 and 5.00 KWh/t (H3; H4). KW at K1 uses both electricity (16.87 KWh/t) and diesel (3.1 L/t). FVW relies only on electricity, at 6.76 and 1.42 KWh/t, V1 and V2 respectively. Sites like K1 and V2 show notably higher energy intensity per transport distance, with 3.66 × 10⁻⁴ KWh/t·km and 6.44 × 10⁻⁵ KWh/t·km.

3.2 Module-wise emission of food waste management

3.2.1 Collection and transportation

The carbon emissions associated with the collection and transportation of FW varied significantly across the FW sites and categories (Figure 4a). KW site K1 was the dominant contributor, accounting for 1,415.03 tCO₂, approximately 70% of the total 2,020.64 tCO₂ emissions of the projects. HHFW sites (H1–H5) together contributed 438.28 tCO₂ (21.7%), while FVW sites (V1–V2) accounted for 167.33 tCO₂ (8.3%). The contribution of FW collection and transport to total emissions in this study ranged from 0.6% at H3 to 9.34% at K1 (KW), with an average contribution of 3.06% across all FW streams (Figure 4e). When disaggregated by waste type, the contributions are comparable for HHFW and FVW, each averaging approximately 3.1%. These values are slightly higher than previously reported for major Chinese cities, including Shenzhen (2.8% in 2020), Beijing (0.56% in 2017) and Shanghai (2.2% in 2015) (Zhang et al., 2024; Xin et al., 2020; Jiang et al., 2020). These underscore the sensitivity of transport-related emissions to local operational conditions, even within decentralized waste management systems.

Figure 4

Figure 4. Total Carbon emissions (tCO₂) and module emission intensity (kg CO₂/t) of FW management modules across eight sites. (a) Collection and treatment, (b) pretreatment module including sorting and crushing preprocessing, and three-phase separation or dry-wet separation, (c) resource recovery module includes grease, compost and organic acid bioproducts recovery, and (d) pollution treatment involves odor, sewage and residue control and treatment. Process contribution to total emission across food FW) groups (e) and module contribution with treatment facilities (f). Redline + circle symbol stands for total emission intensity within each module. H1-H4 stands for Household food waste (HHFW) sites, K1 is the canteen/restaurant kitchen waste (KW) sites and V1-V2 are Fruit and vegetables waste (FVW) sites. Grey Band is a declination of smooth visualization of the three groups of FW.

Despite their lower share in total waste volume, the KW site exhibited a substantially higher unit emission intensity, with K1 recording 11.4 kg CO₂/t, well above those of HHFW sites (0.58 at H3 to 2.36 kg CO₂/t at H4) and FVW sites (0.67 at V2 to 3.19 kg CO₂/t at V1). Notably, H4 had the highest intensity among HHFW sites, though with a moderate total emission of 50.77 tCO₂. The comparatively low emission intensities observed at most sites are consistent with previous findings that transport-related emissions in decentralized systems typically range between 0 and 5.59 kg CO₂/t FW (Zhang L. et al., 2025). However, the elevated value at K1 represents a clear deviation. This discrepancy can be explained by the exclusive use of diesel-powered vehicles at K1, in contrast to the electric vehicles used at all other sites. Given evidence that electric vehicles emit approximately 50% less CO₂ than diesel equivalents (Schmid et al., 2021), the higher emission factor at K1 is consistent with expectations when accounting for vehicle type. The regression analysis further reveals that vehicle type, along with transport distance, FW volume, and population density, significantly influences transport emissions (Supplementary Table S3). Specifically, the analysis shows that transport emissions decrease by 0.5892 units when switching from diesel-powered vehicles to electric vehicles. This suggests that, within decentralized systems, carbon intensity from transport aligns with prior estimates when vehicle characteristics are considered, and highlights fuel type as a key differentiating factor.

3.2.2 Pretreatment

Pretreatment in FW processing serves to improve the efficiency of subsequent treatment stages by removing contaminants, reducing particle size, and separating moisture or recoverable materials. However, this phase can represent a non-negligible environmental burden. Previous studies have highlighted its substantial contribution to energy consumption and global warming potential (Carlsson et al., 2015; Jin et al., 2015). In this study, pretreatment accounted for 2.35% of total carbon emissions, with category-specific contributions of 608.05 tCO₂ (0.92%) for HHFW, 791.35 tCO₂ (1.2%) for KW, and 152.76 tCO₂ (0.23%) for FVW (Figures 4b,e,f). Emissions varied significantly across sites, with KW site K1 emitting the most (791.35 tCO₂), and FVW site V2 the least (10.43 tCO₂). Among HHFW sites, H1 recorded the highest (300.45 tCO₂), while H4 reported the lowest (12.91 tCO₂).

Emission intensities ranged from 0.47 kg CO₂/t at V2 to 8.46 kg CO₂/t at H3, reflecting substantial variability driven by site-specific operational efficiency and waste composition. For instance, dewatering of high-moisture substrates, common in HHFW, has been shown to significantly increase electricity demand. In this context, the three-phase separation process alone contributed 0.73% of total emissions and emphasizes the importance of optimizing this step to improve overall system performance and resource recovery. Supporting this, Carlsson et al. reported that enhanced separation achieved GWP savings of −204 kg CO₂/t HHFW, which was slightly lower than the −248 kg CO₂/t in the reference scenario, underscoring the role of broader system factors, such as biogas utilization and energy mix, in determining the net environmental benefit of pretreatment (Carlsson et al., 2015). Additionally, KW exhibited elevated emission intensity (6.38 kg CO₂/t at K1), likely due to its high fat and oil content, which necessitates more energy-intensive oil–water and solid–liquid separation processes. This aligns with previous findings that link increased energy demand to the complexity of lipid and suspended solid removal during pretreatment (Ren et al., 2009). However, a comparison with lab-scale studies shows a much higher emission intensity of 49 kg CO₂/t kW (Wu et al., 2021). This discrepancy suggests that, in real-world conditions, handling more challenging waste types such as lipid- or moisture-rich fractions results in significantly lower emissions than those observed under controlled experimental settings, where energy recovery and operational efficiencies may not be fully implemented.

3.2.3 Resource utilization

Recovered products from FW treatment, such as compost and grease, can displace conventional products from the agricultural and petrochemical sectors, contributing to overall GHG mitigation. In this study, their recovery benefits were operationally implemented and thus included in the emission accounting (Figures 4c,e). Resource recovery led to a net carbon credit of 799.82 tCO₂, with composting contributing to 2,281.27 tCO₂ of emissions (+3.45%) and grease recovery offsetting −1,481.45 tCO₂ (−2.24%) (Table 2). Process-specific emission intensities for FW treatment across sites (kg CO2/t). The net contribution of recovered products accounted for 1.21% of total system emissions.

Table 2

Table 2. Process-specific emission intensities for FW treatment across sites (kg CO₂/t).

Composting emerges as a significant emissions hotspot. The highest emissions were recorded at V2, reaching 1,647.28 tCO₂, corresponding to an emission intensity of 74.69 kg CO₂/t. For HHFW, site H4 emitted 531.66 tCO₂ (24.69 kg CO₂/t), while H5 generated 94.81 tCO₂ (27.14 kg CO₂/t). The HHFW intensities observed here are lower than the 44.9 kg CO₂/t for decentralized systems [46], suggesting benefits from improved process control or reduced biodegradable content. The FVW value is also far below the ∼403 kg CO₂/t reported for catering waste composting (Fan et al., 2022), likely due to its lower protein and lipid content, which limits nitrogen- and methane-related emissions. These findings reinforce that composting emissions are highly sensitive to feedstock properties.

In contrast, grease and oil recovery for biodiesel production resulted in net negative emissions where applied. The most substantial reduction occurred at KW site K1, which avoided −1,349.23 t CO₂ with a unit emission of −10.87 kg CO₂/t. HHFW category sites H2 and H3 also recorded carbon savings of −112.29 t CO₂ and −12.41 t CO₂, with corresponding intensities of −2.47 and −0.7 kg CO₂/t, respectively. Site H5 achieved a small reduction of −7.52 tCO₂ or −1.99 kg CO₂/t. Such variations are consistent with literature showing that biodiesel production potential depends largely on the lipid content of FW, with high-oil streams (>5%) being most suitable (Ahamed et al., 2016). Previous studies reported crude oil yields from FW ranging from 0.1% to 3.7%, with centralized plants achieving higher extraction rates around 2.9% due to processing of restaurant FW and thermal enhancement (Zhang L. et al., 2025; Jin et al., 2015). In a co-digestion plant in Sichuan Province, local restaurant contributed to a higher oil recovery rate of 4%, resulting in carbon offsets near −58.82 kg CO₂/t (Zeng et al., 2022). The substantial emissions avoidance at K1 likely reflects a combination of high oil content and efficient recovery, while lower savings at other sites may result from reduced oil availability or extraction efficiency.

3.2.4 Pollution treatment

Management of pollutants, including odor control, sewage treatment, and solid residue handling, accounts for 93.38% of total emissions, representing the dominant environmental burden in FW treatment facilities. Residue treatment contributed the largest share (51.77%), followed by sewage treatment (41.14%), with odor control contributing a minor fraction (0.5%) (Figures 4d,e). This pattern aligns with previous findings that digestate and wastewater management are not only major sources of GHG emissions but also significant operational cost drivers, comprising up to 30% of an anaerobic digestion (AD) plant’s total expenditure (Chen et al., 2021; Herbes et al., 2020). These results highlight the necessity of incorporating pollutant treatment within life-cycle system boundaries to achieve robust carbon mitigation and environmental sustainability outcomes in FW utilization pathways.

Odor control contributed only 0.5% of total pollutant-related emissions but remains a critical aspect of environmental management and public acceptance. Poorly enclosed bio-processing units and open dumping have been identified as major odor sources (Xiao et al., 2022), while deodorization systems frequently employing sodium hydroxide or sodium hypochlorite, introduce indirect GHG emissions. Odor treatment emissions were highest at H1, H2, and V1, with intensities of 0.70–1.63 kg CO₂/t (162.1–117.5 t CO₂), while H4 and H5 were lower (0.36–7.43 kg CO₂/t, 53.6–70.8 t CO₂), due to differences in system sealing and gas management. The process-specific emission intensities observed here fall within the range reported for industrial-scale plants, where odor control systems contribute on average 5.8 ± 0.2 kg CO₂/t (Zeng et al., 2022). Such systems, particularly those using centrifugal blowers for gas suction, represent notable emission points and, despite their low overall percentage contribution, can still provide meaningful GHG reduction.

Historically, inadequate leachate control generated severe environmental impacts, including aquatic ecotoxicity from heavy metals and organic pollutants and eutrophication from ammonia nitrogen (NH₃-N) and total phosphorus (TP) (Xiao et al., 2022). Advances in modern systems, such as solid–liquid separation and concentrated leachate treatment (Yan et al., 2023) have substantially mitigated these impacts, yet the high energy demand of this process remains a key target for carbon footprint reduction. Sewage treatment recorded high unit intensities at sites H1 (84.84 kg CO₂/t), H2 (65.91 kg CO₂/t), H3 (50.78 kg CO₂/t), and K1 (20.23 kg CO₂/t). These values are considerably higher than those reported in the literature, where wastewater treatment in industrial-scale FW co-digestion plants has been identified as the largest GHG contributor but with reported impacts of approximately 14.3 ± 0.7 kg CO₂/t (Zeng et al., 2022).

Residue treatment accounted for the largest single share of emissions, with H1 and K1 showing the highest unit intensities at 73.95 and 95 kg CO₂/t, respectively. Among FVW sites, V1 and V2 recorded 36.6 and 8.44 kg CO₂/t, while H3 and H4 had lower values of 36.79 and 16.17 kg CO₂/t, respectively. The highest emissions observed in this study (K1: 99.83 kg CO₂/t) fall within the range reported by Zeng et al. (93–125 kg CO₂/t) for digestate and solid residue management, primarily attributed to dewatering and drying processes. In comparison, the carbon footprint of transporting these residues and related by-products from decentralized facilities is substantially lower. Although source classification can increase emissions from collection and transportation due to more complex logistics and lower vehicle loads (Zhang et al., 2024), secondary transportation typically contributes minimally, usually between 0 and 5.59 kg CO₂/t FW (Zhang L. et al., 2025). This highlights that while residue treatment dominates emissions, transportation plays a relatively minor role in the overall carbon footprint.

Effective management of pollutants, including odor, sewage, and residues, is central to minimizing emissions and improving the sustainability of food waste treatment systems. Integrated strategies such as anaerobic digestion, composting, and advanced biofiltration (e.g., multi-layer biofilters and biotrickling filters) can simultaneously reduce odorous compounds, volatile organic compounds (VOCs), and nutrient loads, while supporting energy and nutrient recovery (Dhamodharan et al., 2019; Wang Z. et al., 2025). Hybrid approaches, including bioelectrochemical systems, further enhance contaminant removal and energy recovery, though challenges related to microbial stability and scalability remain (Kushwaha et al., 2025). Optimization of operational conditions can enhance the removal of carbon, nitrogen, and phosphorus while reducing greenhouse gas emissions (Sweetapple et al., 2014). Advanced monitoring and analytical tools are increasingly important for addressing emerging pollutants, including microplastics and endocrine-disrupting compounds. Future research should prioritize the synergy between pollutant treatment and overall emission reduction, exploring integrated approaches that maximize environmental benefits while maintaining energy and resource recovery. By combining technological innovation with optimized operational strategies, food waste management systems can become more resilient, efficient, and sustainable.

3.3 Emission reduction in food waste treatment operations

When benchmarked against landfill disposal, the project scenario delivered substantial climate benefits across all sites, both in total emissions and on an intensity basis (Figures 5a,b). Total emission reductions ranged from 2,827.3 t CO₂ at H5 (91.8% decrease) to 141,845.1 t CO₂ at H1 (78.9% decrease). Within the HHFW group, H4 achieved one of the largest percentage cuts (93.8%), while H5 despite having the smallest absolute reduction still maintained over 90% savings. K1 avoided 84,114.4 t CO₂ (84.7% reduction), and within the FVW group, V1 and V2 reached reductions of 93.3% and 90.4%, respectively. In terms of emission intensity, the project scenario achieved a 7–16-fold drop compared to landfill baselines, with landfill intensities between 771.5 and 884.2 kg CO₂/t falling to project values as low as 53.56 kg CO₂/t (H4) and 59.18 kg CO₂/t (V1). The largest per-ton reductions were observed at V1 (−817.2 kg CO₂/t) and H4 (−816.1 kg CO₂/t), representing >13-fold improvements. Even the smallest improvement, H1 at −609.4 kg CO₂/t, still reflected a nearly 4.8-fold reduction in carbon intensity. The climate benefits observed in this study where oil recovery and composting replaced landfill disposal are consistent with evidence showing landfilling as the most carbon-intensive FW management option. In Sweden, Eriksson et al. found landfill to be the least favorable for supermarket FW, mainly due to methane emissions, while Zhang et al. reported that landfill alone contributed 92.3% of total life cycle carbon emissions for HHFW (Zhang L. et al., 2025; Zhang Z. et al., 2025). Oil recovery is highly effective, with crude oil extraction offsetting over 140% of emissions through substitution, and converting waste crude oil to biodiesel reducing overall environmental impact by 38.68% (Liu et al., 2023; Zhang L. et al., 2025). Composting, despite variability from nutrient recovery, consistently outperforms landfill showing reductions around −938.8 kg CO₂/t, comparable to the 800 kg CO₂/t reduction seen in our fruit and vegetable waste treatment (Wang Y. et al., 2025).

Figure 5

Figure 5. Total and intensity carbon emission reduction in eight sites of FW management. (a) Total emission reduction both under landfill and incineration scenario and their intensity to landfill (b) and incineration (c).

Relative to incineration, the project scenario also achieved strong carbon savings, though with more variation across sites (Figures 5a,c). Total avoided emissions ranged from −446 t CO₂ at H5 (a 63.7% reduction) to −51,356.9 t CO₂ at H1 (a 57.6% reduction). In the HHFW group, H2 recorded a modest gain of 1,949.4 t CO₂ (5.7% reduction), while H3 and H4 showed slight increases in net emissions relative to incineration, suggesting site-specific constraints. Among KW sites, K1 achieved 8,464.1 t CO₂ avoided (26.4% reduction), whereas FVW sites V1 and V2 cut incineration emissions by 34.4% and 47.8%, respectively. On an intensity basis, the project pathway delivered reductions of up to −220.6 kg CO₂/t at H1 (∼1.7-fold drop) and −124.5 kg CO₂/t at H5 (∼1.6-fold drop), with V2 showing a −40.3 kg CO₂/t improvement (∼1.9-fold). The smallest intensity gains occurred at H2 (+42.9 kg CO₂/t) and H4 (+30.4 kg CO₂/t), where the project scenario marginally exceeded incineration’s baseline efficiency. These patterns indicate that while the landfill comparison reveals across-the-board dominance, the incineration benchmark highlights a few operational contexts where targeted optimization could close remaining performance gaps.

For the project strategy implication, if sent to landfill, the carbon emission reduction in the intensity of FW processing follows this order: V1 (−817.2 kg CO₂/t, −93.3%) > H4 (−816.1 kg CO₂/t, −93.8%) > V2 (−792.1 kg CO₂/t, −90.4%) > H5 (−788.8 kg CO₂/t, −91.8%) > H3 (−780.5 kg CO₂/t, −88.9%) > H2 (−766.7 kg CO₂/t, −86.9%) > K1 (−712.4 kg CO₂/t, −84.7%) > H1 (−609.4 kg CO₂/t, −79.0%). If sent to incineration, the carbon emission reduction intensity ranks as follows: H1 (−220.6 kg CO₂/t, −57.6%) > H5 (−124.5 kg CO₂/t, −63.7%) > H3 (−88.4 kg CO₂/t, −47.6%) > K1 (−71.7 kg CO₂/t, −35.9%) > V2 (−40.3 kg CO₂/t, −40.3%) > V1 (−15.2 kg CO₂/t, −34.6%) > H4 (+30.4 kg CO₂/t, +131.5%) > H2 (+42.9 kg CO₂/t, +57.5%). Based on the top three rankings, to maximize carbon emission reductions, the project should prioritize diverting the FVW stream from landfill. When focusing on incineration avoidance, emphasis should be placed on the HHFW stream.

3.4 Decarbonization pathways scenario optimization

This study further compares in Table 3 the emission of FW management from the Baseline Scenario (S1) with two other strategies. The Moderate Scenario (S2) introduces moderate interventions through anaerobic acid production (AOP) and incineration of residues, treating up to 300 tons per day. The Optimistic Scenario (S3) which incorporates advanced recycling technologies such as AD for biogas and BSFL bioconversion. For HHFW, emission intensities in S1 range from 117.5 (H2) to 162.1 (H1) kg CO₂/t. S2 reduces these emissions by approximately 1.6- to 1.7-fold, to 78.5–98.7 kg CO₂/t. S3 achieves net negative emissions ranging from −259.5 (H2) to −235.0 (H1) kg CO₂/t, representing a 3.3- to 3.6-fold decrease compared to S1. Notably, H4 exhibits an exceptionally low emission intensity of −308.4 kg CO₂/t in S3, nearly 5.7 times lower than its S1 value of 54.3 kg CO₂/t. For FVW, emissions in S1 vary from 59.2 (V1) to 84.3 (V2) kg CO₂/t, decreasing to 9.6–59.2 kg CO₂/t in S2, and further dropping to −117.4 (V2) to −67.8 (V1) kg CO₂/t in S3, indicating up to a 9-fold reduction at V2 compared to S1. KW emissions remain stable at around 122.1 kg CO₂/t in both S1 and S2 but fall sharply to −215.9 kg CO₂/t in S3, nearly a 2.8-fold decrease. These deep reductions are primarily attributed to the integration of Both AD for biogas recovery and BSFL bioconversion, which generate high-value outputs that offset fossil-based products. The coupling of bioenergy production with protein and lipid recovery creates a strong displacement effect on conventional feed, fuel, and fertilizer systems, thereby enabling net-negative outcomes (Singh et al., 2022), consistent with S3 highlighting its potential for carbon neutrality and beyond.

Table 3

Table 3. Carbon footprint comparison of FW treatment scenarios.

Each scenario captures the carbon emissions from three main modes. Emissions from the front-end mode showed no variation across scenarios (Figure 6). However, in the recycling mode, S2 delivered modest gains, with most sites showing 1–2-fold reductions in emission intensity relative to the baseline. For example, H1 shifted from 0 to −4.31 kg CO₂/t, while H4 improved from 24.69 to −31.08 kg CO₂/t, representing more than a 2-fold change in magnitude with a complete sign reversal. Sites already net-negative in S1, such as H2 and K1, deepened their removal capacity slightly. S3 produced transformational changes, the value decreased by 201.69 kg CO₂/t at V2 (74.69 to −127 kg CO₂/t) and by 338 kg CO₂/t at K1 (−10.87 to −348.87 kg CO₂/t), representing a more than 30-fold increase in magnitude. At H1, emissions plunged from zero to −338 kg CO₂/t, marking a dramatic shift from neutrality to a large carbon sink. These results highlight recycling as the most responsive stage to technological and operational upgrades.

Figure 6

Figure 6. Scenario-based analysis of emissions intensity across sites. S1, S2, and S3 denote the Baseline, Moderate, and Optimistic scenarios, respectively. The red circle marks the zero-reference threshold, distinguishing positive emissions from avoided emissions.

In the back-end treatment mode, reductions were more restrained. Under S2, H1 saw a 1.6-fold decrease (159.5–100.33 kg CO₂/t), H2 dropped by 1.5-fold (116.28–77.29 kg CO₂/t), while V1, H5, H4, V2, H3, and K1 remained unchanged from S1. S3 retained these Moderate-level reductions, yielding no further fold change in this stage. This stability suggests that back-end treatment has lower flexibility for emission improvements, and that most system-wide carbon savings are driven by upstream recycling interventions rather than downstream treatment optimizations.

4 Conclusion

This study provides a carbon emission accounting of FW management in a megacity, highlighting the importance of distinguishing between HHFW, KW and FVW. The results show that pollutant treatment is the most carbon-intensive stage, while resource recovery offers significant mitigation potential, especially through grease recycling. Scenario analysis reveals that integrating anaerobic fermentation, insect bioconversion, and targeted recovery strategies can transform FW systems from emission sources into net-negative contributors. These findings underscore the value of multi-stream classification, modular process optimization, and circular economy principles in designing climate-resilient waste management strategies for megacities.

Beyond the local regulatory context in Shenzhen, these findings also hold broader policy relevance. At the national scale, the stream-specific carbon accounting framework aligns closely with China’s “dual-carbon” goals and the “14th Five-Year Plan”, which emphasize source separation, waste reduction, and low-carbon urban infrastructure. The module-level emission estimates generated in this study can support forthcoming national requirements for carbon disclosure in MSW systems and provide a standardized approach for integrating FW treatment into city-level GHG inventories. Internationally, the methodology complements major global frameworks such as the EU Green Deal, the EU Waste Framework Directive, the U.S. EPA’s WARM model, and SDG 12.3 on FW reduction. These existing tools predominantly rely on generalized emission factors, while our stream-specific accounting approach demonstrates how differentiating FW types and pollutant-treatment burdens substantially improves the accuracy of GHG inventories. Finally, because the framework is modular and based on empirical activity data, it is transferable to other megacities and regions. Cities with similar characteristics, including high population density, rapid waste growth, or decentralized treatment infrastructure, can directly adapt this approach by substituting local energy mixes and regulatory baselines. As a result, the Shenzhen case provides a scalable template for designing climate-aligned, circular FW systems in both developed and emerging urban contexts.

Data availability statement

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

Author contributions

GY: Conceptualization, Formal Analysis, Investigation, Visualization, Writing – original draft, Writing – review and editing. ZZ: Funding acquisition, Supervision, Writing – review and editing. FY: Resources, Supervision, Writing – review and editing. JJ: Methodology, Supervision, Writing – review and editing. YG: Data curation, Investigation, Writing – review and editing. SC: Writing – review and editing. FF: Data curation, Validation, Writing – review and editing. XW: Writing – review and editing.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This work was supported by the National Key Research and Development Program of China (Grant No. 2018YFC1902900) and received strong support from the Harbin Institute of Technology, the District Urban Appearance and Environmental Management Center of Bao’an.

Acknowledgements

Special thanks are hereby extended!

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fenvs.2025.1709717/full#supplementary-material

References

Abidin, A. Z., Steven, S., Fadli, R., Nabiel, M. F., Yemensia, E. V., Soekotjo, E. S. A., et al. (2024). Influence of several physical parameters in enzymatic fermentation of vegetable and fruit waste to produce organic liquid fertilizer using MASARO technology. Results Eng. 23, 102567. doi:10.1016/j.rineng.2024.102567

CrossRef Full Text | Google Scholar

Ahamed, A., Yin, K., Ng, B. J. H., Ren, F., Chang, V. W. C., and Wang, J. Y. (2016). Life cycle assessment of the present and proposed food waste management technologies from environmental and economic impact perspectives. J. Clean. Prod. 131, 607–614. doi:10.1016/j.jclepro.2016.04.127

CrossRef Full Text | Google Scholar

Bian, R., Chen, J., Zhang, T., Gao, C., Niu, Y., Sun, Y., et al. (2022). Influence of the classification of municipal solid wastes on the reduction of greenhouse gas emissions: a case study of Qingdao city, China. J. Clean. Prod. 376, 134275. doi:10.1016/j.jclepro.2022.134275

CrossRef Full Text | Google Scholar

Braun, O., Coquery, C., Kieffer, J., Blondel, F., Favero, C., Besset, C., et al. (2021). Spotlight on the life cycle of acrylamide-based polymers supporting reductions in environmental footprint: review and recent advances. Molecules 27 (1), 42. doi:10.3390/molecules27010042

PubMed Abstract | CrossRef Full Text | Google Scholar

Carlsson, M., Naroznova, I., Moller, J., Scheutz, C., and Lagerkvist, A. (2015). Importance of food waste pre-treatment efficiency for global warming potential in life cycle assessment of anaerobic digestion systems. Resour. Conservation Recycl. 102, 58–66. doi:10.1016/j.resconrec.2015.06.012

CrossRef Full Text | Google Scholar

Chen, F., Li, X., Ma, J., Yang, Y., and Liu, G.-J. (2018). An exploration of the impacts of compulsory source-separated policy in improving household solid waste-sorting in pilot megacities, China: a case study of Nanjing. Sustainability 10 (5), 1327. doi:10.3390/su10051327

CrossRef Full Text | Google Scholar

Chen, S., Huang, J., Xiao, T., Gao, J., Bai, J., Luo, W., et al. (2020). Carbon emissions under different domestic waste treatment modes induced by garbage classification: case study in pilot communities in Shanghai, China. Sci. Total Environ. 717, 137193. doi:10.1016/j.scitotenv.2020.137193

PubMed Abstract | CrossRef Full Text | Google Scholar

Chen, T., Qiu, X., Feng, H., Yin, J., and Shen, D. (2021). Solid digestate disposal strategies to reduce the environmental impact and energy consumption of food waste-based biogas systems. Bioresour. Technol. 325, 124706. doi:10.1016/j.biortech.2021.124706

PubMed Abstract | CrossRef Full Text | Google Scholar

City Greenhouse Gas Working Group. (2022). China products carbon footprint factors database (In Chinese). Available online at: http://lca.cityghg.com/ (Accessed September, 2025).

Google Scholar

Corigliano, O., Florio, G., and Fragiacomo, P. (2011). A numerical simulation model of high temperature fuel cells fed by biogas. Energy sources. Part A Recovery, Util. Environ. Eff. 34 (2), 101–110. doi:10.1080/15567036.2011.584116

CrossRef Full Text | Google Scholar

Corigliano, O., Morrone, P., and Algieri, A. (2025). Navigating the challenges of sustainability in the food processing chain: insights into energy interventions to reduce footprint. Energies 18 (2), 296. doi:10.3390/en18020296

CrossRef Full Text | Google Scholar

Cui, G., Lü, F., Lu, T., Zhang, H., and He, P. (2023). Feasibility of housefly larvae-mediated vermicomposting for recycling food waste added digestate as additive. J. Environ. Sci. (China) 128, 150–160. doi:10.1016/j.jes.2022.07.016

PubMed Abstract | CrossRef Full Text | Google Scholar

Dhamodharan, K., Varma, V. S., Veluchamy, C., Pugazhendhi, A., and Rajendran, K. (2019). Emission of volatile organic compounds from composting: a review on assessment, treatment and perspectives. Sci. Total Environ. 695, 695. doi:10.1016/j.scitotenv.2019.133725

PubMed Abstract | CrossRef Full Text | Google Scholar

Eggleston, H., Buendia, L., Miwa, K., Ngara, T., and Tanabe, K. (2006). 2006 IPCC guidelines for national greenhouse gas inventories. Hayama, Japan: Institute for Global Environmental Strategies (IGES), 1–5. Available online at: https://www.ipcc-nggip.iges.or.jp/public/2006gl/index.html.

Google Scholar

Ermolaev, E., Lalander, C., and Vinneras, B. (2019). Greenhouse gas emissions from small-scale fly larvae composting with Hermetia illucens. Waste Manag. 96, 65–74. doi:10.1016/j.wasman.2019.07.011

PubMed Abstract | CrossRef Full Text | Google Scholar

Fan, Z., Dong, H., Geng, Y., and Fujii, M. (2022). Life cycle cost–benefit efficiency of food waste treatment technologies in China. Environ. Dev. Sustain. 25 (6), 4935–4956. doi:10.1007/s10668-022-02251-4

CrossRef Full Text | Google Scholar

Herbes, C., Roth, U., Wulf, S., and Dahlin, J. (2020). Economic assessment of different biogas digestate processing technologies: a scenario-based analysis. J. Clean. Prod. 255, 120282. doi:10.1016/j.jclepro.2020.120282

CrossRef Full Text | Google Scholar

Herzberg, R., Schmidt, T. G., and Schneider, F. (2020). Characteristics and determinants of domestic food waste: a representative diary study across Germany. Sustainability 12 (11), 4702. doi:10.3390/su12114702

CrossRef Full Text | Google Scholar

Ho, K. S., and Chu, L. M. (2018). Characterization of food waste from different sources in Hong Kong. J. Air and Waste Manag. Assoc. 69 (3), 277–288. doi:10.1080/10962247.2018.1526138

PubMed Abstract | CrossRef Full Text | Google Scholar

Intergovernmental Panel on Climate Change (IPCC) (2021). Climate change 2021: the physical science basis. Contribution of working group i to the sixth assessment report of the intergovernmental panel on climate change. Editors V. Masson-Delmotte, P. Zhai, A. Pirani, S. L. Connors, C. Péan, and S. Berger (Cambridge, United Kingdom Cambridge University Press). doi:10.1017/9781009157896

CrossRef Full Text | Google Scholar

Jiang, Y., Zhang, H., He, J., and Zeng, Y. (2020). Carbon emission of municipal solid waste in Shanghai. IOP Conf. Ser. Earth Environ. Sci. 555 (1), 012058. doi:10.1088/1755-1315/555/1/012058

CrossRef Full Text | Google Scholar

Jin, Y., Chen, T., Chen, X., and Yu, Z. (2015). Life-cycle assessment of energy consumption and environmental impact of an integrated food waste-based biogas plant. Appl. Energy 151, 227–236. doi:10.1016/j.apenergy.2015.04.058

CrossRef Full Text | Google Scholar

Kim, M. H., Song, H. B., Song, Y., Jeong, I. T., and Kim, J. W. (2013). Evaluation of food waste disposal options in terms of global warming and energy recovery: Korea. Int. J. Energy Environ. Eng. 4 (1), 1. doi:10.1186/2251-6832-4-1

CrossRef Full Text | Google Scholar

Kushwaha, B., Shukla, R., and Sharma, R. K. (2025). Operational efficiencies and sustainable bioprocessing in electro-fermentation and microbial fuel cells. Bioprocess Biosyst. Eng. doi:10.1007/s00449-025-03228-z

PubMed Abstract | CrossRef Full Text | Google Scholar

Lee, E., Oliveira, D. S. B. L., Oliveira, L. S. B. L., Jimenez, E., Kim, Y., Wang, M., et al. (2020). Comparative environmental and economic life cycle assessment of high solids anaerobic co-digestion for biosolids and organic waste management. Water Res. 171, 115443. doi:10.1016/j.watres.2019.115443

PubMed Abstract | CrossRef Full Text | Google Scholar

Li, H., Ebner, J., Ren, P., Campbell, L., Dizayi, B., and Hadavi, S. (2014). Determination of carbon footprint using LCA method for straight used cooking oil as a fuel in HGVs. SAE Int. J. Fuels Lubr. 7 (2), 623–630. doi:10.4271/2014-01-1948

CrossRef Full Text | Google Scholar

Liu, B., Tian, G., Zhou, K., Yang, Y., Yang, Z., Zhang, T., et al. (2022). Option of treatment scenario for food waste anaerobic digestion by-products based on comprehensive evaluation and scenario promotion benefits analysis. J. Clean. Prod. 374, 133861. doi:10.1016/j.jclepro.2022.133861

CrossRef Full Text | Google Scholar

Liu, T., Liu, H., Wei, G., Zhao, H., Gong, Y., Guo, X., et al. (2023). Comprehensive assessment of food waste treatment emission reduction in China: a life cycle perspective. Environ. Sci. Pollut. Res. Int. 30 (37), 87669–87684. doi:10.1007/s11356-023-28622-6

PubMed Abstract | CrossRef Full Text | Google Scholar

Liu, S.-Y., Liang, S., Chen, Z.-Q., Ma, Y.-G., Gao, W.-C., Tian, X.-Y., et al. (2024). Mathematical modeling and indirect carbon emission reduction analysis of urban wastewater treatment systems under different temperature conditions. Water 16 (21), 3039. doi:10.3390/w16213039

CrossRef Full Text | Google Scholar

Ministry of Ecology and Environment (MEE). (2022). Announcement on the release of carbon dioxide emission factors for electricity in 2022 (Announcement No. 33 of 2024).Available online at: https://lca.cityghg.com/pages/product-view/9087/ (Accessed on 3 July 2025).

Google Scholar

Mondello, G., Salomone, R., Ioppolo, G., Saija, G., Sparacia, S., and Lucchetti, M. (2017). Comparative LCA of alternative scenarios for waste treatment: the case of food waste production by the mass-retail sector. Sustainability 9 (5), 827. doi:10.3390/su9050827

CrossRef Full Text | Google Scholar

Naroznova, I., Møller, J., and Scheutz, C. (2016). Global warming potential of material fractions occurring in source-separated organic household waste treated by anaerobic digestion or incineration under different framework conditions. Waste Manag. 58, 397–407. doi:10.1016/j.wasman.2016.08.020

PubMed Abstract | CrossRef Full Text | Google Scholar

Pang, W., Hou, D., Nowar, E. E., Chen, H., Zhang, J., Zhang, G., et al. (2020). The influence on carbon, nitrogen recycling, and greenhouse gas emissions under different C/N ratios by black soldier fly. Environ. Sci. Pollut. Res. Int. 27 (34), 42767–42777. doi:10.1007/s11356-020-09909-4

PubMed Abstract | CrossRef Full Text | Google Scholar

Perez, T., Vergara, S. E., and Silver, W. L. (2023). Assessing the climate change mitigation potential from food waste composting. Sci. Rep. 13 (1), 7608. doi:10.1038/s41598-023-34174-z

PubMed Abstract | CrossRef Full Text | Google Scholar

Ramzy, R. R., Goenka, V., El-Dakar, M. A., and Lee, J. S. H. (2025). Assessing the environmental impacts of the Black soldier fly-based circular economy and decentralized System in Singapore: a case study. Sustainability 17 (13), 6115. doi:10.3390/su17136115

CrossRef Full Text | Google Scholar

Ren, L., Jin, Y., Liu, J., and Nie, Y. (2009). Factors influencing the liquefaction, separation, and recovery of solid grease from restaurant garbage. Qinghua Daxue Xuebao/Journal Tsinghua Univ. 49 (3), 386–389.

Google Scholar

Schmid, F., Taube, L., Rieck, J., and Behrendt, F. (2021). Electrification of waste collection vehicles: technoeconomic analysis based on an energy demand simulation using real-life operational data. IEEE Trans. Transp. Electrification 7 (2), 604–615. doi:10.1109/tte.2020.3031072

CrossRef Full Text | Google Scholar

Shen, W., Qiao, H., Tong, X., Xiao, Y., and Han, L. (2025). Assessing the sustainability performances of the supply chain of decentralized urban food waste composting for urban green spaces. Resour. Conservation and Recycl. Adv. 25, 200245. doi:10.1016/j.rcradv.2025.200245

CrossRef Full Text | Google Scholar

Singh, P. K., Mohanty, P., Mishra, S., and Adhya, T. K. (2022). Food waste valorisation for biogas-based bioenergy production in circular bioeconomy: opportunities, challenges, and future developments. Front. Energy Res. 10, 903775. doi:10.3389/fenrg.2022.903775

CrossRef Full Text | Google Scholar

Song, Y., Liu, D., Zhu, F., Zhang, C., Zhang, S., Li, T., et al. (2024). Urban food waste and socioeconomic drivers in China. Earth Crit. Zone 1 (1), 100015. doi:10.1016/j.ecz.2024.100015

CrossRef Full Text | Google Scholar

Sweetapple, C., Fu, G., and Butler, D. (2014). Multi-objective optimisation of wastewater treatment plant control to reduce greenhouse gas emissions. Water Res. 55, 52–62. doi:10.1016/j.watres.2014.02.018

PubMed Abstract | CrossRef Full Text | Google Scholar

Tang, J., Wangg, X. C., Hug, Y., Pug, Y., Huangg, J., Hao Ngog, H., et al. (2018). Nitrogen removal enhancement using lactic acid fermentation products from food waste as external carbon sources: performance and microbial communities. Bioresour. Technol. 256, 259–268. doi:10.1016/j.biortech.2018.02.033

PubMed Abstract | CrossRef Full Text | Google Scholar

Tian, H., Wang, X., Lim, E. Y., Lee, J. T. E., Ee, A. W. L., Zhang, J., et al. (2021). Life cycle assessment of food waste to energy and resources: centralized and decentralized anaerobic digestion with different downstream biogas utilization. Renew. Sustain. Energy Rev. 150, 111489. doi:10.1016/j.rser.2021.111489

CrossRef Full Text | Google Scholar

Wang, Y., Shi, Y., Zhou, J., Zhao, J., Maraseni, T., and Qian, G. (2021). Implementation effect of municipal solid waste mandatory sorting policy in Shanghai. J. Environ. Manage 298, 113512. doi:10.1016/j.jenvman.2021.113512

PubMed Abstract | CrossRef Full Text | Google Scholar

Wang, Z., Cui, L., Cui, J., Wang, H., Lou, Z., Cheng, Z., et al. (2025a). Characteristics of odor and health risk assessment in anaerobic digestion plants for food waste: a case study of Shanghai. Chin. J. Environ. Eng. 19 (6), 1351–1361. doi:10.12030/j.cjee.202407046

CrossRef Full Text | Google Scholar

Wang, Y., Ying, H., Stefanovski, D., Shurson, G. C., Chen, T., Wang, Z., et al. (2025b). Food waste used as a resource can reduce climate and resource burdens in agrifood systems. Nat. Food 6 (5), 478–490. doi:10.1038/s43016-025-01140-z

PubMed Abstract | CrossRef Full Text | Google Scholar

Weber, S. (2023). Lipid rich kitchen waste: analytic characterization, anaerobic digestion, and mass balances. Biomass Convers. Biorefinery 14 (17), 21317–21328. doi:10.1007/s13399-023-04466-5

CrossRef Full Text | Google Scholar

Wu, M., Hu, J., Shen, F., Huang, M., Zhao, L., Tian, D., et al. (2021). Conceptually integrating a multi-product strategy for the valorization of kitchen waste towards a more sustainable management. J. Clean. Prod. 306, 127292. doi:10.1016/j.jclepro.2021.127292

CrossRef Full Text | Google Scholar

Wu, R., Chen, M., Qin, Y., Liu, S., and Li, X. (2023). Combined hydrothermal and biological treatments for valorization of fruit and vegetable waste into liquid organic fertilizer. Environ. Res. 221, 115262. doi:10.1016/j.envres.2023.115262

PubMed Abstract | CrossRef Full Text | Google Scholar

Xiao, H., Zhang, D., Tang, Z., Li, K., Guo, H., Niu, X., et al. (2022). Comparative environmental and economic life cycle assessment of dry and wet anaerobic digestion for treating food waste and biogas digestate. J. Clean. Prod. 338, 130674. doi:10.1016/j.jclepro.2022.130674

CrossRef Full Text | Google Scholar

Xin, C., Zhang, T., Tsai, S.-B., Zhai, Y.-M., and Wang, J. (2020). An empirical Study on greenhouse gas emission calculations under different municipal solid waste management strategies. Appl. Sci. 10 (5), 1673. doi:10.3390/app10051673

CrossRef Full Text | Google Scholar

Xu, C., Shi, W., Hong, J., Zhang, F., and Chen, W. (2015). Life cycle assessment of food waste-based biogas generation. Renew. Sustain. Energy Rev. 49, 169–177. doi:10.1016/j.rser.2015.04.164

CrossRef Full Text | Google Scholar

Yan, Y., Gu, R., Zhu, M., Tang, M., He, Q., Tang, Y., et al. (2023). Environmental impacts and optimization simulation of aerobic anaerobic combination treatment technology for food waste with life cycle assessment. Waste Manag. 164, 228–237. doi:10.1016/j.wasman.2023.03.036

PubMed Abstract | CrossRef Full Text | Google Scholar

Yang, X., Liu, Y., Zhu, Z., Gu, C., Rui, Y., Wang, Q., et al. (2020). Life cycle assessment of biodiesel from soybean oil and waste oil. Nongye Gongcheng Xuebao/Transactions Chin. Soc. Agric. Eng. 36 (19), 233–241. doi:10.11975/j.issn.1002-6819.2020.19.027

CrossRef Full Text | Google Scholar

Yang, H., Wang, Z., Zhang, X., and Lu, H. (2025). Assessing greenhouse gas emission factors in wastewater treatment. Sustainability 17 (4), 1488. doi:10.3390/su17041488

CrossRef Full Text | Google Scholar

Yu, Q., and Li, H. (2021). Life cycle environmental performance of two restaurant food waste management strategies at Shenzhen, China. J. Material Cycles Waste Manag. 23 (2), 826–839. doi:10.1007/s10163-020-01157-5

CrossRef Full Text | Google Scholar

Yu, Q., Li, H., Deng, Z., Liao, X., Liu, S., and Liu, J. (2020). Comparative assessment on two full-scale food waste treatment plants with different anaerobic digestion processes. J. Clean. Prod. 263, 121625. doi:10.1016/j.jclepro.2020.121625

CrossRef Full Text | Google Scholar

Zeng, Q., Zhen, S., Liu, J., Ni, Z., Chen, J., Liu, Z., et al. (2022). Impact of solid digestate processing on carbon emission of an industrial-scale food waste co-digestion plant. Bioresour. Technol. 360, 127639. doi:10.1016/j.biortech.2022.127639

PubMed Abstract | CrossRef Full Text | Google Scholar

Zexing, L., Yuejin, L., Lingfei, Z., Weiwen, L., and Yulin, T. (2025). Investigating carbon emissions in drinking water treatment plant through machine learning: a case study in South China. J. Water Process Eng. 75, 108070. doi:10.1016/j.jwpe.2025.108070

CrossRef Full Text | Google Scholar

Zhang, L., Wang, B., Wang, Z., Li, K., Fang, R., Su, Y., et al. (2022). Spatiotemporal footprints of odor compounds in megacity's food waste streams and policy implication. J. Hazard. Mater. 437, 129423. doi:10.1016/j.jhazmat.2022.129423

PubMed Abstract | CrossRef Full Text | Google Scholar

Zhang, L., Li, H., Hu, T., Du, X., Zhou, Y., Sun, G., et al. (2024). The reduction of the carbon footprint of municipal solid waste management via source classification and supporting strategies: an analysis for the megacity of Shenzhen. Waste Manag. 187, 145–155. doi:10.1016/j.wasman.2024.07.012

PubMed Abstract | CrossRef Full Text | Google Scholar

Zhang, L., Li, H., Yang, L., Du, X., Zhou, Y., Sun, G., et al. (2025a). Carbon footprints of centralized and decentralized food waste utilization pathways. Renew. Sustain. Energy Rev. 208, 115040. doi:10.1016/j.rser.2024.115040

CrossRef Full Text | Google Scholar

Zhang, Z., Huang, K., Zhang, J., Xiao, J., Peng, H., Ge, X., et al. (2025b). Study on the carbon emission reduction potential of rural domestic waste based on life cycle approach in Hubei Province, China. Environ. Sci. Pollut. Res. Int. 32 (25), 15243–15257. doi:10.1007/s11356-025-36612-z

PubMed Abstract | CrossRef Full Text | Google Scholar

Zhao, Y., Chang, H., Liu, X., Bisinella, V., and Christensen, T. H. (2022). Climate change impact of the development in household waste management in China. Environ. Sci. and Technol. 56 (12), 8993–9002. doi:10.1021/acs.est.1c07921

PubMed Abstract | CrossRef Full Text | Google Scholar