The United Nations have agreed to negotiate a legally binding instrument to eliminate plastic pollution which includes provisions to reduce environmental emissions of plastic through improved waste management. However, there is a paucity of scientific evidence to prioritize the actions which will have the greatest impact on plastic pollution mitigation and many of the specific emission and transmission pathways are highly challenging to observe and measure. To this, we apply an expert elicitation study to provide a systematic evidence base on which parts of the solid waste management (SWM) system have the highest potential for plastic emission as well as the effectiveness of specific interventions to reduce them. We found that SWM processes closer to the waste generator, such as the waste storage and collection stage, have the greatest potential for plastic emission, with the most effective mitigations often simple solutions, such as provision of rigid containers or ensuring waste is disposed of in bags. Our results contribute to a much-needed evidence base and demonstrate significant reduction in plastic debris emissions is achievable by deploying tangible and effective local SWM infrastructure and service interventions.
Research Article
Plastic pollution emission factors: How much escapes, how far it goes, and how to intervene in the solid waste management system
https://doi.org/10.21203/rs.3.rs-4850420/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Plastic pollution
expert elicitation
solid waste management
plastic emission factors
municipal solid waste
pollution mitigation
plastic debris
plastics treaty
We quantify the effectiveness of local solid waste management interventions for reducing plastic emissions using expert elicitation. We find significant emission mitigation is achievable by deploying simple infrastructure and service interventions.
Plastic pollution is a grand challenge of our times causing damage to ecosystems1, 2 and wider society.3, 4 The observed rapid accumulation of plastic in the natural environment is a consequence of increased growth in plastic production, insufficient solid waste management (SWM) systems and the materials’ inherent engineered longevity.5
Efforts to reduce plastic pollution have been strengthened by the recent ratification of a United Nations draft resolution to create an international legally binding instrument (ILBI: e.g., a treaty – hereafter, “Plastics Treaty”) to eliminate plastic pollution,6 in which waste management of plastics is highlighted as a core component.7 Specifically, the Plastics Treaty drafts call for environmentally sound waste management throughout the major SWM stages of: handling, collection, transportation, storage, recycling and final disposal; as well as the prohibition of waste management practices that may lead to the emission and release of hazardous substances.7
Local SWM practices have been shown to reduce plastic pollution.8 However, there is a lack of empirical data on the effectiveness of physical SWM infrastructure and practices at mitigating emission of plastic waste into the environment. We define plastic emission as “material that has moved from the managed or mismanaged systems (where waste is subject to a form of control, however basic – contained state), to the unmanaged system (the environment – uncontained state) with no control”.9
An evidence base on plastic waste emissions is crucial for successful implementation of the Plastics Treaty because it allows effective targeting of interventions and enables long-term consistent monitoring and reporting, such as with emission inventories.10 Emission data would also benefit plastic pollution quantification models which currently exhibit large uncertainty11 because they do not conceptualize emission processes and lack reliable SWM data.12 These models calculate emissions using “mismanaged waste” - a simplified proxy that does not discern between the SWM stages and the corresponding emission-associated phenomena.13–17
In the absence of experimentally measured emission factors, expert judgement can be obtained to cover knowledge gaps through a methodologically structured expert elicitation process. This type of approach systematically documents knowledge from experts on an uncertainty quantity in the form of probability distributions.18 It has been widely applied in numerous fields including public health,19, 20 water,21 mining,22 energy,23, 24 and plastic pollution,25 and is particularly suited to circumstances where there is little or no measured data, as is the case with plastic pollution emission factors.
Here, we address this fundamental understanding and data paucity by conducting an expert elicitation to ascertain first estimates (priori distributions) of macroplastic debris emission factors from typical SWM system arrangements and the magnitude of potential mitigation achievable by specific engineering and operational choices, associated with infrastructure and service modalities. Our results further conceptualize and detail the direct relationship between local waste management practices and plastic debris emissions, providing timely insights to support the on-going Plastic Treaty negotiations and abatement of plastic pollution via the prioritization of effective MSW systems.
We used the IDEA protocol to structure the expert elicitation approach,26 a summary of the step-by-step process is provided in Figure S1. Experts individually and anonymously provided their central estimate, plausible ranges and justification for each uncertain quantity. Results from this first round were visually summarized back to experts, including relaying any justifications provided. Experts were encouraged to discuss results and clarify different interpretations. They were provided the opportunity to individually adjust their responses considering the prior feedback, with the final responses of all experts combined using mathematic aggregation. Consensus was not sought.
2.1. Selection of variables to elicit
We elicited quantitative information from experts about three plastics system phenomena for which limited data exist:
Phenomenon 1: The relative effectiveness of waste management practices and infrastructure at reducing plastic debris emissions during each stage of the SWM system (hereafter referred to as “mitigation measures”).
Phenomenon 2: The expected proportion of plastic debris emitted into the environment from each stage of the SWM system given the presence of no mitigation measures (No Mitigation Measures Scenario) and all mitigation measures (All Mitigation Measure Scenario)
Phenomenon 3: The relative likelihood of plastic item movement once emitted into the natural or built environment.
Questions (Supporting Information (SI) – Section S3) were aligned with five formal waste system stages (stages A-E, Figure 1) in two plastic pollution models: the ISWA Plastic Pollution Calculator,27 and the Waste Flow Diagram.28 Questions on Phenomenon 1 and Phenomenon 2 were repeatedly asked for each of the five stages and Phenomenon 3 was aligned with the eleven categories (groupings) of plastic items modelled by the ISWA Plastic Pollution Calculator (SI Section S4). The mitigation measures of Phenomenon 1 were assessed independently, so that elicited responses reflected the effectiveness of a mitigation measure only if that measure was implemented in isolation. Independently quantifying mitigation measures effectiveness meant the number of questions could be both minimized and made relevant to any SWM system.
Waste that never enters the SWM system, for example, uncollected waste (either open burned or discarded as physical debris) and litter (isolated items dropped on the go), were excluded from the scope of our study because the entire fraction of these flows can be by default considered to be an emission (Figure 1). Only the macroplastic debris proportion of waste emissions were considered here (physical plastic items >5mm in one dimension) and any other materials (unburned fragments) or chemical compound emissions due to open burning of waste were also excluded.
2.2. Expert selection
We identified 19 experts, of which 18 completed the first elicitation round, and 16 completed both rounds. Each had demonstrable international expertise in the design or operation of solid waste management infrastructure. Additionally, experts featured at least one of the following: membership of an organization dealing with solid waste management and / or plastic pollution; authorship of academic publications related to land-based sources of macroplastic pollution; experience in solid waste management or plastic pollution assessment; or experience in practical solid waste management and / or street / environmental cleansing operations.
2.3. Evidence dossier and ready reckoner
As recommended by 18, an evidence dossier (SI – Section S2) was disseminated to experts prior to the elicitation. This provided background to the elicitation process; clarified key definitions; discussed common cognitive bias that may occur during the process along with procedures to control them; and provided a global overview of solid waste management characteristics to help experts consider all possible realities and avoid availability bias.
A reference mass of one metric ton of mixed municipal solid waste generated in an upper middle-income country was defined for each question to help experts visualize the scenario and on which they could relate their emission factors. Of this, we asked them to assume that 11% of the mass is plastic waste (110 kg). A “ready reckoner” was provided (SI – Ready Reckoner) to help experts translate their estimates between percentage, mass, volume, and plastic item counts.
2.4. Training workshop
Experts attended one of two identical online training workshops to accommodate different time zones. Participants were provided with an overview of the project and offered approaches on how to manage their own potential bias. Experts were then given two days to complete the questionnaire (SI – Section S3) in private.
2.5. Elicitation round one and data analysis
The questionnaire was prepared in a manner to aid in the avoidance of bias. For example, we scheduled the question about the movement of plastic item types (Phenomenon 3) first to encourage experts to consider the wide variety of plastic items when they answered subsequent questions on Phenomenon 1 and 2. Questions for eliciting probabilities of plastic emissions (Phenomenon 2) were structured according to the three-step IDEA protocol,26 which elicits the most likely value (mode) and maximum and minimum plausible values, defined as the range that would cover approximately 95% of cases. Eliciting plausible limits, rather than absolute maximum and minimum values, discourages experts from considering extreme cases that are memorable, but highly improbable. The questionnaire was structured to elicit the most likely value after the plausible limits in order to avoid anchoring estimates.29, 30
Each question had a space allocated for experts to add a justification for their response and, in the case of questions on the mitigation measures, comment on whether they considered any measures to be missing. Any mitigation measures highlighted as missing were presented to the group in the feedback workshop, with a group decision made on whether that mitigation measure should be added to the questionnaire for the second round.
Questionnaire responses were anonymized and corrected for basic errors. For example, experts occasionally provided ambiguous estimates such as a range in values or used greater and less than symbols instead of an exact estimate for the minimum, maximum or most-likely estimates. This ambiguity was corrected to a single value by taking the mid-point of the range and removing any greater or less than symbols. Responses for Phenomenon 1, which aimed to elicit relative effectiveness of mitigation measures (Figure 1), were normalized so that they summed to 100. Corrections were summarized and explained to participants in the subsequent feedback workshop and experts were asked to modify their responses in the second elicitation round to avoid such mistakes or ambiguity.
The median of all responses was calculated for the minimum, most-likely and maximum plausible values for questions on emission factors (Phenomenon 2). We used the median rather than any other type of location statistic (i.e., average) to mitigate the influence of potential extreme outlier estimates, offered by a minority of experts.31, 32 Where experts provided a justification for their responses, we anonymized them for presentation in the feedback workshop.
2.6. Feedback workshop
A feedback workshop was held online consisting of two sessions to accommodate multiple time zones. The workshop allowed clarification of any comments, queries and errors encountered during the first elicitation round. Responses were summarized back to experts using box and whisker plots, accompanied by the anonymized justifications (SI - Responses). In line with the IDEA framework26, we encouraged experts to discuss responses and accompanying justifications, before giving them an opportunity to anonymously modify their own responses.
2.7. Elicitation round two and data analysis
Sixteen experts provided responses for the second round of elicitation, 15 of which modified their original answers based on the discussion in the feedback workshop. The revised responses were subject to the same cleaning and mathematical aggregation, as carried out in the first elicitation round (Figure S1).
3.1 Effectiveness of mitigation measures
Waste storage for collection: A variety of methods are used to store waste whilst it is awaiting collection by the municipal authorities, ranging from storing it loose in piles on the ground, through to storage in dedicated enclosed containers.33 Experts scored the relative effectiveness (on a scale of 0 to 100, with 100 being the most effective score) of mitigation measures (Figure 1: A1 – A7) for reducing emission of plastic at the waste storage stage for two scenarios; one where a rigid container is not provided, and conversely, two, where a rigid container is provided.
The experts judged that the most effective means of reducing plastic emission during the waste storage stage where no rigid containers are provided is to “use rigid containers” (A1) (Figure 2a). This judgement was broadly agreed upon, with 12 out of the 16 experts scoring it as the single most effective intervention, all of whom gave it a score >90. If a rigid container is not provided, for example due to budget restraints or due to colored bags being preferred as often the case for source-separated curbside recycling schemes34, 35, “storing the waste in bags” (A2) is considered the second most effective mitigation measure with a score of 70 (mean). However, some experts raised concerns that bags are susceptible to being opened by animals, for instance, black vultures are reported to open bags in search of food on Brazilian beaches.36
For waste stored in a rigid container (Figure 2b), “sufficient capacity of the container” (A5) was considered most effective mitigation measure, implying that overflowing containers is an important emission source. This agrees with the findings of Wagner and Broaddus37 who studied litter generation following curbside collection of recycling in the city of Portland, USA and noted overflowing open-top containers are a common occurrence that have potential to generate significant volumes of litter. The following three most effective mitigation measures were “fully enclosed sides” (A6), “provision of a container lid” (A7) and “discarding the waste in bags within the rigid container” (A2), demonstrating that experts considered the level of containment of waste as important.
Loading of waste: Experts also assessed the effectiveness of mitigation measures for the loading stage (Figure 1, B1 – B6), where waste stored (either loose, bagged or containerized) is transferred to the refuse collection vehicle. Experts largely agreed (mean score = 84) that although the use of bags (B1) has limitations for long-term storage of waste due to the potential for them to be opened (e.g., by animals), they make the collection (loading) easier than would be the case for loose waste, and therefore substantially reduce the probability of plastic emission (Figure 3a). The care taken by loaders to handle waste without spilling it (B2) was also considered to be highly effective by most of the experts, with a mean score of 80. Although the remaining mitigation measures (B3 – B6) had lower mean scores, these scores still suggest relatively high effectiveness and therefore should not be interpreted as unimportant. Instead, the lower means scores were a result of greater variation in the opinions of experts, particularly for the mitigation measures of “Well operated collection system ensuring a very low number of missed collections” (B3), “Containers exchanged, rather than emptied” (B4), and “Waste easily transportable to collection vehicle” (B5). It was apparent from justifications provided by the experts that opinions differed as to whether the most important phase of loading the waste was the movement of the waste to the vehicle, the loading of that waste into the vehicle, or wider failings in the collection system (e.g., missed collections).
Waste transport: The transport stage (Figure 1, C1 – C8) reflects the manner by which waste is transported in the primary or secondary collection vehicle, and for which vehicle design varies substantially around the world.33 The “covering of waste” during transport (C1), for example by use of tarpaulin or netting, was considered to be the most effective mitigation method by half of the respondents (n =8), resulting in a mean score of 90 (Figure 3b). The next most effective mitigation measures were “waste is not piled higher than the sides of the vehicle” (C2) with a mean score of 80, and that the “vehicle has rigid sides that fully enclose the waste” (C3) with a mean score of 74. A common collection modality in the Global South is where waste is collected from households using manually powered vehicles (e.g., handcarts) before being transferred (typically on the side of roads) to larger motorized vehicles.33 To reflect this reality, an additional mitigation was added to Round 2 at the suggestion of experts: to have a “well-functioning interface between transport vehicles” (C6). There was wide disparity amongst the responses about how important this mitigation measure was, with relative effectiveness estimates scoring between 5-100. Responses on the use of “waste compaction” (C4) in waste collection vehicles (mean = 69) had a similar uncertainty with six experts ranking this as the most effective intervention (score of 100) and four experts scoring 30 or less.
Temporary storage and sorting of waste: Collected waste is often transported to transfer stations where it is temporarily stored before being transferred to its final destination (e.g., disposal sites or treatment plants) in larger vehicles. Occasionally, waste may be sorted to recover recyclables in transfer stations, or at a dedicated site (e.g., Material Recovery Facility – MRF).33 In either case, the mitigation measures for reducing plastic emission are similar, and are grouped here under the temporary storage and sorting stage (Figure 1, D1 – D7).
The most effective mitigation measure for preventing emissions from the temporary storage and sorting of waste is “locating the site indoors” (D1), scoring a mean value of 94 and unanimously scored highly by experts (Figure 4a). The “use of perimeter walls or fencing” (D3) also scored relatively highly (mean score of 68), but experts noted its effectiveness would depend on the height of the perimeter. Likewise, if sorting of recyclables is occurring, then disposing of reject material in the formal system (D2) had a mean effectiveness score of 72, with some experts assuming that practicing otherwise would result in open dumping of the rejects. The “use of bags to store waste” (D5) was rated as having a lower effectiveness (mean score of 51) compared to the earlier SWM stages. We propose the effectiveness of bags may therefore negatively correlate with the volume of waste managed at each stage of the SWM system.
Land disposal: Land disposal mitigation measures (Figure 1, E1 – E8) cover the various techniques that may reduce plastic debris emission across all forms of land disposal (e.g., from open dumpsites to engineered controlled landfills), regardless of whether the practice is solely implemented for the direct control of litter or as part of wider land disposal facility management.
Waste deposited at land disposal sites can be periodically covered by material such as soil to prevent windblown litter; reduce odors; reduce the risk of fire; and deter rodents and birds.38, 39 Experts suggested that “providing daily cover” (E1) is by far the most effective mitigation measure, giving it a mean score of 97 (Figure 4c). By contrast, “providing monthly cover” (E8) was given the lowest relative effectiveness score of 32, demonstrating that the frequency of cover is critical in reducing plastic emissions from land disposal.
Open uncontrolled burning is a form of plastic pollution9 with major negative impacts on human health and the environment.40 However, burning plastic is also a mitigation measure that reduces material available for physical debris emissions. For example, fires are often started deliberately on disposal sites to reduce the volume of waste or to recover metal.40, 41 The mitigation measure “open burning encouraged” (E7) was therefore presented to experts as a possible means of reducing plastic debris emissions. Microplastic emissions as a consequence of open burning were not considered because our scope only included macroplastic emissions (physical objects >5mm in size).
Experts disagreed on the effectiveness of open burning resulting in a wide range of answers (0-90) (Figure 4). Those that scored open burning as highly effective often justified this based on its ability to reduce total material available to be emitted, whereas those that ranked it low claimed burning results in emission itself therefore is not effective. Despite that the scope of the work was limited to debris emissions, as detailed in the training and feedback workshop, we speculate that some experts still considered open burning as an emission.
Additional operational control measures often implemented at land disposal sites include mechanical “compaction of waste” (E2), “designated discharge zones” (E4) for unloading of waste, “considering the wind speed and direction during waste discharge” (E5), and use of “portable litter screens” (E6) for capturing windblown litter.42 Each of these mitigation measures were scored as having mid-to-high relative effectiveness (mean scores of 75, 59, 54 and 53 respectively). Despite that these measures were not considered to be as effective as daily cover, they can be applied easily at low cost with the use of basic or repurposed machinery (e.g., agricultural tractors with attachments can be used at small or rural disposal sites to achieve basic levels of control).33 As such, our results suggest that effective mitigation of plastic emissions from open dumpsites does not require them to be upgraded to full controlled engineered landfills, but instead can be achieved by implementing practices that fall under basic levels of control, as defined by the UN-Habitat Waste Wise Cities Tool (WaCT).43 Such an incremental improvement of open dumpsites to basic control according to the WaCT ladder has also recently been advocated for by Wilson,44 especially when municipalities do not have the ability to afford or implement such measures.
3.2 Plastic debris emission factors
We define plastic emission factors as the percentage of plastic debris emitted into the environment from a SWM stage compared to total plastic waste at that stage.9 Expert estimates of plastic emission factors are shown in Figure 5 for the five main SWM stages and assuming the presence of no mitigation measures (No Mitigation Measure Scenario) and all mitigation measures (All Mitigation Measures Scenario).
All SWM stages were considered to have substantially higher plastic emission factors for the No Mitigation Measures Scenario compared to the All Mitigation Measures Scenario. For instance, the most-likely value for plastic emission factors during the waste storage stage was estimated by experts at 20% for the No Mitigation Measures Scenario compared to just 1% when all mitigation measures are present. Implementing mitigation measures at the waste storage stage would therefore expect to reduce plastic emissions into the environment from this stage by 95%. Similarly high reduction potentials were assessed for the other SWM stages, with land disposal having the lowest reduction potential at 80% based on the median most-likely emission factor value. This implies that implementing mitigations has the potential to reduce plastic emissions into the environment by at least an order of magnitude. In addition, the range between the upper and lower plausible limits was also much smaller for the All Mitigation Measures Scenario compared to the No Mitigation Measures Scenario. This difference would suggest that mitigations not only reduce the most likely emission of plastic, but also reduces overall potential for emission of plastic. These results suggest that solid waste management infrastructure and practice-based interventions are important for reducing plastic pollution debris and therefore are justified as a major consideration in any potential Plastic Treaty.
The plastic emission factors shown in Figure 5 allow a comparison of the magnitude of plastic debris emissions across each SWM stage. For a no mitigation scenario, the storage and loading stages have higher emissions factors than other SWM stages (for both most-likely values and plausible ranges). If correct, this would mean that for absolute emissions reduction, municipalities should prioritize mitigations at these two stages of solid waste handling. As discussed in Section 3.1, possible effective mitigations include: (i) ensuring households have a rigid storage container with high levels of containment, and (ii) encouraging households to dispose of their waste in plastic bags, because it limits dispersion of debris during both storage and loading.
The experts assessed that emission factors for the no mitigation scenario decrease as we progress through the SWM stages from storage for collection to disposal (Figure 5). For example, most likely emission factors show 20% of plastic is expected to be emitted from the storage stage compared to 10% at disposal. This decrease is more prominent for the upper plausible limits (50% at the storage stage, reducing to 20% at the disposal stage). Without any form of mitigation measure, both stages involve similar conditions in that waste is dumped in piles on the ground, therefore, the expected reduction in emissions must come about by some other means. We propose that the reductions observed are due to scale issues. As waste progresses along the SWM system, operations become more centralized and waste is aggregated. The aggregation of material results in fewer, larger piles of waste, with lower surface area to volume ratio exposed to the elements; limiting the amount of plastic emitted.9 By contrast, smaller piles of waste, such as those produced by a single household, would have a much larger surface exposed to the environment and are unlikely to be compacted. A higher proportion of this waste would therefore be expected to be emitted.
Emission factors were also predicted to reduce as waste progresses through the SWM system for the all mitigations scenario. With the exception of land disposal, estimated to have the highest emission factors of all stages. Expert justifications for these high values largely focused on the concept that small amounts of emission are unavoidable even in well-managed sites. It is unclear why experts considered these “unavoidable emissions” to be greater at the disposal stage than at previous stages, especially given the previously discussed aggregation issue. However, one possible explanation is that experts found quantifying accurate values for a small amount of emissions challenging. For instance, experts provided justifications stating that in the loading stage for an All Mitigation Measures Scenario, emissions would be only a small number of items maximum - “Emptying a well-managed wheelie bin of mixed plastic into a refuse truck, I would expect to see no more than one carrier bag or bottle top escape”. Despite this, and considering the scenario provided to experts was 110 kg of plastic for loading, the expert stated their most-likely estimate as 0.2% equating to 220 grams of plastic which at roughly 5 - 15 grams per item, would represent 15 – 44 items. This difficulty in estimating very small percentages was foreseen by the authors, hence the ready reckoner spreadsheet (SI – Ready Reckoner) was provided to experts to assist in converting item estimates to mass or percent estimates; however, it is unclear how many experts relied upon this tool.
The minimum and maximum plausible values in Figure 5 are not symmetric around the most likely value (as would be expected if a Gaussian distribution applied): the maximum plausible values are larger. If we were to plot these results as probability distributions (e.g., triangular distribution, SI – Section S5), they would be positively skewed. A positively skewed distribution implies that the larger extreme emission values occur less frequently than the lower extreme emission values, for example, only occurring during adverse weather events. However, these rare events may still contribute to the overall mass of plastic emitted given their high emission factor. This importance of extreme events is in agreement with findings relating to plastic transmission in the environment, whereby infrequent but extreme events, such as (flash) flooding, can lead to large movement of plastic waste.45
3.3 Comparison with existing approaches to emissions
Comparison of the emission factors in this work to those used in the literature is challenging because the majority of plastic pollution quantification methodologies13-17 rely on the concept of “mismanaged waste” to estimate emissions and do not practically distinguish between the SWM stages and practices. Some of the few notable exceptions use emission factors only for disposal sites and vary considerably (Table 1). For example, emission factors for the maximum amount of debris emitted from open dumpsites ranged between works from just 0.41% wt. to 20% wt. However, our results suggest that emissions are often higher at the collection and loading stages: we therefore recommend that these stages are given more consideration in future plastic pollution quantification models. Furthermore, the emission factors were either estimated by experts in an unstructured manner or based on unvalidated conceptual models. The emission factors defined and quantified here provide a more structured, accountable, and updatable basis for establishing emission factors, offering potential for wider standardization.
Table 1: Emission factors used in other models which are comparable to those quantified in the present study. Emission factors for out-of-scope parts of the solid waste management (SWM) system are not shown (e.g., littering or uncollected waste).
|
No. |
Name of methodology |
Emission factors |
Notes |
|
M1 |
Spatio-temporal quantification of plastic pollution origins and transportation (SPOT)9 |
Collection system emissions: 0.0001% - 4.0% Mismanagement of sorting rejects: 0 – 83% (as a percentage of sorting rejects, includes open burning) Uncontrolled disposal emissions (exc. open burning): 0.006% - 0.41% |
Emissions factors are based on conceptual models that aim to account important phenomena and available literature in emission estimates. The emission factor ranges shown here are those at the national level and relate to emission of MSW. The proportion of the MSW emission that is plastic varies for each emission source depending on composition data. |
|
M2 |
Waste Flow Diagram (WFD)28 |
Collection emissions: 0.1% – 10.8% Informal value chain emissions: 0.1% – 2.5% Formal sorting emissions: 0-100% of rejects Informal sorting emissions: 0-100% of rejects Transportation emissions: 0.001% – 0.5% Disposal emissions (with flooding): 3% – 81% Disposal emissions (no flooding): 0.00075% - 1% |
Emission factors assumed. A decision tree method was used to link emissions potential with on-the-ground observations. The values shown in this table are for the minimum and maximum possible values only and relate to plastic. |
|
M3 |
Plastic Waste Discharges from Rivers and Coastlines in Indonesia46, 47 |
Emissions from disposal sites: Sanitary landfill = 0% Controlled landfill = 2% (low), 3% (mid), and 5% (high) Formal dumpsite = 5% (low), 10% (mid) and 20% (high) |
Authors term emissions as “mismanaged waste available for wash-off”. It is unclear how term relates to the definition of emissions used in this work. Emission factors assumed. |
|
M4 |
Plastic Drawdown48 |
Emissions from disposal sites: 9% (urban waste) 20% (resort waste inc. emissions during transfer by boat) 9% (island rate based on 4% dumped and 5% blown off landfill) |
Emission factors applicable to the Maldives only and based on interviews with local stakeholders. Emission factors appear to include other parts of the SWM system such as transportation. |
|
M5 |
Evaluating scenarios towards zero plastic pollution49 |
Emission from dumpsites = 2% (rigid plastic), 16% (flexible or multi-material plastic) Post-collection emissions = 5% |
Emission factors assumed based on distance to water and reflecting item properties (e.g., likelihood of movement by wind). |
|
M6 |
Plastics in the global environment assessed through material flow analysis, degradation and environmental transportation50 |
Emissions from disposal sites: Sanitary landfill = 0.1% Controlled landfill = 0.5% Open dumpsites = 5% |
Emission factors assumed. Emission factors were also calculated for microplastics and plastics production although these are not listed in this table. Emissions were assigned to different compartments (e.g., residential, soil, water). |
Abbreviations: MSW = Municipal solid waste; SWM = Solid waste management
3.4 Transmission of plastic items
Expert judgements on the transmission of plastic items are defined here as time taken for an item to be displaced 100 m, with results shown in Figure 6. Experts allocated a score of 100 to the item expected to move 100 m first, with all other values assigned relative to this, considering comparative time. Plastic bags were ranked the most mobile item by 75% of the experts (mean score 95), with other plastic film the second most mobile with a mean score of 80. Experts agreed these flexible items were most susceptible to movement by wind, although also commented that they had a higher likelihood of becoming entangled (i.e., trapped in vegetation). Expanded polystyrene containers also scored high for mobility, with experts stating these too could be blown by wind, but also that they would float and move if exposed to surface runoff. This ability to float was raised as a key parameter for movement by many experts (n = 6), hence the relatively high scores for items such as drink bottles, PTT’s and single-use food service items. However, experts noted that it cannot be taken for granted that these items always float, as bottles or containers may be crushed, have part of their contents remaining inside, or not be air-/ water-tight (e.g., bottles missing lids). Each of these would result in changes in the item density (which is of relevance here rather than the plastic material density) and therefore reduced likelihood of floating.
The items with the lowest propensity to move included other dense plastic items, other small dense plastic items and sanitary products, with mean scores of 13.4, 23.5 and 37.1 respectively. It was suggested that other large dense plastic items are unlikely to move by wind or surface runoff, whilst sanitary products may absorb water (e.g., diapers), restricting their movement. These results highlight that plastic items exhibit large variability in both the mechanisms by which they can be transmitted, and the time required to do so. We therefore recommend that plastic pollution models account for item-level variability when predicting plastic emissions into the environment and transmission in it.
All opinions are subject to inherent bias,18 and although we implemented measures to control for this in our study, it is likely that some persisted. For example, to avoid availability bias (considering an event more probable if it is easily remembered), experts were provided with an evidence dossier to help them consider all global realities and were asked to quantify the plausible ranges of emission factors (covering 95% of cases) rather than absolute ranges (100% of cases). However, it is still likely that some experts scored the maximum plausible values by considering rare, but easily remembered extreme events. For instance, one justification stated: “The maximum would only happen during a perfect storm of bad events and would be unlikely”.
Other challenges included accounting for the uncertainty of experts’ estimates. Ideally, the 4-step IDEA protocol would have been followed, rather than the 3-step IDEA protocol used here.26 The 4-step protocol asks experts to quantify their uncertainty around elicited responses. However, we opted against this in an attempt to avoid fatigue from too many questions. Similarly, experts were asked to judge the effectiveness of mitigation measures assuming they were implemented independently to avoid burdening experts with too many questions relating to the independency of mitigation measures. However, many mitigation measures inherently display dependence (for example, if waste was stored in a rigid container, it would also be somewhat protected from animals), meaning experts could have found such confounding difficult to assess.
We anticipated that experts may estimate very small emission factors in some cases. For example, some experts provided justifications such as “only a small number of items would leak”. However, we found they often struggled to attribute a percentage that reflected their reasoning on an item level, despite being provided with a “ready reckoner” to facilitate this (SI – Ready Reckoner). In future studies, we would recommend that researchers perform this conversion and relay it back to the experts in the feedback workshop for consideration.
Despite these challenges, our results provide a scientific evidence base of plastic debris emission factors, especially in the current absence of any field measurements. These estimates provide timely insight for assessing the magnitude of plastic emissions from different parts of the SWM system and the effectiveness of interventions in mitigating these. For example, our results indicate that plastic emissions could be reduced by at least an order of magnitude if mitigation measures are targeted at the SWM system, with even simple cost-efficient mitigations, such as storing waste in bags rather than loose, effective.
Use of emissions factors for quantifying plastic pollution and its mitigation, is ultimately the main scientifically defensible approach and efforts should be intensified to cover the existing gap in understanding and data. We acknowledge that the estimates obtained here are based on experts’ understanding and experience alone, and as such are inherently uncertain. However, we suggest that the approach used here, whereby emission factors are defined and quantified by each SWM stage, provides much needed transparency, detail and structure to enable harmonized quantification of plastic pollution in the future. As a way ahead, we advocate for intensifying scientific efforts to quantify plastic debris emission factors by modeling and field / empirical data validation, so that the original values reported here can be compared and updated.
ASSOCIATED CONTENT
Supplementary Information: Summary illustration of the expert elicitation process, evidence dossier, questionnaire, table showing the categorization of plastic items, and figure showing triangular probability density functions of plastic debris emissions. (PDF)
Ready Reckoner: Toolkit provided to experts to aid in the conversion of percentages, mass, volume and item counts relating to plastic emissions. (Microsoft Excel)
Responses: Anonymized responses to the questionnaire for each expert. Responses are for the second round of the elicitation and have been cleaned and analyzed as discussed in the methods section. (Microsoft Excel)
DATA AVAILABILITY STATEMENT
All data associated with this publication are freely available in the associated content and available as a preprint in Research Square.
AUTHOR INFORMATION
Corresponding Author
*Corresponding author. Email: [email protected]
Author Contributions
The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.
Funding Sources
This research project was funded by the International Solid Waste Association (ISWA) under the auspices of the Marine Litter Task Force (2016-2022) and specifically the project “ISWA Plastic Pollution Calculator validation”.
ACKNOWLEDGMENTS
We would like to thank all experts who participated as independent experts in expert elicitation study for sharing their expertise and time. Participating experts, agreed to be acknowledged (in alphabetical order): Steffen Blume, Britta Denise Hardesty, Mohamed Kamal, David Lerpiniere, Antonis Mavropoulos, Clive Phillips, Arne Ragossnig, Aditi Ramola, Emenda Sembiring, Ed Stentiford, Elizabeth Tilley, Maria Tsakona, Mike Webster, Andrew Whiteman, Andrew Willcott, David Wilson, Imanol Zabaleta. Christian Zurbrügg. We also wish to thank Angeliki Savvantoglou and Nick Rigas for support on illustrations and graphic design. Special thanks to the ISWA Marine Litter Task Force (MLTF) for their support, and Professor John-Paul Gosling for initial advice on conducting expert elicitation studies.
ISWA
International Solid Waste Association
SPOT
Spatio–temporal quantification of plastic pollution origins and transportation
SWM
Solid waste management
WFD
Waste Flow Diagram
PTT
Pots, tubs and trays
- Derraik JGB (2002) The pollution of the marine environment by plastic debris: A review. Mar Pollut Bull 44(9):842–852. 10.1016/S0025-326X(02)00220-5
- Thompson RC, Moore CJ, Saal FSV, Swan SH (2009) Plastics, the environment and human health: Current consensus and future trends. Philos Trans R Soc B 364(1526):2153–2166. 10.1098/rstb.2009.0053
- Beaumont NJ, Aanesen M, Austen MC, Börger T, Clark JR, Cole M, Hooper T, Lindeque PK, Pascoe C, Wyles KJ (2019) Global ecological, social and economic impacts of marine plastic. Mar Pollut Bull 142:189–195. 10.1016/j.marpolbul.2019.03.022
- Prata JC, da Costa JP, Lopes I, Duarte AC, Rocha-Santos T (2020) Environmental exposure to microplastics: An overview on possible human health effects. Sci. Total Environ. 702. 10.1016/j.scitotenv.2019.134455
- Barnes DKA, Galgani F, Thompson RC, Barlaz M (2009) Accumulation and fragmentation of plastic debris in global environments. Philos Trans R Soc B 364(1526):1985–1998. 10.1098/rstb.2008.0205
- United Nations Environment Assembly of the United Nations Environment Programme (2022) Resolution 5/14 - End plastic pollution: towards an international legally binding instrument. UNEP: Nairobi,
- United Nations Environment Assembly of the United Nations Environment Programme (2023) Zero draft text of the international legally binding instrument on plastic pollution, including in the marine environment. UNEP: Nairobi,
- Willis K, Hardesty BD, Vince J, Wilcox C (2022) Local waste management successfully reduces coastal plastic pollution. One Earth 5(6):666–676. 10.1016/j.oneear.2022.05.008
- Cottom JW, Cook E, Velis CA (2024) A local-to-global emissions inventory of macroplastic pollution. Nat [In Press]. 10.1038/s41586-024-07758-6
- Zhu X, Rochman C (2022) Emissions Inventories of Plastic Pollution: A Critical Foundation of an International Agreement to Inform Targets and Quantify Progress. Environ Sci Technol 56(6):3309–3312. 10.1021/acs.est.2c01038
- Edelson M, Håbesland D, Traldi R (2021) Uncertainties in global estimates of plastic waste highlight the need for monitoring frameworks. Mar Pollut Bull 171:112720. 10.1016/j.marpolbul.2021.112720
- Velis CA, Cook E, Cottom J (2021) Waste management needs a data revolution – Is plastic pollution an opportunity? Waste Manag Res 39(9):1113–1115. 10.1177/0734242x211051199
- Lebreton LCM, Van Der Zwet J, Damsteeg JW, Slat B, Andrady A, Reisser J (2017) River plastic emissions to the world's oceans. Nat. Commun. 8. 10.1038/ncomms15611
- Schmidt C, Krauth T, Wagner S (2017) Export of Plastic Debris by Rivers into the Sea. Environ Sci Technol 51(21):12246–12253. 10.1021/acs.est.7b02368
- Jambeck JR, Geyer R, Wilcox C, Siegler TR, Perryman M, Andrady A, Narayan R, Law KL (2015) Plastic waste inputs from land into the ocean. Science 347(6223):768. 10.1126/science.1260352
- Borrelle SB, Ringma J, Law KL, Monnahan CC, Lebreton L, McGivern A, Murphy E, Jambeck J, Leonard GH, Hilleary MA, Eriksen M, Possingham HP, De Frond H, Gerber LR, Polidoro B, Tahir A, Bernard M, Mallos N, Barnes M, Rochman CM (2020) Predicted growth in plastic waste exceeds efforts to mitigate plastic pollution. Science 369(6510):1515–1518. 10.1126/science.aba3656
- Meijer LJJ, van Emmerik T, van der Ent R, Schmidt C, Lebreton L (2021) More than 1000 rivers account for 80% of global riverine plastic emissions into the ocean. Sci Adv 7(18). 10.1126/sciadv.aaz5803
- O’Hagan A (2019) Expert Knowledge Elicitation: Subjective but Scientific. Am Stat 73(sup1):69–81. 10.1080/00031305.2018.1518265
- Knol AB, de Hartog JJ, Boogaard H, Slottje P, van der Sluijs JP, Lebret E, Cassee FR, Wardekker JA, Ayres JG, Borm PJ, Brunekreef B, Donaldson K, Forastiere F, Holgate ST, Kreyling WG, Nemery B, Pekkanen J, Stone V, Wichmann HE, Hoek G (2009) Expert elicitation on ultrafine particles: likelihood of health effects and causal pathways. Part Fibre Toxicol 6(1):19. 10.1186/1743-8977-6-19
- Ferri CP, Prince M, Brayne C, Brodaty H, Fratiglioni L, Ganguli M, Hall K, Hasegawa K, Hendrie H, Huang Y, Jorm A, Mathers C, Menezes PR, Rimmer E, Scazufca M (2005) Global prevalence of dementia: A Delphi consensus study. Lancet 366(9503):2112–2117. 10.1016/S0140-6736(05)67889-0
- Walters JP, Javernick-Will AN (2015) Long-Term Functionality of Rural Water Services in Developing Countries: A System Dynamics Approach to Understanding the Dynamic Interaction of Factors. Eng Sci Technol 49(8):5035–5043. 10.1021/es505975h
- Kaikkonen L, Helle I, Kostamo K, Kuikka S, Törnroos A, Nygård H, Venesjärvi R, Uusitalo L (2021) Causal Approach to Determining the Environmental Risks of Seabed Mining. Eng Sci Technol 55(13):8502–8513. 10.1021/acs.est.1c01241
- Wiser R, Jenni K, Seel J, Baker E, Hand M, Lantz E, Smith A (2016) Expert elicitation survey on future wind energy costs. Nat Energy 1(10):16135. 10.1038/nenergy.2016.135
- Schmidt O, Gambhir A, Staffell I, Hawkes A, Nelson J, Few S (2017) Future cost and performance of water electrolysis: An expert elicitation study. Int J Hydrog Energy 42(52):30470–30492. 10.1016/j.ijhydene.2017.10.045
- Wilcox C, Mallos NJ, Leonard GH, Rodriguez A, Hardesty BD (2016) Using expert elicitation to estimate the impacts of plastic pollution on marine wildlife. Mar Policy 65:107–114. 10.1016/j.marpol.2015.10.014
- Hemming V, Burgman MA, Hanea AM, McBride MF, Wintle BC (2018) A practical guide to structured expert elicitation using the IDEA protocol. Methods Ecol Evol 9(1):169–180. 10.1111/2041-210X.12857
- United Nations Economic and Social Commission for Asia and the Pacific (UNESCAP) (2021) University of Leeds, Module 3: How to measure plastic waste. In Closing the Loop eLearining course, UNESCAP: https://www.unescap.org/projects/ctl/elearning, Principal investigator: Velis C.A., Main researcher: Cottom J.W., Based on the ISWA Plastic Pollution Calculator
- GIZ; University of Leeds, Zabaleta J, Zurbruegg I, Stretz C J. and, Blume S (2020) Eawag-Sandec; Wasteaware, Toolkit: Waste Flow Diagram (WFD): A rapid assessment tool for mapping waste flows and quantifying plastic leakage. Version 1.0. Principal Investigator: Velis C.A. Research team: Cottom. Switzerland and Leeds, UK,, Zurich
- Morgan MG (2014) Use (and abuse) of expert elicitation in support of decision making for public policy. Proc. Natl. Acad. Sci. 111 (20), 7176–7184. 10.1073/pnas.1319946111
- Tversky A, Kahneman D (1974) Judgment under Uncertainty: Heuristics and Biases. Science 185(4157):1124–1131. 10.1126/science.185.4157.1124
- Hora SC, Fransen BR, Hawkins N, Susel IJDA (2013) Median aggregation of distribution functions. Decis Anal 10(4):279–291. 10.1287/deca.2013.0282
- de Vries H, van de Wal RSW (2015) How to interpret expert judgment assessments of 21st century sea-level rise. Clim Change 130(2):87–100. 10.1007/s10584-015-1346-x
- Coffey M, Coad A (2010) Collection of Municipal Solid Waste in Developing Countries. United Nations Human Settlements Programme (UN-Habitat): Malta,
- Sörme L, Voxberg E, Rosenlund J, Jensen S, Augustsson A (2019) Coloured Plastic Bags for Kerbside Collection of Waste from Households—To Improve Waste Recycling. Recycling 4(2):20
- Workentin J, De Sousa JM, Mensink PJ (2022) The use of recycling bags: An overview of collection policies and a spatial assessment of bag use. Environ Chall 8:100535. 10.1016/j.envc.2022.100535
- Sazima I (2007) From carrion-eaters to bathers’ bags plunderers: how Black Vultures (Coragyps atratus) could have found that plastic bags may contain food. Rev Bras Ornitol 15(4):617–620
- Wagner TP, Broaddus N (2016) The generation and cost of litter resulting from the curbside collection of recycling. Waste Manag 50:3–9. 10.1016/j.wasman.2016.02.004
- Daniel DE (2012) Geotechnical Practice for Waste Disposal. Springer US
- Agency EP (2014) Guidance Note on Daily and Intermediate Cover at Landfills. Environmental Protection Agency: Ireland,
- Velis CA, Cook E (2021) Mismanagement of Plastic Waste through Open Burning with Emphasis on the Global South: A Systematic Review of Risks to Occupational and Public Health. Environ Sci Technol 55(11):7186–7207. 10.1021/acs.est.0c08536
- Cook E, Velis C (2021) Global review on safer end of engineered life. Royal Academy of Engineering
- International Solid Waste Association (2019) Landfill Operational Guidelines 3rd Edition.; International Solid Waste Association (ISWA)
- UN-Habitat Waste Wise Cities Tool: Step by Step Guide to Assess a City's Municipal Solid Waste Management Performance through SDG indicator 11.6.1 Monitoring; (2020)
- Wilson DC Learning from the past to plan for the future: An historical review of the evolution of waste and resource management 1970–2020 and reflections on priorities 2020–2030 – The perspective of an involved witness. Waste Manag Res 0 (0), 0734242X231178025. 10.1177/0734242x231178025
- van Emmerik THM, Frings RM, Schreyers LJ, Hauk R, de Lange SI, Mellink YAM (2023) River plastic transport and deposition amplified by extreme flood. Nat Water 1(6):514–522. 10.1038/s44221-023-00092-7
- Veiga JM, van Veen B, Buckman L, van Gils J, Wuriyandoko DT, van der Sluys C, Philp K, Acharya A (2023) Assessing Plastic Waste Discharges into the Sea in Indonesia: An Integrated High-Resolution Modeling Approach That Accounts for Hydrology and Local Waste Handling Practices. Water 15(6):1143. 10.3390/w15061143
- Bank W (2021) Plastic Waste Discharges from Rivers and Coastlines in Indonesia. World Bank
- Royle J, Jack B, Parris H, Elliott T, Castillo AC, Kalawana S, Nashfa H, Woodall LC (2022) Plastic Drawdown: A rapid assessment tool for developing national responses to plastic pollution when data availability is limited, as demonstrated in the Maldives. Glob. Environ. Change 72. 10.1016/j.gloenvcha.2021.102442
- Lau WWY, Shiran Y, Bailey RM, Cook E, Stuchtey MR, Koskella J, Velis CA, Godfrey L, Boucher J, Murphy MB, Thompson RC, Jankowska E, Castillo Castillo A, Pilditch TD, Dixon B, Koerselman L, Kosior E, Favoino E, Gutberlet J, Baulch S, Atreya ME, Fischer D, He KK, Petit MM, Sumaila UR, Neil E, Bernhofen MV, Lawrence K, Palardy JE (2020) Evaluating scenarios toward zero plastic pollution. Science 369(6510):1455–1461. 10.1126/science.aba9475
- Schwarz AE, Lensen SMC, Langeveld E, Parker LA, Urbanus JH (2023) Plastics in the global environment assessed through material flow analysis, degradation and environmental transportation. Sci Total Environ 875:162644. 10.1016/j.scitotenv.2023.162644
The authors declare potential competing interests as follows: Dr Costas Velis, Dr Josh Cottom and Ed Cook are actively involved in the waste, resources and circular economy sector in research, advising, scholarship and consultancy capacities. Dr Costas Velis reports financial support was provided by ISWA. Dr Costas Velis reports a relationship with International Solid Waste Association (ISWA) that includes: consulting or advisory, funding grants, speaking and lecture fees, and travel reimbursement. Dr Costas Velis reports financial support provided by UKRI, GCRF, NERC, ESRC, BBRSC, Royal Academy of Engineering, British Council, Innovate UK, EC H2020, World Bank Group, OECD, GIZ, UN-Habitat, UNESCAP, UNOPS, The PEW Charitable Trusts, IGES, ISWA, Grid-Arendal, Swedish EPA and SYSTEMIQ. A statement of ethics approval - Research ethics for the expert elicitation study was approved by the Faculty of Engineering Physical Science (EPS) and Research Ethics (EPS/FREC) Committee, at the University of Leeds, UK. A statement on participant consent - Participants (experts) consented to participate offering expert opinion in the study; and also separately agreed on having their names included in the acknowledgements of the publication.
- SupportingInformationCottomCookVelis2024PlasticPollutionEmission.pdf
Supporting Information Cottom_Cook_Velis 2024 Plastic Pollution Emission