Global Plastics Outlook

Annex A. Modelling approaches used to compose the OECD Global Plastics Outlook Database

The Annex presents the methodologies applied to provide the estimates contained in this report and in the OECD Global Plastics Outlook Database.1 These estimates include, plastics use, plastic waste generation, plastic waste management and the environmental impacts of plastics, i.e. 1) leakage to the environment, detailing the macroplastics and microplastics fractions; 2) leakage to aquatic environments; 3) particulate matter emissions from tyre and brake abrasion; and 4) greenhouse gas (GHG) emissions.

The Annex contains the following sections:

• Overview of the modelling framework

• Modelling plastics use in ENV-Linkages

• Modelling plastic waste and end-of-life fates in ENV-Linkages

• Modelling plastic leakage to the environment (Technical University of Denmark)

• Modelling plastic leakage to aquatic environments (Laurent Lebreton)

• Modelling plastic leakage to terrestrial and aquatic environments (University of Leeds)

• Modelling plastic leakage to the terrestrial and aquatic environments (OECD Global Plastics Outlook Database)

• Modelling particulate matter emissions to air from tyre and brake wear (Norwegian Institute for Air Research)

• Modelling greenhouse gas emissions from plastics in ENV-Linkages.

1.1. Overview of the modelling framework

This section explains in more detail the methodologies employed to obtain the estimates of plastics use, waste and environmental impacts presented in the Outlook and included in the OECD Global Plastics Outlook Database. Estimates have been generated by building on output from the OECD computable general equilibrium (CGE) model ENV-Linkages (Chateau, Dellink and Lanzi, 2014[1]), by filling existing data gaps and by generating projections on the environmental impacts.

The modelling of economic flows, plastics use, plastic waste and environmental impacts involves different steps, as illustrated in Figure A A.1. Sectoral and regional economic estimates drive the evolution of plastics use over time. Volumes of plastics are then used to calculate generated waste, based on product lifespans of different applications. The waste generated is further broken down by waste treatment, i.e. recycled (collected for recycling), incinerated, landfilled, mismanaged and littered waste (Chapter 2), taking into account differences across regions. Finally, estimates for a subset of environmental impacts are calculated: leakage of microplastics and macroplastics to the environment, leakage to aquatic environments, particulate matter linked to tyre and brake wear and GHG emissions.

The analysis relies on a suite of modelling tools. More specifically, estimates of the economic flows, plastics use, plastic waste (Steps 1-4), and greenhouse gas emissions (within Step 5) rely on the OECD in-house modelling tools, while other environmental impacts (also in Step 5) rely on external models. Some of the information provided by external models in Step 5 have been used to calibrate the ENV-Linkages models in Steps 1-4.

Figure A A.1. Methodological steps

The OECD’s in-house dynamic CGE model ENV-Linkages is used as the basis to estimate the economic activities that drive plastics use in 2019. ENV-Linkages is a multi-sectoral, multi-regional model that links economic activities to energy and environmental issues. A more comprehensive model description is given in Chateau, Dellink and Lanzi (2014[1]). A description of the baseline scenario construction procedure is given in Chateau, Rebolledo and Dellink (2011[2]), while recent baseline results are illustrated in OECD (2019[3]).

The model is based on the Social Accounting Matrices (SAM) contained within the GTAP 10 database (Aguiar et al., 2019[4]). This database describes bilateral trade patterns, production, consumption and intermediate use of commodities and services, including capital, labour, tax revenues and use. The base year of the SAM and of the model is 2014. Therefore, to obtain estimates for 2019, the ENV-Linkages model was run to 2019 (Box A A.1 for an overview of the functioning of the model). The short-term changes to the economy from 2014 to 2019 reflect short-term economic changes captured in international databases: the OECD Economics Department (OECD, 2020[5]) and the International Monetary Fund (2020[6]).

For the development of this Outlook and the Global Plastics Outlook Database, ENV-Linkages has been enhanced to include data on plastics use, waste and waste treatment. In ENV-Linkages, plastics estimates follow economic estimates, and, more precisely, the evolution of the production and consumption of goods in different sectors and regions. Section 2 of this annex describes the implementation of plastics use in ENV-Linkages, while Section 3 of this appendix describes the modelling of waste generation and waste end-of-life fates. A summary of the data sources is available in Chapter 2.

Box A A.1. The ENV-Linkages model

Production in ENV Linkages is assumed to operate under cost minimisation with perfect markets and constant returns-to-scale technology. The production technology is specified as nested Constant Elasticity of Substitution (CES) production functions in a branching hierarchy. This structure is replicated for each output, while the parameterisation of the CES functions may differ across sectors. The model adopts a putty/semi-putty technology specification, where substitution possibilities among factors are assumed to be higher with new vintage capital than with old vintage capital. In the short run this ensures inertia in the economic system, with limited possibilities to substitute away from more expensive inputs, but in the longer run this implies a relatively smooth adjustment of quantities to price changes. Capital accumulation is modelled as in the traditional Solow/Swan neoclassical growth model, where economic growth is assumed to stem from the combination of labour, capital accumulation and technological progress.

Household consumption is the result of within-period maximisation behaviour which is formally implemented as an “Extended Linear Expenditure System”. A representative consumer in each region - who takes prices as given - optimally allocates disposable income among the full set of consumption commodities and savings. Saving is considered as a standard good in the utility function and does not rely on forward looking behaviour by the consumer. The government in each region collects various kinds of taxes in order to finance government expenditures. Assuming fixed public savings (or deficits), the government budget is balanced through the adjustment of the tax on consumer income. In each period, investment net-of-economic depreciation is equal to the sum of government savings, consumer savings and net capital inflows from abroad.

International trade is based on a set of regional bilateral flows. The model adopts the Armington specification, assuming that domestic and imported products are not perfectly substitutable. Moreover, total imports are also imperfectly substitutable between regions of origin. Allocation of trade between partners then responds to relative prices at the equilibrium.

Market goods equilibria imply that, on the one side, the total production of any good or service is equal to the demand addressed to domestic producers plus exports; and, on the other side, the total demand is allocated between the demands (both final and intermediary) by domestic producers and the import demand.

ENV Linkages is fully homogeneous in prices and only relative prices matter. All prices are expressed relative to the numéraire of the price system that is arbitrarily chosen as the index of OECD manufacturing exports prices. Each region runs a current account balance, which is fixed in terms of the numéraire.

As ENV-Linkages is recursive-dynamic and does not incorporate forward-looking behaviour, price-induced changes in innovation patterns are not represented in the model. The model does, however, entail technological progress through an annual adjustment of the various productivity parameters, including e.g. autonomous energy efficiency and labour productivity improvements. Furthermore, as production with new capital has a relatively large degree of flexibility in choice of inputs, existing technologies can diffuse to other firms. Thus, within the CGE framework, firms choose the least-cost combination of inputs, given the existing state of technology. The capital vintage structure also ensures that such flexibilities are larger in the long run than in the short run.

Source: Chateau, Dellink and Lanzi, (2014[1]).

The sectoral aggregation of the model adopted in this report is given in Table A A.1, while the regional aggregation is presented in Table A A.2.

Table A A.1. Sectoral aggregation of ENV-Linkages

Agriculture, fisheries and forestry	Manufacturing
Paddy rice	Food products
Wheat and meslin	Textiles
Other grains	Wood products
Vegetables and fruits	Chemicals
Oil seeds	Basic pharmaceuticals
Sugar cane and sugar beet	Primary rubber and plastic products
Fibres plant	Secondary plastic products
Other crops	Pulp, paper and publishing products
Cattle and raw milk	Non-metallic minerals
Other animal products	Fabricated metal products
Fisheries	Electronics
Forestry	Electrical equipment
Non-manufacturing Industries	Motor vehicles
Coal extraction	Other transport equipment
Crude oil extraction	Other machinery and equipment
Natural gas extraction	Other manufacturing including recycling
Other mining	Iron and steel
Petroleum and coal products	Non-ferrous metals
Gas distribution	Services
Water collection and distribution	Land transport
Construction	Air transport
Electricity transmission and distribution	Water transport
Electricity generation (8 technologies): Nuclear electricity; hydro (and geothermal); solar; wind; coal-powered electricity; gas-powered electricity; oil-powered electricity; other (combustible renewable, waste, etc.).	Insurance Trade services Business services n.e.s. Real estate activities
	Accommodation and food service activities
	Public administration and defence
	Education
	Human health and social work

Source: OECD ENV-Linkages model.

Table A A.2. Regional aggregation of ENV-Linkages

Macro regions		ENV-Linkages countries and regions	Most important comprising countries and territories
OECD	OECD America	Canada	Canada
		USA	United States of America
		Other OECD America	Chile, Colombia, Costa Rica, Mexico
	OECD Europe	OECD EU 22	Austria, Belgium, Czech Republic, Denmark, Estonia, Finland, France, Germany Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Netherlands, Poland, Portugal, Slovak Republic, Slovenia, Spain, Sweden
		Other OECD Europe	Iceland, Israel,1 Norway, Switzerland, Turkey, United Kingdom
	OECD Pacific	OECD Asia	Japan, Korea
		OECD Oceania	Australia, New Zealand
Non-OECD	Other America	Other Latin America	Non-OECD Latin American and Caribbean countries
	Eurasia	Other EU	Bulgaria, Croatia, Cyprus,2 Malta, Romania
		Other Europe and Caspian	Non-OECD European and Caspian countries, incl. Russian Federation
	Middle East and Africa	Middle East and North Africa	Algeria, Bahrain, Egypt, Iraq, Islamic Rep. of Iran, Kuwait, Lebanon, Libya, Morocco, Oman, Qatar, Saudi Arabia, Tunisia, United Arab Emirates, Syrian Arab Rep., Western Sahara, Yemen
		Other Africa	Sub-Saharan Africa
	Other Asia	China	People’s Republic of China, Hong Kong (China)
		India	India
		Other non-OECD Asia	Other non-OECD Asian and Pacific countries

Notes: 1. The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law. 2. Note by Turkey: The information in this document with reference to “Cyprus” relates to the southern part of the Island. There is no single authority representing both Turkish and Greek Cypriot people on the Island. Turkey recognises the Turkish Republic of Northern Cyprus (TRNC). Until a lasting and equitable solution is found within the context of the United Nations, Turkey shall preserve its position concerning the “Cyprus issue”. Note by all the European Union Member States of the OECD and the European Union: The Republic of Cyprus is recognised by all members of the United Nations with the exception of Turkey. The information in this document relates to the area under the effective control of the Government of the Republic of Cyprus.

Source: OECD ENV-Linkages model.

1.2. Modelling plastics use in ENV-Linkages

1.2.1. Volumes

The ENV-Linkages model has been extended to include plastics volumes, for both primary and secondary (recycled) plastics use. The data on plastics volumes is presented in million metric tonnes (Mt) and plastics use is split by region, polymer and application.

Volumes of primary plastics on data from Ryberg at al. (2019[7]), that updates and expands on the seminal work in Geyer, Jambeck and Law (2017[8]), providing a database for 2015. Since the estimates from Ryberg et al. (2019[7]) were provided on the one hand by region and application, and on the other hand by application and polymers, an assumption of homogeneity of polymers by application was taken to estimate the primary plastics use by region, polymer and application.

Secondary plastics volumes for 2015 were estimated following a methodology deriving secondary plastics through waste collected for recycling and recycling losses. Loss rates including sorting losses and reprocessing losses were estimated using a methodology developed by the University of Leeds based on a review of the literature (see Section 1.3.3 in this Annex).

The estimates of plastics use for 2019 are based on the 2015 year, using the link between plastics volumes in Mt and plastic inputs to sectors in USD, which are estimated relying on the OECD ENV-Linkages model. In addition, the OECD Global Plastics Outlook Database is complemented with plastics use for the past between 1950 and 2014, for two reasons. The first reason is to be able to accurately compute waste flows in the future, since plastic lifetimes can span up to decades. The second reason is to form the basis for the computation of environmental impacts, as for instance plastic leaked in the ocean accumulates over time.

The 1950-2014 historical plastics use is calculated following a step-wise approach. First, global plastics use is taken from the Geyer, Jambeck and Law (2017[8]) study. The regional split of plastics use builds then on weight-based estimates of waste, from a cross country regression of municipal solid waste on gross domestic product (GDP) per capita, using What a Waste 2.0 (Kaza et al., 2018[9]), multiplied by the regional consumption shares in 2015. Finally, for each region, the split by polymer and application is assumed to be constant prior to 2014, based on the estimates from Ryberg et al. (2019[7]). This methodology is constrained by data availability (and thus necessarily imperfect) but provides estimates of plastics use by region, polymer and application.

1.2.2. Economic flows

The ENV-Linkages model has been modified to include primary and secondary plastics production. ENV‑Linkages relies on the GTAP 10 database (Aguiar et al., 2019[4]), which provides economic flows by sector and regions for the year 2014. While in the original database primary and secondary plastic production are aggregated in the same sector (Rubber and plastic products [RPP]), the representation of plastic in ENV-Linkages was enhanced to allow the distinction of a technology producing primary plastic and an alternative technology producing secondary plastics.

Similar to coal power plants and gas power plants both providing the same good (electricity), these two technologies produce a similar plastic good, with an elasticity of substitution of 2. The production of plastic goods was thus split with two data sources. First, the total shares in production for primary and secondary plastics were taken from the volumes in tonnes described above (Ryberg et al. (2019[7]) for primary and own estimates for secondary plastics). Table A A.3 describes the calculated share for the secondary plastic production technology. Furthermore, the Exiobase 3 database (Stadler et al., 2018[10]) was used to adapt the cost structures. The main difference stems from the material inputs: the primary technology uses fossil fuels, while the secondary technology uses inputs from the chemical sector.

Table A A.3. Share of the secondary production technology

	Region	Share of secondary technology in 2015 (in tonnes)
OECD America	USA	3.9%
	Canada	4.6%
	Other OECD America	6.8%
OECD Europe	OECD EU countries	9.0%
	OECD Non-EU countries	5.9%
OECD Pacific	OECD Asia	6.3%
	OECD Oceania	2.4%
Other America	Latin America	7.5%
Eurasia	Other EU	4.9%
	Other Eurasia	3.5%
Middle East and Africa	Middle East & North Africa	3.5%
	Other Africa	4.5%
Other Asia	China	7.3%
	India	6.5%
	Other non-OECD Asia	4.8%

Source: OECD ENV-Linkages model. Primary plastics are based on Ryberg et al. (2019[7]). For the calibration of secondary plastics, recycling rate sources are detailed in Table A A.4 and loss rates for the recycling process are from Cottom et al. (2020[11]).

1.2.3. Link between volumes and economic flows

To model plastics use in ENV-Linkages, data on plastics volumes by application and polymer have been linked to the detailed sectoral production structure of the model and the GTAP 10 database that underlies the model and modified to include both primary and secondary plastics. Two main sources of data (volumes and economic flows described above) were used and put in coherence: 1) plastics production and consumption by economic sector by GTAP 10 adapted with a primary and secondary production technology in monetary values; and 2) regional flows of a range of plastic polymers and application-specific flows of plastics in tonnes. Table A A.4 summarises the mapping of the economic sectors and plastics applications. The initial values for this mapping are calibrated using data from Ryberg et al. (2019[7]), combining polymer distribution by application at the global level with distribution of total plastics use by region and application. The polymer distribution was taken from the global averages and applied for each region taking into account the specific economic structures of the various regions.

Table A A.4. Mapping plastics use by application to economic sectors

Input sectors	Applications	Output sectors	Polymers*
Plastic products	Building & construction	Construction	ABS, ASA, SAN; bioplastics; HDPE; LDPE, LLDPE; PP; PS; PUR; PVC; other
	Consumer & institutional products	Accommodation and food service activities; air transport; education; health; insurance; lumber; non-metallic minerals; Business services; other manufacturing; public services; land transport; pulp, paper and publishing; real estate; textile; water transport	ABS, ASA, SAN; bioplastics; HDPE; LDPE, LLDPE; PP; PS; PUR; PVC; other
	Electrical/electronic	Electrical equipment; electronics	ABS, ASA, SAN; bioplastics; HDPE; LDPE, LLDPE; PP; PS; PUR; PVC; other
	Industrial/machinery	Fabricated metal products; iron and steel; nonferrous metal; Machinery and equipment	HDPE; LDPE, LLDPE; PP; PUR
	Packaging	Food products; chemical products	Bioplastics; HDPE; LDPE, LLDPE; PET; PP; PS; PUR; PVC; other
	Personal care products	Chemical products	HDPE; PET
	Transportation - other	Motor vehicles; public services; other transport equipment	ABS, ASA, SAN; bioplastics; Fibres; HDPE; LDPE, LLDPE; PP; PUR; PVC; other
	Other	Other sectors	Other
Chemicals	Marine coatings	Other manufacturing, other transport equipment	Marine coatings
	Road markings	Construction	Road markings
	Textile sector - clothing	Textiles	Bioplastics; fibres
	Textile sector - others	Textiles	Fibres
	Transportation - tyres	Plastic products	Elastomers (tyres)

Note: *See Table 2.2 in Chapter 2 for abbreviations and examples of use for those polymers.

Source: OECD ENV-Linkages model.

Based on the initial GTAP database, which presents data for 2014, primary plastics use is projected following the flows of “plastic products” into the various corresponding demand sectors, from initial values, following the methodology developed for the OECD’s Global Material Resources Outlook (OECD, 2019[3]). In particular, the model incorporates a series of plastics chains from initial production to final demand, either partially or in full depending on the particular structure of each regional economy. The basis for the chain includes flows from “oil” or ”biomass” to “chemicals”, that are then used for the production of “plastic products” which serve as intermediate goods or for sectors such as food product/appliances/motor vehicles/construction, before reaching final demand. The underlying assumption is that the coefficient (USD/tonne per polymer, per application, per region) that links monetary flows to physical flows (in tonnes), is kept constant. Plastics production then follows these demands, based on trade flows and plastics use.

There are three steps to project plastics use and the split of primary and secondary plastics to fulfil demand (after 2015). First, total demand for plastics use is estimated following the evolution of the demand for the plastic commodity (produced by both the primary and secondary technologies). Second, as collected and sorted materials (further referred to as plastic scrap) are – after correcting for loss rates (see Section 1.3.2 on Losses from sorting and reprocessing in this Annex) - used to produce secondary plastics, the tonnes of secondary plastics follow the growth of the secondary sector in the ENV-Linkages projections. Third, the volumes of primary plastics are calculated as a residual between the two such that total demand for plastics is met.

1.3. Modelling plastic waste and end-of-life fates in ENV-Linkages

1.3.1. Plastic waste

Plastic waste is calculated linking plastics use to the lifespan distribution of different applications (representing a large range of products). Specifically, plastic waste is calculated as a function of plastics use (in volumes), following Geyer, Jambeck and Law (2017[8]), using a methodology based on lifespan distributions,2 under the assumption of global homogeneity.

Plastic waste of different applications is grouped into three main categories: Municipal Solid Waste (MSW), Other, and Markings & microbeads. MSW includes packaging, consumer & institutional products, electrical/electronic and textiles. ‘Other’ incorporates waste that is not included in MSW, therefore mostly reflecting waste from industrial applications (including building and construction, industrial and machinery applications, transportation applications). Markings and microbeads is a very small stream that includes marine coatings, road markings and personal care products.

1.3.2. End-of-Life Fates

Plastic waste is divided into different waste management streams (end-of-life fates) by applying end-of-life shares that vary across countries, polymers and waste categories. MSW and Other plastic waste categories can be 1) recycled; 2) incinerated; or 3) discarded. The latter is further disaggregated into waste that is disposed of in sanitary landfills, mismanaged waste and, in the case of MSW, littering. Littering is included as a separate category to reflect the different drivers and geographical distribution compared to mismanaged waste (this occurs mainly in regions that lack basic waste management infrastructure). Littering is set as a constant share of municipal solid waste following the assumption in Jambeck et al. (2015[12]). Markings & Microbeads form a very small stream (by mass) that is assumed not to be managed and to leak directly to the environment.

The sources of end-of-life fate shares for the year 2019 vary across regions. Recycling (defined here as material that has been collected for recycling) shares for plastics are exogenously fixed based on a range of sources, primarily country sources (Table A A.5). To account for unreported informal recycling (which leads to understating plastic recycling rates) or overly optimistic reported recycling rates, all reported recycling rates were sense-tested, adapted and validated leveraging on consultations with experts and modelling carried out by Ed Cook, Josh Cottom and Costas Velis from the University of Leeds.

The recycling shares are further split across polymers by multiplying the recycling shares for plastics by factors that reflect the recyclability and value of individual polymers based on expert consultations and ensuring that the estimated recycled volumes do not exceed the recycling capacities subject to data availability. Overall, PET and HDPE are assumed to have the highest recycling rates, followed by LDPE, PP and PVC (for construction). PUR, fibres, elastomers, bioplastics, marine coatings and road markings are not recycled, while only a very small proportion of PS, ABS, ASA, SAN and other polymers are recycled.

Table A A.5. Data sources for plastic recycling rates in base year

Region	Recycling Rate Source and Assumptions
USA	United States Envrionmental Protection Agency (EPA) (2020[13]; 2020[14])
Canada	Environment and Climate Change Canada (2019[15])
Other OECD America	Based on SEMARNAT (2020[16]) and FCH (2021[17])
OECD EU countries	Polymer-specific recycling rates have been determined based on expert opinion and applied to the volumes of polymers collected for recycling by ENV-Linkages.*
Other EU
OECD Non-EU countries	Based on EU adjusted by the proportion of region’s MSW recycling rate to EU MSW recycling rate from What a Waste 2.0 (Kaza et al., 2018[9])
OECD Pacific	Plastic Waste Management Institute (2019[18]) and expert judgement to account for plastic recycling rates in Korea
OECD Oceania	Australian Government (2020[19])**
Latin America	Based on Other OECD America adjusted to account for a larger informal sector
Other Eurasia	What a Waste 2.0 (Kaza et al., 2018[9])
Middle East & North Africa	What a Waste 2.0 (Kaza et al., 2018[9])
Other Africa	What a Waste 2.0 (Kaza et al., 2018[9])
China	China Recycling Industry Development Report (2013-18) by the Ministry of Commerce (2019[20])
India	Central Pollution Control Board (2019[21]) and UNIDO (2020[22])
Other non-OECD Asia	What a Waste 2.0 (Kaza et al., 2018[9])

Note: *For the EU, the calculated recycling rate for total plastics has been benchmarked with the numbers presented by Plastics Europe (2020[23]). In the OECD Global Plastics Outlook Database, the amount of plastics taken into account for the calculation of the recycling rate is substantially higher (the denominator: total plastics in OECD Global Plastics Outlook Database includes fibres and other rarely recycled plastics). So the total recycling rate of plastics that is included in the OECD Global Plastics Outlook Database is lower than Plastics Europe (2020[23]). ** An updated report is available: Government of Australia (2021[24]).

The use of incineration as a waste treatment type is country-specific and related to historic elements and local population densities. The share of plastic waste that is incinerated is strongly correlated with the share of total solid waste that is incinerated. Therefore, the incineration shares are set so that the ratio of the incineration share to the non-recycled share is equal to the corresponding ratio for total MSW from the What a waste 2.0 database (Kaza et al., 2018[9]). Moreover, the same incineration shares apply for non-MSW plastic waste, namely the ‘Other’ waste category.

Regarding discarded waste, its share is equal to the residual, under the assumption that 2% of MSW is littered at all times (Jambeck et al., 2015[12]). The discarded share is further split into sanitary landfilled and mismanaged waste. In this analysis, mismanaged waste includes open dumping and unaccounted waste treatments for all income levels apart from lower and lower middle income countries, for which also unspecified landfilling, waterway treatment and other categories are included based on country level data for MSW (Kaza et al., 2018[9]) and building on assumptions for the previous version of the database in Jambeck et al. (2015[12]). In general, mismanaged plastic waste, as a share of total plastic waste, is expected to decrease with income level. Following this assumption and using MSW data from Kaza et al. (2018[9]), the share of mismanaged plastic waste was estimated by regressing the ratio of mismanaged waste to discarded waste on GDP per capita, accounting for regulatory differences between OECD and non-OECD countries using an OECD dummy. Specifically, the following regression was estimated for 156 countries for which complete data was available:

${MIS}_i/({MIS}_i+{LAN}_i)\mathrm{ }=\mathrm{ }\alpha \mathrm{ }+\mathrm{ }\beta \mathrm{ }\mathrm{*}\mathrm{ }ln\left({gdp\_pc}_i\right)\mathrm{ }+\mathrm{ }{OECD}_i\mathrm{ }$

where ${MIS}_i$ = mismanaged waste/MSW, ${LAN}_i$ = Landfilled waste/MSW, ${gdp\_pc}_i$ = GDP per capita and ${OECD}_i$ = dummy for OECD countries, $i$= country.

Finally, the share for landfilled waste is equal to the residual.

Historical data for recycling, incineration and discarded shares of plastic waste are taken from Geyer, Jambeck and Law (2017[8]) for the period 1980-90 for four regions – United States, European Union, China and Rest of the World. Following, using granular data for MSW recycling and incineration rates from Kaza et al. (2018[9]), the historical shares for 1990 were mapped to the 15 regions within ENV-Linkages, and were linearly interpolated for the period 1990-2018 in line with the methodology previously applied in Geyer Jambeck and Law (2017[8]). Historical data for mismanaged waste and landfilling followed the same methodology as in the base year.

1.3.3. Losses from sorting and reprocessing (University of Leeds)

Plastic waste that has been collected for recycling almost always includes some non-plastic materials and articles. Moreover, collected plastic waste typically includes a multitude of plastics with varying chemical and physical composition. The degree to which these items, objects and fragments are useful to a plastics reprocessor depends on a wide range of factors that influence the value of the material. In general, high income countries implement recyclate collection schemes (programmes) that are designed to yield high material mass through an accessible and simplified system that is easy for people to understand. Conversely, in low- and middle-income countries, plastic waste collection for recycling is carried out by informal workers (IRS) who selectively collect (cherry pick) items and objects that are most valuable, focusing on quality and concentration rather than high yield. Even with diligent, selective collection, plastic articles contain a multitude of intentionally and non-intentionally appended, entrapped, adhered and entrained materials and objects that must be removed from the dominant plastic before it can be comminuted and re-melted under pressure in an extruder. A list of characteristics of waste plastics and their influence on the value of materials and hence their recyclability is reported by Cottom et al. (2020[11]).

Robust and generalisable loss rates during sorting and reprocessing of plastic waste that has been collected for recycling are not commonly reported. Hestin, Faninger and Milios (2015[25]) proffered 18% and 30% for sorting and reprocessing respectively, based on surveys of European reprocessors. However, the nature of the survey was not reported and it is possible that plastic and non-plastic materials and objects have been reported alongside plastic losses. The ENV-Linkages model is only concerned with plastics so data for non-plastic materials were excluded from this component of the model.

The University of Leeds developed a theoretical model for plastic waste collected for recycling in high-income countries and low- and middle-income countries, based on material value. Acknowledging that collection and sorting systems vary enormously worldwide, these two generalised groups were chosen because high income countries largely operate, either single stream collection of dry recyclate or co‑collection of mixed plastic waste alongside metal packaging. Conversely in low-income and middle-income countries, collection of plastic waste for recycling is largely carried out by the informal recycling sector whose participants selectively collect materials and have much lower loss rates.

Recycling losses for packaging waste collected for recycling in high income countries, are estimated based on a dataset (Chruszcz and Reeve, 2018[26]) that reports a weighted average for all collection scheme types across the United Kingdom. For LDPE, an approximation was made based on data reported by Lau et al. (2020[27]) (P₂O model), because LDPE is predominantly used as a as a flexible foil in packaging. Although LDPE is commonly collected for recycling, it is almost never reprocessed in high income countries when coming from post-consumer household sources due to the challenges associated with surface contamination and selectivity detailed in Table A A.6. On the other hand, post-consumer LDPE from commercial sources is commonly recycled in high-income countries as it is easily collectable and separately and can be extruded dry, often without undergoing substantial cleaning. The result is a low loss rate. The assumptions from Lau et al. (2020[27]) were used to determine the proportion of material that was from commercial/institutional sources compared to household sources.

A probability of plastic waste items being selected at the sorting stage based on material value was applied to each of the packaging and plastic types. Cottom et al (2020[11]) calculate these probabilities for Consumer and Institutional products, Electrical and Electronic products, Packaging, and Textiles. Probabilities for other products have been calculated for this reports and are reported in Table A A.7. The probabilities were estimated using cost data summarised by SystemiQ and the Pew Charitable Trust (2020[28]), recyclability imperatives detailed by Recoup (2019[29]) and data on material actually recycled reported by Antonopoulos, Faraca and Tonini (2021[30]) and Plastics Recyclers Europe (2020[31]). In general HDPE, PET and LDPE were considered to have a 100% chance of being selected for reprocessing at the Materials Recovery Facilities (MRF) and PVC and PS were considered to have 0% chance of being selected for reprocessing at the MRF. Although the evidence for PVC for packaging is more clear cut, Antonopoulos, Faraca and Tonini (2021[30]) reported some post-consumer PS selection taking place in Europe. However, these quantities are reported by Plastics Recyclers Europe (2020) to be small and unusual, there is a likelihood that they do not refer to post-consumer material. The probability was set to zero for packaging but an overall probability of 98.5% was set to allow for some small occurrences of non-packaging material.

The loss rates at the reprocessor stage were approximated using data on plastic content reported by Roosen et al. (2020[32]).Non plastic content was excluded and the relative masses were normalised.

High income countries were assumed to have formal collection and the plastic packaging reported there was subject to loss rates at both sorting and reprocessing. Low and middle income countries were assumed to have informal collection and the loss rates were therefore assumed to occur only at the reprocessing stage as informal actors selectively collect.

The assumptions for non-packaging applications were based largely on estimates from the project expert team, as there are no published data to support them. Consumer and institutional products were assumed to be the same as packaging except for PVC for which evidence from VinylPlus (2019[33]) indicates some recycling takes place. For the textiles (fibres), an estimate of 20% from financial modelling by Thompson et al. (2012[34]) was used in the absence of any other robust data. Readers should note that this loss rate is approximated on the basis that post-consumer textiles have been recycled into shoddy fibres and/or flocking (stuffing) rather than items that have been ‘reused’ and are out of scope of this study.

Table A A.6. Assumptions used to determine loss rates for plastic packaging waste that has been collected for recycling

Plastic item1	Plastic type by dominant polymer1	Weighted composition1	High-income countries			Low- & middle-income countries
			Probability of being rejected before reprocessing2	Loss rate at reprocessor adjusted for wastage3	Net losses after sorting & reprocessing4	Probability of being rejected before reprocessing2	Loss rate at reprocessor adjusted for wastage3	Net losses after sorting & reprocessing4
Film LA recycling sacks	LDPE	2.9	100	0.00	2.90	25	1.00	0.75
FILM Other film	LDPE	11.2	100	0.00	11.20	25	1.00	2.88
FILM Carrier bags	LDPE	1.5	100	0.00	1.50	25	1.00	0.39
B PET NATURAL	PET	26.4	0	13.45	3.55	0	13.45	3.55
B PET JAZZ	PET	3.1	0	13.45	0.42	0	13.45	0.42
B HDPE Milk Bottles	HDPE	13.2	0	15.93	2.10	0	15.93	2.10
B HDPE All non-milk bottles	HDPE	7.7	0	15.93	1.23	0	15.93	1.23
B PVC ALL	PVC	0	100	0.00	0.00	100	0.00	0.00
B PP ALL	PP	0.4	50	21.31	0.24	0	21.31	0.09
Pack PET NATURAL	PET	10.3	0	14.63	1.51	0	14.63	1.51
Pack PET JAZZ	PET	0.5	0	14.63	0.07	0	14.63	0.07
Pack HDPE NATURAL	HDPE	0.1	100	0.00	0.10	0	14.63	0.01
Pack HDPE JAZZ	HDPE	0.6	100	0.00	0.60	0	14.63	0.09
Pack PVC ALL	PVC	0.1	100	0.00	0.10	100	0	0.10
Pack PP NATURAL	PP	4.4	100	0.00	4.40	0	2.08	0.09
Pack PP JAZZ	PP	5.3	100	0.00	5.30	0	2.08	0.11
Pack PS ALL	PS	1.5	100	0.00	1.50	100	0	1.50
Pack EPS ALL	EPS	0.4	100	0.00	0.40	100	0	0.40
Black PET	PET	1.9	100	0.00	1.90	100	0	1.90
Black PP	PP	0.6	100	0.00	0.60	100	0	0.60
Black Other	Mixture	1.1	100	0.00	1.10	100	0	1.10
Other	Mixture	0.2	100	0.00	0.20	100	0	0.20
Unidentified	Mixture	1.9	100	0.00	1.90	100	0	1.90
Plastic non-packaging	Mixture	4.4	100	0.00	4.40	100	0	4.40

Source: 1. Chruszcz and Reeve (2018[26]). 2. Assumptions based on polymer value by SYSTEMIQ & The Pew Charitable Trust (2020[28]), recyclability reported by Recoup (2019[29]), and material reported to have been recycled by Antonopoulos, Faraca and Tonini (2021[30]) and Plastics Recyclers Europe (2020[31]). 3. Roosen et al (2020[32]).

Table A A.7. Average loss rates by plastic type and application for high income countries and low- middle income countries (Non-MSW)

Plastic type by dominant polymer	Building and Construction			Industrial/Machinery			Other			Transportation - Other			Total (t y-1)
	Mass (t y-1)	Loss rate HIC (%)	Loss rate LMIC (%)	Mass (t y-1)	Loss rate HIC (%)	Loss rate LMIC (%)	Mass (t y-1)	Loss rate HIC (%)	Loss rate LMIC (%)	Mass (t y-1)	Loss rate HIC (%)	Loss rate LMIC (%)
Fibres							0.1	100.0	100.0	0.0	100.0	100.0	0.1
HDPE	1.0	20.0	5.0	0.1	20.0	5.0	1.1	20.0	5.0	0.6	98.0	90.0	2.9
LDPE, LLDPE	0.1	2.0	2.0	0.1	2.0	2.0	0.7	2.0	2.0	0.1	98.0	90.0	1.0
Other	0.0	100.0	100.0				0.2	100.0	100.0	0.1	100.0	100.0	0.4
PP	0.2	20.0	5.0	0.1	20.0	5.0				1.2	100.0	100.0	1.5
PS	0.1	100.0	100.0				0.1	100.0	100.0				0.2
PUR	0.1	40.0	10.0	0.0	40.0	10.0	0.4	40.0	10.0	0.2	100.0	100.0	0.6
PVC	0.6	18.73	18.73				0.4	40.0	18.73	0.1	100.03	100.03	1.1

Source: Calculations by Leeds University. 1. Calculated from Chruszcz and Reeve (2018[26]) and Roosen et al. (2020[32]). 2. Calculated from Lau et al. (2020[27]). 3. Approximated from data reported by VinylPlus (2019[33]). 4. Thompson Willis and Morley (2012[34]).

For simplicity, The European Union, the USA, and Canada were considered to have formal collection and all other regions were considered to have predominantly informal collection for recycling. The exception was China which has been undergoing a partial transition from informal to formal collection for recycling. Due to the lack of robust data on the informal recycling sector, this component of the model assumed a 70:30 ratio for informal/formal collection for recycling.

Table A A.8 puts forward the outcome of the technical calculations. The calculations initially suggested loss rates of 100% for PS and of 98.1% for “Other”, but these loss rates have both been set to the loss rate from LDPE, to represent that these polymers are sometimes recycled, but only in small quantities. Furthermore, to reflect that a large share of recycling of PET is rather a downcycling transformation of PET into fibres, the modelling assumes 35% of recycled PET is transformed into fibres.

Table A A.8. Average loss rates by plastic type and region for MSW and non-MSW combined

Loss rates as percentages of collected plastics for recycling, 2016

		ABS, ASA, SAN	Fibers	HDPE	LDPE, LLDPE	Other	PET	PP	PS	PUR	PVC	Average
OECD America	USA	100%	54%	27%	60%	44%	18%	94%	49%	57%	43%	50%
	Canada	100%	55%	26%	60%	43%	18%	95%	49%	57%	43%	49%
	Other OECD America	100%	64%	17%	59%	29%	18%	24%	50%	28%	47%	30%
OECD Europe	OECD EU countries	100%	59%	24%	57%	38%	18%	90%	42%	53%	35%	44%
	OECD Non-EU countries	100%	60%	24%	60%	36%	18%	92%	45%	53%	37%	46%
OECD Pacific	OECD Asia	100%	48%	22%	62%	30%	18%	92%	46%	50%	35%	44%
	OECD Oceania	100%	30%	26%	67%	64%	18%	50%	45%	56%	41%	40%
Other America	Latin America	100%	64%	16%	60%	26%	18%	22%	54%	27%	54%	29%
Eurasia	Other EU	100%	56%	26%	63%	58%	18%	52%	44%	53%	44%	41%
	Other Eurasia	100%	55%	26%	65%	58%	18%	50%	49%	55%	49%	41%
Middle East and Africa	Middle East & North Africa	100%	54%	19%	64%	38%	18%	31%	49%	36%	45%	33%
	Other Africa	100%	71%	16%	61%	22%	18%	21%	53%	24%	56%	30%
Other Asia	China	100%	54%	20%	62%	34%	18%	48%	54%	39%	54%	36%
	India	100%	34%	19%	67%	39%	18%	27%	59%	38%	60%	32%
	Other non-OECD Asia	100%	47%	23%	64%	49%	18%	41%	55%	47%	56%	37%
Global average		100%	53%	22%	61%	38%	18%	63%	49%	46%	43%	40%

Source: OECD ENV-Linkages model, based on Cottom et al. (2020[11]).

1.3.4. Modelling international trade in plastic waste

The ENV-Linkages model has been extended to include international trade in plastic waste per application and polymer type. Volumes of plastic waste exports and imports are calculated based on data from UN Comtrade (United Nations Statistics Division, 2020[35]) following two steps. First, total exports of plastic waste per country and polymer are estimated using the share of plastics exports (Comtrade) to plastic waste (output of ENV-Linkages). Second, exports are split into partner countries and polymers using the projected country and polymer weights for 2019, and historical data for the years before. Bilateral export and import weights per country (row weights) were calculated based on the bilateral data of export and import values for the period 2010-19 (most recent and complete year) and for the four subcategories of plastic waste reported in the UN Comtrade database. The latter were mapped to the polymer types included in ENV-Linkages (Table A A.9), to ensure that global trade balances total imports and exports, bilateral plastic waste imports per reporter-partner pair correspond to the bilateral export of the corresponding partner-reporter pair.

The end-of-life fates of plastic waste that is traded, differ from the domestically treated waste to reflect that a high proportion of traded plastic waste tends to be recyclable. In particular, 50% of traded plastic waste is expected to be recycled, with the remaining being distributed across the other waste streams following the same proportions of end-of-life fates as domestically treated waste excluding littering.

Table A A.9. UN Comtrade plastic waste series mapping to polymers in ENV-Linkages

UN Comtrade code	Series Description	Polymers types in ENVLinkages
3915	Waste, parings and scrap, of plastics
391510	… of polymers of ethylene	HDPE,LDPE, LLDPE, PET, PP,PUR, Elastomers (tyres)
391530	… of polymers of styrene	PS
391530	… of polymers of vinyl chloride	PVC
391590	… of other	Fibres, Marine coatings, Road marking coatings, ABS, ASA, SAN, Other

Source: United Nations Statistics Division (2020[35]) and OECD ENV-Linkages model.

1.4. Modelling plastic leakage to the environment (Technical University of Denmark)

Estimations on the plastic leakage are based on an interaction of the ENV-Linkages Model with other dedicated models. Each of the dedicated models builds on earlier work that has passed peer review with respect to estimations for current plastic leakage. The sources for leakage to the environment are varied. Consequently, the modelling techniques to make projections on these flows differ. This section explains the methodology and parameters employed by Teddy Serrano, Alexis Laurent, and Morten Ryberg from the section for Quantitative Sustainability Assessment at the Technical University of Denmark (DTU) to estimate leakage of macro and micro plastics into the environment, as well as wastewater pathway and losses via sludge application to land.

1.4.1. Leakage of macroplastics

For losses of macroplastics, four main categories have been considered: mismanaged municipal solid waste, mismanaged non-municipal solid waste, littering, and losses from marine activities. Plastic waste generation is calculated by the ENV-Linkages model as explained in the previous sections. The details on the calculations of the four categories are as follows:

• Mismanaged MSW was retrieved from the ENV-Linkages model. In line with Lau (2020[27]), it was assumed that 32% of mismanaged MSW is lost to the environment.

• Mismanaged non-MSW was also retrieved from the ENV-Linkages model. Due to a lack of data on the fate of mismanaged non-MSW, the share of mismanaged non-MSW lost to the environment is assumed to be equal to the share of mismanaged MSW lost to the environment (32%).

• Losses occurring via littering were calculated as a fraction of generated MSW in two steps. First, in line with Jambeck et al. (2015[12]) and studies carried out for the United Kingdom and Belgium (OVAM, 2018[36]; Resource Futures, 2019[37]), it was assumed that 2% of MSW is littered. Second, a substantial fraction of this littered waste happens in an urban environment and is cleaned up before it makes it to the environment. It is assumed that between 15% and 35% of littered waste is not captured by street sweeping, storm drain catchments and pump stations (Jambeck et al., 2015[12]). The estimated share of litter that is lost to the environment in each region, was established according to the income level (as GNI/cap, US dollars), with lower shares for the high-income regions, as illustrated in (Table A A.10.).

• In the ENV-Linkages model, the non-collected share of litter is considered lost to the environment. The collected share of litter is reallocated and added to incineration, landfilling, open-pit burning and dumping in line with the respective waste treatment shares per region (see 1.3.2 in this Annex).

• Losses from marine activities (fishing gear and non-netting waste) were calculated based on the following information: production data of fishing gear in Europe (PRODCOM, 2016) (Eunomia, 2018[38]; Eurostat, n.d.[39]) upscaled to the rest of the world based on the projected growth of fishing activity (from the ENV-Linkages model), the assumption that 28% of plastic waste in the fishing and aquaculture sector comes from netting (Viool et al., 2018[40]), and the assumption that 15% of fishing gear material is lost every year during use (Viool et al., 2018[40]).

Table A A.10. The share of litter lost to the environment in function of the regional income levels

Category	Low and lower-middle income	Upper-middle income	High income
Income level as GNI/cap [USD]	< 4045	4045-12535	> 12535
Share of litter lost to the environment	35%	25%	15%

Note: The World Bank country classifications by income level: 2020-21 were used to allocate regions into low and lower-middle income, upper-middle income, or high income categories.

Source: World Bank (2020[41]).

1.4.2. Leakage of microplastics

For losses of microplastics, ten categories have been considered: microbeads, primary pellets, textile wash, tyre abrasion, road markings, brake dust, artificial turf, marine coatings, microplastics dust and wastewater sludge. This section presents the methodology employed to calculate emissions of microplastics from the sources considered. Some microplastics directly leak into the environment, but others end up in the sewage system. The fate of the different microplastics ending up in municipal wastewater networks, is discussed in the next section.

The category “microbeads” includes losses of microplastics intentionally added to rinse-off personal care and cosmetic products, detergents and maintenance products that are discharged into municipal wastewaters during use. Estimates of microbead consumption in personal care and cosmetic products (PCCPs) are derived from the output of the ENV-Linkages model. All microbeads are assumed to end up in the sewage system in the year that they are consumed.

The category “primary pellets” includes losses of primary plastic pellets occurring during production, transportation, and handling. Eunomia (2018[38]) estimated the losses of plastic pellets in 2015 in the European Union, as originating from pellet production from raw materials, intermediary handling processes, processing and conversion, off-site waste management, and transportation and shipping. Assuming that leakage is proportional to the quantity of plastics produced, losses for the European Union were scaled up to the entire world based on the European production share of plastics in 2015 (Plastics Europe, 2017[42]), and then allocated to geographical regions based on production shares. Losses from producers, recyclers, processors and offsite waste management were assumed to enter the sewage network as part of wastewater. Losses from Intermediary facilities and Shipping were assumed to be directly lost to the environment.

The category “textile wash” includes losses of synthetic microfibres lost during the washing of textile and apparel products. Estimates are computed based on the total volume (tonnes) of plastics used in the category ‘Wearing apparel’ (following the textiles sector in ENV-Linkages) in a given year, and the assumption that during the lifespan of a textile product 0.4% of material is lost during washing. The share of material lost during the lifespan of a textile and apparel product was calculated based on an assessment of existing studies accounting for the share of synthetic material lost due to washings over several wash cycles (De Falco et al., 2019[43]; Pirc et al., 2016[44]). It was assumed that all microfibres released during washing enter the sewage system.

Three sources of microplastics emissions from road transport were taken into account:

• The category “tyre abrasion” includes losses of elastomers originating from the abrasion of tyre treads of cars, trucks, and motorcycles. Emission estimates are derived from traffic data on the yearly activity in vehicle-km for passenger cars and in tonne-km for trucks from 2016 to 2019 in each region (retrieved from the ENV-Linkages model). Wear rates (i.e. average mass of tyre tread lost per vehicle-km, by vehicle type) employed are those reported by Eunomia (2018[38]). For trucks, an average freight tonnage of 16 t/vehicle was estimated, based on data from Eurostat (2018[45]). It was assumed that 46% of tyre treads is of elastomer content (Sommer et al., 2018[46]), and that the fate of the particles is as follows: 45% are retained in the asphalt pavement or remain close to the road, 45% is transported by road runoff and 10% is airborne (OECD, 2021[47]). The share of particles lost into the environment is dependent on the rural/urban population share of each region (as also used in the ENV-Growth and therefore ENV-Linkages model). In rural regions, road runoff and airborne emissions are considered as lost to the environment, whereas the particles trapped in the asphalt/road sides are not. In urban regions, airborne emissions are considered as lost to the environment, particles trapped in the asphalt/road sides are not, and particles as part of road runoff are assumed to go to a sewer system and treated as in wastewater the region where the loss occurs.

• The category “road markings” includes losses from markings applied to road surfaces. Estimates of plastics use for road markings are generated by the ENV-Linkages model, and the fate of road marking particles has been assumed to be similar to that of tyre abrasion particles due to a lack of data.

• The category “brake wear” includes losses of synthetic polymers originating from the wear of brake pads and other components. From the average composition of brake pads described by Hallal et al. (2013[48]), the polymer content of brake pads was assumed to be 23%. Similarly to the methodology used for tyre abrasion, loss estimations were based on annual traffic data and abrasion rates based on calculations by Eunomia (2018[38]).The fate of the microplastics in brake dust was assumed to be similar to that of tyre abrasion particles.

The category “artificial turf” includes losses of plastics from the infill of sport turfs. Estimates in the literature find losses of 300-730 kg / year per field in Denmark and 550 kg/year in Sweden (Løkkegaard, Malmgren-Hansen and Nilsson, 2018[49]; Swedish EPA, 2019[50]). According to ECHA (2020[51]), the number of artificial sport pitches will reach 39 000 by 2020 and average infill use is between 40 and 120 tonnes of material. Assuming that annual infill consumption is 1-4% of the total volume (ECHA, 2020[51]; Eunomia, 2018[38]), average yearly infill is 101 400 tonnes. Estimates for Europe were upscaled to other regions based on artificial turf market size figures (ResearchNester, 2021[52]) and GDP growth estimates (from the ENV-Linkages model). Based on the composition of rubber granulate used as infill, it was assumed that 96% of all infill is microplastics.3 In terms of losses and environmental fate, in line with Løkkegaard, Malmgren-Hansen and Nilsson (2018[49]), it was assumed that:

• 10% of rubber granulate particles are lost to the surrounding soil (and therefore considered as lost to the environment).

• 10% are discharged with water. Based on the rural share of the population in each region provided by the ENV-Linkages model from 2016 to 2019, it was assumed those 10% are considered as directly lost to the environment in rural areas. In urban areas, they are considered to enter the wastewater network. For those reaching a treatment system (primary, secondary, tertiary), all particles are assumed to be removed and therefore end up in sewage sludge, since the relatively large size of turf crumbles allows them to be usually well removed in treatment plants.

The category “marine coatings” includes losses of paint and coatings worn off from ships and marine structures. It is expected that 10% of plastics employed in the production of marine coatings is lost over the lifespan of the product, directly into the environment (Boucher and Friot, 2017[53]).

The category “microplastics dust” is used to refer to unintentional losses of microplastics occurring during the use phase of a number of products. Specifically, in the model five sources were taken into account: microplastics in household textile dust, the wear off of paint from interior surfaces, the wear off of paint from exterior surfaces, losses from construction and demolition activities, and shoe sole abrasion. These categories do not embody an exhaustive list of all remaining microplastics losses, but only put forward those categories for which sufficient literature has been found to quantify the resulting leakage.

For each source, with the exception of household textile dust, estimates are based on reported losses at the scale of a country or the European Union, which have been scaled down to calculate per capita emissions or per USD of GDP at constant Purchasing Power Parity (PPP) created, and finally scaled up to calculate the emissions for the entire world for each year, using data provided by the ENV-Linkages model. For interior and exterior paints, as well as exterior construction and demolition sources of dust, GDP was used as a scaling proxy under the assumption that the use of these materials is correlated to wealth.

For shoe sole abrasion, population data was considered a more relevant proxy. Because a person can only wear one pair of shoes at a time, the wear of shoes is assumed to be dependent on the activity of the person and not on wealth. In lack of better data, trends in the use of shoes are assumed to follow population trends.

The losses estimations of household textile dust are based on a recent study, according to which airborne-emitted synthetic fibres from textile and apparel products could represent a third of the particles lost to water during washings (De Falco et al., 2020[54]). Therefore, the emissions of textile fibres previously calculated during textile wash were used to calculate the losses of household textile dust.

A summary of the references used to calculate the losses from can be found in Table A A.11. It was assumed that 15% of household textile dust (Kawecki and Nowack, 2020[55]) and 100% of microplastics from interior paints end up in wastewater. For other sources, particles emitted in urban areas were also assumed to enter wastewater systems, whereas they were considered lost to the environment for rural areas.

Table A A.11. Sources for losses from microplastics dust

Microplastics dust sources	Reference (Country or region)	Scaling method
Household textile dust	ENV-Linkages model’s textile projections	-
Interior paints	Eunomia (2018[38])(EU)	GDP (USD, PPP)
Exterior paints	Eunomia (2018[38]) (EU)
Exterior construction and demolition	Kawecki and Nowack (2020[55]) (Switzerland)
Shoe sole abrasion	Lassen et al. (2016[56]) (Denmark)	Population

The category “wastewater sludge” includes losses of microplastics occurring via the application of wastewater sludge to land, as detailed in the next section.

1.4.3. The wastewater pathway and losses via sludge application to land

A large share of the emitted microplastics end up in wastewater or stormwater runoff (OECD, 2021[47]). To estimate the quantities of microplastics that reach the environment, a stylised flow of relevant end-of-pipe treatment systems has been developed as illustrated in Figure A A.2. The model considers a number of possible fates for microplastics, in line with Ryberg et al. (2019[7]). Ultimately, microplastics can either be retained by wastewater treatment or be lost to the environment.

Figure A A.2. Fate of microplastics in wastewaters

Source: Methodology adapted from Ryberg et al. (2019[7]).

The share of microplastics emissions ending up in different pathways depends on the state of wastewater infrastructure coverage in different countries. Allocation shares for each fate were estimated on a regional level. For each region, most allocation shares leading to treatments (represented by yellow boxes in Figure A A.2) were calculated using allocation shares averages of the countries composing the region, weighted by the population of each country. An assessment of data for 187 countries showed high variability in data availability and quality across countries. For most OECD countries, as well as Brazil, Colombia, and South Africa, the latest available data from the OECD Environment Database (2017[57]) was used and considered representative for wastewater treatment in 2016. For China and India, allocation shares were based on Kalbar, Muñoz and Birkved (2017[58]).

For other countries, the reference data comes from the WHO/UNICEF Joint Monitoring Programme for Water Supply, Sanitation and Hygiene (JMP, 2020[59]). This data is used for monitoring development in SDG 6.3.1 “Proportion of safely treated domestic wastewater flows (%)”. In the dataset, the following classification is used:

• Safely managed: use of improved facilities which are not shared with other households and where excreta are safely disposed in situ or transported and treated off-site.

• Basic: use of improved facilities which are not shared with other households.

• Limited: use of improved facilities shared between two or more households.

• Unimproved: use of pit latrines without a slab or platform, hanging latrines or bucket latrines.

• Open defecation: disposal of human faeces in fields, forests, bushes, open bodies of water, beaches and other open spaces or with solid waste (JMP 2020).

The “safely managed” share of the wastewater was assumed to at least undergo primary treatment. The remaining share of the wastewater is modelled as being directly released to the environment. Although this is a conservative assumption, it was not possible to retrieve more detailed data on the treatment levels for certain regions.

A microplastics removal rate was assigned to different levels of wastewater treatment (primary, secondary, and tertiary), as illustrated in Table A A.12 and employed to calculate the fate of microplastics passing through wastewater treatment, following the approach by Ryberg et al. (2019[7]). The removal rate of unspecified and independent wastewater treatment was assumed equal to the removal rate for primary treatment. Regional data on loss of wastewater due to overflow (represented by blue boxes in Figure A A.2) is generally lacking and the loss share was therefore modelled using the same loss shares for all regions. It is estimated that 0.6% and 2.4% of the wastewater is lost due to overflow of the sewer system and of the waste water treatment plant (WWTP), respectively (Magnusson et al., 2016[60]; Ryberg et al., 2019[7]).

Table A A.12. Microplastics removal rate for different levels of wastewater treatment

Treatment technology	Microbeads	Fibres	Other microplastics
Primary treatment	86%	87%	69%
Secondary treatment	92%	92%	96%
Tertiary treatment	99%	96%	99%

Source: Calculations from Michielssen et al. (2016[61]).

Wastewater sludge is the waste by-product of wastewater treatment containing the water pollutants removed from the influent. Sludge reuse for agricultural applications is encouraged in several countries, mainly due to the high nutrient content and its beneficial effects on crops, as well as to reduce the need for landfilling or incineration. However, recent evidence suggests that this practice leads to the transfer of a share of the microplastics retained during wastewater treatment to agricultural land (Nizzetto, Futter and Langaas, 2016[62]).

Losses into the environment via agricultural land were calculated based on the share of sludge generated in a given year that is applied on agricultural land. Due to data scarcity on the fate of microplastics during sludge treatment, it was assumed that there is no further removal of microplastics before sludge is applied to land (Ryberg et al. 2019). For Canada, China and the United States, the share of sludge applied to agricultural land follows the fractions reported by Rolsky et al. (2020[63]) (i.e. 43%, 45% and 55% for Canada, China and the United States, respectively). Due to a lack of data, the share of wastewater sludge applied on agricultural fields in all other countries was assumed to be equal to the European average (i.e. 46%) (Eurostat, 2020[64]).

1.5. Modelling plastic leakage to aquatic environments (Laurent Lebreton)

This section explains the methodology and parameters employed by Laurent Lebreton to estimate the amount of leaked plastics ending up in aquatic environments and assesses their mobility as well as degradation in rivers and oceans.

With a wide variety of polymer types, object shapes and sizes, and the dynamic nature of aquatic environments, quantifying sources and the fate of plastics in rivers, lakes, and the ocean is not trivial. Some studies have recently attempted to quantify the amount of mismanaged plastic waste generated by countries worldwide, which likely reach an aquatic environment (Borrelle et al., 2020[65]) and subsequently the ocean (Meijer et al., 2021[66]). These studies utilise spatial models describing the generation of mismanaged plastic waste in relation to topography and other environmental parameters. This section raised country-scale emission results to the modelled global regions represented in the ENV-Linkages model. The transport of emitted plastics was estimated considering geographical variations. Then the fate of plastics for the different regions was modelled as a function of polymer types predicted by projections of waste generation from various sectors of the economy (from ENV-Linkages). Finally, the mass of plastics accumulated in different aquatic environments for each region is reported.

To calculate inputs of plastics by region into aquatic environments, the analysis starts from the leakage to environment calculated by the ENV-Linkages model. The model then computes the probability for leaked plastics to reach an aquatic environment (rivers, lakes, and oceans) as a function of distance and terrain slope direction. The analysis uses the national probability of emissions of plastics into aquatic environments, adapted from Borelle et al. (2020[65]). The probability is independent of the total amount of leaked plastic waste but may differ around the world as a function of population location and topography of countries. In this study, the probability of emissions by region was computed by weighing country scale emission probability by population size and formulating a regional average including confidence intervals (Figure A A.3). The likelihood of plastic waste emissions varies by region. Island nations with predominantly coastal populations have the highest chance that plastics leaked to the environment end up in aquatic environments.

Figure A A.3. Weighted probability of leaked plastics entering aquatic environments

Note: The probability was calculated as the average of the country-scale fraction of emissions weighted by the population size of countries.

Source: Calculations based on Borrelle et al. (2020[65]).

In freshwater, floating plastics may be transported downstream while sinking plastics (plastics with a larger density than freshwater, e.g. PET, PVC or PS) will inevitably reach bottom sediments. Floating plastics may also be retained in freshwater environments in vegetation bordering the river, sediments in the river banks, artificial barriers (e.g. dams), or lakes. Some floating plastics may also be colonised by organisms and sink due to loss of buoyancy. A recent study estimating direct global inputs of plastics into the ocean via waterways reported that only 1% to 2% of plastics leaked annually have a chance to reach the sea globally within a year (Meijer et al., 2021[66]). The study utilised the same probability framework derived from location and quantities of lost plastics to the nearest river network. Still, it computed additional transport probabilities to river mouth from distance to the river mouth, river discharge, and river network order. The results show around one order of magnitude less discarded plastics reaching the ocean than the whole aquatic environment globally, including freshwater ecosystems (Table A A.13). This suggests that a large fraction of emitted plastic waste is likely still retained inland.

Table A A.13. Fractions of leaked macroplastics that enter aquatic environments and that reach the ocean

Macro region	Region	Fraction of leaked macroplastics entering aquatic environments	Fraction of aquatic plastics reaching the ocean
OECD America	USA	32%	3%
	Canada	36%	3%
	Other OECD America	21%	5%
OECD Europe	OECD EU countries	34%	3%
	OECD Non-EU countries	34%	4%
OECD Asia	OECD Pacific	43%	11%
	OECD Oceania	44%	2%
Other America	Latin America	28%	5%
Eurasia	Other EU	27%	1%
	Other Eurasia	32%	1%
Middle East and Africa	Middle East & North Africa	27%	4%
	Other Africa	23%	4%
Other Asia	China	28%	2%
	India	26%	4%
	Other non-OECD Asia	34%	14%

Source: Fraction of mismanaged and littered plastic waste entering aquatic environments (adapted from Borrelle et al. (2020[65])) and fraction of waste in aquatic environment entering the ocean environment (adapted from Meijer et al. (2021[66])) by region.

In the ocean, plastics with a larger density than seawater will sink to the bottom, accumulating in deep-sea canyons and trenches by the action of gravity. Floating plastics, however, will be transported by the action of waves, wind and currents. The largest part of these plastics, however, will rapidly reencounter land and beach on a coastline. A study presenting a model of dispersion of plastics in the ocean from global coastal sources reported that within a year, around 97% of released model particles had resided near a coastline for more than two consecutive days (Lebreton and Andrady, 2019[67]), suggesting a significant fraction had likely beached in that time. Rich coastal ecosystems will also facilitate the retention of floating plastics near the coastline as, similarly to freshwater environments, organisms in the marine environment will colonise floating plastics. Objects with smaller volume to surface ratios, such as plastic films or small microplastics, will likely sink near the coastline. Fragments and objects with a sufficiently large volume to maintain their buoyancy can escape the coastal environments. Over time debris tend to accumulate offshore in subtropical latitudes. Five accumulation zones have been widely reported in the literature from field observations and numerical models. The largest one is located in the North Pacific Ocean between Hawaii and California (Lebreton et al., 2018[68]).

Environmental conditions will also dictate the fate of plastics during their journey in freshwater and marine environments. Particularly under the action of sunlight, plastics degrade by photo-oxidation. As such, it is expected that plastics near the surface in rivers, lakes, or in the ocean are more likely to degrade into smaller particles, commonly referred to as microplastics with varying definitions (usually, particles below 1-5 mm and larger than one micron). Due to the large complexity of mechanisms and under varying conditions, data on the degradation of plastics in natural environments is scarce. Still, results are starting to appear with long-term experiments on the degradation of plastics in controlled environments. Fragmentation rates expressed in the percentage of weight loss per year did not exceed 5% in a laboratory seawater microcosm for various conventional thermoplastics (Gerritse et al., 2020[69]). This is in good agreement with modelled whole-ocean plastic degradation rates expected by numerical models (i.e. 3% of total ocean plastic mass degraded per year from macro- to microplastics, (Lebreton and Andrady, 2019[67]).

For the purpose of this work, the whole-ocean plastic mass budget model presented in Lebreton et al. (2019[70]) was expanded to a simplified representation of the global aquatic environment. The model now differentiates between annual inputs in freshwater and the ocean, allowing floating plastic waste to circulate from one compartment to the other over time. The model was also enhanced by differentiating inputs by polymer types using the OECD waste projections (presented in Chapter 2). The likely fate of emitted plastics was determined depending on their density. Additionally, the degradation rates varied between polymers based on laboratory results (Gerritse et al., 2020[69]). The general model framework is presented in Figure A A.4. To differentiate between freshwater and marine environment inputs, the model uses the results from Meijer et al. (2021[66]), which provides country-scale probabilities of emissions to the ocean. These results were upscaled to the modelled region by following the same weighted method as for inputs into aquatic environments (see the previous section). Thus was estimated the fraction of waste emitted in freshwater and the fraction emitted directly into the ocean for every region and per year. Starting the model in 1951, plastics were emitted into the modelled aquatic environment from every region. Polymers with a density higher than water were assumed to sink on the riverbed, lakebed, or seabed. Floating polymers circulating at the surface could directly reach the coastal ocean surface within the first year or remained in the freshwater system, likely stranded on river and lakeshores. The model also remobilised accumulated waste in river and lakeshores, adding onto inputs from the following year. Floating polymers in the coastal ocean surface followed the same dynamics as in the model presented in Lebreton et al. (2019[70]), with recirculation between the shoreline and the sea surface and transfer from coastal to offshore waters. Floating plastics accumulated in river and lake shore or on the ocean surface and shoreline were considered in contact with sunlight, and a fraction of their mass was degraded yearly to a sink term representing the mass of microplastics accumulated in freshwater and marine environments. For this report, the cycle was repeated for every year until 2019.

Figure A A.4. Mass balance budget model for plastic in global aquatic environments

Note: Mass inputs by modelled region, characterised by polymer types, are accumulated from 1951 to 2019 into the plastic fate model. Plastics with a density higher than water sink and accumulate in riverbed, lakebed and seabed. Floating plastics (density lower than water) are transported between different aquatic compartments and are allowed to degrade into microplastics over time from contact with sunlight. The region-specific parameter ‘i’ is the ratio between plastics remaining in freshwater and the plastics entering the marine environment (Table A A.13). The parameters ‘s’ and ‘r’ represent the fraction of stranding and release from the global shoreline. The parameter ‘t’ is the fraction of floating plastic circulating from the coastal to the offshore ocean. (s = 97%, r = 3%, t = 33%, (Lebreton, Egger and Slat, 2019[70])). Finally, ‘d’ is the mass fraction degrading into microplastics annually and varies with polymer types (Table A A.14).

Source: adapted from Lebreton, Egger and Slat (2019[70]) methodology.

Table A A.14. Parameters for fate of plastic in aquatic environments by polymer type

Polymer type	Micro	Floats	Degradation rate in % of mass per year
HDPE	No	Yes	0.6
LDPE, LLDPE	No	Yes	0.8
PP	No	Yes	0.0
PET	No	No	4.9
PS	No	No	0.1
PUR	No	No	3.0
PVC	No	No	-
ABS, ASA, SAN	No	No	-
Bioplastics	No	No	-
Elastomers (tyres)	No	No	-
Fibres	No	No	-
Road marking coatings	Yes	No	-
Marine coatings	Yes	No	-
Other	No	No	-

Source: Degradation rates are sourced from laboratory experiments (Gerritse et al., 2020[69]).

For this report, the model produced time series from 1951 to 2019, for each of the global regions, of inputs and accumulation of plastic waste into rivers, lakes, and the ocean. The model allows us to produce first-order of magnitude estimates of mass distribution in different compartments of the global aquatic environment.

This simplified model has some limitations, and care should be given in the interpretation of the results. The fate of plastics will vary significantly depending on the situation. These estimates should be seen as a whole, describing the regional quantity of plastic leakage from mismanaged and littered waste expressed by orders of magnitude of mass. Some assumptions were made in the design of the model, which does not always reflect reality. For instance, polymers such as PET, PVC or PUR were considered as sinking plastics, but by design, objects made with these polymers can float for a variable period of time (e.g. empty PET bottles with cap on, PVC buoys, or extended PUR foam). On the contrary, some floating plastics such as HDPE or LDPE may also sink rapidly (e.g. biobased plastic bags) in rivers while still considered movable in the model.

1.6. Modelling plastic leakage to terrestrial and aquatic environments (University of Leeds)

This section explains the methodology and parameters employed by the experts from the University of Leeds to estimate the fate of end-of-life plastics.

1.6.1. Waste management and leakage to environment

The end-of life fate, including leakage to the environment from the waste management system, were quantified using the Spatiotemporal Quantification of Plastic Pollution Origins and Transportation (SPOT) model (Cottom, Cook and Velis, 2020[11]). The SPOT model predominantly estimates material flows at Level 2 and 3 administrative boundary resolution, and therefore it had to be adapted to provide outputs at national (Level 0) which were aggregated to OECD regional level. Material flow analysis following Brunner and Rechberger (2016[71]) was the general methodological approach underpinning the distribution of plastic waste generation estimates provided by the ENV-Linkages model and used to describe its flow through the waste system as illustrated in the conceptual diagram (Figure A A.3). This hybrid model is described hereafter as the ‘ENVLinkages-SPOT plugin’.

Figure A A.5. ENV linkages-SPOT plugin model structure

Source: Cottom, Cook and Velis (2020[11]).

1.6.2. Model structure and basic assumptions

Data were processed using the SPOT model in three stages: 1) Municipal waste generation, composition and management data from 2007 to 2021 from four sources, Waste Wise Cities Tool (WaCT) (UN Habitat, n.d.[72]); Wasteaware Cities Benchmark Indicators (WABI) (Wilson et al., 2012[73]); United Nations Statistical Division (UNSD) (2021[74]); and What a Waste 2.0 (WAW2) (Kaza et al., 2018[9]), were cleaned and normalised according to a common denominator, resulting in approximately 500 data records; 2) Random forest machine learning used predictive variables to model data for the remaining 85 088 global municipalities that had no data; 3) Probabilistic material flow analysis used the interpolated data to allocate the flow of waste from the point of generation through managed, mismanaged and unmanaged process nodes.

The ENVLinkages-SPOT plugin uses the aggregated country level (Level 0) mass of rigid and flexible plastic waste estimated by the SPOT model, to determine transfer coefficients used to allocate material between process nodes. However, the SPOT does not present all data in the format required for the ENVLinkages-SPOT plugin to function, so some adjustments are needed as described in the following paragraphs.

1.6.3. Managed waste - baseline

Incineration data were not specifically reported in this version of the SPOT model due to the lack of spatial granularity in the source data which resulted in their aggregation with other types of recovery. Therefore, data obtained from Kaza et al. (2018[9]) were used in the ENVLinkages-SPOT plugin alongside further research which was used to verify or amend some data points as detailed in Table A A.15.

Table A A.15. Validation of incineration data

Country	Mass of MSW incinerated (t y-1)	Proportion of MSW incinerated (%)	Verification / addition	Source
Liechtenstein	8 268	25.4	add	Liechtenstein Institute for Strategic Development (2020[75])
Azerbaijan	400 000	9.6	add	(Islamic Development Bank, 2020[76])
Vietnam	1 602 764	5.4	add	(Tun et al., 2020[77])
Thailand	1 389 627	5.0	Verified	(Tun et al., 2020[77])
Ethiopia	350 000	2.5	add	(Cleere, 2020[78]; Mubeen and Buekens, 2019[79])
Lao PDR	32 637	2.0	add	(Tun et al., 2020[77])
India	1 916 250	0.7	add	(Central Pollution Control Board, 2021[80])
Myanmar	21 900	0.2	add	(JFE Engineering Corporation, 2017[81])

Source: Kaza et al. (2018[9]).

The proportion of waste collected for recycling by the informal sector was estimated by adapting the P2O model presented by Lau et al. (2020[27]). Additional data was reported by Cottom, Cook and Velis (2020[11]) for average productivity per waste picker, number of waste pickers per head of urban population, proportion of waste collected that is plastic (Table A A.16). It as was also assumed that workers operate for 235 days on average accounting for sickness, vacation and other downtime.

Table A A.16. Data used to model the activities of the informal recycling sector

Income group (World Bank)	Proportion of urban population that is an informal waste worker (%)	Productivity per waste picker (kg d-1)	Proportion of waste collected by informal recycling sector that is plastic (%)
High income	0.01	37.0	5
Upper middle-income	0.26	37.0	28
Lower middle income	0.19	37.0	35
Low income	0.14	37.0	35

Source: Cottom, Cook and Velis (2020[11]).

1.6.4. Dumping mismanaged waste in water

Data to quantify the deliberate dumping of waste into water by waste generators are scarce. Here we present, for the first time, a review of census data that indicate the mass deposited directly into water by householders in the absence of formal waste collection services (Table A A.17). Acknowledging the uncertainty in the data and the high variability in time and across countries, we have taken a conservative approach and approximated the issue by using the mean of the country level median proportion treated in this way: 4.8% of uncollected waste.

Table A A.17. Deliberate dumping into water

Country	Proportion of population engaged in behaviour (median % of uncollected waste)	Source
Malawi	1.0	National Statistical Office (2020[82])
Guatemala	1.8	Guatemala, Instituto Nacional de Estadística (2018[83])
Indonesia	7.6	Sub Direktorat Statistik Lingkungan Hidup (2014[84])
Fiji	0.5	Fiji Bureau of Statistics (2018[85])
Brazil	0.4	Instituto Brasileiro de Geografia e Estatística (2010[86])
Bolivia	15.6	Instituto Nacional de Estadistica (2012[87])
Samoa	0.4	Samoa Bureau of statistics (2019[88])
Ethiopia	10.9	Population Census Commission (2007[89])

1.6.5. Plastic debris emissions to aquatic environments

The leakage transfer from terrestrial to aquatic environments was estimated using transfer ratios suggested by Lau et al. (2020[27]) and detailed in Table A A.18.The GWPv4 (2015) (United Nations, 2019[90]) UNAdj population density map (CIESIN, 2018[91]) was used to estimate the proportion of rural and urban inhabitants in line with Dijkstra and Poelman (2014[92]) using grid cells >300 population and  >5 000 inhabitants in contiguous cells. The urban and rural attribution was mapped onto the HydroSHEDS 30 arc river and coastline dataset. Population data for countries above 60°N latitude were approximated using ratios for the nearest similar countries which were below 60°N.

Waste transfer from the terrestrial to aquatic environment was estimated using transfer ratios suggested by Lau et al. (2020[27]) and detailed in Table A A.18. The GWPv4 (2015) (United Nations, 2019[90]) UNAdj population density map (CIESIN, 2018[91]) was used to estimate the proportion of rural and urban inhabitants using definition from was estimated using Dijkstra and Poelman (2014[92])); that a grid cell has >300 population and >5 000 inhabitants in contiguous cells. The urban and rural attribution was mapped onto the HydroSHEDS 30 arc river and coastline dataset. Population data for countries above 60°N latitude were approximated using ratios for nearest similar countries which were below 60°N.

Table A A.18. Plastic waste transfer rate from terrestrial to aquatic environment (% wt. y^-1)

Flexibility	Distance population to aquatic environment	Proportion of plastic waste in terrestrial environment that transfers to the aquatic environment (% wt.)
Rigid	< 1 km	10
	>1 km	3
Flexible	<1 km	35
	>1 km	8

Source: Lau et al. (2020[27]).

1.7. Plastic leakage to terrestrial and aquatic environments (OECD Global Plastics Outlook Database)

The OECD Global Plastics Outlook Database includes three estimates for leakage to the environment (sum of terrestrial and aquatic environments) and three estimates for leakage to aquatic environments: central, high and low estimates. They are calculated by combining the estimates based on the different methodologies outlined in the previous sections, namely estimates provided by: the DTU, the University of Leeds and Laurent Lebreton.

More specifically, the central estimate of the leakage to the environment corresponds to the average of the leakages to the environment reported by DTU and the University of Leeds. The high and low estimates refer to the maximum and minimum of these values correspondingly.

The leakage to aquatic environments, is calculated by using the central estimate for the leakage to the environment as input in the model of Laurent Lebreton. The central and high estimates for leakage to water come from Laurent Lebreton. The low estimate represents the number reported by the University of Leeds.

The mobility of plastics in aquatic environments uses the central estimate for the leakage to aquatic environments as input for the methodology described earlier (Annex ‘Modelling plastic leakage to aquatic environments – Laurent Lebreton’).

1.8. Modelling particulate matter emissions to air from tyre and brake wear (Norwegian Institute for Air Research)

This section explains the methodology and parameters employed by Nicolaos Evangeliou from the Norwegian Institute for Air Research (NILU) to estimate the emission of airborne road-traffic-related microplastics and their contribution to particulate matter pollution.

1.8.1. Calculation of emissions from tyre and brake wear

Tyre and brake wear particles (TWPs and BWPs) are calculated using the GAINS (Greenhouse gas – Air pollution Interactions and Synergies) model (Amann et al., 2011[93]). GAINS is an integrated assessment model where emissions of air pollutants and greenhouse gases are estimated for nearly two hundred regions globally considering key economic activities, environmental regulation policies and region-specific emission factors. For emissions of particulate matter (PM), GAINS provides PM distinguishing PM₁, PM_2.5, PM₁₀, total PM, as well as carbonaceous particles (BC, OC) that derive from combustion processes, as described in Klimont et al. (2017[94]).

Emissions of non-exhaust PM in GAINS include TWPs, BWPs, as well as road abrasion. The calculation of these emissions is based on region-specific data and estimates of distance driven (km/vehicle-type/year) and vehicle-type specific emission rates (mg/km). The types of vehicles considered include motorcycles, cars, light-duty vehicles, buses, and heavy-duty vehicles. The estimates of distance driven for 2015 are derived using data on fuel use in road transport from the International Energy Agency’s Word Energy Outlook (IEA, 2011[95]), supported by national data on vehicle numbers and assumptions of per-vehicle mileage travelled. Considering vehicle-type specific emission rates and use, allows for better reflection of significant regional differences in fleet structure, e.g. large number of motorcycles in South and South-East Asia and lower car ownership numbers in parts of the developing world. GAINS emissions are estimated globally at the grid level (0.5°×0.5°) using road network data, assumptions about road-type vehicle density, and population data.

The vehicle-type specific TWP and BWP emission factors used in GAINS draw on a review of several measurement papers (Klimont et al., 2002[96]) that were recently updated (Klimont et al., 2017[94]) using primarily van der Gon et al. (2013[97]), EEA (2013[98]) and Harrison et al. (2012[99]). There are large uncertainties in emission factors including concerning the PM size distribution. GAINS provides total suspended particulates (TSP), and then assumes that PM₁₀ from TWPs represent about 10% of TSP, and PM_2.5 about 1% of total TWPs, whereas PM₁₀ from BWPs is about 80% of TSP and PM_2.5 is 40–50% of total BWPs (Klimont et al., 2002[96]).

1.8.2. Atmospheric transport modelling

Emissions of PM₁₀ calculated with the GAINS model are used as input in the FLEXPART (FLEXible PARTicle) atmospheric transport model version 10.4 (Pisso et al., 2019[100]). Atmospheric dispersion of particulate matter, including both transport and deposition of particles, were simulated for the reference year 2014. The FLEXPART model was run in forward mode from 2014. Atmospheric processes affecting particle transport in clouds (e.g. boundary layer turbulent mixing and convection processes) are parameterised in the model (Forster, Stohl and Seibert, 2007[101]). The model was driven by 3-hourly 1°×1° operational analyses from the European Centre for Medium Range Weather Forecast (ECMWF), the spatial output resolution of concentration and deposition fields was set to 0.5°×0.5° in a global domain with a daily temporal resolution. In FLEXPART the dispersion of road microplastics is modelled assuming a spherical shape of particles (Pisso et al., 2019[100]).

The simulations also accounted for below-cloud scavenging and dry deposition, assuming a particle density for TWPs of 1 234 kg/m³, which is in the middle of the densities of 945 kg/m³ for natural rubber and 1 522 kg/m³ for synthetic rubber (Walker, 2019[102]; Federal Highway Administration Research and Technology, 2019[103]). This density is within the reported range for microplastics (940-2 400 kg/m³) (Unice et al., 2019[104]). For BWPs a higher density was assumed (2 000 kg/m³) considering that BWP may also contain metals (Grigoratos and Martini, 2014[105]). Plastics are generally hydrophobic and should therefore be rather inefficient cloud condensation nuclei (CCN) (Di Mundo, Petrella and Notarnicola, 2008[106]; Ganguly and Ariya, 2019[107]). However, coatings may make the particles more hydrophilic with time in the atmosphere (Bond et al., 2013[108]). The efficiency of aerosols to serve as ice nuclei (IN) is also not well known. Based on Evangeliou et al. (2020[109]), it is more realistic to use intermediate scavenging coefficients for CCN/IN in the model.

The emissions of TWPs and BWPs were extrapolated from 2014 to 2019, using the road passenger data from the IEA’s World Energy Outlook (IEA, 2018[110]) for the 15 geographical regions of the OECD ENV‑Linkages model. Year 2014 was taken as base year and the ratio to year 2014 was calculated for each year between 2015 and 2019 and for each of the 15 regions (from now on referred as “regional scaling factor”). This regional scaling factor could be negative, if the road passenger data decreased as compared to 2015, or positive, if an increase is shown.

The FLEXPART model was run with the 2014 emissions for the 15 ENV-Linkages regions; thus creating 15 different model simulations, each representing the dispersion from the respective region. Then, the regional scaling factor was used to scale modelled dispersion that results from each individual regional emission for each year between 2015 and 2019. Finally, the 15 regional annually-scaled modelled dispersions were used to calculate global TWP and BWP estimates.

1.9. Modelling greenhouse gas emissions from primary plastics in ENV-Linkages

This section explains the methodology and parameters employed to estimate the contribution of the lifecycle of plastics to GHG emissions at the global level, based on the OECD ENV-Linkages model.

In the GTAP database (Aguiar et al., 2019[4]), plastics production occurs in two sectors: Chemicals and Rubber and plastics products, along with other products. In ENV-Linkages, the plastic products sectors has been split into primary and secondary plastics production (see Section 1.2 in this Annex). The difficulty in estimating plastic-related greenhouse gas emissions is that in ENV-Linkages there is no direct link between emissions and emissions attributable to plastics. The plastics producing sectors use inputs from the electricity generation sector, fossil fuel extraction sectors and other sectors of the economy. However, these plastics producing sectors also produce other goods, so that not all emissions can be attributed to plastics. Furthermore, emissions outside these sectors can also be attributed to plastics. Furthermore, emissions from the extraction of raw materials and from plastic waste management, are also mixed with emissions not attributable to plastics.

Therefore, to approximate the global lifecycle emissions from plastics, an emission factor-based approach is retained:

${Emi}_{g,t}^{plastics}= {\sum }_p\left({\lambda }_{g,p,t}^{prod}+{\lambda }_{g,p,t}^{conv}\right)C_{p,t}+{\sum }_f{\lambda }_{g,f,t}^{eol}{W}_{f,t}$

where ${Emi}_{g,t}^{plastics}$ are emissions of greenhouse gas $g$ (comprising CO₂, CH₄ and N₂O, measured in CO₂‑equivalents)4 from the plastics lifecycle at time $t$, ${\lambda }_{g,p,t}^{prod}$ and ${\lambda }_{g,p,t}^{conv}$ are respectively the emission factors per tonne of plastic product for production and conversion of plastic for polymer $p$ that are applied to the level of primary plastics consumption $C_{p,t}$ estimated by the model (by convention, emissions from the production of secondary plastics are allocated to waste management emissions). Finally, ${\lambda }_{g,f,t}^{eol}$ is the emission factor for a specific end-of-life fate $f$ (incineration, sanitary landfilling and recycling only are considered, due to data availability), applied to the amount of plastic waste generated $W_{j,t}$.

The literature provides estimates of emission factors for year 2015 (Zheng and Suh, 2019[111])5 that are used to calibrate the emissions for 2015. These emission factors comprise emissions from the whole value-chain of plastics production, and have no reason to be constant over time due to structural changes in the production process. As a consequence, the GHG intensity of plastics production and conversion is updated to 2019 based on the structural change observed in the model. An index based on the global average scope 2 emissions (direct emissions plus emissions from electricity demand) of the sectors related to plastics production and conversion (chemicals, primary rubber and plastics products, oil extraction, gas extraction and petroleum and coal products) represents the evolution over time of production and conversion emissions. Another index based on scope 2 emissions of the secondary plastics sector represents the evolution over time of recycling emission intensity, while emissions factors are constant for incineration and landfilling (Figure A A.6).

Figure A A.6. Greenhouse gas emission factors for plastics lifecycle in ENV-Linkages in 2015

Note: Emissions from recycling and incineration are direct emissions from their respective industrial processes. Avoided emissions (emissions from electricity generation that are replaced by waste incineration and emissions from primary plastics production that are replaced by secondary plastics) are not included in the emission factors depicted here, because they are included directly in the ENV-Linkages model (changes in the input structure of the electricity generation sector and changes in the balance between primary and secondary plastics, both being driven by the change in the relative prices of products, are endogenously determined by the model).

Source: Adapted from Zheng and Suh (2019[111]).

References

[4] Aguiar, A. et al. (2019), “The GTAP Data Base: Version 10”, Journal of Global Economic Analysis, Vol. 4/1, pp. 1-27, https://doi.org/10.21642/jgea.040101af.

[93] Amann, M. et al. (2011), “Cost-effective control of air quality and greenhouse gases in Europe: Modeling and policy applications”, Environ. Model. Softw., Vol. 26, pp. 1489–1501.

[30] Antonopoulos, I., G. Faraca and D. Tonini (2021), “Recycling of post-consumer plastic packaging waste in the EU: Recovery rates, material flows, and barriers”, Waste Management, Vol. 126, pp. 694-705, https://doi.org/10.1016/j.wasman.2021.04.002.

[19] Australian Government, D. (2020), 2018-19 Australian plastics recycling survey - national report, https://www.awe.gov.au/environment/protection/waste/publications/australian-plastics-recycling-survey-report-2018-19 (accessed on 28 October 2021).

[108] Bond, T. et al. (2013), “Bounding the role of black carbon in the climate system: A scientific assessment”, J. Geophys. Res. Atmos., Vol. 118, pp. 5380–5552.

[65] Borrelle, S. et al. (2020), “Predicted growth in plastic waste exceeds efforts to mitigate plastic pollution”, Science, Vol. 369/6510, pp. 1515-1518, https://doi.org/10.1126/science.aba3656.

[53] Boucher, F. and D. Friot (2017), Primary Microplastics in the Oceans: A Global Evaluation of Sources..

[71] Brunner, P. and H. Rechberger (2016), Handbook of Material Flow Analysis: For Environmental, Resource, and Waste Engineers, CRC Press, https://doi.org/doi.org/10.1201/9781315313450.

[80] Central Pollution Control Board (2021), Report of Waste to Energy Plants in Delhi by CPCB in OA No. 640 of 2018 (Earlier O.A. No. 22 of 2013(THC), Sukhdev Vihar Residents Welfare Association Vs State of NCT of Delhi, https://greentribunal.gov.in/report-waste-energy-plants-delhi-cpcb-oa-no-640-2018-earlier-oa-no-22-2013thc-sukhdev-vihar.

[21] Central Pollution Control Board (CPCB) (2019), Annual Report for the year 2018-2019 on Implementation of Plastic Waste Management Rules, Ministry of Environment, Forest and Climate Change, Govt of India.

[1] Chateau, J., R. Dellink and E. Lanzi (2014), “An Overview of the OECD ENV-Linkages Model: Version 3”, OECD Environment Working Papers, No. 65, OECD Publishing, Paris, http://dx.doi.org/10.1787/5jz2qck2b2vd-en.

[2] Chateau, J., C. Rebolledo and R. Dellink (2011), “An Economic Projection to 2050: The OECD “ENV-Linkages” Model Baseline”, OECD Environment Working Papers, No. 41, OECD Publishing, Paris, http://dx.doi.org/10.1787/5kg0ndkjvfhf-en.

[26] Chruszcz, A. and S. Reeve (2018), Composition of plastic waste: Results of a waste compositional analysis of plastics at MRFs and PRFs, WRAP.

[91] CIESIN (2018), Gridded population of the world, version 4 (GPWv4): population count adjusted to Match 2015 revision of UN WPP country totals, revision 11., Center for International Earth Science Information Network - Columbia University, https://sedac.ciesin.columbia.edu/data/set/gpw-v4-population-count-adjusted-to-2015-unwpp-country-totals-rev11.

[78] Cleere, R. (2020), “The New Reppie Incinerator at Koshe Landfill in Addis Ababa, Ethiopia.”, [Online data set] Environmental Justice Atlas, https://ejatlas.org/conflict/the-new-reppie-incinerator-at-koshe-landfill-in-addis-ababa-ethiopia-leaves-the-wastepickers-without-livelihood.

[11] Cottom, J., E. Cook and C. Velis (2020), “Spatio-temporal quantification of plastic pollution origins and transportation (SPOT)”, University of Leeds, UK, https://plasticpollution.leeds.ac.uk/toolkits/spot/.

[54] De Falco, F. et al. (2020), “Microfiber Release to Water, Via Laundering, and to Air, via Everyday Use: A Comparison between Polyester Clothing with Differing Textile Parameters”, Environmental Science & Technology, doi: 10.1021/acs.est.9b06892, pp. 3288-3296, https://doi.org/10.1021/acs.est.9b06892.

[43] De Falco, F. et al. (2019), “The contribution of washing processes of synthetic clothes to microplastic pollution”, Scientific Reports, Vol. 9, p. 6633, https://doi.org/10.1038/s41598-019-43023-x.

[106] Di Mundo, R., A. Petrella and M. Notarnicola (2008), “Surface and bulk hydrophobic cement composites by tyre rubber addition”, Constr. Build. Mater., Vol. 172, pp. 176–184.

[92] Dijkstra, L. and H. Poelman (2014), A harmonised definition of cities and rural areas: the new degree of urbanisation, https://ec.europa.eu/regional_policy/sources/docgener/work/2014_01_new_urban.pdf.

[51] ECHA (2020), Committee for Risk Assessment (RAC) Committee for Socio-economic Analysis (SEAC). Opinion on an Annex XV dossier proposing restrictions on intentionally added microplastics.

[98] EMEP/EEA (2013), Air pollutant emission inventory guidebook 2013: Technical guidance to prepare national emission inventories, https://doi.org/10.2800/92722.

[15] Environment and Climate Change Canada (2019), Economic Study of the Canadian plastic industry, markets and waste, Environment and Climate Change Canada.

[38] Eunomia (2018), “Investigating options for reducing releases in the aquatic environment of microplastics emitted by (but not intentionally added in) products”, Report for DG Env EC, Vol. Vol. 62, N/February, pp. 1596-1605, https://doi.org/10.1002/lsm.22016.

[64] Eurostat (2020), Statistics | Sewage sludge production and disposal., https://ec.europa.eu/eurostat/databrowser/view/ENV_WW_SPD/default/table. (accessed on 28 January 2021).

[45] Eurostat (2018), Average loads for total road freight transport, 2018 (tonnes), https://ec.europa.eu/eurostat/statistics-explained/index.php?title=File:Average_loads_for_total_RFT,_2018_(tonnes).png. (accessed on 21 May 2021).

[39] Eurostat (n.d.), Sold production, exports and imports by PRODCOM list (NACE Rev. 2) - annual data, https://ec.europa.eu/eurostat/web/prodcom/data/database.

[109] Evangeliou, N. et al. (2020), “Atmospheric transport is a major pathway of microplastics to remote regions”, Nat Commun, Vol. 11, p. 3381, https://doi.org/10.1038/s41467-020-17201-9.

[17] FCH (2021), “NEW PLASTICS ECONOMY”, https://fch.cl/en/initiative/new-plastics-economy.

[103] Federal Highway Administration Research and Technology (2019), User Guidelines for Waste and Byproduct Materials in Pavement Construction.

[85] Fiji Bureau of Statistics (2018), 2017 Fiji population and housing census., https://www.statsfiji.gov.fj/component/advlisting/?view=download&format=raw&fileId=5970.

[101] Forster, C., A. Stohl and P. Seibert (2007), “Parameterization of convective transport in a Lagrangian particle dispersion model and its evaluation”, J. Appl. Meteorol. Climatol., Vol. 46, pp. 403–422.

[107] Ganguly, M. and P. Ariya (2019), “Ice Nucleation of Model Nanoplastics and Microplastics: A Novel Synthetic Protocol and the Influence of Particle Capping at Diverse Atmospheric Environments”, ACS Earth Sp. Chem, Vol. 3, pp. 1729–1739.

[69] Gerritse, J. et al. (2020), “Fragmentation of plastic objects in a laboratory seawater microcosm”, Scientific Reports, Vol. 10/1, p. 10945, https://doi.org/10.1038/s41598-020-67927-1.

[8] Geyer, R., J. Jambeck and K. Law (2017), “Production, use, and fate of all plastics ever made”, Science Advances, Vol. 3/7, p. e1700782, https://doi.org/10.1126/sciadv.1700782.

[24] Government of Australia (2021), Australian plastics flows and fates 2019-2020.

[105] Grigoratos, T. and G. Martini (2014), Non-exhaust traffic related emissions. Brake and tyre wear PM, https://doi.org/10.2790/21481.

[83] Guatemala, Instituto Nacional de Estadística (2018), Características generales del hogar. Censo 2018: Cuadro B6.1 - Hogares por forma principal de eliminación de la basura, según departamento. [Online data set], https://www.censopoblacion.gt/explorador.

[48] Hallal, A. et al. (2013), “Overview of Composite Materials and their Automotive Applications”, in Advanced Composite Materials for Automotive Applications, John Wiley & Sons, Ltd, https://doi.org/10.1002/9781118535288.ch1.

[99] Harrison, R. et al. (2012), “Estimation of the contributions of brake dust, tire wear, and resuspension to nonexhaust traffic particles derived from atmospheric measurements”, Environ. Sci. Technol., Vol. 46, pp. 6523–6529.

[25] Hestin, M., T. Faninger and L. Milios (2015), Increased EU Plastics Recycling Targets: Environment, Economic and Social Impact Assessment, https://743c8380-22c6-4457-9895-11872f2a708a.filesusr.com/ugd/0af79c_d3c616e926e24896a8b82b833332242e.pdf.

[110] IEA (2018), World Energy Outlook.

[95] IEA (2011), World Energy Outlook.

[6] IMF (2020), World Economic Outlook, October 2020: A Long and Difficult Ascent, International Monetary Fund, Washington, D.C., https://www.imf.org/en/Publications/WEO/Issues/2020/09/30/world-economic-outlook-october-2020 (accessed on 22 January 2021).

[86] Instituto Brasileiro de Geografia e Estatística (2010), Demographic Census: Table 1395 - Permanent private households, by household situation and existence of bathroom or toilet and number of toilets for the exclusive use of the household, according to the type of household, the form of water supply, the desti, https://sidra.ibge.gov.br/tabela/1395.

[87] Instituto Nacional de Estadistica (2012), Disposal of garbage in the house, according to province and municipality, 2012 census [Online data set], https://www.ine.gob.bo/index.php/estadisticas-sociales/vivienda-y-servicios-basicos/censos-vivienda/.

[112] IPCC (1995), Climate Change 1995: A report of the Intergovernmental Panel on Climate Change - IPCC Second Assessment.

[76] Islamic Development Bank (2020), Waste to Energy: Averting environnmental damage in Azerbaijan., https://www.isdb.org/sites/default/files/media/documents/2020-06/Success_Lflt_Azerbaijan_EN.pdf.

[12] Jambeck, J. et al. (2015), “Plastic waste inputs from land into the ocean”, Science, Vol. 347/6223, pp. 768-771, https://doi.org/10.1126/science.1260352.

[81] JFE Engineering Corporation (2017), Opening Ceremony for Myanmar’s First Waste to Energy Plant, https://www.jfe-eng.co.jp/en/news/2017/20170410.html.

[59] JMP (2020), Wash Data, https://washdata.org/data/household#!/table?geo0=region&geo1=sdg. (accessed on 29 January 2021).

[58] Kalbar, P., I. Muñoz and M. Birkved (2017), “WW LCI v2: A second-generation life cycle inventory model for chemicals discharged to wastewater systems.”, Sci Total Environ., https://doi.org/10.1016/j.scitotenv.2017.10.051.

[55] Kawecki, D. and B. Nowack (2020), “A proxy-based approach to predict spatially resolved emissions of macro- and microplastic to the environment”, Science of The Total Environment, Vol. 748, p. 141137, https://doi.org/10.1016/j.scitotenv.2020.141137.

[9] Kaza, S. et al. (2018), What a Waste 2.0: A Global Snapshot of Solid Waste Management to 2050, The World Bank, https://doi.org/10.1596/978-1-4648-1329-0.

[96] Klimont, Z. et al. (2002), Modelling Particulate Emissions in Europe. A Framework to Estimate Reduction Potential and Control Costs.

[94] Klimont, Z. et al. (2017), “Global anthropogenic emissions of particulate matter including black carbon”, Atmos. Chem. Phys., Vol. 17, pp. 8681–8723.

[56] Lassen, C. et al. (2016), Microplastics Occurrence, effects and sources of releases to the environment in Denmark, Danish Environmental Protection Agency, Copenhagen.

[27] Lau, W. et al. (2020), “Evaluating scenarios toward zero plastic pollution”, Science, Vol. 369/6510, pp. 1455-1461, https://doi.org/10.1126/science.aba9475.

[67] Lebreton, L. and A. Andrady (2019), “Future scenarios of global plastic waste generation and disposal”, Palgrave Communications, Vol. 5/1, p. 6, https://doi.org/10.1057/s41599-018-0212-7.

[70] Lebreton, L., M. Egger and B. Slat (2019), “A global mass budget for positively buoyant macroplastic debris in the ocean”, Scientific Reports, Vol. 9/1, p. 12922, https://doi.org/10.1038/s41598-019-49413-5.

[68] Lebreton, L. et al. (2018), “Evidence that the Great Pacific Garbage Patch is rapidly accumulating plastic”, Scientific Reports, Vol. 8/1, https://doi.org/10.1038/s41598-018-22939-w.

[75] Liechtenstein Institute for Strategic Development (2020), Circular economy strategy for Liechtenstein., https://www.alpine-space.org/projects/greencycle/deliverables/t2/lisd---circular-economy-strategy-for-liechtenstein-vol1-10-03-2020-1.pdf.

[49] Løkkegaard, H., B. Malmgren-Hansen and N. Nilsson (2018), Mass balance of rubber granulate lost from artificial turf fields, focusing on discharge to the aquatic environment.

[60] Magnusson, K. et al. (2016), Swedish Sources and Pathways for Microplastics to the Marine Environment..

[66] Meijer, L. et al. (2021), “More than 1000 rivers account for 80% of global riverine plastic emissions into the ocean”, Science Advances, Vol. 7/18, https://doi.org/10.1126/sciadv.aaz5803.

[61] Michielssen, M. et al. (2016), “Fate of microplastics and other small anthropogenic litter (SAL) in wastewater treatment plants depends on unit processes employed”, Environmental Science: Water Research and Technology, Vol. 2/6, pp. 1064-1073, https://doi.org/10.1039/c6ew00207b.

[20] Ministry of Commerce (2019), The China Recycling Industry Development Report (2013-2018).

[79] Mubeen, I. and A. Buekens (2019), “Chapter 14 - Energy From Waste: Future Prospects Toward Sustainable Development”, in Kumar, S., R. Kumar and A. Pandey (eds.), Current Developments in Biotechnology and Bioengineering, Elsevier, https://doi.org/10.1016/B978-0-444-64083-3.00014-2.

[82] National Statistical Office (2020), 2018 Malawi population and housing census: water and sanitation report Zomba, http://www.nsomalawi.mw/images/stories/data_on_line/demography/census_2018/Thematic_Reports/Water%20and%20Sanitation%20Report.pdf.

[62] Nizzetto, L., M. Futter and S. Langaas (2016), Are Agricultural Soils Dumps for Microplastics of Urban Origin?, American Chemical Society, https://doi.org/10.1021/acs.est.6b04140.

[47] OECD (2021), Policies to Reduce Microplastics Pollution in Water: Focus on Textiles and Tyres, OECD Publishing, Paris, https://dx.doi.org/10.1787/7ec7e5ef-en.

[5] OECD (2020), OECD Economic Outlook, Volume 2020 Issue 2, OECD Publishing, Paris, https://dx.doi.org/10.1787/39a88ab1-en.

[3] OECD (2019), Global Material Resources Outlook to 2060: Economic Drivers and Environmental Consequences, https://doi.org/10.1787/9789264307452-en.

[57] OECD (2017), Environment Database - Wastewater treatment (% population connected) (dataset), http://stats.oecd.org/index.aspx?DatasetCode=WATER_TREAT. (accessed on 29 January 2021).

[36] OVAM (2018), Huishoudelijk afval en gelikkaardig bedrijfsafval., https://www.ovam.be/inventarisatie-huishoudelijke-afvalstoffen.

[44] Pirc, U. et al. (2016), “Emissions of microplastic fibers from microfiber fleece during domestic washing”, Environ Sci Pollut Res, Vol. 23, pp. 22206–22211, https://doi.org/10.1007/s11356-016-7703-0.

[100] Pisso, I. et al. (2019), “The Lagrangian particle dispersion model FLEXPART version 10.4”, Geosci. Model Dev., Vol. 12, pp. 4955–4997.

[18] Plastic Waste Management Institute (2019), An Introduction to Plastic Recycling.

[23] Plastics Europe (2020), “Plastics – the Facts 2020”.

[42] Plastics Europe (2017), Plastics: the Facts (2017) An analysis of European plastics production, demand and waste data, Plastics Europe.

[31] Plastics Recyclers Europe (2020), Report on Plastics Recycling Statistics, http://743c8380-22c6-4457-9895-11872f2a708a.filesusr.com/ugd/dda42a_2544b63cfb5847e39034fadafbac71bf.pdf.

[89] Population Census Commission (2007), 2007 Population and Housing Census of Ethiopia., https://microdata.worldbank.org/index.php/catalog/2747/download/39216.

[29] Recoup (2019), Recyclability by Design, https://www.bpf.co.uk/design/recyclability-by-design.

[52] ResearchNester (2021), Artificial Turf: Market Insights, Demand & Growth Forecast 2027., https://www.researchnester.com/reports/artificial-turf-market/995. (accessed on 28 January 2021).

[37] Resource Futures (2019), Composition analysis of litter waste in Wales.

[63] Rolsky, C. et al. (2020), “Municipal sewage sludge as a source of microplastics in the environment.”, Curr. Opin. Environ. Sci. Heal..

[32] Roosen, M. et al. (2020), “Detailed Analysis of the Composition of Selected Plastic Packaging Waste Products and Its Implications for Mechanical and Thermochemical Recycling”, Environmental Science & Technology, doi: 10.1021/acs.est.0c03371, pp. 13282-13293, https://doi.org/10.1021/acs.est.0c03371.

[7] Ryberg, M. et al. (2019), “Global environmental losses of plastics across their value chains”, Resources, Conservation and Recycling, Vol. 151, p. 104459, https://doi.org/10.1016/j.resconrec.2019.104459.

[88] Samoa Bureau of statistics (2019), Samoa’s Experimental Solid Waste Accounts FY2013-14 to FY2015-16, https://www.sbs.gov.ws/digi/Samoa's%20Experimental%20Solid%20Waste%20Arrounts_2013-2014%20to%202015-2016.pdf.

[16] SEMARNAT (2020), Diagnostico basico para la gestion integral de los residuos.

[46] Sommer, F. et al. (2018), “Tire Abrasion as a Major Source of Microplastics in the Environment”, Aerosol and Air Quality Research, Vol. 18/8, pp. 2014-2028, https://doi.org/10.4209/aaqr.2018.03.0099.

[10] Stadler, K. et al. (2018), “EXIOBASE 3: Developing a Time Series of Detailed Environmentally Extended Multi-Regional Input-Output Tables”, Journal of Industrial Ecology, Vol. 22/3, pp. 502-515, https://doi.org/10.1111/jiec.12715.

[84] Sub Direktorat Statistik Lingkungan Hidup (2014), Indikator Perilaku Peduli Lingkungan Hidup (2014 Environmental Care Behavior Indicators, https://www.bps.go.id/publication/2015/12/23/2cdc2ef08c706d6f205c69fc/indikator-perilaku-peduli-lingkungan-hidup-2014.html.

[50] Swedish EPA (2019), Microplastics in the Environment 2019, http://www.naturvardsverket.se/Om-Naturvardsverket/Publikationer/ISBN/6900/978-91-620-6957-5/.

[28] SystemIQ and the Pew Charitable Trust (2020), Breaking the Plastic Wave: A Comprehensive Assessment of Pathways Towards Stopping Ocean Plastic Pollution, https://www.systemiq.earth/breakingtheplasticwave/.

[34] Thompson, P., P. Willis and N. Morley (2012), A review of commercial textile fibre recycling technologies, Waste and Resources Action Programme (WRAP), UK, https://refashion.fr/eco-design/sites/default/files/fichiers/A%20review%20of%20commercial%20textile%20fibre%20recycling%20technologies.pdf.

[77] Tun, M. et al. (2020), “Renewable Waste-to-Energy in Southeast Asia: Status, Challenges, Opportunities, and Selection of Waste-to-Energy Technologies”, Applied Science, Vol. 10/20, p. 7312, https://doi.org/10.3390/app10207312.

[72] UN Habitat (n.d.), Cities’ Waste Data, https://unhabitat.org/waste-wise-cities-waste-data (accessed on 20 September 2021).

[104] Unice, K. et al. (2019), “Characterizing export of land-based microplastics to the estuary - Part I: Application of integrated geospatial microplastic transport models to assess tire and road wear particles in the Seine watershed”, Sci. Total Environ., Vol. 646, pp. 1639–1649.

[22] UNIDO (2020), Recycling of plastics in Indian perspective, UNIDO Office, VIC, Vienna, https://www.unido.org/sites/default/files/files/2018-11/Plenary%202%20-%20Plastics%20-%20Mohanty.pdf.

[90] United Nations (2019), World urbanization prospects: The 2018 revision., https://population.un.org/wup/Publications/Files/WUP2018-Report.pdf.

[35] United Nations Statistics Division (2020), UN Comtrade, https://comtrade.un.org (accessed on 21 September 2020).

[13] United States Envrionmental Protection Agency (EPA) (2020), “Advancing Sustainable Materials Management: 2018 Tables and Figures”, https://www.epa.gov/sites/default/files/2021-01/documents/2018_tables_and_figures_dec_2020_fnl_508.pdf.

[14] United States Envrionmental Protection Agency (EPA) (2020), Plastics: Material-Specific Data, https://www.epa.gov/facts-and-figures-about-materials-waste-and-recycling/plastics-material-specific-data.

[74] UNSD (2021), UNSD Environmental Indicators: Waste In Environment Statistics, https://unstats.un.org/unsd/envstats/qindicators.cshtml.

[97] van der Gon, H. et al. (2013), “The policy relevance of wear emissions from road transport, now and in the future--an international workshop report and consensus statement”, Air Waste Manag Assoc., Vol. 63/2, pp. 136-49, https://doi.org/10.1080/10962247.2012.741055.

[33] VinylPlus (2019), PVC Recycling in Action, https://vinylplus.eu/uploads/images/Leaflets/Recovinyl_21x21_04-05_web.pdf.

[40] Viool, V. et al. (2018), Study to support impact assessment for options to reduce the level of ALDFG.

[102] Walker, R. (2019), The mass of 300 different ‘dry’ materials.

[73] Wilson, D. et al. (2012), “Comparative analysis of solid waste management in 20 cities”, Waste Management & Research, doi: 10.1177/0734242X12437569, pp. 237-254, https://doi.org/10.1177/0734242X12437569.

[41] World Bank (2020), New World Bank country classifications by income level: 2020-2021., https://blogs.worldbank.org/opendata/new-world-bank-country-classifications-income-level-2020-2021 (accessed on 28 January 2021).

[111] Zheng, J. and S. Suh (2019), “Strategies to reduce the global carbon footprint of plastics”, Nature Climate Change, Vol. 9/5, pp. 374-378, https://doi.org/10.1038/s41558-019-0459-z.

Notes

← 1.  https://doi.org/10.1787/c0821f81-en.

← 2.  To simulate lifespan distributions for plastic applications from historical years, an exogenous component of waste generated by earlier produced commodities is added in the first years.

← 3.  In particular, ECHA (2020[51]) reports that the share of end-of-life tyre-derived granules would represent 78% on the infill, whereas EPDM and TPE would account for 18%, and cork 4%, by 2028. As artificial turf is only made up of the rubber part of tyres (EuRIC MTR 2020), 96% of all infill is assumed to be microplastics.

← 4.  The nominal emissions of CH₄ and N₂O are converted to CO₂-equivalents using the 100-year GWP from 2^nd assessment report (IPCC, 1995[112]).

← 5.  We would like to thank the authors for providing greenhouse-gas specific emission factors that are not available directly in their paper.