Can the implementation of household waste classification mitigate greenhouse gas emissions in Beijing? A comprehensive analysis of recent trends and future scenarios

Household waste contributes significantly to global greenhouse gas (GHG) emissions, and waste classification is crucial for reducing emissions. This study focuses on Beijing and utilizes life cycle assessment (LCA) and material flow analysis (MFA) to calculate GHG emissions in waste management systems and quantify emission reduction potential of different measures. The results show that net emissions from the classification system in 2021 are 116.77 kg CO2-eq/t waste, reducing 61.82 % compared to the traditional mixed collection and transportation system. Waste volume, classification efficiency, and treatment strategies are the primary factors affecting emissions in classification systems. Recycling is identified as effective treatment methods. Three scenarios are designed to explore emission pathway of the system toward 2060. In the business-as-usual (BAU) Scenario, emissions will continue to grow to 108.57 × 104 t CO2-eq/yr in 2060. In the Classification Efficiency Scenario and the Comprehensive Scenario, emissions in 2060 will be cut to −177.26 × 104 t CO2-eq/yr and −702.00 × 104 t CO2-eq/yr, respectively. These results underscore the critical role of waste classification and recycling in mitigating the negative impacts of increasing waste volume. By 2060, combining waste classification with recycling can offset emissions by 803.51 × 104 t CO2-eq/yr, contributing 99 % to emission reduction potential. Improving classification efficiency and recycling ratio are key measures for achieving this reduction goal. Meanwhile, treatment methods and technologies should prioritize classification and recycling. Aiming at carbon neutrality, the study proposes several recommendations to improve classification systems, including enhancing classification efficiency, optimizing treatment facilities and strategies, and establishing recycling and utilization systems, etc.


Introduction
GHG from waste has been recognized as one of the major sources of GHG emissions [1,2], accounting for 5 % of the total at the global level [2,3].In China, emissions reached about 203.54 million tons in 2019, accounting for 1.68 % of total emissions [4], and system in Beijing, China.Section 5 is policy implications.Finally, Section 6 draws the research conclusions.

Household waste classification in Beijing
In this study, Beijing, with 16 municipal districts is selected as a case.The household waste in this study is generated from the residential, office, and public areas, but excludes the construction waste and the medical waste.The focus is mainly on household waste generated by units and individuals in their daily lives.The Regulation has come into force in May 2020.According to the Regulation, household waste has been classified into four categories, namely food waste, recyclables, hazardous waste, and residual waste.Each category should be collected, transported, and treated separately.
The household waste flow in the waste system is listed as follows (shown in Fig. 1).In the waste classification system (shown in Fig. 1(a)), the household waste is collected from communities and other units first, and then transported, and sent to the treatment facilities.The separated food waste is sent to the anaerobic digestion plant for treatment.Part of the impurities (10 %) [29] are Fig. 1.Source and sink of household waste in the waste systems (t/d).
Z. Wen et al. separated in the pretreatment process and then sent to the landfill plant for burial, while the remaining part (90 %) is treated by anaerobic digestion.The residual waste is sent directly for incineration, and after that, the fly ash and slag (10 % each) [5] are sent to the landfill plant for burial (not included in Fig. 1).At present, the recycling rate of recyclables in Beijing is about 35 % [30], which means that 65 % sorted recyclables need to be sent to the incineration plants for treatment.To calculate simply, the waste is divided according to its component ratio.
The traditional mixed collection and transportation system (shown in Fig. 1(b)) is established for comparison.Assume that in this process, only 10 % of food waste is separated and no recyclables are recovered.Other technical conditions, such as the waste components, the efficiency of power generation, biogas utilization, etc., are the same as the waste classification system.
In addition, hazardous waste has not been considered in this study as it is generally not disposed of by traditional treatment facilities.Considering the quantity of household waste was reduced after the implementation of the Regulation, the study uses the data from May 2020 to July 2021 as an example (shown in Fig. 2).Meanwhile, according to the design capacity of the transfer station (see Table 1) and the household waste amount in 2021 (see Fig. 1), it is assumed that the indirect and direct transportation ratio of household waste is 1:1.
2. In this study, the term "food waste" encompasses both "Food waste 1" and "Food waste 2"."Residual waste" refers to the portion that excludes recyclables."Household waste" refers to the sum of "food waste", "recyclables" and "residual waste" (excluding recyclables).3. The waste volume in this study is based on the median values of "food waste", "recyclables" and "residual waste".The total amount of household waste is the sum of these three types, rather than the median and mean values directly from the table.4. The data cover a period of 15 months, from May 2020 to July 2021, with some months missing.The "Count" line indicates the number of months for which data was used.
Fig. 2. Household waste volume in Beijing in 2020-2021.Note: 1. Waste volume is classified according to the Beijing Municipal Commission of Urban Management."Food waste 1" refers to waste generated by units, families, communities, etc., while "Food waste 2" refers to restaurant kitchen waste, etc.
Data source: Beijing Municipal Commission of Urban Management.

Analytical framework
At present, LCA [2,5,20,31], Hybrid Life Cycle Assessment (HLCA) [32], Input-Output model [33], Pinch Analysis [6,16,34] are the main methods used to study the environmental benefits such as GHG emissions of household waste system.In recent years, LCA, as a common method to calculate GHG emissions, has attracted attention in the waste industry.Meanwhile, studies on GHG emissions from the waste industry have often combined material flow analysis (MFA) with LCA [21,35].In addition, some scholars have also combined the embodied carbon/energy method with LCA to calculate the energy-saving and emission reduction potential of material recycling or waste management systems [36,37].This study will integrate LCA, MFA, IPCC (The Intergovernmental Panel on Climate Change) guidelines, and the embodied carbon/energy method to examine the GHG emissions and mitigation potentials across various aspects of the waste classification system.The entire process, from community collection to waste treatment, will be evaluated using LCA.The MFA method will be utilized to analyze the distribution of waste quantities.GHG emissions resulting from treatment, transportation processes, and energy consumption from equipment will be calculated following the calculation methods outlined in the IPCC guidelines.In addition, the reduction of GHG emissions through the recycling of recyclable materials will be assessed using the embodied carbon/energy method.It is worth noting that this life cycle assessment is not comprehensive and does not encompass the impact of other pollutants; instead, it primarily focuses on carbon footprint research through the utilization of LCA.
The system is shown in Fig. 3, including the collection and transportation, treatment, and recyclables utilization.Household waste and energy (fuels and power) are regarded as the input items of the system, and GHG emissions and the offset are treated as the output items.The life cycle of pre-collection systems and vehicles is excluded, which is a topic worthy of further discussion.Given the data availability, obtaining specific information about transportation vehicles can be quite challenging, and that's why they are often overlooked in similar studies [38].
The objectives of the classification system include the following aspects: (1) quantify the GHG emissions and offsets for the entire life cycle of the waste classification system; (2) compare with the traditional mixed collection and transportation system; (3) predict the emission reduction potential of future waste classification systems.The functional unit is defined as the treatment of 1 t of household waste in Beijing.

Calculation method
The methodology and data for quantifying GHG emissions related to waste classification system are thoroughly described in this part.In this study, the GHG emissions (CO 2 , N 2 O, CH 4 ) are all from the following three aspects (see Table 2): (1) GHG emissions from three treatment methods: incineration, landfill, and anaerobic digestion; (2) Fuel consumption including the landfill process and the vehicle transportation, etc.; (3) Power consumption including the power consumed from transfer stations and compression stations, power consumed by the incineration and anaerobic digestion plant, power consumed by the compression fluid and the leachate treatment.The GHG emissions offsets include recovery of energy and material recycling.(1) Incineration GHG from the incineration include CO 2 , CH 4, and N 2 O.In general, incineration produces more CO 2 than CH 4 and N 2 O [40].In this study, CO 2 emissions are calculated based on the components of the incinerated waste (such as paper/cardboard, and plastic, etc.), as outlined in Table 3.The calculation formulas presented in equations (1) and ( 2).   3) and ( 4).
(3) (2) Landfill GHG emissions from landfill mainly originate from CH 4 , which is produced by the degradation of organic matter under anaerobic conditions during the landfilling process.Decomposable portion of household waste encompasses food waste, wood/bamboo, paper/ cardboard, etc.The data required for the calculations can be found in Table 4.The calculation formulas are presented in equations ( 5)-( 9).
E lan.− CH4 = CO 2 -eq emissions (CH 4 ) from landfill in the inventory year (t CO 2 -eq/d; t CO 2 -eq/yr); CH 4 generated i = CH 4 generated from various decomposable matter (t CH 4 ); OX = oxidation factor, 0.1 [39]; (3) Anaerobic digestion Since the N 2 O emission factor for anaerobic digestion is assumed to be negligible [39], CH 4 is considered the primary GHG emission in this section.The quantity of CH 4 recovered is accounted for in the biogas power generation of the emissions offsets in Section 3.2.2.The CH 4 emissions from the anaerobic digestion can be estimated by equation (10).The primary GHG present in compressed fluid and leachate are CH 4 and N 2 O.The compressed liquid is produced during compression.Leachate is produced during the stacking process, such as at the transfer stations, incineration plants, landfill sites, anaerobic digestion plants, and recycling facilities.The pertinent formulas for calculating GHG emissions (equations 11-13) are listed.
E wat.− CH4 = CO 2 -eq emissions (CH 4 ) from wastewater in the inventory year (t CO 2 -eq/d; t CO 2 -eq/yr); V i = the volumes of compressed fluid and leachate (m 3 /d).The calculation parameters for the volumes of compressed fluid and leachate are provided in Table 5.

Fuel consumption.
The fuel consumption mainly comes from three aspects: (1) fuel consumption of vehicles during the waste collection and transportation; (2) transportation of incineration residues and fly ash, impurities from anaerobic digestion, biogas residues from anaerobic digestion, and recyclables which are not recycled; (3) the use of landfill equipment.The fuel is mainly diesel.The residues, fly ash, impurities, and biogas residues accounted for 10 %, 10 %, 10 %, and 20 % of the wet waste volume, respectively [5,29].The fuel consumption of vehicles during the collection and transportation is calculated in equation (14).

Table 5
The amount of compressed fluid and leachate from household waste.The average distance from district centers to different treatment facilities is calculated based on the average distance between the 16 district centers and various treatment facilities.These data can be obtained from the Amap.The average distances from the center of each district to the incineration, landfill, anaerobic digestion, and recycling station, as well as the average distances from the center of each district to the transfer station, and the transfer station to the four treatment facilities are 60, 71, 59, 63, 49 and 14 km, respectively.The vehicles exhibit a fuel consumption rate of 8 L/100 km when they are not carrying a load, and 11 L/100 km when they are fully loaded [29].
The fuel consumption during the landfill process is 6.72 MJ/ton [5].The calorific value of diesel is 0.038 GJ/L [42].The basic data for the total fuel consumption during vehicle transportation are shown in Table 6 The calculation of GHG emissions can be expressed by equation (15).
E fuel,i = CO 2 -eq emissions from different aspects of fuel consumption (t CO 2 -eq/d; t CO 2 -eq/yr); C i = number of different aspects of fuel consumption (t/d; t/yr); EF fuel = the carbon emission factor of diesel.The carbon emission factor of diesel is 2.63 kg/L [29]; i = different aspects of fuel consumption.

Power consumption.
The power consumption mainly comes from four aspects: (1) the equipment of the incineration plant.
After the power generation, 30 % of the electricity will be returned to the plants [5]; (2) the equipment of the crushing and pulping process in the anaerobic digestion, which is 32 kWh/t waste [43]; (3) the equipment of compression in transport station, which is 3.3 kWh/t waste [5]; (4) the treatment of compressed liquid and leachate, which is 0.13 kWh/m 3 wastewater [5,44].The calculation of GHG emissions can be expressed by equation ( 16).
E elec.− i = CO 2 -eq from different aspects of power consumption (t CO 2 -eq/d; t CO 2 -eq/yr); P t,i = the number of different aspects of power consumption (kWh/d); i = different aspects of power consumption; EF elec.= electricity grid emission factor, 0.5810 kg CO 2 /kWh [45].

Recovery of energy.
(1) Incineration Energy recovery involves household waste incineration for power generation.In this study, the heat energy generated during the incineration is first calculated based on the waste components.Subsequently, power generation is obtained using the thermoelectric conversion coefficient and power generation recovery efficiency.GHG emissions offsets need to be calculated based on electricity grid emission factor.The lower heating value takes into account the effects of humidity and hydrogen content, thus being derived by Note: The fully loaded weight of the vehicles is 3490 kg [29].
Z. Wen et al.
subtracting these influences from the higher heating value.The calculation formula for incineration power generation is presented in equations is presented in equation ( 17).
E re.− power = GHG emissions offsets from incineration power generation (t CO 2 -eq/d; t CO 2 -eq/yr); W i,1 = weight of different components of household waste incinerated (t/d; t/yr); EF efficiency = power generation recovery efficiency (%), 22.05 % [27,46]; HHV i = higher heating value of material i, dry matter (kJ/kg); H i = content of element H in material i, dry matter (%); W w,i = amount of moisture in waste i incinerated, dry matter (%).
The values and sources of HHV i , H i and W w,i can be found in Table 7.
(2) Anaerobic digestion The end products of anaerobic digestion include biogas, biogas residues, and leachate.Biogas is mainly composed of CH 4 (main component), CO, CO 2 , and H 2 S. Considering potential leakage ranging from 0 % to 10 % during anaerobic digestion, it is assumed that 95 % of the generated methane can ultimately be utilized for recovery and power generation [39].Assuming that the power generation efficiency is 50 % [47,48], the GHG offsets achieved by the biogas power generation can be calculated using equation (18) [5,48].
E re.− biogas = using biogas to generate electricity achieves offsets in GHG emissions compared to releasing methane (CH 4 ) into the atmosphere (t CO 2 -eq/d; t CO 2 -eq/yr); It means the mass of available CH 4 from anaerobic digestion for recovery (t CH 4 ); 2.75 × M AD− CH4 = CO 2 -eq emissions from methane combustion.The CO 2 -eq emissions per unit of methane combustion are approximately 2.75.

Material recycling.
(1) Recyclables Recycling of recyclables has the potential to replace certain production processes, thereby reducing energy consumption.The proportions of various recyclable materials are outlined in Table 3.In this case, the embodied energy/carbon method is utilized to calculate the GHG emissions offsets [37,49], as demonstrated in equation (19).Table 8 presents the embodied carbon coefficients for both raw materials and recyclables.

Table 7
The average higher heating value (HHV), hydrogen (H) element content, and moisture content of various components in household waste.One of the significant methods for recycling anaerobic biogas residues involves its transformation into organic fertilizer through composting, enabling the reduction of carbon emissions.Taking into account the carbon transfer and storage during the composting process, the net carbon flux in this study is − 0.055 t CE/t waste [5,50].The calculation method is depicted in equation (20).

Scenario design
To analyze the emission reduction potential under different measures, three scenarios including the BAU Scenario, the Classification Efficiency Scenario, and the Comprehensive Scenario are set up to explore the emission trajectories of the waste classification system until 2060.Firstly, in the BAU Scenario, the parameters such as annual classification efficiency are the same as the base year 2021.In this scenario, changes in future waste volumes are projected to estimate the associated GHG emissions and emissions offsets under future waste classification.Secondly, in the Classification Efficiency Scenario, the classification rate will continue to increase, and it will affect the proportion of waste in different treatment methods and the proportion of each component in residual waste.Thirdly, the Comprehensive Scenario further incorporates energy technologies.

BAU scenario
In the BAU Scenario, the waste growth rate is the key variable, and the classification efficiency and energy technologies improvements are not considered.In addition to the increase in waste volume, the classification efficiency and the energy efficiency of each inventory year are the same as the base year 2021, and other coefficients such as emission factors are consistent.Waste quantity prediction referenced approaches from similar research [51,52] and established a multiple linear regression model with gross domestic product (GDP), population, and electricity consumption as predictive variables to forecast household waste generation in Beijing up to 2060.The model details are presented in Table 9 and Table 10.

Classification Efficiency Scenario
In addition to the waste growth rate, classification efficiency is another key variable in the waste classification system, and it is also an important variable that mainly affects waste volume for terminal treatment.It affects the waste volume of transportation, the amount of terminal treatment, and the proportion of waste components.
Based on the waste components in 2021, it is found that the maximum classification rates of food waste and recyclables are about 54 % and 37 %, respectively.It is expected that the waste classification rate will increase apace in the short term, and waste classification work will be easy to advance.In the long term, the growth rate of the waste classification will slow down.The classification efficiency is shown in Table 11.
The proportion of waste components corresponding to different classification efficiencies can be seen in Table 12 and Table 13.The proportion of the treatment methods corresponding to different classification efficiencies can be seen in Table 14.

Comprehensive Scenario
According to the different characteristics of the technologies and their different positions in the waste classification system, the progress of energy technologies under the Comprehensive Scenario can be divided into three different categories: the improvements of existing energy utilization technologies, the substitution of energy technologies, and the application of untapped energy technologies.
The first category is the improvements of existing energy utilization technologies, including improving the recycling rate of recyclables; improving the recovery efficiency of incineration power generation, and improving the conversion efficiency of biogas power generation.The parameters of the energy technologies are shown in Table 15.
From the perspective of energy technologies substitution, the fuel of transshipment vehicles will switch from diesel to electricity.With the reduction of diesel vehicles and the increase of electric vehicles, the penetration rate of electricity technology will gradually increase, affecting the emission factors of transportation.The transportation emission factors influenced by technological penetration are calculated as shown in equations ( 21) and ( 22) (referring to Ref. [57].The parameters are presented in Table 15. EF i,1 = emission factor of transportation (transport process before treatment) (kg/L); EF i0 = carbon emission factor of diesel (kg/L), 2.63 kg/L [29]; T = technology penetration rate; ( 2 × 15.5

19
) = conversion coefficient.This coefficient is calculated based on data obtained from China's automotive energy consumption.
EF i,2 = emission factor of transportation (transport process after treatment) (kg/L); ) = conversion coefficient.This coefficient is calculated based on data obtained from China's automotive energy consumption.The energy technologies in the future include recovering landfill gas to generate electricity and converting incinerated slag into building materials.The parameters are shown in Table 15.The formula and parameters for calculating landfill gas recovery electricity generation can be found in equation ( 23) (referring to Ref. [58]).The method to calculate the emissions offsets of the slag is given in Section 3.2.2.

GHG emission reduction benefits from waste classification
Based on the statistical data from 11 months after the new round of waste classification in Beijing, we have calculated the GHG emissions of the classification system and compared its results with those of the traditional mixed collection and transportation system to determine whether waste classification has a positive impact on GHG emissions.The calculation includes three parts: the collection and transportation, the treatment processes, and the resource recovery.In addition, we have factored in the potential emission reduction through energy recovery and material recycling.

Avoid GHG burdens in the classification system
Fig. 4 shows the comparison of system-wide GHG emissions between the two systems in 2021.It can be seen that the net emissions from the waste classification system are 116.77kg CO 2 -eq/t waste, which is 61.82 % less than that of the traditional mixed collection and transportation system by 189.11 kg CO 2 -eq/t waste.This is due to a 120.44 kg CO 2 -eq/t waste reduction in GHG emissions from classification system and a 68.67 kg CO 2 -eq/t waste increase in GHG offsets.This shows that improvements in system-wide GHG emissions of waste classification systems not only reduce GHG emissions but also increase the number of GHG offsets.

GHG in each link of classification system
According to the data presented in Fig. 5, the emission reduction potential of the two systems differs in terms of GHG emissions and GHG emissions offsets.
For Links 1 to 7, the GHG emissions of the two systems are different, and therefore the emission potential of the two systems in these links is also different.Specifically, the GHG emissions of the traditional mixed collection and transportation system in these links are higher than those of the waste classification system, so its emission potential is also higher.Link 1 is the main contributor to the GHG Note: Due to the improved efficiency of waste classification, there have been changes in the composition of residue waste.This portion is calculated based on the information in Table 10 and the projected amount of waste.Note: Due to the enhanced efficiency of waste classification, the proportions of waste treated by different methods have undergone changes.This portion is calculated using the data from Table 10 and the projected amount of waste.
Z. Wen et al.
emissions at 361-473 kg CO 2 -eq/t waste.In terms of emission reduction potential, Link 1 also has a higher emission reduction potential.The waste classification system reduces about 112.49kg CO 2 -eq/t waste in this link, which is much higher than other links.Emissions in Link 1 are related to waste treatment methods, so the emission reduction potential of the waste classification system from GHG emissions mainly depends on the waste flow and treatment strategies.
For Links 8 to 9, there are also some differences in the emission reduction potential of the two systems in terms of GHG offsets.Specifically, the waste classification system has a higher potential for emission reduction in these links, as the total GHG offsets in these links are higher.In these links, the waste classification system can obtain a higher amount of GHG offsets through material recycling.The amount of GHG offsets by material recycling is about 30 times that of the traditional mixed collection and transportation system, increasing 108.06 kg CO 2 -eq/t waste.However, the energy recovery from incineration power generation may be limited.
In different links, the emission potential and emission reduction potential of the two systems are different.When formulating emission reduction strategies, it is necessary to take different emission reduction measures for different links and treatment methods to maximize the emission reduction potential of the waste classification system.

Avoid GHG burdens in each treatment method of classification system
To compare the differences between GHG emissions and offsets of the four treatments and determine the emission reduction potential of the four treatments, the system-wide GHG emissions of each link in Fig. 5 are divided into four items, namely, incineration, landfill, anaerobic digestion, and recycling (see Fig. 6).
Fig. 6 shows the GHG emissions, GHG offsets, and net emissions of the four treatment methods in each link of the two systems.
In terms of GHG emissions, both incineration, and landfill have significant GHG emissions, ranging from 160 to 343 kg CO 2 -eq/t waste.The emissions from anaerobic digestion and recycling are relatively low, mostly below 20 kg CO 2 -eq/t waste.
In terms of GHG offsets, both incineration and recycling have significant offsets of − 221.52 kg CO 2 -eq/t waste and − 102.10 kg CO 2eq/t waste respectively, making a significant contribution to the net emissions of classification system.This is followed by anaerobic digestion, which is − 9.84 kg CO 2 -eq/t waste.Whereas the landfill has no GHG offsets.
In terms of net emissions, the net emissions from incineration have decreased, but both the emissions and the offsets are relatively high.The net emissions from the landfill have also decreased.It has slightly lower emissions than incineration but has no GHG offsets.The net emissions from anaerobic digestion are 2.2-5.8kg CO 2 -eq/t waste.Recycling has the lowest net emissions.It is essentially zero emissions and has higher offsets.This underscores the importance of energy recovery and material recycling to reduce net emissions.
In addition, comparing the four treatments, we speculate that recycling may have the highest emission reduction potential, followed by incineration.The anaerobic digestion and landfill have the lowest emission reduction potential.
Comparing the two systems, the net emissions of the waste classification system are lower than those of the traditional mixed

The conversion efficiency of biogas power generation:
A comparative review of biogas generation technologies has indicated that thermal efficiency can be enhanced up to 90 % through cogeneration [54].Therefore, the calculations in this study assume 70 % as the technical limit for energy recovery from biogas generation.
4. The emission factor of transportation (before treatment and after treatment): The current CO2 emission factor for diesel is 2.63 kg/L.In theory, electric vehicles can fully replace diesel vehicles for waste transportation.Therefore, this study set the proportions of electric vehicles replacement for diesel vehicles at 0 %, 15 %, 30 %, 45 %, and 60 % for the years 2021, 2030, 2040, 2050, and 2060 respectively.The CO2 emission factor of waste transportation was then calculated using the technological penetration formula presented in Section 3.3.3.

Power generation efficiency of the ICE used for LFG and the efficiency of gas collection from the landfills: Currently, the internal combustion engine (ICE)
used for landfill gas (LFG) utilization has a power generation efficiency of about 40 %.Therefore, this study sets a long-term development goal for ICE's power generation efficiency with LFG at 55 % by 2060 [55].According to data from a survey of gas collection systems at 23 landfills in Denmark, the gas collection efficiency of these landfills reached up to 86 %, with an average of 50 % [56].Therefore, this study establishes a long-term development goal of 85 % for landfill gas collection efficiency by 2060.
6.The recovery rate of incinerated slag: After reviewing the literature, no specific data is available to determine the percentage of incineration slag that is converted into building materials.Due to this lack of data, the parameters of this study were established by referencing the recyclable percentage of sustainable building materials from Sahlol et al., 2021.
Z. Wen et al. collection and transportation system in three treatment methods.The net emissions of anaerobic digestion in the waste classification system are slightly higher than the traditional collection system, but the net emission intensity of anaerobic digestion is lower than that of landfill and incineration.Generally speaking, the two treatment methods of anaerobic digestion and recycling are the best in the waste classification system, followed by incineration, and landfill should gradually shift to the other three treatment methods.
In general, the existing waste classification system in Beijing provides 62 % more net emissions benefits than the traditional mixed collection and transportation system.Waste classification reduces GHG emissions (120.44 kg CO 2 -eq/t waste) and increases GHG offsets (− 68.67 kg CO 2 -eq/t waste).Waste treated by incineration and landfill decreases by 3637.89t waste/d and 1041.75 t waste/d, corresponding to an emission reduction of 711.08 t CO 2 -eq/d and 1350.89t CO 2 -eq/d.Waste treated by anaerobic digestion and recycling increases by 3208.59 t waste/d and 1471.05t waste/d, corresponding to an emission increase of 78.46 t CO 2 /d and an emission reduction of 2127.53 t CO 2 /d.

GHG emissions reduction potential from household waste classification 4.2.1. Total GHG emissions of different scenarios
Three scenarios, namely the BAU Scenario, the Classification Efficiency Scenario, and the Comprehensive Scenario, are developed in this study to explore the emission trajectories of the waste classification system to 2060.Fig. 7 shows the net emissions from 2021 to 2060 under three scenarios of China's waste classification system.
In the BAU Scenario, according to the multivariate regression model, it is anticipated that waste generation will gradually and slowly increase from 2021 to 2060, reaching 929.46 × 10 4 t/yr in 2060.During this period, the average growth rate of net emissions is approximately 4 %.By 2060, the net GHG emissions are projected to reach 108.57× 10 4 t CO 2 -eq/yr, which is 1.17 times the 2021 emissions.
In the Classification Efficiency Scenario, net-zero emissions will be achieved before 2035.Compared with the BAU Scenario, the Classification Efficiency Scenario has a significant and sustained reduction in net emissions from 2030 to 2060 and a significant increase in emission reduction.By 2060, the emissions will be reduced to − 177.26 × 10 4 t CO 2 -eq/yr, representing a reduction of 285.83 × 10 4 t CO 2 -eq/yr relative to the BAU Scenario.
In the Comprehensive Scenario, net-zero emissions can be achieved earlier than in the Classification Efficiency Scenario and a larger range of net emission reduction can be achieved through further emission reduction measures.By 2060, the emissions will be further reduced to − 702.00 × 10 4 t CO 2 -eq/yr, representing a net emissions reduction of 524.74 × 10 4 t CO 2 -eq/yr compared to the Classification Efficiency Scenario.
Overall, there is a large difference in net emissions between the Classification Efficiency Scenario and the Comprehensive Scenario, and their emission reduction effects are significant.Over time, emission reduction measures such as improving waste classification efficiency and increasing recycling rate have produced different degrees of emission reduction benefits, so the overall net emissions of the waste classification system are showing a continuous decreasing trend.Since the emission reduction measures adopted in each scenario are different, it is necessary to analyze the emission reduction potential of the emission reduction measures in each scenario to determine the optimal emission reduction strategy.

Reduction potential of different measures
In the scenario setting, a total of seven different measures are set up to identify the emission reduction potential of different measures in the waste classification system.The result is shown in Fig. 8.
During the whole study period, the improvements in waste classification efficiency and the increase in recycling rate are the most important measures to promote emission reduction in the waste classification system.Its emission reduction increases from − 167.32 × 10 4 CO 2 -eq/yr in 2030 to − 803.51 × 10 4 CO 2 -eq/yr in 2060, accounting for 82 %-99 % of the overall emission reduction.
Improving the efficiency of incineration power generation and the collection and utilization of landfill gas are the secondary contribution measures to the emission reduction potential of the waste classification system.However, with the improvements in the waste classification efficiency and the recycling rate, its emission reduction advantages are no longer obvious.It accounts for less than 0.3 % of the overall emission reduction by 2060.
In the future, with the advancement of classification, the technical development of incineration and landfill will have a limited role in promoting the emission reduction of the waste classification system.The development of waste treatment methods and technologies should be inclined towards classification and recycling.

Policy implications
Reducing GHG through waste classification is a systematic project that needs to be carried out from three aspects, which are to improve efficiency, implement treatment measures adapted to waste classification, and promote resource recovery.Therefore, with the goal of carbon neutrality, several recommendations are put forward to improve the household waste classification system.

Ensuring the improvement of waste classification efficiency
Waste classification facilitates emission reduction in waste management systems, and enhancing the efficiency of waste classification can further amplify the benefits of the reduction.To attain carbon neutrality, it is imperative to reinforce the efficacy of waste classification.Beijing initiated waste classification pilot demonstrations in 2000 but faces challenges such as inadequate resident participation, imperfect facilities, limited recycling capacity, and policy issues [59,60], resulting in low classification accuracy with a rate of only 16.92 %-34.56 % [61].This leaves a significant gap for improvement.
Policy support is a crucial for waste classification, as government policies exert a direct or indirect influence on waste management [62].The primary challenge confronting our country's waste classification efforts is how to leverage policy support to enhance the efficiency of waste classification.Mandatory policies can expedite the attainment of waste classification efficiency standards.Shanghai's compulsory household waste classification policy has resulted in the effective classification of over 80 % of household waste [63].As one of the four municipalities directly under the central government, Beijing shares commonalities and similarities with Shanghai.The successful experience of waste classification in Shanghai can be integrated with "the Regulations" to effectively promote waste classification.

Optimizing facilities and strategies for treatment
Improving waste classification efficiency reduces landfill and incineration waste while increasing waste for anaerobic digestion and recycling (see Fig. 9).This promotes waste reduction, resource utilization, and recycling, impacting the emission reduction potential from treatment.Incineration power generation is widely employed due to its advantages of harmlessness, reduction, and resource utilization.Additionally, the classification of food waste can decrease waste moisture content, facilitating energy recovery in incineration [64,65].However, as waste classification efficiency improves, the average lower calorific value (LHV) decreases, and the economic benefits of incineration diminish, limiting its applicability.Research indicates that when Beijing's waste classification rate increases from 45 % to 91 % (the separation rate of food waste and recyclables), the moisture content of incinerated waste drops from 44.42 % to 2.98 %, while its average LHV drops from 13965.12 kJ/kg to 875.85 kJ/kg (see Table 16).The average LHV of waste approaches or even falls below the minimum standard for incineration [66], requiring additional energy input.Higher waste classification efficiency compromised the operation stability of incineration, leading to reduced power generation benefits.Excessively high waste classification efficiency may not necessarily contribute to net emissions reduction [5,67].Therefore, future efforts should focus on enhancing the technical capabilities of low calorific value waste incineration [68], while gradually shifting towards anaerobic Fig. 6.Each treatment method's GHG in two systems.

Z. Wen et al.
digestion and recycling in waste management system.Additionally, greater attention should be given to anaerobic digestion and recycling, and to utilize waste more effectively rather than raw materials for new products to achieve a greater emission reduction potential and carbon neutrality goals.
2. Landfill includes impurities derived from anaerobic digestion, which accounts for 10 % from 2021 to 2060 in all three scenarios.This part will be presented separately in the sixth column of the figure.

Establishing a waste recycling and utilization system
As waste classification efficiency improves, there are corresponding changes in the waste composition (see Fig. 10).The most significant change is observed in food waste, which constitutes a major portion of household waste in China.Additionally, four categories of recyclables, including paper/cardboard, plastics, textiles, and wood/bamboo, will also increase.Improving the recycling rate effectively increases GHG offsets for material recycling.Studies have shown that recovery rates of different components have impacted GHG emissions differently [2,20,69].Currently, the contribution of GHG offsets can be listed in the following order plastic > wood/bamboo > paper/cardboard > food waste > textile > rubber and leather.Textiles and rubber have greater emission reduction potential through material recycling than incineration and landfill (see Table 17).
With the increasing proportion of recyclable resources in household waste, it has considerable utilization value for resource utilization [60].This study confirms the GHG reduction benefits of promoting recycling and resource utilization through waste classification.However, the implementation of waste classification may result in the loss of economic benefits for incineration enterprises, and recycling lacks cost competitiveness compared with raw materials usage [70].Therefore, the reuse of recyclables based on waste classification is strongly opposed by stakeholders [59].Chen (2016) proposed that "polluter pays" has promoted Taiwan's recycling industry development.Beijing can learn from Taiwan's experience and develop a suitable plan for the city to ensure sustainable waste classification.Additionally, encouraging and mandatory policies have different impacts on the effectiveness of waste classification [71].Encouraging policies such as cognition training, normative constraints, publicity and education, and improved facilities and service can directly influence citizens' behavioral decisions and willingness to participate in waste classification [72].Economic measures can boost waste classification efficiency.For recyclables with high economic value and strong emission reduction potential, measures such as social funds and market-oriented management can promote their utilization.For those with low economic value but who have a strong contribution to emission reduction, non-market methods such as financial subsidies and carbon tax support can be provided to ensure its reuse.

Conclusions
This study compares the GHG emissions between the traditional mixed collection and transportation system and the waste classification system, and explores the net emissions of the waste classification system in Beijing from 2021 to 2060, considering reduction potential from the improved classification system and advanced technology.The results indicate that waste classification can lower the net emissions of the entire waste system, and its GHG reduction potential depends on the waste flow, recycling, and treatment strategies.The order of net emission intensity of the four methods is landfill > incineration > anaerobic digestion > recycling.The future development of waste treatment should be inclined to recycling.
To further study the emission reduction potential of Beijing's waste classification system, three scenarios are set up and the GHG offset of seven measures are compared.The Classification Efficiency Scenario shows a significant and sustained reduction in net emissions compared to the BAU Scenario.Among the seven measures, improving waste classification efficiency and recycling rate are the most effective measures to reduce net emissions.
Waste classification has significant benefits for reducing GHG emissions, and source classification also enables the emission reduction potential from recycling.It should be noted that increasing recycling only reduces the embodied carbon emissions of raw materials rather than reducing net emissions from their production.Changes in waste flow, composition, moisture content, and calorific value, coupled with improved waste classification efficiency, will affect the existing waste management strategies.GHG emissions from incineration will inevitably increase, and waste management strategies that prioritize recycling are crucial for realizing potential climate benefits.Furthermore, the emission reduction potential of various components in the recycling process varies.Attention should be paid to the recyclables with high economic value and strong emission reduction potential, and the recyclables with low economic value but a great contribution to emission reduction should be guaranteed through government support.Policymakers should take into account the emission reduction potential from the classification system.Through the joint efforts of the government, enterprises, and the public, waste classification can be further consolidated to achieve emission reduction targets.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to Fig. 9. Waste volume sent to treatment facilities under three scenarios from 2021 to 2060 ( × 10 4 t/yr).Note: 1. Incineration includes the non-recyclables portion.In the BAU scenario, 65 % of the recyclables cannot be recycled from 2021 to 2060.In the other two scenarios, theses proportions are 65.00 %, 48.75 %, 32.5 %, 16.25 %, and 0 %.This quantity of non-recyclables sent to incineration will be separately indicated in the fifth column of the figure.
× (MSW-MSW*5.6 %)÷30 × 7 Data source: Chen et al., 2020.Z. Wen et al.C fuel− trans.= fuel consumption of vehicles during waste collection and transportation (L/d); D i = average distance from district centers to different treatment facilities (km); C empty = fuel consumption of a transport vehicle when unloaded (L/100 km); C full = fuel consumption of a transport vehicle when fully loaded (L/100 km).

Fig. 4 .
Fig. 4. Comparison of system-wide GHG emissions between the waste classification system and the traditional mixed collection and transportation system in 2021.

3 .
The waste amount sent to anaerobic digestion is the remaining 90 % from 2021 to 2060 under all three scenarios.4. The recyclable portion is 35 % from 2021 to 2060 in BAU scenarios.In the other two scenarios, these proportions are 35.00%, 51.25 %, 67.50 %, 83.75 %, and 100.00 %. 5. Vertical lines with arrows indicate changes in the amount of waste treated by the four disposal methods under the three scenarios.Arrows pointing down indicate a decrease, pointing up indicate an increase, and arrows pointing right indicate no change.6.The data label in the figure represents the data of the BAU Scenario.Data source: By calculation.

Fig. 10 .
Fig. 10.Waste volume of different components sent to treatment facilities under three scenarios from 2021 to 2060 ( × 10 4 t/yr).Note: The data label in the k-ure represents the data of the BAU Scenario.Data source: By calculation.

Table 1
Basic situation of household waste treatment facilities in each district in Beijing, 2021.
Z. Wen et al.

Table 2
Accounting links of GHG emissions and emissions offsets.

Table 3
Data for estimating CO 2 emissions from incineration based on household waste components.= the proportion of components in the recyclables (as wet weight) (%); dm i = fraction of dry matter content in the component of the household waste incinerated (%).CF i , FCF i and dm i data are available in IPCC 2006.N 2 O and CH 4 in this study are calculated based on the amount of household waste incinerated and the default emission factors of N 2 O and CH 4 .These values are then converted into their respective global warming potential.The calculation formulas are provided in equations (

Table 4
[39] for estimating CH 4 emissions from landfill based on household waste components. 1 = CH 4 correction factor for aerobic decomposition in the year of deposition, 1.0[39]; DOC = degradable organic carbon in the year of deposition (kg C/kg waste).The default values can be found in IPCC 2006; DOC f ,i = fraction of DOC that can decompose, 0.5[39]; W i,2 = number of different components of household waste treated by landfill (t/d; t/yr); M 2 = total amount of household waste as wet weight landfilled (t/d; t/yr); WF i,1 = proportion of components in the residual waste (as wet weight) (%).The data is presented in Table3; 16/12 = molecular weight ratio CH 4 /C.

Table 6
Basic data of vehicle transportation and total fuel consumption.

Table 9
Parameters of the household waste prediction model.
a. Independent variable variables: (Intercept), Electricity consumption, GDP, population b.Dependent variable: waste removal volume Z. Wen et al.

Table 10
Results of the household waste prediction model.

Table 11
Classification rate of the food waste, recyclables and residual waste from 2021 to 2060.
Note: Based on the waste quantity and composition from 2021, calculate the maximum separation rate for kitchen waste and recyclable materials.Z. Wen et al.

Table 12
Components of residual waste at different classification efficiencies.

Table 13
Components of recyclables at different classification efficiencies.The increased efficiency in waste classification has led to changes in the composition of recyclables.This data is calculated based on the data from Table10and the projected amount of waste.

Table 14
The proportion of treatment methods at different classification efficiencies.

Table 15
[53] et al., 2018ferent energy technologies.The recycling rate of recyclables: Based on the scenario for increasing recycling rates set byDong et al., 2018(assuming all recyclables can be recycled), this study establishes a target vision of achieving a 100 % recycling rate by the year 2060 for long-term development.2.The recovery efficiency of incineration power generation: According to the latest European policies, waste incineration facilities installed after 2009 should achieve an energy recovery rate of at least 65 %[53].Hence, this study establishes an incineration energy recovery rate of 40.05 %, which falls within the feasible scope of the long-term development goal.

Table 16
Moisture content (%) and average LHV (kJ/kg) of residual waste in Classification Efficiency Scenario.

Table 17
GHG emissions offsets of six components.